Byrne - Learning & Memory PDF

Learning
& Memory
SECOND EDITION
EDITOR IN CHIEF
John H. Byrne
University of Texas Medical School at Houston
ASSOCIATE EDITORS
Howard Eichenbaum
Boston University Laboratory of Cognitive Neurobiology
Henry L. Roediger, III

Washington University Department of Psychology
Richard F. Thompson
University of Southern California Program in Neuroscience
MACMILLAN REFERENCE USA

Frank Menchaca, Vice President
Jill Lectka, Director of Publishing Operations
MACMILLAN PSYCHOLOGY
REFERENCE SERIES
The Macmillan Psychology Reference Series is an on-going collec-
tion of single-volume overviews of the disciplines of psychology.
Each volume in the series presents a comprehensive and accessible
snapshot of a psychology field for general readers, high school sen-
iors, and college freshmen. Child Development (2001), the first vol-
ume in the series, examines the contemporary issues in the field of
child and adolescent development. Each work will feature the latest
research in the field along with numerous real-world applications
and examples.
Learning
& Memory
SECOND EDITION
John H. Byrne, Editor in Chief

Learning and Memory, Second Edition
2003 by Macmillan Reference USA. ALL RIGHTS RESERVED Cover images reproduced by permission of
Macmillan Reference USA is an imprint of No part of this work covered by the copy- William T. Greenough and James D. Churchill
The Gale Group, Inc., a division of Thomson right hereon may be reproduced or used in (synapses), Photodisc (MRI profile), and Corbis
Learning, Inc. any form or by any meansgraphic, electron- (brain).
ic, or mechanical, including photocopying,
Macmillan Reference USA and Thomson recording, taping, Web distribution, or infor- Since this page cannot legibly accommo-
Learning are trademarks used herein under mation storage retrieval systemswithout date all copyright notices, the acknowledg-
license. the written permission of the publisher. ments constitute an extension of the copy-
right notice.
For more information, contact For permission to use material from this
Macmillan Reference USA product, submit your request via Web at
300 Park Avenue South http://www.gale-edit.com/permissions, or you
New York, NY 10010 may download our Permissions Request form
Or you can visit our Internet site at and submit your request by fax or mail to:
http://www.gale.com
Permissions Department
The Gale Group, Inc.
27500 Drake Rd.
Farmington Hills, MI 48331-3535
Permissions Hotline:
248-699-8006 or 800-877-4253, ext. 8006
Fax: 248-699-8074 or 800-762-4058
LIBRARY OF CONGRESS CATALOG-IN-PUBLICATION DATA
Learning & Memory / edited by John H. Byrne. - - 2nd ed.

p. cm. - - (Macmillan psychology reference series)
Previous ed. published as: Encyclopedia of learning and memory. New
York : Macmillan, c1992.
ISBN 0-02-865619-9
1. Learning, Psychology of- -Encyclopedias. 2. Memory- -
Encyclopedias. 3. Learning in animals- -Encyclopedias. 4. Animal
memory- -Encyclopedias. 5. Neuropsychology- -Encyclopedias. I. Title:
Learning and memory. II. Byrne, John H. III. Series.
BF318 .E53 2002

153 . 103- -dc21 2002008357
Printed in the United States of America

10 9 8 7 6 5 4 3 2 1
CONTENTS
Preface........................................................................................vii
List of Articles ..............................................................................ix
List of Contributors.......................................................................xv
Learning and Memory, Second Edition ..........................................1

Index........................................................................................677
EDITORIAL AND PRODUCTION STAFF
PROJECT EDITOR
Ken Wachsberger
EDITORIAL
Denise Evans
Gretchen Gordon
Bill Kaufman
Gina Misiroglu
Patricia Onorato
Angela Pilchak
Gregory Teague
IMAGING AND MULTIMEDIA

Dean Dauphinais
Mary Grimes
Lezlie Light
PRODUCT DESIGN
Cindy Baldwin
PERMISSIONS
Lori Hines
COMPOSITION
Evi Seoud
MANUFACTURING
Rita Wimberley
INDEXER
Linda K. Fetters
vi
PREFACE
Learning refers to the process of acquiring new infor- Despite the great progress that has been made,
mation, whereas memory refers to the retention of that many questions remain. Why will we remember exact-
information so that it can be retrieved at a later time. ly where we were on September 11, 2001 for the rest
The topics of learning and memory have intrigued of our lives, but we cannot remember what we had for
philosophers and writers for centuries, and for good lunch on this day two weeks ago? Why are we unable
reason. Learning and memory are so central to our to recall events from our infancy? Are there any drugs
daily lives that disruption in these functions can inter- that can improve our memory? What is Alzheimers
rupt our most routine activities. For example, consid- disease and how can it be treated? Learning and
er the extraordinary memory processes that must take Memory, Second Edition contains articles on these key
place for you to successfully drive from home to work. aspects of memory. In general, its goal is to provide
When you wake up, you need to recall whether or not readers with a comprehensive overview of key topics
that day is a workday. When you enter your car, you in learning and memory through brief articles written
need to remember the best route to work and any traf- by selected experts in the field.
fic information that you may have heard to expedite All of the entries are original contributions from
your arrival. Next, you have to recall how to drive your scholars and researchers, written for a readership of
car and the complex rules of the road. When you students, teachers, journalists, and members of the
arrive at work, you need to remember the names of educated public. The articles range in length from
your colleagues and draw upon a memory for job- 800 to 3,000 words. Entries are arranged alphabetical-
related vocabulary and common past experiences. ly; each is accompanied by a bibliography listing sug-
Most importantly, you need to remember the skills gestions for further reading. The entries are also
that allow you to perform your job. linked by a comprehensive set of cross-references that
The centrality of learning and memory to our appear at the end of many of the entries. For exam-
daily lives has led to intense analysis by psychologists ple, in the entry on infantile amnesia, one finds cross-
and neurobiologists for the past century, and it will references to the following entries: CHILDHOOD,
undoubtedly remain at the forefront of research DEVELOPMENT OF MEMORY IN; CODING PROCESSES:
throughout this new century as well. Learning and ORGANIZATION OF MEMORY; EPISODIC MEMORY; EXPERTS
memory systems are vast and diverse, yet scientists MEMORIES; GUIDE TO THE ANATOMY OF THE BRAIN;
have determined that memory can be divided into NATURAL SETTINGS, MEMORY IN; PROSE RETENTION.
two major types: memory for skills and habits, and In addition, blind entries facilitate access to main
memory for facts and events. Moreover, significant articles (for example, Drugs. See: COGNITIVE
progress has been made in understanding the parts of ENHANCERS; DRUGS AND MEMORY; ELECTROCONVULSIVE
the brain involved in learning and memory as well as THERAPY AND MEMORY LOSS; PHARMACOLOGICAL
in acquiring a basic understanding of the genes, pro- TREATMENT OF MEMORY DEFICITS.). These blind
teins, and signaling molecules that mediate memory entries are arranged alphabetically throughout
acquisition and storage. Learning and Memory, Second Learning and Memory and provide alternate access
Edition contains articles that discuss these findings. It points to direct the reader to appropriate articles An
also contains biographical sketches of some of the key additional feature is the use of guideposts or series
individuals, now deceased, who have contributed to introductions. Where there is a group of related arti-
the current understanding of learning and memory. cles on a single topic, a short guidepost is available to
vii
viii PREFACE
orient the reader to the topic. For example, for the tion would not have been possible without the
area of invertebrate learning, a short introduction tremendous work of the current Editorial Board, who
presents an overview of the six entries that follow. identified the topics and their authors, and reviewed
This second edition of Learning and Memory builds each contribution. Special thanks also go to Jill
upon the success of the first, which was called Lectka, Director of Publishing Operations at
Encyclopedia of Learning and Memory when it was pub- Macmillan Reference USA, and Project Editor Ken
lished by Macmillan in 1992. We are indebted to Larry Wachsberger for supporting the concept of a second
Squire, the previous editor, for establishing the over- edition as well as ensuring that the production sched-
all direction and organization of the Encyclopedia, ule was maintained.
which this second edition maintains. This second edi- John H. Byrne
LIST OF ARTICLES
Active and Passive Avoidance DEVELOPMENT OF PROCESSES LEVELS OF PROCESSING

Learning: Behavioral Phenomena UNDERLYING LEARNING Robert S. Lockhart
F. Robert Brush Thomas J. Carew ORGANIZATION OF MEMORY
Activity-Dependent Regulation of MOLECULAR BASIS OF LONG-TERM Roger W. Schvaneveldt

SENSITIZATION
Neurotransmitter Synthesis Cognitive Enhancers
Jack C. Waymire Gregg A. Phares
Gregor y M. Rose
John H. Byrne
Aging and Memory in Animals Collective Memory
Diana S. Woodruff-Pak Aristotle (384322 B.C.E.)
James V. Wertsch
Patricia Smith Churchland
Aging and Memory in Humans Comparative Cognition
Fergus I. M. Craik Attention and Memory Anthony A. Wright
Harold Pashler
Algorithms, Learning Concepts and Categories,
Andrew G. Barto Autobiographical Memory Learning of
Martin A. Conway Edward Wisniewski
Alzheimers Disease
BEHAVIORAL ASPECTS Bartlett, Frederic (18861969) Conditioning, Cellular and
Marilyn S. Albert David J. Murray Network Schemes for Higher-
HUMAN DISEASE AND THE GENETICALLY Behaviorism Order Features of
ENGINEERED ANIMAL MODELS Philip N. Hineline Dean V. Buonomano
Philip C. Wong
Behavior Therapy Conditioning, Classical and
Donald L. Price
Stanley J. Rachman Instrumental
Amnesia, Functional I. Gormezano
John F. Kihlstrom Birdsong Learning
Gregor y F. Ball Declarative Memory
Daniel L. Schacter
Jacques Balthazart Neal Cohen
Amnesia, Infantile
Children, Development of Dj Vu
Harlene Hayne
Memory in Alan S. Brown
Amnesia, Organic
Robert V. Kail Dementia
Kelly Sullivan Giovanello
Meghan Saweikis Cynthia A. Munro
Mieke Verfaellie
Classical Conditioning: Discrimination and
Amnesia, Transient Global
Behavioral Phenomena Generalization
Mark Kritchevsky
John T. Green John Moore
Aplysia Joseph E. Steinmetz
Distributed Practice Effects
CLASSICAL CONDITIONING AND
Coding Processes Robert L. Greene
OPERANT CONDITIONING
IMAGERY
Fred D. Lorenzetti Drugs and Memory
John H. Byrne Marc Marschark Gregor y M. Rose
ix
x LIST OF ARTICLES
Early Experience and Learning Glutamate Receptors and Their Hunter, Walter S. (18891954)
Seymour Levine Characterization Norman E. Spear
Deborah Suchecki M. N. Waxham
Hypnosis and Memory
Ebbinghaus, Hermann Guide to the Anatomy of the John F. Kihlstrom
(18501909) Brain
Implicit Memory
Endel Tulving OVERVIEW
Neil W. Mulligan
Yuri Koutcherov
Eccles, John (19031997)
George Paxinos Imprinting
Per O. Andersen
AMYGDALA Johan J. Bolhuis
Eidetic Imagery Alexander J. McDonald G. Horn
Robert G. Crowder BASAL FOREBRAIN Individual Differences in
Electroconvulsive Therapy and M-Marsel Mesulam Learning and Memory
Memory Loss BASAL GANGLIA Colin M. MacLeod
Larr y R. Squire Ann M. Graybiel
Infancy, Memory in
CEREBELLUM
Emotion, Mood, and Memory Andrew N. Meltzoff
Paula Hertel Michael Mauk
Leslie J. Carver
CEREBRAL CORTEX
Episodic Memory Helen Barbas Insect Learning
Endel Tulving Stewart Hendr y Brian H. Smith
Evolution and Learning HIPPOCAMPUS AND Charles I. Abramson
David F. Sherr y PARAHIPPOCAMPAL REGION
Intelligence and Memory
Menno P. Witter Anna T. Cianciolo
Experts Memories
NEURON Robert J. Sternberg
K. Anders Ericsson
Peter Somogyi
False Memories Interference and Forgetting
OLFACTORY CORTEX
Kathleen B. McDermott Robert A. Bjork
Michael E. Hasselmo
Jason C. K. Chan PERIRHINAL CORTEX AND Invertebrate Learning
Food Aversion and Preference ASSOCIATED CORTICAL AREAS ASSOCIATIVE LEARNING AND
Learning in Humans Rebecca D. Burwell MEMORY PROCESSING IN BEES
Bennett G. Galef, Jr. SYNAPSE Randolf Menzel
Paul Rozin Kristen M. Harris ASSOCIATIVE LEARNING IN
Guthrie, Edwin R. (18861959) HERMISSENDA

Foraging
Sara J. Shettleworth Ernest R. Hilgard Terr y J. Crow
ASSOCIATIVE LEARNING IN LIMAX
Forgetting Habituation and Sensitization in
Alan Gelperin
Ian Neath Vertebrates
C. ELEGANS
Michael Davis
Freud, Sigmund (18561939) M. David Egger Stephan Steidl
Matthew Hugh Erdelyi Catharine H. Rankin
Harlow, Harry F. (19051981)
Frontal Lobes and Episodic HABITUATION AND SENSITIZATION
Br yan E. Kolb
Memory IN TRITONIA
Morris Moscovitch Head Injury William N. Frost

Gordon Winocur Gerri Hanten NEUROGENETICS OF MEMORY IN
Harvey Levin DROSOPHILA
Genetic Substrates of Memory
Hebb, Donald (19041985) Tim Tully
AMYGDALA
Peter M. Milner James, William (18421910)
Fred J. Helmstetter
CEREBELLUM Hormones and Memory Sheldon White
Shaowen Bao James L. McGaugh Kamins Blocking Effect:
Lu Chen Benno Roozendaal
Neuronal Substrates
HIPPOCAMPUS Hull, Clark L. (18841952) Jeansok J. Kim
Alcino J. Silva Abram Amsel Mark G. Baxter
LIST OF ARTICLES xi
Knowledge Systems and Material- Mathematical Learning Theory Multiple-Memory Systems

Specific Memory Deficits W. K. Estes Mark G. Packard
Elkhonon Goldberg
McGeoch, John A. (18971942) Natural Settings, Memory in
William B. Barr
Robert G. Crowder Ulric Neisser
Konorski, Jerzy (19031973)
Measurement of Memory Neocortical Plasticity
Anthony Dickinson
Robert S. Lockhart
ADULT VISUAL CORTEX
Language Learning: Humans
Memory Consolidation: ADAPTATION AND
Peter W. Culicover
Molecular and Cellular Processes REORGANIZATION
Language Learning: Nonhuman Alcino J. Silva Daniel J. Felleman
Primates
Memory Consolidation: AUDITORY CORTEX
Duane M. Rumbaugh
Prolonged Process of Norman M. Weinberger
E. Sue Savage-Rumbaugh
Reorganization DEVELOPMENT OF THE VISUAL
Jared P. Taglialatela
Larr y R. Squire SYSTEM
Lashley, Karl (18901958) Paul G. Shinkman
Memory Search
Richard F. Thompson
Barbara Anne Dosher MOTOR CORTEX
Learned Helplessness Brian McElree Jeffrey A. Kleim
Steven E. Maier
Memory Span SOMATOSENSORY CORTEX
Learning Disabilities Michael J. Watkins Peter W. Hickmott
R. Holly Fitch
Mental Retardation (Intellectual Neural Computation
Learning Theory: A History Disabilities) APPROACHES TO LEARNING
Robert C. Bolles Robert M. Hodapp Terrence J. Sejnowski
Learning Theory: Current Status Metacognition about Memory CEREBELLUM
Steve Maren Thomas O. Nelson Masao Ito
Petra Scheck HIPPOCAMPUS
Localization of Memory Traces
David J. Krupa Metaplasticity Bruce L. McNaughton
Thomas J. Carew NEOCORTEX
Long-Term Depression in the
Thomas M. Fischer Leif H. Finkel
Cerebellum, Hippocampus, and
Neocortex Migration, Navigation, and OLFACTORY CORTEX AS A MODEL
Masao Ito Homing FOR TELENCEPHALIC PROCESSING
Verner P. Bingman Richard H. Granger, Jr.
Long-Term Potentiation
OVERVIEW: COOPERATIVITY AND Mnemonic Devices Neural Substrates of Avoidance
ASSOCIATIVITY Mark A. McDaniel Learning
Bengt Gustafsson Mnemonists Michael Gabriel
Eric Hanse Charles P. Thompson Neural Substrates of Classical
AMYGDALA
Modality Effects Conditioning
Thomas H.Brown
Derick H. Lindquist John M. Gardiner CARDIOVASCULAR RESPONSES
Nelson Cowan Donald A. Powell
BEHAVIORAL ROLES
Howard Eichenbaum Morphological Basis of Learning DISCRETE BEHAVIORAL RESPONSES
MAINTENANCE and Memory Joseph E. Steinmetz

Michel Baudr y INVERTEBRATES FEAR CONDITIONING, FREEZING
Gar y Lynch Len Clear y Bill P. Godsil
SIGNAL TRANSDUCTION VERTEBRATES Michael S. Fanselow
MECHANISMS AND EARLY EVENTS William T. Greenough FEAR-POTENTIATED STARTLE
J. David Sweatt James D. Churchill Michael Davis
Lorenz, Konrad (19031989) Motor Skill Learning Neural Substrates of Emotional
David F. Sherr y James C. Houk Memory
Luria, A .R. (19021977) Mller, Georg Elias (18501934) Glenn E. Schafe
Elkhonon Goldberg Hilde A.E. Lechner Joseph E. LeDoux
xii LIST OF ARTICLES
Neurogenesis Prefrontal Cortex and Memory in Savant Syndrome

Gerd Kempermann Primates Darold A. Treffert
Fred H. Gage Joaquin M. Fuster
Schizophrenia and Memory
Neuroimaging Primates, Visual Attention in Stephan Heckers
Steven E. Petersen Bharathi Jagadeesh Anthony P. Weiss
Neurotransmitter Systems and Primates, Visual Perception and School Learning
Memory Memory in Nonhuman M. C. Wittrock
Michel Baudr y Elisabeth A. Murray
Joel L. Davis Second Messenger Systems
Procedural Learning James H. Schwartz
Object Concept, Development of ANIMALS
Jeanne L. Shinskey Semantic Memory
Mark G. Packard
COGNITIVE EFFECTS
Observational Learning HUMANS

Indre V. Viskontas Claude G. Cech
Albert Bandura
Barbara J. Knowlton NEUROBIOLOGICAL PERSPECTIVE
Olds, James (19221973) Sharon L. Thompson-Schill
John F. Disterhoft Prose Retention
Mark A. McDaniel Semon, Richard (18591918)
Olton, David (19431994) Daniel L. Schacter
Matthew L. Shapiro Protein Synthesis in Long-Term
Memory in Vertebrates Sensory Memory
Operant Behavior
Steven P. R. Rose Nelson Cowan
W. K. Honig
Brent Alsop Ramn y Cajal, Santiago Serial Organization
(18521934) Alice F. Healy
Oral Traditions
Larr y W. Swanson
David C. Rubin Sex Differences in Learning
Reconstructive Memory Jocelyne B. Bachevalier
Orienting Reflex Habituation
Daniel M. Bernstein
E. N. Sokolov Skinner, B. F. (19041990)
Elizabeth F. Loftus
Oscillations, Synchrony, and Kurt Salzinger
Neuronal Codes Rehabilitation of Memory
Disorders Sleep and Memory Consolidation
Wolf Singer Robert Stickgold
Elizabeth L. Glisky
Parallel Distributed Processing Social Memory Processes
Models of Memory Reinforcement
Andrew G. Barto Barbara H. Basden
James L. McClelland
Reinforcement or Reward in Sometimes Opponent Process
Passive (Inhibitory) Avoidance,
Learning (SOP) Model, in Conditioning
Fear Learning
ANATOMICAL SUBSTRATES
Susan E. Brandon
Almira Vazdarjanova
Richard H. Thompson Allan R. Wagner
Pavlov, Ivan (18491936)
CEREBELLUM Source Monitoring
John J. Furedy
Rodney A. Swain Marcia K. Johnson
Pharmacological Treatment of ELECTRICAL SELF-STIMULATION, Karen J. Mitchell
Memory Deficits BRAIN
Gregor y M. Rose Spatial Learning: Animals
Andreas Arvanitogiannis
Thomas S. Collett
Phobias STRIATUM
Stanley J. Rachman Wolfram Schultz Spatial Memory
Timothy P. McNamara
Piaget, Jean (18961980) Repetition and Learning
Howard E. Gruber Robert L. Greene Spence, Kenneth (19071967)
Howard H. Kendler
Place Cells Retrieval Processes in Memory
Edvard I. Moser Henr y L. Roediger III Spinal Plasticity
Michelle L. Meade Michael M. Patterson
Place versus Response Learning
Revisited in the Brain Ribot, Thodule (18391916) Stress and Memory
Mark G. Packard Jean-Louis Signoret Tracey J. Shors
LIST OF ARTICLES xiii
Taste Aversion and Preference Underwood, Benton (19151994) Working Memory

Learning in Animals Eugene B. Zechmeister ANIMALS
Diego E. Berman Anthony A. Wright
Vestibulo-Ocular Reflex (VOR)
Thorndike, Edward (18741949) Plasticity Humans
Gilbert Gottlieb Masao Ito Todd S. Braver
Tip-of-the-Tongue Phenomenon
Janet M. Duchek Visual Memory, Brightness and
Jessica M. Logan Flux in
David Lavond
Tolman, Edward C. (18861959)
Mark R. Rosenzweig Watson, John B. (18781958)
Donald A. Riley Eliot Hearst
LIST OF CONTRIBUTORS
Charles I. Abramson Shaowen Bao Diego E. Berman

Oklahoma State University University of California, San Weizmann Institute of Science,
Insect Learning Francisco Israel
Genetic Substrates of Taste Aversion and
Marilyn S. Albert
Memory: Cerebellum Preference Learning in
Harvard Medical School
Animals
Alzheimers Disease: Helen Barbas
Behavioral Aspects Daniel M. Bernstein
Boston University
Brent Alsop University of Washington, Seattle
Guide to the Anatomy of the
University of Otago, New Reconstructive Memory
Brain: Cerebral Cortex
Zealand Verner P. Bingman
Operant Behavior William B. Barr
Bowling Green State University
New York University School of
Abram Amsel Migration, Navigation, and
Medicine Homing
University of Texas Knowledge Systems and
Hull, Clark L. (18841952) Robert A. Bjork
Material-Specific Memory
Per O. Andersen Deficits UCLA
University of Oslo, Norway Interference and Forgetting
Eccles, John (19031997) Andrew G. Barto
Johan J. Bolhuis
University of Massachussetts
Andreas Arvanitogiannis Utrecht University, The
Algorithms, Learning
Concordia University, Montreal Netherlands
Reinforcement
Reinforcement or Reward in Imprinting
Learning: Electrical Self- Barbara H. Basden Robert C. Bolles
Stimulation, Brain California State University, University of Washington (ret.)
Fresno Learning Theory: A History
Jocelyne B. Bachevalier
University of Texas Medical Social Memory Processes
Susan E. Brandon
School at Houston Michel Baudry Yale University and American
Sex Differences in Learning University of Southern Psychological Association
Gregory F. Ball California Sometimes Opponent
Johns Hopkins University Long-Term Potentiation: Process (SOP) Model, in
Birdsong Learning Maintenance Conditioning
Jacques Balthazart Neurotransmitter Systems Todd S. Braver
University of Lige, Belgium and Memory Washington University, St. Louis
Birdsong Learning Mark G. Baxter Working Memory: Humans
Albert Bandura Harvard University Alan S. Brown
Stanford University Kamins Blocking Effect: Southern Methodist University
Observational Learning Neuronal Substrates Dj Vu
xv
xvi LIST OF CONTRIBUTORS
Thomas H. Brown Genetic Substrates of McGeoch, John A.

Yale University Memory: Cerebellum (18971942)
Long-Term Potentiation:
James D. Churchill Peter W. Culicover
Amygdala
University of Illinois at Urbana- Ohio State University
F. Robert Brush Champaign Language Learning: Humans
Purdue University Morphological Basis of
Active and Passive Avoidance Learning and Memory: Joel L. Davis
Learning: Behavioral Vertebrates Office of Naval Research
Phenomena Neurotransmitter Systems
Patricia Smith Churchland and Memory
Dean V. Buonomano University of California, San
UCLA Diego Michael Davis
Conditioning, Cellular and Aristotle (384322 B.C.E.) Emory University
Network Schemes for Habituation and Sensitization
Anna T. Cianciolo in Vertebrates
Higher-Order Features of
Yale University Neural Substrates of Classical
Classical
Intelligence and Memory Conditioning: Fear-
Rebecca D. Burwell Potentiated Startle
Len Cleary
Brown University
University of Texas Medical Anthony Dickinson
School at Houston University of Cambridge, UK
Brain: Perirhinal Cortex
Morphological Basis of Konorski, Jerzy (19031973)
and Associated Cortical
Learning and Memory:
Areas John F. Disterhoft
Invertebrates
John H. Byrne Northwestern University Medical
Neal Cohen
University of Texas Medical School
University of Illinois
School at Houston Olds, James (19221973)
Declarative Memory
Aplysia: Classical
Barbara Anne Dosher
Conditioning and Operant Thomas S. Collett
University of California, Irvine
Conditioning University of Sussex, Brighton,
Memory Search
Aplysia: Molecular Basis of UK
Long-Term Sensitization Spatial Learning: Animals Janet M. Duchek
Martin A. Conway Washington University School of
Thomas J. Carew
University of Durham, UK Medicine, St. Louis
Aplysia: Development of Autobiographical Memory Tip-of-the-Tongue
Processes Underlying Phenomenon
Nelson Cowan
Learning M. David Egger
University of Missouri
Metaplasticity University of Medicine and
Modality Effects
Leslie J. Carver Sensory Memory Dentistry of New Jersey
University of California, Habituation and Sensitization
Fergus I. M. Craik in Vertebrates
San Diego
Rotman Research Institute of
Infancy, Memory in Howard Eichenbaum
Baycrest Centre, Toronto

Claude G. Cech Aging and Memory in Boston University
University of Louisiana at Humans Long-Term Potentiation:
Lafayette Behavioral
Terry J. Crow
Semantic Memory: Cognitive Matthew Hugh Erdelyi
University of Texas Medical
Effects School at Houston Brooklyn College and the
Jason C. K. Chan Invertebrate Learning: Graduate School, City
Washington University, St. Louis Associative Learning in University of New York
False Memories Hermissenda Freud, Sigmund (18561939)
Lu Chen Robert G. Crowder K. Anders Ericsson
University of California, San Yale University (ret.) Florida State University
Francisco Eidetic Imagery Experts Memories
LIST OF CONTRIBUTORS xvii
W. K. Estes Bennett G. Galef, Jr. Ann M. Graybiel

Indiana University McMaster University, Ontario Massachusetts Institute of
Mathematical Learning Food Aversion and Technology
Theory Preference Learning in Guide to the Anatomy of the
Humans Brain: Basal Ganglia
Michael S. Fanselow
UCLA John M. Gardiner John T. Green
University of Sussex, UK Indiana University
Neural Substrates of Classical
Modality Effects Classical Conditioning:
Conditioning: Fear
Behavioral Phenomena
Conditioning, Freezing Alan Gelperin
Monell Chemical Senses Center, Robert L. Greene
Daniel J. Felleman
Philadelphia Case Western Reserve University
University of Texas Medical Distributed Practice Effects
Invertebrate Learning:
School at Houston Associative Learning in Repetition and Learning
Neocortical Plasticity: Adult Limax
Visual CortexAdaptation William T. Greenough
Kelly Sullivan Giovanello University of Illinois at Urbana-
and Reorganization
Boston University School of Champaign
Leif H. Finkel Medicine Morphological Basis of
University of Pennsylvania Amnesia, Organic Learning and Memory:
Neural Computation: Vertebrates
Elizabeth L. Glisky
Neocortex Howard E. Gruber
University of Arizona
Thomas M. Fischer Rehabilitation of Memory Teachers College, Columbia
Wayne State University Disorders University (ret.)
Piaget, Jean (18961980)
Metaplasticity Bill P. Godsil
UCLA Bengt Gustafsson
R. Holly Fitch
Neural Substrates of Classical Gteborg University, Sweden
University of Connecticut, Storrs
Conditioning: Fear Long-Term Potentiation:
Learning Disabilities
Conditioning, Freezing Overview: Cooperativity and
William N. Frost Associativity
Elkhonon Goldberg
Chicago Medical School Eric Hanse
New York University School of
Invertebrate Learning: Gteborg University, Sweden
Medicine
Habituation and Long-Term Potentiation:
Knowledge Systems and
Sensitization in Tritonia Material-Specific Memory Overview: Cooperativity and
Deficits Associativity
John J. Furedy
University of Toronto Luria, A .R. (19021977) Gerri Hanten
Pavlov, Ivan (18491936) I. Gormezano Baylor College of Medicine,
University of Iowa (ret.) Houston
Joaquin M. Fuster Head Injury
UCLA Conditioning, Classical and
Instrumental Kristen M. Harris
Prefrontal Cortex and
Gilbert Gottlieb Boston University
Memory in
University of North Carolina at Guide to the Anatomy of the
Michael Gabriel Brain: Synapse
Chapel Hill
University of Illinois and Thorndike, Edward Michael E. Hasselmo
Beckman Institute (18741949) Boston University
Neural Substrates of Guide to the Anatomy of the
Avoidance Learning Richard H. Granger, Jr.
Brain: Olfactory Cortex
Fred H. Gage Neural Computation: Harlene Hayne
Salk Institute for Biological Olfactory Cortex as a Model University of Otago, New
Studies, La Jolla, CA for Telencephalic Zealand
Neurogenesis Processing Amnesia, Infantile
xviii LIST OF CONTRIBUTORS
Alice F. Healy Masao Ito Yuri Koutcherov

University of Colorado Brain Science Institute, RIKEN, Prince of Wales Medical Research
Serial Organization Japan Institute, Australia
Long-Term Depression in the Guide to the Anatomy of the
Eliot Hearst
Cerebellum, Hippocampus, Brain: Overview
University of Arizona
and Neocortex
Watson, John B. (18781958) Mark Kritchevsky
Neural Computation:
University of California, San
Stephan Heckers Cerebellum
Diego
Massachusetts General Hospital Vestibulo-Ocular Reflex Amnesia, Transient Global
Schizophrenia and Memory (VOR) Plasticity
David J. Krupa
Fred J. Helmstetter Bharathi Jagadeesh Duke University
University of Wisconsin, University of Washington, Seattle Localization of Memory
Milwaukee Primates, Visual Attention in Traces
Genetic Substrates of
Marcia K. Johnson David Lavond
Memory: Amygdala
Yale University University of Southern California
Stewart Hendry Source Monitoring Visual Memory, Brightness
Johns Hopkins University and Flux in
Robert V. Kail
Purdue University Hilde A. E. Lechner
Brain: Cerebral Cortex
Children, Development of Salk Institute for Biological
Paula Hertel Memory in Studies, La Jolla, CA
Trinity University Mller, Georg Elias
Gerd Kempermann
Emotion, Mood, and (18501934)
Max Delbrck Center for
Memory
Molecular Medicine and Joseph E. LeDoux
Peter W. Hickmott Humboldt University, Berlin New York University
University of California, Neurogenesis Neural Substrates of
Riverside Emotional Memory
Howard H. Kendler
Neocortical Plasticity:
University of California, Santa Harvey Levin
Somatosensory Cortex
Barbara (ret.) Baylor College of Medicine,
Ernest R. Hilgard Spence, Kenneth Houston
Stanford University (ret.) (19071967) Head Injury
Guthrie, Edwin R.
John F. Kihlstrom Seymour Levine
(18861959)
University of California, Berkeley University of California, Davis
Philip N. Hineline Amnesia, Functional Early Experience and
Temple University Hypnosis and Memory Learning
Behaviorism
Jeansok J. Kim Derick H. Lindquist
Robert M. Hodapp Yale University Yale University
UCLA Kamins Blocking Effect: Long-Term Potentiation:
Mental Retardation Neuronal Substrates Amygdala
(Intellectual Disabilities)
Jeffrey A. Kleim Robert S. Lockhart
W. K. Honig University of Lethbridge, Alberta, University of Toronto
Dalhousie University, Halifax, Canada Coding Processes: Levels of
Canada (ret.) Neocortical Plasticity: Motor Processing
Operant Behavior Cortex Measurement of Memory
G. Horn Barbara J. Knowlton Elizabeth F. Loftus
Cambridge University UCLA University of Washington, Seattle
Imprinting Procedural Learning: Reconstructive Memory
Humans
James C. Houk Jessica M. Logan
Northwestern University Medical Bryan E. Kolb Washington University, St. Louis
School University of Lethbridge, Canada Tip-of-the-Tongue
Motor Skill Learning Harlow, Harry F. (19051981) Phenomenon
LIST OF CONTRIBUTORS xix
Fred D. Lorenzetti James L. McGaugh Neil W. Mulligan

University of Texas Medical University of California, Irvine Southern Methodist University
School at Houston Hormones and Memory Implicit Memory
Aplysia: Classical
Timothy McNamara Cynthia A. Munro
Conditioning and Operant
Conditioning Vanderbilt University Johns Hopkins University School
Spatial Memory of Medicine
Gary Lynch Dementia
University of California, Irvine Bruce L. McNaughton
Long-Term Potentiation: University of Arizona David J. Murray
Maintenance Neural Computation: Queens University at Kingston,
Hippocampus Canada
Colin M. MacLeod Bartlett, Frederic
University of Toronto at Michelle L. Meade
(18861969)
Scarborough Washington University, St. Louis
Individual Differences in Retrieval Processes in Elisabeth A. Murray
Learning and Memory Memory National Institute of Mental
Health
Steven E. Maier Andrew N. Meltzoff Primates, Visual Perception
University of Colorado University of Washington, Seattle and Memory in Nonhuman
Learned Helplessness Infancy, Memory in
Ian Neath
Steve Maren Randolf Menzel Purdue University
University of Michigan Freie Universitt Berlin Forgetting
Learning Theory: Current Invertebrate Learning:
Status Associative Learning and Ulric Neisser
Marc Marschark Memory Processing in Bees Cornell University
Rochester Institute of Technology Natural Settings, Memory in
M-Marsel Mesulam
Coding Processes: Imagery Thomas O. Nelson
Feinberg Medical School of
Michael Mauk Northwestern University University of Maryland, College
University of Texas Medical Guide to the Anatomy of the Park
School at Houston Brain: Basal Forebrain Metacognition about
Guide to the Anatomy of the Memory
Brain: Cerebellum Peter M. Milner
McGill University, Montreal Mark G. Packard
James L. McClelland Hebb, Donald (19041985) Yale University
Carnegie Mellon University Multiple-Memory Systems
Parallel Distributed Karen J. Mitchell Place versus Response
Processing Models of Yale University Learning Revisited in the
Memory Source Monitoring Brain
John Moore Procedural Learning:
Mark A. McDaniel
University of New Mexico University of Massachusetts- Animals
Mnemonic Devices Amherst Harold Pashler
Prose Retention Discrimination and University of California, San
Generalization Diego
Kathleen B. McDermott
Washington University, St. Louis Morris Moscovitch Attention and Memory
False Memories University of Toronto and Michael M. Patterson
Alexander J. McDonald Rotman Research Institute, University of California, San
University of South Carolina Toronto Diego
School of Medicine Frontal Lobes and Episodic Spinal Plasticity
Guide to the Anatomy of the Memory
George Paxinos
Brain: Amygdala Edvard I. Moser Prince of Wales Medical Research
Brian McElree Norwegian University of Science Institute, Australia
New York University and Technology Guide to the Anatomy of the
Memory Search Place Cells Brain: Overview
xx LIST OF CONTRIBUTORS
Steven E. Petersen Mark R. Rosenzweig James H. Schwartz

Washington University, St. Louis University of California, Berkeley Columbia University
Neuroimaging Tolman, Edward C. Second Messenger Systems
Gregg A. Phares (18861959)
Terrence J. Sejnowski
University of Texas Medical Paul Rozin Salk Institute for Biological
School at Houston University of Pennsylvania Studies, San Diego
Aplysia: Molecular Basis of Food Aversion and Neural Computation:
Long-Term Sensitization Preference Learning in Approaches to Learning
Donald A. Powell Humans
Matthew L. Shapiro
Dorn VA Medical Center and
David C. Rubin Mount Sinai School of Medicine,
University of South Carolina
Duke University New York
Neural Substrates of Classical
Conditioning: Oral Traditions Olton, David (19431994)
Cardiovascular Responses Duane M. Rumbaugh David F. Sherry
Donald L. Price Georgia State University University of Western Ontario
Johns Hopkins University School Language Learning: Evolution and Learning
of Medicine Nonhuman Primates Lorenz, Konrad (19031989)
Alzheimers Disease: Human Kurt Salzinger Sara J. Shettleworth
Disease and the Genetically American Psychological University of Toronto
Engineered Animal Models
Association Foraging
Stanley J. Rachman Skinner, B. F. (19041990)
University of British Columbia Paul G. Shinkman
E. Sue Savage-Rumbaugh University of North Carolina at
Behavior Therapy
Phobias Georgia State University Chapel Hill
Language Learning: Neocortical Plasticity:
Catharine H. Rankin Nonhuman Primates Development of the Visual
University of British Columbia
Meghan Saweikis System
Invertebrate Learning: C. ele-
gans Purdue University Jeanne L. Shinskey
Children, Development of University of Denver
Donald A. Riley
Memory in Object Concept,
University of California, Berkeley
Tolman, Edward C. Daniel L. Schacter Development of
(18861959) Harvard University Tracey J. Shors
Henry L. Roediger III Amnesia, Functional Rutgers University, Picataway,
Washington University, St. Louis Semon, Richard (18591918) NJ
Retrieval Processes in Glenn E. Schafe Stress and Memory
Memory New York University Jean-Louis Signoret
Benno Roozendaal Neural Substrates of Salpetriere Hospital, Paris (ret.)
University of California, Irvine Emotional Memory Ribot, Thodule (18391916)
Hormones and Memory Petra Scheck Alcino J. Silva
Gregory M. Rose University of Maryland, College UCLA
Memory Pharmaceuticals Corp., Park Genetic Substrates of
Montvale, NJ Metacognition about Memory Memory: Hippocampus
Cognitive Enhancers
Wolfram Schultz Memory Consolidation:
Drugs and Memory
Pharmacological Treatment University of Cambridge Molecular and Cellular
of Memory Deficits Reinforcement or Reward in Processes
Learning: Striatum
Steven P. R. Rose Wolf Singer
Open University, UK Roger W. Schvaneveldt Max Planck Institute for Brain
Protein Synthesis in Long- Arizona State University, East Research, Frankfurt, Germany
Term Memory in Coding Processes: Oscillations, Synchrony, and
Vertebrates Organization of Memory Neuronal Codes
LIST OF CONTRIBUTORS xxi
Brian H. Smith Rodney A. Swain Almira Vazdarjanova

Ohio State University University of Wisconsin- University of Arizona
Insect Learning Milwaukee Passive (Inhibitory)
Reinforcement or Reward in Avoidance, Fear Learning
E. N. Sokolov Learning: Cerebellum
Moscow Lomonosov State Mieke Verfaellie
Larry W. Swanson
University (ret.) Boston University School of
University of Southern California
Orienting Reflex Habituation Medicine and Boston VA
Ramn y Cajal, Santiago
Peter Somogyi (18521934) Healthcare System
University of Oxford, UK Amnesia, Organic
J. David Sweatt
Guide to the Anatomy of the Baylor College of Medicine, Indre V. Viskontas
Brain: Neuron Houston UCLA
Norman E. Spear Long-Term Potentiation: Procedural Learning:
Signal Transduction Humans
Binghamton University
Mechanisms and Early
Hunter, Walter S.
Events Allan R. Wagner
(18891954)
Jared P. Taglialatela Yale University
Larry R. Squire Georgia State University Sometimes Opponent
VA Medical Center, San Diego Language Learning: Process (SOP) Model, in
and University of California, Nonhuman Primates Conditioning
San Diego
Charles P. Thompson Michael J. Watkins
Electroconvulsive Therapy
Kansas State University Rice University
and Memory Loss
Mnemonists Memory Span
Memory Consolidation:
Prolonged Process of Richard F. Thompson M. N. Waxham
Reorganization University of Southern California
University of Texas Medical
Lashley, Karl (18901958)
Stephan Steidl School at Houston
Richard H. Thompson Glutamate Receptors and
University of British Columbia
University of British Columbia Their Characterization
Invertebrate Learning: C. ele-
Reinforcement or Reward in
gans Jack C. Waymire
Learning: Anatomical
Joseph E. Steinmetz Substrates University of Texas Medical
Indiana University Sharon L. Thompson-Schill School at Houston
Classical Conditioning: University of Pennsylvania Activity-Dependent
Behavioral Phenomena Semantic Memory: Regulation of
Neural Substrates of Classical Neurobiological Perspective Neurotransmitter Synthesis
Conditioning: Discrete
Darold A. Treffert Norman M. Weinberger
Behavioral Responses St. Agnes Hospital, WI University of California, Irvine
Robert J. Sternberg Savant Syndrome Neocortical Plasticity:
Yale University Tim Tully Auditory Cortex
Intelligence and Memory Cold Spring Harbor Laboratory,
NY Anthony P. Weiss
Robert Stickgold Massachusetts General Hospital
Invertebrate Learning:
Harvard Medical School Schizophrenia and Memory
Neurogenetics of Memory
Sleep and Memory
in Drosophila
Consolidation James V. Wertsch
Endel Tulving Washington University, St. Louis
Deborah Suchecki Rotman Research Institute of Collective Memory
Universidade Federal de So Baycrest Centre, Toronto
Paulo Ebbinghaus, Hermann Sheldon White
Early Experience and (18501909) Harvard University
Learning Episodic Memory James, William (18421910)
xxii LIST OF CONTRIBUTORS
Gordon Winocur Brain: Hippocampus and Anthony A. Wright

Trent University, Peterborough, Parahippocampal Region University of Texas Medical
and Rotman Research Institute, School at Houston
M.C. Wittrock
Toronto Comparative Cognition
UCLA
Frontal Lobes and Episodic Working Memory: Animals
School Learning
Memory Eugene B. Zechmeister
Philip C. Wong
Edward Wisniewski Loyola University
Johns Hopkins University School
University of North Carolina at Underwood, Benton
of Medicine
Greensboro (19151994)
Alzheimers Disease: Human
Concepts and Categories, Disease and the Genetically
Learning of Engineered Animal Models
Menno P. Witter Diana S. Woodruff-Pak
Vrije Universiteit Medical Center, Temple University
Amsterdam Aging and Memory in
Guide to the Anatomy of the Animals
A
ACTIVE AND PASSIVE AVOIDANCE the specified response during the WS-shock interval
LEARNING: BEHAVIORAL terminates the WS and prevents the occurrence of the
shock. In the passive form, suppression of the speci-
PHENOMENA
fied response during the WS-shock interval prevents
Avoidance learning is the behavioral product of an in- the occurrence of the shock. In both forms an inter-
strumental (operant) training procedure in which a trial interval (ITI) intervenes between successive pre-
predictable aversive event, typically electric shock, sentations of the WS, usually in the range of 0.5 to 5.0
does not occur contingent upon the occurrence or minutes.
nonoccurrence of a specified response by the learning
In the active free-operant procedure there are no
organism. Avoidance training occurs in two forms: ac-
discrete trials signaled by WSs. Instead, the avoidance
tive and passive. In the active form, the avoidance contingency is dependent on time. Specifically, two
contingency depends on the occurrence of a specified timers control events: a response-shock (R-S) timer
response on the part of the organism; in the passive (e.g., set for thirty seconds) and a shock-shock (S-S)
form, the avoidance contingency depends on the non- timer (e.g., set for five seconds). Training starts with
occurrence (i.e., the suppression) of some specified re- the S-S timer operating. Every time it runs out, it re-
sponse. The response to be suppressed may be either starts and delivers an inescapable shock of some dura-
spontaneous or learned by virtue of prior reward tion (e.g., 0.5 second). The specified response turns
training. In both forms, however, the avoidance con- off the S-S timer and starts the R-S timer. Every addi-
tingency consists of the prevention or omission of a tional response resets the R-S timer to its full value.
predictable noxious event. Noxious events are de- If the R-S timer runs out, it presents a shock and starts
fined in terms of the preference relation in which the the S-S timer (Sidman, 1953). This procedure has
absence of the event is preferred (measured by been used only in the active form. A variation of this
choice) to the presence of the event. Usually the nox- procedure eliminates the S-S timer and makes shock
ious event is electric shock, but loud noise, blasts of termination contingent upon the specified response
air, and high and low temperatures have been used. rather than upon a fixed duration of shock.
Avoidance training also utilizes one of two proce- In addition, a free-operant passive procedure
dures: discrete-trial or free-operant. In the discrete- known as punishment simply takes a response, which
trial procedure a distinctive stimulus, called a warning occurs spontaneously or by virtue of prior reward
signal (WS), signals the organism that the occurrence training, and makes shock or some other aversive
of the aversive event (e.g., electric shock) is imminent. event contingent on the occurrence of that response.
In most experiments the WS-shock interval is five to The response is usually suppressed. This is also called
sixty seconds in duration. In the active form, making passive-avoidance training. It has been used in a proce-
1
2 ACTIVE AND PASSIVE AVOIDANCE LEARNING: Behavioral Phenomena
dure in which an animal such as a mouse or rat runs This theoretical problem was ostensibly solved by
from a brightly lighted elevated platform into a dark Mowrer (1950), supported by Solomon and Wynne
compartment where it receives a single electric shock. (1954) and Rescorla and Solomon (1967), by postulat-
The tendency to enter the dark compartment is ining that Pavlovian conditioning of fear on early es-
nate, and the single punishment results in subsequent cape trials, in which the WS is paired with shock, pro-
long latencies to reenter the dark compartment. This vided the acquired motivation to terminate the WS
is called one-trial passive-avoidance training, and it (now a conditioned aversive stimulus), thus providing
has been used extensively in the study of memory be- secondary (acquired) negative reinforcement for the
cause the learning event is fixed in time, which allows escape-from-fear response (i.e., the avoidance re-
analysis and manipulation of temporarily constrained sponse). Others thought that the fear response was in-
neuropharmacological and endocrine processes asso- strumentally reinforced by the termination of shock
ciated with learning. Alternatively, a hungry animal (Miller, 1951), but the upshot was the same: Reduc-
may initially be rewarded with food for pressing a tion of fear by termination of the WS, whether ac-
lever and subsequently shocked for making that same quired by Pavlovian or instrumental means, was the
response. Usually several shocks are required to sup- source of the acquired negative reinforcement for the
press the lever pressing. avoidance response. Thus, two processes were postu-
lated: acquisition of fear during escape trials (by Pav-
Warner (1932) was the first to use a discrete-trial
lovian or operant conditioning) and acquisition of the
active-avoidance procedure to study the association
span of the white rat (using WS-shock intervals of one instrumental avoidance response, reinforced by fear
to thirty seconds); he used what has become known as reduction. This theoretical interpretation was sup-
a shuttle box, a two-compartment box in which the ported by the results of an elegant experiment by
animal is required to run or jump back and forth be- Kamin (1956). Additional research in support of two-
tween the two compartments to avoid the shock. process theory used a transfer paradigm in which ani-
mals were given Pavlovian conditioning in one situa-
These procedures and the behaviors they pro- tion, and the effects of those conditioned stimuli were
duce have been of interest to psychologists since the observed when they were subsequently superimposed
early studies of behaviorism in the United States. on an operant baseline of responding in another situ-
John Watson and, especially, Edward Thorndike pos- ation (Solomon and Turner, 1960). This two-process
tulated that learned responses were a product of their theory provides the best account of avoidance learn-
consequences. That is, a response occurs, a pleasur- ing in its various forms.
able or aversive event ensues, and the response is re-
inforced (increases) if the event is pleasurable or pun- Some animals of most species learn the avoidance
ished (decreases) if the event is aversive. contingency, whether in the active or the passive
form, using discrete-trial or free-operant procedures.
Hilgard and Marquis (1940), two early behavioral Dogs are particularly adept at avoidance learning in
theorists, had trouble accounting for avoidance learn- an active, discrete-trial shuttle-box procedure and
ing because it was a product of a procedure where the typically show strong resistance to extinction (Solo-
reinforcing event was the response-contingent absence mon and Wynne, 1954). In contrast, rats are particu-
of an event, not the response-contingent presence of an larly difficult to train in an active, lever-press, dis-
event: crete-trial procedure and require special training
procedures (Berger and Brush, 1975). Thus, there
Learning in this [avoidance] situation ap- are important differences among species and re-
pears to be based in a real sense on the avoid-
sponse requirements. Additionally, in all forms of
ance of the shock. It differs clearly from other
avoidance learningactive and passive, discrete-trial
types of instrumental training in which the
and free-operantthere are enormous individual dif-
conditioned response is followed by a definite
ferences. Some individuals of whatever species learn
stimulus changefood or the cessation of
rapidly and well, whereas others do not (Brush,
shock [reward training (positive reinforce-
1966).
ment) or escape training (negative reinforce-
ment)]. In instrumental avoidance training In view of these findings it is not surprising that
the new response is strengthened in the ab- several investigators have genetically selected for dif-
sence of any such stimulus; indeed, it is ferences in avoidance learning. Bignami (1965) re-
strengthened because of the absence of such ported the first experiment with Wistar albino rats in
a stimulus. Absence of stimulation can obvi- which the selectively bred phenotypes were good or
ously have an influence on behavior only if poor at avoidance learning in a shuttle box. The re-
there exists some sort of preparation for or sulting strains are known as the Roman High Avoid-
expectation of the stimulation. (pp. 5859) ance and Roman Low Avoidance strains (RHA and
ACTIVE AND PASSIVE AVOIDANCE LEARNING: Behavioral Phenomena 3
RLA, respectively). Training consisted of five daily avoidance learning is strongly influenced by genetic
sessions of fifty trials each. Selection was based on the factors, and many behavioral, physiological, and ana-
number of avoidance responses during the first two tomical correlates of avoidance learning have been
sessions (many or few) and on good or poor retention identified. Several of those correlates appear to be
from each session to the next. Selection was highly ef- closely linked, genetically, to the avoidance pheno-
fective because, by the fifth generation, the RHA and types. Researchers are trying to identify the mecha-
RLA animals avoided, respectively, on 68 percent and nisms by which genes determine avoidance learning.
20 percent of the trials. Modern molecular-genetic technology might enable
In 1977 Brush reported on the development of them to identify those genes.
the Syracuse High Avoidance and Syracuse Low
Avoidance strains (SHA and SLA, respectively). Long- See also: NEURAL SUBSTRATES OF AVOIDANCE
Evans hooded rats were trained for sixty trials in auto- LEARNING; PASSIVE (INHIBITORY) AVOIDANCE,
mated shuttle boxes. The data from over twenty gen- FEAR LEARNING
erations of selection indicated that shuttle-box avoid-
ance learning is heritable: SHA and SLA animals
Bibliography
avoided on 67 percent and 0 percent of the sixty trials
Bammer, G. (1978). Studies on two new strains of rats selectively
of training. Realized heritability (h2, which can range bred for high or low conditioned avoidance responding.
between 0.0 and 1.0; Falconer, 1960) was estimated Paper presented at the Annual Meeting of the Australian Soci-
to be 0.16 in each strain, a value comparable with that ety for the Study of Animal Behavior, Brisbane.
found in other selection studies (Brush, Froehlich, Berger, D. F., and Brush, F. R. (1975). Rapid acquisition of dis-
and Sakellaris, 1979). crete-trial lever-press avoidance: Effects of signal-shock inter-
val. Journal of the Experimental Analysis of Behavior 24, 227239.
In 1978 Bammer reported on the first six genera- Bignami, G. (1965). Selection for high rates and low rates of avoid-
tions of selective breeding of Sprague-Dawley albino ance conditioning in the rat. Animal Behavior 13, 221227.
rats for high and low levels of avoidance responding Brush, F. R. (1966). On the differences between animals that learn
and do not learn to avoid electric shock. Psychonomic Science
in a shuttle box. The resulting strains are known as
5, 123124.
the Australian High Avoidance and Australian Low (1977). Behavioral and endocrine characteristics of rats se-
Avoidance strains (AHA and ALA, respectively). lectively bred for good and poor avoidance behavior. Activitas
Training consisted of fifty trials in one or more daily Nervosa Superioris 19, 254255.
sessions. Realized heritability over the first five gener- Brush, F. R., Froehlich, J. C., and Sakellaris, P. C. (1979). Genetic
ations of selection was 0.18 and 0.27 for the AHA and selection for avoidance behavior in the rat. Behavior Genetics
9, 309316.
ALA strains, respectively. Falconer, D. S. (1960). Introduction to quantitative genetics. London:
A unidirectionally selected strain, known as the Oliver and Boyd.
Tokai High Avoider (THA), was bred in Japan from Hilgard, E. R., and Marquis, D. G. (1940). Conditioning and learning.
New York: Appleton-Century-Crofts.
Wistar stock using a lever-press response and a free- Kamin, L. J. (1956). The effect of termination of the CS and avoid-
operant procedure (S-S = 5 seconds, R-S = 30 sec- ance of the US on avoidance learning. Journal of Comparative
onds, shock duration = 0.5 second). The selection cri- and Physiological Psychology 49, 420424.
terion was an avoidance rate of more than ninety-five Miller, N. E. (1951). Learnable drives and rewards. In S. S. Stevens,
percent in the last five of ten daily one-hour training ed., Handbook of experimental psychology. New York: Wiley.
Mowrer, O. H. (1950). On the dual nature of learninga reinter-
sessions. Selection was successful: THA males and fe-
pretation of conditioning and problem solving. In Mo-
males learn faster and to a higher level of perfor- wrers Learning theory and personality dynamics. New York: Ron-
mance than unselected control animals from the orig- ald Press.
inal stock. Rescorla, R. A., and Solomon, R. L. (1967). Two-process learning
theory: Relationships between Pavlovian conditioning and in-
The fact that so many selective breeding experi- strumental learning. Psychological Review 74, 151182.
ments for avoidance behavior have been successful is Sidman, M. (1953). Two temporal parameters of the maintenance
a clear indicator of the extent to which this kind of be- of avoidance behavior by the white rat. Journal of Comparative
havior is under genetic control. In each experiment and Physiological Psychology 46, 253261.
the individual variability within each strain becomes Solomon, R. L., and Turner, L. H. (1960). Discriminative classical
conditioning under curare can later control discriminative
less as selection progresses, and it appears not to mat- avoidance responses in the normal state. Science 132, 1,499
ter what the details of the training procedures are. 1,500.
For example, SHA animals do better than controls in Solomon, R. L., and Wynne, L. C. (1954). Traumatic avoidance
a free-operant procedure, and THA animals do better learning: The principles of anxiety conservation and partial
than controls in discrete-trial, shuttle-box training. irreversibility. Psychological Review 61, 353385.
Warner, L. H. (1932). The association span of the white rat. Journal
Similarly, AHA animals outperform ALA animals in
of Genetic Psychology 39, 5789.
a discrete-trial avoidance task quite different from the
one in which they were selected. Thus, it is clear that F. Robert Brush
4 ACTIVITY-DEPENDENT REGULATION OF NEUROTRANSMITTER SYNTHESIS
ACTIVITY-DEPENDENT REGULATION equilibrium with choline acetyltransferase (CAT), the

OF NEUROTRANSMITTER enzyme that catalyzes acetylcholine synthesis (Jope,
1979). Consequently, acetylcholine synthesis is well
SYNTHESIS
below its maximal possible rate. Therefore any
Activity-dependent regulation of neurotransmitter synthesis change in the concentration of acetylcholine or its
refers to the ability of some nerve cells to change the precursors produces a change in acetylcholine syn-
amount of neurotransmitter synthesized in response thesis. This has been verified by the demonstration
to activity. Study of this regulation is prompted by the that the transport of choline into the cholinergic neu-
belief that it is important not only for maintaining a ron controls acetylcholine synthesis. For example,
source of neurotransmitter but also for adaptive choline addition to slices of brain tissue markedly in-
changes that take place in certain nerve cells during creases the acetylcholine synthesis rate. In addition,
learning and memory. A basic postulate necessary for increased neuronal activity increases choline uptake
neurotransmitter synthesis regulation to be a mecha- (Simon and Kuhar, 1975). This increased uptake oc-
nism for learning and memory is that increased curs in a manner that persists far beyond the period
neurotransmitter synthesis acts to increase neuro- of increased activity. Thus, stimulation of choline up-
transmitter secretion and, as a consequence, synaptic take is not due merely to a shift in the equilibrium of
strength. It has not been technically possible thus far the CAT-catalyzed reaction; instead, choline uptake
to demonstrate a causal relationship between activity- is regulated by neural activity per se. This regulation
dependent regulation of neurotransmitter synthesis of choline uptake by nerve activity is predicted to
and an increase in neurotransmitter secretion. None- maintain the strength of cholinergic synapses and is
theless, activity-dependent regulation of neurotran- a candidate to increase the capacity of these synapses
smitter synthesis remains a candidate for the cause of to secrete acetylcholine. Despite this evidence for ac-
neuroplastic changes that underlie learning, memo- tivity-dependent regulation of choline uptake, little is
ry, and neuroplasticity. Neurotransmitters that have known concerning how this regulation occurs. Thus,
shown activity-dependent regulation of their biosyn- the link between nerve activity and choline transport
thesis are acetylcholine, dopamine, and norepineph- remains unknown.
rine. This entry reviews the mechanisms of the syn- Catecholaminesdopamine, norepinephrine, and
thesis of these three neurotransmitters and possible epinephrineare neurotransmitters in the sympa-
roles of their regulatory mechanisms in learning and thetic limb of the autonomic nervous system and in
memory. several groups of neurons in the CNS. In contrast
Depending upon both the type of nerve cell and with the lack of a mechanism linking nerve activity
the time scale over which adaptation occurs, the cellu- and the regulation of choline uptake, catecholamine-
lar and biochemical mechanisms responsible for ac- synthesizing cells have several mechanisms in place to
tivity-dependent regulation of neurotransmitter syn- regulate catecholamine levels in response to nerve ac-
thesis vary. The time scale of changes ranges from tivity, both in peripheral and central nervous systems
very rapid changes (seconds), in which covalent modi- and at short- and long-term levels. As one review stat-
fication of enzyme protein structure is involved, to ed, An intricate scheme has evolved whereby tyro-
more delayed, longer-term changes (days). The latter sine hydroxylase activity is modulated by nearly every
involve alterations in genetic expression and turnover documented form of regulation (Kumer and Vrana,
of enzymes responsible for neurotransmitter biosyn- 1996). Regulation occurs at the step in catechol-
thesis. Catecholamines are regulated at both short- amines synthesis where tyrosine is hydroxylated to
and long-term levels, while acetylcholine is regulated form L-dopa. The enzyme catalyzing this reaction, ty-
only at a short-term level. The regulatory mecha- rosine hydroxylase, is the first of four enzymatic steps
nisms are often similar to cellular and biochemical in the catecholamine synthesis pathway. Because tyro-
mechanisms used in nonneural cells to regulate the sine hydroxylase is present in lower concentration
than the other enzymes of the synthesis pathway, it re-
synthesis of hormones or to regulate the biochemical
stricts the amount of neurotransmitter synthesized. In
pathways of intermediary metabolism.
addition, tyrosine hydroxylase is constitutively inhib-
Acetylcholine is the neurotransmitter in the auto- ited by the binding of a mole of catecholamine to each
nomic nervous system, the central nervous system mole of enzyme. Two interacting mechanisms are im-
(CNS), and at the neuromuscular junction. Even portant in short-term, nervous-activity regulation of
though acetylcholine has one of the highest rates of catecholamine synthesis. One mechanism is catechol-
turnover of any neurotransmitter, its concentration amine end-product inhibition of tyrosine hydroxylase
within nerve tissue fluctuates very little. This is be- activity by the catecholamine products of the pathway.
cause the precursors for acetylcholine synthesis, ace- The second is modification of tyrosine hydroxylase
tylcoenzyme A and choline, exist in a steady-state structure by the placement of phosphate groups on
ACTIVITY-DEPENDENT REGULATION OF NEUROTRANSMITTER SYNTHESIS 5
the tyrosine hydroxylase molecule. The latter pro- Catecholamine synthesis is also regulated at a
cess, termed phosphorylation, is catalyzed by at least chronic, long-term level in response to persistent or
four separate protein kinases (Kumer and Vrana, extreme neural activation. In this case the amount of
1996), which phosphorylate tyrosine hydroxylase on the enzymes in the catecholamine synthetic pathway,
a combination of three serine residues in the N termi- and especially tyrosine hydroxylase, is elevated in re-
nal domain of the enzyme. The phosphorylation al- sponse to increased synaptic activity. For example,
ters the properties of tyrosine hydroxylase, increasing drugs or conditions such as stress that increase the au-
its catalytic activity. Phosphorylation is commonly tonomic nerve activity increase the level of tyrosine
used to modify proteins involved in regulation. hydroxylase in peripheral autonomic cells (Thoenen,
Mueller, and Axelrod, 1969). Because cutting the in-
The protein kinases that phosphorylate tyrosine nervation to these cells blocks the increase, the influ-
hydroxylase are cyclic adenosine monophosphate- ence is thought to be transynaptic. Because a rise in
dependent protein kinase (PKA), which phosphoryl- tyrosine hydroxylase mRNA precedes the increase in
ates serine 40; calcium-calmodulin-dependent pro- protein, the mechanism is believed to be increased
tein kinase II (CaM KII), which phosphorylates serine transcription of the mRNA encoding tyrosine hy-
19; and extracellular receptor activated protein ki- droxylase. The observation that drugs that increase
nase (ERK) (Haycock, Ahn, Cobb, and Krebs, 1992), CNS neuronal activity also induce increased levels of
which phosphorylates serine 31. This phosphoryl- CNS tyrosine hydroxylase shows that central tyrosine
ation has three major effects on the enzymes func- hydroxylase levels are also regulated by neuronal ac-
tion; all three changes enhance tyrosine hydroxylase tivity. This nerve activitydependent regulation of ty-
activity and increase catecholamine synthesis. Serine rosine hydroxylase synthesis is an attractive candidate
40 phosphorylation increases the affinity of the en- for learning and memory because the increase in ty-
zyme for tetrahydrobiopterin cofactor (the cofactor is rosine hydroxylase level is expected to increase the
normally present below optimal concentrations) and strength of the activated synapses. Whether or not
reduces the catecholamine binding and inhibition of this is the case is not yet known. Even so, a consider-
tyrosine hydroxylase (Daubner, Lauriano, Haycock, able effort is being carried out to understand as much
and Fitzpatrick, 1992). Serine 19 phosphorylation as possible about transynaptic regulation of tyrosine
causes the enzyme to interact with an activating pro- hydroxylase level because it is likely the best example
tein termed 14-3-3 protein (Ichimura et al., 1987; Ita- known of synaptically mediated regulation of protein
gaki et al., 1999). Serine 31 phosphorylation in- synthesis.
creases the catalytic activity by an as yet undefined Among the issues that remain unresolved con-
mechanism (Haycock, Ahn, Cobb, and Krebs, 1992). cerning the mechanism of the long-term regulation
of tyrosine hydroxylase is the nature of the intracellu-
Which of these mechanisms act to regulate the
lar mechanisms responsible for increased transcrip-
synthesis of catecholamines in response to neural ac-
tion. In a model tissue, the adrenal medulla, acetyl-
tivity, and how may they relate to learning and memo- choline and pituitary adenylyl cyclase activating
ry? Although these questions have been difficult to polypeptide (PACAP) are each able to modulate tyro-
answer, some conclusions are possible. Under circum- sine hydroxylase level. In this tissue either PACAP or
stances of cell depolarization, such as when a nerve acetylcholine stimulates cAMP level and activates
impulse invades the nerve terminal or during cholin- PKA. One hypothesis is that PKA migrates to the cell
ergic stimulation at the adrenal medulla, the phos- nucleus to regulate the rate of tyrosine hydroxylase
phorylation and activation of tyrosine hydroxylase by transcription by phosphorylating CREB (Cyclic AMP
Ca/CAM kinase II appear to be responsible for initial Response Element Binding), a regulatory protein as-
activation of tyrosine hydroxylase (Waymire et al., sociated with the tyrosine hydroxylase gene (Kuro-
1988; Waymire and Craviso, 1993) with a later activa- sawa, Guidotti, and Costa, 1976). Unresolved issues
tion through both PKA and ERK. In addition, evi- in this simple model are 1. whether acetylcholine
dence indicates that the phosphorylation on serine 19 stimulates a rise in cAMP or 2. whether other trans-
may facilitate the phosphorylation on serine 40 mitters, such as PACAP, are involved. Because
(Bevilaqua et al., 2001). Thus there appears to be a neuropeptides are secreted along with acetylcholine
time-dependent hierarchical phosphorylation and at some cholinergic synapses, it has been suggested
activation of tyrosine hydroxylase that occurs in steps that these agonists are responsible for long-term reg-
to activate catecholamine biosynthesis. Any condition ulation of tyrosine hydroxylase level (Wessels-Reiker,
that increases the size of these phosphorylation Haycock, Howlett, and Strong, 1991). Also, it is not
events is predicted to enhance the synthesis of cat- clear whether transcription is regulated solely
echolamines. This may translate into increased through PKA. Because several regulatory domains
neurotransmitter release and synaptic strength. exist in the tyrosine hydroxylase gene, it appears that
6 AGING
protein kinases other than the PKA are likely to be in- Haycock, J. W. (1990). Phosphorylation of tyrosine hydroxylase in
volved. In addition it is possible that the rapidly syn- situ at serine 8, 19, 31 and 40. Journal of Biological Chemistry
265, 11,68211,691.
thesized protein c-Fos may serve as a protein factor Haycock, J. W., Ahn, N. G., Cobb, M. H., and Krebs, E. G. (1992).
regulating tyrosine hydroxylase transcription. An ad- ERK1 and ERK2, two microtubule-associated protein 2 ki-
ditional question is whether increased transcription is nases, mediate the phosphorylation of tyrosine hydroxylase at
the principal mechanism of regulation. In isolated serine-31 in situ. Proceedings of the National Academy of Sciences
adrenal medullary chromaffin cells, transcription in- of the United States of America 89, 2,3652,369.
Ichimura, T., Isobe, T., Okuyama, T., Yamauchi, T. and Fujisawa,
creases for only a few hours following either acetyl- H. (1987). Brain 14-3-3 protein is an activator protein that ac-
choline of PACAP stimulation, whereas the increase tivates tryptophan 5-monooxygenase and tyrosine 3-
in mRNA occurs over several days and remains elevat- monooxygenase in the presence of Ca2+, calmodulin-
ed long after transcription has subsided. This raises dependent protein kinase II. FEBS Letters 219, 7982.
the issue of the stabilization of the tyrosine hydroxy- Itagaki, C., Isobe, T., Taoka, M., Natsume, T., Nomura, N., Hori-
gome, T., Omata, S., Ichinose, H., Nagatsu, T., Greene, L. A.,
lase mRNA as a major component of the synaptic reg- and Ichimura, T. (1999). Stimulus-coupled interaction of ty-
ulation. Indeed, tyrosine hydroxylase mRNA is capa- rosine hydroxylase with 14-3-3 proteins. Biochemical Journal
ble of regulation through stability, as has been 38, 15,67315,680.
demonstrated for its stabilization by elevated oxygen Jope, R. S. (1997). High affinity choline transport and acetylCoA
tension (Paulding and Czyzyk-Krzeska, 1999). production in brain and their roles in the regulation of acetyl-
choline synthesis. Brain Research Review 1, 313344.
The understanding of the activity-dependent Kumer, S. C., and Vrana, K. E. (1996). Intricate regulation of tyro-
regulation of catecholamines is much more complete sine hydroxylase activity and gene expression. Journal of
than for other neurotransmitters, such as acetylcho- Neurochemistry 67, 443462.
Kurosawa, A., Guidotti, A., and Costa, E. (1976). Induction of tyro-
line. For some neurotransmitter systemsthe amino sine 3-monooxygenase elicited by carbamylcholine in intact
acids and purines, for examplealmost nothing is and denervated adrenal medulla: Role of protein kinase acti-
known about their synthesis regulation, so it is not vation and translocation. Molecular Pharmacology 12, 420432.
clear whether it is activity-dependent. This is primari- Paulding, W. R., and Czyzyk-Krzeska, M. F. (1999). Regulation of
ly because these compounds are so intimately associ- tyrosine hydroxylase mRNA stability by protein-binding, py-
rimidine-rich sequence in the 3-untranslated region. Journal
ated with intermediary metabolism that it is difficult of Biological Chemistry 274, 2,5322,538.
to separate their neurotransmitter-related metabo- Simon, J. R., and Kuhar, M. J. (1975). Impulse-flow regulation of
lism from that associated with general cell function. high affinity choline uptake in brain cholinergic nerve termi-
One generalization emerging from the studies of the nals. Nature 255, 162163.
mechanisms of activity-dependent regulation of cat- Thoenen, H., Mueller, R. A., and Axelrod, J. (1969). Trans-
synaptic induction of adrenal tyrosine hydroxylase. Journal of
echolamine synthesis is the prominent position pro- Pharmacology and Experimental Therapeutics 169, 249254.
tein phosphorylation plays in both short- and long- Waymire, J. C., and Craviso, G. L. (1993). Multiple site phosphory-
term regulation. In the future, studies will likely be di- altion and activation of tyrosine hydroxylase. Advances in Pro-
rected to applying the understanding being gained of tein Phosphatases 7, 495506.
the mechanisms regulating catecholamine synthesis Waymire, J. C., Johnston, J. P., Hummer-Lickteig, K., Lloyd, A.,
Vigny, A., and Craviso, G. L. (1988). Phosphorylation of bo-
to other neurotransmitters. And although it is impor- vine adrenal chromaffin cell tyrosine hydroxylase: Temporal
tant to continue to investigate the mechanisms in- correlation of acetylcholines effect on site phosphorylation,
volved in activity-dependent regulation of neurotran- enzyme activation, and catecholamine synthesis. Journal of Bi-
smitter synthesis, it is also important to recognize that ological Chemistry 263, 12,43912,447.
the role of neurotransmitter synthesis regulation in Wessels-Reiker, M, Haycock, J. W., Howlett, A. C., and Strong, R.
(1991). Vasoactive intestinal polypeptide induces tyrosine hy-
higher functions, such as learning and memory, is still droxylase in PC12 cells. Journal of Biological Chemistry 266,
hypothetical. 9,3479,350.
Jack C. Waymire
See also: PROTEIN SYNTHESIS IN LONG-TERM
MEMORY IN VERTEBRATES
Bibliography
Bevilaqua, L. R., Graham, M. E., Dunkley, P. R., Nagy-Felsobuki,
E. I., and Dickson, P. W. (2001). Phosphorylation of Ser(19) AGING
alters the conformation of tyrosine hydroxylase to increase See: AGING AND MEMORY IN ANIMALS; AGING
the rate of phosphorylation of Ser(40). Journal of Biological
AND MEMORY IN HUMANS; ALZHEIMERS
Chemistry 276, 40,41140,416.
Daubner, S. C., Lauriano, C., Haycock, J. W., and Fitzpatrick, P.
DISEASE: BEHAVIORAL ASPECTS; ALZHEIMERS
F. (1992). Site-directed mutagenesis of serine 40 of rat tyro- DISEASE: HUMAN DISEASE AND THE
sine hydroxylase. Effects of dopamine and cAMP-dependent GENETICALLY ENGINEERED ANIMAL MODELS;
phosphorylation on enzyme activity. Journal of Biological Chem- PHARMACOLOGICAL TREATMENT OF
istry 267, 12,63912,646. MEMORY DEFICITS
AGING AND MEMORY IN ANIMALS 7
AGING AND MEMORY IN ANIMALS conditioning is a simple form of learning that can be
studied with little modification across a variety of spe-
The passage of time produces changes in both the be- cies, including humans. There are at least four advan-
havior and the brains of organisms. A number of use- tages to using eyeblink conditioning as an animal
ful animal models of learning and memory in normal model of the effects of aging on learning and memory
aging have expanded the knowledge base and ex- in humans. First, both animals and humans show age-
tended the prospects for ameliorating learning and associated deficits in conditioning, and these can be
memory deficits. Completion of the mapping of the easily dissociated from age-associated changes in sen-
human and mouse genome and the development of sory systems (i.e., differences in CS thresholds) or
transgenic mouse models in the 1990s have accelerat- motor systems (differences in UR amplitude). Sec-
ed insights about mechanisms of learning, memory, ond, both animals and humans show age-associated
and aging. Since the mid-1990s, mouse models of changes in the neural substrates critical for eyeblink
neuropathology in Alzheimers disease have become conditioning, the cerebellum and the hippocampus.
available for behavioral testing. Two features of ani- Third, age-associated deficits have been artificially in-
mal models make them invaluable: First, the life duced with drugs in both young animals and young
spans of most animals are considerably shorter than humans. Finally, age-associated deficits can be re-
the human life span, compressing the time required versed in normal older animals using cognition-
to observe processes of aging. Second, invasive or enhancing drugs.
high-risk observations and experimental manipula-
tions are feasible with animals but not with humans. The critical substrate for eyeblink conditioning is
the cerebellum. The hippocampus plays a modulato-
Aging is most typically associated with declines in ry role in acquisition of CRs. Abnormal functioning
functioning, both neural and behavioral. However, of the hippocampus retards the rate of eyeblink con-
individual organisms age at different rates. One or- ditioning in the delay procedure. In addition, the
ganism may show a steady decline in functioning, hippocampus is essential for eyeblink classical condi-
whereas another shows only slight changes over the tioning procedures involving greater complexity,
years. An important goal toward an understanding of such as trace conditioning. In the trace procedure,
aging processes is to determine how changes in neu- the CS is presented and then turned off, and a blank
ral structures impact behavior. As such, behavioral period ensues before the onset of the US. The blank
paradigms that engage well-defined neural substrates period is called the trace and is shown in the left
are particularly valuable. Two such neurobiologically panel of Figure 1.
well-characterized paradigms will be highlighted:
eyeblink classical conditioning and spatial learning In eyeblink classical conditioning, intervals be-
and memory. tween the CS and US (called interstimulus intervals,
ISIs) of 250 and 750 milliseconds are the most com-
mon, with the most rapid conditioning between 250
Eyeblink Classical Conditioning and 500 milliseconds. The trace procedure extends
The basic classical conditioning procedure, the ISI in addition to inserting the blank trace period.
named the delay procedure by Ivan Pavlov, in- Thus, it increases difficulty in two ways. Holding the
volves repeated trials in which the presentation of an ISI constant at 750 milliseconds, researchers com-
initially neural stimulus, such as a tone (the condi- pared the delay and trace procedures in three-
tioned stimulus, or CS), is followed after approxi- month-old and twenty-four-month-old rabbits. There
mately half a second by a stimulus that evokes a re- were significant effects of age and procedure (see Fig-
flexive eyelid closure, such as a corneal air puff (the ure 1). Older rabbits performed more poorly in both
unconditioned stimulus, or US). The CS turns on and procedures, and both age groups performed more
remains on while the US is delivered and the two poorly in the trace than in the delay procedure. Com-
stimuli coterminate. Initially, eye blinks occur only parison of a number of studies testing the delay and
after the US, and the blinks are a reflexive response trace procedure in older rabbits indicated that age
(the unconditioned response, or UR). Eventually, eye differences in conditioning appeared earlier when
blinks occur after the CS but before the US. This is a the trace procedure was used.
learned response (the conditioned response, or CR). Cerebellar Substrates of Impaired
Thus, learning is defined as the acquisition of CRs. Conditioning in Older Animals
Richard F. Thompson suggested that the eye- Purkinje cells in the cerebellum integrate CS and
blink classical conditioning paradigm might be the US input and show patterns of engagement during
Rosetta Stone for brain substrates of age-related defi- eyeblink conditioning. In rabbits, Purkinje cell counts
cits in learning and memory (Thompson, 1988). have been carried out using histological techniques
Thompsons major point was that eyeblink classical after behavioral testing with eyeblink classical condi-
8 AGING AND MEMORY IN ANIMALS
Figure 1
Left: Delay and trace procedures using same 750 milliseconds ISI (interstimulus intervals). Right: The number of trials it took rabbits
to attain a learning criterion of eight conditioned responses in nine consecutive trials. The higher the trials to criterion measure, the
slower the animals were to learn. A total of twenty-four young and seventeen old rabbits were tested in either the 750-milliseconds
delay or the 750-milliseconds trace (250 milliseconds CS, 500 milliseconds trace) procedure. There were statistically significant
differences between the younger and older rabbits, and there were statistically significant differences between the delay and trace
procedures. Older rabbits learned more slowly in both the delay and trace procedures, and both younger and older rabbits learned
more slowly in the trace than in the delay procedure.
tioning. The correlations between Purkinje cell num- Logan, and Thompson, 1987). Older rabbits were sig-
ber and eyeblink classical conditioning in rabbits were nificantly impaired in acquiring CRs. Furthermore,
high and statistically significant (Woodruff-Pak, older rabbits showed significantly less neural activity
Cronholm, and Sheffield, 1990). The fewer Purkinje in the US period than young rabbits by the second
cells a rabbit had, the longer it took it to acquire CRs. session of training. Matthew McEchron and John Dis-
Further analysis demonstrated that this relationship terhoft also reported that older rabbits had less hip-
was relatively independent of age because there was pocampal responsivity in the US period.
a highly significant correlation between Purkinje cell
number and conditioning when only young rabbits In a series of experiments in young and older rab-
were included. Mutant mice without Purkinje cells bits, Disterhoft and McEchron (2000) found that con-
condition slowly and produce fewer CRs, whereas ditioning-related hippocampal pyramidal-cell activity
their wild type littermates with Purkinje cells condi- varied across cells and that the different response
tion normally (Chen et al., 1996). Given the essential profiles were differentially affected by aging. Patterns
role of the cerebellum in the acquisition of CRs, Pur- of hippocampal pyramidal cell activation associated
kinje cell loss may account for a significant portion of with acquisition of trace eyeblink conditioning were
the age-related difference in eyeblink classical condi- different from activity recorded after CRs became as-
tioning. ymptotic. Pyramidal-cell activity associated with ac-
quisition was more sensitive to the effects of aging
Hippocampal Substrates of Impaired (McEchron, Weible, and Disterhoft, 2001). Various
Conditioning in Older Animals patterns of single-unit pyramidal-cell activity were
There is some evidence of age differences in hip- identified, and three response patterns were different
pocampal activity during eyeblink conditioning. Neu- between young and older rabbits that learned and
ronal responses were recorded from the dorsal hippo- those aged rabbits that did not. The patterns showed
campus of young (three-month-old), middle-aged significant changes during the first five days of condi-
(twenty-six-to-thirty-three-month-old), and older tioning in the young and aged learners, but the pat-
(thirty-nine-to-fifty-month-old) rabbits during eye- terns showed no change in the aged nonlearning
blink conditioning in the 750-millisecond trace pro- group. If these cells function to hold important infor-
cedure shown in Figure 1 (Woodruff-Pak, Lavond, mation for consolidation in other neural structures,
AGING AND MEMORY IN ANIMALS 9
age-related deficits in conditioning may be ameliorat- trieve the correct map may explain why old rats are
ed by enhancing the function of these cells. more likely to make behavioral map-retrieval errors.
Older organisms, including rats, monkeys, and hu-
mans, have a greater tendency to become lost. Altered
Spatial Learning and Memory hippocampal plasticity mechanisms may underlie
Spatial behavioral tests evaluate the ability of the these changes in cognition that occur during normal
organism to know or to have a representation of its aging.
location in the environment and thus to navigate ef-
fectively. The intact functioning of the hippocampus Transgenic Mouse Models of Alzheimers
is necessary for learning and remembering spatial Disease
tasks. In old age, spatial memory is less efficient in hu-
mans and animals. Severe memory loss is the most prominent cogni-
tive symptom of Alzheimers disease (AD), and a fun-
When single cells are recorded in the hippocam- damental role in the pathogenesis of AD is brain de-
pus of a behaving rat, firing rate increases when the position of -amyloid (A). Mutations in the amyloid
animal is in a particular place in the environment. precursor protein (APP) and presenilin-1 (PS1) genes
These cells have been called place cells, and the are linked to forms of AD that are carried in families
area over which these cells show increased firing rates and called familial AD. By altering APP metabolism,
are called the cells place fields. Carol Barnes these mutations result in increased brain levels of A
(2001) described deficits that occur in the develop- peptide. Transgenic mice harboring mutant forms of
ment and maintenance of hippocampal place fields in the APP and/or PS1 gene associated with AD in hu-
old rats. As rats explore the environment, there is a mans are valid tools in the study of pathophysiologi-
change in the pattern of hippocampal place-cell dis- cal and behavioral effects of those genes in AD. Due
charge that occurs as a consequence of experience. to the relatively short life span of mice (two to three
For example, when rats run laps around a track in years), a high overexpression of the transgene is nec-
one direction, there is an expansion of the place fields essary to achieve the development of AD-like symp-
and a shift in their centers of mass toward the origin toms in the animals. The first transgenic mouse mod-
of the route. In old rats, there is a striking reduction els of AD were produced in the mid-1990s, and
in this experience-dependent form of plasticity in the thereafter there were a number of mouse models of
old place cells. Barnes suggested that the lack of field AD that developed A-containing plaques in the hip-
broadening observed in old rats might be expected to pocampus and neocortex, thus modeling human AD.
lead to a loss of precision in the information transmit- The hippocampus is engaged in eyeblink classical
ted as a consequence of experience. Ability to remem- conditioning and in spatial learning and memory,
ber routes may also be impaired by this deficit in and both of these behaviors are profoundly impaired
place-cell field broadening. in AD. Spatial learning and memory is the behavior
Barnes also found that the same memory- most commonly tested in transgenic mouse models of
impaired old rats that showed deficits in experience- AD, although eyeblink classical conditioning is a use-
dependent place-field expansion also retrieved inap- ful alternative behavioral measure that has direct par-
propriate hippocampal maps on some occasions. allels and can be tested in humans diagnosed with
Even when these memory-impaired old rats were in AD. A frequently used behavioral test for AD mouse
familiar environments, they retrieved inappropriate models is the Morris water maze, in which mice are
hippocampal maps from time to time. When a young placed in a pool of water that is opaque (to hide a plat-
rat is exposed to a familiar environment on one day form) and must learn the location of that platform to
and then exposed to that environment again a second escape from the water. Impairment in this and other
time later that day, the same place-field map will be forms of spatial learning and memory are observed in
recorded from hippocampal place cells on both ses- various transgenic mouse models of AD.
sions. Testing old rats in this two-session recording Dale Schenk and colleagues (1999) made a re-
procedure, Barnes and her colleagues (1997) ob- markable discovery that vaccinations with A peptide
served a bimodal distribution of responses. For two- can dramatically reduce amyloid deposition in trans-
thirds of the double-session recordings, the old rats genic mouse models of AD. A peptide vaccination
retrieved the same map on both occasions, perform- prevents spatial learning and memory loss (Morgan
ing normally as young rats. However, on one-third of et al., 2000). The long-term behavioral results of A
the double-session recordings, the old rats exhibited peptide vaccinations indicate that the behavioral pro-
a complete rearrangement of the place-field map be- tection of the vaccinations is task-specific, with preser-
tween the two sessions. They apparently retrieved the vation of hippocampal-associated spatial-memory
wrong map on one of the sessions. This failure to re- tasks most likely to occur (Arendash et al., 2001).
10 AGING AND MEMORY IN HUMANS
Evidence suggests that normal aging affects forty-five-month-old rabbits: Behavioral learning and hippo-
mammalian eyeblink conditioning through age- campal unit activity. Neurobiology of Aging 8, 101108.
related deficits in the cerebellum and hippocampus. Diana S. Woodruff-Pak
Age-related changes in spatial learning and memory
also rely on hippocampal mechanisms. AD exacer-
bates impairment in learning and memory and pro-
foundly disrupts hippocampal function early in its AGING AND MEMORY IN HUMANS
course. Transgenic models of AD provide a means to
test therapeutic interventions, such as vaccination One of the commonest complaints of older people is
with A peptide, that might protect against the cogni- that their memory is not what it used to be. The validi-
tive and neural impairment characteristic of this neu- ty of such subjective reports is borne out by the scien-
rodegenerative disease. tific literature: Memory performance does decline as
a function of the normal aging process in healthy
adults, although the decline is much more evident
See also: AGING AND MEMORY IN HUMANS with some materials and tasks than it is with others.
This variability has given researchers useful clues to
Bibliography the specific memory components or processes that
Arendash, G. W., Gordon, M. N., Diamond, D. M., Austin, L. A., are particularly vulnerable to the effects of aging.
Hatcher, J. M., Jantzen, P., DiCarlo, G., Wilcock, D., and Mor- Some of this work is described below.
gan, D. (2001). Behavioral assessment of Alzheimers trans- Nearly all the studies described use the cross-
genic mice following long-term A vaccination: Task specifici-
ty and correlations between A deposition and spatial sectional method of age comparison; that is, a group
memory. DNA and Cell Biology 20, 737744. of young adults (often college students in their early
Barnes, C. A. (2001). Plasticity in the aging central nervous system. twenties) is compared with a group of older adults
International Review of Neurobiology 45, 339355. (usually community-dwelling volunteers in their six-
Barnes, C. A., Suster, M. S., Shen, J., and McNaughton, B. L. ties and seventies). Additionally, some studies incor-
(1997). Cognitive map multistability in aged rat hippocam-
pus. Nature 388, 272275.
porate middle-aged groups of people in their forties
Chen, L., Bao, S., Lockard, J. M., Kim, J. J., and Thompson, R. F. and fifties. The cross-sectional method is much more
(1996). Impaired classical eyeblink conditioning in cerebellar practicable than within-subject longitudinal studies,
lesioned and Purkinje cell degeneration (pcd) mutant mice. but it does leave open the possibility that the differ-
Journal of Neuroscience 16, 2,8292,838. ences observed between the groups may be attribut-
Disterhoft, J. F., and McEchron, M. D. (2000). Cellular alterations
in hippocampus during acquisition and consolidation of hip-
able to causes other than agingto differences in ed-
pocampus-dependent trace eyeblink conditioning. In D. S. ucation or motivation, for example. Obviously
Woodruff-Pak and J. E. Steinmetz, eds., Eyeblink classical condi- researchers take pains to minimize the possibility of
tioning, Vol. 2: Animal models, pp. 313334. Boston: Kluwer Ac- such artifacts, and they do this by matching the
ademic Publishers. groups by educational level, by vocabulary (a rough
McEchron, M. D., Weible, A. P., and Disterhoft, J. F. (2001). Aging
and learning-specific changes in single-neuron activity in CA1
measure of verbal intelligence), and by other indica-
hippocampus during rabbit trace eyeblink conditioning. Jour- tors of intelligence and socioeconomic status. Many
nal of Neurophysiology 86, 1,8391,857. crucial experimental results take the form of interac-
Morgan, D., Diamond, D. M., Gottschall, P. E., Ugen, K. E., Dick- tions between age and some experimental variable;
ey, C., Hardy, J., Duff, K., Jantzen, P., DiCarlo, G., Wilcock, that is, one condition of the experiment is associated
D., Connor, K., Hatcher, J., Hope, C., Gordon, M., and Aren-
dash, G. W. (2000). A peptide vaccination prevents memory with large age-related differences, whereas another
loss in an animal model of Alzheimers disease. Nature 408, condition is associated with much smaller differences,
982985. even though the same subjects are used. The finding
Schenk, D., Barbour, R., Dunn, W., Gordon, G., Grajeda, H., of such differential effects makes it harder to argue
Guido, T., Hu, K., Huang, J., Johnson-Wood, K., Kahn, K.,
that group differences are a function, say, of reduced
Kholodenko, D., Lee, M., Liao, Z., Lieberburg, I., Motter, R.
Mutter, L., Soriano, F., Shopp, G., Vasquez, N., Vandevert, motivation in the older sample.
C., Walker, S., Wogulis, M., Yednock, T., Games, D., and Seu- Like other experimental explorations of individ-
bert, P. (1999). Immunization with amyloid- attenuates Al-
ual differences in memory ability, the work on aging
zheimer-disease-like pathology in the PDAPP mouse. Nature
400, 173177. has been carried out within various theoretical frame-
Thompson, R. F. (1988). The Rosetta stone for brain substrates of works and with a view to establishing some theoretical
age-related deficits in learning and memory? Neurobiology of point. In this brief article it is not possible to go into
Aging 9, 547548. details of theoretical motivation in most cases, but, in
Woodruff-Pak, D. S., Cronholm, J. F., and Sheffield, J. B. (1990).
summary, four main age-related changes that argu-
Purkinje cell number related to rate of eyeblink classical con-
ditioning. NeuroReport 1, 165168. ably underlie changes in memory performance are a
Woodruff-Pak, D. S., Lavond, D. G., Logan, C. G., and Thompson, decline in general processing speed; a decline in pro-
R. F. (1987). Classical conditioning in three-, thirty-, and cessing resources or attentional energy (Craik and
AGING AND MEMORY IN HUMANS 11
Byrd, 1982); an age-related reduction in the efficien- of time and place are crucial. There are different
cy of inhibitory processes; and an impairment in the views on whether episodic and semantic memory are
executive control of cognitive processing. These four different memory systems or whether they simply
theoretical viewpoints, along with supporting evi- exemplify different degrees of abstraction from the
dence, are discussed in Kester, Benjamin, Castel, and original events that gave rise to the encoded knowl-
Craik (2002); and in Zacks, Hasher, and Li (2000). edge (Craik, 2002). Whatever the resolution of this
The present article focuses on empirical work since debate, older adults typically show fewer losses in se-
the 1980s, pointing out the implications for theory mantic memory than in episodic memory, provided
where relevant. that the knowledge in question is used on a regular
basis. Thus, older people show minimal losses in vo-
cabulary and in knowledge of facts and concepts.
Indirect Memory Tests On the other hand, some aspects of semantic
When we think about memory, it is usually in the memory do appear to decline with agethe ability to
sense of the conscious retrieval of a past event or the learn completely new facts, for example. Even well-
retrieval of a previously learned fact. However, there learned material can be more difficult to retrieve for
are also many cases in which previous events can af- older adults, and difficulty in remembering names is
fect present behavior in the absence of conscious possibly the most frequent age-related memory com-
awareness of the past event in question. Such cases of plaint. It is unclear, however, whether names show a
implicit memory are revealed by indirect memory disproportional impairment with age (Maylor, 1997);
testsexamples include word-stem completion and a more general statement might be that older adults
word-fragment completion. In word-stem comple- have sporadic difficulty in retrieving any information
tion, the participant is given the first few letters of a that they use infrequently. Having a word or name
word (e.g., MAR___ or DRA___) and asked to com- on the tip of the tongue is an experience that occurs
plete the stem with the first word that comes to mind. more frequently as people grow older, and older
Word-fragment completion is similar; here the partic- adults report less partial information about the target
ipant is given words with letters missing and is asked word. In addition, speed of retrieval slows with ad-
to complete them (e.g., M _ _ K _ T or _ R A _ _ R). vancing age, in line with the slowing of many other
The general finding in such experiments is that par- cognitive processes. In summary, then, whereas the
ticipants can complete the words more readily if they representation of learned knowledge remains reason-
have studied the target words (e.g., MARKET, ably intact into old age, older adults experience occa-
DRAWER) previously, even though the person may sional difficulty in assessing that knowledge.
be quite unaware that the words are drawn from the
studied list. This phenomenon is known as prim-
ing, and it has typically been found that age differ- Episodic Memory
ences in tests of priming are much smaller than those Episodic memory refers to the ability to recollect
found in direct tests (Fleischman and Gabrieli, 1998). specific events, and this form of memory has been ex-
A similar discrepancy between direct and indirect tensively studied in the laboratory using tests of recall
tests is found in amnesic patients, who do poorly on and recognition. Many such studies have used rather
explicit or direct tests but comparatively well on im- artificial materialslists of unrelated words, for ex-
plicit or indirect tests. These results are sometimes ample. The benefit has been greater experimental
taken as evidence that encoding processes are there- control over encoding and retrieval processes, but a
fore intact in elderly individuals and amnesic pa- possible drawback is that the principles emerging
tients, and that the observed decrements are failures from such studies might not apply to real-life remem-
of retrieval. But encoding processes may be somewhat bering. Work carried out beginning in the 1980s on
impaired in older people and amnesics; they might be autobiographical memory has allayed these fears to a
sufficient to support later indirect tests, but insuffi- large extent, however. In these studies, memory for
cient to support later direct tests. personally experienced events in the subjects life
have shown that real-life memories are affected by the
same factors and subject to the same laws as are mate-
Semantic Memory rials learned in the laboratory. Age-related differ-
Semantic memory refers to a persons learned ences in autobiographical memory have been studied
knowledge of facts and concepts. Typically we are un- by presenting people of different ages with cue words
aware of where and when we first learned that Paris such a flag or school and asking participants to gener-
is the capital of France, the meaning of rhinoceros, and ate a personal memory that each word evokes. For
that 7 x 9 = 63; the notion of semantic memory thus adults of all ages, recent memories were generated
stands in contrast to episodic memory, where details most often, and the incidence of recollected events
12 AGING AND MEMORY IN HUMANS
declined from the present to the past. One interesting gion is associated with the processing of semantic in-
exception to this general trend is that all adults formation and with good later memory of processed
showed a disproportionate bump of generated items. Other studies have shown that, whereas young
memories from late teenage and early adult years adults activate regions that are predominantly lateral-
(Rubin, Wetzler, and Nebes, 1986). Presumably this ized in the right prefrontal cortex during retrieval,
bump reflects the many important career-related and older adults show activations in both right and left
emotionally significant events that occur during this prefrontal regions when retrieving information. The
period of our lives. additional left-sided activations in older people may
reflect the brains attempt to compensate for the de-
Episodic memory for events occurring in the last
clining efficiency of regions that subserve retrieval in
few minutes, hours, or even weeks typically shows
young adulthood. An account of this exciting line of
large age decrements, and much of the effort of cog-
research is provided by Prull, Gabrieli, and Bunge
nitive aging researchers has been directed to under-
(2000).
standing the factors underlying this problem. One set
of findings points to an age-related decline in the effi- One further topic pertaining to episodic memory
ciency of retrieval processes. Several investigators is the ability to remember the source of encoded infor-
have shown larger age differences in tests of free re- mation, or details of the context in which an event oc-
call (in which participants must recall a list of words curred. Everyone has experienced the situation in
or a paragraph of text with no cues or reminders) which a persons face seems very familiar, and yet we
than in tests of recognition memory (in which partici- cannot say how we know the person. Usually it turns
pants must pick out the originally studied items from out that we have encountered the familiar person in
a mixed set of targets and distractors). Details of these a very different context from his or her habitual loca-
experiments are summarized by Kester, Benjamin, tion, leading George Mandler to refer to the experi-
Castel, and Craik (2002). The fact that older adults ence as the butcher on the bus phenomenon. A sim-
have difficulty recalling studied items but can recog- ilar failure of memory for context occurs when we
nize them suggests that the difficulty lies at the re- know a fact but cannot recollect whether someone
trieval stage, although recognition is usually the easi- told us the information or whether we read it in the
er test. It seems likely that there are also problems at newspaper or heard it on the radio. Older adults are
encoding, however. One piece of evidence supporting particularly prone to such failures to bind contextual
the encoding hypothesis is that, when encoding pro- information to the core aspects of the event, or to the
cesses are guided appropriately by means of ques- item of information (see Kester, Benjamin, Castel,
tions that emphasize the semantic aspects of the item, and Craik, 2002, for details). The effects are also seen
age-related differences are often reduced (Craik and during the output of information; for example, an
older person may be cued by a conversational com-
Jennings, 1992). One similarity between encoding
panions background to exclaim, Ahyou will be in-
and retrieval is thus that, when appropriate processes
terested to hear this! and then proceed to retell the
are supported or guided by additional information
same story that the hapless listener has endured many
(e.g., semantic orienting tasks at encoding, retrieval
times previously.
cues or a recognition test at retrieval), age differences
are typically reduced. This pattern of findings led What lies behind the phenomenon of source for-
Craik (1983) to suggest that, whereas unsupported getting? Some researchers have linked the failure to
encoding and retrieval processes are often inefficient bind contextual details to core features to an ineffi-
in older people, possibly due to a reduction in avail- ciency of frontal lobe function. Other cognitive
able processing resources, these inefficiencies can be neuroscientists have suggested that the hippocampus
overcome through the environmental support pro- is centrally concerned with such binding functions
vided by the experimental situation or by the context (Prull, Gabrieli, and Bunge, 2000). At the behavioral
of a persons familiar surroundings. level, source forgetting may be viewed as one instance
of a general age-related difficulty of association or in-
The conclusion that age-related difficulties in ep- tegration. Naveh-Benjamin (2000) lays out convinc-
isodic memory are consequences of impaired processing evidence that older adults have particular difficul-
ing at both encoding and retrieval is supported by ty remembering the associative links between items of
findings from studies of functional neuroimaging. information, although the items themselves may be
Studies using PET (positron emission tomography) remembered quite well.
and fMRI (functional magnetic resonance imagery)
have shown that older adults exhibit less activation of
the ventral left prefrontal cortex during memory en- Short-Term and Working Memory
coding than do their younger counterparts. Many The phrase short-term memory has unfortunately
studies have demonstrated that this left prefrontal re- been used in a number of slightly different ways, and
AGING AND MEMORY IN HUMANS 13
this can give rise to confusions about findings. Clini- Prospective Memory
cians typically use the phrase to mean recent memo- Prospective memory refers to the situation in
ryevents that have happened in the last few hours which a person intends to carry out some action at a
or dayswhereas experimental psychologists have future time and then either performs the action suc-
used the phrase to refer to information still held in cessfully or forgets to perform it. Such situations are
mind, as when we look up a telephone number and common in everyday life, as are failures of prospec-
rehearse the information until we have dialed it. This tive memoryforgetting to make a phone call, mail
latter type of memory (sometimes also referred to as a letter, or to pass on a message, for example. Re-
primary memory) shows very little decline with age. searchers have made the useful distinction between
Similarly, memory spanthe longest list of digits or event-based and time-based prospective memory, the
words that a person can repeat back accurately first referring to situations in which the intended ac-
declines only slightly from the twenties to the tion should be cued by an event, such as seeing the let-
eighties. If short-term memory is used in the first sense, ter to be mailed or meeting the colleague for whom
however, it falls into the general category of episodic the message was intended. On the other hand, time-
memory, and substantial age-related decrements are based prospective actions are cued by times: I should
found, as discussed in the previous section. call home at 3:30 P.M., for example.
The term working memory has been adopted wide- Prospective memory failure generally increases
ly to refer to information held and manipulated in with age. As one example, Mntyl and Nilsson
mind. Thus, solving a verbal problem or performing (1997) reported a study in which participants were
mental arithmetic is considered to involve working asked to remind the experimenter at the end of a test-
memory (WM). In this sense WM incorporates execu- ing session that they should sign a form. Successful
tive processes as well as relatively automatic auxiliary performance dropped from 61 percent of partici-
systems such as the articulatory loop and visuo-spatial pants aged thirty-five to forty-five to only 25 percent
sketchpad (Baddeley, 1986). With regard to aging, of participants aged seventy to eighty. Other research
performance relying largely on auxiliary systems demonstrates that older adults do worse on time-
holds up well; memory span falls into this category. based than on event-based, prospective memory
But when good performance requires executive pro- tasks, arguably because the former type of task is less
cesses or complex manipulations of information held well supported by environmental cues (Craik, 1983).
in mind, then older people typically do less well than Finally, older people do better on real-life prospec-
their younger counterparts (Craik and Jennings, tive memory tasks than on laboratory-based tasks
1992; Zacks, Hasher, and Li, 2000). It seems possible (Kester, Benjamin, Castel, and Craik, 2002). This
that this age-related decline reflects the reduced effi- finding may reflect greater motivation on the part of
ciency of frontal lobe processes in older adults older adults, or it may reflect their greater use of daily
(Glisky, Polster, and Routhieaux, 1995). structures and routines.
Older adults often have difficulty dealing with
dual-task situations in which they must divide their at-
Summary
tention between two simultaneous activities. In one
such demonstration Anderson, Craik, and Naveh- Memory performance does decline with age, but
Benjamin (1998) found that, when younger and older the decline is greater in some tasks than in others.
adults divided their attention between a memory task Performance is often poor on episodic memory tasks,
and an ongoing reaction-time task, memory perfor- especially if the person must recall the information
mance dropped equally for the younger and older without cues. Performance is also comparatively poor
groups (relative to performing the memory task on its on source memory, on prospective memory, and on
own) but that performance on the reaction-time task working memory tasks. On the other hand, memory
dropped much more for older than for younger for general knowledge and for routine activities holds
adults, especially during the retrieval phase of the up well with age, as does primary memory for infor-
memory task. In a similar demonstration, Lindenber- mation held briefly in mind. Finally, environmental
ger, Marsiske, and Baltes (2000) measured walking support from familiar surroundings can be particu-
accuracy and speed in adults of different ages while larly helpful as an aid to remembering in older adults.
they learned a list of words. They found greater dual-
task costs in the older group, partly reflecting the in- See also: AGING AND MEMORY IN ANIMALS;
creased need for executive processes in word learning AUTOBIOGRAPHICAL MEMORY; EPISODIC
but also partly reflecting the older adults greater MEMORY; INDIVIDUAL DIFFERENCES IN
need to deploy executive processes to control accu- LEARNING AND MEMORY; TIP-OF-THE-TONGUE
rate walking. PHENOMENON; WORKING MEMORY: HUMANS
14 ALGORITHMS, LEARNING
Bibliography formance on the basis of its own experience. Also

Anderson, N. D., Craik, F. I. M., and Naveh-Benjamin, M. (1998). called learning procedures, methods, or rules, learn-
The attentional demands of encoding and retrieval in youn- ing algorithms are key components of mathematical
ger and older adults: 1. Evidence from divided attention
models of animal learning and of technological de-
costs. Psychology and Aging 13, 405423.
Baddeley, A. D. (1986). Working memory. London: Oxford Univer- vices engineered to improve their behavior as they
sity Press. operate. Learning algorithms have been studied ex-
Craik, F. I. M. (1983). On the transfer of information from tempo- tensively in artificial intelligence, psychology, compu-
rary to permanent memory. Philosophical Transactions of the tational neuroscience, statistics, and engineering.
Royal Society of London, ser. B, 302, 341359.
They are used in many applications in science, indus-
(2002). Human memory and aging. In L. Bckman and C.
von Hofsten, eds., Psychology at the turn of the millennium, Vol. try, and finance that require fitting models to or find-
1: Cognitive, biological, and health perspectives, pp. 261280. ing patterns in collections of data or that require
Hove, Sussex, UK: Psychology Press. learning skills needed to solve specific tasks. Learning
Craik, F. I. M., and Byrd, M. (1982). Aging and cognitive deficits: algorithms play major roles in artificial neural net-
The role of attentional resources. In F. I. M. Craik and S. Tre-
work systems, where they provide rules for adjusting
hub, eds., Aging and cognitive processes, pp. 191211. New
York: Plenum. the strengths, or weights, of connections between net-
Craik, F. I. M., and Jennings, J. M. (1992). Human memory. In F. work elements. These connections correspond to syn-
I. M. Craik, and T. A. Salthouse, eds., The handbook of aging apses by which neurons communicate with other neu-
and cognition, pp. 51110. Hillsdale, NJ: Erlbaum. rons and with sensory and motor mechanisms.
Fleischman, D. A., and Gabrieli, J. D. E. (1998). Repetition prim-
ing in normal aging and Alzheimers disease: A review of find-
ings and theories. Psychology and Aging 13, 88119.
Glisky, E. L., Polster, M. R., and Routhieaux, B. C. (1995). Double
Hebbian Learning
dissociation between item and source memory. Neuropsy- In 1949 the psychologist Donald Hebb hypothe-
chology 9, 229235. sized that when a neuron, A, repeatedly and persis-
Kester, J. D., Benjamin, A. S., Castel, A. D., and Craik, F. I. M.
tently takes part in firing another neuron, B, the syn-
(2002). Memory in the elderly. In A. Baddeley, B. Wilson, and
M. Kopelman, eds., Handbook of memory disorders, 2nd edition. apses by which A stimulates B are strengthened. It
London: Wiley. then becomes easier for neuron A to fire neuron B.
Lindenberger, U., Marsiske, M., and Baltes, P. B. (2000). Memoriz- Consequently, any event that stimulates neuron A
ing while walking: Increase in dual-task costs from young may also stimulate neuron B and thus become associ-
adulthood to old age. Psychology and Aging 15, 417436.
ated with the consequences of Bs firing as well as with
Mntyl, T., and Nilsson, L. G. (1997). Remembering to remember
in adulthood: A population-based study on aging and pro- other events that stimulate B. This is one of the earli-
spective memory. Aging, Neuropsychology, and Cognition 4, 81 est and most influential ideas contributing to neural-
92. network learning algorithms.
Maylor, E. A. (1997). Proper name retrieval in old age: Converging
evidence against disproportionate impairment. Aging, Turning this hypothesis into a complete learning
Neuropsychology, and Cognition 4, 211266. algorithm requires precise definitions of the variables
Naveh-Benjamin, M. (2000). Adult age differences in memory per- involved and how they interact. One of the simplest
formance: Tests of an associative deficit hypothesis. Journal of ways to do this is to consider the situation in which
Experimental Psychology: Learning, Memory, and Cognition 26,
neuron B receives input from other neurons, labeled
1,1701,187.
Prull, M. W., Gabrieli, J. D. E., and Bunge, S. A. (2000). Age-related 1, 2, . . . , n, any of which can play the role of neuron
changes in memory: A cognitive neuroscience perspective. In A in Hebbs hypothesis. Let XB(t) be a positive num-
F. I. M. Craik and T. A. Salthouse, eds., The handbook of aging ber representing the activity level (the instantaneous
and cognition, 2nd edition, pp. 91153. Mahwah, NJ: Erlbaum. rate of firing) of neuron B at time t. Similarly, for i =
Rubin, D. C., Wetzler, S. E., and Nebes, R. D. (1986). Autobio-
1, 2, . . . , n, let Xi(t) represent the activity levels at
graphical memory across the adult life-span. In D. C. Rubin,
ed., Autobiographical memory, pp. 202221. Cambridge, UK: time t of neurons 1, 2, . . . , n (see Figure l). In repre-
Cambridge University Press. senting how the activity of neuron B depends on the
Zacks, R. T., Hasher, L., and Li, K. Z. H. (2000). Human memory. activities of the other neurons, the simplest assump-
In F. I. M. Craik and T. A. Salthouse eds., The handbook of tion is that XB(t) is a weighted sum of the activities of
aging and cognition, pp. 293357. Mahwah, NJ: Erlbaum.
the other neurons, where each weight represents the
Fergus I. M. Craik current strength of a synapse. If Wi(t) is the strength
at time t of the synapse by which neuron i influences
neuron B, this means that
XB(t) = W1(t)X1(t) + . . . + Wn(t)Xn(t). (1)

ALGORITHMS, LEARNING
Learning algorithms are sets of rules, usually ex- Hebbs hypothesis suggests how the synaptic
pressed using mathematical equations or computer strengths change over time, depending on the other
instructions, that enable a system to improve its per- variables. The extent to which neuron i takes part in
ALGORITHMS, LEARNING 15
firing neuron B at a time t is often represented by the Figure 1

product of the activity levels of neuron B and neuron
i: XB(t) Xi(t). This product is large when the activity
levels of both neurons are high, and it is small when
the activity level of one or both neurons is close to
zero. Thus, if neuron i persistently takes part in firing
neuron B, this product will be large for many times
t. This leads to a rule for changing synaptic strengths
that can be written for each synapse i as:
Wi(t + t) = Wi(t) + cXB(t)Xi(t), (2)
where t is a small time increment and c is a small pos-

itive number. Selecting values for t and c determines
how rapidly the synaptic strengths change. According
to Equation 2, at any time, t, when the activity levels
of neurons i and B are both greater than zero, XB (t)Xi
(t) is greater than zero, and the synaptic strength Wi Neurons 1, 2, . . . n provide input to neuron B.
(t) increases from time t to time t + t. According to
Equation 1, this larger synaptic weight means that
neuron is activity will contribute more to the activity tative and categorical outcomes). Solving it requires
of neuron B in the future. Equations 1 and 2 consti- finding a rule (mathematically, a function) that best
tute a learning algorithm; both are needed to com- fits the data in the training set (where best is pre-
pute the changes in the synaptic strengths when the cisely defined in some way). The resulting rule is then
values Xi (t), i = l, 2, . . . , n, are given for each time used to provide the desired predictions. Many meth-
t. This is the simplest of many learning algorithms ods for solving these problems are formulated as
based on Hebbs hypothesis, and, despite many short- learning algorithms. The training set represents the
comings, it has played an important role in theories experience of the learning system, and the rule being
of learning in neural networks. Brown et al. (1990) learned improves in its ability to provide valid predic-
discuss this and many other Hebb-inspired learning tions as the data is processed. Learning researchers
algorithms and how well they model synaptic proper- think of the training set as a set of examples provided
ties actually observed in regions of the brain such as by a teacher, and they call this process learning by
the hippocampus. example, learning with a teacher, or, most often, su-
pervised learning. Many learning algorithms have
been devised for supervised learning problems, and
Statistical Learning
modern research on the theoretical properties of
Other learning algorithms are motivated by a de- these algorithms is strongly tied to the field of statis-
sire to solve problems that occur frequently in a vari- tics. Hastie et al. (2001) is a good reference for the sta-
ety of fields. One type of problem that has received tistical view of supervised learning algorithms.
much attention, especially in the field of statistics, is
that of function fitting. Suppose you observe the in-
puts and outputs of some system under study and as- Least Mean Square Algorithm
semble a training set of observations (Xi, Zi), i = 1, 2, The least mean square (LMS) algorithm is an in-
. . . , m, where Zi is the systems output for input Xi. fluential supervised learning algorithm proposed by
(The superscripts are used to avoid confusion with the the electrical engineers Bernard Widrow and Marcian
neuron-activity levels used above.) On the basis of Hoff in 1960. Like the Hebbian algorithm described
these data, you would like to predict what the systems above, it can be expressed as a rule for changing the
output will be for new inputs. For example, the system synaptic weights of a neuronlike element, but it re-
might be a medical treatment, with inputs giving fea- quires another variable to provide training input to
tures of patients (such as the results of clinical tests) the element. This signal tells the element what its ac-
and outputs describing treatment outcome (which can tivity should be; it is said to provide the desired out-
be a quantitative measure or a nonnumeric, or cate- puts or target outputs. During learning, input signals
gorical, description, such as recovery /no-recovery). and target outputs are selected from a training set of
You would like to predict the outcomes of treating example input/output pairs. If Z(t) denotes the target
new patients. output at time t for an element whose actual output
In statistics, this is known as a regression or a clas- is X(t), then Z(t) X(t) gives the error in the elements
sification problem (for the respective cases of quanti- output: the discrepancy between what the elements
16 ALGORITHMS, LEARNING
activity should be and what it actually is. The LMS al- will eventually find a local minimum, meaning a point
gorithm is an error-correction algorithm because it that is better than all points in the immediate neigh-
changes synaptic weights to reduce the sizes of the er- borhood, although a better point may exist. When the
rors gradually over time. elements output is a linear function of its weights, as
given by Equation 1, all local minima are also global
If we suppose that X(t) is a weighted sum of inputs
minima, meaning that the error surface does not get
as given by Equation 1, then the LMS algorithm says
any lower than at these points. This means that the
that a synaptic weight i changes as follows:
LMS algorithm finds a set of weights that produces
the least mean-square error over the relevant inputs,
Wi(t + t) = Wi(t) + c [Z(t) X(t)]Xi(t). (3) thus justifying the algorithms name.
This means that when the elements output, X(t),

is too low, the error is positive and the synaptic Error Backpropagation
weight, Wi(t), increases (assuming both c and Xi(t) are A generalization of the LMS algorithm known as
positive). On the other hand, when the elements out- the error back-propagation algorithm has been influ-
put is too high, the error is negative, and the weight ential in artificial neural network research. (Haykin
decreases. In either case, this change in synaptic [1999] provides details about this and other network
weight tends to make the elements output more learning algorithms.) In its simplest form, this is a
nearly equal to the target output when input element gradient-descent algorithm that applies to networks
i is active in the future. of neuronlike elements that are connected in multiple
The LMS algorithm is closely related to several layers and that are capable of computing nonlinear
other error-correction algorithms. The perceptron functions of the networks input. An error for each el-
learning algorithm differs from the LMS algorithm ement is computed by propagating the error of the
only in using an error that is sensitive to the differ- network as a whole back through the network in the
ence in the signs of the target and actual outputs but direction opposite to the flow of element activity.
is not sensitive to the difference in their sizes. The Mathematically, this process computes the gradient
LMS algorithm is also nearly identical in mathemati- of the error surface at the point corresponding to all
cal form to an influential model of Pavlovian, or clas- the current weights in the network. Unlike the linear
sical, conditioning proposed by psychologists Robert LMS algorithm, the error back-propagation algo-
Rescorla and Allan Wagner (1972). Instead of rithm is not guaranteed to find a best set of weights
changes in synaptic strengths, this model defines because the error surface can have a complex topog-
changes in behavioral variables called associative raphy with many local minima.
strengths that link conditioned stimuli with an uncon-
ditioned response. The LMS algorithm is the basis of
Other Learning Algorithms
many models of the cerebellum, where climbing fi-
bers may provide the error signals. A learning algorithm is unsupervised, or self-
organizing, if its experience comes from a training set
containing inputs but no target outputs. Instead of
Gradient Descent trying to match target outputs, it tries to recode the
Theories of the LMS algorithm and many other inputs according to some built-in principle. For ex-
learning algorithms depend on viewing the learning ample, some unsupervised learning algorithms re-
process as a search for a set of weights that is best ac- code input signals in order to reduce their complexity
cording to some measure of the algorithms perfor- while trying to preserve their information content.
mance. For the LMS algorithm, this measure assigns Often this takes the form of learning how to form
to each possible set of weights the average, or mean, clusters of input signals so that signals within a cluster
of the squares of the errors that would result from are more similar to one another than they are to sig-
using that set of weights for all relevant inputs. If we nals in other clusters. The Hebbian learning algo-
visualize this measure as a surface over weight rithm is most often used for unsupervised clustering
space, the error measure for the current weights corin networks where elements compete for the opportu-
responds to a point on this surface, and the LMS algo- nity to represent different inputs.
rithm changes the weights by moving this point down Reinforcement learning algorithms rely on train-
the slope of the surface, thus improving the systems ing input called reward signals, which evaluate the
ability to match the target outputs. This is called a quality of the learning systems behavior without ex-
gradient-descent algorithm because the direction of plicitly telling the system what its outputs should be.
steepest slope is given mathematically by the gradient Reinforcement learning algorithms have to discover
of the error measure. A gradient-descent algorithm how to improve their behavior by trying various out-
ALZHEIMERS DISEASE: Behavioral Aspects 17
puts and comparing the resulting evaluations. This is ALZHEIMERS DISEASE

a form of trial-and-error learning, which should not
be confused with error-correction learning, such as [Alzheimers disease was first described by the German physi-
that performed by the LMS algorithm, which requires cian Alois Alzheimer in 1906. The disease is associated with
target outputs. Many reinforcement learning systems a progressive decline in general cognitive function and a
use temporal difference (TD) algorithms to learn pre- specific impairment in the ability to form new memories (an-
dictions of the amount of reward a reinforcement terograde amnesia). Approximately 4 million individuals in
learning algorithm will accumulate over an extended the United States suffer from Alzheimers disease at an esti-
period. TD algorithms reduce errors in predictions mated cost for treatment and care that exceeds $100 billion
made at different times by adjusting earlier predic- per year. Two entries on Alzheimers disease follow. BEHAV-
IORAL ASPECTS describes the cognitive defects associated
tions to be closer to laterand therefore more reli-
ablepredictions. The computer scientists Richard with the disease and the ways in which it is diagnosed;
Sutton and Andrew Barto (1998) extensively discuss HUMAN DISEASE AND THE GENETICALLY ENGINEERED
reinforcement learning and TD algorithms. Physio- ANIMAL MODELS describes recent progress in elucidating
the genes and proteins implicated in the disease, the develop-
logical experiments by the neuroscientist Wolfram
ment of an animal model for the disease, and some exciting
Schultz (1998) reveal intriguing parallels between the
alternative treatment possibilities. For a discussion of treat-
behavior of TD algorithms and the activity of domp-
ments for Alzheimers disease see PHARMACOLOGICAL
amine-producing neurons in the brain.
TREATMENTS OF MEMORY DEFECTS.]
Conclusion
Learning algorithms have been devised for very BEHAVIORAL ASPECTS
different purposes: to model hypotheses about neural Alzheimers disease (AD) is the most common cause
mechanisms of learning, to model the learning be- of dementia in the elderly, affecting 50 percent to 70
havior of animals, and to solve important statistical percent of dementia patients. Dementia involves im-
and engineering problems. Despite these different pairments in daily functioning related to a progres-
purposes, the resulting algorithms show a remarkable sive decline in two or more areas of mental ability.
degree of similarity. Perhaps this is not so surprising These mental and functional deficits are traceable to
because animal behavior and its neural basis are na- neuritic plaques and neurofibrillary tangles, which
tures solutions to problems that have much in com- cause cell loss in the brain. As the disease progresses,
mon with those studied by statisticians and engineers. it affects ever larger areas of the brain. Eventually, so
many parts of the brain are involved that patients can-
not move around normally and feed themselves. They
Bibliography then become susceptible to other potentially fatal dis-
Brown, T. H., Kairiss, E. W., and Keenan, C. L. (1990). Hebbian
eases such as pneumonia. The disease can last from
synapses: Biophysical mechanisms and algorithms. Annual Re-
view of Neuroscience 13, 475511.
five to fifteen years. Many drugs are being developed
Hastie, T., Tibshirani, R., and Friedman, J. (2001). The elements of to treat it, but there is no way of preventing the pro-
statistical learning. New York: Springer-Verlag. gression of the disease.
Haykin, S. (1999). Neural networks: A comprehensive foundation.
Upper Saddle River, NJ: Prentice Hall. When the disease was first described by Alois Alz-
Hebb, D. O. (1949). The organization of behavior. New York: Wiley. heimer in 1906, it seemed to affect only persons
Rescorla, R. A., and Wagner. A. R. (1972). A theory of Pavlovian under the age of sixty-five. In the 1960s, researchers
conditioning: Variations in the effectiveness of reinforcement discovered during autopsies that patients whose se-
and non-reinforcement. In A. H. Black and W. F. Prokasy, vere cognitive decline had been attributed to diseases
eds., Classical conditioning, Vol. 2: Current research and theory,
in the blood vessels of the brain (cerebrovascular dis-
pp. 6499. New York: Appleton.
Schultz, W. (1998). Predictive reward signals of dopamine neurons.
ease or hardening of the arteries) in fact exhibited
Journal of Neurophysiology 80, 127. the pathological hallmarks of AD (neurofibrillary tan-
Sutton, R. S. and Barto, A. G. (1998). Reinforcement learning: An in- gles and neuritic plaques). The realization that AD af-
troduction. Cambridge, MA: MIT Press. fects persons of all ages is, therefore, a recent one. Be-
Widrow, B., and Hoff, M. E. (1960). Adaptive switching circuits. In cause of the rarity of the disease among those under
1960 IRE WESCON convention record, pp. 96104. New York:
sixtyless than 2 percentit used to be considered
IRE. Reprinted in J. A. Anderson and E. Rosenfeld, eds.,
Neurocomputing: Foundations of research. Cambridge, MA: MIT
an uncommon disorder. Now it is understood that the
Press, 1988. likelihood of contracting AD increases with age.
Among people sixty-five to eighty-five, its prevalence
Andrew G. Barto is 5 percent to l5 percent; 30 to 47 percent of those
18 ALZHEIMERS DISEASE: Behavioral Aspects
over eighty-five have AD. It is the fourth-leading the concentration of a number of neurotransmitters,
cause of death among those sixty-five or older. the chemicals that are responsible for the transmis-
The earliest symptom of AD in most patients is se- sion of nerve signals in the brain. One neurotransmit-
vere difficulty in learning new information (i.e., an- ter, acetylcholine, which is important for normal
terograde memory impairment), especially increas- human memory, declines by up to 70 percent in cases
ing forgetfulness about day-to-day events. First, of severe impairment. The combination of these
patients may forget a recent event from one week to structural and chemical changes seems to be responsi-
the next, then from one day to the next, and finally ble for the unique severity of anterograde memory
from one minute to the next. AD impairs nearly all as- impairment in AD.
pects of new learning. It is not, for example, limited The cognitive impairments of AD are not con-
to information that the patient is trying to learn (i.e., fined to memory. These other difficulties interact
explicit or episodic memory), as is true in some amne- with the memory impairment to increase its severity.
sic disorders. Patients also have difficulty with implicit Early on in the disease the other main area of impair-
memorylearning information that impinges on ment is the executive functions, which pertain to con-
consciousness in the absence of any effort at master- cept formation, directed attention, and the concur-
ing it. rent manipulation of information. Studies suggest
AD does not uniformly impair all implicit memo- that the early stages of AD present problems with the
ry tasks. The same is true for the ability to learn new concurrent manipulation of information and a conse-
general skills (i.e., procedural knowledge). Patients quent difficulty with tasks that require simultaneous
have difficulty on some kinds of skill learning tasks manipulation of several different components, even
but not others; for example, learning of motor skills when the individual components of the tasks are
is frequently spared. Early in the disease patients deeply ingrained (e.g., preparing meals, paying bills,
memory for remote events is usually intact. As the dis- balancing a checkbook, and so on). That is why it
ease progresses, however, remote memories also re- sometimes seems as if the patients have difficulty with
cede, but in reverse chronological order; the most re- remote memory early in the disease. However, when
cent memories evaporate first. Only the most severe the tasks are broken down into their constituent parts,
impairment obliterates memories of earliest child- patients can perform them.
hood. Recent evidence suggests that problems with ex-
The rapidity of anterograde memory impairment ecutive function are at least in part the result of pa-
in AD is dramatic: It occurs after a delay of ten min- thology in the central portion of the cingulate gyrus.
utes or less after exposure to new information. The Studies in animals demonstrate that damage to the
best way to reveal this rapid loss of information is to middle portion of the cingulate alters executive func-
give a patient something new to learnfor example, tion. Another possible cause of executive-function im-
a storyand to ask him/her to state how much is re- pairment may be the loss of the corticocortical fibers,
membered immediately, and then after a delay. The which connect different parts of the cortex to one an-
delay should be less than ten minutes, because after other. The partial degeneration of this intracortical
ten minutes most types of patients, and even normal projection system in the early stages of AD could im-
older persons, forget a substantial amount of infor- pair the performance that requires the rapid and si-
mation. For example, patients with other forms of de- multaneous integration of multiple types of informa-
mentia, such as Picks disease or Huntingtons dis- tion.
ease, also have difficulty learning and retaining new The further progression of AD often severs the
information, but if delays last less than ten minutes, meaningful associations among wordshence a loss
their retention is significantly better than that of AD of semantic knowledge or semantic memory. Patients
patients. If delays are longer than ten minutes, then with AD begin to have difficulty with a number of lin-
normal patients and those with other forms of de- guistic tasks when they are mildly-to-moderately im-
mentia do not differ significantly from one another. paired. These difficulties affect the ability to produce
Two types of changes in the brain are primarily the name of an object when shown the object itself or
responsible for the severe anterograde memory im- a picture of the object (confrontation naming). They
pairment of AD. The first is damage to the brain re- also affect the ability to produce rapidly a series of
gions in the medial temporal lobe that are essential words that belong to a particular category, such as
for normal memory, specifically the entorhinal cortex vegetables or words starting with the letter s (verbal
and the hippocampus. Even in very mildly impaired fluency).
patients, there is approximately a 30-percent neuro- Some researchers argue that AD involves a break-
nal loss in the entorhinal cortex, the primary input down in the structure of semantic knowledge and that
path to the hippocampus. There are also declines in patients actually lose knowledge they once possessed.
ALZHEIMERS DISEASE: Human Disease and the Genetically Engineered Animal Models 19
Others argue that semantic knowledge remains intact McKhann, G., Drachman, D., Folstein, M. F., Katzman, R., Price,
in mildly and moderately impaired patients but is D., and Stadlan, E. (1984). Clinical diagnosis of Alzheimers
disease. Report of the NINCDS-ADRDA Workgroup under
more difficult to access; thus, when one asks them to the auspices of Department of Health and Human Services
intentionally search their semantic memory for infor- Task Force. Neurology 34, 939944.
mation, they have difficulty, but if their semantic Petersen, R., Smith, G., Waring, S., Ivnik, R., Tangalos, E., and
knowledge is evaluated indirectly, it appears to be Kokmen, E. (1999). Mild cognitive impairment, clinical char-
preserved. It has been difficult to resolve this debate acterization and outcome. Archives of Neurology 56, 303308.
because research, although intensive, often yields Marilyn S. Albert
contradictory findings.
AD is still difficult to diagnose during life. No test
can definitively verify the onset of the disease short of HUMAN DISEASE AND THE GENETICALLY
examining brain tissue through a biopsy. Researchers ENGINEERED ANIMAL MODELS
have therefore developed a number of conventions to
Alzheimers disease (AD) represents a great challenge
communicate the degree to which patients have been
for science and medicine because of its prevalence,
examined and the diagnostic criteria that the patients
cost, lack of reliable treatments, and often devastating
meet. Patients who have been carefully examined and
impact on individuals and caregivers. This age-
who meet clinical research criteria but have not had
associated chronic illness involves genetic risk factors,
a brain biopsy are said to have probable AD, or proba-
a well-defined clinical syndrome with a progressive
ble dementia of the Alzheimer type. Patients who
course, evidence of dysfunction and/or death of popu-
have had dementia during life and in whom results of
lations of neurons, pathological and biochemical ab-
an examination of brain tissue meet pathological re-
normalities, and intra- or extra-cellular protein ag-
search criteria for the disease are said to have definite
gregates (Price, Tanzi, Borchelt, and Sisodia, 1998).
AD.
Patients become severely disabled and often die of in-
Because novel treatments are under investiga- tercurrent illnesses. There are treatments for symp-
tion, there has been increasing interest in determin- toms but no cure. However, recent research, particu-
ing the earliest possible diagnosis of AD. The term larly in animal models, has begun to provide new
mild cognitive impairment (MCI) has been developed to insights into the mechanisms of Alzheimers disease
describe individuals with evidence of a functional dif- and has identified new targets for therapy.
ficulty in daily life that does not warrant a diagnosis
of dementia. A large number of individuals with MCI
(approximately 15 percent per year) are later diag- Clinical-Pathological Features of
nosed with AD. Drug trials are underway to see if Alzheimers Disease
treating patients with MCI reduces the likelihood that In most cases of AD, the initial impairments of
they will be diagnosed with AD. memory and cognition appear gradually during the
AD is a common and serious disorder that is seventh decade. The accuracy of clinical diagnoses
caused by damage to the structure and function of the improved from 1980 to 2000, and early diagnosis will
brain. There are several hypotheses concerning the become increasingly important as mechanism-based
underlying cause of AD, but these remain unproved, treatments become available.
and there are no effective treatments. The disease AD involves the brain (and not other organs) and
produces a gradual decline in cognitive function. The certain neuronal populations are selectively vulnera-
most common early symptom is a dramatic loss of new ble. AD results from the selective degeneration of
information after a brief delay, but AD can impair all nerve cells in the brains regions and neural circuits
aspects of mental ability. that are critical for memory, cognitive performance,
and personality (Albert, 1996). The dysfunction and/
See also: DEMENTIA; PHARMACOLOGICAL
TREATMENTS OF MEMORY DEFICITS or death of these neurons reduces the numbers of ge-
neric and transmitter specific-synaptic markers in
Bibliography their target fields; the disruption of synaptic commu-
Albert, M., and Moss, B. (1999). Early features of Alzheimers dis- nication in affected regions/circuits can lead to men-
ease. In A. Peters and J. Morrison, eds. Cerebral Cortex, Vol. 14, tal impairments and, finally, severe dementia.
pp. 461471, New York: Plenum.
Gomez-Isla, T., Price, J., McKeel, D., Morris, J., Growdon, J., and AD typically involves intracellular or extracellular
Hyman, B. (1996). Profound loss of layer II entorhinal cortex protein aggregates in brain. Neurofibrillary tangles
neurons occurs in very mild Alzheimers disease. Journal of (NFTs), inclusions located within cell bodies and
Neuroscience 16, 4,4914,500.
Katzman, R. and Kawas, C. (1984). The epidemiology of dementia proximal dendrites, are composed of poorly soluble
and Alzheimers disease. In R. D. Terry, R. Katzman, and K. paired helical filaments (PHF), which are, in turn,
L. Bick, eds., Alzheimer disease. New York: Raven. composed principally of hyperphosphorylated isofor-
20 ALZHEIMERS DISEASE: Human Disease and the Genetically Engineered Animal Models
ms of tau. PHF are also present in dystrophic neu- thought to be either tumor necrosis factor (TNF)
rites, the filamentous swellings of distal axons/ converting enzyme (TACE) or a disintegrin and met-
terminals (usually seen in proximity to A deposits). alloproteinase 10 (ADAM 10), which cut between resi-
Perturbations related to hyperphosphorylated tau dues 16 and 17 of A, and by BACE2, a protease shar-
seem to play a role in disturbances in intracellular ing features with BACE1, but cleaving APP after
transport. residues 19 and 20 of A (i.e., within the A domain).
The extracellular aggregates in brain are abnor- These different endoproteolytic cleavages generate
mal accumulations of A, a 4kD pleated sheet amy- various C-terminal peptides, including the APP intra-
loid peptide, derived by - and -secretase cleavages cellular domain (C60), which may play a role in the
of the amyloid precursor protein (APP). Levels of A activation of transcription.
are elevated in brain, and A monomers form oli- A variety of APP mutations, including APPswe (a
gomers and multimers that assemble into protofila- double mutation at the N-terminus of A) and APP-
ments and then fibrils. Eventually, A fibrils are de- 717 (near the C-terminus of A), have been reported
posited as the amyloid cores of neuritic or senile in cases of FAD. These mutations, strikingly situated
plaques, which are complex structures also containing near several secretase cleavage sites, are proamyloi-
dystrophic neurites, astrocytes, and microglia. dogenic, and cells that express mutant APP show ab-
Plaques are preferentially localized to the cortex, hip- errant APP processing: the APPswe mutation, which
pocampus, and amygdala. enhances BACE1 cleavage, is associated with elevated
The levels and distributions of APP and its cleav- levels of A; the APP 717 mutations, which affect -
age enzymes in neurons lead to the selective appear- secretase activity, lead to a higher secreted fraction of
ance of A in brain. It seems that toxic A peptides, longer, more toxic A peptides (A42) relative to cells
particularly oligomers, accumulate near synapses and that express wild-type APP (Price, Tanzi, Borchelt,
may impair transsynaptic communication, eventuat- and Sisodia, 1998; Citron et al., 1992).
ing in the disruption of synaptic connections between
PS1 and PS2
neurons and their targets (other nerve cells). Nerve
cells are functionally damaged, changes occur in tau Localized to chromosomes 14 (PS1) and 1 (PS2),
phosphorylation; microtubule stability is compro- respectively, these genes encode highly homologous
mised; intracellular transport processes are impaired; 43- to 50-kD proteins with multiple transmembrane
intracellular PHF appear; and cell geometry is al- (TM) domains. Oriented toward cytoplasm are a hy-
tered, with synapses, axon terminals, and dendrites drophilic acidic loop region, an N-terminal region,
appearing to be most vulnerable. The neuron is inca- and a C-terminal domain. PS1 is synthesized as an
pable of performing its normal functions for a signifi- 42- to 43-kD polypeptide, but the preponderant
cant interval before the ensuing demise of the cell. PS1-related species that accumulate in vitro and in
vivo are 27- to 28-kD N-terminal and 16- to 17-kD
C-terminal derivatives, which accumulate and/or as-
Mutant Genes/Proteins Implicated in sociate in a 1:1 stochiometry and are stable, tightly
Familial Alzheimers Disease (FAD) regulated, and saturable. PS genes are widely ex-
In some individuals with early onset AD, the ill- pressed at low abundance in the CNS.
ness may be inherited as an autosomal dominant with PS1 influences APP processing, but it is not clear
mutations in three different genes: the APP; PS1; and whether PS1 itself is an aspartyl protease (i.e., -
PS2 (Price, Tanzi, Borchelt, and Sisodia, 1998). secretase), is a cofactor critical for the activity of -
APP secretase, or plays a role in trafficking of APP to the
proper compartment for -secretase cleavage (De
Encoded by a gene on chromosome 21, APP is ex-
Strooper et al., 1998).
pressed in many cells and tissues but is particularly
abundant in neurons. This type-1 transmembrane The PS1 gene harbors more than fifty different
protein is cleaved endoproteolytically by an enzyme, FAD mutations in more than eighty families, whereas
-site APP-cleaving enzyme 1 (BACE1), and by an ac- only a small number of mutations have been found in
tivity termed -secretase, which, in concert, gener- PS2-linked families. The vast majority of the abnor-
ate the N- and C-termini of the A peptide, respec- malities in PS genes are missense mutations that re-
tively. The levels and distributions of APP and the sult in single amino-acid substitutions. However, re-
activities of proamyloidogenic cleavage enzymes, par- searchers have found a mutation that deletes exon 9
ticularly BACE1, in neurons seem to be lead to the from PS1 in several different FAD families. The vari-
formation of A in brain. The formation of A 1- ous PS mutations appear to influence -secretase ac-
40,42 is precluded by the endoproteolytic cleavage of tivity and increase the generation of the A 42 pep-
APP within the A sequence by -secretase, now tide.
ALZHEIMERS DISEASE: Human Disease and the Genetically Engineered Animal Models 21
Transgenic Models of A Amyloidogenesis Consistent with this idea are observations that
APLP2-/- mice appear normal but that mice with ei-
Some of the lines of mutant APP mice, although
ther both APP and APLP2 targeted alleles or both
they do not reproduce the full phenotype of AD, rep-
APLP1 and APLP2 null alleles show significant post-
resent excellent models of A amyloidosis and are of
natal lethality.
great value for testing causal effects of mutant genes,
analyses of pathogenic pathways, determination of BACE1-/- Mice
the molecules participating in A amyloidogenesis,
and identification of therapeutic targets. What follows These null mice are viable and healthy, have no
is a review of selected examples of lines of mice ex- obvious phenotype or pathology, and can mate suc-
pressing autosomal dominant FAD-linked mutant cessfully (Cai et al., 2001). In cortical neurons from
transgenes. BACE1 null embryos, there is no cleavage at the +1
and + 11 sites of A (Cai et al., 2001), and the secre-
Mutant APP Mice tion of peptides is abolished even in the presence
Several promoters have been used to drive the ex- transfected mutant APP transgenes. Moreover, A
pression of an APP minigenes that encode the FAD- peptides are not produced in the brains of BACE1
linked APP mutants (swe and 717) in strains of mice. null mice. These results establish that BACE1 is the
The pathology is influenced by the level of transgene neuronal -secretase required to cleave APP to gener-
product and the specific mutation. The hippocampi ate the N-termini of A they further establish that
and cortices of these mice show elevated levels of A, BACE1 is an excellent therapeutic target for drug de-
diffuse A deposits, and plaques consisting of dystro- velopment for AD.
phic neurites displayed around an A core. Astrocytes
and microglia are clustered in and around plaques. PS1 and PS2 Null Mice
NFT are not apparent. In some lines of mice there To examine the roles of PS1 in development, sev-
may be mild loss of neurons. Some mice show abnor- eral groups have produced PS1-/- mice (4). Homozy-
malities of synapses in hippocampal circuits that pre- gous mutant mice fail to survive beyond the early
cede the deposition of A. In some lines, mice may postnatal period and show severe perturbations in the
exhibit learning deficits, problems in object- development of the axial skeleton and ribs (defects in
recognition memory (related to the number of amy- somitogenesis) that resemble a particular Notch1 null
loid deposits in specific regions) and impairments of phenotype. Because PS1 homologues interact with
alternation/spatial-reference and working memory Notch1, a receptor protein involved in critical cell-
(perhaps related to reductions in synaptic densities in fate decisions during development, it is not surprising
the hippocampus). that cells lacking PS1 show reductions in proteolytic
release of the Notch1 intracellular domain (NICD), a
APPswe/PS1 Mutant Transgenic Mice cleavage that is thought to be critical for Notch1 sig-
Transgenic mice that coexpress A246E HuPS1 naling. Both wild type and mutant human PS1 trans-
and Mouse/Human-APPswe have elevated levels of A genes rescue the spectrum of developmental defects
in the brain and develop numerous amyloid deposits in PS1 null mice. These results indicate that the FAD-
in the hippocampus and cortex (Borchelt et al., 1997; linked PS1 variants retain sufficient normal function
Borchelt et al., 1996). The A deposits are associated to allow normal mammalian embryonic development.
with dystrophic neurites that contain APP, PS1, and With regard to the role of PS proteins in A biology,
BACE1 immunoreactivities; thus, the key participants mutations in PS genes increase the formation of
in amyloidosis appear locally at some of the possible A42, and ablation of PS1 reduces the secretion of A.
sites of formation of A. Significantly, cells from PS1-/- mice show reductions
in the levels of -secretase cleavage products and le-
vels of A
Gene-Targeted Mice Particularly Relevant
to AD
APP and APLP2 Null Mice Vulnerability of Neurons in Alzheimers
Homozygous APP-/- mice are viable and fertile, Disease
but they appear to have subtle decreases in locomotor Among the most challenging mysteries of neu-
activity and forelimb grip strength. The absence of rodegenerative diseases is the identification of factors
substantial phenotypes in APP-/- mice may be related that render neurons susceptible in specific diseases
to the functional redundancy provided by homolo- (the principle of selective vulnerability). For example,
gous amyloid precursorlike proteins (APLP1 and APP and SOD1 are abundant in many cells. Why,
APlP2), molecules expressed at high levels with devel- then, do mutations of genes encoding these proteins
opmental and cellular distributions similar to APP. cause neurological diseases? And why are mutations
22 ALZHEIMERS DISEASE: Human Disease and the Genetically Engineered Animal Models
in APP associated with the development of a demen- dogenic enzyme, wherease BACE2 is an antiamyloi-
tia syndrome and mutations in SOD1 with an MND dogenic protease, and that the relative levels of
phenotype? Recent research has begun to provide ex- BACE1 and BACE2 are major determinants of A
citing new clues concerning the biological basis for amyloidosis. Significantly, -secretase, which may be
vulnerabilities of neurons. AD serves as an illustra- PS1 (or has its activity influenced by PS), is present in
tion. We suggested that the basis for the vulnerabili- a relatively low level in brain and does not form A
ties of the brain to AD are related to the levels and dis- without BACE1 cleavages.
tributions of APP and its cleavage enzymes in neurons
as opposed to other cells (Cai et al., 2001). APP is one In this model (Cai et al., 2001), the secretion of
of the most abundant proteins in neurons, and avail- A peptides would be highest in neurons/brain as
able evidence indicates that neurons are the principal compared to other cell types/organs because neurons
source of A. In nerve cell, APP is transported within express high levels of BACE1 coupled with low ex-
axons by the fast anterograde system. APP processing pression of BACE2. If the ratio of the level of BACE1
can occur at nerve terminals. In the terminal fields of to BACE2 is a critical factor that selectively predis-
the perforant pathway, BACE1 cleavage generates poses the brain to A amyloidosis, AD would likely in-
soluble C-terminally truncated APPs and amyloido- volve the brain selectively as opposed to other organs.
genic C-terminal fragments. Moreover, in mutant
A seemingly contradictory study shows a high level of
APP transgenic mice, APP, BACE1, and PS1, the key
BACE1 mRNA expression in the pancreas. Since APP
proteins in the formation of amyloid, have been seen
is expressed in the pancreas, why do AD and diabetes
in swollen neurites in immediate proximity to A de-
mellitus not occur together? It now appears that some
posits. These observations conform to the idea that
neurons and their processes, including axon termi- of the pancreatic BACE1 mRNAs are alternatively
nals, are one source of the APP that gives rise to A spliced to generate a BACE1 isoform that is incapable
peptide species. of cleaving APP. Taken together with the observations
that the pancreas possesses low levels of BACE1 and
However, the presence of APP in neurons, allow amounts of BACE1 activity, these results are con-
though necessary, is not sufficient to explain why the sistent with the view that a high ratio of BACE1 to
brain is particularly vulnerable to A amyloidogenesis
BACE2 activity leads to selective vulnerability of neu-
whereas other organs, such as the pancreas, are not
rons and not pancreatic cells to A amyloidosis. To
. The patterns of APP-cleavage enzymes in different
test this hypothesis at the level of specific cell popula-
cell populations are of equal importance. The cellular
tions, it will be important to define the levels and dis-
distributions, relative levels, and sites of APP cleavage
of BACE1, BACE2, and -secretase are principal de- tributions of BACE1 and BACE2 in specific brain re-
terminants of such vulnerability. Although both gions, circuits, and neurons using specific BACE1 and
BACE1 and BACE2 are expressed ubiquitously, 2 antisera and to attempt to correlate these measures
BACE1 mRNA levels are particularly high in the with the regional vulnerabilities to A amyloidosis
brain and pancreas, whereas the levels of BACE2 seen in AD.
mRNA are relatively low in all tissues except the
brain, where it is nearly undetectable. As indicated
above, A is generated by biochemical pathways in- Treatment in Models of A
volving the endoproteolytic cleavages carried out by Amyloidogenesis
BACE1 and by an activity termed -secretase, which Research on model systems relevant to AD illus-
generate the N- and C-termini of the A peptide, re- trates the value of studies of transgenic and gene-
spectively. Most importantly, BACE1 is the principal
targeted mice for experimental treatments. Although
-secretase necessary to cleave APP at the +1 and
they do not model the full phenotype of AD, these
+11 sites of A in neurons (Cai et al., 2001). In con-
mutant mice represent excellent models of A
trast, BACE2 cleaves APP more efficiently at residues
amyloidogenesis and are highly suitable for analyses
+19 and +20 of APP compared to the +1 site of A.
Significantly, levels of A1-19 and A1-20 are undet- of pathogenic pathways, determination of the molec-
ectable in brain. APP can also be cleaved endoproteo- ular participants in amyloidogenesis, and identifica-
lytically before residue +17 within the A sequence by tion of therapeutic targets. Moreover, they are invalu-
-secretases, either TACE or ADAM10. These able for examining the effects of the introduction
three cleavages within the A domain of APP pre- and/or ablation of specific genes, administration of
clude the formation of A 1- 40,42. Because BACE1 pharmacological agents (secretase inhibitors), and A
is the principal -secretase in neurons (Cai et al., vaccination or passive transfer of A antibody. Re-
2002) and BACE2 may serve to limit the secretion of searchers need to assess the efficacy of anti-A thera-
A peptides, we suggest that BACE1 is a proamyloi- pies in transgenic mice exhibiting tau pathology.
AMNESIA, FUNCTIONAL 23
Bibliography Amnesia in the Dissociative Disorders

Albert, M. S. Cognitive and neurobiologic markers of early Al-
zheimer disease (1996). Proceedings of the National Academy of
One major category of functional amnesia occurs
Sciences of the United States of America 93, 13,54713,551. within the context of diagnosable psychopathology,
Borchelt, D. R., Ratovitski, T., Van Lare, J., Lee, M. K., Gonzales, especially the dramatic Dissociative Disorders listed
V. B., Jenkins, N. A., Copeland, N. G., Price, D. L., and Si- in the Diagnostic and Statistical Manual (DSM) (4th
sodia, S. S. (1997). Accelerated amyloid deposition in the edition) of the American Psychiatric Association. In
brains of transgenic mice co-expressing mutant presenilin 1
and amyloid precursor proteins. Neuron 19, 939945.
current diagnostic nosology, this category includes a
Borchelt, D. R., Thinakaran, G, Eckman, C. B., Lee, M. K., Daven- wide variety of syndromes whose common core is an
port, F. Ratovitsky, F. Prada, C.-M., Kim, G. Seekins, S., alteration in consciousness affecting memory and
Yager, D., Slunt, H. H., Wang, R., Seeger, M. Levey, A. I., identity.
Gandy, S. E., Copeland, N. G., Jenkins, N. A., Price, D. L.,
Younkin, S. G., and Sisodia, S. S. (1996). Familial Alzheimers In dissociative amnesia (also known as psycho-
disease-linked presenilin 1 variants elevate A142/140 ratio genic amnesia, limited amnesia), the patient suffers
in vitro and in vivo. Neuron 17, 1,0051,1013. a loss of autobiographical memory for specific past
Cai, H., Wang, Y., McCarthy, D., Wen, H., Borchelt, D. R., Price,
D. L., and Wong, P. C. (2001). BACE1 is the major -secretase
experiences. It sometimes occurs in cases of violent
for generation of A peptides by neurons. Nature Neuroscience crime (interestingly, affecting either victims or perpe-
4, 233234. trators), war neurosis, and other types of posttraumat-
Citron, M., Oltersdorf, T., Haass, C., McConlogue, L., Hung, A. ic stress disorder. Unfortunately, many reports of dis-
H., Seubert, P., Vigo-Pelfrey, C., Lieberburg, I., and Selkoe, sociative amnesia are anecdotal and lack independent
D. J. (1992). Mutation of the -amyloid precursor protein in
familial Alzheimers disease increases -protein production.
corroboration of the purported instigating event.
Nature 360, 672674. Moreover, it is not possible to reliably distinguish
De Strooper, B., Saftig, P., Craessaerts, K., Vanderstichele, H., genuine cases of psychogenic amnesia from simulated
Guhde, G., Annaert, W., Von Figura, K., and Van Leuven, F. cases.
(1998). Deficiency of presenilin-1 inhibits the normal cleav-
age of amyloid precursor protein. Nature 391, 387390. In dissociative fugue (psychogenic fugue, func-
Price, D. l., Tanzi, R. E., Borchelt, D. R., and Sisodia, S. S. (1998). tional retrograde amnesia), the amnesia covers the
Alzheimers disease, genetic studies and transgenic models. whole of the individuals past life and his or her per-
Annual Review of Genetics 32, 461493.
sonal identity; there may also be physical relocation
Donald L. Price (which gives the syndrome its name). The condition
Sangram S. Sisodia may go unnoticed until the patients are asked person-
Revised by Philip C. Wong and Donald L. Price al questions that they cannot satisfactorily answer. Re-
covery typically begins with the patients recognition
of loss of identity. Recovery of identity and memory
per se may occur spontaneously or in response to the
AMNESIA appearance of a relative or other salient cue. When
See: ALZHEIMERS DISEASE: BEHAVIORAL ASPECTS; the fugue is resolved, the patient is typically left with
ALZHEIMERS DISEASE: HUMAN DISEASE AND a limited amnesia covering the period of the fugue.
THE GENETICALLY ENGINEERED ANIMAL
In dissociative identity disorder (multiple-
MODELS; AMNESIA, FUNCTIONAL; AMNESIA,
personality disorder), a single individual appears to
INFANTILE; AMNESIA, ORGANIC; AMNESIA,
TRANSIENT GLOBAL; HEAD INJURY; manifest two or more distinct identities, each alter-
KNOWLEDGE SYSTEMS AND MATERIAL- nating in control over conscious experience, thought,
SPECIFIC MEMORY DEFICITS; and action. Before World War II, the typical case in-
REHABILITATION OF MEMORY DISORDERS volved only two or three such ego states; more re-
cent cases have tended to present more alter egos,
leading some to speculate that iatrogenic and so-
ciocultural factors may account for much of the
AMNESIA, FUNCTIONAL multiple-personality epidemic of the 1980s. In genu-
ine cases, the personalities are separated by an amne-
Analyses of learning and memory increasingly at-
sic barrier. The dissociation may be symmetrical, in
tempt to take account of clinical and experimental re-
which each ego state is ignorant of the other(s) or,
search on victims of amnesia. Most of this literature
more commonly, asymmetrical, in which case an ego
has focused on pathologies of memory associated with
state may be aware of some of its counterparts but ig-
demonstrable brain lesions or the administration of
norant of others.
centrally acting drugs. The functional amnesias are a
collection of memory disorders instigated by process- In depersonalization the person believes that he
es that do not result in damage or injury to the brain or she has changed in some way or is somehow unreal;
but that do engender a marked increase in forgetting. in derealization the same beliefs are held about ones
24 AMNESIA, FUNCTIONAL
surroundings. Because these beliefs are objectively in- Explicit and Implicit Memory
appropriate, these experiences can be construed as While the functional amnesias by definition im-
disorders of memory: the person fails to recognize pair explicit memory, some anecdotal and experi-
some object, self, or situation with which he or she is mental evidence suggests that the amnesia may spare
objectively quite familiar. Episodes of depersonaliza- implicit memory, or the unconscious influence of past
tion and derealization frequently occur in response to events on subsequent experience, thought, or action.
stress and in association with anxiety disorders; they In dissociative identity disorder, for example, both
may also be induced by psychedelic drugs and occur procedural learning and priming effects may transfer
spontaneously in a substantial proportion of the nor- between personalities, so that one alter ego is influ-
mal population. enced by the experiences of another even though the
Although dissociative disorders have been of in- amnesic barrier prevents conscious recollection. The
terest at least since the time of Freud and Janet, they situation is complicated, however, because not all
rarely have been studied with controlled experimen- forms of implicit memory are equally spared. There
tal procedures. For example, little is known about has been no experimental corroboration of clinical
psychogenic amnesia beyond anecdotes. A few cases claims that special procedures such as hypnosis and
of fugue and multiple personality have been studied barbiturate narcosis can promote conscious access to
in the laboratory, but we have no idea how represen- the lost memories.
tative they are. Nevertheless, the available evidence Among nonpathological amnesias, the dissocia-
suggests a pattern of selective memory deficit in some tion between explicit and implicit memory is especial-
respects resembles that observed in organic amnesia. ly well documented in posthypnotic amnesia. Prever-
Thus, psychogenic fugue impairs memory for past ex- bal infants can show long-term retention of new
periences and other aspects of self-knowledge but learning in a manner that suggests implicit memory.
leaves the patients repertoire of impersonal proce- However, sleep-learning procedures do not appear to
dural and semantic knowledge largely intact. Disso- leave any traces, even in implicit memory.
ciative identity disorder displays a similar pattern.
Trauma, Repression, and Dissociation

Nonpathological Amnesias
The lack of reliable evidence of brain damage has
In other forms of functional amnesia, dramatic fostered a tendency to account for functional amne-
forgetting occurs in the ordinary course of everyday sias in purely psychological terms. Since the nine-
living, albeit with no implication of pathology. For ex- teenth century, repression and dissociation have been
ample, people commonly fail to remember their the favored explanations. Repression, as defined by
dreams and other events of the nights sleep; attempts Freud, is the motivated forgetting of material (typi-
to demonstrate sleep learning have been almost uni- cally, relating to sexual or aggressive ideas and im-
formly unsuccessful. Theoretical accounts of this pulses) that conflicts with physical reality or social
memory deficit usually revolve around encoding fac- sanctions. Dissociation, as discussed by Janet and
tors. For example, one hypothesis holds that sleep in- Prince, is a more adventitious splitting off from
hibits the higher cortical centers that support percep- awareness of a set of percepts, memories, thoughts, or
tual processing. feelings. While Freud argued that repressed contents
Another example of nonpathological functional could be known only by inference (because they were
amnesia is the general paucity of memory for infancy expressed only symbolically, as in dream contents),
and childhood. As with sleep, most theoretical ac- Janet argued that dissociated contents could be recov-
counts of this developmental amnesia focus on encod- ered directly, by hypnosis and other means. Given the
ing factors. For example, infantile amnesia (covering pervasive influence of Freudian psychoanalysis in
the first two years of life) may reflect the childs rela- twentieth-century discourse, the repression thesis
tive inability to encode symbolic and especially lin- long held sway. There has been a subsequent revival
guistic representations of events; even older children of the concept of dissociation, as indicated by adop-
may lack the information-processing capacity to en- tion of Dissociative Disorder as a category in the
code retrievable memories. DSM. An eclectic combination of Freudian and Jun-
gian theories formed the basis of clinical theories
A dramatic form of forgetting known as post- about trauma and memory that emerged in the late
hypnotic amnesia occurs in some hypnotized subjects. twentieth century. These theories, in turn, promoted
In some respects, posthypnotic amnesia may serve as the 1980s revival of therapeutic strategies based on
a laboratory analogue of the dissociative amnesias the recovery and working through of traumatic
seen in the clinic. memories.
AMNESIA, FUNCTIONAL 25
Paradoxically, the idea that trauma plays a role in Bibliography

amnesia, while a salient aspect of clinical folklore, is Bootzin, R. R., Kihlstrom, J. F., and Schacter, D. L. (1990). Sleep
not supported by solid empirical evidence. Animal and cognition. Washington, DC: American Psychological Asso-
studies indicate that high levels of emotional arousal ciation.
Eich, E., Macaulay, D., Lowewenstein, R. J., and Dihle, P. H.
stimulate the release of stress hormones that activate (1996). Memory, amnesia, and dissociative identity disorder.
the amygdala, leading to enhanced memory. A review Psychological Science 8, 417422.
of more than sixty longitudinal studies of document- Ellman, S. J., and Antrobus, J. S. (1991). The mind in sleep: Psychology
ed trauma survivors yielded not a single instance of and psychophysiology. New York: Wiley.
amnesia that could not be accounted for by nontrau- Howe, M. L. (2000). The fate of early memories: Developmental science
and the retention of childhood experiences. Washington, DC:
matic infantile and childhood amnesia or organic
American Psychological Association.
factors such as intoxication, anoxia, or head injury. Kapur, N. (1999). Syndromes of retrograde amnesia: A conceptual
Many clinical studies alleging repression of trauma and empirical synthesis. Psychological Bulletin 125, 800825.
also fail to rule out such factors, especially in cases of Kihlstrom, J. F. (1994). One hundred years of hysteria, Dissociation:
infantile and childhood amnesia that occur indepen- Clinical and theoretical perspectives. New York: The Guilford
dently of trauma. Others fail to confirm that the os- Press.
(1996). The trauma-memory argument and recovered
tensibly traumatic event even occurred or misinter-
memory therapy. In K. Pezdek and W. P. Banks, eds.,The re-
pret subjects failure to disclose their trauma as a covered memory/false memory debate. San Diego: Academic Press.
failure to remember it. (1997). Suffering from reminiscences: Exhumed memory,
There are very few corroborated reports of the implicit memory, and the return of the repressed. In M. A.
Conway, ed., Recovered memories and false memories. Oxford:
therapeutic recovery of traumatically lost memories. Oxford University Press.
Most clinicians apparently assume their patients (2001). Dissociative disorders. In P. B. Sutker and H. E.
memories are valid or refrain from challenging them Adams, eds., Comprehensive handbook of psychopathology. New
in order to protect the therapeutic alliance. The York: Plenum.
techniques of memory work used therapeutically to Kihlstrom, J. F., and Barnhardt, T. M. (1993). The self-regulation
of memory: For better and for worse, with and without hypno-
help patients recover traumatic memories also in-
sis, In D. M. Wegner and J. W. Pennebaker, eds., Handbook of
crease the risk of memory distortion and confabula- mental control. Englewood Cliffs, NJ: Prentice-Hall.
tion. There is no evidence that either hypnosis or se- Kihlstrom, J. F., and Schacter, D. L. (2000). Functional amnesia.
dation by barbiturate drugs facilitates the recovery of In F. Boller and J. Grafman, eds., Handbook of neuropsychology.
valid traumatic memories. Whenever recovered Amsterdam: Elsevier.
memories are taken as evidence, whether in the clinic Kopelman, M. D. (1995). The assessment of psychogenic amnesia.
In A. D. Baddeley, B. A. Wilson, and F. N. Watts, eds., Hand-
or the courtroom, it is critical that they be accompa-
book of memory disorders. Chichester, UK: Wiley.
nied by independent, objective, corroborative evi- (1997). Anomalies of autobiographical memory: Retro-
dence. grade amnesia, confabulation, delusional memory, psycho-
genic amnesia, and false memories. In J. D. Read and D. S.
Lindsay, eds., Recollections of trauma: Scientific evidence and clini-
Neuropsychology of Functional Amnesia cal practice. New York: Plenum.
Markowitsch, H. J. (1999). Functional neuroimaging correlates of
Beginning with the discovery of the syphilis spiro- functional amnesia. Memory 7, 561583.
chete, a major goal of psychiatry has been to transfer McNally, R.J. (2002). Remembering Trauma., Cambridge, MA: Har-
syndromes from the functional to the organic catego- vard University Press.
ry as their neural bases are revealed. It seems likely Piper, A. (1993). Truth serum and recovered memories of sex-
that neuroscientific research will pinpoint such causal ual abuse: A review of the evidence. Journal of Psychiatry and
the Law 21, 447471.
relationships for the dissociative amnesias as well. For
Piper, A., Pope, H. G., and Borowiecki, B. S. (2000). Custers last
example, a recent PET study of dissociative fugue stand: Brown, Schefflin, and Whtfields latest attempt to sal-
found no sign of the activity in the right posterior vage dissociative amnesia. Journal of Psychiatry and Law 28
frontal and anterior temporal lobes that usually ac- 149213.
companies recollection of emotionally salient person- Pope, H. G., Oliva, P. S., and Hudson, J. I. (2000). Repressed mem-
al experiences but did find activation in the left fron- ories: B. Scientific status. In D. L. Faigman, D. H. Kaye, M.
J. Saks, and J. Sanders, eds., Modern scientific evidence: The law
tal and temporal regions. Although functional may and science of expert testimony. St. Paul, MN: West Publisher.
eventually prove to be a misnomer, for now the term Reed, G. (1988). The psychology of anomalous experience. Buffalo, NY:
delineates a class of amnesias in which psychological Prometheus.
processes rather than brain insult, injury, or disease Schacter, D. L. (1986). Amnesia and crime: How much do we really
are the immediate causes of the memory failure. know? American Psychologist 41, 286295.
(1987). Implicit memory: History and current status. Jour-
nal of Experimental Psychology: Learning, Memory, and Cognition
See also: AMNESIA, INFANTILE; AMNESIA, ORGANIC; 13, 501518.
AMNESIA, TRANSIENT GLOBAL; HYPNOSIS AND (2001). The seven sins of memory: How the mind forgets and re-
MEMORY members. Boston: Houghton-Mifflin.
26 AMNESIA, INFANTILE
Shobe, K. K., and Kihlstrom, J. F. (2002). Interrogative suggestibil- The Lower Boundary for Long-Term Recall
ity and memory work. In M. L. Eisen, J. Quas, and G. S.
Goodman, eds., Memory and suggestibility in the forensic inter-
of Early Experiences
view. Mahwah, NJ: Erlbaum. Maturation of the Central Nervous System
Sullivan, M. D. (1990). Organic or functional? Why psychiatry
needs a philosophy of mind. Psychiatric Annals 20, 271277. Maturation of the human brain begins at concep-
tion, but continues throughout childhood (and be-
John F. Kihlstrom
yond). Although our understanding of the time
Daniel L. Schacter
course of human brain development is not complete,
we do know that many of the brain areas that play a
role in long-term memory are not fully mature during
infancy and early childhood. Thus, although learning
occurs rapidly during this phase of development, the
AMNESIA, INFANTILE
ability to retain and use information over a lifetime
Do you remember being born? Your first birthday may be precluded by the immaturity of the brain
party? Your first day of school? Despite the signifi- (Campbell and Spear, 1972).
cance of these early experiences, most adults recall lit- Maturation of two areas of the brainin particu-
tle or nothing about them. The absence of autobio- lar, the medial temporal lobe (including the hippo-
graphical memory for events that occurred during campus) and the frontal cortexis thought to play a
infancy and early childhood is commonly referred to particularly important role in the phenomenon of in-
as infantile (or childhood) amnesia. Sigmund Freud fantile amnesia (Bachevalier, 1992). Maturation of
originally identified the phenomenon of infantile am- the hippocampus occurs relatively early in develop-
nesia by asking his patients to describe their earliest ment and may be sufficient to support some of the so-
personal memories in the course of therapy. On the phisticated memory skills exhibited by infants; how-
basis of these patient reports, Freud argued that the ever, maturation of the higher-association areas of
period of infantile amnesia extended into the sixth or the frontal cortex continues well into childhood and
eighth year of life. Freuds most often-cited explana- may be required for the maintenance and retrieval of
tion of infantile amnesia was highly influenced by his memories over the long term (Hayne, Boniface, and
patient population. He believed that memories for Barr, 2000; C. Nelson, 1995).
our infancy and early childhood were stored in pris-
tine condition, but were actively repressed due to The Development of Language
their emotionally and sexually charged content. In When we ask adults to recall their earliest person-
fact, one goal of Freuds psychoanalytic process was al memories, we commonly ask them to provide a ver-
to unlock these hidden memories to allow patients bal report of what they can rememberboth the in-
to come to terms with the traumatic thoughts and ex- structions they are given (tell me about your earliest
periences of their childhood. memory) and their response to those instructions re-
quire sophisticated language skills. Infants and chil-
Subsequent normative studies of adults earliest
dren, on the other hand, typically express their mem-
memories have shown that Freud probably overesti-
ories, by necessity, through nonverbal behaviors.
mated the period of infantile amnesia. There is now
Even once they have acquired conversational lan-
general consensus that adults earliest autobiographi-
guage skills, children still rely primarily on their non-
cal memories are for events that occurred when they
verbal skills to solve tasks that require memory. Fur-
were approximately three to four years of age (Bruce,
thermore, the ability to translate early, preverbal
Dolan, and Phillips-Grant, 2000; Dudycha and Dudy-
experiences into language is extremely limited, if not
cha, 1941; Mullen, 1994; Sheingold and Tenney,
impossible (Simcock and Hayne, 2002). Although an
1982; Waldfogel, 1948) or even slightly younger
early preverbal memory may be reflected in some as-
(MacDonald, Uesiliana, and Hayne, 2000; Usher and
pect of an adults behavior (Newcombe, Drummey,
Neisser, 1993). Furthermore, normative studies of
Fox, Lie, and Ottinger-Alberts, 2000), he or she will
adults earliest memories have failed to provide any
be unable to provide a verbal report of the original
empirical evidence in support of Freuds repression
experience. In this way, language development is an-
model (Pillemer and White, 1989).
other rate-limiting step in the offset of infantile am-
Thus, the fundamental question remains: Why is nesia.
it that we have little or no recollection of events that
occurred during our infancy and early childhood? Al-
though repression does not provide an adequate ex-
Beyond the Basic Ingredients: The
planation for the phenomenon, empirical studies Emergence of Autobiographical Memory
point to a number of other factors that might account The basic ingredients for long-term verbal mem-
for infantile amnesia (Howe and Courage, 1993). ory are in place by the end of the second year of life.
AMNESIA, INFANTILE 27
For example, children as young as two and a half can ory for a particular experience (e.g., Remember
provide a verbal report of an event that occurred when we went to the zoo last year?). Furthermore, an
eighteen months earlier (Fivush and Hammond, adults account of an event may augment the childs
1990). Despite this, most adults can recall nothing memory for the same experience, yielding a richer,
about events that occurred prior to their third or more integrated (and thus more memorable) repre-
fourth birthday, and even those memories are few sentation. The frequency and content of these conver-
and far between and are significantly less vivid or de- sations ultimately shape the number and clarity of our
tailed than events that occurred later in childhood. earliest recollections (MacDonald, Uesiliana, and
How can we account for this aspect of infantile amne- Hayne, 2000; Mullen, 1994).
sia? In conclusion, no single explanation of infantile
A Framework for Organization amnesia can account for all of the available data. In-
stead, it is likely that brain maturation and language
Many lines of research have shown that memory
acquisition define the lower limit for our earliest rec-
is enhanced when the to-be-remembered information
ollections, but the number and quality of memories
can be organized according to some cognitive frame-
that actually survive will be determined by the way in
work, or schema. Whether the material to be remem-
which those memories have been structured and or-
bered consists of stories, baseball statistics, or lists of
ganized. Conversations about the past during early
words, it is learned and retrieved most effectively by
childhood are a driving force in this process.
people who can systematically organize the material
by relating it to an existing framework of knowledge.
Without this kind of schematic organization, recall is See also: CHILDREN, DEVELOPMENT OF MEMORY IN;
only partial and fragmentary, if it occurs at all. CODING PROCESSES: ORGANIZATION OF
MEMORY; EPISODIC MEMORY; EXPERTS
This same principle also applies to the recall of MEMORIES; GUIDE TO THE ANATOMY OF THE
autobiographical information. When we think or talk BRAIN; NATURAL SETTINGS, MEMORY IN; PROSE
about our past experiences, for example, we typically RETENTION
recall that information in chunks that reflect mile-
stones in our lives: graduation from high school, at-
tending college, getting married, the birth of our chil- Bibliography
dren and grandchildren, and so on. Our schema for Bachevalier, J. (1992). Cortical versus limbic immaturity: Relation-
ship to infantile amnesia. In M. R. Gunnar and C. A. Nelson,
these stages in our lives helps us to retrieve individual eds., Developmental behavioral neuroscience, pp. 129153. Hills-
memories that occurred during each stage. Further- dale, NJ: Erlbaum.
more, thinking or talking about experiences from a Bruce, D., Dolan, A., and Phillips-Grant, K. (2000). On the transi-
particular stage often reminds us of other events that tion from childhood amnesia to the recall of personal memo-
occurred during the same stage even if those events ries. Psychological Science 11, 360364.
Campbell, B. A., and Spear, N. E. (1972). Ontogeny of memory.
were greatly separated in time. Psychological Review 79, 215236.
Children, on the other hand, do not organize Dudycha, G. J., and Dudycha, M. M. (1941). Childhood memories:
their memories on the basis of milestones that reflect A review of the literature. Psychological Review 38, 668682.
Fivush, R., and Hammond, N. R. (1990). Autobiographical memo-
important stages of development or pivotal life ry across the preschool years: Towards reconceptualizing
events. Because these early experiences are not linked childhood amnesia. In R. Fivush and J. A. Hudson, eds.,
to the same schemata that adults use when they at- Knowing and remembering in young children, pp. 223248. New
tempt to retrieve them, individual memories become York: Cambridge University Press.
difficult, if not impossible to recall (Pillemer and Fivush, R., and Reese, E. (1992). The social construction of autobi-
ographical memory. In M. A. Conway, D. C. Rubin, H. Spinn-
White, 1989). ler, and W. A. Wagenaar, eds., Theoretical perspectives on autobi-
ographical memory, pp. 115132. Dordrecht: Kluwer Academic
Parent-Child Conversations about the Past Publishers.
Childrens emerging ability to organize and re- Freud, S. (1960). Three essays on the theory of sexuality: 2. Infan-
count their life history is shaped through conversa- tile sexuality. In The complete psychological writings of Sigmund
Freud, vol. 7, pp. 173206. London: Hogarth. The original
tions about the past that occur in the context of their
definition of infantile amnesia. First published in 1905.
family (Fivush and Reese, 1992; Hudson, 1990; K. Hayne, H., Boniface, J., and Barr, R. (2000). The development of
Nelson, 1993). In the course of reminiscing about declarative memory in human infants: Age-related changes in
mutually experienced events with parents or other deferred imitation. Behavioral Neuroscience 114, 7783.
significant social partners, children acquire the sche- Howe, M., and Courage, M. (1993). On resolving the enigma of in-
fantile amnesia. Psychological Bulletin 113, 305326.
ma that are necessary to catalog their autobiographi-
Hudson, J. A. (1990). The emergence of autobiographical memory
cal memories. Additionally, conversations about the in mother-child conversation. In R. Fivush and J. A. Hudson,
past allow children the opportunity to practice using eds., Knowing and remembering in young children, pp. 166196.
another persons language to retrieve their own mem- New York: Cambridge University Press.
28 AMNESIA, ORGANIC
MacDonald, S., Uesiliana, K., and Hayne, H. (2000). Cross-cultural learning word pairs), and recognition memory. His
and gender differences in childhood amnesia. Memory 8, 365 impairment is global in nature: information received
376.
Mullen, M. K. (1994). Earliest recollections of childhood: A demo-
from all sensory modalities is affected, and both ver-
graphic analysis. Cognition 52, 5579. bal and nonverbal (e.g., spatial) memory are im-
Nelson, C. (1995). The ontogeny of human memory: A cognitive paired. H.M. also evidences deficient recall of remote
neuroscience perspective. Developmental Psychology 31, 723 memories antedating the surgery (retrograde amne-
738. sia). For instance, H.M. could not recall the death of
Nelson, K. (1993). The psychological and social origins of autobio-
graphical memory. Psychological Science 4, 714.
a favorite uncle who had died three years prior to the
Newcombe, N. S., Drummey, A. B., Fox, N. A., Lie, E., and Ot- surgery. Moreover, nearly all of his personal memo-
tinger-Alberts, W. (2000). Remembering early childhood: ries are from the age of sixteen or earlier (Sagar,
How much, how, and why (or why not). Current Directions in Cohen, Corkin, and Growdon, 1985).
Psychological Science 9, 5558.
Pillemer, D. B., and White, S. H. (1989). Childhood events recalled Despite the severity of his amnesia, H.M. exhibits
by children and adults. In H. W. Reese, ed., Advances in child normal attention, language, and general intellectual
development and behavior, Vol. 21, pp. 297340. Orlando, FL:
abilities. His performance on tasks of immediate, or
Academic Press.
Sheingold, K., and Tenney, Y. J. (1982). Memory for a salient working, memory is relatively preserved. That is, he
childhood event. In U. Neisser, ed., Memory observed, pp. 201 can temporarily store and maintain information over
212. San Francisco: Freeman. a brief interval, such as that required when rehearsing
Simcock, G., and Hayne, H. (2002). Breaking the barrier? Children a telephone number. But if he is distracted or pre-
fail to translate their preverbal memories into language. Psy-
vented from continually rehearsing the material, the
chological Science 13, 225231.
Usher, J. A., and Neisser, U. (1993). Childhood amnesia and the information is forgotten. H.M. exhibits preservation
beginnings of memory for four early life events. Journal of Ex- of some long-term memory functions, such as skill
perimental Psychology: General 122, 155165. and habit learning, where learning is expressed as en-
Waldfogel, S. (1948). The frequency and affective character of hanced task performance. The detailed examination
childhood memories. Psychological Monographs 62, 139.
of H.M.s amnesia has served as a milestone in the
Ulric Neisser quest to elucidate the functional and neuroanatomi-
Revised by Harlene Hayne cal basis of memory and has spawned decades of re-
search into the memory processes that are impaired
and preserved in amnesia.
AMNESIA, ORGANIC
Etiologies of Amnesia
Organic amnesia is a neurological disorder character-
ized by a dense impairment of memory in the context Although H.M. still serves as a benchmark for
of normal intelligence and other preserved mental characterizing global amnesia, it has also become
abilities. Investigations of patients with this disorder clear that the disorder is composed of a number of
have enhanced the understanding of the psychologi- different patterns of memory loss that may be linked
cal processes involved in learning and remembering, to distinct etiologies and associated patterns of brain
as well as the brain organization of human memory. damage. Organic amnesia has been associated with a
wide range of medical conditions including vascular
Much of the current interest in memory and brain accidents (i.e., strokes), ischemia (e.g., loss of blood
function finds its origin in the study of patient H.M., flow to the brain), anoxia (i.e., loss of oxygen to the
a man who, in 1953, underwent surgery for treatment brain), viral infections, and Wernicke-Korsakoff syn-
of refractory seizures (Scoville and Milner, 1957). The drome. As with H.M., patients with these medical con-
surgery involved bilateral resection of a large portion ditions typically show preserved attention, working
of the medial temporal region, which includes the memory, and general intellectual abilities.
amygdala, hippocampus, and hippocampal gyrus. Al-
though the surgery was successful in substantially re- Not all forms of amnesia are permanent. For ex-
ducing H.M.s seizures, the procedure produced a ample, head injury can cause transient and selective
pervasive impairment of memory (Milner, Corkin, memory impairment. Anterograde amnesia following
and Teuber, 1968). Since the time of his surgery at head trauma can last minutes, days, or even weeks.
the age of twenty-seven, H.M. has been unable to con- Depending on whether the trauma is mild or severe,
sciously learn and remember new information (an- patients may completely regain their learning ability
terograde amnesia). For example, thirty minutes or may suffer long-lasting and sometimes permanent
after eating lunch, H.M. cannot recall what he ate for impairment. Retrograde amnesia may also occur, and
lunch, nor can he recall if in fact he ate lunch at all. the temporal extent of retrograde amnesia is often
He exhibits severe impairment on laboratory tests of correlated with the severity of anterograde amnesia.
word and picture recall, cued-word learning (e.g., Quite severe memory problems also occur after elec-
AMNESIA, ORGANIC 29
troconvulsive therapy (ECT), a procedure sometimes disorder resulting from the convergent effects of
prescribed for treatment of severe depression. How- chronic alcohol abuse and malnutrition. Studies of
ever, ECT-induced amnesia is usually transient, and postmortem brain tissue by Victor, Adams, and Col-
extensive recovery of memory occurs with time. lins (1989) have revealed bilateral damage involving
the dorsomedial nucleus of the thalamus and a sub-
thalamic nucleus called the mamillary bodies. Similar
The Anatomy of Memory pathology can also occur in nonamnesic alcoholics,
Analysis of the locus of brain lesions in patients but what distinguishes patients with Wernicke-
such as H.M. has underscored the importance of the Korsakoff syndrome is that they also show neuronal
medial temporal region, particularly the hippocam- loss in anterior thalamic nuclei (Harding, Halliday,
pus and adjacent regions, in new learning. Further in- Caine, and Kril, 2000). Cortical atrophy (i.e., brain-
sight into the prominent role of the hippocampus in cell loss) and cerebellar damage are also often ob-
memory comes from the study of patient R.B. by Zola- served. On neuropsychological tests Wernicke-
Morgan, Squire, and Amaral (1986). R.B. became am- Korsakoff patients evidence both anterograde amne-
nesic in 1978 when he suffered an ischemic episode sia and retrograde amnesia. In addition, these pa-
during open-heart surgery. Neuropsychological as- tients may exhibit attention and problem-solving
sessment of R.B. revealed moderate level antero- impairments, as well as impaired insight. These addi-
grade amnesia alongside mild retrograde amnesia. In tional difficulties may occur as a result of cortical atro-
1983, R.B. suffered a fatal cardiac arrest and his brain phy, especially in prefrontal areas.
was brought to autopsy. Examination of his brain re-
vealed that R.B. had sustained discrete bilateral brain
damage restricted to a portion of the hippocampus Declarative Memory in Amnesia
called the CA1 subfield. Several other cases of amne- Globally amnesic patients are severely impaired
sia resulting from ischemia have come to autopsy at consciously and intentionally remembering infor-
since then (Rempel-Clower, Zola, Squire, and Ama- mation. This deficit encompasses the acquisition,
ral, 1996). These cases have revealed that more ex- long-term retention, and retrieval of both personally
tensive damage to the hippocampal formation pro- experienced events (i.e., episodic memory) and im-
duces more severe anterograde amnesia, as well as personal information (i.e., semantic memory). Collec-
extensive retrograde amnesia for memories predat- tively, these forms of memory are referred to as declar-
ing the brain injury up to fifteen years or more. In ative memory.
many patients, lesions extend beyond the hippocam- Episodic memory enables individuals to remem-
pal formation to include adjacent regions (i.e., en- ber experiences from their personal past (e.g., re-
torhinal and perirhinal). This typically leads to dense membering what they ate for breakfast this morning).
anterograde and retrograde amnesia. Importantly, Episodic memories contain a multitude of sources of
comparisons between patients with restricted versus information including perceptual, conceptual, and
extensive medial temporal lesions suggest that the emotional components. Episodic memories are not
hippocampal formation and adjacent regions make stored in isolation, but are placed within a context of
qualitatively different contributions to memory. It has personally relevant information. A pervasive deficit in
been suggested that one neural circuit centered in the episodic memory is dramatically exemplified by glob-
hippocampus supports recollection (i.e., the con- ally amnesic patients. In the clinic, episodic memory
scious or intentional retrieval of past experiences) is assessed by tests of recall and recognition. Because
whereas another neural circuit centered in adjacent these tasks require intentional retrieval of recent ex-
perirhinal regions mediates overall feelings of famil- periences, they are referred to as tests of explicit memo-
iarity (i.e., the subjective sense that something was en- ry. Although both recall and recognition require in-
countered previously) (Aggleton and Brown, 1999; tentional retrieval, their processing demands are not
Brown and Aggleton, 2001). identical. Recognition memory is based on two dis-
Damage to another region of the brain, the dien- tinct memory processes: recollection, an effort-
cephalic midline, can also produce organic amnesia. dependent process by which information is deliber-
This brain region includes various midline thalamic ately brought to mind (e.g., remembering what book
nuclei (nuclei are groups of neurons), as well as sub- you read last night); and familiarity, a facilitation of
thalamic nuclei. These neurons serve as relay groups, or fluency in stimulus processing that results from
sending and receiving projections to numerous parts prior experience with that stimulus (e.g., the sense of
of the brain, including the medial temporal region. familiarity when seeing a person on the bus, even
The best-studied cases of amnesia resulting from though you may not remember when you last saw him
damage to diencephalic midline structures are pa- or her). By contrast, recall depends largely on recol-
tients with Wernicke-Korsakoff syndrome, an amnesic lection. Most globally amnesic patients evidence defi-
30 AMNESIA, ORGANIC
cits both on recall and on recognition tests, but their amnesic individuals. Strikingly, however, H.M. had
recall performance is typically worse than their recog- no awareness of having practiced the task.
nition performance (Giovanello and Verfaellie,
2001). This finding suggests that recollection may be Another form of nondeclarative memory is repe-
more severely disrupted in amnesia than is familiarity tition priming, the facilitation in performance in-
(Yonelinas et al., 1998). duced by previous exposure to stimuli. A typical prim-
ing experiment is composed of a study phase
Whereas episodic memory is concerned with per- followed by a test phase. During the study phase,
sonally experienced events, semantic memory refers participants are exposed to a list of words or pictures.
to the acquisition and long-term retention of generic For example, participants might see a list containing
factual information. Semantic knowledge encompass- the word apricot. During the subsequent test phase,
es a wide range of information, including facts about participants are asked to perform a seemingly unre-
the world (e.g., the knowledge that Rome is the capi- lated task. For example, they may be asked to identify
tal of Italy), the meanings of words and concepts (e.g., briefly flashed words or to generate as many words as
an understanding of the concept Website), and the possible when cued with the semantic category
names attached to objects and people (e.g., the fruit. Priming is measured as the change in accura-
knowledge that William Shakespeare wrote Hamlet). cy or the bias in task performance induced by recent
Studies of H.M. (Gabrieli, Cohen, and Corkin, 1988) exposure to task stimuli (e.g., enhanced accuracy at
and other amnesic patients (Verfaellie, Reiss, and identifying the word apricot or enhanced likelihood of
Roth, 1995) suggest that semantic learning is im- generating the word apricot), compared to a test con-
paired in amnesia. For example, H.M. has been un- dition in which apricot did not appear on the prior
able to acquire any new vocabulary words that entered study list.
the language since his surgery. Other amnesic pa-
tients, however, appear able to acquire some new Priming in amnesia has been assessed using tasks
semantic information, and occasionally semantic that require analysis of the perceptual attributes of a
learning is quite good despite a patients severe im- stimulus (perceptual priming), such as identification
pairment in episodic memory (Verfaellie, 2000). For of briefly presented words or pictures and speeded
example, Jon, a young patient who became amnesic reading of words. Priming has also been assessed
following an anoxic episode at birth, has been able to using tasks that require analysis of the meaning of a
acquire a considerable amount of semantic knowl- stimulus (conceptual priming), such as category
edge, as demonstrated by his ability to attend main- exemplar generation and production of word asso-
stream schools, despite his pronounced amnesia for ciates. Amnesic patients show intact priming on both
day-to-day events, such as remembering conversa- perceptual and conceptual priming tasks, suggesting
tions or television programs (Baddeley, Vargha- that these forms of unaware memory do not depend
Khadem, and Mishkin, 2001). The degree to which on the medial temporal regions implicated in amne-
semantic learning is still possible in amnesic patients sia. The neural basis of conceptual priming is not well
likely depends on the extent of lesion in the temporal understood, but findings of impaired visual perceptu-
lobe. al priming in patients with occipital lesions suggest
that perceptual priming is mediated by posterior cor-
tical areas that process modality-specific information
Nondeclarative Memory in Amnesia (Verfaellie and Keane, 1997).
One of the most striking findings about amnesia
is that, despite their severe impairment in conscious,
intentional retrieval of information (declarative Remote Memory in Amnesia
memory), amnesic patients can acquire several forms Globally amnesic patients evidence deficits in
of memory normally. These forms of memory are col- memory for events and for facts predating the onset
lectively referred to as nondeclarative memory. One of their amnesia, although the extent of retrograde
form of nondeclarative memory is procedural learn- memory loss varies from patient to patient. Retro-
ing, the ability to acquire new perceptual and motor grade amnesia tends to follow Ribots law, which
skills on the basis of repeated practice (e.g., learning states that memory for the recent past is more severe-
to ride a bicycle). Studies of H.M. were among the ly affected than memory for the remote past. This
first to establish the preservation of procedural learn- pattern of retrograde memory loss provides clues
ing in patients with global amnesia. For instance, dur- about the mechanisms and time course involved in
ing one task, H.M. had to learn how to keep a metal the storage and retrieval of new memories. Although
stylus in contact with a revolving disc. He learned to the medial temporal region plays a critical role in the
adapt his motor movement to the movement of the permanent laying down of information, these brain
disc and gradually got better at the task, just like non- areas are not the ultimate repositories for new memo-
AMNESIA, TRANSIENT GLOBAL 31
ries. Storage of new memories requires interaction Brown, M. W., and Aggleton, J. P. (2001). Recognition memory:
between medial temporal and neocortical (outer What are the roles of the perirhinal cortex and hippocampus?
Nature Review Neuroscience 2, 5161.
brain) regions. The hippocampal formation receives Gabrieli, J. D. E., Cohen, N. J., and Corkin, S. (1988). The im-
information from a number of neocortical sites and paired learning of semantic knowledge following bilateral
binds together the spatial, perceptual, conceptual, medial temporal-lobe resection. Brain and Cognition 7, 525
and emotional components of an experience. Subse- 539.
quent attempts to retrieve this experience cause the Giovanello, K. S., and Verfaellie, M. (2001). The relationship be-
tween recall and recognition in amnesia: Effects of matching
hippocampus to reactivate the neocortical sites that recognition between patients with amnesia and controls.
contain the information. Reactivation of the neocorti- Neuropsychology 15, 444451.
cal sites leads to a strengthening of the connections Harding, A., Halliday, G., Caine, D., and Kril, J. (2000). Degenera-
between these sites and, over time, allows memories tion of anterior thalamic nuclei differentiates alcoholics with
to be retrieved without assistance from the hippocam- amnesia. Brain 123 (1), 141154.
Milner, B., Corkin, S., and Teuber, H.-L. (1968). Further analysis
pal formation (Alvarez and Squire, 1994). Conse- of the hippocampal amnesia syndrome: Fourteen-year follow-
quently, in patients with damage to medial temporal up study of H.M. Neuropsychologia 6, 215234.
or connected diencephalic structures, information Rempel-Clower, N. L., Zola, S. M., Squire, L. R., and Amaral, D.
that has not been fully consolidated is vulnerable, G. (1996). Three cases of enduring memory impairment after
whereas fully consolidated older memories can still be bilateral damage limited to the hippocampal formation. Jour-
nal of Neuroscience 16, 5,2335,255.
retrieved. Sagar, H. J., Cohen, N. J., Corkin, S., and Growdon, J. H. (1985).
Dissociations among processes in remote memory. In D. S.
Olton, E. Gamzu, and S. Corkin, eds., Memory dysfunctions: An
Conclusion integration of animal and human research from preclinical and clini-
Neuropsychological investigations of patients cal perspectives, Vol. 4, pp. 533535. New York: New York
Academy of Sciences.
with organic amnesia have contributed greatly to the
Scoville, W. B., and Milner, B. (1957). Loss of recent memory after
understanding of memory processes and the brain re- bilateral hippocampal lesions. Journal of Neurology, Neurosur-
gions that mediate them. Damage to medial temporal gery, and Psychiatry 20, 1121.
regions or diencephalic midline structures produces Verfaellie, M. (2000). Semantic learning in amnesia. In Handbook
a pervasive amnesic disorder in which conscious or of neuropsychology, 2nd edition, Vol. 2: Memory and its disorders,
ed. L. S. Cermak, pp. 335354. Amsterdam: Elsevier Science.
declarative memory is severely impaired. In contrast,
Verfaellie, M., and Keane, M. M. (1997). The neural basis of
various forms of nondeclarative memory such as pro- aware and unaware forms of memory. Seminars in Neurology
cedural memory and repetition priming are pre- 17, 7583.
served, suggesting that nondeclarative memory is not Verfaellie, M., Reiss, L., and Roth, H. (1995). Knowledge of new
mediated by the medial temporal region implicated English vocabulary in amnesia: An examination of premor-
bidly acquired semantic memory. Journal of the International
in amnesia. These theoretical advances provide the
Neuropsychological Society 1, 443453.
opportunity for more sophisticated assessment of pa- Victor, M., Adams, R., and Collins, G. (1989). The Wernicke-
tients with memory impairments and may lead to the Korsakoff syndrome and related neurologic disorders due to alcohol-
development of new rehabilitation techniques that ism and malnutrition. Philadelphia: F. A. Davis.
capitalize on preserved forms of memory. Yonelinas, A. P., Kroll, N. E., Dobbins, I., Lazzara, M., and Knight,
R. T. (1998). Recollection and familiarity deficits in amnesia:
Convergence of remember-know, process dissociation, and
See also: DECLARATIVE MEMORY;
receiver operating characteristic data. Neuropsychology 12,
ELECTROCONVULSIVE THERAPY AND MEMORY 323339.
LOSS; EPISODIC MEMORY; IMPLICIT MEMORY; Zola-Morgan, S., Squire, L. R., and Amaral, D. G. (1986). Human
MEMORY SPAN; PROCEDURAL LEARNING: amnesia and the medial temporal region: Enduring memory
HUMANS; RIBOT, THODULE; SEMANTIC impairment following bilateral lesion limited to the field CA1
MEMORY: COGNITIVE ASPECTS; SEMANTIC of the hippocampus. Journal of Neuroscience 6, 2,9502,967.
MEMORY: NEUROBIOLOGICAL PERSPECTIVE;
Arthur P. Shimamura
WORKING MEMORY: HUMANS
Revised by Kelly Sullivan Giovanello and Mieke Verfaellie
Bibliography
Aggleton, J. P., and Brown, M. W. (1999). Episodic memory, amne-
sia, and the hippocampal-anterior thalamic axis. Behavioral
and Brain Sciences 22, 425489. AMNESIA, TRANSIENT GLOBAL
Alvarez, P., and Squire, L. R. (1994). Memory consolidation and
the medial temporal lobe: A simple network model. Proceed- Transient global amnesia (TGA) is a benign neuro-
ings of the National Academy of Sciences of the United States of Amer- logical condition in which the prominent deficit is a
ica 91, 7,0417,045. temporary organic amnesic syndrome. The episode
Baddeley, A., Vargha-Khadem, F., and Mishkin, M. (2001). Pre-
served recognition in a case of developmental amnesia: Impli- of TGA is stereotyped. It usually begins suddenly,
cations for the acquisition of semantic memory? Journal of Cog- lasts for at least several hours, and resolves gradually
nitive Neuroscience 13, 357369. over several hours to a day. Careful examination dur-
32 AMYGDALA
ing an episode of TGA shows that the patient has a tained by well-studied patients with chronic organic
relatively isolated amnesic syndrome. Vision, hear- amnesia. The severity of the anterograde amnesia ap-
ing, sensation, strength, and coordination are nor- peared to correlate with the time since onset of TGA.
mal. Language, spatial abilities, and general intellec- There was no evidence for material-specific, or par-
tual function also are normal, and the TGA patient tial, amnesiano TGA patient had a significant dis-
can repeat a list of numbers or words. In contrast, the parity between the degree of anterograde amnesia for
patient can recall little of any verbal or nonverbal ma- verbal and nonverbal material. The patients also had
terial presented minutes before. The TGA patient a temporally graded retrograde amnesia (i.e., inabili-
often repeats the same question many times because ty to recall events that occurred before the onset of
of an inability to remember the answer that was just amnesia) covering at least twenty years prior to TGA
given. Frequently repeated questions include, Is onset. Again, their test scores were similar to the
there something the matter with me? and Whats scores of many well-studied patients with chronic or-
wrong, have I had a stroke? During TGA the patient ganic amnesia. The retrograde amnesia was patchy
has a patchy loss of recall for events dating from sev- all patients were able to recall some events that oc-
eral hours to many years before the attack. Older curred within the time interval affected by retrograde
memories are spared, and the patient does not lose amnesia. Indeed, during TGA some memories were
personal identity. TGA patients examined carefully recalled, albeit incompletely, that were less than one
after the episode are normal except for an inability to to two months old. The temporally graded retrograde
recall the episode. amnesia was similar for both public and personal
events. During the episode, TGA patients performed
TGA generally occurs in persons over the age
normally on almost all other formal neuropsychologi-
fifty, and 75 percent of patients are fifty to sixty-nine
cal tests.
years old. Men and women are affected with nearly
equal frequency. The estimated incidence is 5.2 per TGA is likely caused by temporary dysfunction of
100,000 per year for persons of all ages and 23.5 per either bilateral medial temporal lobe structures, in-
100,000 per year for persons older than fifty. A third cluding field CA1 of the hippocampus and adjacent,
to a half of TGA attacks are precipitated by physical anatomically related, structures; or bilateral medial
or psychological stress, including strenuous exertion, diencephalic structures, including the dorsomedial
sexual intercourse, intense emotion, pain, exposure nucleus of the thalamus, the mammillothalamic tract,
to intense heat or cold, and minor or major medical and the mammillary bodies. The exact cause of TGA
procedures. TGA has a recurrence rate of 3 to 5 per- is not known, although it appears to be a benign con-
cent per year for at least five years after the initial epi- dition that requires no medical treatment.
sode. However, the patient with TGA does not appear
to have an increased risk of developing significant
permanent memory deficit or other cognitive dys- Bibliography
function, and TGA patients have an incidence of sub- Hodges, J. R. (1998). Unraveling the enigma of transient global
sequent stroke equal to the incidence of a comparable amnesia. Annals of Neurology 43, 151153.
population. Kritchevsky, M. (1987). Transient global amnesia: When memory
temporarily disappears. Postgraduate Medicine 82, 95100.
Few laboratory abnormalities are associated with (1989). Transient global amnesia. In F. Boller and J. Graff-
TGA. Blood, urine, and spinal fluid examination, man, eds., Handbook of neuropsychology, Vol. 3, pp. 167182.
electrocardiogram, electroencephalogram, and brain Amsterdam: Elsevier.
computerized tomography are normal. Some investi-
Mark Kritchevsky
gators have noted abnormalities, particularly in one
or both medial temporal lobes, on single-photon
emission computerized tomography (SPECT) brain
scans and on diffusion-weighted magnetic resonance
imaging (MRI) brain scans performed during the epi-
sode of TGA, which resolve following the episode. AMYGDALA
Only a small number of patients have been exam- See: GENETIC SUBSTRATES OF MEMORY:
ined with formal neuropsychological tests during AMYGDALA; GUIDE TO THE ANATOMY OF THE
BRAIN; LONG-TERM POTENTIATION:
TGA. Of these, eleven were studied with a single bat-
AMYGDALA
tery of memory tests that also had been given to pa-
tients with chronic organic amnesia. All of the eleven
patients had severe anterograde amnesia (i.e., inabili-
ANIMAL COGNITION
ty to learn new material) for verbal and nonverbal ma-
terial, and their test scores were similar to scores ob- See: COMPARATIVE COGNITION
APLYSIA: Classical Conditioning and Operant Conditioning 33
APLYSIA the salient event with a fixed latency, then the animal
learns that the stimulus can serve as a predictor.
[Since the mid-1960s, the marine mollusk Aplysia has
proved to be an extremely useful model system for gaining Behavioral Studies
insights into the neural and molecular mechanisms of simple A tactile or electrical stimulus delivered to the si-
forms of memory. Indeed, the pioneering discoveries of Eric phon results in a reflex withdrawal of the gill and si-
Kandel using this animal were recognized by his receipt of phon, which presumably serves the defensive role of
the Nobel Prize in physiology or medicine in 2000. A num- protecting these sensitive structures from potentially
ber of characteristics make Aplysia well suited to the exami- harmful stimuli. The reflex exhibits several simple
nation of the molecular, cellular, morphological, and net- forms of learning, including habituation, sensitiza-
work mechanisms underlying neuronal modifications tion, and classical conditioning. In studies of aversive
(plasticity) and learning and memory. The animal has a rel- classical conditioning, the conditioned stimulus (CS)
atively simple nervous system with large, individually identi- is a brief, weak tactile stimulus to the siphon, which
fiable neurons that are accessible for detailed anatomical, by itself produces a small siphon withdrawal. The un-
biophysical, biochemical, and molecular studies. Neurons conditioned stimulus (US) is a short-duration, strong
and neural circuits that mediate many behaviors in Aplysia (noxious) electric shock to the tail, which by itself pro-
have been identified. In several cases, these behaviors have duces a large withdrawal of the siphon (the uncondi-
been shown to be modified by learning. Moreover, specific tioned response, UR). After repeated pairings, the
loci within neural circuits where modifications occur during ability of the CS to produce siphon withdrawal (the
learning have been identified, and aspects of the cellular conditioned response, CR) is enhanced beyond that
mechanisms underlying these modifications have been ana- produced by presentations of the US alone (sensitiza-
lyzed. The three entries that follow review several aspects of tion control) or of explicitly unpaired or random pre-
research on Aplysia. CLASSICAL CONDITIONING AND OP- sentations of the CS and the US (Carew, Walters, and
ERANT CONDITIONING describes several of the behaviors of Kandel, 1981). The conditioning persists for as long
Aplysia that exhibit these associative forms of learning and as four days. Thomas J. Carew and colleagues (Carew,
the progress that has been made in examining the underlying Hawkins, and Kandel, 1983) also found that this re-
mechanisms. MOLECULAR BASIS OF LONG-TERM flex exhibits differential classical conditioning. Dif-
SENSITIZATION describes the elucidation of second- ferential classical conditioning can be produced by
messenger cascades and the roles of specific genes and pro- delivering one CS to the siphon and another to the
teins in the induction and maintenance of this example of mantle region. One CS is paired with the US (the
nonassociative learning. DEVELOPMENT OF PROCESSES CS+) and the other is explicitly unpaired (the CS-).
UNDERLYING LEARNING describes the ways in which it has As in the previous studies, the US is an electric shock
been possible to identify and dissociate multiple components delivered to the tail. After conditioning, the CS+ will
of nonassociative learning on both behavioral and cellular produce more withdrawal than the CS-.
levels. For entries on other invertebrates that have proved
Other behaviors of Aplysia can also be classically
useful for examining mechanisms of learning and memory,
conditioned. For example, feeding behavior can be
see INSECT LEARNING, INVERTEBRATE LEARNING, and
classically conditioned with an appetitive protocol
MORPHOLOGICAL BASIS OF LEARNING AND MEMORY:
(Lechner, Baxter, and Byrne, 2000).
INVERTEBRATES.]
Neural Mechanisms of Aversive Classical
Conditioning in Aplysia
CLASSICAL CONDITIONING AND A cellular mechanism called activity-dependent
OPERANT CONDITIONING neuromodulation contributes to associative learning
in Aplysia (Hawkins, Abrams, Carew, and Kandel,
The simple nervous system and the relatively large 1983; Walters and Byrne, 1983; Antonov, Antonova,
identifiable neurons of the marine mollusk Aplysia Kandel, and Hawkins, 2001). A general cellular
provide a useful model system in which to examine scheme of activity-dependent neuromodulation is il-
the cellular mechanisms of two forms of associative lustrated in Figure 1. Two sensory neurons (SN1 and
learning: classical conditioning and operant (instru- SN2), which constitute the pathways for the condi-
mental) conditioning. tioned stimuli (CS+ and CS-), make weak subthresh-
old connections to a motor neuron. Delivering a rein-
forcing or unconditioned stimulus (US) alone has two
Classical Conditioning effects. First, the US activates the motor neuron and
Classical conditioning occurs when an animal produces the unconditioned response (UR). Second,
learns to associate a typically neutral stimulus with a the US activates a diffuse modulatory system that
later salient event. If the neutral stimulus precedes nonspecifically enhances transmitter release from all
34 APLYSIA: Classical Conditioning and Operant Conditioning
Figure 1 dependent protein kinase; the subsequent protein

phosphorylation leads to a modulation in several
properties of the sensory neurons. These changes in
the sensory neuron include modulation of membrane
conductances and other processes, which facilitate
synaptic transmission. This facilitation results in the
increased activation of the motor neuron and, thus,
sensitization of the reflex. The pairing specificity of
the associative conditioning is due, at least in part, to
an increase in the level of cAMP beyond that pro-
duced by serotonin alone (Abrams and Kandel, 1988;
Ocorr, Walters, and Byrne, 1985). The influx of Ca+2
associated with the CS (spike activity) amplifies the
US-mediated modulatory effect by interacting with a
Ca+2-sensitive component of the adenylyl cyclase
(Abrams and Kandel, 1988; Schwartz et al., 1983). A
critical role for Ca+2-stimulated cyclase is also sug-
gested by studies of Drosophila showing that the ad-
General model of activity-dependent neuromodulation. A. enylyl cyclase of a mutant deficient in associative
Learning: Shading indicates paired activity. A motivationally learning exhibits a loss of Ca+2/calmodulin sensitivity.
potent reinforcing stimulus (US) activates a motor neuron to
There are contributions to the plasticity of the
produce the unconditioned response (UR) and a facilitatory
synapse occurring in the motor neuron, as well (Mur-
neuron (FN) or modulatory system that regulates the strength
of the connection between sensory neurons (SN1 and SN2) and
phy and Glanzman, 1997; Bao, Kandel, and Hawkins,
the motor neuron. Increased spike activity in one sensory 1998). The postsynaptic membrane of the motor neu-
neuron (SN1) immediately before the modulatory signal ron contains NMDA receptors. If these receptors are
amplifies the degree and duration of the modulatory effects, blocked, then the associative modification of the syn-
perhaps through the Ca+2 sensitivity of the modulatory evoked apse is disrupted. NMDA receptors require concur-
second messenger, with possible contributions from the rent delivery of glutamate and depolarization in
postsynaptic neuron. The unpaired sensory neuron (SN2) does order to allow the entry of calcium. Activity in the sen-
not show an amplification of the modulatory effects. B. sory neuron (CS) provides the glutamate and the US
Memory: The amplified modulatory effects cause long-term depolarizes the cell. The subsequent increase in intra-
increases in transmitter release and/or excitability of the paired
cellular Ca+2 putatively releases a retrograde signal
neuron, which in turn strengthens the functional connection
from the postsynaptic cell to the presynaptic termi-
between the paired sensory neuron (SN1) and the motor
neuron. The associative enhancement of synaptic strength
nal. This retrograde signal would then act to further
represents the conditioned response (CR). enhance the cAMP cascade in the sensory neuron.
An important conclusion is that this mechanism
for associative learning is an elaboration of a process
the sensory neurons. This nonspecific enhancement already in place that mediates sensitization, a simpler
contributes to sensitization. Temporal specificity, form of learning. This finding raises the interesting
characteristic of associative learning, occurs when possibility that even more complex forms of learning
there is pairing of the CS (spike activity in SN1) with may use simpler forms as building blocks, an idea that
the US, causing a selective amplification of the modu- has been suggested by some psychologists for many
latory effects in SN1. Unpaired activity does not am- years but one that until the early years of the twenty-
plify the effects of the US in SN2. The amplification first century has not been testable at the cellular level.
of the modulatory effects in SN1 leads to an enhance-
ment of the ability of SN1 to activate the motor neu- Operant Conditioning
ron and produce the CR.
Operant conditioning is a process by which an an-
As discussed in the entry Molecular Basis of imal learns the consequences of its own behavior. In
Long-Term Sensitization, experimental analyses of an operant-conditioning paradigm, the delivery of a
sensitization of defensive reflexes in Aplysia have reinforcing stimulus is contingent upon the expres-
shown that the neuromodulator released by the rein- sion of a designated behavior. The probability of ex-
forcing stimulus, which is believed to be serotonin, ac- pression of this behavior will then be altered. In other
tivates the enzyme adenylyl cyclase in the sensory words, the animal learns to associate the behavior
neuron and thereby increases the synthesis of the sec- with the contingent reinforcement and modifies its
ond messenger cAMP, which activates the cAMP- behavior accordingly.
APLYSIA: Classical Conditioning and Operant Conditioning 35
Behavioral Studies Figure 2

Feeding behavior in Aplysia has been used to gain
insights into the modification of a behavioral re-
sponse by reinforcement. Aplysia feed by protracting
a toothed structure called the radula into contact with
seaweed. The radula grasps seaweed by closing and
retracting. This sequence results in the ingestion of
the seaweed. Inedible objects can be rejected if the
radula protracts while closed (grasping the object)
and then opens as it retracts, releasing the object.
Thus, the timing of radula closure determines which
behavior will be elicited.
Feeding behavior in Aplysia can be modified by
pairing feeding with an aversive stimulus. In the pres-
ence of food wrapped in a tough plastic net, Aplysia
bite and attempt to swallow the food. However, netted
food cannot be swallowed, and it is rejected. The in-
ability to consume food appears to be an aversive
stimulus that modifies feeding behavior, since trained
animals do not attempt to bite netted food (Susswein,
Schwarz, and Feldman, 1986).
Feeding behavior can also be operantly condi-
tioned with an appetitive stimulus (Brembs et al., Model of operant conditioning of feeding in Aplysia. The
2002). Animals receiving positive reinforcement that cellular network that mediates feeding behavior is represented
by the elements in circles. Motor activity comprising two basic
is contingent on biting will learn to bite more than an-
feeding patterns is depicted above. A. At first, the radula
imals receiving either reinforcement that is not con-
protraction-generating element (Prot.) is active, followed by the
tingent on their behavior or no reinforcement at all. radula retraction element (Ret.). In the nave state, neuron B51
Neural Mechanisms of Appetitive Operant has a low probability for recruitment and thus does not take
part in the feeding motor program. Radula closure occurs
Conditioning in Aplysia
during the protraction phase. Consequently, the pattern elicited
The feeding system of Aplysia has many advan- is a rejection. B. Neuron B51 now has a higher probability for
tages. For example, much of the cellular circuitry con- recruitment following contingent reinforcement. B51 is now
trolling feeding behavior has been identified. Thus, active during the motor program, leading to radula closure
it is possible for researchers to study neurons with occurring primarily during the retraction phase. Thus, the
known behavioral significance. One of these neurons, pattern elicited will now be an ingestion.
denoted as B51, is implicated in the expression of in-
gestive behavior. B51 is active predominately during
the retraction phase. Furthermore, when B51 is re- patterns being produced. The contingent reinforce-
cruited into a pattern, it recruits radula closure motor ment also results in the modulation of the membrane
neurons (see Figure 2). properties of neuron B51. The input resistance in-
creases and a smaller stimulus is required to elicit
An in vitro analogue of operant conditioning has
electrical activity. Thus, the cell is more likely to be
been developed using only the isolated buccal gan-
active in the future. This change in the likelihood of
glia, which is responsible for generating the motor
B51 activation contributes to the conditioned in-
programs involved in feeding (Nargeot, Baxter, and
crease in ingestionlike patterns. Furthermore, these
Byrne, 1999a). The ganglion expresses motor pat-
results can be replicated when induced electrical ac-
terns that are analogous to feeding behavior. These
tivity in B51 is substituted as the analogue of the be-
motor patterns can either be ingestionlike or rejec-
havior, instead of an ingestionlike motor pattern
tionlike and the type that will be expressed is not pre-
(Nargeot, Baxter, and Byrne, 1999b).
dictable. In the analogue of operant conditioning,
motor patterns corresponding to ingestion are selec- The analogue can be further reduced by remov-
tively reinforced by contingently shocking the eso- ing neuron B51 from the ganglia and placing it in cul-
phageal nerve. The esophageal nerve contains dop- ture (Lorenzetti, Baxter, and Byrne, 2000; Brembs et
aminergic processes and is believed to be part of the al., 2002). This single, isolated neuron can be condi-
pathway mediating food reward. The conditioning tioned by contingently reinforcing induced electrical
results in an increase in the likelihood of ingestionlike activity (the analogue of behavior) with a direct and
36 APLYSIA: Classical Conditioning and Operant Conditioning
temporally discrete application of dopamine (the an- common both within any one species and between dif-
alogue of reinforcement). After conditioning, the ferent species? What is the relationship between the
membrane properties of B51 are again modulated initial induction of neuronal change (acquisition of
such that the cell is more likely to be active in the fu- learning) and the maintenance of the associative
ture. Such a highly reduced preparation is a promis- change? What are the relationships among different
ing candidate to study the mechanisms of dopamine- forms of learning, such as sensitization, classical con-
mediated reward and the conditioned expression of ditioning, and operant conditioning?
behavior at the level of the intracellular signaling cas-
cades. See also: APLYSIA: MOLECULAR BASIS OF LONG-TERM
SENSITIZATION; CONDITIONING, CELLULAR
AND NETWORK SCHEMES FOR HIGHER-ORDER
Conclusions FEATURES OF; CONDITIONING, CLASSICAL AND
One of the important findings to emerge from re- INSTRUMENTAL; GLUTAMATE RECEPTORS AND
cent studies on invertebrates is their capacity to ex- THEIR CHARACTERIZATION; INVERTEBRATE
LEARNING: NEUROGENETICS OF MEMORY IN
hibit various forms of associative learning. Of particu-
DROSOPHILA; REINFORCEMENT OR REWARD IN
lar significance is the finding that at least some
LEARNING
mollusks, such as Limax, exhibit higher-order features
of classical conditioning, such as second-order condi-
tioning and blocking. Contextual conditioning, con- Bibliography
ditioned discrimination learning, and contingency ef- Abrams, T. W., and Kandel, E. R. (1988). Is contiguity detection in
classical conditioning a system or cellular property? Learning
fects have been described in Aplysia (Colwill, Absher, in Aplysia suggests a possible site. Trends in Neurosciences 11,
and Roberts, 1988; Hawkins, Carew, and Kandel, 128135.
1986). Such higher-order features can be viewed in a Antonov, I., Antonova, I., Kandel, E. R., and Hawkins, R. D. (2001).
cognitive context, and raise the interesting possibility The contribution of activity-dependent synaptic plasticity to
that other complex behavioral phenomena will be classical conditioning in Aplysia. Journal of Neuroscience 21,
6,4136,422.
identified as the capabilities of these animals are in- Bao, J. X., Kandel, E. R., and Hawkins, R. D. (1998). Involvement
vestigated further. of presynaptic and postsynaptic mechanisms in a cellular ana-
log of classical conditioning at Aplysia sensory-motor neuron
The possibility of relating cellular changes to synapses in isolated cell culture. Journal of Neuroscience 18,
complex behavior in invertebrates is encouraged by 458466.
the progress that has already been made in examin- Brembs, B., Lorenzetti, F. D., Reyes, F. D., Baxter, D. A., and
ing the neural mechanisms of simple forms of nonas- Byrne J. H. (2002). Operant reward learning in Aplysia: Neu-
sociative and associative learning. The results of these ronal correlates and mechanisms. Science 296, 1,7061,709.
Carew, T. J., Hawkins, R. D., and Kandel, E. R. (1983). Differential
analyses of Aplysia have shown that: 1. learning in- classical conditioning of a defensive withdrawal reflex in Aply-
volves changes in existing neural circuitry (one does sia californica. Science 219, 397400.
not need the growth of new synapses and the forma- Carew, T. J., Walters, E. T., and Kandel, E. R. (1981). Classical con-
tion of new circuits for learning and short-term mem- ditioning in a simple withdrawal reflex in Aplysia californica.
ory to occur); 2. learning involves the activation of Journal of Neuroscience 1, 1,4261,437.
Colwill, R. M., Absher, R. A., and Roberts, M. L. (1988). Context-
second messenger systems; 3. the second messengers US learning in Aplysia californica. Journal of Neuroscience 8,
affect multiple subcellular processes to alter the re- 4,4344,439.
sponsiveness of the neuron (at least one locus for the Hawkins, R. D., Abrams, T. W., Carew, T. J., and Kandel, E. R.
storage and readout of memory is the alteration of (1983). A cellular mechanism of classical conditioning in Aply-
specific membrane currents); and 4. long-term mem- sia: Activity-dependent amplification of presynaptic facilita-
tion. Science 219, 400405.
ory requires new protein synthesis, whereas short- Hawkins, R. D., Carew, T. J., and Kandel, E. R. (1986). Effects of
term memory does not. interstimulus interval and contingency on classical condition-
ing of the Aplysia siphon withdrawal reflex. Journal of Neuro-
While researchers have made considerable prog-
science 6, 1,0951,701.
ress in the analysis of simple forms of learning in Aply- Lechner, H. A., Baxter, D. A., and Byrne, J. H. (2000). Classical
sia, other invertebrates, and vertebrate model sys- conditioning of feeding in Aplysia: I. Behavioral analysis. Jour-
tems, there is still no complete mechanistic analysis nal of Neuroscience 20, 3,3693,376.
available for any single example of simple learning. Lorenzetti, F. D., Baxter, D. A., and Byrne, J. H. (2000). Contin-
gent reinforcement with dopamine modifies the properties of
Many of the technical obstacles are being overcome,
an individual neuron in Aplysia. Society for Neuroscience Abstracts
however, and it is likely that the analyses of several ex- 26, 1,524.
amples of learning will reach completion. Murphy, G. G., and Glanzman, D. L. (1997). Mediation of classical
conditioning in Aplysia californica by long-term potentiation
For the near future, major questions to be an- of sensorimotor synapses. Science 278, 467471.
swered include the following: To what extent are Nargeot, R., Baxter, D. A., and Byrne J. H. (1999a). In vitro analog
mechanisms for classical and operant conditioning of operant conditioning in Aplysia: I. Contingent reinforce-
APLYSIA: Development of Processes Underlying Learning 37
ment modifies the functional dynamics of an identified neu- important insights into the final phenotypic expres-
ron. Journal of Neuroscience 19, 2,2472,260. sion of learning in the adult.
(1999b). In vitro analog of operant conditioning in Aplysia:
II. Modifications of the functional dynamics of an identified
neuron contribute to motor pattern selection. Journal of
The Development of Aplysia
Ocorr, K. A., Walters, E. T., and Byrne, J. H. (1985). Associative The life cycle of Aplysia can be divided into five
conditioning analog selectively increases cAMP levels of tail phases:
sensory neurons in Aplysia. Proceedings of the National Academy
of Sciences of the United States of America 82, 2,5482,552. 1. an embryonic phase (lasting about ten days, from
Schwartz, J. H., Bernier, L., Castellucci, V. F., Polazzolo, M., Sai- fertilization to hatching)
toh, T., Stapleton, A., and Kandel, E. R. (1983). What molecu-
lar steps determine the time course of the memory for short- 2. a planktonic larval phase (lasting about thirty-five
term sensitization in Aplysia? Cold Spring Harbor Symposium on days)
Quantitative Biology 48, 811819.
Susswein, A. J., Schwarz, M., and Feldman, E. (1986). Learned
3. a metamorphic phase (lasting only two to three
changes of feeding behavior in Aplysia in response to edible days)
and inedible foods. Journal of Neuroscience 6, 1,5131,527. 4. a juvenile phase (lasting at least ninety days)
Walters, E. T., and Byrne, J. H. (1983). Associative conditioning of
single sensory neurons suggests a cellular mechanism for 5. the adult phase, defined as the onset of reproduc-
learning. Science 219, 405408. tive maturity.
John H. Byrne These five phases can be further divided into thirteen
Revised by Fred D. Lorenzetti and John H. Byrne discrete stages, each defined by a specific set of mor-
phological criteria (Kriegstein, 1977). In the analysis
of learning that will be described here, most work
has focused on the juvenile phase of development
DEVELOPMENT OF PROCESSES (stages nine through twelve), since it is during this
UNDERLYING LEARNING time that many of the behavioral systems of interest
emerge.
In the 1980s exciting progress was made in under-
standing a variety of developmental processes, rang-
ing from principles governing the birth, differentia- Forms of Learning and Developmental
tion, and migration of nerve cells to the mechanisms Timetables
underlying the functional assembly of complex neural
circuits. In addition to the intrinsic interest in devel- In adult Aplysia the siphon withdrawal reflex ex-
opment as a fundamental field of inquiry, the analysis hibits both nonassociative and associative forms of
of development has a secondary gain: By affording learning. The developmental analysis thus far carried
the experimental opportunity of examining early- out in Aplysia has focused primarily on nonassociative
emerging processes in functional isolation from later- learning in that reflex. The three most common
emerging ones, development can serve as a powerful forms of nonassociative learning are habituation, dis-
analytic tool with which to dissect and examine specif- habituation, and sensitization. Habituation refers to
ic behavioral, cellular, and molecular processes as a decrease in response magnitude occurring as a
they are expressed and integrated during ontogeny. function of repeated stimulation to a single site; dis-
habituation describes the facilitation of a habituated
A developmental strategy such as that described response by the presentation of a strong or novel
above has been useful in furthering the analysis of stimulus, usually to another site; sensitization refers
learning and memory in the marine mollusk Aplysia, to the facilitation of nondecremented responses by a
a preparation that has proved to be quite powerful for similar strong or novel stimulus. Using a behavioral
cellular and molecular studies of several forms of preparation that permitted quantification of siphon
learning. Specifically, using this strategy, it has been withdrawal throughout juvenile development, Rankin
possible to identify and dissociate multiple compo- and Carew (1987, 1988) found that these three forms
nents of nonassociative learning on both behavioral of nonassociative learning emerged according to dif-
and cellular levels. This type of analysis has also re- ferent developmental timetables. Habituation of si-
vealed previously unappreciated behavioral and cel- phon withdrawal was present very early (in stage nine)
lular processes in Aplysia. Moreover, the developmen- and progressively matured across all juvenile stages
tal dissociation of different components of learning in in terms of its interstimulus interval (ISI) function: In
juvenile Aplysia prompted a similar analysis in adult young animals, extremely short ISIs were necessary to
animals, where the same clearly dissociable compo- produce habituation, whereas in older animals, pro-
nents of learning were identified. Thus an analysis of gressively longer ISIs could be used to produce habit-
the developmental assembly of learning has provided uation. Dishabituation (produced by tail shock)
38 APLYSIA: Development of Processes Underlying Learning
Figure 1
Summary of developmental timetables for different forms of learning and their respective cellular analogues. Behavioral analysis:
Habituation is present as early as has been examined, in stage 9; dishabituation emerges soon after habituation, in stage 10;
sensitization does not emerge until sixty days after dishabituation, in mid to late stage 12. Behavioral inhibition has been measured as
early as stage 11 (indicated by shading) but may emerge even earlier. Cellular analysis: The cellular analogue of habituation
(homosynaptic decrement of EPSPs in neuron R2) is present in stage 9; the analogue of dishabituation (facilitation of decremented
EPSPs in R2) emerges in stage 10; the analogue of sensitization (facilitation of nondecremented EPSPs in R2) emerges between early
and mid stage 12. Inhibition of nondecremented EPSPs in R2 has been detected as early as early stage 12 (indicated by shading) but,
as with behavioral inhibition, may emerge even earlier. Thus there is a close developmental parallel between the emergence of each
behavioral form of learning (as well as behavioral inhibition) and its respective cellular analogue.
emerged soon after habituation, in a distinct and later The observation that dishabituation and sensiti-
stage (stage ten). However, sensitization (also pro- zation can be developmentally dissociated raises im-
duced by tail shock) did not emerge until surprisingly portant theoretical questions for a complete explana-
late in juvenile development (stage twelve), at least tion of nonassociative learning. For example, until
sixty days after the emergence of dishabituation (see recently a commonly held view was that nonassocia-
Figure 1). tive learning could be accounted for by a dual process
APLYSIA: Development of Processes Underlying Learning 39
theory involving two opposing processes: a single results illustrate two important points. First, the cellu-
decrementing process that gives rise to habituation, lar analogue of each form of learning emerges in
and a single facilitatory process that underlies both close temporal register with its respective behavioral
dishabituation and sensitization (Carew, Castellucci, form (see Figure 1). Second, just as dishabituation
and Kandel, 1971; Groves and Thompson, 1970). A and sensitization can be developmentally dissociated
key prediction of this view is that dishabituation and on a behavioral level, so can their cellular analogues
sensitization should always occur together. However, be developmentally dissociated.
the developmental dissociation of these processes, to-
gether with recent behavioral and cellular evidence in
adult Aplysia, suggests that a dual-process view is inad- A Novel Inhibitory Process
equate to account for nonassociative learning in Aply- When the effects of sensitization training (i.e., the
sia. effects of tail shock on nondecremented reflex re-
The emergence of sensitization in stage twelve is sponses) in different developmental stages were ex-
not confined to the siphon withdrawal reflex in Aply- amined, an unexpected effect of tail shock was discov-
sia. Stopfer and Carew (1988) examined another re- ered: Prior to the emergence of sensitization in stage
sponse system, escape locomotion, and found that twelve, tail shock had an inhibitory effect on reflex re-
sensitization in that system also emerges in stage sponsiveness (Rankin and Carew, 1988). The proper-
twelve. Thus sensitization is expressed in two differ- ties of this inhibitory process in juvenile Aplysia have
ent systems, one a graded reflex and the other a cen- been studied by Rankin and Carew (1989), who found
trally programmed cyclical behavior, at the same time that tail shock-induced inhibition of siphon withdraw-
in development. This raises the interesting hypothe- al can be detected in two ways: 1. by reduction of re-
sis that one or more developmental signals may flex responsiveness; and 2. by the apparent competi-
switch on the general process of sensitization in stage tion of the inhibitory process with the facilitatory
twelve, not only in individual response systems but in process of dishabituation. Specifically, they found
the whole animal. that as levels of tail shock were increased, progressive-
ly more inhibition resulted and, concomitantly, pro-
gressively less dishabituation was produced, suggest-
Cellular Analogues of Learning and ing the hypothesis that the tail shock-induced
Behavioral Learning inhibition could significantly retard the expression of
dishabituation in early developmental stages. Finally,
The developmental separation of different learn- as the process of sensitization matured, there was a
ing processes described above provides the opportu- clear transition from the inhibitory effect of tail shock
nity to examine the unique contribution of specific to reflex facilitation between early and late stage
cellular mechanisms to each form of learning. An im- twelve.
portant step in such a cellular analysis is to show that
the cellular analogue of each form of learning can be In parallel with the behavioral reflection of inhi-
identified in the central nervous system of juvenile bition described above, Nolen and Carew (1988)
Aplysia and that these analogues exhibit a develop- identified a clear cellular analogue of this inhibitory
mental time course parallel to the behavioral expres- process. Specifically, they found that prior to the
sion of the learning. The identified motor neuron can emergence of the cellular analogue of sensitization in
serve as a reliable cellular monitor of plasticity in the mid to late stage twelve, activation of the pathway
afferent input to the siphon-withdrawal reflex. from the tail produced significant inhibition of non-
decremented synaptic responses in neuron R2 (see
The developmental emergence of the cellular an- Figure 1). As with the behavior, there was a clear tran-
alogue of habituation (synaptic decrement) and of sition from inhibition to facilitation in mid to late
dishabituation (facilitation of decremented synaptic stage twelve.
potentials) was first examined by Rayport and Camar-
do (1984). They found that synaptic decrement could The inhibitory process first identified in juvenile
be observed in neuron R2 as early as stage nine, and Aplysia has received considerable attention in the
that facilitation of depressed synaptic potentials adult. Several laboratories have observed behavioral
emerged in stage ten. Nolen and Carew (1988) then tail shock-induced inhibition of the siphon withdraw-
examined the emergence of the cellular analogue of al reflex (Krontiris-Litowitz, Erikson, and Walters,
sensitization (facilitation of nondecremented synaptic 1987; Mackey et al., 1987; Marcus et al., 1988), and
potentials) in R2. They found that the analogue of important progress has been made in studying the
sensitization emerged between early and middle cellular mechanisms underlying the inhibitory pro-
stage twelve, many weeks after the emergence of the cess. For example, Mackey et al. (1987) found that tail
analogue of dishabituation. Taken collectively, these shock produced presynaptic inhibition of the trans-
40 APLYSIA: Development of Processes Underlying Learning
mission from siphon sensory neurons. Wright, Mar- stronger stimuli. Moreover, the stimulus intensity
cus, and Carew (1991) found that polysynaptic input that was most effective in producing dishabituation
to the siphon motor neurons plays an important role produced no sensitization, and the intensity that was
in mediating tail shock-induced inhibition, and most effective in producing sensitization produced no
Wright and Carew (1990) found that a single identi- significant dishabituation. Thus, as in juvenile Aplysia
fied inhibitory interneuron in the abdominal gangli- (Rankin and Carew, 1989), the processes of dishabi-
on, cell L16, can account for most, if not all, of the in- tuation, sensitization, and inhibition can be behavior-
hibition of siphon withdrawal following tail shock. ally dissociated in adult animals.
Finally, Fitzgerald and Carew (1991) found that sero-
tonin, a known facilitatory neuromodulator in Aplysia, The behavioral observations described above
can also mimic the inhibitory effects of tail shock. It raise important questions about the cellular processes
will be of considerable interest to study the develop- underlying the dissociation of dishabituation and sen-
ment of these inhibitory mechanisms and examine sitization. One possibility is that these two forms of
the way in which they are integrated with facilitatory learning reflect different underlying cellular mecha-
forms of behavioral plasticity. nisms. Alternatively, the same or related mechanisms
may be involved in both forms of learning, and the
dissociation observed could be due to differential in-
Behavioral Dissociation of Dishabituation, teraction of the inhibitory process with dishabituation
Sensitization, and Inhibition in Adults and sensitization. Behavioral results alone cannot dis-
tinguish between these possibilities. However, prog-
The developmental studies described above show ress has been made in elucidating the cellular mecha-
that dishabituation, sensitization, and a novel inhibi- nisms underlying dishabituation and sensitization
tory process, as well as their respective cellular ana- (Hochner et al., 1986) as well as inhibition (Bellardet-
logues, can each be dissociated in juvenile animals. It ti, Kandel, and Siegelbaum, 1987; Mackey et al.,
is possible, however, that these processes, although 1987; Wright and Carew, 1990; Wright, Marcus, and
separable during ontogeny, are not distinct in the Carew, 1991). Thus it will be important to determine
final adult phenotype. Thus an important question the degree to which these different cellular processes
arose as to whether the same forms of behavioral plas- can account for the behavioral dissociations that are
ticity could be identified and separated in adult ani- observed in both developing and adult Aplysia.
mals. Marcus et al. (1988) addressed this issue by ex-
amining, in adult Aplysia, the effects of a wide range
of tail-shock intensities, at several times after tail Conclusion
shock, on both habituated and nonhabituated siphon
A developmental analysis in Aplysia has shown
withdrawal responses. They found that dishabituation
that different forms of learning, as well as their cellu-
and sensitization could be clearly dissociated in adult
lar analogues at central synapses, emerge according
animals in two ways.
to very different developmental timetables. These
First was time of onset. When tested soon (ninety studies have allowed the dissociation of four behavior-
seconds) after tail shock, dishabituation was evident al processes (see Figure 1): two decrementing (habitu-
at a variety of stimulus intensities, whereas, in this ation and inhibition) and two facilitatory (dishabitua-
early test, sensitization was not exhibited at any stimu- tion and sensitization). Whether these dissociations
lus intensity. In fact, examining nondecremented re- are produced by different facilitatory mechanisms, by
sponses revealed that tail shock produced inhibition differential interactions of inhibition with decre-
of reflex amplitude. Although no sensitization was ev- mented and nondecremented responses, or by some
ident in the ninety-second test, in subsequent tests combination of these alternatives, results suggest that
(twenty to thirty minutes after tail shock) significant a dual-process view of nonassociative learning, which
sensitization was observed. Thus, dishabituation has postulates a single decremental and a single incre-
an early onset (within ninety seconds), whereas sensi- mental process, requires revision, and that a multiple-
tization has a very delayed onset (twenty to thirty min- process view, which includes the possibility of inhibi-
utes) after tail shock. Juvenile Aplysia also exhibit de- tory as well as facilitatory interactions, is necessary to
layed-onset sensitization that emerges in early stage account adequately for the mechanisms underlying
twelve, at least thirty days after the emergence of dis- nonassociative learning.
habituation.
The developmental studies discussed in this brief
Second was stimulus intensity. When a range of review have focused only on nonassociative learning
stimulus intensities to the tail was examined, maximal in Aplysia, and only on short-term learning, which is
dishabituation was produced by relatively weak stimu- retained for a relatively brief time (minutes to hours).
li, whereas maximal sensitization was produced by However, Aplysia is also capable of exhibiting a variety
APLYSIA: Molecular Basis of Long-Term Sensitization 41
of forms of associative learning. Moreover, in addi- Marcus, E. A., Nolen, T. G., Rankin, C. H., and Carew, T. J. (1988).
tion to short-term forms, both nonassociative and as- Behavioral dissociation of dishabituation, sensitization, and
inhibition in Aplysia. Science 241, 210213.
sociative learning in Aplysia can exist in long-term Nolen, T. G., and Carew, T. J. (1988). A cellular analog of sensitiza-
forms lasting days to weeks. It will be of interest to ex- tion emerges at the same time in development as behavioral
amine the development of these additional processes sensitization in Aplysia. Journal of Neuroscience 8, 212222.
in order to gain insights into theoretically important Rankin, C. H., and Carew, T. J. (1987). Development of learning
questions such as the relationships between nonasso- and memory in Aplysia: II. Habituation and dishabituation.
Journal of Neuroscience 7, 133144.
ciative and associative learning and between short- (1988). Dishabituation and sensitization emerge as separate
term and long-term memory. As a step in this direc- processes during development in Aplysia. Journal of Neuro-
tion, Carew, Wright, and McCance (1989) have estab- science 8, 197211.
lished that long-term memory for sensitization (1989). Developmental analysis in Aplysia reveals inhibitory
emerges in exactly the same stage as short-term mem- as well as facilitatory effects of tail shock. Behavior and Neuro-
science 103, 334344.
ory, stage twelve. This observation lends support to Rayport, S. G., and Camardo, J. S. (1984). Differential emergence
the notion that short- and long-term memory may be of cellular mechanisms mediating habituation and sensitiza-
mechanistically interrelated in Aplysia (Golet et al., tion in the developing Aplysia nervous system. Journal of
1986). By analyzing the development of these diverse Neuroscience 4, 2,5282,532.
processes at synaptic, biophysical, and molecular le- Stopfer, M., and Carew, T. J. (1988). Development of sensitization
of escape locomotion in Aplysia. Journal of Neuroscience 8, 223
vels, it may be possible to gain unique insights into 230.
the substrates underlying learning and memory by Wright, W. G., and Carew, T. J. (1990). Contributions of interneu-
examining their developmental assembly. rons to tail-shock induced inhibition of the siphon withdrawal
reflex in Aplysia. Society for Neuroscience Abstracts 16, 20.
Wright, W. G., Marcus, E. A., and Carew, T. J. (1991). A cellular
See also: APLYSIA: CLASSICAL CONDITIONING AND analysis of inhibition in the siphon withdrawal reflex of Aply-
OPERANT CONDITIONING; APLYSIA: MOLECULAR sia. Journal of Neuroscience 11, 2,4982,509.
BASIS OF LONG-TERM SENSITIZATION
Thomas J. Carew
Bibliography
Bellardetti, F., Kandel, E. R., and Siegelbaum, S. (1987). Neuronal
inhibition by the peptide FMRFamide involves opening of S
K+ channels. Nature 325, 153156.
MOLECULAR BASIS OF LONG-TERM
Carew, T. J., Castellucci, V. F., and Kandel, E. R. (1971). Analysis SENSITIZATION
of dishabituation and sensitization of the gill withdrawal re-
flex in Aplysia. International Journal of Neuroscience 2, 7998. Sensitization is a simple form of nonassociative learn-
Carew, T. J., Wright, W. G., and McCance, E. F. (1989). Develop- ing that involves the enhancement of the response to
ment of long-term memory in Aplysia: Long-term sensitiza- a weak stimulus that occurs after the presentation of
tion is present when short-term sensitization first emerges. So- a strong or noxious stimulus. Sensitization usually oc-
ciety for Neuroscience Abstracts 15, 1,285. curs in two forms that differ in their duration and un-
Fitzgerald, K., and Carew, T. J. (1991). Serotonin mimics tail shock
in producing transient inhibition in the siphon withdrawal re- derlying mechanisms. Short-term sensitization lasts
flex of Aplysia. Journal of Neuroscience 11, 2,5102,518. seconds to minutes and involves the modification of
Golet, P., Castellucci, V. F., Schacher, S., and Kandel, E. R. (1986). neuronal membrane properties and synaptic efficacy,
The long and short of long-term memorya molecular often through the alteration of the phosphorylation
framework. Nature 322, 419422. state of existing proteins. Long-term sensitization
Groves, P. M., and Thompson, R. F. (1970). Habituation: A dual
process theory. Psychological Review 77, 419450. lasts from days to weeks, depending on the training
Hochner, B., Klein, M., Schacher, S., and Kandel, E. R. (1986). Ad- protocol. Unlike the short-term version, long-term
ditional component in the cellular mechanism of presynaptic sensitization requires synthesis of new macromo-
facilitation contributes to behavioral dishabituation in Aplysia. leculesthe inhibition of either gene transcription
Proceedings of the National Academy of Sciences of the United States into mRNA or translation of mRNA into protein
of America 83, 8,7948,798.
Kriegstein, A. R. (1977). Stages in post-hatching development of blocks long-term sensitization. In its most persistent
Aplysia californica. Journal of Experimental Zoology 199, 275288. form, long-term sensitization involves morphological
Krontiris-Litowitz, J. K., Erikson, M. T., and Walters, E. T. (1987). changes and neuronal growth.
Central suppression of defensive reflexes in Aplysia by noxious
stimulation and by factors released from body wall. Society for The marine mollusk Aplysia has proved a useful
Neuroscience Abstracts 13, 815. model for gaining insights into the underlying neural
Mackey, S. L., Glanzman, D. L., Small, S. A., Dyke, A. M., Kandel, and molecular mechanisms of long-term sensitiza-
E. R., and Hawkins, R. D. (1987). Tail shock produces inhibition. Aplysia has a simple nervous system with large,
tion as well as sensitization of the siphon-withdrawal reflex of individually identifiable neurons that are accessible
Aplysia: Possible behavioral role for presynaptic inhibition me-
diated by the peptide Phe-Met-Arg-Phe-NH2. Proceedings of for detailed anatomical, biophysical, biochemical,
the National Academy of Sciences of the United States of America 84, and molecular studies. Researchers have identified
8,7308,734. many of the neural circuits involved in sensitization
42 APLYSIA: Molecular Basis of Long-Term Sensitization
of several reflexive withdrawal behaviors and have duced preparations derived from sensitized and con-
characterized individual neurons within these cir- trol animals have shown that changes in the animals
cuits. They have also identified specific locations of behavior correlate with changes in the properties of
learning-related modifications in these circuits and both the presynaptic sensory neuron and the postsyn-
have analyzed the underlying cellular mechanisms. aptic motor neuron. Although these are unlikely to be
the only sites of plasticity associated with sensitization
in the nervous system, changes in the sensory and
The Siphon Withdrawal Reflexes of Aplysia motor neurons are likely to be important because
Sensitization studies in Aplysia have focused chief- they occur in the basic circuit of the reflex arc.
ly on the withdrawal reflexes of the siphon gill and
the tail siphon. The gill, the animals respiratory
organ, is in the mantle cavity. The siphon is a fleshy, Modulation of the Presynaptic Sensory
tubelike extension of the body wall protruding from Neuron
the mantle cavity. A stimulus delivered to the siphon The first correlate of long-term sensitization to
elicits withdrawal of both the siphon and the gill: This be described was a decrease in sensory-neuron-
is the siphon-gill withdrawal reflex. Eliciting the tail- membrane potassium current (Scholz and Byrne,
siphon withdrawal reflex requires the delivery of a 1987). Researchers had found that modulation of the
stimulus to the tail, causing withdrawal of both the tail same current for short-term sensitization was sensi-
and the siphon. In the wild, sensitization might be in- tive to serotonin (5-hydroxytryptamine, 5-HT) and
duced by an unsuccessful attack by a predator (e. g., cyclic adenosine monophosphate (cAMP) treatments.
pinching or scratching by a crab). In the laboratory, The decrease in membrane current after long-term
electrical shocks to the body wall are the most com- sensitization training leads to an increase in the num-
mon means of sensitizing the animal because it is easi- ber of action potentials elicited for a fixed amount of
er to deliver the same stimulus repeatedly. Long-term injected current. Thus, sensory neurons fire more ac-
sensitization requires prolonged training (multiple tion potentials in response to a given stimulus after
shocks delivered over an extended intervale.g., one sensitization (Cleary, Lee, and Byrne 1998).
to two hours). One training session can sensitize the
animal for several days, whereas multiple training ses- Another correlate of long-term sensitization is an
sions repeated over several days can sensitize the ani- increase in synaptic efficacy at the synapses between
mal for weeks. sensory neurons and motor neurons (Cleary, Lee, and
Byrne 1998; Frost, Castellucci, Hawkins, and Kandel,
The afferent signal is carried by mechanosensory
1985). A major component of this synaptic modula-
neurons. About twenty-four sensory neurons inner-
tion appears to be due to an increase transmitter re-
vate the siphon, and another 200 sensory neurons in-
lease from the presynaptic sensory neurons (Dale,
nervate the body wall (a subset of only about twenty
Schacher, and Kandel, 1988), although the nature of
of these innervate the tail). These sensory neurons
this increase is still poorly understood. The prime tar-
make excitatory synapses onto motor neurons that
gets of of modulation are as follows: broadening ac-
cause contractions of muscles in the gill, siphon, and
tion potential, increased numbers of release-ready
tail. Sensory neuron membrane properties and pre-
synaptic vesicles, and increased efficiency of the syn-
synaptic release machinery are the principal sites of
aptic-release machinery. It is not yet known whether
plasticity during sensitization. Sensory neurons also
other synapses made by sensory neurons (e.g., onto
synapse with a variety of interneurons. Some inter-
interneurons) are also enhanced. In addition, post-
neurons convey sensory information from one body
synaptic mechanisms may also contribute to in-
region to the circuits that control other regions.
creased synaptic efficacy (Cleary, Lee, and Byrne
Other interneurons secrete modulatory transmitters
1998; see below).
such as serotonin that modify the biophysical proper-
ties of sensory and motor neurons and the character- More extensive training (e.g., four days) pro-
istics of neurotransmitter release from the sensory duces more persistent sensitization. After four days of
neuron (i.e., heterosynaptic facilitation, discussed in training, anatomical studies have revealed remodel-
more detail below). ing of sensory neuron release sites and the growth of
additional axonal and dendritic branches as well as
new synaptic contacts that may contribute to in-
Cellular Correlates of Long-Term creased synaptic efficacy (Bailey and Chen, 1983;
Sensitization Wainwright, Zhang, Byrne, and Cleary, 2002). The
One advantage of Aplysia for studies on long-term recruitment of this growth requires several days of
learning is that the nervous system, or portions of it, training rather than a single day (Wainwright, Zhang,
can be removed for studies in vitro. Studies in re- Byrne, and Cleary, 2002). The down regulation of the
APLYSIA: Molecular Basis of Long-Term Sensitization 43
Aplysia neuronal cell adhesion molecule, ApCAM, 1999; Zhang et al., 1997). Furthermore, a scavenger
plays an important role in this growth-associated molecule specific for transforming growth factor
plasticity (Abel and Kandel, 1998). Sensory neurons blocks long-term facilitation induced by 5-HT treat-
reduce their synthesis of new ApCAM protein and in- ment. These data suggest that the Aplysia homologue
ternalize ApCAM already present at their plasma of the growth factor may be necessary for the induc-
membranes. The signal for internalization of ApCAM tion or expression of long-term plasticity by acting
is phosphorylation by extracellular signal-regulated downstream of 5-HT. Transforming growth factor
kinase (ERK, discussed below). The decrease in activates two protein kinases (PKC and ERK, see
ApCAM in sensory-neuron axonal membranes may below) in Aplysia sensory neurons (Chin et al., 2002;
allow the formation of new branches and additional Farr, Mathews, Zhu, and Ambron 1999)
neurotransmitter release sites.
Second Messengers and Protein Kinase

Modulation of the Postsynaptic Neuron Cascades
Changes in the properties of the postsynaptic In the early 1980s, researchers found that sensiti-
neuron may also contribute to the increase in synaptic zation training and 5-HT each increased cytoplasmic
efficacy underlying long-term sensitization. After levels of the intracellular second messenger, cAMP
training, there are changes in two motor-neuron (Bernier, Castellucci, Kandel, and Schwartz, 1982;
membrane properties: an increase in the resting- Ocorr, Tabata, and Byrne, 1986). Regulation of this
membrane potential and a decrease in the spike second messenger is critical for long-term memory in
threshold (Cleary et al, 1998). In addition, motor fruit flies, bees, and rodents. In Aplysia, injection of
neurons show a protein-synthesis-dependent in- cAMP into sensory neurons produces many of the
creased responsiveness to glutamate twenty-four long-term changes previously correlated with sensiti-
hours after 5-HT treatment (Trudeau and Castelluc- zation training, including decreased potassium con-
ci, 1995). This result suggests an increase in the num- ductance, increased membrane excitability, and neu-
ber of postsynaptic receptors that accompanies the in- rite growth (OLeary, Byrne, and Cleary, 1995;
crease in presynaptic transmitter release. However, Schacher et al., 1988, 1993; Scholz and Byrne, 1988).
blocking this increase in postsynaptic responsiveness Indeed, cAMP appears to be sufficient for the induc-
does not block long-term facilitation. tion of several of the major forms of long-term neuro-
nal plasticity associated with sensitization.
Heterosynaptic Modulation The key effector of cAMP for long-term neuronal
What is the modulatory signal that produces plasticity is the cAMP-dependent protein kinase
these presynaptic and postsynaptic changes? It seems (PKA). Inhibition of PKA during the induction and
that the sensory neurons that convey information early stages of consolidation of long-term memory
about the sensitizing stimulus into the central nervous blocks synaptic facilitation one day after 5-HT treat-
system synapse onto one or more interneurons. These ment (Abel and Kandel, 1998), and injection of the
interneurons, in turn, release a modulatory transmit- active catalytic subunit of PKA can induce long-term
ter to produce the changes described above. A grow- facilitation in the absence of upstream signaling
ing body of evidence indicates that the transmitter (Chain et al., 1999). Among the substrates of PKA that
used by these interneurons is serotonin (5-HT) are critical for the induction of long-term plasticity
(Byrne and Kandel, 1996). Indeed, 5-HT mimics the are several transcription factors that regulate gene
many effects of sensitization training: Prolonged or expression. The principal target of PKA is the cAMP/
multiple applications of 5-HT are necessary to pro- Ca2+-responsive element binding protein (CREB,
duce long-term changes in vitro. As an analog of sen- discussed below), a transcription factor that plays im-
sitization, 5-HT use has proved invaluable in studies portant roles in many forms of long-term memory. In
seeking to delineate the cellular mechanisms that un- addition, PKA activity persists at least twenty-four
derlie long-term sensitization. hours after sensitization training or treatment with
5-HT (Chain et al., 1999; Mller and Carew, 1998).
In addition to 5-HT, other transmitter molecules
It is possible, then, that persistent kinase activity is
may induce long-term sensitization. In the late 1990s,
also important for long-term changes.
researchers found evidence that a member of the
transforming growth factor family was involved in Although the activation of the cAMP/PKA cascade
long-term sensitization (Zhang et al., 1997). Indeed, appears to be sufficient to produce long-term plastici-
treatment of Aplysia neurons with human transform- ty associated with sensitization, other kinases play im-
ing growth factor results in long-term facilitation portant roles in the induction of sensitization. One of
and increased sensory-neuron excitability (Chin et al., the kinases is protein kinase C (PKC). Sensitization
44 APLYSIA: Molecular Basis of Long-Term Sensitization
training or extended 5-HT treatment results in a pro- Immediate Early Genes

longed activation of PKC that lasts about two hours CREB-dependent transcription leads to the ex-
after training or treatment. During this period PKC
pression of immediate early genes. Researchers have
contributes to the regulation of translation and the
identified two of these genes. One is a key component
activation of a transcription factor (C/EBP, see below)
of the ubiquitin-proteosome pathway for the degra-
that regulates the expression of late genes necessary
dation of proteins, Aplysia ubiquitin C-terminal hy-
for long-term sensitization.
drolase (ApUCH), and the other is another transcrip-
Another kinase cascade that plays an essential tion factor, CCAAT/enhancer binding protein (C/
role in the induction of long-term facilitation is the EBP). ApUCH is an enzyme that associates with the
extracellular signal-regulated kinase (ERK), a mem- proteosome and increases its proteolytic activity.
ber of the mitogen-activated protein kinase family. ApUCH is an immediate early gene involved in the
ERK can be activated by several intracellular cas- induction of long-term sensitization (Abel and Kan-
cades. Researchers have yet to pinpoint the one that del, 1998; Alberini, 1999). CREB activation increases
leads to its activation during sensitization training or the expression of ApUCH. An important substrate of
by 5-HT treatment. Yet in Aplysia sensory neurons the proteosome in sensory neurons is the regulatory
cAMP can mediate both the activation of ERK and its subunit of PKA. Increased expression of ApUCH re-
translocation to the nucleus. The cAMP cascade ap- sults in the proteosome-mediated degradation of a
pears to be important for activation of ERK during subset of the PKA regulatory subunits (those binding
the early stages of long-term sensitization induction. cAMP). As the ratio of the regulatory to catalytic sub-
ERK can also be activated by PKC, a cascade that may units decreases, PKA activity becomes independent of
be important during later stages of induction. cAMP and is prolonged. Thus, CREB activation leads
The protein synthesis required for long-term sen- to degradation of the PKA regulatory subunit and
sitization occurs during a critical period that begins thereby prolongs the activity of the PKA catalytic sub-
at the onset of training and continues until training unit, which is now independent of cAMP.
ends (Castellucci, Blumenfeld, Goelet, and Kandel, C/EBP, the CCAAT/enhancer binding protein, is
1989; Levenson et al., 2000). This period corre- another transcription factor. In Aplysia sensory neu-
sponds to the time when CREB-dependent transcrip- rons, C/EBP is an immediate early gene expressed in
tion occurs. response to CREB activation (Alberini, 1999). C/EBP
There is evidence that a second round of protein can form both homodimers with other C/EBP pro-
synthesis occurs during the formation of long-term teins or heterodimers with other transcription factors
sensitization. Growth of sensory neurons in response to activate transcription of late genes. In order to acti-
to cAMP injections is blocked by protein-synthesis in- vate transcription of its target genes, C/EBP must be
hibitor up to seven hours after injection (OLeary, phosphorylated by ERK. Although the synthesis of a
Byrne, and Cleary, 1995). This second round of pro- number of proteins are known to be regulated by sen-
tein synthesis may correspond to the gene expression sitization, researchers have yet to identify C/EBP tar-
that stabilizes long-term sensitization. get genes in Aplysia.
Although there are other sites of neural plasticity,

CREB and Memory the detailed study of the modifications that occur in
sensory neurons and their synapses with motor neu-
As noted above, long-term memory differs from
rons (i. e., the fundamental reflex arc) during long-
short-term memory chiefly in requiring protein syn-
term sensitization have led to an understanding of the
thesis. During the induction of long-term memory,
basic mechanisms that underlie this simple form of
activated PKA translocates from the sensory neuron
learning. Research using the relatively simple ner-
cytoplasm into the nucleus, where it phosphorylates
vous system of Aplysia has provided insights into
the transcriptional activator CREB1 (Abel and Kan-
memory formation at the cellular, molecular, and bio-
del, 1998). CREB1 activates transcription by binding
to the DNA of certain genes that contain cAMP/Ca2+ physical levels. These basic mechanisms appear to
responsive element (CRE) sequences in their regula- form the platform upon which more complex forms
tory domains. Another form of CREB, CREB2, re- of learning (i. e., classical and operant conditioning)
presses transcription of these genes. CREB2 is a sub- develop.
strate of ERK, which also translocates into the nucleus
during the induction phase. Thus, expression of See also: MORPHOLOGICAL BASIS OF LEARNING AND
CRE-containing genes requires both activation by MEMORY: INVERTEBRATES; PROTEIN SYNTHESIS
CREB1 and derepression by CREB2. IN LONG-TERM MEMORY IN VERTEBRATES
ARISTOTLE 45
Bibliography
Abel, T., and Kandel, E. R. (1998). Positive and negative regulatory
mechanisms that mediate long-term memory storage. Brain
Research Reviews 26, 360378.
Alberini, C. (1999). Genes to remember. Journal of Experimental Bi-
ology 202, 2,8872,891.
Bailey, C. H., and Chen, M. (1983). Morphological basis of long-
term habituation and sensitization in Aplysia. Science 220, 91
93.
Bernier, L., Castellucci, V. F., Kandel, E. R,. and Schwartz, J. H.
(1982). Facilitatory transmitter causes a selective and pro-
longed increase in adenosine 3,5-monophosphate in sensory
neurons mediating the gill and siphon withdrawal reflex in
Aplysia. Journal of Neuroscience 2, 1,6821,691.
Byrne, J. H., and Kandel, E. R. (1996). Presynaptic facilitation re-
visited: State and time dependence. Journal of Neuroscience 16,
425435.
Castellucci, V. F., Blumenfeld, H., Goelet, P., and Kandel, E. R.
(1989). Inhibitor of protein synthesis blocks long-term behav-
ioral sensitization in the isolated gill-withdrawal reflex of Aply-
sia. Journal of Neurobiology 20, 19.
Chain, D. G., Casadio, A., Schacher, S., Hegde, A. N., Valbrun, M.,
Yamamoto, N., Goldberg, A. L., Bartsch, D., Kandel, E. R.,
and Schwartz, J. H. (1999). Mechanisms for generating the
autonomous cAMP-dependent protein kinase required for
long-term facilitation in Aplysia. Neuron 22, 147156.
Chin, J., Angers, A., Cleary, L. J., Eskin, A., and Byrne, J. H.
(1999). TGF-1 in Aplysia: Role of long-term changes in the
excitability of sensory neurons and distribution of TR-II-like
immunoreactivity. Learning and Memory 6, 317330.
(2002). TGF-1 alters synapsin distribution and modulates Aristotle (The Library of Congress)
synaptic depression in Aplysia. Journal of Neuroscience 22,
RC220.
Cleary, L. J., Lee, W. L., and Byrne, J. H. (1998). Cellular corre-
changes produced by neurotransmitters in Aplysia sensory
lates of long-term sensitization in Aplysia. Journal of Neuro-
neurons. Neuron 10, 1,0791,088.
science 18, 5,9885,998.
Scholz, K. P., and Byrne, J. H. (1987). Long-term sensitization in
Dale, N., Schacher, S., and Kandel, E. R. (1988). Long-term facili-
Aplysia: biophysical correlates in tail sensory neurons. Science
tation in Aplysia involves increase in transmitter release. Sci- 235, 685687.
ence 239, 282285. (1988). Intracellular injection of cAMP induces a long-term
Farr, M., Mathews, J., Zhu, D. F., and Ambron, R. T. (1999). In- reduction of neuronal K+ currents. Science 240, 1,6641,666.
flammation causes a long-term hyperexcitability in the noci- Trudeau, L.-E., and Castellucci, V.F. (1995). Postsynaptic modifi-
ceptive sensory neurons of Aplysia. Learning & Memory 6, 331 cations in long-term facilitation in Aplysia: Upregulation of
340. excitatory amino acid receptors. Journal of Neuroscience 15,
Frost, W. N., Castellucci, V. F., Hawkins, R. D., and Kandel, E. R. 1,2751,284.
(1985). Monosynaptic connections made by the sensory neu- Wainwright, M. L., Zhang, H., Byrne, J. H., and Cleary, L. J.
rons of the gill- and siphon-withdrawal reflex in Aplysia partic- (2002). Localized neuronal outgrowth induced by long-term
ipate in the storage of long-term memory for sensitization. sensitization training in Aplysia. Journal of Neuroscience 22,
Proceedings of the National Academy of Sciences of the United States 4,1324,141.
of America 82, 8,2668,269. Zhang, F., Endo, S., Cleary, L. J., Eskin, A., and Byrne, J. H.
Levenson, J., Endo, S., Kategaya, L. S., Fernandez, R. I., Brabham, (1997). Role of transforming growth factor- in long-term
D. G., Chin, J., Byrne, J. H., and Eskin, A. (2000). Long-term synaptic facilitation in Aplysia. Science 275, 1,3181,320.
regulation of neuronal high-affinity glutamate and glutamine
uptake in Aplysia. Proceedings of the National Academy of Sciences Vincent Castellucci
of the United States of America 97, 12,85812,863. Revised by Gregg A. Phares and John H. Byrne
Ocorr, K. A., Tabata, A. M., and Byrne, J. H. (1986). Stimuli that
produce sensitization lead to an elevation of cyclic AMP levels
in tail sensory neurons of Aplysia. Brain Research 371, 190192.
OLeary, F. A., Byrne, J. H., and Cleary, L. J. (1995). Long-term
structural remodeling in Aplysia sensory neurons requires de
novo protein synthesis during a critical time period. Journal
ARISTOTLE (384322 B.C.E.)
of Neuroscience 15, 3,5193,525. Aristotle was born in northern Greece, in the town of
Schacher, S., Castellucci, V. F., and Kandel, E. R. (1988). cAMP
Stagira, in 384 B.C.E. At seventeen, he went to Athens
evokes long-term facilitation in Aplysia sensory neurons that
requires new protein synthesis. Science 240, 1,6671,669. and became a student in Platos Academy, where he
Schacher, S., Kandel, E. R., and Montarolo, P. (1993). cAMP and remained for twenty years. Although greatly influ-
arachidonic acid simulate long-term structural and functional enced by Plato and by the pre-Socratic philosophers,
46 ARISTOTLE
especially Empedocles, Aristotle was a highly original ception of a state of affairs can be stored, 2. how it can
thinker and a disciple of no one. In 347 B.C.E. he left be brought to mind later, and 3. how it happens that,
Athens and traveled extensively in Asia Minor, be- when it is brought to mind, the relation between the
coming tutor to Alexander the Great in 342 B.C.E. representation and the original state of affairs, now
Seven years later he returned to Athens and began his absent, is such that the first is a memory of the second
own school, the Lyceum. After the death of Alexander and is known to be such. In contemporary dress, these
the Great in 323 B.C.E., he left Athens, and he died are the problems of information storage, information
the following year in Khalks, a few miles north of retrieval, and the general problem of how representa-
Athens. tions represent.
In his main work on memory, De memoria et re- Aristotle tries to explain information storage by
miniscentia, Aristotle tries to dissect out the central appeal to the analogy of imprinting soft wax with a
phenomena to be explained, and suggests mechani- seal. He reasons that sense perception is somehow
cal explanations of a general sort to account for them. like a picture and that it is the perception picture that
In his scientific works, Aristotle typically seeks the re- stamps its likeness to create a memory. Apparently
ality behind the appearances, and he expects that the the perception is stamped on the soul (Aristotle has
reality may be different from what it seems. This is es- a physical, not a supernatural, conception of the soul),
pecially forward-looking in the case of mental phe- or at any rate, it is stamped on some sort of physical
nomena, where subsequent thinkers, such as Ren stuff that can be in causal interaction with it and can
Descartes (in the seventeenth century) and Zeno take on some of its properties. This helps address the
Vendler (in the late twentieth century), insist that representation problem. The imprint (memory rep-
mental reality must be exactly as it seems. Aristotles resentation) resembles, physically, the perception
collection of memory phenomena displays some sys- (perceptual representation), which in turn resembles,
tematicity, and with characteristic insight, he lights on physically, that of which it is a perception. So by tran-
several basically correct classifications. Nevertheless, sitivity of resemblance, there is a correlation between
to modern eyes some of his collection is a bit of a jum- stored representation and original state of affairs. Ar-
ble, and the mechanical explanations tendered are so istotles conclusion that there must be a resemblance
implausible that they must have been no more than was taken as axiomatic by most subsequent thinkers,
helpful metaphors to him. and they searched for the parameters of physical re-
semblance. Research since the 1970s, especially in
Aristotles relentlessly naturalistic perspective, computer science and neuroscience, has revealed that
however, gives him a decidedly modern stamp. That representation does not require resemblance in any
is, he sought physical rather than supernatural or straightforward sense, a radical departure from earli-
spiritual explanations for memory phenomena, and er theories.
he well knew the importance of observations even
though his own were occasionally mere assumptions. In asking how representations represent, Aristot-
(For example, he thought women had fewer teeth le identified a truly fundamental problem. Still only
than men.) In the absence of a developed biology, ex- partially solved, it remains a central problem, though
perimental psychology, or neuroscience, he could it is now addressed within the framework of modern
hardly be expected either to envisage explanations in psychology, philosophy, neuroscience, and computer
terms of neuronal connectivity or to know how to pen- science.
etrate learning phenomena at the behavioral level. Understanding the importance of broad systema-
ticity in a theory, Aristotle tests a theorys strength by
seeing how much can be encompassed within its
Observations and Explanations ambit. Thus he claims that the stuff that receives the
In commenting upon memory and learning phe- imprint may have varying degrees of imprintability.
nomena, Aristotles fundamental distinction is be- Explanations are then forthcoming for ones poor
tween recalling information to mind and storing in- recollection of early childhood and for declining
formation, or, as he puts it, between remembering, memory in the elderly: In very young children the
which is the reinstatement in consciousness of some- stuff is too much like running water to take the im-
thing that was there before (451b6), and memory, print; in older humans, the stuff hardens and no lon-
the existence, potentially, in the mind (452a10), of ger is very impressible. Extending this idea further,
an earlier perception or conception. In modern par- Aristotle thinks a related explanation will apply to his
lance, this is the distinction between remembering in observation that those who are too quick and those
the occurrent sense and remembering in the stored who are too slow also have poor memories. Exactly
or dispositional sense. The central problems, in Aris- what phenomenon he is addressing here is unclear,
totles view, are to explain three things: 1. how a per- and this may be one of those inexplicable Aristotelian
ARISTOTLE 47
observations that need a much broadened base of with the result that the animal can recognize different
data. individuals as belonging to the same category. In hu-
The representation problem, Aristotle notices, mans this means, for example, that a pine tree, a yew,
has a further dimension. When an image from memo- and an olive tree may all be recognized as similar de-
ry comes to mind, how do we know that it is a memo- spite differences in shape, size, and color. He says that
ry, rather than a thought or image without relation to the soul is so constituted that the universal tree can
bygone events? That is, how does the occurrent pre- be developed from the stored perceptions of individ-
sentation carry the information that it is a memory? ually distinct items. A slug, on the other hand, lacks
His answer has two parts. First, sometimes we do get the capacity to generalize across individuals because
confused, and we think a presentation is a memory it lacks the capacity to store information.
when it is not (false memory); and sometimes we have In Aristotles view, storing information provides
a memory presentation but are unaware that it is a the similarity substructure that underpins both scien-
memory. So the system is imperfect. Second, when tific categorization and the skilled knowledge dis-
the system does work, it is because for animals with played by craftsmen who can make many different
memory, the organ whereby they perceive time is clay pots or ships captains who can sail under many
also that whereby they remember (449b30). The different conditions. In modern guise, his idea is that
idea here is that when perceptions are stored as mem- generalization to items that are relevantly similar but
ories, they are also somehow indexed as to time, so incidentally different, both perceptually and behav-
that the imprint bears not only the perceptions shape iorally, requires information storage. Additionally, he
but also its whenness. regards this capacity as enabling experience, the rea-
Retrieval appears to require something like an son being that experience requires understanding,
image or an iconic presentation that resembles the which in turn requires categorization of perceptions.
original perception. The mechanism of retrieval Consequently, animals such as humans have genuine
should, one surmises, have to do with something tak- experience; animals such as slugs do not (Posterior An-
ing up the stored imprint and re-presenting it, but in alytics, 99b36;l00a5).
fact Aristotle says nothing of this. Instead he discusses
the phenomenon of association, noting that events Conclusion
experienced together are often remembered togeth- Any inclination to feel smug about Aristotles
er. He explains associated recollections by saying that shortcomings should be tempered by noting that
the movement of a perception causes the move- even current classifications of learning phenomena
ment of the memory. He sees, therefore, that part of are controversial and tentative, and experimental
the theory of storage will include the relations be- psychologists are sometimes chided for doing little
tween associated memories, but he neither provides more than codifying common sense. Nor, of course,
an account of those storage relations nor elaborates should Aristotle himself be blamed for the slavish
on how information is retrieved by the movements adoption of his every word by uncritical monks in the
(451b1530). Middle Ages. Aristotle the scientist-philosopher was
In Historia animalium Aristotle suggests that hu- anything but dogmatic. Twentieth-century physical
mans and animals differ in that humans alone can re- explanationsalthough not mechanical, but electri-
member something at will (488b25), though he also cal and biochemicalsit well with his abiding natural-
notes in De memoria et reminiscentia that recollections ism.
can occur without effort. Indeed, he observes that For a long period in the history of thought, Aris-
melancholics often have obsessive memories, try totles views on nearly everything were taken as au-
though they might to repress them. In the physicalist thoritative. His Metaphysics probably had the greatest
spirit, he conjectures that melancholics have more impact; however, the work on memory was not espe-
moisture around their sense perception center, which cially influential.
is easily set in motion, thus explaining the memorys
being presented again and again despite ones will. Bibliography
Aristotle (1941). De anima, trans. J. A. Smith. In Richard Mc-
Aristotle believed that animals differ in whether Keon, ed., The basic works of Aristotle. New York: Random
they have the capacity to store their perceptions; ani- House.
mals with the capacity to do so have genuine knowl- (1941). De memoria et reminiscentia, trans. by J. I. Beare. In
edge of their world, whereas animals lacking the ca- Richard McKeon, ed., The basic works of Aristotle. New York:
Random House.
pacity merely respond to their current perceptions on
(1941). Historia animalium. In Richard McKeon, ed., The
the basis of their innate dispositions. The advantage basic works of Aristotle. New York: Random House.
of storing perceptions is that the stored items may (1975). Aristotles Posterior analytics, trans. J. Barnes. Oxford:
come to have systematic relations among themselves, Clarendon Press.
48 ATTENTION AND MEMORY
Beare, J. I. (1906). Greek theories of elementary cognition. Oxford: tention left over for other inputs or other tasks. Ordi-
Clarendon Press. nary usage also seems to imply that limitations of at-
Churchland, P. S. (2002). Brain-wise: Studies in neurophilosophy.
Cambridge, MA: MIT Press.
tention are a key factor restricting ones ability not
Descartes, R. (1649; reprint 1968). Les passions de lme. English only to perceive many stimuli at once, but also to per-
translation in E. S. Haldane and G. R. T. Ross, trans. (1911), form two tasks at the same time. These commonplace
The philosophical works of Descartes. Cambridge, UK: Cam- notions from folk psychology may or may not accu-
bridge University Press. rately describe humans information-processing ma-
Edel, A. (1982). Aristotle and his philosophy. Chapel Hill: University
of North Carolina Press. chinery.
Kahn, C. H. (1966). Sensation and consciousness in Aristotles psy-
Late-twentieth-century research argues that the
chology. Archiv fr Geschichte der Philosophie 48, 4381.
Ross, W. D. (1923; reprint 1959). Aristotle. Cleveland: Meridian phenomena of attention are not as unitary as com-
Books. mon sense might suggest (Pashler, 1998). This is
(1955). Aristotles Parva naturalia. Oxford: Oxford Univer- scarcely surprising given the complexities of the
sity Press. bodys underlying neural machinery and of the tasks
Sorabji, R. (1972). Aristotle on memory. London: Trinity Press.
to which the human brain is adapted. Evidence for
Patricia Smith Churchland distinct attentional mechanisms is seen most clearly
Georgios Anagnostopoulos in relation to divided attention: the performance limita-
Revised by Patricia Smith Churchland tions that arise when a person attempts to process
more than one stimulus at a time. On the one hand,
there appear to be a set of processing limitations asso-
ciated with perceptual analysis of inputs in different
sense modalities. If a person is confronted with differ-
ATTENTION AND MEMORY
ent sensory inputs at the same time, it is more difficult
It seems to be a tenet of ordinary common sense that to perceive them when they arrive through the same
people remember what they attend to and forget what sense modality rather than through different modali-
they do not. Not surprisingly, researchers have noted ties. This was first demonstrated by A.Treisman and
the very close relationship between attention and A. Davies (1973), who showed that people were better
memory for a very long time, and some empirical evi- able to monitor animal names when some of the
dence for the linkage was offered as far back as the words to be monitored were presented visually and
late nineteenth century (Smith, 1895). However, it others auditorily (as compared to both in the same
was only during the twentieth century, with the ad- modality). These modality-restricted perceptual ca-
vent of cognitive psychology and its relatively rich pacity limitations are wholly governed by the volun-
array of methods for studying human information tary direction of selective attention: It is only the stim-
processing over fine time scales, that it became possi- uli that are attended to that compete for limited
ble to begin to analyze this connection in more detail. resources, not all the stimuli that may be impinging
To do so, researchers have used taxonomies of me- on the senses.
morial and attentional processes that emerge from
laboratory studies in each of these areas. In addition to the perceptual processing limita-
tions tied to particular input modalities, research
points to a separate set of limitations that become evi-
Forms of Attention dent only when a person tries to perform multiple
The term attention as used in everyday language tasks at the same time. These central attentional limi-
is a diffuse and global term that alludes to both selec- tations have their locus in the more cognitive stages
tivity and capacity limitations. The potential for selectiv- of processing, especially the planning of actions and
ity is evident in the fact that of the great multitude of the retrieval of information in memory. When two
stimuli impinging on the sense organs at any one in- tasks each require mental operations of these types,
stant, human beings are usually vividly aware of only the processing in the two tasks is normally subject to
a fairly small subset. Capacity limitations are evident queuing: selection of a response in one task must be
whenever people try to attend to more than one completed before selection of the response in the
stream of inputs, particularly if comprehension or re- other task can commence (Pashler and Johnston,
sponse is required: for example, trying to listen to the 1998). Interestingly, these limitations seem quite in-
radio at the same time as one reads a newspaper. Ca- different to what modality the information arrives in.
sual usage seems to imply that attention refers to a If a person must make a speeded response to a tone,
single process or substance that accounts for both se- for example, this will delay his or her ability to make
lectivity and capacity limitations. Thus, people speak a rapid response to a concurrently presented color
of focusing attention on one sensory input or one patch. The central interference is also independent of
task, with the result that one does not have enough at- response modalities; for example, if one response is
ATTENTION AND MEMORY 49
vocal and one manual, the interference is still ob- information is normally retained in echoic memory
served. for one to two seconds. Sensory memory systems,
When one seeks to understand the relation of at- while impressive in capacity, do not retain informa-
tention to memory, it is fruitful to inquire both about tion long enough to be useful for most purposes. If
the role of modality-specific perceptual attention the information is not transferred to short-term and/
mechanisms and the role of central attentional mech- or long-term memory, it is lost.
anisms. Does information get into sensory memory even
when a person attempts to ignore it at the time it is
presented? For auditory sensory memory, the answer
Forms of Memory
is evidently yes. This is seen, for example, in early ob-
Memory, too, appears to be composed of a num- servations by Broadbent and others that when people
ber of relatively separate functional systems. Several shadow spoken input to one ear, they can, if inter-
key distinctions have been proposed since the mid- rupted, abruptly switch over and (relying on echoic
1960s. The most important and well validated of memory) recall the last bit of information presented
these is the distinction between three broad memory to the other ear. Daily life offers many examples of
systems that hold information over different time the same phenomenon. Although there is little re-
scales (and differ in certain other properties). This search that directly addresses the question of atten-
analysis, sometimes called the modal model, postu- tional involvement in iconic memory storage, the data
lates separate sensory memory, short-term memory that do exist suggest that even unattended informa-
(STM), and long-term memory (LTM) systems. The tion is briefly represented in this literal memory stor-
model arose out of the pioneering work of N. Waugh age.
and D. A. Norman (1965) and M. Glanzer and A. R.
Cunitz (1966). Sensory memory systems seem to pro-
duce something akin to a brief literal persistence of Attention and Short-Term Memory
a stimulus for a very short time beyond its actual pre-
Primary or short-term memory maintains infor-
sentation. This probably reflects continued firing of
mation in an active state, but only so long as it is at
neurons in sensory pathways after the offset of a stim-
least periodically rehearsed or refreshed. While some
ulus. Short-term (or working) memory refers to a set
theorists originally proposed a single unitary STM, it
of very limited-capacity storage mechanisms. It is
has become clear that there are a number of separate
often assumed to reflect an actively refreshed neural
and independent short-term memory systems. One
representation in sensory, perceptual, and motor-
system (sometimes termed the articulatory buffer and
control areas, perhaps maintained by a reverberative
familiar to everyone who has ever remembered a
process involving frontal brain structures. Long-term
phone number for a half minute or so) holds speech-
memory, on the other hand, refers to the relatively
like representations of verbal material. Another kind
more permanent set of memory traces that is almost
of STM system holds visual information in a schemat-
surely encoded by changes in synaptic weights. In ad-
ic form. There are hints that other independent sys-
dition to the three-part demarcation, many investiga-
tems may also exist, holding, for example, nonvisual
tors propose a distinction within the realm of long-
spatial representations, acoustic imagery, semantic
term memory, distinguishing between so-called ex-
representations, and tactile or kinesthetic patterns.
plicit memory (underlying conscious recollections, as
The clearest evidence that different STM systems are
in recall or recognition tasks) and implicit memory
indeed independent systems comes from experi-
(e.g., changes caused by exposure to a stimulus that
ments showing that people can hold onto more than
modulate later processing without causing conscious
one type of information at the same time. For exam-
recollection).
ple, people can retain spoken digits and visual mate-
rials with little mutual interference. Further clinching
Attention and Sensory Memory the case is the finding that some patients, after suffer-
ing brain damage, have lost one form of STM without
Sensory memory systems can potentially retain a any loss of the others (Basso, Spinnler, Vallar, and
large amount of detailed sensory information about Zanobio, 1982).
an input for an extremely short period. For visual in-
puts, the sensory memory is called iconic memory. Icon- Many researchers refer to working memory to
ic memory seems to hold onto inputs for only about indicate that information in STM may be actively ma-
100 to 400 milliseconds, depending on physical prop- nipulated or transformed during the time it is re-
erties of the stimulus and the visual input that follows tained. Whatever purpose information is put to and
it. Sensory memory for auditory sensations is usually whatever term may be used for it, the most notable
called echoic memory, and most investigators agree that characteristic of STM storage is limited capacity. For
50 ATTENTION AND MEMORY
spoken material, a very limited number of words (or getting it out (retrieval). Devoting perceptual atten-
perhaps more critically, syllables) can be retained. tion to a stimulus is by no means sufficient for LTM
For visual patterns, even a four-by-four checkerboard storage (whereas it may be sufficient for STM). A per-
grid is enough to overload visual short-term memory. son can see one hundred words exposed one at a time
Access time for STM is quite good, however. For ex- in rapid succession (at a rate of, say, four per second),
ample, it appears that the image of a letter can be successfully detecting every word in the list that is an
transferred into visual short-term memory in a matter animal name, and at the end he or she will often re-
of a few hundred milliseconds, and the rate for speech member scarcely anything except animal names. Ob-
is probably not much different. What is the connec- viously the person attended to all the words in order
tion between attention and STM storage? Attending to determine their semantic category, but the process
to a stimulus is clearly a necessary condition for stor- left no permanent residue.
ing that stimulus in short-term memory. Less clear is
What does produce memory storage, then? Many
whether attending to a stimulus is a sufficient condi-
studies have concluded that the key factor is elabora-
tion. The critical experiment would involve asking
tion: active processing that uncovers connections be-
whether someone can retain one set of information
tween a to-be-stored item and other information in
(A) in STM and then attend to additional information
long-term store (Craik and Lockhart, 1972). By con-
arriving in the same modality (B) without having B
trast, mere intention or desire to remember seems
overwrite A. Using visual patterns, W. A. Phillips and
relatively unimportant.
F. M. Christie (1977) concluded that this may be pos-
sible but that the issue requires further study. As for Whereas central attentional mechanisms may
storage of spoken information, attending to irrele- play only a fairly minor role in the encoding of infor-
vant speech often compromises short-term memory mation into STM, for LTM they appear critical. Near-
quite considerably, suggesting that mere attention to ly every study that has looked at peoples ability to
a new input overwrites the existing contents of STM. form long-term memory traces while performing a
concurrent task has found a substantial impairment.
Does the storage and retention of information in This occurs even when there is no discernible similari-
STM require central attentional mechanisms dis- ty between the stimuli to be remembered and those
cussed above? For articulatory STM, the involvement used in the concurrent task: for example, remember-
appears to be intermittent rather than continuous ing odors while playing a computer game (Perkins
(perhaps arising in some initial consolidation in STM, and Cook, 1990).
and then later in scheduling periodic rehearsals). A
person can retain a phone number, for example, and In the late 1990s there was a lively controversy
perform another brief demanding task involving un- about the role of central attentional mechanisms in
related materials without losing the phone number. retrieval of information from long-term memory. On
However, if the demanding task is initiated immedi- the one hand, when people are given unlimited time
ately after storing some spoken material, memory for to retrieve materials, a concurrent task often fails to
this material may suffer (Naveh-Benjamin and Jo- much dent the number of items they can produce
nides, 1984), and the same can happen when people (Craik, Govoni, Naveh-Benjamin, and Anderson,
undertake a continuously demanding task for some 1996). On the other hand, a demanding concurrent
sustained period. task markedly reduces the rate at which information
can be retrieved from LTM (Hicks and Marsh, 2000).
The picture that emerges, then, is of short-term Furthermore, when an unrelated speeded-choice re-
memory systems closely tied to perceptual input sys- action-time task is performed concurrently with a
tems (and corresponding perceptual attention ma- paired-associate retrieval task, there is evidence that
chinery). Central attentional limitations may have the memory retrieval is completely delayed by central
some involvement, but it is relatively indirect or inter- attention to the speeded task (Carrier and Pashler,
mittent. 1995).
Attention and Long-Term Memory Attention and Implicit Memory

To store information in long-term memory, one The idea of a separate implicit memory system
need not do anything active to maintain it. Nonethe- grew out of data showing amnesics may have a spared
less, recollecting a memory tends to strengthen its ability to form certain kinds of memories: those for
long-term memory representation, often dramatical- which explicit recollection is not required (or at least
ly so; conversely, memories are probably subject to a subset of such memories). A number of studies re-
erosion with the mere passage of time. What seems to ported that formation of implicit memories also does
be most difficult is getting information into LTM, and not require limited-capacity attentional resources.
AUTOBIOGRAPHICAL MEMORY 51
For example, M. E. Smith and M. Oscar-Berman Rajaram, S., Srinivas, K., and Travers, S. (2001). The effects of at-
(1990) observed that a concurrent task at the time of tention on perceptual implicit memory. Memory & Cognition
29, 920930.
word encoding did not reduce the priming effect Smith, M. E., and Oscar-Berman, M. (1990). Repetition priming
found for repeated words as tested in a later lexical of words and pseudowords in divided attention and in amne-
decision task. However, more recent studies tend to sia. Journal of Experimental Psychology: Learning, Memory, and
find that there are considerable costs when more de- Cognition 16, 1,0331,042.
manding secondary tasks are performed at the time Smith, W. G. (1895). The relation of attention to memory. Mind 4,
4743.
of encoding (Mulligan and Hornstein, 2000; Ra- Treisman, A., and Davies, A. (1973). Dividing attention to ear and
jaram, Srinivas, and Travers, 2001). Researchers have eye. In S. Kornblum, ed., Attention and performance IV. New
more to learn about this conflict. It is possible that ex- York: Academic Press.
plicit memory measures are simply more sensitive to Waugh, N., and Norman, D. A. (1965). Primary memory. Psycholog-
impaired storage produced by divided attention, but ical Review 72, 89104.
it is also possible that implicit memory storage reflects Harold Pashler
mechanisms that are independent of central atten-
tional capacity.
See also: LANGUAGE LEARNING: HUMANS; WORKING ATTENTION DEFICIT DISORDER

MEMORY: ANIMALS; WORKING MEMORY: See: LEARNING DISABILITIES
HUMANS
AUDITORY MEMORY
Bibliography
Basso, A., Spinnler, H., Vallar, G., and Zanobio, M. E. (1982). Left See: MODALITY EFFECTS
hemisphere damage and selective impairment of auditory
verbal short-term memory: A case study. Neuropsychologia 20,
263274.
Broadbent, D. E. (1957). Immediate memory and simultaneous
stimuli. Quarterly Journal of Experimental Psychology 9, 111. AUTOBIOGRAPHICAL MEMORY
Carrier, L. M., and Pashler, H. (1995). Attentional limits in memo-
Autobiographical memory is the psychological history
ry retrieval. Journal of Experimental Psychology: Learning, Memo-
ry, and Cognition 21, 1,3391,348. of the self. It consists of memories of personal experi-
Craik, F. I. M., Govoni, R., Naveh-Benjamin, M., and Anderson, encesepisodic memoriesand knowledge of the
N. D. (1996). The effects of divided attention on encoding self or autobiographical knowledge: for example,
and retrieval processes in human memory. Journal of Experi- schools we attended, people we had relationships
mental Psychology: General 125, 159180.
with, places we have lived, places we have worked, and
Craik, F. I. M., and Lockhart, R. S. (1972). Levels of processing:
A framework for memory research. Journal of Verbal Learning so on (Conway, 2001; Conway and Pleydell-Pearce,
and Verbal Behavior 11, 671684. 2000; and McAdams, 2001). It is critical for personal
Glanzer, M., and Cunitz, A. R. (1966). Two storage mechanisms in identity, forming the basis of the self and binding self-
free recall. Journal of Verbal Learning and Verbal Behavior 5, conceptions to reality. Psychiatric illnesses or brain
351360.
damage can disrupt the connections that bind self to
Hicks, J. L., and Marsh, R. L. (2000). Toward specifying the atten-
tional demands of recognition memory. Journal of Experimen- reality through memory, leading to a loss of personal
tal Psychology: Learning, Memory, and Cognition 26, 1,483 history and the attendant delusions, confabulations,
1,498. and false beliefs.
Mulligan, S., and Hornstein, S. (2000). Attention and perceptual
priming in the perceptual identification task. Journal of Experi-
mental Psychology: Learning, Memory, and Cognition 26, 626 The Nature of Autobiographical Memory
637. and Its Relation to Self
Naveh-Benjamin, M., and Jonides, J. (1984). Maintenance rehears-
al: A two-component analysis. Journal of Experimental Psycholo- Autobiographical knowledge encompasses far
gy: Learning, Memory, and Cognition 10, 369385. more than memory: It includes statements, proposi-
Pashler, H. E. (1998). The psychology of attention. Cambridge, MA: tions, declarations, and beliefs about the self, often
MIT Press.
Pashler, H., and Johnston, J. C. (1998). Attentional limitations in accompanied by generic and/or specific (mainly visu-
dual-task performance. In H. Pashler, ed., Attention. Hove, al) images of details of prior experience. Autobio-
East Sussex, UK: Psychology Press. graphical knowledge is distinct from sensory percep-
Perkins, J., and Cook, N. M. (1990). Recognition and recall of tual episodic memories, which represent details of
odours: The effects of suppressing visual and verbal encoding actual experience (Conway, 2001). In the formation
processes. British Journal of Psychology 81, 221226.
Phillips, W. A., and Christie, F. M. (1977). Interference with visual- of an autobiographical memory, autobiographical
ization. Quarterly Journal of Experimental Psychology 29, 637 knowledge becomes linked to episodic memories
650. (Conway and Pleydell-Pearce, 2000). Memory forma-
52 AUTOBIOGRAPHICAL MEMORY
tion leads to recollective experience, a sense or feel- There have been many attempted theoretical ex-
ing of the self in the past (Tulving, 1985; Wheeler, planations of childhood amnesia (Pillemer and
Stuss, and Tulving, 1997), and attention turns inward White, 1989), but most founder on young childrens
to the autobiographical memory and, perhaps, to capacity for a wide range of specific episodic memo-
other episodic memories and autobiographical ries and detailed autobiographical knowledge (Fivush
knowledge. Of course, full autobiographical memory et al., 1996). Explanations that center on changes in
formation does not have to take place, and autobio- intellect, language, and emotional development fail
graphical knowledge and episodic memory can be simply because apparently normal autobiographical
processed independently. memories are typically accessible before the age of
Conway and Pleydell-Pearce (2000) apply the five and only after that seem to submerge in a general
term the working self to the control structure that mod- forgetfulness; it seems unlikely that an increase in
ulates this whole system of autobiographical memory general functioning would make unavailable already
formationthe dynamic combining of autobiograph- accessible memories. Hence there is no compelling
ical knowledge with episodic memories. The working explanation for this component of the lifespan-
self consists of an active goal hierarchy (only parts of retrieval curve, which remains a challenge to autobio-
which are consciously accessible), models or concep- graphical memory researchers.
tions of the self, and other forms of self-knowledge
that facilitate access to autobiographical knowledge The reminiscence bump has also attracted its
structures. New autobiographical knowledge and epi- share of theories (Rubin, Rahhal, and Poon, 1998).
sodic memories are formed (encoded) through the One is that this period endures in memory because it
working self, which also influences memory construc- is suffused with novel experiences. An alternative ex-
tion by controlling input to the knowledge base and planation holds that although only a small percentage
by evaluating output (activated autobiographical of experiences during the reminiscence bump are
knowledge). The working self may even exercise in- novel events, they survive because of their uniqueness
hibitory control over the knowledge base (Conway
in the formation of a persons life circumstances and
and Pleydell-Pearce, 2000).
interests (Fitzgerald, 1988); on this view, it is the high
accessibility of memories from this period that ac-
Autobiographical Memory Across the counts for their durability (Conway and Pleydell-
Lifespan Pearce, 2000). Perhaps many memories from the pe-
The working selfgoal hierarchy and self- riod of the reminiscence bump are of self-defining
conceptionsprobably first emerges in some coher- experiences (Fitzgerald, 1988; Singer and Salovey,
ent form as the infant develops, in its second year, the 1993) and have a powerful effect in binding the work-
capacity for objective and subjective self-awareness in ing self to a specific reality. But this period, like chlid-
the form of conceptions of I and me. (Howe and hood amnesia, has yet to find a definitive explanation
Courage, 1997). Children as young as thirty months for its relation to memory.
have detailed autobiographical memories (Fivush,
Hadden, and Reese, 1996), although these are not The recency component of the lifespan-retrieval
ususally accessible in adulthood. Undoubtedly the curve (see Figure 1) can be simply explained as a peri-
working self and its relation to autobiographical od of forgetting: recently encoded memories, whose
memory changes during childhood and perhaps sta- accessibility is retained for a longer interval, are sub-
bilises into an enduring form only in late adolescence ject to decay and/or interference and so become pro-
and early adulthood (Erikson, 1950). gressively less accessible. This is a familiar, often-
These periods of development of the self are re- observed pattern of retention. One might wonder,
flected in the lifespan-retrieval curve that arises from however, why such memories or salient experiences
older adults (about thirty-five years and older) free or should be forgotten in this way. If older adults are
cued recall of autobiographical memories (Franklin expressly instructed to recall autobiographical mem-
and Holding, 1977; Fitzgerald and Lawrence, 1984; ories from this period of forgetting, there are appar-
Rubin, Wetzler, and Nebes, 1986; Rubin, Rahhal, and ently plenty of available memories (Holmes and Con-
Poon, 1998). This technique plots the age of encod- way, 1999). Thus, this seems a matter not of
ing of memories; as shown in Figure 1, the lifespan- forgetting but rather of bias or preference in access
retrieval curve consists of three periods: childhood to memories. It may be that the recency portion of the
amnesia, (from birth to approximately five years of lifespan-retrieval curve reflects a lowering in the self-
age), reminiscence bump (from ten to thirty years), relevance of memories of recent experiences and
and the period of recency (from the present declining hence a lowering in their accessibility rather than
back to the period of the reminiscence bump). complete forgetting.
AUTOBIOGRAPHICAL MEMORY 53
Figure 1
Autobiographical Memory and Personality erativity. In contrast, those participants without a

The working self increases the accessibility of prominent commitment story showed no such bias. In
a similar way work by Woike and her colleagues has
goal-related autobiographical knowledge. Markus
further established the connection between personali-
(1977) found preferential access to memories of expe-
ty and memory (Woike, 1995; Woike, Gershkovich,
riences congruent with central self-schema, those that
Piorkowski, and Polo, 1999). Woike and colleagues
are critical to someones sense of independence or de-
(1999) investigated groups of individuals classified as
pendence. McAdams (1982) identified individuals
agentic (concerned with personal power, achieve-
with a strong intimacy motivation or with a distinctive
ment, and independence) or as communion (con-
power motivation and found that the intimacy-
cerned with relationships, interdependence, and oth-
motivation group recalled peak experiences with a ers). Agentic types consistently recalled emotional
preponderance of intimacy themes compared to indi- memories of events that involved issues of agency
viduals who scored lower on this motivation, who (mastery, humiliation). In contrast, communal types
showed no memory bias. Similarly, the power- recalled emotional memories featuring others, often
motivation group recalled peak experiences with significant others, in acts of love and friendship.
strong themes of power and satisfaction. Subsequent- These and a range of findings from other studies
ly, McAdams, Diamond, de Aubin, and Mansfield, (McAdams, 2001) all show a that the dominant mo-
(1997) examined the influence of the Eriksonian no- tives or goals of the self make memories of goal-
tion of generativity (Erikson, 1963) on the life stories relevant experiences highly accessible.
of middle-aged adults. Generativity refers to nurtur-
ing and caring for those things, products, and people
that have the potential to outlast the self. Those indi- Autobiographical Memory in Distress
viduals who were judged high in generativity, who Brain injury can impair autobiographical memo-
had a commitment story, were found to recall a pre- ry in various ways (Conway and Fthenaki, 2000). Inju-
ponderance of events highly related to aspects of gen- ries to regions of the frontal lobes often lead to a
54 AUTOBIOGRAPHICAL MEMORY
clouding or loss of detailed memories. In more ex- Fitzgerald, J. M., and Lawrence, R. (1984). Autobiographical mem-
treme cases patients may confabulate, constructing ory across the lifespan. Journal of Gerontology 39, 692698.
Fivush, R., Haden, C., and Reese, E. (1996). Remembering, re-
autobiographical knowledge into plausible but false
counting, and reminiscing: The development of autobio-
memories. Patients with damage to the temporal graphical memory in social context. In D.C. Rubin, ed., Re-
lobes and underlying structures in the limbic system, membering our past: Studies in autobiographiocal memory, pp. 341
especially the hippocampal formation, may lose the 359. Cambridge: Cambridge University Press.
ability to form new memories while retaining access Franklin, H. C., and Holding, D. H. (1977). Personal memories at
to at least some preinjury memories. Those with dam- different ages. Quarterly Journal of Experimental Psychology 29,
age to posterior regions of the brain, regions involved 527532.
Holmes, A., and Conway, M. A. (1999). Generation identity and the
in visual processing (occipital lobes) may lose the abil-
reminiscence bump: Memories for public and private events.
ity to generate visual images of the past and, because Journal of Adult Development 6, 2134.
of this, become amnesic. Their amnesia occurs be- Howe, M. L., and Courage, M. L. (1997). The emergence and early
cause the episodic content of autobiographical mem- development of autobiographical memory. Psychological Re-
ories is predominantly encoded in the form of visual view 104, 499523.
images. When the ability to generate visual images is Markus, H. (1977). Self-schemata and processing information
compromised or lost because of brain damage, then about the self. Journal of Personality and Social Psychology 35,
6378.
access to specific details of the past held in episodic
McAdams, D. P. (1982). Experiences of intimacy and power: Rela-
images is also lost. tionships between social motives and autobiographical memo-
In psychiatric illness a common occurrence is that ry. Journal of Personality and Social Psychology 42, 292302.
of a severe clouding of autobiographical memory, re- (2001). The psychology of life stories. Review of General Psy-
sulting in overgeneral memories. For instance, in chology 5, 100122.
McAdams, D. P., Diamond, A., de Aubin E., and Mansfield, E.
clinical depression patients recall many memories
(1997). Stories of commitment: The psychosocial construction
that lack detail and are much more schematic than of generative lives. Journal of Personality and Social Psychology
typical autobiographical memories. Thus, a patient 72, 678694.
asked to recall specific memories of his father could Rubin, D. C., Rahhal, T. A., Poon, L. W. (1998). Things learned in
only recall general events such as walks in the park early adulthood are remembered best. Memory & Cognition 26,
after Sunday lunch but was unable to generate a sin- 319.
gle specific memory of a single walk (Williams, 1996). Rubin, D. C., Wetzler, S. E., and Nebes, R. D. (1986). Autobio-
graphical memory across the adult lifespan. In D. C. Rubin,
Clouded, overgeneral memories have also been ob-
ed., Autobiographical memory. Cambridge, UK: Cambridge
served in schizophrenic patients and in patients suf- University Press.
fering from obsessional-compulsive disorder. One Singer, J. A., and Salovey, P. (1993). The remembered self. New York:
possibility is that the complex control processes that The Free Press.
modulate memory construction (working self) be- Tulving, E. (1985). Memory and consciousness. Canadian Psycholo-
come attenuated in psychiatric illnesses and so can no gist 26, 112.
longer form fully detailed memories. Wheeler, M. A., Stuss, D. T., and Tulving, E. (1997). Towards a the-
ory of episodic memory: The frontal lobes and autonoetic
consciousness. Psychological Bulletin 121, 351354.
Bibliography
Williams, J. M. G. (1996). Depression and the specificity of autobio-
Conway, M. A. (2001). Sensory perceptual episodic memory and its
graphical memory. In D. C. Rubin, ed., Remembering our past:
context: Autobiographical memory. Philosophical Transactions
Studies in autobiographical memory. Cambridge, UK: Cambridge
of the Royal Society of London 356, 1,2971,306.
University Press.
Conway, M.A., and Fthenaki, A. (2000). Disruption and loss of au-
Woike, B. (1995). Most-memorable experiences: Evidence for a
tobiographical memory. In L. S. Cermak, ed., Handbook of
link between implicit and explicit motives and social cognitive
neuropsychology: Memory and its disorders. Amsterdam: Elsevier.
Conway, M. A., and Pleydell-Pearce, C. W. (2000). The construc- processes in everyday life. Journal of Personality and Social Psy-
tion of autobiographical memories in the self memory system. chology 68, 1,0811,091.
Psychological Review 107, 261288. Woike, B., Gershkovich, I., Piorkowski, R., and Polo, M. (1999).
Erikson, E.H. (1950). Childhood and society. New York: W. W. Nor- The role of motives in the content and structure of autobio-
ton. graphical memory. Journal of Personality and Social Psychology
Fitzgerald, J. M. (1988). Vivid memories and the reminiscence 76, 600612.
phenomenon: The role of a self narrative. Human Development
31, 261273. Martin A. Conway
B
BARTLETT, FREDERIC (18861969) tion but also in the retrieval of memories. Another
interest concerned perception and memory perfor-
Frederic Charles Bartlett was born on October 20, mance in people of other cultures, including South
1886, in Stow on the Wold, Gloucestershire. He stud- Africa, a country he had visited. His discussions of
ied literature, logic, and philosophy before becoming conventionalization leaned heavily on evidence col-
a tutor at the University of Cambridge in 1909. At lected in other societies; in the late twentieth century
Cambridge, his interests turned to psychology, partly there was a revival of interest in Bartletts contribu-
through his acquaintance with James Ward; he was tions to cross-cultural psychology (Saito, 1999).
awarded a fellowship at St. Johns College in 1913
and obtained a first-class degree in moral sciences in In 1922 Myers left Cambridge to head the Na-
1914. Cambridge then was in the forefront of the tional Institute for Industrial Psychology, and Bartlett
movement to make experimental psychology a recog- was appointed reader and director of the Cambridge
nized branch of science in the British university sys- laboratory; two years later he also became editor of
tem; C. S. Myers (18731947), a lecturer in experi- the British Journal of Psychology, a position he held for
mental psychology there, not only had campaigned twenty-four years. In 1925 he wrote an important
for Cambridge to build a laboratory, a wish fulfilled paper on the role of the difficult word feeling in scien-
in 1912, but also had helped to found the British Jour- tific psychology. In 1931 he was elected to a chair in
nal of Psychology in 1904. Bartlett wrote an account experimental psychology at Cambridge. During his
(1937) of the early history of the Cambridge laborato- term as laboratory director, the number of faculty
ry. members grew and an increasing number of students
graduated in experimental psychology. An indication
When World War I began in 1914, Myers ap- of the success of the program was that, of the sixteen
pointed Bartlett relief director of the laboratory. professorships of psychology in Great Britain in 1957,
Bartlett instigated research into a variety of topics, inten were held by students of Myers and Bartlett. From
cluding studies of the detection of faint sounds, a 1922 to the beginning of World War II, Bartlett con-
project in which he collaborated with Emily Mary tinued to study conventionalization and memorizing
Smith, whom he married in 1920, and studies of indi- and wrote three books that show his interest in ap-
vidual differences in how subjects described pictures. plied psychology: Psychology and the Soldier (1927),
These individual differences, Bartlett believed, re- which dealt with personnel selection and war neuro-
flected above all subjective interests and socially de- ses, among other topics; Psychology and Primitive Cul-
termined interpretations: He ascribed the latter to ture (1923), in which he stressed the similarities rather
conventionalization; over the next few years he fo- than the differences between people in different so-
cused on the role of conventionalization in percep- cieties; and The Problem of Noise (1934).
55
56 BARTLETT, FREDERIC
always be at the mercy of old habits; but with a schema

one could revive individual memories that had been
laid down at widely varying periods of time and from
them form new combinations. He believed that con-
sciousness had evolved for this purpose, the looking
at or turning round on ones own schemata; the
ability to do this was greatly aided by the use of visual
images in addition to speech memory.
This insight could not be gained, argued Bartlett,
from studies of rote memory along the lines of Her-
mann Ebbinghauss experiments; further, the schema
theory allowed closer ties to be formed between ex-
perimental psychology and social psychology. Sche-
mata, he believed, were linked by appetites, in-
stincts, interests, and ideals, the first two laid down
particularly in childhood, the latter two in later life.
At the end of the twentieth century memory research-
ers still used the word schema, though some criticisms
had been raised about Bartletts terms schema and re-
construction (reviewed in Zangwill, 1972; Zangwill pre-
ferred the term abstraction).
Bartlett was made a fellow of the Royal Society in
1932, the year Remembering was published. His later
career was mainly devoted to applied psychology. He
was director of the Applied Psychology Unit at Cam-
bridge from 1944 to 1953; he was appointed to the
order of Commanders of the British Empire in 1941;
and he was knighted in 1948. His best-known work
during this period concerned fatigue states following
extended periods of skilled work. His book The Mind
The work for which Bartlett is best known is Re- at Work and Play (1951) is unusual in that it was in-
membering (1932), an elaboration of his research on tended for a juvenile audience. In Thinking (1958) he
conventionalization. In it he describes how he used discussed the development of the schema theory as an
two experimental paradigms to study memory: the example of scientific thinking. He died on September
method of repeated reproduction, in which a partici- 30, 1969.
pant studied a story or a picture and then reproduced Bibliography
it several times over a period of weeks or months; and Bartlett, F. C. (1923). Psychology and primitive culture. London: Cam-
the method of serial reproduction, in which a partici- bridge University Press.
pant recalled a story or a picture, then passed this (1925). Feeling, imaging, and thinking. British Journal of
production on to a second participant, who studied it, Psychology 16, 1628.
(1927). Psychology and the soldier. London: Cambridge Uni-
and so on down a chain of participants. Bartlett ob-
versity Press.
served that the two methods yielded similar results: (1932; reprint 1964). Remembering: A study in experimental
Recall was not duplicative but represented a recon- and social psychology. London: Cambridge University Press.
struction of the original story or picture based on (1934). The problem of noise. London: Cambridge University
memories of key details; the reconstruction could be Press.
(1936). Autobiography. In C. Murchison, ed., History of psy-
biased by conventionalization and importation.
chology in autobiography, Vol. 3. Worcester, MA: Clark Universi-
Given the fact that recall was not simply a duplica- ty Press.
(1937). Cambridge, England: 18871937. American Journal
tion of the same pattern over and over again, Bartlett, of Psychology 50, 97110.
following the suggestion of his neurologist friend (1951). The mind at work and play. London: Allen and Unwin.
Henry Head, argued that memories were not stored (1958). Thinking: An experimental and social study. London:
as static traces waiting to be revived; instead they Allen and Unwin.
formed parts of large complexes, called schemata, in Broadbent, D. E. (1970). Obituary of Sir F. C. Bartlett. Biographical
Memoirs of Fellows of the Royal Society 16, 116.
which individual components could be changed any Crampton, C. (1978). The Cambridge school: The life, works, and influ-
time there was a retrieval act. He argued that if traces ence of James Ward, W. H. R. Rivers, C. S. Myers, and Sir Frederic
were lifeless entities waiting to be revived, we should Bartlett. Ph.D. diss., University of Edinburgh, Scotland.
BEHAVIORISM 57
Harris, A. D., and Zangwill, O. L. (1973). The writings of Sir Fred- functions through techniques of introspection. Wat-
eric Bartlett, C.B.E., F.R.S.: An annotated handlist. British son boldly rejected this, asserting that behavior, per
Journal of Psychology 64, 493510.
Saito, A., ed. (1999). Bartlett: Culture and cognition. New York: Rout-
se, is the proper domain of psychology. For Watson,
ledge. prediction and control of overt behavior, rather than
Zangwill, O. L. (1972). Remembering revisited. Quarterly Journal of introspection of mental processes, formed the basis
Experimental Psychology 24, 123138. for an objective, scientific psychology. Behavior was
David J. Murray to be analyzed into stimulus-response (S-R) units
without appeal to hypothetical activities of brain or
mind. The units could be of widely varying size, from
the relatively molecular eyeblink elicited by a flash of
light to the more molar shopping trip as response
BASAL FOREBRAIN
to an empty cupboard. Watson emphasized the conti-
See: AMNESIA, ORGANIC; GUIDE TO THE nuity between human and nonhuman species, and he
ANATOMY OF THE BRAIN stressed the importance of learning, in animals as well
as in humans, as the fundamental basis for under-
BASAL GANGLIA standing psychological process.
See: GUIDE TO THE ANATOMY OF THE BRAIN A neobehaviorism that came to the fore in the
1930s, that of Clark L. Hull and his student Kenneth
Spence, dominated until mid-century. Like Watson,
Hull described behavior as composed of S-R units,
but whereas Watson had presented S-R analyses as ad-
BEHAVIORISM justable in scale, the Hull-Spence approach focused
Most generally, behaviorism is a viewpoint that takes on molecular building blocks that were described as
psychological phenomena as physical activity rather forming chains of connecting events between envi-
than as belonging to a special domain of mental ronmental stimuli and observed behavior. These me-
events. For a behaviorist, then, psychology is the diating events included hypothetical (but presumably
study of behavior and its physical, mainly environ- physical) stimulus traces, covert responses, and re-
mental, determinants rather than of the nature of ex- sponse-produced stimuli. Learned S-R units were
perience or of mental process. Behaviorism originat- called habits. Hull contrived an elaborate theory
ed in natural-science traditions of the late nineteenth whose theorems and postulates, presented in geome-
century, and precursors of its methods and concepts ters style, were concerned with the formation of habit
developed at the turn of the century in the work of E. strength and with the mechanistic conversion of habit
L. Thorndike and Russian physiologist I. P. Pavlov, as strength into overt action. The theory was published
well as of several other psychologists and physiolo- as essentially complete in 1943. Although highly tout-
gists (Day, 1980; Herrnstein, 1969). ed, it proved ponderous, with numerous terms that
But behaviorism as a distinct viewpoint came to were difficult to evaluate; it fell of its own weight with-
be recognized with the publication of American psy- in a decade. Nevertheless, Hullian students gained
chologist John B. Watsons article Psychology as the dominant positions within academic psychology, and
Behaviorist Views It (1913). Identification of behav- elements of that approach can be discerned to this
iorism with the controversial Watson persists despite day in theorizing that rests on the metaphor of me-
the fact that it developed into several distinct tradi- chanical associative connections. Hulls emphasis on
tions that bear only a family resemblance to Watsons formal hypothesis testing, directed at hypothetical
views and to each other (Malone, 1990; Zuriff, 1985). constructs that are anchored to observable events as
The leading contemporary behaviorist position de- specified by operational definitions, also survives as
rives from the work of B. F. Skinner, which differs a methodological behaviorism (Skinner, 1945) that
from other behaviorisms in its detailed account of ver- has permeated much of psychology.
bal functioning and in its inclusion of activities such A counterpoint to Hulls views in the 1930s and
as thinking and feeling as behavior to be accounted 1940s was provided by Edward C. Tolman, who at-
for, while maintaining a primary focus on behavior- tempted to include purposive, intentional language
environment relations rather than upon processes in- within a behavioristic system. He invoked terms like
ferred as underlying those relations. purpose, expectation, and cognition to capture the larger-
Behaviorism originated in opposition to an or- scale, goal-oriented molar organization of behav-
thodox psychology that attempted to analyze con- ior. Tolman asserted that these terms need not imply
scious experience by focusing upon reports by observ- anything nonphysical or mentalistic; indeed, he em-
ers who were trained to examine their own mental ployed them in accounting for behavior of laboratory
58 BEHAVIORISM
rats as well as of humans. But Tolman undermined tion, examining its properties and its broad implica-
such disclaimers by characterizing his view as S-O-R tions. Empirically, he asked what would happen if
theory, with the O denoting a special role for pro- only some occurrences of a response are reinforced;
cesses within the behaving organism. The learning of he devised schedules of reinforcement to explore the
complex relationships, often characterized as cogni- many ways in which this can happen and their effects
tive maps, was said to mediate between environment on rates, patterning, and persistence of behavior.
and behavior. Critics of Tolmans account suggested Contemporary research has extended this to examine
that it left the organism buried in thought. To the issues such as the conditions of self-control and pref-
extent that he addressed the sources of action, Tol- erences among schedules that are relevant to microe-
man placed them within the organism, which tended conomics (e.g., Rachlin, 1989) and to biological theo-
to link his account with traditional mentalistic expla- ries of foraging (e.g., Fantino and Abarca, 1985).
nations of action. Thus it is not surprising that Tol- Interpretatively, Skinner addressed the functional
mans inclusion of intentional language never was ac- characteristics of verbal behavior, describing how an
cepted by the broader behavioristic community. individual affects the behavior of others and how oth-
B. F. Skinner also departed from the S-R behav- ers teach the individuals verbal discriminations
ioral mainstream of the 1930s and 1940s, but in dif- (Skinner, 1957). His approach initially was not wel-
ferent ways. He rejected mentalistic terms as mis- comed by linguists, but later developments in linguis-
leading fictions while including the relationships tics are more congenial to it (Andresen, 1990). The
that were Tolmans primary concern. Skinners first analysis includes activities like thinking, feeling, and
conceptual innovation was to reformulate the reflex; even introspecting as behavior to be accounted for
he described this simplest unit of behavior not as rather than as special bases for explaining overt ac-
stimulus-response connection but as directly observ- tion. It asserts that individuals know their private
able abstraction, a correlation between classes of stim- thoughts and feelings less well than they know exter-
uli and classes of responses. Then he distanced his nal events, because the world cannot as accurately
theory further from mediational notions of mecha- teach individuals to discriminate the former (Skinner,
nism and associative connection by delineating non- 1963). This provocative position gains independent
reflexive behavioral units that he called operants. Op- support from the philosophies of Ryle (Schnaitter,
erants act upon the environment; they are selected by 1985) and Wittgenstein (Day, 1969).
their consequences through processes denoted as re- Skinner also addressed ethical and social issues in
inforcement, punishment, and extinction. Operants can light of reinforcement-based principles, speculatively
range from small to large, and are defined by the con- in Walden Two, a utopian novel that sketches an ex-
sequences that shape or maintain them but also by the perimental approach to communal living, and analy-
contexts within which the selecting consequences tically in essays such as The Ethics of Helping Peo-
have occurred. The result is a three-term relationship ple (Skinner, 1978), which asserts that human rights
composed of classes of responses, consequences, and properly concern empowerment of effective action
discriminative stimuli. rather than access to things or services. Extensive dis-
Operant behavior often is characterized in ordi- cussions of these and other implications of reinforce-
nary language as intentional and purposive, thus hav- ment theory are provided by Skinner (1971, 1974),
ing the molar characteristics that were Tolmans and by Catania and Harnad (1988). Contemporary
primary concern. But the traditional appeal to men- behavioral research related to language emphasizes
talistic intention is replaced by environment-based se- relationships between verbal and nonverbal behavior
lection in this account of action, just as in Darwinian (Cerutti, 1989; Hayes, 1989) and issues such as the
biology natural selection replaces divine intention in nature and origins of symbolic functioning (Sidman,
the account of new species. Learning of new behavior 1986).
is readily demonstrated by rapidly shaping new pat- Of contemporary approaches, the most distinctly
terns through differential reinforcement and through behavioral one is behavior analysis. Extending from
gradual fading of discriminative stimuli. In later
Skinners work, it differs philosophically and concep-
work, Skinner (1984) described selectionist principles
tually from other behaviorisms as well as from main-
as applying to behavior patterns at the evolutionary
stream psychology (Lee, 1988). Its pragmatic contri-
level as well as at the level of cultural practice, giving
butions have proved effective in such diverse settings
accounts with close affinity to contemporary work in
as health-maintenance programs concerned with
anthropology (e.g. Lloyd, 1985) and biology
weight control, smoking, and wearing of automobile
(Dawkins, 1982; Smith, 1986).
seat belts; frequent flier marketing techniques
While his theory also included other principles, launched by airlines; techniques for basic research on
Skinner emphasized reinforcement as a basic rela- drug addiction as well as for its treatment; education-
BEHAVIOR THERAPY 59
al techniques of documented effectiveness for handi- (1945). The operational analysis of psychological terms.
capped and disadvantaged, as well as for mainstream, Psychological Review 52, 270277, 291294.
(1948). Walden two. New York: Macmillan.
children (Becker, 1978; Wolf et al., 1987); and inno- (1957). Verbal behavior. New York: Appleton-Century-
vative formats for personalized instruction at the col- Crofts.
lege level (Keller, 1977). Contemporary behaviorists (1963). Behaviorism at fifty. Science 134, 566602.
are represented by professional organizations that in- (1971). Beyond freedom and dignity. New York: Knopf.
clude several thousand researchers, scholars, and (1974). About behaviorism. New York: Knopf.
(1978). Reflections on behaviorism and society. Englewood
practitioners (Thompson, 1988) whose work is repre- Cliffs, NJ: Prentice-Hall.
sented by more than a score of primarily behavioral (1984). Selection by consequences. Science 213, 501504.
journals (Wyatt et al., 1986). Thus, behaviorism has Smith, T. L. (1986). Biology as allegory: A review of Elliott Sobers
extended well beyond, while continuing an apposi- The nature of selection. Journal of the Experimental Analysis of Be-
tional role within, the specialized field where it began. havior 46, 105112.
Thompson, T. (1988). Benedictus behavior analysis: B. F. Skinners
See also: OPERANT BEHAVIOR; SKINNER, B. F.; magnum opus at fifty. Contemporary Psychology 33, 397402.
Tolman, E. C. (1932). Purposive behavior in animals and men. New
TOLMAN, EDWARD C.
York: Appleton-Century-Crofts.
Watson, J. B. (1913). Psychology as the behaviorist views it. Psycho-
Bibliography
logical Review 20, 158177.
Andresen, J. T. (1990). Skinner and Chomsky thirty years later. Hi- (1919). Psychology from the standpoint of a behaviorist. Philadel-
storiographia Linguistica 17, 145165. phia: J. B. Lippincott.
Becker, W. C. (1978). The national evaluation of follow-through: Wolf, M. M., Braukmann, C. J., and Ramp, K. A. (1987). Serious
Behavior theory-based programs come out on top. Education delinquent behavior as part of a significantly handicapping
and Urban Society 10, 431458. condition: Cures and supporting environments. Journal of Ap-
Catania, A. C., and Harnad, S. (1988). The selection of behavior, the plied Behavior Analysis 20, 347359.
operant behaviorism of B. F. Skinner: Comments and consequences. Wyatt, W. J., Hawkins, R. P., and Davis, P. (1986). Behaviorism: Are
New York: Cambridge University Press. reports of its death exaggerated? The Behavior Analyst 9, 101
Cerutti, D. T. (1989). Discrimination theory of rule-governed be- 105.
havior. Journal of the Experimental Analysis of Behavior 51, 257 Zuriff, G. E. (1985). Behaviorism: A conceptual reconstruction. New
276. York: Columbia University Press.
Dawkins, R. (1982). The extended phenotype. San Francisco: Freeman.
Day, W. F. (1969). On certain similarities between the Philosophical Philip N. Hineline
investigations of Ludwig Wittgenstein and the operationism of
B. F. Skinner. Journal of the Experimental Analysis of Behavior 12,
489506.
(1980). The historical antecedents of contemporary behav-
iorism. In R. W. Rieber and K. Salzinger, eds., Psychology: The- BEHAVIOR THERAPY
oretical-historical perspectives, pp. 203262. New York: Academ-
ic Press. Behavior therapy is a term used to describe a number
Fantino, E., and Abarca, N. (1985). Choice, optimal foraging, and of therapeutic procedures that share certain assump-
the delayreduction hypothesis. Behavioral and Brain Sciences tions about the nature of behavioral and psychologi-
8, 315-362.
Hayes, S. C., ed. (1989). Rule-governed behavior: Cognition, contingen- cal problems and how they can best be overcome. The
cies, and instructional control. New York: Plenum. procedures can be classified into three main groups:
Herrnstein, R. J. (1969). Behaviorism. In D. L. Krantz, ed., Schools fear-reduction procedures, operant conditioning
of psychology: A symposium, pp. 5168. New York: Appleton- procedures, and aversive techniques.
Century-Crofts.
Hull, C. L. (1943). Principles of behavior. New York: Appleton- The major fear-reduction procedures consist of
Century-Crofts. systematic desensitization, in which the person is trained
Keller, F. S. (1977). Summers and sabbaticals: Selected papers on psy- to imagine a series of increasingly fearful images
chology and education. Champaign, IL: Research Press. while in a state of relaxation; therapeutic modeling, in
Lee, V. (1988). Beyond behaviorism. Hillsdale, NJ: Erlbaum.
Lloyd, K. E. (1985). Behavioral anthropology: A review of Marvin which the person observes and then imitates a thera-
Harriss Cultural materialism. Journal of the Experimental Analysis pist model engaging in increasingly close contact with
of Behavior 43, 279287. the frightening object or situation; and flooding, in
Malone, J. C. (1990). Theories of learning: A historical approach. Bel- which the phobic person is exposed to intensely fear-
mont, CA: Wadsworth. ful stimulus situations for prolonged periods. In all of
Rachlin, H. (1989). Judgment, decision, and choice: A cognitive/
behavioral synthesis. New York: W. H. Freeman. these methods, the phobic person is exposed to the
Schnaitter, R. (1985). The haunted clockwork: Reflections on Gil- real or imagined fear stimulus repeatedly and/or for
bert Ryles The concept of mind. Journal of the Experimental Analy- prolonged periods, and in all of them attempts to es-
sis of Behavior 43, 145153. cape from or avoid the fear stimulus are discouraged
Sidman, M. (1986). Functional analysis of emergent verbal classes. (response prevention). The combination of exposure
In T. Thompson and M. D. Zeiler, eds., Analysis and integration
of behavioral units, pp. 213245. Hillsdale, NJ: Erlbaum.
and response prevention has proved to be a robust
Skinner, B. F. (1938). The behavior of organisms: An experimental anal- and dependable means for reducing fear, and the
ysis. New York: Appleton-Century-Crofts. clinical efficacy of this combination in each of the
60 BEHAVIOR THERAPY
three forms of fear-reduction procedures has been desensitization remains useful for numerous prob-
confirmed in numerous controlled clinical trials lems, especially those in which direct exposure is im-
(Marks, 1987; OLeary and Wilson, 1987). practical or unacceptable, as in the treatment of cer-
tain sexual disorders and social phobias.
All three methods can be traced back to the work
of the Russian physiologist Ivan Petrovich Pavlov on All of the fear-reduction techniques are applica-
conditioning, and especially to his research on exper- tions of learning procedures to clinical problems,
imental neurosis. In developing the first of the mod- and, in common with the other forms of behavior
ern methods, systematic desensitization, Wolpe therapy, are based on the assumption that most psy-
(1958) was influenced by the writings of Pavlov and chological problems can be overcome by the use of
the modern learning theorists, especially Clark Hull. conditioning or other learning processes. In the early
Having rejected the psychodynamic approach, Wolpe stages of behavior therapy, it was assumed that most
attempted to apply modern learning techniques to psychological problems are the result of faulty learn-
psychological problems, particularly those in which ing (for instance, Symptoms are unadaptive re-
anxiety is prominent. After completing a series of ani- sponses and Symptoms are evidence of faulty learn-
mal experiments, he concluded that graded, gradual ing [Eysenck and Rachman, 1965, p. 12]).
re-exposures to a fearful stimulus are the best way to Furthermore, it was argued that problems that are the
weaken or eliminate the fear. He also concluded that result of faulty learning can be unlearned. In due
the fear-reduction process can be facilitated by the course, more complex explanations were substituted.
deliberate superimposition on the evoked fear re- The second form of behavior therapy, consisting
sponse of a competing incompatible response (such of the application of operant conditioning ideas and
as relaxation imposed on a fear response). Each occa- procedures to clinical problems, was engineered in
sion on which an incompatible response is imposed the United States and applied mainly to the psycho-
over the fear response is an instance of reciprocal in- logical problems of children and adults with severe
hibition. Wolpe argued that repeated instances of handicaps (e.g., mentally retarded people in or out of
such reciprocal inhibition give rise to a relatively per- institutions and people with chronic and serious psy-
manent form of conditioned inhibition (of fear), and chiatric disturbances, especially chronic schizophre-
that the therapeutic effects are a direct consequence nia). The fear-reduction techniques, developed main-
of the reciprocal inhibition. ly in Britain, are still used predominantly in dealing
Desensitization is the earliest and best- with adult neurotic problems, especially anxiety dis-
established of the methods, but Wolpe introduced a orders such as obsessional disorders, panic disorders,
number of other therapeutic procedures. Most atten- social phobias, and circumscribed phobias.
tion, however, has been devoted to desensitization, The clinical application of operant conditioning
which has been the subject of considerable experi- later referred to as reinforcement therapy was a direct
mental work and testing. This research advanced the application of Skinnerian ideas, with emphasis on the
therapeutic efficacy of these results and gave rise to consequences of behavior. Behavior that is followed
a fresh view of fear itself. by a reward will be strengthened and behavior that is
Additions to and improvements of the clinical followed by nonreward will be weakened.
techniques were introduced in the early 1970s, and The first application consisted of a series of at-
for certain types of anxiety disorders (such as panic tempts to treat schizophrenic problems in laboratory
disorder, agoraphobia, obsessional disorders, simple settings, but only limited success was achieved until
phobias), therapeutic modeling replaced desensitiza- the methods were applied, often with considerable in-
tion as the method of choice. In most cases, therapeu- genuity, to the maladaptive behavior of institutional-
tic modeling and flooding are carried out in vivo, an ized patients with chronic psychiatric disorders (e.g.,
unfortunately chosen term that here means exposure Ayllon and Azrin, 1968). Notable advances were
to the fear stimulus rather than to an imagined repre- made by the selective reinforcement of desirable be-
sentation of the stimulus (as in the desensitization havior (e.g., self-care, eating) and the withholding of
procedure). The development of therapeutic model- (social) reinforcement after undesirable behavior.
ing, largely the result of Banduras work (1969), con- Many of these useful advances were later incorporat-
sists of repeated exposures to the fear stimulus. The ed into an institutional program that Ayllon and
phobic person first watches and then imitates the ap- Azrin, two pioneers of this work, called the token econo-
proach behavior of a therapeutic model. The method my (exchangeable tokens having replaced tangible re-
is effective and well accepted by most subjects, clients, wards). The fact that the large ambitions of the earlier
and patients. Flooding is seldom the first choice of workers who expected reinforcement therapy to elim-
treatment but can be used in certain cases such as ex- inate these psychiatric problems were never achieved
tensive obsessional/compulsive problems. Systematic does not detract from the contribution that this form
BIRDSONG LEARNING 61
of therapy continues to make. It is widely used to OLeary, K., and Wilson, G. T. (1987). Behavior therapy, 2nd edi-
modify the maladaptive behavior of retarded chil- tion. Englewood Cliffs, NJ: Prentice-Hall.
Rachman, S. (1990). Fear and courage, 2nd edition. New York: W.
dren and adults, patients with chronic psychiatric dis- H. Freeman.
orders, children with speech or other behavioral defi- Wolpe, J. (1958). Psychotherapy by reciprocal inhibition. Stanford, CA:
cits, and in educational settings. Stanford University Press.
Aversion therapy is a direct application of Pavlovian Stanley J. Rachman
conditioning to appetitive but maladaptive behavior
such as the excessive use of alcohol and was originally
prompted by the discovery that laboratory animals
can develop conditioned nausea reactions to the stim- BIRDSONG LEARNING
uli that are associated with the administration of Song behavior refers to complex vocalizations used in
drugs that induce nausea. In order to convert the ab- the context of mate attraction and territorial defense.
errant stimuli (such as inappropriate sexual stimuli or Birds that produce such sounds are commonly called
alcohol) into conditioned stimuli for nausea or other songbirds. Technically songbirds constitute species in
unpleasant responses, the alcohol or sexual stimuli the avian order Passeriformes. This is by far the larg-
are contingently followed by aversive stimulation. est avian order and contains about half of the more
The earlier and still most common form of aversion than 9,000 living bird species. The songbird order,
therapy consists of pairing an alcohol-conditioned one of the most recently evolved, includes familiar
stimulus with aversive nausea induced by drugs. This avian groups such as sparrows, swallows, starlings, ca-
form of chemical aversion therapy is used mainly in naries, finches, warblers, jays, titmice, crows, wrens,
the treatment of alcoholism; the other technique, elec- robins, and buntings. This order can be further divid-
trical aversion therapy, in which the aversive stimulus is ed into two suborders, the Oscines (members of the
an electric shock, is used mainly in the treatment of suborder Passeres) and the sub-Oscines (a much smal-
aberrant sexual behavior such as pedophilia. Al- ler group that includes the flycatchers of North Amer-
though aversion therapy appears to make a useful ica), which appeared earlier in evolutionary history
contribution to dealing with alcohol and sexual prob- and is thought to be more primitive. All songbirds
lems, it is rarely sufficient and nowadays is used as produce complex vocalizations, but there do appear
part of a wider therapeutic program that typically in- to be qualitative differences between Oscine species
cludes counseling, group therapy, and family thera- and subOscine species in vocal development and in
py. This insufficiency, the ethical problems involved the neural substrate mediating vocal learning and
in the deliberate application of aversive stimuli, and production.
the fact that the precise nature of the conditioning
processes involved in aversion therapy is not fully un-
derstood have combined to limit its use. The Basics of Birdsong
The other two forms of behavior therapy also All songbirds have a repertoire of up to twenty or
have their limitations, so clinicians have been recep- so distinct vocal sounds that they use for communica-
tive to the growing influence of cognitive analyses of tion about danger, food, sex, group movements, and
psychological and clinical phenomena. Behavior for many other purposes. One can usually make a dis-
therapy has been expanded to include the influence tinction between a birds calls, which are usually brief
of cognitive factors, and most practitioners now favor and monosyllabic, and its songs, which are more ex-
cognitive behavior therapy, which combines the original tended patterns of sound and often tonal and melod-
forms of behavior therapy with cognitive analyses of ic. The decision to classify a vocalization as a song or
the problem and cognitive procedures for helping to a call is usually based on the perceived function. The
deal with them. Cognitive behavior therapy appears main functions that have been ascribed to song be-
to have achieved most success in the treatment of de- havior are territory defense (or spacing behavior) and
pression (Beck, 1976; OLeary and Wilson, 1987). mate attraction, as opposed to calls, which are in-
volved in such functions as signaling danger or food
Bibliography and maintaining flock cohesion. Songs, especially
Ayllon, T., and Azrin, N. (1968). The token economy. New York: Ap- among songbird species in the temperate zone, are
pleton.
Bandura, A. (1969). Principles of behavior modification. New York:
usually produced by males. Among tropical species
Holt. females as well as males often sing. In some species,
Beck, A. (1976). Cognitive therapy and the emotional disorders. New males and females sing a coordinated song that is
York: International University Press. known as a duet.
Eysenck, H., and Rachman, S. (1965). The causes and cures of narco-
sis. London: Routledge and Kegan Paul. Unlike most calls, songs are learned: They devel-
Marks, I. (1987). Fears, phobias, and rituals. Oxford: Oxford Univer- op abnormally if a young male is reared out of hear-
sity Press. ing of the sounds of adults (see Figure 1). Among spe-
62 BIRDSONG LEARNING
cies in the songbird order only the Oscines clearly to rely on learning for vocal development. The avian
learn their songs; in the few sub-Oscine species inves- groups known to have learned songs include hum-
tigated song learning does not appear to occur thus mingbirds, parrots, and all Oscine songbirds.
providing a natural comparison between closely relat-
ed species that vary in one key aspect of their vocal
behavior. A common consequence of this dependence Sensitive Periods and the Timing of Song
on learning among Oscines is the emergence of local Learning
song dialects, varying on much the same geographic There is an underlying pattern in the steps typi-
scale as dialects in human speech. These dialect cally required in learning to sing. First is the acquisi-
boundaries have been shown to serve as imperfect tion phase, when a bird hears songs and commits
barriers for gene flow in species such as the white- some to memory. These are stored for a period that
crowned sparrow indicating that this intraspecific varies in duration from species to speciesfrom days
geographic variation is meaningful (MacDougall- to monthsuntil the bird begins to recall songs from
Shackleton and MacDougall-Shackleton, 2001). memory and starts to produce imitations, sometimes
faithful to the original model, sometimes far differ-
Songbirds are unique among animals in the many
ent. Thus there is a separation in time between the ac-
analogies that can be struck between song learning
quisition or sensory phase and the production or sen-
and the acquisition of human speech (Doupe and
sorimotor phase, ending with the production of
Kuhl, 1999). Avian vocal development provides one
crystallized, adult song (Marler, 1997).
of the few tractable animal models for studying the
behavioral, hormonal, and neural bases of vocal plas- There are often sensitive periods for song acquisi-
ticity (Ball and Hulse, 1998). No known nonhuman tion, sometimes restricted to a short period early in
primate depends upon learning to develop its natural life, and sometimes extending into adulthood. Even
vocal repertoire. Other than humans, cetaceans and close relatives, such as sparrows and canaries, may dif-
perhaps some bats are the only mammals that appear fer in this respect. Several species of sparrows have a
sensitive period for acquiring songs beginning at or guided learning. In several sparrow species, for
about twenty days of age, soon after young males be- example, species-typical vocalizations appear to be
come independent from their parents, and ending privileged because they are learned preferentially.
four to six weeks later (Nelson, 1997). Such sensitive Such preferences have even been observed for specif-
periods for learning are variable within limits, de- ic subspecies (Nelson, 2000). Another aspect of this
pending on the strength of song stimulation and the memory, strongly advocated by Peter Marler and col-
influence of physiological factors, such as hormonal leagues, is that the memory encodes species-typical
states, that vary with the season. If young are hatched patterns of vocal behavior even before it hears the
late in the season and singing has already ceased for song of its own species. This idea of an innate specifi-
that year, closure of the sensitive period may be de- cation of species-typical vocalizations would be one
layed until the following spring. The experimental way to explain why even in isolated-reared sparrows
withholding of stimulation by songs of the birds own many species-typical attributes of their vocalizations
species can also delay closure of the learning (Kroods- are still apparent in the abnormal songs produced by
ma and Pickert, 1980; Baptista and Gaunt, 1997). these birds. According to this view of song learning,
Species that learn song only during a sensitive period the formation of the auditory memory guiding song
early in ontogeny are known as age-limited learners. learning would result from a memorization by selec-
Species that continue to learn new songs throughout tive activation and attrition of innate circuits rather
their lives (such as canaries and European starlings) than from a selective instruction process (Marler,
are referred to as age-independent learners. In addi- 1997). Researchers have not yet resolved the relative
tion, there are species that fall in between these ex- validity of these two opposing models of auditory
treme cases in terms of song development. memory formation.
Effects of Isolation and Deafness Song Overproduction, Social Interactions,

Regardless of whether or not they have had the and Action-Based Learning
chance to learn, songbirds can always produce some
The young male songbird typically begins singing
aspects of the normal song of their species. When
sometime after he has memorized learned songs, but
sparrows are raised in isolation, for example, the note
the imitations are not fully formed. Instead, the
structure and tonal quality of their songs is abnormal,
young male starts with subsong, an amorphous, noisy
but each species still produces some basic features of
twittering that changes gradually, first into plastic
normal song syntax (see Figure 1). These features
song, which also is highly variable but contains the
are produced irrespective of whether they were repre-
first obvious signs of mature song structure, finally
sented in any songs a male may have learned.
crystallizing into the stable patterns of mature, adult
Insight into the basis of this ability to produce song (see Figure 2). Subsong resembles the babbling
certain normal song features is gained by studying the of human infants and may be important for develop-
songs of deaf birds, which are highly abnormal and ing the motor skills of singing and other prerequisites
variable in structure (see Figure 1). This degraded for song learning, such as the ability to guide the
form of singing is observed both if a male becomes voice by the ear. Rehearsal of previously memorized
deaf before song stimulation and if he is deafened song patterns begins in plastic song. Often more plas-
after song stimulation but before the development of tic song themes are produced than are needed for
singing. There seems to be no internal brain circuitry mature singing (see Figure 3), and many are discard-
that makes memorized songs directly available to ed when song crystallization occurs. Interactions be-
guide motor development. To transform a memo- tween a male and his neighbors can influence which
rized song into a produced song, the bird must be of the overproduced songs will be selected for crystal-
able to hear its own voice. One can infer that there are lization and inclusion into the final repertoire (Nel-
auditory memories for song, involved in guiding song son, 1997). Songs most similar to the neighbors
production, conceived of as neural mechanisms in the songs are retained; those that are different are reject-
auditory circuitry of the songbird brain that must vary ed. It is as if a process akin to operant conditioning
in their specifications from species to species (Kon- influences which songs are selected for later crystalli-
ishi, 1965; Marler, 1997). zation among those in auditory memory. Marler and
Douglas A. Nelson (1993) refer to this process as ac-
tion-based learning to contrast it with memory-
The Nature of the Auditory Memory based learning. The latter refers to cases where crys-
Guiding Song Learning tallized songs are produced in reference to previously
One aspect of the process of memory formation formed memories independently of social interac-
needed for song learning is that it involves selective tions during the sensorimotor period.
then to nXIIts constitutes a motor pathway special-

ized for song production. RA also projects to brain-
stem areas controlling respiration, which needs to be
coordinated with song (Wild, 1997). The anterior
forebrain song nuclei lMAN (the lateral part of the
magnocellular nucleus of the anterior neostriatum)
and area X, as well as the thalamic nucleus that inter-
connects them, DLM (the medial portion of the dor-
solateral nucleus of the thalamus), form a second
pathway connecting HVc to RA. All of the song nuclei
are present bilaterally, and the thalamic song nucleus
Uva (the uvaeform nucleus) forms a connection be-
tween the two sides.
The presence of a song system is particularly as-
sociated with the learning of song, not just with song
production. These brain areas are found only in birds
that sing and that learn their song by reference to au-
ditory information. For example, sub-Oscine song-
birds that do not learn their vocalizations do not seem
to have telencephalic song nuclei such as HVc or RA.
The vast majority of vocal learners are Passerine
songbirds, but evolutionarily unrelated species that
are vocal mimics, such as parrots or hummingbirds,
which require conspecific learning to produce their
complex vocalizations, are also found to have fore-
brain structures resembling nuclei in the song system
including HVc and RA (Jarvis and Mello, 2000; Jarvis
et al., 2000; Gahr, 2000).
A Motor Pathway for Song

Evidence for a specialized motor pathway for
song comes from three sources. Behavioral studies
show that lesions of HVc, RA, or the hypoglossal
nerve cause severe disruption of adult song. Also,
electrophysiological recordings in Nif, HVc, and RA
of awake birds reveal neurons whose activity is highly
correlated with singing (Yu and Margoliash, 1996).
These data indicate a hierarchy in motor processing
with activity in HVc correlated with syllable produc-
tion and activity in RA correlated with note produc-
The Song System tion (Yu and Margoliash, 1996). Nif may be the bird-
A discrete set of brain areas discovered by F. Not- song pattern generator, because section of the tract
tebohm (Nottebohm, 1996), often called the song sys- from Nif to HVc destroys the patterning of song,
tem, controls song learning and production. The while lesions of Uva do not (McCasland, 1987).
song system suggests a possible location for the neural Motor-driven immediate early gene induction also
changes associated with song learning. The song sys- occurs in nuclei such as HVc and RA in that song pro-
tem consists of numerous interconnected nuclei and duction is correlated with such induction in deaf birds
occupies a relatively large volume of brain (see Figure induced to sing (Jarvis and Nottebohm, 1997).
4). The brain area nXIIts (the tracheosyringeal por- The song motor commands must ultimately be
tion of the hypoglossal nucleus) contains the motor translated into a pattern of syringeal muscle activa-
neurons that control the musculature of the birds tion and respiratory control. The output nucleus of
vocal organ, the syrinx. The pathway from Nif (the the song system, nXIIts, has a map of the muscles of
nucleus interface) through HVc (or high vocal center) the syrinx: All motor neurons projecting to a particu-
to RA (the robust nucleus of the archistriatum) and lar syringeal muscle are clustered together in discrete
Figure 3
Diagram illustrating the process of song development in the male swamp sparrow. The horizontal axis represents the birds age from
hatching in months; the successive rows in the vertical axis represent the different processes involved in establishing the adult
crystallized song and how these processes were experimentally investigated. A male successively exposed to different song types at
various ages (top row, A through G) will only memorize song types presented during the sensory acquisition phase. After a period of
subsong during which no song type is recognizable, the bird will start producing during the period of plastic song all song types that
were stored in memory. Selective attrition will then take place, in part as a consequence of social interactions with neighbors, and
crystallized adult song will then only include one or two selected song types.
zones within the motor nucleus. RA also has zones of two separate sources are used to produce a species-
premotor neurons that project to the corresponding typical vocalization. Thus initial claims that control of
muscle control area in nXIIts. The dorsal portion of song production is lateralized similarly to speech pro-
RA does not project to nXIIts but to the dorsomedial duction is an oversimplification. It seems that asym-
nucleus (DM) of the midbrain, an area thought to be metries in syringeal control when they do occur re-
involved in vocalization and respiration in all birds, flect a peripheral asymmetry rather than a
including nonsongbirds (see Figure 4; Vicario, 1991). hemispheric specialization, as is the case in humans.
Neurons in DM then project to nXIIts, suggesting
that RA has two parallel and perhaps functionally dif-
Song-Selective Auditory Pathways and
ferent inputs to the syringeal motor neurons.
Feedback Mechanisms
Intrabronchial measurements of both airflow and Because auditory feedback is used to correct vocal
sound from each syringeal side have clarified how the motor output during song learning, there must be a
syrinx functions to produce sound (Suthers, Goller, link between the auditory system and the vocal motor
and Pytte, 1999). The syrinx contains two separate pathway. In addition, there must be brain mecha-
sound sources, each with an independent motor con- nisms for very specific recognition of song and song-
trol. There is extensive species variability in how these like vocalizations. Auditory information is relayed to
Figure 4
Schematic representation of the song system in Oscines. The diagram illustrates many of the known nuclei and connections in the
song system. Different fill patterns have been used to highlight nuclei that are part of the motor pathway (black), the anterior
forebrain pathway mostly involved in song learning (cross-hatching) and the parts of the auditory pathway that convey auditory
information to nuclei of the song system (white).
the songbird forebrain as in all other birds, traveling auditory stimuli. Some of these HVc auditory neurons
from the cochlear nuclei through the thalamus to the are song-selective: They respond best of all to the
forebrain primary auditory area, Field L. Anatomical birds own song, even in comparison with very similar
experiments show that subdivisions of Field L proj- songs of conspecifics (Margoliash, 1986). HVc song-
ects to the vicinity of HVc and RA (Vates, Broome, selective neurons are also sensitive to temporal order:
Mello, and Nottebohm, 1996; see Figure 4). Howev- They are activated more strongly by the birds own
er, electrophysiological investigations also suggest song when the syllables are in the normal sequence
that auditory information gains access to HVc via Nif than when the identical song components are played
(Janata and Margoliash, 1999). Field L projects to the out of order or in reverse. This high degree of selec-
caudolateral ventral hyperstriatum (clHV) that in tivity in individual neurons provides a possible mech-
turn projects to Nif (Vates, Broome, Mello, and Not- anism for specific recognition of song. Because the
tebohm, 1996). In addition to song-related motor birds own song is learned during development, this
neurons, HVc also contains neurons that respond to song selectivity must also be learned. In fact, studies
in young birds have shown that these neurons acquire learning did not require slow NMDA-EPSCs at syn-
their specificity during sensorimotor learning. apses critical for song development.
Lesion studies have shown that the anterior fore-
brain pathway containing MAN and X (see Figure 4) The Development of the Song System,
plays an important role in song learning but is not an Hormonal Effects, and Sex Differences in
essential part of the motor pathway for song in the Brain and Behavior
adult. Lesions of MAN or X have no apparent effect It is a striking feature of the song system that it
on adult song production, but destruction of either of continues to develop after hatching, during song
these areas in young birds results in markedly abnor- learning. Administration of 3H-thymidine can be
mal song (Bottjer and Johnson, 1997). There are used to label neurons undergoing their last cell divi-
song-selective auditory neurons, similar to those in sion, or birthdate. Such birthdating of song nuclei
HVc, in every nucleus in this loopX, DLM, and shows that MAN and RA are born before hatching,
MAN. Like the neurons in HVc, these auditory neu- but that there is significant neurogenesis in HVc and
rons acquire their song selectivity in parallel with X in the first several months after hatching (Alvarez-
song acquisition. If the syrinx is selectively dennervat- Buylla and Kirn, 1997). There is also naturally occur-
ed prior to the sensorimotor phase of song learning, ring cell death during postnatal development. In
male zebra finches in many cases are unable to match male zebra finches, many MAN neurons die around
their vocal output to the tutors song (Solis and five weeks after hatching. Synaptic connectivity con-
Doupe, 2000). In such birds, many neurons in area X tinues to develop at these late stages. The motor pro-
exhibited a dual selectivity and responded equally jection from HVc reaches its target nucleus RA by
well to the birds own song and to the tutor song. The postnatal day fifteen, but then waits outside RA for
degree of selectivity for these stimuli as compared to about ten days before growing in and completing the
conspecific song or reversed song was considerably circuit (Konishi and Akutagawa, 1985). Interestingly,
less than in normal adults (Solis and Doupe, 2000). male zebra finches first begin to sing at the time of in-
growth of HVc axons into RA. Connections from HVc
One possible function of this song-selective audi-
to MAN via X and DLM are present and functional
tory pathway is to act as a correction signal provided
by day fifteen, but topographic features of the projec-
by auditory feedback for the learning of song and the
tion from lMAN to RA occur at days twenty to twenty-
maintenance of learned song. There is evidence that
five during the early stages of vocal learning. Timing
the delayed auditory feedback can cause a gradual de-
of this projection is delayed if juvenile birds are de-
terioration of adult zebra finch song (Leonardo and
prived of normal conspecific auditory experience
Konishi, 1999). Deafening adult male zebra finches
(Iyengar and Bottjer, 2002).
also results in a deterioration of song (Nordeen and
Nordeen, 1992). However, lesioning a nucleus in the Bird song can vary among the sexes. Typically in
anterior forebrain pathway such as lMAN prevents temperate zone species, male birds sing for courtship
the deterioration in song resulting from deafening and territorial defense, while female birds sing much
(Brainard and Doupe, 2000). These findings suggest less or not at all. In the tropics the pattern is quite dif-
that an active process is required for the maintenance ferent: males and females often stay together on terri-
of song and that the anterior forebrain pathway is es- tories all year round, and both sexes will sing to de-
sential to this process. fend these territories. The behavioral dimorphism
between males and females has a striking correlate in
The song-selective auditory loop from HVc to the the sexual dimorphism of the song system itself. Ini-
anterior forebrain eventually projects back into the tial studies of species the exhibit extreme differences
motor pathway through its connection to RA (see Fig- in song behavior: For example, zebra finches and ca-
ure 4). This convergence of auditory and motor in- naries revealed marked male-biased differences be-
puts makes RA a possible site for the auditory guid- tween males and females (Nottebohm and Arnold,
ance of vocal motor development during learning. 1976). Nottebohm and Arnolds discovery provided
The NMDA subclass of glutamate receptors is an opportunity to explore sex differences in the brain
thought to play a role in some forms of synaptic plas- in a truly comparative sense. Comparative studies
ticity. NMDA receptor-mediated EPSCs (NMDA- have been employed to understand the function of
EPSCs) become fast during song development, a tran- these sex differences in the brain (see Figure 5). In
sition posited to limit learning. However, manipula- some songbird species, females sing rarely or not at
tions of the sensitive period for song learning by all, and the brain nuclei that control song are many
isolating nestling zebra finches delayed NMDA- times larger in volume in males than in females. In
EPSCs but did not prevent the birds from learning other species, males and females sing approximately
(Livingston, White, and Mooney, 2000). Thus song equally, and the brain nuclei that control song are ap-
Figure 5
Interspecific variations in the volumetric sex difference of the song control nucleus HVc in songbirds and their relationship with the
sex difference in singing behavior. The figure illustrates the ratio of the HVc volume in males and females in species where the female
normally never sings (very large HVC in males compared to females), where females sing but less than males (intermediate ratio of
HVc volumes), and where females sing about as often as males (duetting species, ratio of HVc volumes is close to 1). The volume ratio
in three species of the first group has been indicated as approximately infinite () because HVc cannot be detected in females.
proximately equal between the sexes. Recently statis- Steroid hormones influence both singing and the
tical methods have been employed control for phylo- song system. Singing is much increased when andro-
genetic effects while comparing the coevolution of gen levels are high, for instance during the breeding
traits. This analysis indicates that the evolution of sex season in the spring. In several songbird species,
differences in song has coevolved with the evolution there is also a seasonal variation in the size of the song
of sex differences in singing behavior in songbird nuclei (Tramontin and Brenowitz, 2000). In canaries,
species (Figure 5; MacDougall-Shackleton and Ball, where researchers first described this trait, HVc and
1999). It is unclear whether these morphological RA enlarge by approximately 50 percent in the spring
differences related to variation in song are just (Nottebohm, 1981). These seasonal changes in vol-
related to differences in production or if differences ume involve changes in cell size and cell number
in song learning also occur. Sex differences in (Tramontin and Brenowitz, 2000). These seasonal
volume in nuclei such as HVc and RA are asso- morphological changes were initially thought to be
ciated with differences in cell size and cell num- related to seasonal changes in song learning but cur-
ber. Other attributes of the phenotype of cells in rent data support the notion that the changes are
these nuclei are different in males and females, such more closely related to seasonal changes in song per-
as the number of cells expressing androgen recep- formance (Tramontin and Brenowitz, 2000). In some
tors. species, such as canaries and white-crowned sparrows,
adult females will respond to the administration of and synaptic changes. There is a neural reorganiza-
testosterone by beginning to sing and rapidly going tion that includes massive synaptogenesis associated
through the sensorimotor phase of learning. These with the incorporation of new neurons into the vocal
injections also induce marked growth of the song nu- pathways. Behavioral evidence clearly implicates
clei. The effects of testosterone in stimulating these NMDA receptor activation in specific song nuclei as
cellular changes in HVc are mediated at least in part being required for song learning (Basham, Nordeen,
via the upregulation of brain-derived neurotrophic and Nordeen, 1996). However, scientists have yet to
factor, or BDNF (Rasika, Alvarez-Buylla, and Notte- identify the precise cellular events that occur develop-
bohm, 1999). Interestingly, there is also evidence that mentally that mediate song learning rather than the
BDNF is released in HVc in response to singing itself. maturation of the song system.
Thus testosterone can promote morphological
changes in HVc by acting directly on cells in that nu-
cleus or by acting elsewhere in the brain (such as the Conclusion
preoptic area) to induce increased singing behavior Bird song is a complex motor behavior that is
that results in increased BDNF in HVc (Ball, Riters, learned by matching vocal output to an auditory
and Balthazart, 2002). memory. This memory has an innate component but
is also modified by experience early in ontogeny. In
The influence of sex steroids on the development
some species, only songs learned early in life are pro-
of the song system has been extensively studied in
duced in adulthood. In other species, learning con-
zebra finches, where the differences between the male
tinues throughout adult life. Birds tend to learn more
and female song systems are especially pronounced
songs than will be used throughout adult life, and
(see Figure 5). Female zebra finches have very small
these are selected based on social interactions among
and shrunken song nuclei and, unlike canaries, do
conspecifics in the area they settle.
not respond with song to testosterone administration
in adulthood. If given estrogen early in life, however, A specialized neural circuit has evolved in associa-
female zebra finch chicks develop masculinized song tion with the occurrence of song learning. This circuit
nuclei (Schlinger, 1998). Based on many studies of contains nuclei primarily involved in motor produc-
zebra finches, researchers are clear that this differen- tion or in the auditory feedback needed for the learn-
tiation process that occurs early in ontogeny is not the ing and maintenance of song. Neurons within this cir-
result of sex differences in gonadal steroid hormone cuit appear to be feature detectors that exhibit highly
action as is the case in the mammalian brain and in selective response to conspecific song. The selectivity
other aspects of sexually dimorphic behaviors in birds of these neurons develops in parallel with song learn-
(Schlinger, 1998). Rather, it is either the result of es- ing. The morphology of the song system varies be-
trogen synthesized by the brain acting early in ontog- tween males and females in a systematic fashion
eny to masculinize the system (Holloway and Clayton, among different species that reflects species-
2001) or of sex differences in gene expression in the variability in song behavior. Steroid hormones regu-
brain that is triggered independently of the gonad. late these sex differences as well as the striking sea-
sonal plasticity in adulthood. Songbirds exhibit adult
Another unusual feature of the songbird brain neurogenesis that contributes to the unusual adult
and avian brains in general is that neurogenesis con- plasticity. Scientists have described cellular changes
tinues to occur in portions of the periventricular zone in synaptogenesis and neuronal incorporation that
in adult birds, and these newly generated neurons mi- correlate with song learning. This system will contin-
grate out into the forebrain and are incorporated into ue to be a valuable model for the investigation of the
the neural circuitry. Some of these neurons in HVc neurobiology of species-typical learning.
project to RA (Alvarez-Buylla and Kirn, 1997). Re-
searchers are not clear, however, what significance See also: IMPRINTING; NEUROGENESIS
this phenomenon has for song learning. The new
neurons occur throughout the forebrain, in females Bibliography
as well as males. They are found in one-time learners Alvarez-Buylla, A., and Kirn, J. R. (1997). Birth, migration, incor-
such as zebra finches as well as in open-ended learn- poration, and death of vocal control neurons in adult song-
ers like canaries, and even in nonsongbirds. Nonethe- birds. Journal of Neurobiology 33, 585601.
Ball, G. F., and Hulse, S. H. (1998). Bird song. Am. Psych. 53, 37
less, this phenomenon suggests that adult birds retain
58.
a remarkable ability to generate new neurons as well Ball, G. F., Riters, L. V., and Balthazart, J. (2002). Neuroen-
as a preservation of the cues for axon guidance and docrinology of song behavior and avian brain plasticity: Mul-
neuronal specification in their brains. tiple sites of action of sex steroid hormones. Frontiers in
Neuroendocrinology 23, 137178.
Vocal learning early in development during the Baptista, L. F., and Gaunt, S. L. L. (1997). Social interactions and
sensory phase is associated with prominent cellular vocal development in birds. In C. T. Snowdon and M. Haus-
70 BLOCKING
berger, eds., Social influences on vocal development. Cambridge, McCasland, J. S. (1987). Neuronal control of bird song production.
UK: Cambridge University Press. Journal of Neuroscience 7, 2339.
Basham, M. E., Nordeen, E. J., and Nordeen, K. W. (1996). Block- Mooney, R., and Konishi, M. (1991). Two distinct inputs to an
ade of NMDA receptors in the anterior forebrain impairs sen- avian song nucleus activate different glutamate receptor sub-
sory acquisition in the zebra finch (Poephila guttata). Neuro- types on individual neurons. Proceedings of the National Acade-
biol. Learn. Mem. 66, 295304. my of Sciences of the United States of America 88, 4,0754,079.
Bottjer, S. W., and Johnson, F. (1997). Circuits, hormones, and Nelson, D. A. (1997). Social interactions and sensitive phases for
learning: Vocal behavior in songbirds. Journal of Neurobiology song learning: A critical review. In C. T. Snowdon and M.
33, 602618. Hausberger, eds., Social influences on vocal development. Cam-
bridge, UK: Cambridge University Press.
Brainard, M. S., and Doupe, A. J. (2000). Interruption of a basal
(2000). A preference for own-subspecies song guides vocal
ganglia-forebrain circuit prevents plasticity of learned vocal-
learning in a song bird. Proceedings of the National Academy of
izations. Nature 404, 762766.
Sciences of the United States of America 97 (24), 13,34813,353.
Doupe, A. J., and Kuhl, P. K. (1999). Birdsong and human speech:
Nordeen, K. W., and Nordeen, E. J. (1992). Auditory feedback in
Common themes and mechanisms. Annual Review of Neuro-
necessary for the maintenance of stereotyped song in adult
science 22, 567631.
zebra finches. Behavioral and Neural Biology 57, 5866.
Gahr, M. (2000). Neural song control system of hummingbirds:
Nottebohm, F. (1981). A brain for all seasons: Cyclical anatomical
Comparison to swifts, vocal learning (songbirds) and non- changes in song-control nuclei of the canary brain. Science
learning (suboscines) passerines, and vocal learning (budgeri- 214, 1,3681,370.
gars) and nonlearning (dove, owl, gull, quail, chicken) non- (1996). The King Solomon lectures in neuroethology: A
passerines. Journal of Comparative Neurology 426, 182196. white canary on Mount Acropolis. Journal of Comparative
Holloway, C. C., and Clayton, D. E. (2001). Estrogen synthesis in Neurophysiology A. 179, 149156.
the male brain triggers development of the avian song control Nottebohm, F., and Arnold, A. P. (1976). Sexual dimorphism in
pathway in vitro. Nature Neuroscience 4 (2), 170175. the vocal control areas in the song bird brain. Science 194,
Iyengar, S., and Bottjer, S. W. (2002). The role of auditory experi- 211213.
ence in the formation of neural circuits underlying vocal Rasika, S., Alvarez-Buylla, A., and Nottebohm, F. (1999). BDNF
learning in zebra finches. Journal of Neuroscience 22 (3), 946 mediates the effects of testosterone on the survival of new
958. neurons in an adult brain. Neuron 22, 5362.
Janata, P, and Margoliash, D. (1999). Gradual emergence of song Schlinger, B. A. (1998). Sexual differentiation of avian brain and
selectivity in sensorimotor structures of the male zebra finch behavior: Current views on gonadal hormone-dependent and
song system. Journal of Neuroscience 19, 5,1085,118. independent mechanisms. Annual Review of Physiology 60,
Jarvis, E. D., and Mello, C. V. (2000). Molecular mapping of brain 407429.
areas involved in parrot vocal communication. Journal of Com- Solis, M. M., and Doupe, A. J. (2000). Compromised neural selec-
parative Neurology 419, 131. tivity for song in birds with impaired sensorimotor learning.
Jarvis, E. D., and Nottebohm, F. (1997). Motor-driven gene ex- Neuron 25 (1), 109121.
pression. Proceedings of the National Academy of Sciences of the Suthers, R. A., Goller, F., and Pytte, C. (1999). The neuromuscular
United States of America 94, 4,0974,102. control of birdsong. Philosophical Transactions of the Royal Soci-
Jarvis, E. D., Ribeiro, S., Da Silva, M. L., Ventura, D., Vielliard, J., ety of London [Biol.] 354, 927939.
and Mello, C. V. (2000). Behaviourally driven gene expres- Tramontin, A. D., and Brenowitz, E. A. (2000). Seasonal plasticity
sion reveals song nuclei in hummingbird brain. Nature 406, in the adult brain. TINS 23, 251258.
628632. Vates, G. E., Broome, B. M., Mello, C. V., and Nottebohm, F.
Konishi, M., and Akutagawa, E. (1985). Neuronal growth, atrophy (1996). Auditory pathways of caudal telencephalon and their
and death in a sexually dimorphic song nucleus in the zebra relation to the song system of adult male zebra finches
finch brain. Nature 315, 145147. (Taenopygia guttata). Journal of Comparative Neurology 366,
Leonardo, A., and Konishi, M. (1999). Decrystallization of adult 613642.
birdsong by perturbation of auditory feedback. Nature 399 Vicario, D. S. (1991). Organization of the zebra finch song control
(6,735), 466470. system II: Functional organization of outputs from nucleus
Robustus archistriatalis. Journal of Comparative Neurology 309,
Livingston, F. S., White, S. A., and Mooney, R. (2000). Slow
486494.
NMDA-EPSCs at synapses critical for song development are
Wild, J. M. (1997). Neural pathways for the control of birdsong
not required for song learning in zebra finches. Nature Neuro-
production. Journal of Neurobiology 33, 653670.
science 3,482488.
Yu, A. C., and Margoliash, D. (1996). Temporal hierarchical con-
MacDougall-Shackleton, E. A., and MacDougall-Shackleton, S. A.
trol of singing in birds. Science 273, 1,8711,875.
(2001). Cultural and genetic evolution in mountain white-
crowned sparrows: Song dialects are associated with popula- Peter R. Marler
tion structure. Evolution 55 (12), 2,5682,575. Allison J. Doupe
MacDougall-Shackleton, S. A., and Ball, G. F. (1999). Comparative
Revised by Gregory F. Ball and Jacques Balthazart
studies of sex differences in the song-control system of song-
birds. Trends in Neuroscience 22, 432436.
Margoliash, D. (1986). Preference for autogenous song by auditory
neurons in a song system nucleus of the white-crowned spar-
row. Journal of Neuroscience 6, 1,6431,661. BLOCKING
Marler, P. (1997). Three models of song learning: Evidence from
behavior. Journal of Neurobiology 33, 501516. See: CONDITIONING, CELLULAR AND NETWORK
Marler, P., and Nelson, D. A. (1993). Action-based learning: A new SCHEMES FOR HIGHER-ORDER FEATURES OF;
form of developmental plasticity in bird song. Netherlands KAMINS BLOCKING EFFECT: NEURONAL
Journal of Zoology 43 (12), 91103. SUBSTRATES
C
CAJAL, RAMN Y SANTIAGO How Memory Develops During Childhood
See: RAMN Y CAJAL, SANTIAGO and Adolescence
Two factors play a critical role in the develop-
ment of memory during childhood and adolescence:
greater use of strategies and greater task-relevant
CATEGORIES
knowledge.
See: CONCEPTS AND CATEGORIES, LEARNING OF
Developmental Change in Use of Memory
Strategies
A strategy is any deliberate act that is designed to
CEREBELLUM improve retention. One common strategy is rehears-
See: GUIDE TO THE ANATOMY OF THE BRAIN; al, in which people spontaneously repeat (either
NEURAL COMPUTATION aloud or silently) information to be remembered. As
children grow they are more likely to use strategies to
help them remember. Rehearsal, for example, typi-
CEREBRAL NEOCORTEX cally emerges at about six or seven years of age. In ad-
dition, as children and adolescents develop, they tend
See: GUIDE TO THE ANATOMY OF THE BRAIN to abandon simple but relatively ineffective strategies
in favor of complicated but more effective strategies.
In the case of rehearsal, school-age children often
simply repeat the information exactly as it was pres-
ented. In contrast, older children and adolescents are
more likely to modify their rehearsal as necessary to
CHILDREN, DEVELOPMENT OF fit the nature of the material. Instead of repeating
MEMORY IN words in the order presented, older children and ado-
Memory improves dramatically as children develop. lescents may reorganize the words, rehearsing togeth-
The first section of this entry describes the factors that er words that are related semantically.
contribute to improved memory during childhood The general trend for children to use strategies
and adolescence. With this general developmental more often and more effectively as they grow appar-
profile as a backdrop, several special topics in chil- ently reflects parallel developmental changes in chil-
drens memory are discussed in the remainder of the drens ability to diagnose memory tasks and to moni-
entry. tor the effectiveness of their chosen strategy. That is,
71
72 CHILDREN, DEVELOPMENT OF MEMORY IN
Figure 1 that they are more likely to be able to provide a code

that reduces the amount of material that must be re-
membered.
Another way that childrens growing knowledge
benefits memory is by providing more retrieval cues.
Each of the links in Figure 1 provides an alternative
way, a cue, to gain access to a word that was presented.
Thus, the more links the better. When people are ex-
pert in some domain, their knowledge will be linked
extensively. If they know little, their knowledge will
have relatively few links. Consequently, older chil-
dren and adolescents will typically have more cues
available to recall information than will younger chil-
dren.
Knowledge is not always beneficial, because it can
distort memories. Information to be remembered is
often changed so that it conforms to ones existing
knowledge or stereotypes. To illustrate, when chil-
dren hear stories in which superheroes are said to be
weak or ugly, they tend to recall the heroes as strong
and attractive. Similarly, if asked to remember the ac-
tions of boys and girls whose behaviors violate sex-
role stereotypes (e.g., a girl sawing wood), children
will change the sex of the actor to make it consistent
A representation of a portion of a persons knowledge of with sex-role knowledge.
animals.
Special Topics
young children often underestimate the difficulty of The remainder of this entry considers three spe-
memory problems, and consequently may underesti- cial topics: origins of memory, childrens eyewitness
mate the need for a strategy to remember effectively. testimony, and working memory.
Also, young children are less likely than older chil-
dren to realize that a strategy is not working and Origins and Early Development of Memory
should be abandoned for a better strategy (Kail, Most methods used to study memory in children
1990). and adolescents cannot be used with infants because
they require speaking or writing. Consequently, the
Knowledge and Memory study of infants memory became possible only after
As children develop, they acquire more and more psychologists devised techniques that took advantage
knowledge of their worlds. Such understanding is an of infants abilities to respond in other ways. One ap-
important aid to memory. To explain the impact of proach is to show a stimulus to an infant; immediately
knowledge on memory, psychologists represent a thereafter, the infant is shown the same stimulus with
persons knowledge as a network like the one shown a novel stimulus. Infants typically look longer at the
in Figure 1. This network consists of nodes (the ellip- novel stimulus, behavior possible only if they remem-
ses) linked by different types of associations: isa links ber the one stimulus from its original presentation.
denote category membership; can and has denote Newborns tested on novelty-preference tasks look
properties of the node. longer at novel stimuli, indicating that some form of
memory is functional at birth.
Knowledge like that depicted in Figure 1 can aid
memory because it provides special codes that simpli- In another method used to assess infants memo-
fy memorization. To illustrate, suppose that people ry, a mobile is placed over an infants crib and a rib-
were asked to remember a list of words like collie, terri- bon connects the infants leg to the mobile. Infants
er, Dalmatian, poodle. In this case, the category name kick vigorously just a few minutes after their leg is
dogs serves as a code for the list. People can recall the connected to the mobile, demonstrating that they
category name, then scan memory for words associat- have learned the relation between their kicking and
ed with that name and decide if they were presented. the mobiles movement. To study infant memory,
Thus, one benefit of childrens growing knowledge is time is allowed to pass, then the infants leg is recon-
CHILDREN, DEVELOPMENT OF MEMORY IN 73
nected to the mobile. If the infant immediately begins age. Third, with development, children can recall
to kick vigorously, then the infant apparently has re- events over increasingly large intervals, from four
membered the relation between his or her kicking weeks at thirteen months of age to one year at twenty
and the mobiles movement. If, instead, kicking grad- months of age. Fourth, infants and toddlers recall
ually becomes more vigorousas if the infant is re- more accurately when they experience events more
learning the relationthen the infant has forgotten than once and when they actively participate in the
the relation. events (e.g., imitate the actions once before the
delay).
Research based on techniques like these has dem-
onstrated that by the age of three months infants will Childrens Eyewitness Testimony
remember the link between kicking and the moving Memory for past events is particularly important
mobile for up to fourteen days. That is, when an in- when children are asked to provide eyewitness testi-
fants leg is reconnected to the mobile within fourteen mony, as they sometimes are in cases of child abuse
days of the original learning, the infant will kick vig- (Bruck and Ceci, 1999). Although children, even
orously, without the need to relearn the relation. three- to five-year-olds, can remember past events ac-
When a sufficiently long interval had elapsed such curately, many commonly used legal procedures can
that infants no longer kick (i.e., they apparently have lead children to confuse what actually happened with
forgotten that kicking moves the mobile), a cue helps what others suggest may have happened. This is espe-
infant to remember. If the experimenter moves the cially true when interviewers are biasedthat is, when
mobile for the infant, who is later connected to the interviewers believe they know what happened and
mobile with the ribbon, the infant will again kick, ap- use questioning to attempt to confirm these beliefs.
parently signifying that the infant remembers the link The typical result is a highly suggestive interviewing
between kicking and the moving mobiles. Thus, ex- technique that is unlikely to help children recall
periments show that three important features of events accurately. For example, interviewers often ask
memory exist as early as two and three months of age: misleading questions (e.g., ones that imply events
an event from the past is remembered; over time, the happened when they may not have happened) or ask
event can no longer be recalled; and a cue can serve the same question repeatedly (implying that the
to dredge up a memory that seems to have been for- childs previous answers were wrong); both proce-
gotten (Rovee-Collier, 1999). dures increase the chances that a child will describe
Both of these memory tasks demonstrate that in- events that never happened.
fants are able to recognize events experienced previ- One particularly controversial practice is the use
ously. Other methods are used to assess recall, which of anatomically correct dolls to aid childrens recall.
refers to retrieving from memory a representation of Many professionals are strong advocates of these
a past experience that is not currently perceptible dolls, claiming that they improve memory by over-
(e.g., remembering what one had for dinner yester- coming childrens embarrassment over talking about
day). With older children and adults, recall is usually private events or by overcoming their ignorance of
assessed by having people describe verbally what hap- the names of body parts. In fact, research indicates
pened in the past. This method is not appropriate for that the use of anatomically correct dolls increases the
infants and toddlers with limited language; instead, number of false reports, apparently because the ex-
researchers have relied upon imitation tasks. An examiners description of the dolls implicitly encour-
perimenter demonstrates novel actions with simple ages children to create stories about body parts.
toys; the toys are then presented to childreneither
immediately or after a delayand children are en- To improve the reliability of childrens testimony,
couraged to play with the toys. Accurate imitation of they should not be questioned repeatedly on a single
the actions is taken as evidence of recall memory, par- issue and they should be warned that interviewers
ticularly when children reproduce multiple actions in may sometimes try to trick them (e.g., suggest things
the correct order. that didnt happen). In addition, interviewers should
be neutral in their questioning and evaluate alterna-
Research with this method has revealed several tive explanations of a particular event and who was in-
important features of early recall memory. First, re- volved.
call emerges at about nine months: At this age infants
can imitate novel events after seeing them modeled Working Memory
once. Second, the amount of information that chil- Working memory denotes a cognitive structure
dren can recall immediatelydefined as the number that includes ongoing information processing as well
of actions from the sequence that are reproduced ac- as the data required for those processes. According to
curatelyincreases steadily from two actions at elev- one widely accepted view (Baddeley, 1992), working
en months of age to eight actions at thirty months of memory includes a central executive as well as two
74 CLASSICAL CONDITIONING: Behavioral Phenomena
separate subsystems for storing verbal and spatial in- EXPERIENCE AND LEARNING; FALSE MEMORIES;
formation. To measure working memory, investiga- INFANCY, MEMORY IN; WORKING MEMORY:
tors often ask people to remember information while HUMANS
concurrently performing other processing tasks. Par-
Bibliography
ticipants may be asked to read sets of sentences and
Baddeley, A. (1992). Working memory. Science 255, 556559.
concurrently remember the last word from each sen- Bruck, M. B., and Ceci, S. J. (1999). The suggestibility of childrens
tence in the set. Testing begins with one or two sen- memory. Annual Review of Psychology 50, 419439.
tences (hence, participants must remember one or Kail, R. (1990). The development of memory in children, 3rd edition.
two sentence-ending words) and continues until par- New York: W. H. Freeman.
(2002). Information processing and memory. In M. Born-
ticipants are no longer able to recall the sentence-
stein, L. Davidson, C. L. M. Keyes, K. A. Moore, and The Cen-
ending words in the correct sequence. Assessed in this ter for Child Well-Being, eds., Well-being: Positive development
manner, working memory increases substantially dur- across the life course. Mahwah, NJ: Erlbaum.
ing childhood and adolescence. Rovee-Collier, C. (1999). The development of infant memory. Cur-
rent Directions in Psychological Science 8, 8085.
Researchers have identified the causes and conse-
quences of age-related change in working memory. Robert V. Kail
Regarding the causes, age-related changes in working Revised by Robert V. Kail and Meghan Saweikis
memory are due primarily to age-related increases in
the speed with which children can execute basic cog-
nitive processes. As children develop, their process-
ing speed increases, which allows them to execute the
various updating functions of working memory more CLASSICAL CONDITIONING:
rapidly. For example, more rapid processing speed BEHAVIORAL PHENOMENA
results in more rapid rehearsal and more rapid re- Classical conditioning involves learning the relations
trieval of items in working memory. between stimuli. In its simplest form, a neutral stimu-
Regarding the consequences, age-related change lus precedes a stimulus (the unconditioned stimulus,
in working memory contributes to improved reason- or US) that elicits a response (the unconditioned re-
ing and problem solving during infancy, childhood, sponse, or UR). Learning is indexed by the develop-
and adolescence. That is, in the course of reasoning ment of a response (the conditioned response, or CR)
and solving problems, individuals must remember to the neutral stimulus (which is now a conditioned
task elements and coordinate task-relevant opera- stimulus, or CS). The interval between the onset of
tions, which are functions attributable to working the CS and the onset of the US is called the intersti-
memory. Consequently, as the capacity of working mulus interval (ISI). Stimuli that can become CSs may
memory increases with age, children have more re- be discrete or more contextual, and they need not
sources available to storing and processing operations even be external (Bouton, Mineka, and Barlow,
during reasoning and problem solving, resulting in 2001). Responses to stimuli (both CRs and URs) may
improved performance (Kail, 2002). Thus, as chil- be as simple as an eye blink or more complex, such
dren develop, they process information more rapidly, as approach and withdrawal. Originally thought to be
which allows them to use working memory more ef- due simply to contiguity between the CS and US,
fectively. This, in turn, produces improved reasoning modern conceptions of learning in classical condi-
and problem solving. tioning emphasize that the CS must provide informa-
tion about the US, and that the CR is both elicited by
the CS and anticipates the US.
Conclusion
Memory improves during childhood and adoles- Excitatory Classical Conditioning
cence, due to age-related increases in use of memory
In excitatory procedures, the CS signals that a US
strategies and age-related increases in knowledge.
will follow. Learning is indexed by the probability,
Recognition and recall are both evident in the first
magnitude, and latency of CRs to the CS. In the delay
year of life. Young children can recall past events ac-
procedure, CS onset precedes US onset and the CS
curately during legal proceedings, but biased inter-
remains on when the US is delivered. In the trace pro-
viewers may alter childrens testimony. Working
cedure, CS onset precedes US onset but the CS termi-
memory increases with age and contributes to age-
nates before the US is delivered, leaving an empty
related improvement in reasoning and problem solv-
trace interval between CS and US. Evidence sug-
ing.
gests that CS offset, as well as the CS itself, may elicit
See also: AMNESIA, INFANTILE; CODING PROCESSES: CRs in the trace procedure. Kehoe and Weidemann
ORGANIZATION OF MEMORY; EARLY (1999) tested rabbits in eyeblink conditioning using
CLASSICAL CONDITIONING: Behavioral Phenomena 75
a tone CS and a corneal air puff US in a trace proce- US (CS1-US trials) and becomes a conditioned exci-
dure. Typically, longer CS-US intervals slow condi- tor. On other trials, the two CSs are presented simul-
tioning relative to shorter CS-US intervals. However, taneously (CS1/CS2-alone trials) and no US is pres-
across CS-US intervals of 400 milliseconds to thirty ented. The CS that is never paired with the US (CS2)
seconds, the learning rate was similar, as long as the becomes a conditioned inhibitor. Another common
interval between CS offset and the US was held cons- procedure for producing inhibitory CSs is the dis-
tant at 400 milliseconds. crimination procedure, in which one CS (CS+) is fol-
lowed by the US and the other CS (CS) is not. In this
Another procedure for producing excitatory con- case, CS+ becomes excitatory and CS becomes in-
ditioning is the discrimination procedure. In simple hibitory.
discrimination conditioning, two CSs that differ along
some dimension are used. One CS is always followed Because conditioned inhibition involves the sup-
by the US (CS+) whereas the other CS is never fol- pression of CRs, which renders them unobservable,
lowed by the US (CS). Typically, CRs develop to both special tests have been developed to detect condi-
CSs initially, and then decline to the CS. This proce- tioned inhibition (Rescorla, 1969). In the retardation
dure is thought to produce both an excitatory CS test, the potential inhibitory CS, when paired with a
(CS+) and an inhibitory CS (CS, discussed further US, must take longer to elicit CRs than a new CS in
below). A more complicated type of discrimination is order to be considered a conditioned inhibitor. In the
occasion setting. In occasion setting, one CS is used. An- summation test, the potential inhibitory CS is presented
other stimulus (the occasion setter) signals whether simultaneously with an excitatory CS. If fewer CRs are
the CS will be paired or unpaired on a particular trial. observed to this compound than to the excitatory CS
This occasion setter is not considered a CS because it alone, the test CS is considered a conditioned inhibi-
does not appear to form a direct association with the tor.
US. An example of this procedure, in which a contex-
tual stimulus served as the occasion setter, is provided Unlike the widespread acceptance of excitatory
by Rogers and Steinmetz (1998). In that study, a flash- associations between a CS and US (but see Gallistel
ing chamber light was the occasion setter. Whenever and Gibbon, 2000, for an alternative), inhibitory asso-
the flashing chamber light was on, the context was lit ciations have been controversial (Savastano, Cole,
and a tone CS was followed by a corneal air puff US. Barnet, and Miller, 1999). The most common view
Whenever the flashing chamber light was off, the con- has been that inhibitory associations are simply the
text was darkened, and the same tone CS was not fol- opposite of excitatory associations (Rescorla and
lowed by the US. Rabbits learned CRs when the Wagner, 1972). In this view, a single continuum for
chamber was lit and suppressed CRs when the cham- association exists, ranging from positive (excitatory)
ber was darkened. The light or dark chamber by itself to negative (inhibitory). Other researchers have pro-
cannot elicit CRs, so it is not a CS. Rather, the state posed two separate continuums, one for excitatory as-
of the context served to modulate the development of sociations and the other for inhibitory associations
CRs to the tone CS. (Pearce and Hall, 1980). These two continuums op-
pose each other and the presence or absence of CRs
is a reflection of this opposition. Finally, Miller and
Inhibitory Classical Conditioning Matzel (1988) have developed a model of classical
conditioning in which inhibitory associations do not
In inhibitory procedures, the CS signals that a US exist at all, and the presence or suppression of CRs
will not follow. It is important to note that an inhibito- is determined by the inequality of excitatory associa-
ry CS, like an excitatory CS, provides information tions for different stimuli.
about the US. Thus, a US must be present somewhere
in the conditioning situation for inhibitory condition-
ing to occur. This is in contrast to latent inhibition Stimulus Preexposure
(discussed under Stimulus Preexposure, below).
The fact that a US must be present somewhere in the Exposure to the stimuli by themselves prior to
conditioning situation means that some other stimu- classical conditioning can change the rate at which
lus in the conditioning situation, either discrete or these stimuli become CSs. The most well-studied ex-
contextual, becomes a conditioned excitor. ample of stimulus preexposure involves preexposure
to the CS, or latent inhibition (Lubow and Moore,
A number of procedures are thought to be able 1959). In contrast to a new CS, a preexposed CS that
to produce inhibitory CSs. The most common proce- is subsequently paired with a US will come to elicit
dure was originally used by Ivan Pavlov (1927). In CRs only slowly. Because no US is present during pre-
Pavlovs conditioned inhibition procedure, two CSs are exposure, the CS does not provide information on ei-
used. On some trials, one of the CSs is paired with a ther the presence or absence of the US during pre-
76 CLASSICAL CONDITIONING: Behavioral Phenomena
exposure and, therefore, forms neither an excitatory text-US association, showed less blocking than rabbits
nor an inhibitory association. Rather, the preexposed that did not receive context-alone exposure. In addi-
CS is thought to undergo a loss of salience, and is tion, rabbits shifted to a different context between
treated as irrelevant. phases 1 and 2 also showed less blocking than rabbits
Schmajuk, Lam, and Gray (1996) developed a de- training in the same context in phases 1 and 2.
tailed neural network model in which latent inhibi-
tion is explained by a reduction in the ability to store
CR Timing
and retrieve CS-US associations. Preexposure to a CS
reduces both attention to and the magnitude of the In classical conditioning, a US elicits a UR and a
representation of that CS. During subsequent CS-US CS comes to elicit a CR. However, the CR is not just
pairings, less attention to the CS slows the storage of elicited by the CS. In addition, the CR is made in an-
an excitatory CS-US association and a less intense ticipation of the US. In many classical conditioning
representation slows the ability of the CS to elicit re- procedures, such as eyeblink conditioning, this antici-
trieval of the CS-US association. pation takes the form of precise timing of the CR to
occur just before the onset of the US.
Cue Competition There are several ways to demonstrate timing of

CRs. The simplest method is ISI shift. In an ISI-shift
Multiple stimuli may be paired with a US, but not
procedure, training initially occurs at one CS-US in-
all of them will develop the ability to elicit a CR. Con-
terval. When the CS-US interval is lengthened or
sider two CSs presented simultaneously and followed
shortened, the latency of CRs to the CS lengthens or
by a US. If the CSs are similar in salience, each of the
shortens to coincide with the new CS-US interval. An-
CSs will acquire the ability to elicit a CR when pres-
other procedure involves a dual ISI. In a dual-ISI pro-
ented separately. However, if the CSs differ greatly in
cedure, the CS-US interval varies between two values
salience (for example, a bright light and a soft tone),
across trials. Typically, this procedure produces a
only one of the two CSs will come to elicit CRs. The
double-peaked CR, with the peaks corresponding to
more salient CS is said to overshadow, or disrupt, the
the two CS-US intervals. Finally, timing of CRs can be
formation of an association between the less salient
demonstrated with ISI discrimination. In an ISI-
CS and the US. This phenomenon demonstrates that
discrimination procedure, two CSs that differ along
mere contiguity of a CS and a US is not always enough
some dimension are used, as in a standard discrimi-
to produce an association between them.
nation procedure. However, in contrast to a standard
The insufficiency of contiguity between a CS and discrimination procedure, each CS is paired with a
a US for learning in classical conditioning was force- US, but at a different CS-US interval. Mauk and Ruiz
fully demonstrated by Kamin (1969). Kamins proce- (1992) used tones of different frequencies for the two
dure, which is known as blocking, involves three CSs and a corneal air puff as the US. Rabbits learned
phases. In phase 1, a CS is paired with a US (CS1-US a CR to each CS with a latency appropriate to the
trials) until it becomes a strong conditioned excitor. CS-US interval.
In phase 2, the excitatory CS is presented in com-
pound with a new CS, and the entire compound is
paired with a US (CS1/CS2-US trials). In phase 3, the Conclusion
two CSs are tested separately (CS1-alone and CS2- Although classical conditioning involves learning
alone trials) for the ability to elicit a CR. The typical the relations between two stimuli, these relations can
result is that CRs occur only to the CS used in phase be complex. The rate and content of learning in clas-
1 (CS1). This demonstrates that a CS must provide sical conditioning depends upon a number of factors.
new information about the US in order to form an as- These include the following: 1. whether the CS is fol-
sociation with it. Mere contiguity between a CS and lowed by a US or not; 2. how many CSs are present;
a US is not sufficient for classical conditioning. 3. whether the organism has any previous experience
Contextual cues may contribute to the blocking with the CS, the US, or the relationship between the
effect. In essence, it has been proposed that the CS two; and 4. how much time elapses between the deliv-
and the context in phase 1 form a combined associa- ery of the CS and the delivery of US. In addition, the
tion with the US and this combined association blocks stimuli used as CSs can be as complex as an entire
learning to the other CS in phase 2. Giftakis and Tait context. Although the responses examined in labora-
(1998) tested rabbits using a tone CS, a flashing light tory studies of classical conditioning are often simple
CS, and a periorbital shock US. Rabbits given several (such as an eye blink), this is simply a control and
sessions of exposure to the context alone between measurement issue. Classical conditioning can in-
phases 1 and 2, which serves to extinguish any con- volve complex learned responses as well, and is be-
CODING PROCESSES: Imagery 77
lieved to play an important role in everyday human CODING PROCESSES

behavior.
[Coding processes refers to the ways in which information
may be represented in memory. Events in the world strike our
See also: CONDITIONING, CELLULAR AND NETWORK senses and may be well perceived, but the mental operations
SCHEMES FOR HIGHER-ORDER FEATURES OF;
that ensue determine whether the events will be remembered.
CONDITIONING, CLASSICAL AND
INSTRUMENTAL; DISCRIMINATION AND
These mental operations are referred to as coding processes
GENERALIZATION; KAMINS BLOCKING EFFECT: and have been studied in several different ways. Three types
NEURONAL SUBSTRATES; NEURAL SUBSTRATES of coding processes are discussed in the entries in this section.
OF CLASSICAL CONDITIONING: DISCRETE The first entry is on coding processes that involve mental
BEHAVIORAL RESPONSES IMAGERY. People tend to remember information better if they
convert the information to mental pictures while they study
it. This technique is frequently used by experts who can mem-
Bibliography orize huge amounts of information (see MNEMONISTS) and
Bouton, M. E., Mineka, S., and Barlow, D. H. (2001). A modern
by all people who employ effective memory strategies (see
learning theory perspective on the etiology of panic disorder.
Psychological Review 108, 432. MNEMONIC DEVICES). In the LEVELS OF PROCESSING
Gallistel, C. R., and Gibbon, J. (2000). Time, rate, and condition- approach, people are directed to think about different aspects
ing. Psychological Review 107, 289344. of events (attention is directed to superficial properties of
Giftakis, J. E., and Tait, R. W. (1998). Blocking of the rabbits classi- events or their meanings) and memory is tested later. The
cally conditioned nictitating membrane response: Effects of
deeper or more meaningful the level of processing, the better
modifications of contextual associative strength. Learning and
Motivation 29, 2348. the later memory under most circumstances. The final topic
Kamin, L. J. (1969). Predictability, surprise, attention, and condi- in this section is more general, about ORGANIZATION OF
tioning. In B. A. Campbell and R. M. Church, eds., Punishment MEMORY. One effective strategy to code information is to or-
and aversive behavior, pp. 279296. New York: Appleton- ganize it in terms of knowledge we already have. The study
Century-Crofts.
of memory organization is concerned with both how knowl-
Kehoe, E. J., and Weidemann, G. (1999). Within-stimulus competi-
tion in trace conditioning of the rabbits nictitating mem- edge is organized and how people use their knowledge to en-
brane response. Psychobiology 27, 7284. code new information in memory. The study of coding pro-
Lubow, R. E., and Moore, A. U. (1959). Latent inhibition: The ef- cesses is central to the study of human memory and ramifies
fect of nonreinforced preexposure to the conditioned stimu- through most other topics. For example, whether some bit of
lus. Journal of Comparative and Physiological Psychology 52, 415
information can be retrieved from memory depends on how
419.
Mauk, M. D., and Ruiz, B. P. (1992). Learning-dependent timing it was encoded when it was learned (see RETRIEVAL PRO-
of Pavlovian eyelid responses: Differential conditioning using CESSES IN MEMORY).]
multiple interstimulus intervals. Behavioral Neuroscience 106,
666681.
Miller, R. R., and Matzel, L. D. (1988). The comparator hypothesis:
A response rule for the expression of associations. In G. H. IMAGERY
Bower, ed., The psychology of learning and motivation, pp. 5192.
San Diego, CA: Academic Press. According to Cicero, it was the Greek poet Simonides
Pavlov, I. P. (1927). Conditioned reflexes. New York: Dover. who first recognized the utility of mental imagery for
Pearce, J. M., and Hall, G. (1980). A model for Pavlovian learning: memory. During a brief absence from a banquet at
Variations in the effectiveness of conditioned but not of un-
conditioned stimuli. Psychological Review 87, 532552.
which the poet recited the entertainment, the roof of
Rescorla, R. A. (1969). Pavlovian conditioned inhibition. Psychologi- the building caved in, mangling the guests so badly
cal Bulletin 72, 7794. that recognition of the bodies was impossible. Simoni-
Rescorla, R. A., and Wagner, A. R. (1972). A theory of Pavlovian des (who likely already had developed a fairly good
conditioning: Variations in the effectiveness of reinforcement memory in the context of his job) realized that he
and nonreinforcement. In A. H. Black and W. F. Prokasy,
eds., Classical conditioning II: Current research and theory, pp.
could remember the faces and clothing of the guests
6499. New York: Appleton-Century-Crofts. in various locations around the room; thus he was
Rogers, R. F., and Steinmetz, J. E. (1998). Contextually based con- able to identify the corpses for their families. Second-
ditional discrimination of the rabbit eyeblink response. Neuro- arily, perhaps, the art of memory was born (see Yates,
biology of Learning and Memory 69, 307319. 1966, for full histories).
Savastano, H. I., Cole, R. P., Barnet, R. C., and Miller, R. R. (1999).
Reconsidering conditioned inhibition. Learning and Motiva- Image and imagery generally are used to refer to
tion 30, 101127. those concrete, perceptual, and usually visual modes
Schmajuk, N. A., Lam, Y-W, and Gray, J. A. (1996). Latent inhibi- of thought that appear to represent the physical
tion: A neural network approach. Journal of Experimental Psy-
chology: Animal Behavior Processes 22, 321349.
world relatively directly. These are clearly distin-
guishable from verbal thought processes, which are
John T. Green arbitrary in the linguistic sense of there being no nec-
Joseph E. Steinmetz essary relationship (other than social agreement) be-
78 CODING PROCESSES: Imagery
Figure 1 required for subjects to scan from one location to an-

other on their images is a direct function of the linear
distance between the two locations on the original
map, thus reflecting the analogue character of mental
images. Further studies by Kosslyn, Finke, and others
have shown that people can mentally construct two-
and three-dimensional figures that have emergent
properties not predictable from their component
parts. Such findings indicate a perceptionlike quality
of mental imagery that has been supported by some
neurophysiological evidence (Farah, 1984).
Although the psychological processes underlying
these phenomena may not be entirely clear, the utility
of mental imagery and pictures to facilitate memory
has been recognized for centuries. Simonidess method
of loci, for example, in which successive items are im-
aged at specific locations along a familiar route, has
long been used for learning ordered lists of unrelated
items or remembering a series of points to be made
during a speech. Imagery also can be helpful in learn-
ing foreign vocabulary, as in forming an image of a
cadaver falling from a table for the Italian cadere (to
Schematic diagram of the dual coding model and one possible fall) or someone harvesting cogs in a field for cogliere
alternative. (to pick or gather). In addition, imagery seems cen-
tral to the development of thinking and memory in
children, and a progression from motor processes to
tween words and their referents. Mental images of visual images to verbal processes is assumed by most
things experienced thus often appear to be funda- major theories of cognitive development.
mentally related to their meanings. It is difficult to
In scientific studies of memory, there are several
think of a pizza without some olfactory or gustatory
typical findings that indicate an important role of im-
image or to think of ones mother without experienc-
agery: Words rated as referring to more concrete,
ing a visual image of her face.
highly imageable things are better remembered than
Consistent with subjective impressions, a variety words referring to abstract, nonimageable things
of studies have indicated that visual imagery and visu- the concreteness effect; the use of imagery in learning
al perception have some similar qualities. Several (either by instruction or spontaneously) leads to bet-
studies by Shepard and his colleagues, for example, ter memory than nonimaginal strategiesthe imagery
have examined mental rotation (Cooper and Shepard, effect; and pictures are better remembered than
1973). This paradigm typically involves asking people wordsthe picture superiority effect.
to judge whether two visually presented stimuli (e.g.,
This apparent centrality of mental imagery in
letters or three-dimensional shapes) are identical or
thinking and memory has led to several theoretical
mirror reflections of each other. Shepards results
frameworks for studying human memory and cogni-
have consistently indicated that, in order to decide on
tion that give images an essential role. The most com-
the possible identity of the two stimuli, subjects first
pletely articulated of these, the dual coding model of
mentally rotate one of them to the same orientation
Paivio (1971, 1986), is depicted in Figure 1(a). The
as the other. This is evidenced by the fact that reac-
model includes a verbal system specialized for dealing
tion times to make the identity judgments increase
with sequential, especially linguistic information and
regularly as the angular difference between the two
an imagery system specialized for dealing with non-
stimuli increases from 0 to 180 degrees and then de-
verbal, holistic information. It explicitly entails that
creases to 360 degrees (as the rotation can be made
concrete experiences generate perceptual memory
backward).
traces (images) that preserve the structural attributes
Kosslyn and his colleagues have demonstrated of the input. Those generated memory representa-
similar findings in a series of studies on mental scan- tions are assumed to be functionally equivalent to per-
ning (Kosslyn, 1973). In this paradigm, subjects mem- ceptual representations in the sense that an image
orize several landmarks on a simple map and then are of an object generated to its name has the same mne-
asked to form an image of it. Consistently, the time monic properties as the image evoked by the object
CODING PROCESSES: Imagery 79
itself (Paivio, 1986, p. 144). Among other conse- One difficulty in explaining the effects of imagea-
quences, the storage of information-rich perceptual bility on memory is that several other variables co-
images is assumed to account for the findings that occur with it. Dimensions such as concreteness and
mental comparisons of things on physical dimensions word frequency have been shown to be positively re-
(such as size) yield similar patterns of response times lated to rated imagery values, whereas others, such as
regardless of whether the objects, their pictures, or generality of reference and associative set size, have
their printed names are presented. been shown to be negatively related to it. The effect
of imagery can be empirically or statistically separat-
The hypothesized existence of separate but inter- ed from all of those other variables, however, and re-
connected verbal and imaginal systems in the dual mains a significant predictor of memory when the
coding model explains the concreteness effect in others are controlled (Paivio, 1986). Exceptions begin
terms of coding redundancy: Concrete words usually re- to occur only when memory for more complex mate-
sult in dual verbal and imaginal memory codes (via rials like concrete and abstract texts is considered or
the intersystem, referential crossover), whereas ab- when other meaningful contexts make the impor-
stract words usually result in only a single, verbal tance of imaginal processing less essential (Marschark
code. Such redundancy provides two alternative and Cornoldi, 1990).
routes to the traces for concrete words during recall The precise role of imagery in language compre-
and provides a backup if one code is forgotten. In hension and problem solving remains unclear, al-
learning of concrete word pairs (i.e., in paired-associate though it clearly plays a role in visually and spatially
learning), imagery is assumed to provide a means of oriented situations, such as route learning and chess
integrating the meanings of the two words into a sin- playing, which seem to involve analogue mental ma-
gle, imaginal unit that can be retrieved later in the nipulation. Distinguishing imagery from the nature
context of one of the words as a cue. Integration of of long-term memory codes highlights the role of in-
this sort is assumed to be unavailable for abstract dividual differences in imagery ability. A variety of
word pairs, which must be stored as two-unit verbal standardized and well-documented tests that involve
strings. Finally, pictures are assumed by the dual code tasks such as mental paper-folding, mental rotation,
model to be better remembered than words in part and mental scanning have provided evidence of large
because images may be inherently more memorable differences between individuals in the vividness,
than verbal traces but primarily because pictures are speed, and frequency of image generation. The imag-
likely to elicit verbal naming (and hence dual memory ery abilities tapped by these tests generally are not
codes) with a somewhat higher probability than words predictive of memory ability even for concrete materi-
elicit imagery. als. Nonetheless, it is clear that variability in some im-
agery abilities can have specific and marked effects in
One modified version of the dual coding model a variety of cognitive domains, including memory. In-
is depicted in Figure 1(b). In this alternative, dual ver- dividuals who have suffered head injuries, for exam-
bal and imaginal processing systems are assumed to ple, typically exhibit deficits in visual-spatial tasks
operate at a level akin to Baddeleys (1986) working such as finding hidden figures or reconstructing pre-
memory, which has both visual and verbal components. viously presented pictures. They may also fail to ex-
The dual processing systems can account for most of hibit concreteness effects in memory and be particu-
the results concerning verbal processes and image larly slow at making imaginal comparisons
manipulation and inspection. In contrast with the (Richardson, 1990). People who score high on tests of
multiple, modality-specific memory codes of the orig- image manipulation skill tend to be faster in compar-
inal dual coding model, however, this version as- ing their images on particular dimensions, although
sumes that long-term memory involves some more they may be no faster in generating those images in
generic, semantic, or propositional code common to the first place.
both concrete and abstract information (Potter, Among the more puzzling imagery findings yet to
1979). Information is retrieved from this conceptual be explained is why people who are totally, congeni-
memory and images are constructed in visual working tally blind still show apparent imagery and concrete-
memory. Imagery is still invoked to explain the bene- ness effects (DeBeni and Cornoldi, 1988). These ef-
ficial effects of imagery instructions, material con- fects appear not to be attributable to tactile
creteness, and picture presentation, but the locus of knowledge, because they are just as readily obtained
their effects on memory is placed on the distinctiveness when the to-be-remembered items are things for
of imaginally processed items within the encoding which tactile experience is unlikely (e.g., a moon, a
and retrieval contexts (Marschark et al., 1987). Inter- tiger, or a tower). Clearly, imagery is a multidimen-
item integration is assumed to be possible for abstract sional construct that has great theoretical and practi-
as well as concrete materials via conceptual relations. cal utility but no simple psychological explanation.
80 CODING PROCESSES: Levels of Processing
Bibliography such as whether the word is a synonym for a specified

Baddeley, A. (1986). Working memory. Oxford: Clarendon Press. word. The assumption is that these three types of
Cooper, L. A., and Shepard, R. N. (1973). The time required to judgments require increasingly deep levels of pro-
prepare for a rotated stimulus. Memory & Cognition 1, 246 cessing. In a subsequent memory test in which partici-
250.
DeBeni, R., and Cornoldi, C. (1988). Imagery limitations in totally pants are asked to recall the words, deeper levels of
congenitally blind subjects. Journal of Experimental Psychology: processing are associated with higher levels of recall:
Learning, Memory, and Cognition 14, 650655. those words that required a semantic-level judgment
Farah, M. J. (1984). The neurological basis of mental imagery: A are recalled best, whereas those requiring ortho-
componential analysis. Cognition 18, 245272.
graphic processing yield the lowest recall level.
Kosslyn, S. M. (1973). Scanning visual images: Some structural im-
plications. Perception and Psychophysics 14, 9094. Craik and Lockhart (1972) proposed the general
Marschark, M., and Cornoldi, C. (1990). Imagery and verbal mem- concept of levels of processing not as a theory of
ory. In C. Cornoldi and M. A. McDaniel, eds., Imagery and cog-
memory but rather as a framework for future research
nition, pp. 133182. New York: Springer-Verlag.
Marschark, M., Richman, C. L., Yuille, J. C., and Hunt, R. R. into the relationship between coding processes and
(1987). The role of imagery in memory: On shared and dis- memory. Lockhart and Craik have written a retro-
tinctive information. Psychological Bulletin 102, 2841. spective commentary on the significance of this pro-
Paivio, A. (1971; reprint 1979). Imagery and verbal processes. Hills- posal (1990). The basic claim of levels of processing
dale, NJ: Erlbaum.
as a research framework is that a thorough under-
(1986). Mental representations: A dual coding approach. Ox-
ford: Oxford University Press. standing of remembering requires a careful analysis
Potter, M. C. (1979). Mundane symbolism: The relations among of the way in which (and the degree to which) coding
objects, names, and ideas. In N. R. Smith and M. B. Franklin, processes involve the construction of meaning. Ac-
eds., Symbolic functioning in childhood, pp. 4165. Hillsdale, NJ: cording to this view, there is no distinct coding pro-
Erlbaum. cess that can be identified as committing to memo-
Richardson, J. T. E. (1990). Imagery and the brain. In C. Cornoldi
and M. A. McDaniel, eds., Imagery and cognition, pp. 145. ry. Rather, memory codingthe memory traceis
New York: Springer-Verlag. constructed as a byproduct of the everyday mental op-
Yates, F. A. (1966). The art of memory. Chicago: University of Chica- erations we perform as we interact with our environ-
go Press. ment and attempt to understand it. One implication
Marc Marschark of this claim is that a proper account of the relation-
ship between coding processes and memory will in-
volve an understanding of the broader issues of per-
ception and comprehension.
LEVELS OF PROCESSING
Orienting Tasks
Processing and Recall
Experimental conditions (orienting tasks) such as
The term levels of processing was introduced by Craik those involving judgments of case, rhyme, or syn-
and Lockhart (1972) to describe the way in which the onymity are but three examples of the large number
information contained in a stimulus can be analyzed of orienting tasks that have been used in experiments.
at levels ranging from surface physical properties to Other examples are rating a words pleasantness, de-
deeper levels involving its meaning. Thus in reading ciding how many syllables it has, and whether it is
the printed word clever, the reader might process or- spoken by a male or a female voice. Indeed, we can
thographic features, such as its being in capital let- think of our everyday cognitive activity in terms of a
ters, or phonemic features, such as that it rhymes with continuous sequence of orienting tasks as our knowl-
ever, or semantic features, such as that it is a synonym edge, goals, and needs interact with the circum-
for skilled. stances of the moment. Our goals will influence our
The level of processing is a powerful determinant orientation toward a stimulusthe specific informa-
of how well an event will be remembered (Craik and tion that we selectively attend to and analyze. The
Tulving, 1975). A simple demonstration experiment particular form of the analysis will strongly influence
illustrates this point. Participants in the experiment later remembering. In listening to a conversation, for
are presented with a sequence of common words and example, our goal may be to comprehend what is
asked to make one of three possible judgments about being said, but it may also be to infer the speakers in-
each word. For some words participants are asked to telligence, mood, or intentions, or the origin of a dis-
make a judgment at the orthographic level, such as tinctive accent. When we read a restaurant menu, our
whether the word is printed in capital letters; other goal may involve judgments of taste, preference, cost,
words require a phonemic-level judgment, such as or nutritional value. The fundamental principle of le-
whether the word rhymes with a certain word; a third vels of processing as a research framework is the claim
set of words requires a judgment involving meaning, that since the memory trace is the byproduct of these
CODING PROCESSES: Levels of Processing 81
analyses, the key to predicting subsequent memory Skills and the Material to Be Remembered
performance is a matter of gaining increased under- Another determinant of the level of processing is
standing of the nature and level of these analyses. the nature of the material to be remembered. Pictures
Orienting tasks used in experimentstasks such as and common words afford the rapid analysis of mean-
judging rhyme or synonymity (meaning)are the sci- ing. Other material, such as proper nouns, can pose
entists effort to gain tight control over the processing greater difficulty. One reason for the common experi-
that a participant applies to a stimulus to evaluate the ence of rapidly forgetting names following introduc-
impact that different processes have on memory. tions at a party is that proper nouns do not trigger
rapid deep processing. Hence most mnemonic tech-
niques for remembering names are essentially strate-
Incidental versus Intentional Processing
gies for converting proper names into a visual image
Levels of processing have important implications that embodies deeper semantic-level processing. In-
for the distinction between incidental and intentional teracting with the nature of the material is a second
processing. Because we do not usually make an inten- determinant: the skill of the rememberer. To the
tional effort to commit everyday experiences to mem- nonspeaker of French the letter string chien is mean-
ory, much of our normal remembering is incidental. ingless, as is a musical score to anyone who has not
You can probably recall what you were doing exactly been trained to read music, or the game position of
twenty-four hours ago even though, at the time, you chess pieces to one who does not play chess. But to the
were not making a conscious effort to commit the ac- speaker of French, to the trained musician, or to the
tivity to memory. This aspect of everyday life can be skilled chess player, the word, the score, and the
captured in an experimental setting by using inciden- board position, respectively, afford deep processing
tal orienting tasks in which participants are not in- and hence better remembering.
formed that they will receive a subsequent memory
test. Participants might perform the judgments be-
lieving that the only purpose of the experiment is to Conclusion
see how quickly they can respond. According to the Levels of processing is a theoretical framework
levels of processing approach, such participants that claims that memory coding is a byproduct of cog-
should not be disadvantaged relative to those in- nitive and interpretive processes. Memory depends
structed to expect the memory test, provided the level heavily on the degree to which the processing in-
of processing for the two groups is comparable. That volves the construction and elaboration of meaning.
is to say, the important determinant of remembering Tasks such as judging case, rhyme, or synonymity
is not the conscious effort of committing something provide a clear example of three distinct levels of pro-
to memory but the level of processing that the orient- cessing. Depth of processing, however, can vary over
ing task induces. This conclusion is supported by ex- a wide range. According to levels of processing as a
perimental findings. For example, participants who research framework, research into coding processes
are asked to rate a word for pleasantness as an inci- seeks to provide a precise account of this variation
dental orienting task perform as well on a subsequent and its impact on memory. Recent developments of
unexpected memory test as those who are given prior the original idea of levels of processing and critical
warning of the test. analyses are discussed in Naveh-Benjamin, Mos-
What happens when a researcher exhorts partici- covitch, and Roediger (2001).
pants in an experiment simply to try to remember?
Presumably the subjects will process the material in Bibliography
whatever way they think will most effectively support Craik, F. I. M., and Lockhart, R. S. (1972). Levels of processing:
later remembering. A common strategy with verbal A framework for memory research. Journal of Verbal Learning
material is to rehearse it by silently repeating a word and Verbal Behavior 11, 671684.
Craik, F. I. M., and Tulving, E. (1975). Depth of processing and the
or phrase over and over. Such a strategy is relatively
retention of words in episodic memory. Journal of Experimental
ineffective for long-term remembering, since such re- Psychology: General 104, 268294.
petitive processing typically involves no further analy- Lockhart, R. S., and Craik, F. I. M. (1990). Levels of processing:
sis of meaning (no deeper-level processing) but con- A retrospective commentary on a framework for memory re-
sists of maintaining phonemic recollection. Again, the search. Canadian Journal of Psychology 44, 87112.
important point is that successful remembering is less Naveh-Benjamin, M., Moscovitch, M., and Roediger, H. L. III, eds.
(2001). Perspectives on human memory and cognitive aging: Essays
a matter of conscious effort than of being led to perin honor of Fergus Craik. Philadelphia: Psychology Press.
form an orienting task that demands deep levels of
processing. Robert S. Lockhart
82 CODING PROCESSES: Organization of Memory
ORGANIZATION OF MEMORY at least one fundamental process of organization. As-

sociations derive from the frequent temporal cluster-
ing of events. In the early part of the twentieth centu-
Coding and Organization ry, Pavlov (1927) discovered classical conditioning.
Coding refers to the interpretations a person gives to This discovery led to extensive investigations of the
experiences. The significance of experience for mem- formation and maintenance of associations. Pavlov
ory and action depends on the interpretation of the found that after frequently presenting a neutral stim-
experience. The same events can be interpreted in ulus (e.g., a tone) in close proximity to the presenta-
dramatically different ways depending on a persons tion of food, a dog would salivate at the sound of the
knowledge and expectations. To understand coding tone even in the absence of food. Thus, an association
we must understand the organization and use of formed between the tone and the food.
knowledge in interpreting experience. The interre-
Garcia and Koelling (1966) found that some asso-
latedness of ideas is one of the most compelling facts
ciations are learned more easily than others. Their
of mental life. In personal memories, a single associa-
laboratory rats learned to associate a novel taste with
tion with some present event can trigger detailed
gastrointestinal illness much more easily than they
memories of past experiences. Psychology has devel-
learned the association between a flashing light and
oped several ideas about the nature of organization
gastrointestinal illness. This result suggests that vari-
in memory.
ous constraints influence the formation of associa-
We can illustrate the influence of coding by com- tions.
paring the memories of two people with different de-
grees of knowledge: in this case, an expert and a non-
expert about cars. They both see the same small red Associative Networks
car. The expert identifies it as a Miata; the nonexpert In the direct representation of associations in the
can identify it only as a small red car. Would it sur- form of a network, concepts are shown as nodes and
prise you if later the expert was able to state with some associations are shown by lines (or links) connecting
confidence that a small red Triumph was not the car the nodes. Schvaneveldt, Durso, and Dearholt (1989)
seen earlier, while the nonexpert had more difficulty presented a method of deriving such networks from
in making this discrimination? Each individuals proximity data such as judgments of relatedness
knowledge influences the coding and thus the memo- among sets of concepts. Cooke, Durso, and Schvane-
ry of the experience. veldt (1986) found that networks can predict the way
Human memory imposes organization on our ex- people organize the concepts when they learn a list of
periences. Tulving (1962) and others have shown that words. Goldsmith and Johnson (1990) were able to
when people learn a list of randomly selected words, predict students grades in a course on experimental
they organize the words in recalling the list. As the list methods from the degree of similarity of the students
is learned, there is more and more consistency in the and the instructors networks of important concepts.
grouping of the words in recall.
Earlier, Bousfield (1953) showed that subjects re-
Semantic Networks and Semantic Features
call lists of words as clusters of related words. For ex-
ample, if the list contained some names of flowers, Semantic networks also use network representa-
some names of people, some types of buildings, and tions, but they specify more about the relations be-
so on, then the free recall of these words would group tween concepts by using labeled links (Collins and
the similar items. This grouping occurs even though Quillian, 1969; Meyer and Schvaneveldt, 1976; Quil-
the words are presented in random order. Later lian, 1969). For example, such a network would show
Bower and his colleagues (Bower, 1970) showed that that robin is a member of the class bird with an isa
theories about the structure of memory could predict link (A robin is a bird). It would also show that a deer
the organization of material to be learned. Bransford has antlers, and so on. Such networks can also support
and Johnson (1972) studied passages that are difficult inferences such as concluding that a robin is an ani-
to remember unless people are led to give them ap- mal by retrieving a robin is a bird and a bird is an ani-
propriate interpretations. Their work is an impressive mal. Semantic networks have been used to explain ex-
demonstration of the role of interpretation in re- perimental data from studies in language
membering. understanding and category judgments. Such net-
works are also often a part of computer programs de-
signed to exhibit artificial intelligence (Quillian,
Organization of Memory 1989). Other theories propose that concepts consist
What leads to the organization of memories? of collections of features that define the concepts
Most answers to this question refer to association as (Smith and Medin, 1981). The concept bird, for exam-
CODING PROCESSES: Organization of Memory 83
ple, might consist of features such as has wings, flies, constrained by the history and situation of the indi-
lays eggs, has feathers, and so on. According to feature vidual.
theories, when people reason about concepts, they re-
trieve features from memory and use them to draw
conclusions. Conclusion
Coding is the interpretation of events in light of
what we know. Such interpretation can have benefi-
Schemata cial consequences, as in the superiority of the memory
Schemata are general representations of several of chess masters for real board positions. Sometimes
different items of information together with the speci- interpretation leads to false memories of related in-
fication of the relations among the items (Bartlett, formation that was not actually experienced (Loftus
1932; Minsky, 1975). For example, the schema for a and Ketcham, 1991). Understanding the memory of
room might specify that it must have a floor, a ceiling, an event requires an understanding of the coding that
walls, and a door as well as some spatial relations arises from cumulative knowledge. An important
among these. Optionally, it might have additional question for theory and research concerns the extent
doors and windows. Scripts are examples of schemata to which memory depends on stored representations
where actions are organized in familiar sequences as opposed to cues available from the body and the
such as going to a restaurant or visiting the doctor. environment.
Schemata invite inferences. Several studies suggest
that memory includes inferred information (defaults) See also: CODING PROCESSES: IMAGERY; CODING
in addition to what we actually experience. For exam- PROCESSES: LEVELS OF PROCESSING; FALSE
ple, if we hear the sentence, Fred drove the nail into MEMORIES
the board, we are likely to infer that he used a ham-
mer even though the sentence does not mention a Bibliography
hammer. If someone eats in a restaurant, we assume Bartlett, F. C. (1932). Remembering: A study in experimental and social
that he or she paid for the meal. psychology. Cambridge, UK: Cambridge University Press.
Bickard, M. H. (2000). Dynamic representing and representational
Chase and Simon (1973) reported a classic dem- dynamics. In E. Dietrich and A. Markman, eds., Cognitive dy-
onstration of the power of schemata using memory namics: Conceptual and representational change in humans and ma-
for the positions of pieces on a chess board. They chines. Mahwah, NJ: Erlbaum.
Bousfield, W. A. (1953). The occurrence of clustering in the recall
found that chess masters were no better than novices
of randomly arranged associates. Journal of General Psychology
at reconstructing a board with randomly placed 49, 229240.
pieces, but the masters were far superior in recalling Bower, G. H. (1970). Organizational factors in memory. Cognitive
the positions of pieces from the middle of an actual Psychology 1, 1846.
chess game. Experts presumably have elaborate sche- Bransford, J. D., and Johnson, M. K. (1972). Contextual prerequi-
sites for understanding: Some investigations of comprehen-
mata that can code the positions of the pieces on the
sion and recall. Journal of Verbal Learning and Verbal Behavior
board when the positions make sense. 11, 717726.
Chase, W. G., and Simon, H. A. (1973). Perception in chess. Cogni-
tive Psychology 4, 5581.
Embodiment and the Need for Collins, A. M., and Quillian, M. R. (1969). Retrieval time from se-
mantic memory. Journal of Verbal Learning and Verbal Behavior
Representations 8, 240247.
In recent years challenges to traditional ideas Cooke, N. M., Durso, F. T., and Schvaneveldt, R. W. (1986). Recall
and measures of memory organization. Journal of Experimental
about the role of mental representations have arisen
Psychology: Learning, Memory, and Cognition 12, 538549.
from researchers in cognitive science. A major con- Edelman, G. M. (1992). Bright air, brilliant fire: On the matter of mind.
cern is that traditional approaches have neglected the New York: Basic Books.
constraints imposed on learning and development Freeman, W. J. (1995). Societies of brains. Hillsdale, NJ: Erlbaum.
that stem from the physical body and from the envi- Garcia, J., and Koelling, R. A. (1966). Relation of cue to conse-
quence in avoidance learning. Psychonomic Science 4, 123124.
ronment. At the extreme, theorists advocating a dy-
Goldsmith, T. E., and Johnson, P. J. (1990). A structural assess-
namic systems approach claim that grounding cogni- ment of classroom learning. In R. Schvaneveldt, ed., Pathfind-
tion in the interaction of the body and the world er associative networks: Studies in knowledge organization. Nor-
obviates the need to propose mental representations wood, NJ: Ablex.
that mediate perception and action (Edelman, 1992; Johnson, M. (1987). The body in the mind: The bodily basis of meaning,
imagination, and reason. Chicago: University of Chicago Press.
Freeman, 1995; Johnson, 1987; Thelen and Smith,
Loftus, E. F., and Ketcham, K. (1991). Witness for the defense. New
1994; van Gelder, 1997). The grounding of concepts York: St. Martins.
in perception and action helps explain how concepts Meyer, D. E., and Schvaneveldt, R. W. (1976). Meaning, memory
are learned (Bickard, 2000). Consequently, coding is structure, and mental processes. Science 192, 2733.
84 COGNITIVE ENHANCERS
Minsky, M. (1975). A framework for representing knowledge. In P. impaired subjects cognitive function has slipped back
Winston, ed., The psychology of computer vision. New York: onto the rising phase of the dose-response function.
McGraw-Hill.
Pavlov, I. P. (1927). Conditioned reflexes, trans. G. V. Anrep. London:
In this case, treatments to facilitate cognition in im-
Oxford University Press. paired subjects act to restore optimal performance,
Quillian, M. R. (1969). The teachable language comprehender. but the same treatments in normal subjects drive
Communications of the ACM 12, 459476. them past the optimal level, resulting in impairment.
Schvaneveldt, R. W., Durso, F. T., and Dearholt, D. W. (1989). Net- The inverted-U hypothesis suggests that it is easier to
work structures in proximity data. In G. H. Bower, ed., The
psychology of learning and motivation: Advances in research and the-
treat memory deficits than to improve normal memo-
ory, Vol. 24. New York: Academic Press. ry. The results of many studies in animals and hu-
Smith, E. E., and Medin, D. L. (1981). Categories and concepts. Cam- mans support this idea.
bridge, MA: Harvard University Press.
Thelen, E., and Smith, L. B. (1994). A dynamic systems approach to Still, the possibility of memory enhancement in
the development of cognition and action. Cambridge, MA: MIT normal healthy people remains very attractive. A ca-
Press. sual search of the Internet leads to hundreds of sites
Tulving, E. (1962). Subjective organization in free recall of unre- offering memory-boosting aids. In some cases the
lated words. Psychological Review 69, 344354.
agents are described only as herbal mixtures, while in
van Gelder, T. (1997). Dynamics and cognition. In J. Haugland,
ed., Mind design II. Cambridge, MA: MIT Press. others a single compound, or one or more specific
components of the blend, is touted. The major con-
Roger W. Schvaneveldt
stituents of these products are discussed in detail later
in this entry. First, this entry reviews some drugs that
are known to be able to enhance cognition and mem-
ory but are not generally available or usually sold for
COGNITIVE ENHANCERS this purpose.
Virtually everyone has occasionally wished that he or
she could think faster, had a better memory, or was
CNS Stimulants
simply smarter. It is possible to improve these facets of
cognitive function through practice, but most people The common feature of these agents is that they
are looking for an easier solution. At the beginning stimulate the central nervous system. Picrotoxin,
of the twenty-first century, the answer is usually imag- bicuculline, and strychnine are drugs that block in-
ined to be in the form of a pill. hibitory systems in the brain, leading to a generalized
increase in excitability. Picrotoxin was one of the first
Cognition has many elements. Key aspects of cog-
agents recognized to be a memory enhancer. All three
nition include sensory perception, attention and con-
drugs have been shown in animal studies to improve
centration, immediate (or working) memory, and
memory at low doses, but higher doses cause seizures
long-term memory. More complex and integrated
and death. These drugs are too dangerous for experi-
facets of cognition include abstract reasoning and
mentation in humans.
planning. Thus, there are a number of ways for phar-
macological treatments to enhance cognition. Howev- Another stimulant drug, amphetamine, has reli-
er, for most people cognitive enhancement is synony- ably been shown to improve memory in both animals
mous with having a better long-term memory, so this and humans. Amphetamine causes the release of
topic is the primary focus of this entry. neurotransmitters that promote arousal, including
epinephrine and norepinephrine. The resulting in-
Memory itself is a complex phenomenon that is
crease in attention plays a major role in the memory-
still not completely understood by the medical and
enhancing properties of amphetamine. In addition,
scientific community. Much of the research into un-
amphetamine improves memory consolidation, the
derstanding memory mechanisms is motivated by the
process that leads to long-term memory storage. The
need to develop treatments for memory impairments
major problems with this drug are that tolerance to
caused by diseases such as Alzheimers disease. An im-
its beneficial effects develops quickly, along with side
portant question in the present context is whether
effects that include physical dependence and addic-
treatments that correct memory deficits can also en-
tion. These liabilities greatly outweigh amphet-
hance memory in normal healthy individuals. Virtual-
amines utility as a memory enhancer.
ly all drugs that improve memory show an invert-
ed-U dose-response relationship: Doses of a Caffeine is the worlds most widely used stimu-
compound up to a certain level enhance perfor- lant. It also increases alertness and attention, thereby
mance, but at still higher doses the enhancing effect enhancing cognitive performance. However, caf-
is lost. The argument has been made that cognitive feines effects on memory appear to be small. Vaso-
function in normal healthy subjects is at the crest of pressin is a pituitary hormone that plays an important
the inverted-U, while in aged, diseased, or otherwise role in the bodys regulation of water. In addition, va-
COGNITIVE ENHANCERS 85
sopressin has been shown to enhance memory in both non of the inverted-U dosing effect, could explain the
humans and experimental animals. The precise inconsistent results of the few formal tests of these
mechanism is unknown, but since vasopressin affects agents that have been performed.
norepinephrine utilization in the brain researchers
believe that increased arousal plays a role. Five compounds seem to be most frequently sold
as cognition/memory enhancers: huperzine A, vin-
pocetine, Ginkgo biloba, ginseng, and pregnenolone.
Cholinergic Agents Huperzine A, an extract derived from a particular
type of club moss, is an acetylcholinesterase inhibitor
The neurotransmitter acetylcholine is known to
that is far more potent than physostigmine. The ace-
play an important role in learning and memory. Neu-
tylcholine plays an important role in the brain circuit-
rons that contain acetylcholine degenerate in patients
ry that encodes memories and is used as a neuro-
with Alzheimers disease, an illness that is character-
transmitter in the peripheral nervous system. Overac-
ized by profound cognitive impairments. In addition,
tivation of these peripheral cholinergic connections
drugs that block the action of acetylcholine in the
can have profound and potentially dangerous conse-
brain impair cognition in healthy humans and in ex-
quences. Because of this problem, huperzine A is not
perimental animals. Researchers have long hypothe-
a safe drug and should not be used without a physi-
sized that improving cholinergic system function
cians supervision.
would be a good way to treat the symptoms of Al-
zheimers disease, and this is the primary mechanism Vinpocetine is derived from the periwinkle plant,
of action of drugs that are marketed for this purpose. Cricoceras longiflorus. It is widely used as a memory en-
Cholinergic function can be enhanced either by hancer, and has been found to improve memory in
preventing the breakdown of the acetylcholine to pro- healthy people as well as those with impairments
long its action or by directly activating the receptors caused by aging or disease. It is not clear how vin-
for neurotransmitter. Physostigmine is an inhibitor of pocetine works, but there is evidence that vinpocetine
acetylcholinesterase, the enzyme that degrades ace- enhances blood flow in the brain, which, in turn, pro-
tylcholine. Physostigmine and related compounds vides the basis for observations of increased glucose
have been shown to improve memory in normal hu- utilization, a sign of generally increased cerebral ac-
mans and animals as well as in Alzheimers disease pa- tivity. In addition, studies have shown that vinpoce-
tients. Activators of either the muscarinic or nicotinic tine increases the production of norephinephrine,
receptor subtypes of cholinergic receptors also im- which mediates arousal, and acetylcholine. Vinpoce-
prove memory. Nicotine is well recognized as a cogni- tine is a very safe compound at therapeutic doses.
tion enhancer, although it is debated whether its ef-
Extracts from the leaves of the Ginkgo biloba tree
fect is mediated by increasing attention or through
have been used in traditional Chinese medicine for
another mechanism. Drugs that act at muscarinic re-
thousands of years. These extracts consist of many
ceptors have also been shown to be effective cognitive
compounds, the behavioral effects of which have yet
enhancers in animal studies. Unfortunately, several
to be individually characterized. Ginkgo extracts have
muscarinic receptor activators have failed in clinical
been widely used as memory enhancers, particularly
trials with Alzheimer patients because of side effects.
in the elderly, but studies have shown that ginkgo also
No compound in this class has yet been developed
improves memory in healthy young people. The pri-
that is safe enough for human use.
mary effect of ginkgo appears to be increased cere-
bral circulation.
Components of Smart Drugs The ginseng plant, Panax ginseng, also has a long
The vast number of concoctions sold over-the- history of use in traditional Chinese medicine. Ex-
counter to improve memory provides an excellent in- tracts from the roots or the entire plant contain over
dication of both the demand and the perceived need a dozen chemical compounds, collectively termed
for these products. Many of these compounds are ex- ginsenosides. Ginsenosides have been used to treat a
tracted from plants or animal tissues and consist of number of diseases ranging from diabetes to insom-
mixtures of ingredients. A major problem with such nia, and to enhance physical and mental perfor-
products is that claims concerning amounts and puri- mance. While the memory-enhancing activity of gin-
ty of the supposedly active elements have unusually seng is not as well documented as that of ginkgo,
not been verified by an independent agency. While studies have shown beneficial effects in both young
this lack of oversight could lead to safety problems and elderly subjects. Like ginkgo, ginseng also pro-
due to contaminants in the products, a more likely motes cerebral blood flow, but it also may enhance
concern is that dosages may not be accurate or consis- specific neurotransmitter systems. There is evidence
tent. Such unreliability, coupled with the phenome- that the combination of ginseng and ginkgo synergis-
86 COLLECTIVE MEMORY
tically enhance memory. There are no known major McGaugh, J. L., and Izquierdo, I. (2000). The contribution of
health risks associated with either agent. pharmacology to research on the mechanisms of memory for-
mation. Trends in Pharmacological Science 21 (6), 208210.
Pregnenolone is a neurosteroid, a hormone that Rees, K., Allen, D., and Lader, M. (1999). The influences of age
is synthesized in the brain from cholesterol. Studies and caffeine on psychomotor and cognitive function. Psycho-
have shown pregnenolone levels decline with aging pharmacology 145 (2), 181188.
Soetens, E., Casaer, S., DHooge, R., and Hueting, J. E. (1995). Ef-
and are correlated with impaired cognitive perfor- fect of amphetamine on long-term retention of verbal materi-
mance. In addition, an injection of pregnenolone dial. Psychopharmacology 119 (2), 155162.
rectly into the brains of aged animals improves mem- Vallee, M., Mayo, W., and Le Moal, M. (2001). Role of pregneno-
ory. Pregnenolone has also been shown to enhance lone, dehydroepiandrosterone and their sulfate esters on
acetylcholine release in the brain. Taken together, learning and memory in cognitive aging. Brain Resolution Re-
view 37 (13), 301312.
these findings have been used to promote the value Warburton, D. M. (1995). Effects of caffeine on cognition and
of taking pregnenolone supplements. While pre- mood without caffeine abstinence. Psychopharmacology 119 (1),
gnenolone is safe, studies of the compound in hu- 6670.
mans have yet to provide consistent support for the Gregory M. Rose
idea that it is an effective cognitive enhancer.
Conclusion
COLLECTIVE MEMORY
Enhancing cognitive function remains both a
dream and a challenge. There are a number of agents Collective memory is a representation of the past that
that appear to at least improve memory to some de- is shared by members of a group, such as a generation
gree, but none of these compounds is so effective that or nation-state. The concept is usually traced to writ-
it can be distinguished from the others. What is clear ings of the French sociologist Maurice Halbwachs
is that some degree of cognitive enhancement is pos- (18871945), who argued that remembering is
sible even in young healthy people. Maximizing this shaped by participation in collective life and that dif-
effect will require systematic research to better under- ferent groups generate different accounts of the past
stand how pharmacological treatments affect the (Halbwachs, 1952).
basic neurobiological mechanisms that are responsi- Collective memory and related notions such as
ble for particular aspects of cognition. A more imme- public memory (Bodnar, 1992) have been examined
diate goal, and probably one that will be more quickly in academic disciplines including anthropology
achieved, is to develop treatments to improve the (Cole, 2001), history (Novick, 1999), and sociology
cognitive impairments in aged or diseased individu- (Schudson, 1992). Collective memory is also a part of
als. popular culture discussions about the Vietnam War
See also: DRUGS AND MEMORY; PHARMACOLOGICAL and the Holocaust. One of the hallmarks of collective
TREATMENT OF MEMORY DEFICITS memory is that it is tied to identity. Deeply held no-
tions about the past are often the source or pride or
Bibliography shame, and they can give rise to legal, and even
Born, J., Pietrowsky, R., and Fehm, H. L. (1998). Neuropsychologi- armed conflict. For example, the Serbs collective
cal effects of vasopressin in healthy humans. Progressive Brain memory of the Battle of Kosovo in 1389 has often
Research 119, 619643.
Furey, M. L., Pietrini, P., and Haxby, J. V. (2000). Cholinergic en-
been cited as a factor that made it possible to mobilize
hancement and increased selectivity of perceptual processing against Bosnian Muslims in the twentieth century.
during working memory. Science 290 (5,500), 2,3152,319.
Izquierdo, I., and Medina, J. H. (1991). GABAA receptor modula-
tion of memory: The role of endogenous benzodiazepines. Strong and Distributed Accounts of
Trends in Pharmacological Science 12 (7), 260265. Collective Memory
Kennedy, D. O., Scholey, A. B., and Wesnes, K. A. (2000). The
dose-dependent cognitive effects of acute administration of Collective memory is often understood in terms
Ginkgo biloba to healthy young volunteers. Psychopharma- of loose analogies with memory in the individual.
cology, 151 (4), 416423. Many discussions of Americas memory about Viet-
(2001). Dose dependent changes in cognitive performance nam, for example, seem to presuppose that America
and mood following acute administration of Ginseng to is some sort of a large organism that has intentions,
healthy young volunteers. Nutritional Neuroscience 4 (5), 295
310. desires, memories, and beliefs just as individuals do,
Kidd, P. M. (1999). A review of nutrients and botanicals in the inte- something reflected in assertions such as, Our mem-
grative management of cognitive dysfunction. Alternative Med- ory of Vietnam makes us unwilling to accept combat
icine Review 4 (3), 144161. deaths.
Mattay, V. S. et al. (2000). Effects of dextroamphetamine on cogni-
tive performance and cortical activation. Neuroimage 12 (3), Assumptions about this issue are often not well
268275. grounded and as a result have been the object of criti-
COLLECTIVE MEMORY 87
cism. For example, one of the fathers of modern sphere (Bodnar, 1992; Wertsch, 2002). Reflecting its
memory studies in psychology, Frederic Bartlett ties to identity, it is often assessed in terms of its ability
(18861969), was critical of the more or less absolute to provide a usable past, even when this representa-
likeness [that] has been drawn between social groups tion comes at the expense of accuracy. In contrast,
and the human individual (Bartlett, p. 293), and he while recognizing a connection to identity studies of
warned that collectives do not have some sort of mem- individual memory tend to focus on a criterion of ac-
ory in their own right. Bartlett did believe, however, curacy. This is not to say that psychologists believe
that memory of individuals is fundamentally influ- memory is fundamentally accurate. Indeed, many
enced by the social context in which they function. In- studies point to the myriad ways humans can generate
deed, a central point of his argument is that social incorrect accounts of what actually occurred. The
organization gives a persistent framework into which point remains, however, that psychological studies
all detailed recall must fit, and it very powerfully in- formulate memory largely in terms of its degree of ac-
fluences both the manner and the matter of recall (p. curacy.
296). In short, Bartlett accepted the notion of mem- A question that is sometimes asked about collec-
ory in the group, and not memory of the group (p. tive memory is whether it is anything more than a set
294). of individual memories and hence whether it could be
Claims about memory of the group constitute reduced to a personal level. Any attempt to answer
a strong version of collective memory (Wertsch, this question must say something about the social pro-
2002), and, when made explicit, they have usually cesses involved in collective memory without falling
been rejected. An alternative that recognizes memory into a strong version. This is often done by focusing
in the group without slipping into questionable as- on how narratives and other textual resources for
sumptions about memory of the group is a distribut- memory are produced, discussed, and understood by
ed version. From this perspective, memory is viewed members of a collective.
as being distributed socially in small group interac- For example, modern states devote major atten-
tion; as well as instrumentally (Wertsch, 2002). In tion to these issues. They produce official accounts of
the case of social distribution, for example, Mary Sue the past through textbooks and other materials used
Weldon (2001) has examined the collaborative rein schools, national commemorations, the media, and
membering that occurs when groups of individuals other forms of popular culture. In carrying out this
work together to recall information or events from project, states seek not only to promulgate their own
the past. account but also to control access to alternative ac-
In the case of instrumental distribution, memory counts. As depicted by George Orwell in his futuristic
is viewed as involving an active agent and one or more novel 1984 (1949), such control can be almost total,
cultural tools such as calendars or other written re- but in fact every modern state exerts some control
cords. An important transformation of memory in over the account of the past presented to its citizens.
human cognitive evolution occurred with the emer- This focus on the textual mediation of collective
gence of external symbolic storage (Donald, 1991). memory raises the question of whether collective
This does not mean that such memory somehow re- memory is really memory at all. What does it mean to
sides in texts or records, but it does mean that with the say, for example, that a generation of American teen-
rise of new forms of external symbolic storage such as agers remembers World War II when they were born
written texts or the Internet the possibilities for re- four decades after the event? This might involve
membering undergo fundamental change. memory for textual resources, or perhaps semantic
Such change has both psychological and social di- memory, but is it really the kind of episodic memory
mensions. By becoming skilled at using a certain set often assumed?
of cultural tools, new mental habits and schemata Such questions provide a reminder of the prob-
(Bartlett, 1995) emerge that shape remembering for lems that arise when loose analogies between individ-
members of a collective. New forms of collective life ual and collective remembering are employed. In
also derive from the appearance of novel cultural order to address them, one must again turn to the in-
tools. For example, Benedict Anderson (1991) has strumental distribution of collective memory, and in
tied the development of modern nations and other the end it may be more appropriate to discuss this in
imagined communities to the rise of print media. terms of knowledge about a set of shared textual re-
sources rather than memory per se. What is striking
about collective memory, however, is the ways in
Collective versus Individual Memory which it goes beyond what would normally be de-
Collective memory is widely understood as inher- scribed as knowledge or semantic memory. This is so
ently involving conflict and negotiation in the social first because of the social conflict and control involved
88 COMPARATIVE COGNITION
in producing the textual resources, and second be- A final property that characterizes collective re-
cause much more than neutral knowledge is involved. membering is that it tends to be heavily shaped by
When discussing highly charged historical events, it schemata (Bartlett, 1995), implicit theories
is very easy to slip into heated discussions about what (Ross, 1989), or other simplifying and organizing
really happened and deny others accounts in such frameworks. Such frameworks also shape individual
a way that narrative texts become emotionally power- memory and history, but in the case of collective
ful tools of collective identity and memory. memory they take on a particularly important role
One of the ways that studies of individual and col- due to the processes of conflict and negotiation in-
lective memory have overlapped concerns genera- volved. What is shared in collective memory is often
tional differences. Martin A. Conway (1997) and oth- little more than a schematic narrative template
ers have examined the reminiscence bump that (Wertsch, 2002) in which detailed information, espe-
exists for members of a generation for events that oc- cially contradictory information, is distorted, simpli-
curred when they were between fifteen and twenty- fied, and ignored; this schema stands in contrast to
five years of age. Research suggests that such events what are at least the aspirations of analytic history. It
provide the foundation for a generations lifelong col- is for this reason that collective memory often appears
lective memory. Providing another hint about the to be impervious to new information that might chal-
close tie between identity and collective memory, lenge it.
Conway argues that the reminiscence bump and its Bibliography
power to shape a generation are inherently tied to Anderson, B. (1991). Imagined communities: Reflections on the origin
processes of identity development in young adult- and spread of nationalism. London: Verso.
hood. Bartlett, F. C. (1932; reprint 1995). Remembering: A study in experi-
mental and social psychology. Cambridge, UK: Cambridge Uni-
versity Press.
Collective Remembering versus History Bodnar, J. (1992). Remaking America: Public memory, commemoration,
and patriotism in the twentieth century. Princeton, NJ: Princeton
Notions of collective memory and history often University Press.
overlap. In both cases, the events involved are likely Chang, I. (1997). The rape of Nanking: The forgotten holocaust of World
to have occurred before the lifetime of the individuals War II. New York: Basic Books.
Cole, J. (2001). Forget colonialism? Sacrifice and the art of memory in
or group doing the remembering, and in both cases
Madagascar. Berkeley: University of California Press.
there is a tendency to assert that the account being Conway, M. A. (1997). The inventory of experience: Memory and
presented is true. Furthermore, both rely on narra- identity. In J. W. Pennebaker, D. Paez, and B. Rim, eds., Col-
tive as a cultural tool. Despite such shared properties, lective memory of political events: Social psychological perspectives.
it is possible, indeed essential, to distinguish between Mahwah, NJ: Erlbaum.
Donald, M. (1991). Origins of the modern mind: Three stages in the evo-
them.
lution of culture and cognition. Cambridge, MA: Harvard Uni-
Among the properties of collective memory that versity Press.
tend to distinguish it from history are the following: Halbwachs, M. (1952; reprint 1992). On collective memory, ed. and
trans. Lewis A. Coser. Chicago: University of Chicago Press.
1. it reflects a single, subjective, committed perspec- Novick, P. (1999). The Holocaust in American life. Boston: Houghton
tive of a collective, whereas analytic history strives Mifflin.
Ross, M. (1989). Relation of implicit theories to the construction
to be objective and distance itself from any partic- of personal histories. Psychological Review 96 (2), 341357.
ular perspective; Schudson, M. (1992). Watergate in American memory: How we remem-
2. collective memory leaves little room for doubt or ber, forget, and reconstruct the past. New York: Basic Books.
Weldon, M. S. (2001). Remembering as a social process. In D. L.
ambiguity about events and the motivations of ac-
Medin, ed., The psychology of learning and motivation. San
tors (Novick, 1999), whereas analytic history Diego, CA: Academic Press.
strives to take into account multiple, complex fac- Wertsch, J. V. (2002). Voices of collective remembering. New York:
tors; Cambridge University Press.
3. collective memory presupposes an unchanging James V. Wertsch

essence of a group across time, whereas analytic
history focuses on the transformations that collec-
tives undergo.
These three properties of collective memory charac-
COMPARATIVE COGNITION
terize, for example, the highly charged dispute be- Comparative cognition is the comparison of how ani-
tween China and Japan over whether events in 1937 mals (including humans) process and interact with
constitute the rape of Nanking (Chang, 1997) or the world. What sets animal cognition apart from the
the incident in Nanking (as mentioned in some behavioristic tradition of stimulus-response (S-R)
Japanese textbooks). learning is the recognition that (mental) processes in-
COMPARATIVE COGNITION 89
Figure 1
An information-processing model of temporal processing representative of scalar timing theory.
tervene between the stimulus and response. Some an- creases. The peak indicates maximum expectancy of
imal-cognitive procedures may be complex (e.g., reward. Interval timing typically displays the scalar
feeling of knowing experiments by Hampton, propertyvariability (spread) increases proportion-
2001). But even simple Pavlovian conditioning pro- ally with time. The interval-timing theory receiving
cedures can deal with cognitive processes: for exam- the most attention is shown in Figure 1. It includes a
ple, taste-aversion conditioning to the image of an clock (pacemaker), a switch (gate) to an accumulator
expected type of food (e.g., Holland, 1990), or (integrator), reference memory for the expected
higher-order and backward fear conditioning to a time, and a comparator for the reference time versus
neutral stimulus never paired with shock (Cole, Bar- accumulated time to see if times up. Converging
net, and Miller, 1995). Techniques developed in the evidence has been accumulating for each of these hy-
late twentieth century to study cognitive processes pothetical processes. One example is that drug ma-
have produced exciting revelations about how ani- nipulations of the clock produce abrupt changes in
mals perceive, learn, remember, navigate, communi- peak time (see Figure 2) with gradual compensation
cate, time, and count. This entry highlights some of by storage of new accumulator values in reference
these developments. memory, whereas drug manipulations of the memory
produce gradual changes (see Figure 3) due to dis-
torted intervals accumulating in reference memory
Interval Timing (Meck, 1996).
A tremendous amount of information has accu-
mulated since the 1980s about the interval-timing
mechanisms of rats, pigeons, and primates. In a pop- Numerosity: A Scale of Numbers
ular peak procedure, a stimulus cue is presented Learning a number scale is fundamental to math-
and after a fixed elapsed time the next response is ematics. Evidence indicates that rhesus monkeys can
rewarded. On some trials, reward is omitted and learn a number scale (Brannon and Terrace, 1998,
responding increases to a peak rate and then de- 2000). Monkeys touched randomly placed areas on a
Figure 2
Mean peak times from four different groups of rats (n=10/group) initially trained on either a 40-s (squares) or 20-s (circles) PI-timing
procedure and tested under the influence of dopaminergic drugs that affect clock speed (left panel) or after the removal of the drugs
(right panel). The closed symbols represent rats treated with methamphetamine (1.5 mg/kg i.p.), and the open symbols represent rats
treated with haloperidol (0.12 mg/kg i.p.).
computer screen in order: first the area containing Landmark Navigation and Cognitive Maps
one object, then two, then three, and finally four ob- Clarks nutcrackers apparently use directional
jects. Objects varied in size, shape, or color and com- bearings from several landmarks to fix the location
binations of these dimensions within and across trials. of seed caches (Kamil and Jones, 2000; Kamil and
The most remarkable result was that this training Cheng, 2001). Nutcrackers were more accurate using
spontaneously generalized to larger novel numbers of multiple bearings than any single bearing with
five to nine objects beyond the range of training num- a distance measure along the bearing line. This
bers. remarkable finding came from manipulating the
Figure 3
Mean peak times from four different groups of rats (n=10/group) initially trained on either a 40-s (squares) or 20-s (circles) PI-timing
procedure and tested under the influence of cholinergic drugs that affect memory-storage speed (left panel) or after the removal of the
drugs (right panel). The closed symbols represent rats treated with physostigmine (0.01 mg/kg i.p.) and the open symbols represent
rats treated with atropine (0.05 mg/kg i.p.).
goal site in relationship to geometrical properties than configural-landmark navigation (Shettleworth,

of fixed landmarks rather than manipulating the 1998).
landmarks themselves. Landmarks seem to function
as a configural whole, not as independent elements.
Young children, but not older children or adults, use Abstract Concepts
similar geometric navigation strategies (Hermer- Researchers have long debated which species are
Vazquez, Moffet, Munkholm, 2001). The issue of capable of abstract concept learning. Abstract con-
cognitive maps, however, is complex and is more cepts are rules (same versus different, identity match-
ing) that transcend the training stimuli. Novel stimuli ented with pairs of them (behind a window). They
are used to test for abstract concept learning. Pigeons were required to pull one of two ropes to indicate
(like humans, apes, dolphins, and monkeys) are capa- whether they were of the same or different categories
ble of learning abstract concepts. One study showed (food versus nonfood). Novel item transfer was better
matching-to-sample (MTS) concept learning by pi- than 80 percent correct. Rhesus monkeys have also
geons (Wright, 1997). In MTS, a sample stimulus is showed category matching (Neiworth and Wright,
followed by two choice stimuli, one of which is correct 1994).
and matches the sample. Pigeons required to respond
to (peck) the sample twenty times before getting the
choice stimuli learned the concept fully (as opposed Relations Between Relations: Analogical
to partially). Other pigeons required to respond fewer Reasoning
times (e.g., once) did not learn the concept at all, and If abstract concept learning is higher order, then
instead learned the configural pattern of each display analogical reasoning is higher-higher order. The Mil-
(sample plus choice stimuli). lers Analogies Test is a test of relations between rela-
tions. A recent MTS study showed that baboons could
Pigeons are also capable of learning a same/
learn a relation between relations (Fagot, Wasserman,
different (S/D) abstract concept. In S/D, subjects iden-
and Young, 2001). Baboons saw a brief sixteen-icon
tify stimuli as same or different. Pigeons most rapidly
sample, which was then replaced by two sixteen-icon
learn S/D (and the S/D concept partially) with large
choice displays. Neither choice display had any icons
384 element arrays and contrasting target regions on
in common with the sample. The baboons successfully
different trials (Cook, Katz, and Cavoto, 1997), or
learned and transferred this performance to novel
with sixteen-element arrays that are composed of all
stimuli, which could not be based on stimulus identity
different or all same elements (Young, Wasserman,
or functional categories. It is difficult to escape the
and Garner, 1997). But when the number of elements
conclusion that this behavior is based on a relation be-
in these tasks is reduced to two (minimum number),
tween relations. A similar conclusion came from a
performance falls to chance. Nevertheless, pigeons
study on rhesus monkey music perception (Wright et
are capable of learning a S/D task when trained with
al., 2000). Six-note tonal (childhood songs, tonal al-
two stimuli from the beginning and attain full S/D
gorithm) musical passages, separated by one or two
concept learning when the training set is expanded
octaves were judged same better than 80 percent of
to 128 to 512 pictures. Rhesus monkeys attained full
the time. Tonal passages form a relation (tune), and
S/D concept learning with somewhat smaller set sizes
thus this octave generalization (no common frequen-
than pigeons, and chimpanzees remarkably learned
cies) is a relation between relations.
an S/D concept with only two training stimuli (Oden,
Thompson, and Premack, 1988). Great apes are particularly remarkable in their
ability to form relations between relations. Chimpan-
Equally remarkable is the S/D concept learning zees with no relations-between-relations training
by a parrot, Alex, with stimuli that varied along three spontaneously performed correct analogical
dimensions (Pepperberg, 1987). Alex responded with choices without reward (Thompson, Oden, and
the English words color, shape, or mah-mah Boysen, 1997). If anything could be more remark-
(matter). When Alex was presented with a red wooden able, Sarah, a highly trained chimpanzee arranged
triangle and a green rawhide triangle, for example, four out of five scattered geometric forms in the
and asked, Whats same? he typically said, Shape. unique analogical relationship (and set aside the dis-
When presented with a red wooden square and a blue tractor) without explicit training (Thompson and
wooden square and asked, Whats different? he typ- Oden, 2000).
ically said, Color. Alexs novel object transfer per-
formance was 85 percent correct. What makes this ex-
periment so remarkable was that Alex had no way of Episodic Memory
telling which dimension (color, shape, or matter) Episodic memory is memory for a particular
would be questioned and whether he would be asked, event and is related to the issue of consciousness in
Whats same? or Whats different? animals. Recent experiments show that scrub jays
have episodic memory (Clayton and Dickinson, 1998;
Clayton and Dickinson, 1999; Emery and Clayton,
Abstract Concepts Based on Categories 2001). In one experiment, jays cached wax worms and
Research developments have shown that mon- peanuts in separate and distinctive halves of two dis-
keys could learn an S/D concept with food versus non- tinctive sand-filled ice cube trays (two-by-seven ar-
food categories (Bovet and Vauclair, 2001). Baboons rays). After four hours, the jays recovered the more
were given experience with the items and then pres- desirable wax worms first from one tray. After 124
hours, they recovered the peanuts first from the other ditions would not show this ability. The study of quan-
tray. In pretraining, they had learned that the wax titative differences encounters the same difficulty.
worms deteriorated after 124 hours. Reversal of their Some argue that comparisons should be among close-
earlier preference for wax worms means that they rely related species to study cognitive evolution (Cam-
membered what (peanuts versus wax worms), when bell and Hodos, 1991) or environmentally specialized
(four versus 124 hours), and where (which tray side) cognitive behavior (Sherry and Schacter, 1987). But
the foods were stored. Cache recovery by Clarks nut- even testing closely related species (e.g., Clarks nut-
crackers in the forest must also be episodic memory crackers, Mexican jays, pinyon jays, and scrub jays) in
because they use different cache sites each year and the same task does not ensure that quantitative differ-
overcome potential proactive interference (i.e., con- ences represent cognitive differences, as opposed to
fusions) from previous years sites. Thus, they too the task fitting the predispositions of some of the spe-
know what, when, and where. cies better. Nevertheless, testing the species in a vari-
ety of tasks and varying critical parameters can con-
verge on persuasive evidence for quantitative
Feeling of Knowing: Metacognition
differences (Olson, Kamil, Balda, and Nims, 1995).
Exciting research developments with monkeys in-
dicate that they know when they remember (Hamp- A more overarching goal would seem to under-
ton, 2001). This ability in humans indicates conscious stand the mechanisms and processes involved; in
cognition and is closely aligned with the issue of con- short, how cognition works. Any thorough under-
sciousness and recognition of self. In a MTS experi- standing will need evidence from at least three do-
ment, rhesus monkeys saw a sample (clip-art image), mains: conditions (i.e., natural context) best suited to
touched the sample, had a retention delay, and then express this cognitive ability; laboratory tests that ma-
decided whether or not to take a memory test. If they nipulate critical parameters over a substantial range;
chose to take the memory test, four choices were pres- and neurobiology and neural circuitry that mediates
ented on the four corners of the monitor and reward this behavior. One can begin in any domain, but will
for the correct choice was highly desired peanuts. If need regular input from the other domainsa sort of
they declined to take the test, they (automatically) re- cumulative upward spiral and a revisit of issues as evi-
ceived a less preferred pellet. The monkeys were dence accumulates. Ultimately, quantitative differ-
more accurate when they chose to take the test than ences (e.g., interval timing [Church, Meck, and Gib-
when they were required to take the test, and this dif- bon, 1994]) and even qualitative differences (e.g.,
ference increased with delay. An important aspect of serial-order learning [Terrace, Chen, and Newman,
this study was requiring the monkeys to monitor their 1995]; memory rehearsal [Cook, Wright, and Sands,
memory before the test was presented. It is difficult 1991]) should be shown to be individual points of
to escape the conclusion that these monkeys know comparison along much larger functional relation-
when they remember. Pigeons do not show this ships revealed by parametric studies and quantitative
choice-test advantage (Inman and Shettleworth, models.
1999), and thus may lack metacognition. The Cognitive Issue
Issues associated with the cognition part of com-
Directions for Future Research parative cognition mainly revolve around the
Animals have cognitive abilities that were previ- strength of evidence for private events (cognitive pro-
ously thought to be uniquely human. Other qualita- cesses) between the stimulus and response. There will
tive differences may disappear as procedures better certainly be skeptics of the feeling-of-knowing experi-
fit the predispositions of species being tested (Clayton ments with animals (Hampton, 2001), but this is only
and Dickinson, 1999). But there are issues other than the tip of the iceberg. It gets only better (or worse de-
finding human cognitive abilities in animals and pending on ones persuasion) with experiments on
these issues continue to revolve around what is meant beliefs, intentions, wants, desires, altruism, concept of
by comparative and cognition. self, and knowledge about anothers beliefs and inten-
tionsthe so-called theory of mind experiments
The Comparative Issue (Povinelli, 2001; Tomasello and Call, 1997). These
Criticisms of comparative cognition have been investigations have stimulated a great deal of enthusi-
that they are not comparative (Shettleworth, 1993). asm among scientists. Continued growth and accep-
However, researchers have seen species comparisons tance will likely depend on parametric studies (i.e.,
in interval timing, landmark navigation, abstract con- manipulating critical variables over a substantial
cept learning, analogical reasoning, and metacogni- range) and converging evidence from several experi-
tion. A problem with proving qualitative differences ments like the blue jay and wax-worm studies. Fur-
is that one can never be sure that some other test con- thermore, researchers will need to show that their
findings cannot be accounted for by simpler associa- Hermer-Vazquez, L. Moffet, A., and Munkholm, P. (2001). Lan-
tive learning principles involving observable stimuli guage, space, and the development of cognitive flexibility in
humans: The case of two spatial memory tasks. Cognition 79,
and responses, like behavior-based theories timing 263299.
(Killeen and Fetterman, 1988; Machado, 1997) as op- Holland, P. C. (1990). Event representation in Pavlovian condi-
posed to scalar-timing theory (see Figure 1). A wide tioning: Image and action. Cognition 37, 105131.
variety of learning, cognitive, and memory phenome- Inman, A., and Shettleworth, S. J. (1999). Detecting metamemory
na can be accounted for by careful parametric experi- in nonverbal subjects: A test with pigeons. Journal of Experi-
mental Psychology: Animal Behavior Processes 25, 389395.
mentation coupled with powerful mathematical mod-
Kamil, A. C., and Cheng, K. (2001). Way-finding and landmarks:
els to reveal underlying processes (Gallistel and The multiple-bearings hypothesis. Journal of Experimental Biol-
Gibbon, 2000; White and Wixted, 1999). Any time we ogy 204, 103113.
are tempted to get too heady over some new animal- Kamil, A. C., and Jones, J. E. (2000). Geometric rule learning by
cognitive phenomenon, we might reflect on the exu- Clarks nutcrackers (Nucifraga columbiana). Journal of Experi-
mental Psychology: Animal Behavior Processes 26, 439453.
berance originally generated by ape-language
Katz, J. S., Wright, A. A., and Bachevalier, J. (2002). Same/different
studies, and recall the instructive (if not tongue-in- concept learning in Rhesus monkeys. Journal of Experimental
cheek) S-R conditioning demonstrations of the pi- Psychology: Animal Behavior Processes 28.
geons (cognitive) concepts of insight, symbolic com- Killeen, P. R., and Fetterman, J. G. A behavioral theory of timing.
munication, and self conducted by Epstein, Skinner, Psychological Review 95, 274295.
and their collaborators (1981). Machado, A. (1997). Learning the temporal dynamics of behavior.
Psychological Review 104, 241265.
Meck, W. H. (1996). Neuropharmacology of timing and time per-
Bibliography ception. Cognitive Brain Research 3, 227242.
Neiworth, J. J., and Wright, A. A. (1994). Monkeys (Macaca mulat-
Bovet, D., and Vauclair, J. (2001). Judgment of conceptual identity
in monkeys. Psychonomic Bulletin & Review 8, 470475. ta) learn category matching in a nonidentical same-different
Brannon, E. M., and Terrace, H. S. (1998). Ordering of the task. Journal of Experimental Psychology: Animal Behavior Process-
numerosities 1 to 9 by monkeys. Science 282, 746749. es 20, 429435.
(2000). Representation of the numerosities 19 by rhesus Oden, D. L., Thompson, R. K., and Premack, D. (1988). Spontane-
macaques (Macaca mulatta). Journal of Experimental Psychology: ous transfer of matching by infant chimpanzees (Pan troglo-
Animal Behavior Processes 26, 3149. dytes). Journal of Experimental Psychology: Animal Behavior Pro-
Campbell, C. B. G., and Hodos, W. (1991). The scala naturae revis- cesses 14, 140145.
ited: Evolutionary scales and anagenesis in comparative psy- Olson, D. J., Kamil, A. C., Balda, R. P., and Nims, P. J. (1995). Per-
chology. Journal of Comparative Psychology 105, 211221. formance of four seed-caching corvid species in operant tests
Church, R. M., Meck, W. H., and Gibbon, J. (1994). Application of of nonspatial and spatial memory. Journal of Comparative Psy-
scalar timing theory to individual trials. Journal of Experimental chology 109, 173181.
Psychology: Animal Behavior Processes 20, 135155. Pepperberg, I. M. (1987). Acquisition of the same/different concept
Clayton, N. S., and Dickinson, A. (1998). Episodic-like memory by an African Grey parrot (Psittacus erithacus): Learning with
during cache recovery by scrub jays. Nature 395, 272274. respect to categories of color, shape, and material. Animal
(1999). Memory for the content of caches by scrub jays Learning and Behavior 15, 423432.
(Aphelocoma coerulescens). Journal of Experimental Psychology: Povinelli, D. J. (2001). On the possibilities of detecting intentions
Animal Behavior Processes 25, 8291. prior to understanding them. In B. F. Malle, L. J. Moses, and
Cole, R. P., Barnet, R. C., and Miller, R. R. (1995). Temporal en- D. A. Baldwin, eds., Intentions and intentionality: Foundations of
coding in trace conditioning. Animal Learning and Behavior 23, social cognition. Cambridge, MA: MIT Press.
144153. Sherry, D. F., and Schacter, D. L. (1987). The evolution of multiple
Cook, R. G., Katz, J. S., and Cavoto, B. R. (1997). Pigeon same- memory systems. Psychological Review 94, 439454.
different concept learning with multiple stimulus classes. Jour- Shettleworth, S. J. (1993). Where is the comparison in comparative
nal of Experimental Psychology: Animal Behavior Processes 23, cognition? Alternative research programs. Psychological Science
417433. 4, 179184.
Cook, R. G., Wright, A. A., and Sands, S. F. (1991). Interstimulus (1998). Cognition, evolution, and behavior. New York: Oxford
interval and viewing time effects in monkey list memory. Ani- University.
mal Learning and Behavior 19, 153163. Terrace, H. S., Chen, S., and Newman, A. B. (1995). Serial learning
Emery, N. J., and Clayton, N. S. (2001). It takes a thief to know a with a wild card by pigeons (Columba livia): Effect of list
thief: Effects of social context on prospective caching strate- length. Journal of Comparative Psychology 109, 1,1621,172.
gies in scrub jays. Nature 414, 443446. Thompson, R. K. R., and Oden, D. L. (2000). Categorical percep-
Epstein, R., Lanza, R. P., and Skinner, B. F. (1981). Self- tion and conceptual judgments by nonhuman primates: The
awareness in the pigeon. Science 212, 695696. paleological monkey and the analogical ape. Cognitive Science
Fagot, J., Wasserman, E. A., and Young, M. E. (2001). Discriminat- 24, 363396.
ing the relation between relations: The role of entropy in ab- Thompson, R. K. R., Oden, D. L., and Boysen, S. T. (1997). Lan-
stract conceptualization by baboons (Papio papio) and hu- guage-nave chimpanzees (Pan troglodytes) judge relations
mans (Homo sapiens). Journal of Experimental Psychology: between relations in a conceptual matching-to-sample task.
Animal Behavior Processes 27, 316328. Journal of Experimental Psychology: Animal Behavior Processes 23,
Gallistel, C. R., and Gibbon, J. (2000). Time, rate, and condition- 3143.
ing. Psychological Review 107, 289344. Tomasello, M., and Call, J. (1997). Primate cognition. New York: Ox-
Hampton, R. R. (2001). Rhesus monkeys know when they remem- ford University Press.
ber. Proceeding of the National Academy of Sciences of the United White, K. G., and Wixted, J. T. (1999). Psychophysics of remember-
States of America 98, 5,3595,362. ing. Journal of the Experimental Analysis of Behavior 71, 91113.
CONCEPTS AND CATEGORIES, LEARNING OF 95
Wright, A. A. (1997). Concept learning and learning strategies. Psy- Medin, 1981) have led to a shift in attention from the
chological Science 8, 119123. classical view to the probabilistic view.
Wright, A. A., Rivera, J. J., Hulse, S. H., Shyan, M., and Neiworth,
J. J[GU2]. (2000). Music perception and octave generalization The probabilistic view argues that most concepts
in rhesus monkeys. Journal of Experimental Psychology: General are organized around properties that are only charac-
129, 291307. teristic or typical of a category (rather than defining).
Young, M. E., Wasserman, E. A., and Garner, K. L. (1997). Effects
One specific account of the probabilistic view assumes
of number of items on the pigeons discrimination of same
from different visual displays. Journal of Experimental Psycholo- that a concept is a summary representation or proto-
gy: Animal Behavior Processes 23, 491501. type that indicates what is, on the average, true of a
category (e.g., the bird prototype would include fea-
Anthony A. Wright
tures such as flies and sings because the properties are
true of most though not all birds). The prototype view
readily handles goodness-of-example effects; the
more similar an item is to the prototype, the more
CONCEPTS AND CATEGORIES, typical it will be of the category. For example, robins
LEARNING OF would be more similar to the bird prototype than are
Concepts are a fundamental aspect of intelligent be- ostriches because they have more features that are
havior. Traditionally, a concept has been viewed as a generally true of birds. Thus, they would be better ex-
mental representation that picks out a group of equiv- amples of birds.
alent items or a category. For example, every person However, the prototype view also has problems.
has a concept of dog and can use that concept to pick Prototype representations alone are not rich enough
out a category of things that one would call dogs. Some to capture peoples knowledge about a category. For
of the most fundamental questions about the mind in- example, people are sensitive to the number of in-
clude the following: What do human concepts consist stances in a category (e.g., there are many more hous-
of (i.e, what is their structure)? How are they are ac- es than igloos), the variability of features (e.g., the size
quired? Why do humans have concepts (i.e, What of quarters varies less than the size of pizzas), and the
functions do they have)? correlations between features (e.g., wooden spoons
tend to be big whereas metal spoons tend to be small
[Medin, 1989]). To overcome such difficulties, re-
What Do Human Concepts Consist Of?
searchers have suggested that instead of (or in addi-
An early, popular view of concepts was the classi- tion to) a prototype people may simply store memory
cal view. For a variety of reasons this approach was instances or examples of the category and reason with
unsatisfactory and gave way to the probabilistic view. them (Brooks, 1978). Thus, according to the exem-
These views are described and compared below. plar view, for instance, the concept of a chair would
Classical versus Probabilistic View include memory traces of particular chairs and their
associated features.
The classical view argues that concepts are struc-
tured around defining features (Bruner, Goodnow, The Theory View
and Austin, 1956). Defining features are features that Another approach to concepts is the theory view.
are singly necessary and jointly sufficient to define the As described below, this approach developed as a re-
concept. For example, the concept bachelor has the action to certain limitations of the classical and proba-
defining features human, unmarried, and male. bilistic approaches. In particular, the classical and
However, the classical view appears to have a probabilistic views failed to take into account peoples
number of serious problems. First, if concepts have background knowledge or theories of the world.
defining features, then one ought to be able to specify The shift from the classical view to the probabilis-
what they are. But many common concepts, such as tic view was motivated by a detailed analysis of natural
game or chair, seem to have no defining features. In- object categories. Associated with this analysis is the
stead, instances of these concepts have characteristic view that concepts or mental representations of cate-
features that are neither necessary nor sufficient for gories closely mirror the structure afforded by prop-
category membership (e.g., has four legs and has a back erties of category members. It seems almost a tautolo-
for chairs). Second, not all instances of a concept are gy that if the structure of examples does not have
equally good examples of that concept. For example, defining features then the corresponding mental rep-
people judge robins to be better examples of the con- resentations cannot conform to the classical view.
cept bird than ostriches. If both robins and ostriches Similarly, a probabilistic category structure suggests
have the defining features of birds, then why should a probabilistic concept representation. In brief, re-
robins be considered better examples of birds than searchers assume that mental representations are de-
ostriches? These and other problems (Smith and termined by the structure of examples in the world.
96 CONCEPTS AND CATEGORIES, LEARNING OF
However, the classical and probabilistic views bers (Posner and Keele, 1968), by attending to fea-
tend to ignore the role of the learner in their accounts tures commonly shared by members and discarding
of concepts. According to the theory view, learners features varying among members (Elio and Ander-
also impose structure on their concepts. That is, con- son, 1981), or by noting the most common value on
cepts are based on a learners general knowledge and each dimension. The basic idea behind these models
theories of the world together with information pro- can be traced to Galtons composite photograph
vided by the environment (Carey, 1985; Murphy and theory (Galton, 1879). Galton superimposed several
Medin, 1985; Rips, 1989). For example, Susan Carey faces to make a composite photograph in which com-
showed that childrens biological theories influence mon properties were accentuated and variant proper-
their patterns of inductions at a very early age. To il- ties were attenuated. Such a process is assumed in
lustrate, a mechanical monkey is rated by both chil- prototype theories. On the other hand, exemplar
dren and adults to be more similar to a human being models assume category instances are stored but gen-
than is a worm, yet even young children infer that erally have not specified detailed learning mecha-
worms rather than toy monkeys have a spleen after nisms (see Kruschke, 1992, for an exception). These
being told that people have a spleen, a round and green views assume that the learner begins with features of
[thing] . . . in the persons body. In this example, the fea- the entities and then learns which features are impor-
ture has a spleen, which is more consistent with a tant for the concept. However, research conducted in
childs background knowledge or theory about ani- the 1990s suggests that an important part of concept
mate things than inanimate things. learning is learning to identify the features them-
Theories themselves may be anchored by how selves (Schyns, Goldstone, Thibaut, 1998).
well their predictions receive support from the world. The theory-based view of concepts takes a differ-
Gregory Murphy and Douglas Medin (1985) suggest ent perspective on concept formation. Several re-
that the relation between concept and examples is searchers have proposed that humans may be born
like that between theory and data. Thus, concepts with a naive physics and a naive biology or psychology
would not necessarily consist of features that are also (Carey, 1985; Keil, 1989; Spelke, 1990) that act as ini-
in examples; rather, the constituents of examples tial theories to organize conceptual knowledge. A
would only need to support more abstract constituents major implication of the theory-based view is that
of concepts (Wisniewski and Medin, 1994). For exam- concept learning involves integrating new examples
ple, one may infer that a man is drunk because one with prior knowledge. In particular, prior knowledge
sees him jump into a pool fully clothed. If one does may influence the identification of features and, in
so, it is probably not because the feature jumps into turn, information about examples may modify a per-
pools, clothed is listed with the concept drunk. Rather, sons prior knowledge (Wisniewski and Medin, 1994).
it is because part of ones concept of drunk involves a
theory of impaired judgment that serves to explain Taking the theory-based view, a group of re-
the mans behavior. searchers in artificial intelligence (an area in comput-
er science the goal of which is to develop computers
to do intelligent things) have developed models of
How Are Concepts Acquired? concept formation called explanation-based learning
Young children probably enter the world with few (Mitchell, Keller, and Kedar-Cabelli, 1986; DeJong
preexisting concepts. Instead, they must acquire or and Mooney, 1986). These models suggest that the
form concepts from experiences (e.g., form a concept most important aspect of concept learning is to ex-
of dog from existing experiences with dogs). As de- plain why a given example is an instance of the con-
scribed below, the classical, probabilistic, and theory cept. Construction of the explanation is carried out by
views of concepts propose different ways in which causally connecting known concepts. For example,
concepts are acquired. suppose a computer is to learn a concept cup and it
According to the classical view, the process of con- already knows such concepts as liftable, handle, liquid
cept formation is one of discovering necessary and container, and stable. Seeing an object that can be lift-
sufficient attributes by observing which attributes ed, has a handle, contains liquid, and is stable, the sys-
occur in all members and only in members of the cate- tem uses its background knowledge to construct an
gory. Research associated with the classical view has explanation about why one can drink from this object.
been directed at investigating hypothesis testing Then it generalizes this explanation to develop its
strategies, with each hypothesis being a guess as to concept of cup.
which features are part of the definition (Levine, Learning by analogy is another form of theory or
1971). knowledge-driven learning in which a known similar
In contrast, according to the probabilistic view, concept is modified. For example, one can learn
concept learning occurs by averaging values of mem- about the internal structure of atoms by applying
CONDITIONING 97
ones knowledge of solar system (e.g., electrons re- Kruschke, J. K. (1992). ALCOVE: An exemplar-based connection-
volve around the nucleus as planets revolve around ist model of category learning. Psychological Review 99, 2244.
Levine, M. (1971). Hypothesis theory and non-learning despite
the sun [Gentner, 1989]). One can also discover new
ideal S-R reinforcement contingencies. Psychological Review
features through analogy or metaphors (e.g., given a 45, 626632.
smile is like a magnet, one can learn that a smile at- Malt, B. C., Sloman, S. A., Gennari, S., Shi, M., and Wang, Y.
tracts). (1999). Knowing versus naming: Similarity and the linguistic
categorization of artifacts. Journal of Memory and Language 40,
230262.
The Why of Concept Learning Markman, A. B., and Makin, V. S. (1998). Referential communica-
There are many reasons why humans have con- tion and category acquisition. Journal of Experimental Psycho-
logy: General 127, 331354.
cepts. They allow people to classify things (e.g., recog-
Medin, D. L. (1989). Concepts and conceptual structure. American
nize that something is snake) and to make important Psychologist 12, 1,4691,481.
predictions or inferences (e.g., that snake may be poi- Mitchell, T. M., Keller, R. M., and Kedar-Cabelli, S. T. (1986). Ex-
sonous). John Anderson (1990) has developed a theo- planation-based generalization: A unifying view. Machine
ry of concepts that emphasizes this prediction func- Learning 1, 4780.
tion. Other functions include explanation (e.g., the Murphy, G. L., and Medin, D. L. (1985). The role of theories in
concept introvert might help to explain why some per- conceptual coherence. Psychological Review 92, 289316.
Posner, M. L., and Keele, S. W. (1968). On the genesis of abstract
son did not attend a party), reasoning (deriving
ideas. Journal of Experimental Psychology 77, 353363.
knowledge from the stored information), communi- Rips, L. J. (1989). Similarity, typicality, and categorization. In S.
cation, and conceptual combination (e.g., from the Vosniadou and A. Ortony, eds., Similarity and analogical reason-
concepts glass and elephant one might construct the ing. Cambridge, UK: Cambridge University Press.
combined concept glass elephant). Many approaches to Ross, B. H. (1997). The use of categories affect classification. Jour-
concept learning have focused only on the classifica- nal of Memory and Language 37, 240267.
tion function. However, functions of concepts interact Schyns, P. G., Goldstone, R. L., and Thibaut, J. P. (1998). The de-
velopment of features in object concepts. Behavioral and Brain
such that it is important to study multiple functions
Sciences 21, 154.
together. For example, Brian Ross (1997) found that Smith, E. E., and Medin, D. L. (1981). Categories and concepts. Cam-
the diagnosis (classification) of a disease was impor- bridge, MA: Harvard University Press.
tantly influenced by features that were relevant to its Spelke, E. S. (1990). Principles of object perception. Cognitive Sci-
treatment (Malt et al., 1999; Markman and Makin, ence 14, 2956.
1998, for other examples of interactions). Research- Wisniewski, E. J., and Medin, D. L. (1994). On the interaction of
theory and data in concept learning. Cognitive Science 18, 221
ers have begun to appreciate and investigate the vari-
281.
ety of functions that concepts have.
Douglas L. Medin
See also: SEMANTIC MEMORY: COGNITIVE ASPECTS;
SEMANTIC MEMORY: NEUROBIOLOGICAL Wou-Kyoung Abn
PERSPECTIVE Revised by Edward Wisniewski
Bibliography
Anderson, J. R. (1990). The adaptive characteristic of thought. Hills-
dale, NJ: Erlbaum.
Brooks, L. R. (1978). Non-analytic concept formation and memory
for instances. In E. Rosch and B. Lloyd, eds., Cognition and cat- CONDITIONING
egorization. Hillsdale, NJ: Erlbaum.
Bruner, J. S., Goodnow, J. J., and Austin, G. A. (1956). A study of
[A survey of the forms of learning called conditioning will
thinking. New York: Wiley. be found under CONDITIONING, CLASSICAL AND IN-
Carey, S. (1985). Conceptual change in childhood. Cambridge, MA: STRUMENTAL. The following articles discuss specific as-
MIT Press. pects of the classical conditioning paradigm in greater de-
DeJong, G. F., and Mooney, R. J. (1986). Explanation-based learn-
tail:
ing: An alternative view. Machine Learning 1, (2) 145176.
Elio, R., and Anderson, J. R. (1981). The effects of category gener- CLASSICAL CONDITIONING: BEHAVIORAL
alizations and instance similarity on schema abstraction. Jour-
PHENOMENA
nal of Experimental Psychology: Human Learning and Memory 7,
CONDITIONING, CELLULAR AND NETWORK
397417.
Galton, F. (1879). Composite portraits, made by combining those SCHEMES FOR HIGHER-ORDER FEATURES OF
of many different persons into a single, resultant figure. Jour- CLASSICAL
nal of the Anthropological Institute 8, 132144. NEURAL SUBSTRATES OF CLASSICAL
Gentner, D. (1989). The mechanisms of analogical learning. In S. CONDITIONING
Vosniadou and A. Ortony, eds., Similarity and analogical reason-
ing. Cambridge, UK: Cambridge University Press.
Keil, F. C. (1989). Concepts, kinds, and cognitive development. Cam- A particular type of instrumental conditioning is more close-
bridge, MA: MIT Press. ly examined under OPERANT BEHAVIOR.]
98 CONDITIONING, CELLULAR AND NETWORK SCHEMES
CONDITIONING, CELLULAR AND generally activated by the US or whether qualitatively

NETWORK SCHEMES FOR HIGHER- different associations are formed between the CSs.
The notion that the same associative forms of synaptic
ORDER FEATURES OF CLASSICAL
plasticity are involved in both the formation of the
Experimental work on invertebrates and vertebrates CS1-US and CS2-CS1 associations has been corrobo-
has begun to allow the analysis of the neural mecha- rated by studies in rat fear conditioning that have
nisms underlying second-order conditioning, block- shown that the same manipulation that blocks first-
ing, and contingency. Here we look at the experimen- order conditioning also blocks second-order condi-
tal and theoretical data that provide a cellular- and tioning without blocking the conditioned response
network-level description of higher-order forms of (Gewirtz and Davis, 1997). It was shown that infusion
classical conditioning. of an NMDA antagonist (which blocks associative syn-
aptic plasticity) into the amygdala prevented second-
order conditioning. This finding supports the notion
Experimental Models of Higher-Order that second-order conditioning relies on the same
Forms of Classical Conditioning forms of synaptic plasticity as classical conditioning.
The development of neural analogs of habitua-
tion, sensitization, and classical conditioning have
greatly enhanced the understanding of the neural Blocking
mechanisms responsible for these forms of learning. Blocking, along with contingency, establishes that
A logical extension of these studies is to develop neu- pairing of a CS with a US is not sufficient to produce
ral analogs of higher-order forms of classical condi- conditioning. Blocking shows that it is important that
tioning. Research has shown that because many the CS contains novel information about the US. A
higher-order forms of conditioning are present in in- typical blocking protocol (Kamin, 1968) consists of
vertebrates, these preparations may lead to the estab- two phases. First CS1 is paired with the US. In Phase
lishment of neural analogs of higher-order condition- II a compound stimulus CS1+/CS2 is paired with the
ing. Moreover, recent research in vertebrates has also US. In the case of complete blocking, CS2 does not
begun to examine the neural mechanisms of higher- undergo any conditioning. The argument is that even
order conditioning. though it was explicitly paired with the US, CS2 does
not provide any information that was not already pre-
dicted by CS1+.
Second-Order Conditioning
Blocking has been described in many vertebrates,
In second-order conditioning, a CS produces a in the snail Limax (Sahley et al., 1982) and in bees
conditioned response not by direct pairing with the (Smith and Cobey, 1994). A critical question regard-
US but by pairing with another CS that has previously ing the mechanisms underlying blocking is, How does
been paired with the US. The training protocol for preconditioning of CS1 prevent or block the condi-
second-order conditioning proceeds in two phases. tioning of CS2, even though CS2 is paired with the
During Phase I, CS1 is paired with the US (resulting US? One study on eyeblink conditioning in rabbits
in CS1+). During Phase II, a novel CS2 is paired with has suggested that blocking relies on active inhibition
CS1+. The defining feature of second-order condi- of the US pathway (Kim et al., 1998). Eyeblink condi-
tioning is that a conditioned stimulus (CS1+) functioning relies on the cerebellum. It was shown that
tions as a reinforcing stimulus for the conditioning of complex spikes in Purkinje cells of the cerebellum
a second conditioned stimulus (CS2). (which represent activity in the inferior olive) are elic-
Second-order conditioning has been demonstrat- ited by the CS-US pairing in nave animals but not by
ed in various invertebrate (e.g., Menzel, 1981; Sahley CS-US presentations in trained animals. Since inferi-
et al., 1982) and vertebrate preparations (Rescorla, or olive activity seems to mediate the US, these
1980). It has also been observed in the gill-withdrawal studies suggest that the US pathway is actively inhibit-
reflex in a reduced preparation in Aplysia (Hawkins et ed during Phase II of a blocking protocol. The active
al., 1998). In this preparation the mantle organs are suppression of the US response seems to be mediated
isolated together with the abdominal ganglia. Tactile by the inhibition in the inferior olive produced by the
and electrical stimulation of the siphon and gill can cerebellar output (representing CS1+). Indeed,
be used as CSs and US. In this study extinction of CS1 blocking of this inhibition during Phase II of training
resulted in extinction of CS2, suggesting the forma- prevented blockingthat is, it allowed CS2 to under-
tion of a direct association between both conditioned go conditioning. This manipulation did not prevent
stimuli. Indeed an important question regarding the first-order conditioning to the CS1 in a separate
underlying mechanisms of second-order condition- group of control animals. These experiments suggest
ing is whether CS1+ essentially takes over the system that first-order conditioning not only results in the
CONDITIONING, CELLULAR AND NETWORK SCHEMES 99
production of a CR but also causes active inhibition Figure 1

of the US pathway; they also suggest that this nega-
tive-feedback loop is involved in blocking (see below).
Contingency
Rescorla (1968) suggested that it is not just the
number of CS-US pairings but also the correlation be-
tween the CS and the US, or the ability of the CS to
predict the US, that determines the occurrence of
conditioning. He demonstrated that if one group of
animals was presented with ten CS-US pairings, and
one group was trained with ten extra USs in addition
to the ten CS-US pairings (i.e., the probability of a US
given a CS equals .5), the latter group displayed less
conditioning. Despite both groups having received an
equal number of CS-US pairings, the group that re-
ceived the additional US presentations exhibited less
conditioning.
In most invertebrate (Abramson and Bitterman,
1986; Farley, 1987) and vertebrate preparations in
Schematic circuit of a model which exhibits second-order
which it has been examined, contingency has been
conditioning and blocking. Conditioned stimuli (CS) activate
observed. However, it is important to distinguish be-
sensory neurons (SN). In the nave state the SNs make weak
tween the different protocols used to study contingen- synapses unto both the motor neuron (MN) and the facilitatory
cy. The extra US presentations can take place before neuron (FN). These synapses can be potentiated by activity-
(US preexposure), interspersed with, or after (US dependent plasticity: Synapses undergo potentiation if they are
postexposure) the CS-US pairings. In Aplysia it has coactive with the (FN). A key element of the model is that
been shown that extra US presentations interspersed plasticity is occurring in parallel at two sites: SNMN and
with training decreases conditioning (Hawkins et al., SNFN. It is this feature that allows a previously conditioned
1986). Understanding the mechanisms underlying stimulus to function as a US.
contingency requires a determination of whether
both US pre- and postexposures decrease condition-
ing. This is because both US pre-and postexposure ef- Consider two CS pathways, CS1 and CS2, each
fects are difficult to explain with the same mecha- represented by a sensory neuron (SN1 and SN2), a
nism. For example, Hawkins and Kandel (1984) US pathway represented modulatory or facilitatory
suggested that contingency could result from habitua- neuron (FN), and a single output unit that generates
tion of the US pathway; this phenomenon, however, both the CR and UR, represented by a motor neuron
would not account for postexposure effects. Thus, it (MN). An important aspect of the network is that the
is possible that what is often thought of as a single CS neurons not only connect to the MN but also to the
higher-order form of conditioning may actually re- FN (see Figure 1). During first-order conditioning in
flect multiple mechanisms. which the SN1 and FN are paired (for example with
a 100-ms interval), the facilitatory neuron will pro-
duce associative plasticity at both the SN1MN and
Theoretical Models of Higher-Order Forms SN1FN synapse.
of Conditioning
Some researchers have proposed numerous
mechanistic models that address both the cellular and Second-Order Conditioning
network mechanisms underlying higher-order forms It is easy to see how this circuit is intrinsically able
of conditioning (Hawkins and Kandel, 1984; Gluck to account for second-order conditioning. Since plas-
and Thompson, 1987; Buonomano et al., 1990). ticity is also occurring at the SN1FN synapse, CS1
Many of these models share common conceptual fea- can take over the FN and become a US. When CS2
tures that we will examine with respect to a model of and CS1 are paired, CS1 will activate the FN and re-
second-order conditioning and blocking originally sult in conditioning of CS2. In this model, second-
developed in the context of the Aplysia siphon- order conditioning results directly from associative
withdrawal reflex (Buonomano et al., 1990). synaptic plasticity occurring at two synapses. Plasticity
100 CONDITIONING, CLASSICAL AND INSTRUMENTAL
at the SN1MN synapse is responsible for the gener- complex forms of learning may differ from simple
ation of the conditioned response, whereas plasticity forms of learning in either the form(s) of cellular plas-
at the SN1FN synapses accounts for second-order ticity involved and/or in the circuitry. However, the
conditioning. One important prediction of this type studies described above indicate that when the same
of model is that it should be possible to block second- forms of synaptic plasticity responsible for simpler
order conditioning without interfering with first- forms of associative learning are embedded in more
order conditioning by specifically blocking plasticity complex networks, more complex forms of learning
at the SN1FN synapse. may emerge, including blocking and second-order
conditioning.
Blocking See also: APLYSIA: CLASSICAL CONDITIONING AND

OPERANT CONDITIONING; INVERTEBRATE
Like second-order conditioning, blocking could
LEARNING: ASSOCIATIVE LEARNING AND
arise from the ability of a previously conditioned MEMORY PROCESSING IN BEES; INVERTEBRATE
CS1+ (SN1+) to take control of the FN. As in sec- LEARNING: ASSOCIATIVE LEARNING IN LIMAX
ond-order conditioning, during Phase I of training,
the synaptic strength of SN1 would increase and be- Bibliography
come strong enough to activate the FN. There are two Buonomano, D. V., Baxter, D. A., and Byrne, J. H. (1990). Small
consequences of this CS1-induced activity in the FN networks of empirically derived adaptive elements simulate
higher-order features of classical conditioning. Neural Net-
that would contribute to blocking. Note that if SN1+
works 3, 507523.
effectively activates the FN (and activity in the FN un- Gewirtz, J. C., and Davis, M., (1997) Second-order fear condition-
dergoes rapid depression), during the Phase II of ing prevented by blocking NMDA receptors in amygdala. Na-
blocking in which CS1+/CS2 are paired with the US, ture 368, 471474.
two important things happen: First, the FN will be ac- Gluck, M. A., and Thompson, R. F. (1987). Modeling the neural
substrates of associative learning and memory, A computa-
tivated approximately simultaneously with the onset
tional approach. Psychological Review 94, 176191.
of CS1+/CS2, essentially resulting in a 0-ms intersti- Hawkins, R. D., Carew, T. J., and Kandel, E. R. (1986). Effects of
mulus interval that should not result in conditioning interstimulus interval and contingency on classical condition-
of CS2. Second, since the FN undergoes depression, ing of the Aplysia siphon withdrawal reflex. Journal of Neuro-
the output of the FN in response to a US that followed science 6, 1,6951,701.
Hawkins, R. D., Cohen, T. E., Greene, W., and Kandel, E. R.
CS1/CS2 would be small or zero and would not sup-
(1998). Classical conditioning, differential conditioning, and
port associative conditioning of SN2. One subtle as- second-order conditioning in the Aplysia gill-withdrawal re-
pect of blocking pertains to the question of extinction flex in a simplified mangle organ preparation. Behavioral
of CS1+: if CS2 does not undergo conditioning be- Neuroscience 112, 636645.
cause the US in ineffective, why doesnt CS1+ under- Hawkins, R. D., and Kandel, E. R. (1984). Is there a cell-biological
alphabet for simple forms of learning? Psychological Review 91,
go extinction? Plasticity in the Aplysia sensory neurons
376391.
predicts that the effective interstimulus interval func- Kim, J. J., Krupa, D. J., and Thompson, R. F. (1988). Inhibitory
tion of plasticity will be state-dependent. That is, cerebello-olivary projections and blocking effect in classical
while a nonconditioned SN may not exhibit plasticity conditioning. Science 279, 570573.
with a 0-ms ISI, a conditioned SN can exhibit plastici- Rescorla, R. A. (1968). Probability of shock in the presence and ab-
sence of CS in fear conditioning. Journal of Comparative and
ty at 0 ms and thus prevent extinction of CS1+ during
Physiological Psychology 66, 15.
Phase II of blocking. (1980). Pavlovian second-order conditioning, studies in associa-
Therefore, in this model, although CS2 would be tive learning. New York: Erlbaum.
Smith, B. H., and Boey, S. (1994). The olfactory memory of the
paired with the US during Phase II of training, prior
honeybee Apis mellifera II Blocking between odorants in bi-
conditioning of CS1 would block associative enhance- nary mixtures. Journal of Experimental Biology 195, 91108.
ment of SN2. Thus, blocking could be supported by
Jennifer L. Raymond
activity-dependent neuromodulation (the mechanism
John H. Byrne
of classical conditioning) in combination with accom-
Revised by Dean V. Buonomano
modation or synaptic depression.
Conclusion
The results of both empirical and theoretical CONDITIONING, CLASSICAL AND
studies indicate that at least two factors determine the
ability of neural circuits to support various features of
INSTRUMENTAL
learning: the learning rules or forms of synaptic plas- Classical (Pavlovian) and instrumental (Thorndikian)
ticity and the connectivity of the circuitry in which conditioning are the two most widely employed para-
those forms of plasticity are embedded. Therefore, digms for studying simple, associative learning result-
CONDITIONING, CLASSICAL AND INSTRUMENTAL 101
ing from the organisms exposure to the temporal are obtained. Autoshaping consists of response-
conjunction of two or more events. The fully specified independent presentations of a lighted manipu-
classical conditioning paradigm consists of a set of op- landum (e.g., lighted key) as a CS and activation of a
erations involving an unconditioned stimulus (US) reli- food magazine as the US; the target response is con-
ably producing an unconditioned response (UR) and a tact with the manipulandum (e.g., key pecking). Key
conditioned stimulus (CS) initially shown not to produce pecking is not an instrumental response, nor is it a
a response resembling the UR. The CS and US are UR appearing in the constellation of URs to food in
then presented repeatedly to the organism in a speci- the mouth (Woodruff and Williams, 1976). Hence, ac-
fied order and temporal spacing, and a response sim- quisition of a response in an effector system not elicit-
ilar to the UR develops to the CS that is called the con- ed by the US qualifies autoshaping as a new associa-
ditioned response (CR); that is, CS-CR functions are tive learning paradigm.
obtained. Control over the temporal conjunction of Some discriminative approach procedures have
the CS, US, and UR makes classical conditioning been designated as Pavlovian simply because an ex-
preparations ideal vehicles for studying associative plicit cue (CS) is presented and food or water, desig-
learning because they can uniquely specify stimulus nated the US, is made available at a fixed time follow-
antecedents to the target response. Various temporal ing CS onset (e.g., Holland and Rescorla, 1975) and
arrangements of the CS and US give rise to different the approach behavior, by definition instrumental to
forms of classical conditioning (e.g., delay, trace, simul- receipt of the reinforcing event, has been erroneously
taneous). Classical conditioning is called classical re- designated a CR. At present, the preceding para-
ward conditioning if the US is a positive stimulus and digms are widely employed in the study of associative
classical defense conditioning if it is negative stimulus. learning, but whether they will converge with the em-
The positive or negative designation depends on the pirical laws of CS-CR paradigms has yet to be deter-
independent demonstration of the organisms per- mined systematically. The S-S and discriminative ap-
forming instrumental responses necessary to obtain proach paradigms lack the capability of CS-CR
the US or to remove itself from the US, respectively. paradigms to exercise absolute control over the tim-
What distinguishes classical from instrumental condi- ing and sequencing of stimulus events; and to identify
tioning is that presentation or omission of the US is the stimulus antecedents to the target response from
independent of CR occurrence; and the definition of the outset of training. In addition, with CS-IR and dis-
a CR is restricted to a target response selected from criminative approach procedures, the target response
among those effector systems elicited as URs by the is instrumentally conditioned. Consequently, these
US. Adherence to both components of the definition paradigms might be expected to be even less likely to
of classical conditioning avoids common confusions display convergence with the empirical laws of CS-CR
and ambiguities with other associative learning para- paradigms. In any event, despite the greater technical
digms commonly designated as classical condition- demands of measuring URs and CRs in CS-CR re-
ing. search, their methodological characteristics favor
their use in the study of associative learning.
Stimulus-Stimulus Paradigms
The designation classical conditioning has Biological Substrates
been applied to paradigms meeting only the require- The CS-CR paradigms are ideally suited for the
ment that CS and US be administered independent study of the biological substrates of associative learn-
of the target response, ignoring selection of the UR. ing because the target response is defined anatomi-
As a consequence, the term classical conditioning cally by a set of movements or secretions. The UCSs
has been extended from Russian physiologist Ivan Pe- elicitation of the UCR permits identification of the
trovich Pavlovs CS-CR to stimulusstimulus (S-S) target responses final common neural pathway(s)
conditioning paradigms involving principally condi- outside the conditioning situation and, thereby, af-
tioned stimulus-instrumental response (CS-IR) and fords the opportunity to observe changes in its activity
autoshaping procedures. The CS-IR paradigms in- from the start of conditioning (Thompson, 1976). In
clude conditioned suppression and other classical- contrast, S-S contingency and discriminative ap-
instrumental transfer procedures in which the stimulus- proach paradigms are inherently unsuitable for
stimulus pairings of classical conditioning are con- studying the biological basis of learning. First, in
ducted with a CS and a biologically significant event CS-IR paradigms, changes in the instrumental target
(e.g., shock) but without measurement of a UR or CR. response are not the consequences of its participation
The CS is then presented during ongoing instrumen- in the learning process. Rather, the changes are the
tal behavior and its facilitory or disruptive effect on result of interactions of hypothetical (unobserved)
responding is measured; therefore, CS-IR functions CRs with the CS that are governed by prior CS-US
102 CONDITIONING, CLASSICAL AND INSTRUMENTAL
pairings. As a consequence, any neural analysis of and US is necessary for CR acquisition; and respond-
learning that is directed at changes in the target re- ing produced by the unpaired control is nonassocia-
sponse is pointless. Second, since the target response tive, since the randomized sequencing of CSs and USs
in the discriminative approach paradigm is outcome- at exceedingly long, random intervals prevents any
defined, a wide variety of different body movements CS-US contiguity effects. However, associative theory
can yield the required outcome. Therefore, it is virtu- and its unpaired control methodology have been
ally impossible to identify a final common pathway for challenged by a contingency hypothesis (Prokasy, 1965;
the movements that make up the response. Rescorla, 1967), which asserts that associative learn-
ing in classical conditioning can be viewed as deter-
Control Methodology mined by the statistical relationship between the CS
The associative nature of a conditioning prepara- and US. The hypothesis assumes that if US probabili-
tion has come to be determined by the contiguous oc- ty is greater in the presence of the CS than in its ab-
currence of the CS and US and a set of control opera- sence, a positive contingency prevails and excitatory asso-
tions intended to estimate the contribution of other ciative effects would accrue to the CS; conversely, if
possible processes to responding. All response sys- US probability is higher in the absence than in the
tems show some level of baseline activity, often raised presence of the CS, the negative contingency would
by UCS presentations, which can produce an acciden- yield inhibitory associative effects. Moreover, the con-
tal coincidence of the CS and target response. More- tingency hypothesis assumes the unpaired controls
over, the likelihood of a target response to the CS perfectly negative contingency would lead the CS to
may be systematically affected by alpha responses, acquire inhibitory associative effects. Hence, Rescorla
which are URs to the CS in the same effector system (1967) proposed a truly random control to provide an
as the target response; and pseudo-conditioned and sen- associatively neutral condition for assessing excitatory
sitized responses established on the basis of prior US- and inhibitory conditioning.
alone presentations. A detailing of the latency, dura-
tion, amplitude, and course of habituation of the Rescorla (1967) specified the truly random con-
alpha response with a control group given CS-alone trol as involving independent programming of the CS
presentations can provide a basis for eliminating al- and US or equal US probabilities in the presence and
phas from consideration as a CR, since they are usual- absence of the CS. However, delineating pairing/
ly of a shorter latency than CRs. Hence, if a sufficient- unpairing cannot be determined a priori but only em-
ly long CS-US interval is employed, both alphas and pirically. CS-US pairing is specified by the CS-US in-
CRs can be observed in the interval and scored ac- tervals demonstrated to produce CR acquisition for a
cordingly (Gormezano, 1966; Gormezano et al., specific preparation, while explicitly unpaired de-
1983). notes the use of stimulus intervals outside the effec-
The reinstatement or augmentation of alphas to tive CS-US intervals. Consequently, in the absence of
the CS through US-alone or CS-US pairings is re- an empirically derived metric (i.e., effective CS-US
ferred to as sensitization. After eliminating alphas from conditioning intervals) to designate paired and un-
consideration, the contribution of pseudo-CRs to CR paired conditions, it is virtually impossible to pro-
measurement can be assessed by presentations of the gram an associatively neutral truly random control or
US one or more times prior to CS presentations. The predictable excitatory or inhibitory conditioning
procedure frequently results in responses to the CS, groups. Rescorla, seeking to validate the truly ran-
labeled pseudo-CRs, which are treated separately dom control, reported that its CS had no effect upon
from CRs because of their occurrence in the absence avoidance conditioning (Rescorla, 1966) or upon re-
of CS-US pairings. However, the US-alone procedure sponding in a CS-IR study where shock-US probabili-
precludes trial-by-trial assessment of pseudo-CRs for ty in the presence and absence of the CS were equal
comparison with CRs. Accordingly, a single unpaired (Rescorla, 1968). However, these findings were chal-
control procedure has evolved in which CS-alone and lenged by CS-IR studies revealing that trial number
US-alone trials are presented randomly the same and frequency of chance CS-US pairings under a truly
number of times as the paired CS-US group, but at random control could substantially affect (excitatory)
variable CS-US intervals exceeding those effective for conditioning (Gormezano and Kehoe, 1975). Subse-
CR acquisition. Under the unpaired control, re- quently, Rescorla (1972) disavowed the contingency
sponses on CS trials (excluding alphas) provide a hypothesis and truly random control and reverted to
summative measure of pseudo-CRs and baseline re- the use of the unpaired control. Nevertheless, the
sponses. truly random control is still widely employed despite
Use of the unpaired control is based on associa- the detailing of additional methodological limitations
tive assumptions that temporal contiguity of the CS (Papini and Bitterman, 1990; Wasserman, 1989).
CONDITIONING, CLASSICAL AND INSTRUMENTAL 103
Instrumental Conditioning when the experimental subject performs the target

response, the contingent event is received by both
Instrumental conditioning procedures are all
subjects. Therefore, both members of the pair receive
characterized by a contingent relationship between
the same number and temporal distribution of stimu-
the organisms response and a stimulus. Typically, if
lus events; the only difference is that the experimen-
the stimulus increases, decreases, or leaves unaffected
tal but not the control subject always receives the rein-
the probability of the response, it is identified as posi-
forcing event after execution of the target response,
tive, negative, or neutral, respectively. Although such
while the yoked partner receives the reinforcing event
labeling appears to be circular, Edward Thorndikes
independent of execution of the response.
(1913) characterizations of stimuli as satisfying
(positive) and annoying (negative) were not circular Thus, the yoked-control design appears to be ad-
because they were specified by behavior changes in- mirably suited to test the (null) hypothesis that the
dependent of the target response. Noncircularity can temporal relationship between the response and sub-
also be achieved by demonstrating transitivity of stim- sequent stimulus event is irrelevant to the observed
ulus effects on the target response to other (new) re- behavior change. Unfortunately, the design con-
sponses. founds within-subject sources of random error with
the treatment effect: Control of stimulus events by ex-
A positive or negative contingency between the perimental subjects can allow for systematic differ-
target response and reinforcing stimulus gives rise to ences in the number of experimental subjects that are
a variety of instrumental conditioning paradigms. more affected by the stimulus event than their yoked
The five most extensively studied are reward, punish- partners. The possibility of such confounding has
ment, omission, escape, and avoidance, which derive from rendered the results of yoked-control designs neces-
responses producing a positive (reward) or a negative sarily ambiguous. As a consequence, a means for as-
(punishment) stimulus; preventing a positive (omis- sessing the contribution of the third nonassociative
sion) or negative (avoidance) stimulus from occur- factor to instrumental conditioning has not yet been
ring; and terminating a negative stimulus (escape). achieved.
Woods (1974), employing a classification schema that
includes operant conditioning and presence or absence See also: OPERANT BEHAVIOR; PAVLOV, IVAN;
of a discriminative stimulus, enumerated sixteen in- THORNDIKE, EDWARD
strumental conditioning paradigms. However, de-
Bibliography
spite repeated attempts at conceptual clarification,
Coleman, S. R. (1981). Historical context and systematic functions
the operant remains devoid of causal stimulus ante- of the concept of the operant. Behaviorism 9, 207226.
cedents (Coleman, 1981) and, consequently, it cannot Gormezano, I. (1966). Classical conditioning. In J. B. Sidowski, ed.,
be employed to study associative learning. The oper- Experimental methods and instrumentation in psychology. New
ant is applicable only to the study of performance York: McGraw-Hill.
Gormezano, I., and Kehoe, E. J. (1975). Classical conditioning:
variables affecting postasymptotic or steady-state re-
Some methodological-conceptual issues. In W. K. Estes, ed.,
sponding. Aside from the study of discrimination Handbook of learning and cognitive processes, Vol. 2: Conditioning
learning processes, discriminative instrumental condi- and behavior theory. Hillsdale, NJ: Erlbaum.
tioning paradigms are widely employed to assess the Gormezano, I., Kehoe, E. J., and Marshall, B. S. (1983). Twenty
effects of concurrent classical conditioning of fear years of classical conditioning research with the rabbit. In J.
M. Sprague and A. N. Epstein, eds., Progress in psychobiology
or incentive motivation to the discriminative stimu-
and physiological psychology, Vol. 10. New York: Academic Press.
lus upon the instrumental target response. Holland, P. C., and Rescorla, R. A. (1975). Second-order condi-
tioning with food unconditioned stimulus. Journal of Compara-
tive and Physiological Psychology 88, 459467.
Control Methodology Papini, M. R., and Bitterman, M. E. (1990). The role of contingen-
cy in classical conditioning. Psychological Review 97, 396403.
Any occurrence of the target response without Prokasy, W. F. (1965). Classical eyelid conditioning. Experimenter
prior conjunction with the reinforcing stimulus is des- operations, task demands, and response shaping. In W. F.
ignated a nonassociative response attributable to 1. Prokasy, ed., Classical conditioning: A symposium. New York: Ap-
pleton-Century-Crofts.
base rate; 2. independent presentations of the rein-
Rescorla, R. A. (1966). Predictability and number of pairings in
forcing stimulus; and 3. presentations of the reinforc- Pavlovian fear conditioning. Psychonomic Science 4, 383384.
ing stimulus independent of the target response. Im- (1967). Pavlovian conditioning and its proper control pro-
plementing controls for the first two factors are self- cedures. Psychological Review 74, 7180.
evident, and achieving a control for the third factor (1968). Probability of shock in the presence and absence of
CS in fear conditioning. Journal of Comparative and Physiologi-
has been essentially limited to the yoked-control design.
cal Psychology 66, 15.
In the design, pairs of subjects are selected and one (1972). Informational variables in Pavlovian conditioning.
of them is randomly designated the experimental and In G. H. Bower and J. T. Spence, eds., Psychology of Learning
the other the control subject. During conditioning, and Motivation. New York: Academic Press.
104 CONSOLIDATION
Thompson, R. F. (1976). The search for the engram. American Psy- CONSOLIDATION
chologist 31, 209227.
Thorndike, E. L. (1913). Educational psychology. New York: Teach- See: AMNESIA, ORGANIC; ELECTROCONVULSIVE
ers College, Columbia University. THERAPY AND MEMORY LOSS; KAMINS
Wasserman, E. A. (1989). Pavlovian conditioning: Is contiguity ir- BLOCKING EFFECT: NEURONAL SUBSTRATES;
relevant? American Psychologist 44, 1,5501,551. MEMORY CONSOLIDATION: MOLECULAR AND
Woodruff, G., and Williams, D. R. (1976). The associative relation CELLULAR PROCESSES; MEMORY
underlying autoshaping in the pigeon. Journal of the Experi-
CONSOLIDATION: PROLONGED PROCESS OF
mental Analysis of Behavior 26, 113.
REORGANIZATION
Woods, P. J. (1974). A taxonomy of instrumental conditioning.
American Psychologist 29, 584597.
I. Gormezano
D
DECLARATIVE MEMORY experience. It also supports representations of rela-
tionships among various events, providing the larger
Memory is the process or processes by which the brain record of ones experience. It provides the critical
enables us to represent experience and permits expe- means for rapidly representing events, those one-
rience to shape us. Rather than a unitary capacity sup- time arbitrary or accidental concurrences of people,
ported by a single set of processes, however, there are places, and things, and the spatial, temporal, and in-
different forms of memory, supported by multiple, teractional relations among them.
functionally, and anatomically distinct memory sys-
tems. The form of memory upon which we seem to Moreover, declarative memory enables one to
depend most in the activities of everyday life and learn arbitrary, nonderivable associations through
about which we can most readily reflect is declarative experiencefor example, learning the names con-
memory. nected with peoples faces, or their addresses and
telephone numbers. Declarative memory thereby
provides for representations of relations beyond the
Declarative and Procedural Memory
province of events, encompassing the relations
There are various proposed taxonomies of mem- among the facts that constitute our knowledge of the
ory, each offering a different account of the divisions world. This point leads to a further critical distinc-
among the memory systems of the brain. Most such tion: between episodic memory, which contains auto-
accounts distinguish between declarative memory biographical records of personally experienced
and procedural memory (Cohen and Squire, 1980; events, and semantic memory, consisting of world
Cohen, 1984). Declarative memory supports the on- knowledge stored outside of personal contexts (Tulv-
demand accumulation, storage, and retrieval of new ing, 1972). As fundamentally relational, capturing the
data about facts and eventsthe information that we relations among many different elements of knowl-
capture from our experiences through our represen- edge, both episodic and semantic memory are sup-
tations of it. In contrast, procedural memory supports ported by the declarative memory system.
the shaping of behavioral repertoires acquired
through experience. Declarative memory differs from A second critical property of declarative memory
procedural memory in being a relational memory sys- is representational flexibility (Cohen, 1984; Cohen
tem. and Eichenbaum, 1993). Declarative memories can
be activated by all manner of external sensory or even
purely internal inputs, regardless of the current con-
The Nature of Declarative Memory text. And they can be accessed by any number of dif-
Declarative memory supports representations of ferent brain processors, not only the ones involved in
relationships among the constituent elements of an initially acquiring the memories. Once accessed, they
105
106 DECLARATIVE MEMORY
can be manipulated and used flexibly to guide perfor- Schacter, 1987; Richardson-Klavehn and Bjork,
mance under an enormous range of testing condi- 1988), in which performance depends on using the
tions, including those differing significantly from the test item to permit conscious recollection of some spe-
circumstances of original learning. In this manner, cific prior learning experience and then inspecting
declarative memory serves as the relational database the contents. A successful outcome requires recall of
on which so much of cognitive processing and behav- the relation between the items to be tested and the
ioral performance depends. Among the brain systems study list or study experience.
that access and manipulate the declarative-memory
database are the frontal-lobe systems that support The deficit in amnesia is evident in all manner of
cognitively mediated and consciously aware process- relations, whether verbal or nonverbal, spatial or
es, including conscious introspection and verbal re- nonspatial, or episodic or semantic. As regards the
ports of the contents of ones memories. last, hippocampal amnesia typically affects both per-
sonal and public events (Sagar, Cohen, Corkin, and
Growdon, 1985; Zola-Morgan, Cohen, and Squire,
Deficit of Declarative Memory in Amnesia 1984); it includes not only autobiographical but also
world knowledge. One example is the profound defi-
Amnesia is a devastating memory impairment fol- cit shown by the patient H.M. in learning new vocabu-
lowing damage to the hippocampal system. Patients lary (word-definition relations) that has entered the
with hippocampal amnesia typically have a combina- language since the onset of his amnesia (Gabrieli,
tion of anterograde amnesia, an impairment in ac- Cohen, and Corkin, 1988).
quiring new memory, and retrograde amnesia, loss of
memories preceding the trauma. The deficits seem Despite profound and pervasive impairment of
confined to the domain of declarative memory memory, amnesic patients show impressive preserved
(Cohen and Squire, 1980; Cohen, 1984; Ryan, Alth- learning and memory abilities. Such patients can
off, Whitlow, and Cohen, 2000). Thus, amnesic pa- learn motor, perceptual, and cognitive skills even
tients show marked impairment in tasks or situations though they are unable to remember the experiences
in which performance depends on learning the rela- during which they acquired the skills. For example,
tions among items, especially items associated only amnesic patients were able to learn to read mirror-
arbitrarily or accidentally. For example, such patients reversed text in training extended over several days;
have great difficulty in remembering the events of and they retained the skill after three months despite
daily life. The amnesic patient H.M., who has been marked impairment in recollecting the training expe-
studied fifty years, since undergoing a surgical resec- riences or recognizing the words on which they were
tion of medial temporal lobe structures (Scoville and actually trained (Cohen and Squire, 1980). Preserved
Milner, 1957; Corkin, 1984), exhibits marked impair- memory is characteristic of performance that can be
ment on various tests of memory for public events based on tuning of skills in particular domains, built
that occurred after the onset of amnesia and can bare- up of incremental improvements in performance with
ly recall any personal events since the time of his sur- each exposure, and expressed in a repetition of the
gery (Sagar, Cohen, Corkin, and Growdon, 1985). original learning situationsuccessful performance
in this case does not require the flexible, relational
Formal laboratory testing confirms the deficit in representations of declarative memory (Cohen, 1984;
memory for relations. Paired-associate learning is es- Schacter, 1987; Gabrieli, 1998; Eichenbaum and
pecially useful in diagnosing amnesia; in this proce- Cohen, 2001).
dure, in which one must learn a set of arbitrarily
paired words, amnesic patients show severe impair-
ment, as they do on most list-based recall or recogni- Declarative Memory and Consciousness
tion-memory tasks, in which they are asked to commit
to memory a set of common words, faces, or visual ob- Declarative memory is critical for conscious intro-
jects presented in a study list and then to report (in spection and conscious recollection. But this system
recall tests) or to judge (in recognition tests) which does not mediate any particular aspect of conscious
items appeared on that particular study list. Because processing; rather, it provides the flexible access to
such common stimuli are familiar from a lifetime of information about relations among people, places,
previous experience, remembering of specific study objects, and actionsthe relational databasethat
items requires the linkage of their identity to this par- undergirds conscious recollection and introspective
ticular study list or learning experience, thereby call- reports. This view accounts for the amnesic deficits in
ing on declarative memory. memory for relations, even those do not enter into
the conscious awareness of normal subjects (Ryan, Al-
Amnesic patients are usually impaired on explicit thoff, Whitlow, and Cohen, 2000; Chun and Phelps,
or direct tests of memory (Graf and Schacter, 1985; 1999). It also underscores the affinities between
DECLARATIVE MEMORY 107
human and animal models of amnesia. Hippocampal memories are then flexibly accessible to various corti-
amnesia in rodents and nonhuman primates pro- cal processors in supporting cognitive processing and
duces the same dissociation among memory capaci- behavioral performance.
ties that are typical of human amnesia. Such animals
show impairments in learning and remembering spa- See also: AMNESIA, FUNCTIONAL; AMNESIA,
tial relations among environmental cues, configura- INFANTILE; AMNESIA, ORGANIC; AMNESIA,
tions of multiple perceptually independent cues, con- TRANSIENT GLOBAL; EPISODIC MEMORY; GUIDE
textual or conditional relations, and comparisons TO THE ANATOMY OF THE BRAIN; SEMANTIC
among temporally discontinuous eventsall of which MEMORY: COGNITIVE ASPECTS; SEMANTIC
require a relational form of memory. Yet the same MEMORY: NEUROBIOLOGICAL PERSPECTIVE
animals can show normal learning and remembering
of a large variety of conditioning, discrimination, and Bibliography
skill tasks, none of which require a relational form of Chun, M. M., and Phelps, E. A. (1999). Memory deficits for implicit
memory but rather only gradual, incremental contextual information in amnesic subjects with hippocampal
changes in bias or reactivity to individual items with damage. Nature Neuroscience 2, 844847.
Cohen, N. J. (1984). Preserved learning capacity in amnesia: Evi-
repeated exposure. dence for multiple memory systems. In N. Butters and L. R.
Squire, eds., The neuropsychology of memory. New York: Guilford
Press, pp. 83103.
Brain Mechanisms of Declarative Memory Cohen, N. J., and Eichenbaum, H. (1993). Memory, amnesia and the
hippocampal system. Cambridge, MA: MIT Press.
The critical role of the hippocampal system in
Cohen, N. J., Ryan, J., Hunt, C., Romine, L., Wszalek, T., and
declarative memory is evident in the phenomenology Nash, C. (1999). Hippocampal system and declarative memo-
of amnesia. Neurophysiological and neuroimaging ry (relational) memory: Summarizing the data from function-
studies of the hippocampal system also demonstrate al neuroimaging studies. Hippocampus 9, 8398.
its association with memory for relations. Hippocam- Cohen, N. J., and Squire, L. R. (1980). Preserved learning and re-
tention of a pattern-analyzing skill in amnesia: Dissociation of
pal neurons encode various relationships among sig-
knowing how and knowing that. Science 210, 207210.
nificant elements of an experience, firing preferen- Corkin, S. (1984). Lasting consequences of bilateral medial tempo-
tially for particular conjunctions of the elements in ral lobectomy: Clinical course and experimental findings in
studies of freely behaving rodents (Wood, Dudc- H.M. Seminars in Neurology 4, 249259.
henko, and Eichenbaum, 1999; Eichenbaum et al., Eichenbaum, H., and Cohen, N. J. (2001). From conditioning to con-
scious recollection: Memory systems of the brain. NY: Oxford Uni-
2000). In functional neuroimaging studies of hu-
versity Press.
mans, hippocampal system activation arises whenever Eichenbaum, H., Dudchenko, P., Wood, E., Shapiro, M., and
the task engages memory for the relations among Tanila, H. (1999). The hippocampus, memory, and place
items (Henke, Buck, Weber, and Wieser, 1997; Cohen cells: Is it spatial memory or a memory space? Neuron 23,
et al., 1999). 209226.
Gabrieli, J. D. E. (1998). Cognitive neuroscience of human memo-
Amnesia indicates that the hippocampal system ry. Annual Review of Psychology 49, 87115.
must interact with other brain systems to effect declar- Gabrieli, J. D. E., Cohen, N. J., and Corkin, S. (1988). The im-
ative memory. Retrograde amnesia in cases of hippo- paired learning of semantic knowledge following bilateral
medial temporal-lobe resection. Brain and Cognition, 7, 157
campal amnesia can extend backward over variable
177.
lengths of time, but it is never total; the storage of Graf, P., and Schacter, D. L. (1985). Implicit and explicit memory
long-term memory is never completely lost. Hence for new associations in normal and amnesic subjects. Journal
the hippocampal system cannot be the repository, or of Experimental Psychology: Learning, Memory, and Cognition 11,
permanent storage site, of all long-term memory. In- 501518.
Henke, K., Buck, A., Weber, B., and Wieser, H. G. (1997). Human
stead, the reciprocal connections of the hippocampal
hippocampus establishes associations in memory. Hippocam-
system with all the higher-order cortical processors pus 7, 249256.
allow it to mediate storage in interaction with neocor- Richardson-Klavehn, A., and Bjork, R. A. (1988). Measures of
tical sites. After the various cortical processors identi- memory. Annual Review of Psychology 39, 475543.
fy the constituent elements of the event or experi- Ryan, J. D., Althoff, R. R., Whitlow, S., and Cohen, N. J. (2000).
Amnesia is a deficit in relational memory. Psychological Science
ence, the hippocampal system binds together the
11, 454461.
multiple elements into long-term declarative memory Sagar, H. J., Cohen, N. J., Corkin, S., and Growdon, J. M. (1985).
representations that capture the relations among the Dissociations among processes in remote memory. Annals of
elements, with the individual elements or attributes the New York Academy of Sciences 444, 533535.
represented in distributed fashion in the relevant cor- Schacter, D. L. (1987). Implicit memory: History and current sta-
tus. Journal of Experimental Psychology: Learning, Memory, and
tical processors. Thus, the interaction of the hippo-
Cognition 13, 501518.
campal system with neocortical processors and stor- Scoville, W. B., and Milner, B. (1957). Loss of recent memory after
age sites mediates the relational memory binding that bilateral hippocampal lesions. Journal of Neurology Neurosur-
allows the formation of declarative memory. Such gery and Psychiatry 20, 1121.
108 DJ VU
Tulving, E. (1972). Episodic and semantic memory. In E. Tulving Explanations of Dj Vu

and W. Donaldson, eds., Organization of Memory. New York:
Academic Press, 382403. Interpretations of dj have varied from the
Wood, E. R., Dudchenko, P. A., and Eichenbaum, H. (1999). The parapsychological (a telepathic reversion to a previ-
global record of memory in hippocampal neuronal activity. ous lifetime) to the psychodynamic (the mind neutral-
Nature 397, 613616. izes an emotional situation by displacing it into the
Zola-Morgan, S., Cohen, N. J., and Squire, L. R. (1983). Recall of
remote episodic memory in amnesia. Neuropsychologia 21,
past). These theses are thoroughly discussed by Sno
487500. and Linszen (1990) and Neppe (1983). We will exam-
ine several possible scientific interpretations.
Neal J. Cohen
Neurological Explanations
For more than a century, researchers have sug-
gested that a dj vu experience reflects a neurologi-
cal dysfunction. Since dj vu is part of the preseizure
DJ VU aura in some TLEs, it seems reasonable that dj vu
in nonepileptics may result from small temporal lobe
Dj vu is a memory phenomenon widely experi- seizures. Another interpretation is that the dj vu ex-
enced by the general public and is often cited in the perience results from a momentary delay in neuronal
popular literature. While the experience was de- transmission from the perceptual organ to the higher
scribed in the scientific literature as early as the mid- processing centers of the brain. This slight (several
1800s, researchers used a variety of different terms to millisecond) increase in the normal time taken to
describe the experience (e.g., paramnesia) until the transmit the message because of a synaptic dysfunc-
middle of the twentieth century. A standard definition may lead to a misinterpretation of the informa-
tion of dj vu used today was provided by Neppe tion as being old.
(1983, p. 3): any subjectively inappropriate impres-
Another neurological interpretation involves two
sion of familiarity of a present experience with an un-
pathways rather than one. In the visual system, infor-
defined past.
mation is received in the cortex first from the primary
Survey research indicates that between half to and then a secondary pathway. When the normally
two-thirds of individuals have had a dj vu experi- brief delay between the two tracks lengthens, the usu-
ence, although this estimate varies considerably (30 to ally seamless integration of the two messages is dis-
97 percent) across three dozen investigations. Of rupted and experienced as two separate messages. A
those who experience dj vu, nearly all have had variation on this position is that the primary rather
more than one, and most experience one or more than the secondary neuronal path is inordinately
each year. A dj vu experience is typically triggered slowed, resulting in a reversal of the normal sequence
by a visual scene, lasts for a few seconds, and is associ- of messages. We routinely interpret information from
ated with mild stress, anxiety, or fatigue. The primary the primary pathway as our initial perception, so
psychological reaction is surprise, and the time sense when the secondary pathway arrives slightly before
seems to slow down. the primary pathway, the information from the pri-
mary pathway feels familiar because a memory
The incidence of dj vu decreases systematically match already existsit was established moments be-
with age, and this is probably due to an increase in so- fore.
cietal awareness across recent years. For example,
Memory Explanations
Gallup and Newport (1991) found that belief in dj
vu nearly doubled between 1978 and 1990. There It is possible that the individual experienced the
may be a decrease in reports of dj vu with age be- present situation not directly but through a magazine,
cause older adults grew up in a time when belief in movie, TV show, or dream. Considerable research
dj vu was not as widely accepted as it is today. suggests that source confusions are routine and that
dj vu may reflect a match with a memory created
Dj vu appears to be more common in better- by media or imagination. Another memory perspec-
educated and better-traveled individuals; there is no tive is that the type of cognitive processing is being
evidence of gender differences. Dj vu also occurs in duplicated rather than the actual memory. Retrieval
the aura preceding a seizure in some temporal lobe success often depends upon the correspondence be-
epileptics (TLEs). This trend motivated research to tween the way information is processed during input
determine whether dj vu is a potential diagnostic and retrieval. If the cognitive processing of new infor-
tool in seizure activity, epilepsy, and brain pathology. mation follows a set of mental procedures similar to
Accumulating evidence, however, does not support those of a prior experience, an unexpected sense of
such speculation. familiarity may result.
DEMENTIA 109
Another memory perspective is that one aspect of coding and retrieval are simultaneously activated, this
the present setting is objectively familiar, but con- spurious familiarity results in dj vu.
scious recognition fails when the object appears in a Finally, we may have two different varieties of
different context. The familiarity elicited by the un- consciousness: One processes information from the
identified object is then overgeneralized to the entire environment, whereas the other monitors our inner,
setting. On a visit to a friends home for the first time, mental world. Dj vu may occur when normal con-
for example, the grandfather clock in the corner is sciousness is diminished by distraction, fatigue, or sei-
identical to one in your aunts home. The implicit fa- zure, forcing reliance on the internal consciousness
miliarity of this object is not connected to the old operating from internally generated images, result-
object but instead misattributed to the objectively new ing in a misreading of a new experience as old.
setting.
In summary, dj vu is experienced by most peo-
Finally, the present setting may evoke a sense of ple on an average of once per year. Its incidence de-
dj vu because of a familiar organization of the ele- creases with age, increases with education, and is
ments in a scene. It is not the grandfather clock in the more common under stress or fatigue. Likely expla-
living room of your friends new home that is familiar nations include neurological dysfunction (seizure,
but rather that the rooms layout: a sofa to the right synaptic slowdown), implicit familiarity of objects, di-
of the love seat, with a stairway to the left and a grand- vided attention followed by full perception, and/or al-
father clock against the back wall, the same as in your teration in the normal function of two cognitive pro-
aunts house. cesses.
Attentional Explanations Bibliography
An ongoing perceptual experience may occasion- Gallup, G. H., and Newport, F. (1991). Belief in paranormal phe-
ally be divided into separate perceptions through dis- nomena among adult Americans. The Skeptical Inquirer 15,
137146.
traction or inattention. In this explanation, first pro-
Mack, A., and Rock, I. (1998). Inattentional blindness. Cambridge,
posed by Titchener in 1924, a brief initial perception MA: MIT Press.
under diminished attention is followed immediately Neppe, V. M. (1983). The psychology of dj vu: Have I been here be-
by a second perception under full attention. The sec- fore? Johannesburg: Witwatersrand University Press.
ond impression matches the ignored first glance Sno, H. N., and Linszen, D. H. (1990). The dj vu experience: Re-
membrance of things past? American Journal of Psychiatry 147,
taken moments earlier, giving rise to the feeling that
1,5871,595.
the current experience duplicates something experi- Titchener, E. B. (1924). A beginners psychology. New York: Macmil-
enced before. As you walk into a new hotel lobby talk- lan
ing on your cell phone, you process the scene without
Alan S. Brown
full attention. When you hang up, you perceive the
situation with full awareness and get the feeling that
you have been here before. Recent research on inat-
tentional blindness by Mack and Rock (1998) suggests
that individuals often miss clearly visible objects if
DEMENTIA
they are focused on something else, even when the The term dementia describes a decline of previously
unattended object is directly in front of them. acquired intellectual, or cognitive, skills. Memory loss
is the primary symptom of dementia, but other cogni-
Dual-Processing Explanations tive symptoms exist as well. A diagnosis of dementia
Routine cognitive experience may often involve requires an impairment in memory and at least one
the operation of two separate but interactive process- other cognitive domain (American Psychiatric Associ-
es. While both processes usually operate in concert, ation, 1994). This can include an impairment in com-
occasionally they might shift out of synchrony, or one prehending or expressing language, sustaining atten-
process might occur in the absence of the other. For tion, orientation (knowing where one is and the date
example, retrieval and familiarity usually operate in and time), visual perception, visual construction, or
an independent but coordinated manner, with recall executive (planning and organizing) skills. In addi-
accompanied by familiarity. When retrieval is activat- tion to cognitive deficits, dementia is often associated
ed without familiarity, a familiar setting seems mo- with behavioral and/or psychiatric changes such as
mentarily unfamiliar (jamais vu). Conversely, when poor judgment (for example, spending money reck-
familiarity is activated in the absence of retrieval, dj lessly or dressing inappropriately), delusions (false
vu is the result. Another interpretation is that memo- beliefs), or hallucinations (seeing, feeling, or hearing
ry encoding and retrieval usually operate indepen- things that are not actually there). Disinhibition, ex-
dently of each other. We experience a new situation; emplified by the use of inappropriate language (i.e.,
then we encode it into memory. However, if both en- swearing), the telling of off-color jokes, or acting
110 DEMENTIA
overly familiar with others, can also occur. Mood Dementia, Delirium, Senility, or Mental
symptoms, such as depression, apathy, or increased Retardation?
irritability, can accompany dementia as well.
Dementia is diagnosed only in the absence of de-
Dementia is not a disease, but a syndrome that lirium. Like dementia, delirium is associated with
has many causes, including head injury, vitamin defi- memory loss, but it is also associated with disorienta-
ciency, hydrocephalus, epilepsy, depression, medication, confusion, and alterations in consciousness that
tion effects, and toxic exposure. In addition, there are fluctuate throughout the day. Unlike dementia, delir-
a number of diseases that can cause dementia. These ium typically begins suddenly and is usually the result
include Alzheimers disease, vascular disease, HIV in- of a treatable cause such as illness or the effects of
fection, multiple sclerosis, Parkinsons disease, Hun-
medications.
tingtons disease, Lewy body disease, and others. De-
pending on the cause of dementia, its onset may be Dementia and advancing age often co-occur be-
gradual or sudden, and its course may be progressive, cause the biggest risk factor for the development of
stable, or reversible. dementia is advancing age. Dementia is not, however,
an inevitable consequence of aging. The term senility
Pathophysiology of Dementia has been incorrectly applied to memory impairment
that accompanies old age. Senility actually refers only
The pathology associated with dementia varies,
to the age of onset of dementia; the term senile demen-
depending on the cause of dementia. The term neu-
rodegenerative is used to describe dementia caused by tia is sometimes used to refer to dementia that occurs
diseases that lead to the progressive death of nerve after age sixty-five. The term presenile dementia, in
cells (neurons) in the brain. Some diseases are associ- contrast, signifies dementia whose onset is prior to
ated with the development of abnormalities within or age sixty-five.
around neurons that interfere with communication Mental retardation differs from dementia in that
between brain cells and/or cause cell death. For exam- it describes intellectual and adaptive functioning that
ple, the amyloid plaques (insoluble deposits of a type
has developed subnormally. Thus, functioning has
of protein found between neurons) and neurofibril-
not declined from a previously higher level, as is the
lary tangles (abnormal collections of twisted threads
case with dementia. It is possible, however, for some-
found within neurons) associated with Alzheimers
one with mental retardation to acquire dementia. The
disease (AD) are found primarily in the cerebral cor-
most common example of this occurs in people with
tex (outer covering of the brain), hippocampus, and
Downs syndrome, most of whom will develop AD by
entorhinal cortex. Parkinsons disease, in contrast, is
the time they are in their fifties (Mann, 1993).
associated with Lewy bodies (accumulations of neuro-
filamentous material located within neurons) in the
substantia nigra and locus ceruleus (Kish, Shannak,
and Horneykiewicz, 1988). Pick bodies (a common
Memory Functioning in Dementia
neuropathological finding in frontotemporal demen- Due to similarities in the clinical presentation of
tia) are concentrated primarily in the frontal and/or patients with disorders affecting primarily subcortical
temporal cortex of the brain (McKhann et al., 2001). structures, which differs from that observed in de-
Cerebrovascular disease causes dementia by dis- mentia caused by cortical damage, the terms cortical
rupting blood flow to brain cells and is not considered dementia and subcortical dementia are often used. A cor-
to be neurodegenerative. If blood vessels in the brain tical pattern of cognitive deficits is characterized by
are blocked (e.g., by cholesterol), blood flow is con- an inability to perform purposeful motor movements
stricted. Such an event, called ischemia, results in oxy- (apraxia), language disturbance (aphasia), memory
gen deprivation, which can cause neuron death. Vas- impairment (amnesia), and difficulty with object rec-
cular dementia is the term used to refer to dementia ognition (agnosia). The most common disease caus-
caused by multiple vascular accidents, such as small ing cortical dementia is AD. Subcortical dementia, in
strokes (in which case it is sometimes referred to as contrast, has many more causes, such as Huntingtons
multiinfarct dementia), or a single stroke. Whether disease (HD) or Parkinsons disease (PD), and is char-
numerous small vessel changes can cause dementia is acterized by slowed mentation (bradyphrenia), for-
unclear, but with sufficient accumulation these getfulness, motor abnormalities, impaired planning
changes have been associated with significant cogni- and judgment (so-called executive dysfunction),
tive changes (Boone et al., 1992). There are, of and mood changes such as depression or apathy
course, many other mechanisms leading to death of (Cummings and Benson, 1984). Memory perfor-
brain tissue and dementia, but these examples ex- mance also differs between cortical and subcortical
plain some of the ways dementia can be caused. dementia (see Table 1).
DEMENTIA 111
Short-Term Memory Table 1

Short-term, or working, memory refers to informa-
tion that is stored for only a few seconds, such as when
one repeats a list of numbers immediately after hear-
ing them. This skill does not reliably differentiate pa-
tients with cortical from those with subcortical de-
mentia, as both have been shown to perform as well
as normal control subjects in repeating digits for-
wards, and deficits in their ability to repeat digits in
reverse order (Calderon et al., 2001; Redondo Verge,
Brown, and Chacon Pena., 2001). Patients with Lewy
body dementia, however, are impaired on both por-
tions of this task (Calderon et al., 2001).
Long-Term Memory
Long-term, or secondary, memory refers to memory
for information that must be recalled after some peri-
od of time, such as after few minutes or after many
years. Semantic memory, a type of long-term memory
relating to meanings of words and their relationships,
is typically impaired in patients with cortical demen- dementia is not improved. Patients with subcortical
tia. Perhaps the most impressive example of this defi- dementia, however, benefit from cues, and might
cit is observed in patients with fronto-temporal de- even perform as well as normal healthy adults. This
mentia (FTD), who can have very severe language difference is thought to demonstrate that cortical de-
impairment, such as word-finding difficulty, inability mentia causes an inability to encode new information,
to appreciate syntax, and/or impaired language com- whereas subcortical dementia causes a deficit in re-
prehension, with relative sparing of other cognitive trieving information. Tests of implicit (or procedural)
abilities (Mesulam, 2000). The word-finding difficulty memory (memory that occurs without conscious aware-
observed in AD is also an example of the impaired se- ness, such as the learning of motor movements) elicit
mantic memory seen among patients with cortical de- the opposite pattern. That is, subcortical dementia
mentia. Semantic memory is not typically impaired in
produces impairment on tests of implicit memory,
patients with subcortical dementia.
such as a mirror-tracing task, whereas patients with
Another type of long-term memory, episodic mem- cortical dementia improve as much as normal healthy
ory, concerns memory for events. Patients with AD are adults with practice, even if they have no recollection
typically severely impaired in their ability to recall re- of having performed the task previously (Butters,
cent events, but are much better able to recall events Heindel, and Salmon, 1990).
that occurred many years ago (Dorrego et al., 1999).
Patients with HD, however, have been found to
exhibit a flat gradient of remote memory, with Recent Developments in Dementia
equal impairment for all previous years (Albert, But- Research
ters, and Brandt, 1981). Whether this is the case for
patients with other types of subcortical dementia New Drug Treatments
(e.g., PD) has not been consistently shown (Fama et In 1996, the Food and Drug Administration ap-
al., 2000). proved the first drug (tacrine) to treat the memory
Long-term memory has been further divided into impairment that results from AD. The drug works by
explicit and implicit memory. Explicit memory (some- increasing the amount of a particular chemical, called
times used interchangeably with the term declarative a neurotransmitter, in the brain. This neurotransmit-
memory) is typically more severely impaired in cortical ter, called acetylcholine, is related to memory func-
than subcortical dementia. Both types of dementia tioning, and is depleted in the brains of people with
cause an impairment in this ability to recall informa- AD. Since approval of the first drug to treat AD, three
tion (e.g. a list of words, a story, or simple line draw- new drugs with similar mechanisms of action have
ings) that was presented earlier. Provided with cues, also been approved. Another development in clinical
however, such as the category from which the words trials research for memory disorders has been the in-
belong, or a recognition (e.g., multiple choice or yes/ vestigation of over-the-counter herbs such as Ginkgo
no) paradigm, the performance of those with cortical biloba and the antioxidant vitamin E.
112 DEMENTIA
Refining Diagnostic Criteria for Types of Beck, C., Cody, M., Souder, E., Zhang, M., and Small, G. W. (2000).
Dementia Dementia diagnostic guidelines: Methodologies, results, and
implementation costs. Journal of the American Geriatrics Society
Because dementia-causing disorders can be diffi- 48, 1,1951,203.
cult to diagnose during life, and because of new devel- Boone, K. B., Miller, B. L., Lesser, I. M., Mehringer, C. M., Hill-
opments in neuropathological techniques, criteria for Guttierrez, E., Goldberg, M. A., and Berman, N. G. (1992).
Neuropsychological correlates of white-matter lesions in
dementing disorders are evolving. To achieve consis-
healthy elderly subjects: A threshold effect. Archives of Neurolo-
tency, experts in these disorders meet to establish a gy 49, 549554.
consensus regarding the features that distinguish a Butters, N., Heindel, W. C., and Salmon, D. P. (1990). Dissociation
particular disorder from others. From 1984 to 1999, of implicit memory in dementia: Neurological implications.
fourteen consensus guidelines concerning dementing Bulletin of the Psychonomic Society 28, 359366.
Calderon, J., Perry, R. J., Erzinclioglu, S. W., Berrios, G. E., Den-
disorders were published (Beck et al., 2000). Exam-
ing, T. R., and Hodges, J. R. (2001). Perception, attention,
ples of these include Lewy body dementia (McKeith and working memory are disproportionately impaired in de-
et al., 1997) and FTD (McKhann et al., 2001). mentia with Lewy bodies compared with Alzheimers disease.
Journal of Neurology, Neurosurgery, and Psychiatry 70, 157164.
Mild Cognitive Impairment Cummings, J. L., and Benson, D. F. (1984). Subcortical dementia:
Sometimes patients do not meet criteria for de- Review of an emerging concept. Archives of Neurology 41, 874
mentia because they have a mild impairment in mem- 879.
Dorrego, M. F., Sabe, L., Cuerva, A. G., Kuzis, G., Tiberti, C., Bol-
ory but in no other cognitive domain, and their every- ler, F., and Starkstein, S. E. (1999). Remote memory in Al-
day functioning is not impaired. Scientists have zheimers disease. Journal of Neuropsychiatry and Clinical Neuro-
introduced the term mild cognitive impairment (MCI) to sciences 11, 490497.
describe this syndrome, and in 2001 consensus Fama, R., Sullivan, E. V., Shear, P. K., Stein, M., Yesavage, J. A.,
criteria for MCI were published (Petersen et al., Tinklenberg, J. R., Pfefferbaum, A. (2000). Extent, pattern,
and correlates of remote memory impairment in Alzheimers
2001). This disorder has received a great deal of at- disease and Parkinsons disease. Neuropsychology 14, 265276.
tention because, in many cases, it is considered to be Kish, S. J., Shannak, K., and Horneykiewicz, O. (1988). Uneven
a harbinger of AD (Morris et al., 2001). Because of the pattern of dopamine loss in the striatum of patients with idio-
availability of new drug treatments for the memory pathic Parkinsons disease: Pathophysiologic and clinical im-
impairment caused by AD, there is a great deal of in- plications. New England Journal of Medicine 318, 876880.
Mann, D. M. A. (1993). Association between Alzheimer disease and
terest in diagnosing this disease before it has prog- Down syndrome: Neuropathological observations. In J. M.
ressed to dementia. Berg, H. Karlinsky, and A. J. Holland, eds., Alzheimer disease,
Down syndrome and their relationship. Oxford: Oxford Universi-
ty Press.
Conclusion McKeith, I. G. et al. (1997). Dementia with Lewy bodies: Trying to
define a disease. Neurology 47, 1,1131,124.
Burgeoning interest in dementia research in the
McKhann, G. M., Albert, M. S., Grossman, M., Miller, B., Dickson,
late twentieth century has led to better understanding D., and Trojanowski, J. Q. (2001). Clinical and pathological
of its neuropathological underpinnings, treatments diagnosis of frontotemporal dementia. Report of the work
for its memory impairments, and refinements in the group on frontotemporal dementia and Picks disease. Ar-
diagnostic criteria for the disorders causing it. The chives of Neurology 58, 1,8031,809.
Mesulam, M. (2000). Aging, Alzheimers disease, and dementia:
importance of early diagnosis of neurodegenerative
Clinical and neurobiological perspectives. In Principles of Be-
disorders has led to an increasing efforts to identify havioral and Cognitive Neurology, 2nd edition. New York: Ox-
very mild, objective memory problems. Thus, charac- ford University Press.
terization of memory functioning continues to be an Morris, J. C., Storandt, M., Miller, J. P., McKell, D. W., Price, J. L.,
essential aspect for the diagnosis of dementia syn- Rubin, E. H., and Berg, L. (2001). Mild cognitive impairment
represents early-stage Alzheimer disease. Archives of Neurology
dromes.
58, 397405.
Petersen, R. C., Stevens, J. C., Banguli, M., Tangalos, E. G., Cum-
See also: ALZHEIMERS DISEASE: BEHAVIORAL mings, J. L., and DeKosky, S. (2001). Practice parameter:
ASPECTS; ALZHEIMERS DISEASE: HUMAN Early detection of dementia: Mild cognitive impairment (an
DISEASE AND THE GENETICALLY ENGINEERED evidenced-based review). Neurology 56, 1,1331,142.
ANIMAL MODELS; COGNITIVE ENHANCERS; Redondo Verge, L., Brown, R. G., and Chacon Pena, J. R. (2001).
PHARMACOLOGICAL TREATMENT OF MEMORY Executive dysfunction in Huntingtons disease. Reviews in
DEFICITS Neurology 32, 923929.
Jason Brandt
Bibliography Ralph H. B. Benedict
Albert, M. S., Butters, N., and Brandt, J. (1981). Patterns of remote Revised by Cynthia A. Munro
memory in amnesic and demented patients. Archives of Neurol-
ogy 38, 495500.
American Psychiatric Association (1994). Diagnostic and Statistical See also: APLYSIA: DEVELOPMENT OF PROCESSES
Manual of Mental Disorders, 4th edition. Washington, DC: UNDERLYING LEARNING; CHILDREN,
American Psychiatric Association. DEVELOPMENT OF MEMORY IN; EARLY
DISCRIMINATION AND GENERALIZATION 113
EXPERIENCE AND MEMORY; LANGUAGE dimension, with gradients forming an inverted V

LEARNING; SCHOOL LEARNING shape (Siegel, 1972). Latent inhibition refers to re-
tarded classical conditioning as a result of preexpo-
sure to the CS. Inverted-V-shape generalization gra-
dients have also been observed with tonal stimuli
DIENCEPHALON trained with Pavlovs conditioned inhibition proce-
See: AMNESIA, ORGANIC dure (Mis, Lumia, and Moore, 1972). Inverted-V gen-
eralization gradients have been observed in pigeon
operant tasks (Hearst, 1969).
DISCRIMINATION AND Paradigms for Occasion Setting

GENERALIZATION
Occasion setters are stimulus features, such as the
The decade of the 1990s witnessed acceleration in the presence of a light or tone, that serve as discrimina-
convergence of theoretical and experimental studies tive stimuli. For example, the presence of a light
of discrimination and generalization from the do- might signal that operant responses will be rein-
mains of classical conditioning and instrumental (op- forced. The absence of this light would signal that op-
erant) learning. Classical conditioning refers to the es- erant responses are not reinforced. In general, a
tablishment of behavioral adaptations (conditioned feature-positive paradigm is one in which the OS
responses; CRs) by the methods of Pavlov. Instrumen- signals reinforcement; a feature-negative paradigm is
tal learning is a general term for goal-seeking behav- one in which the OS signals the absence of reinforce-
ior, and operant conditioning refers to reinforcement ment. The presence of a light or tone might signal re-
learning by the methods of Skinner. The term discrim- inforcement, whereas its absence signals nonreinfor-
ination refers to the capacity of organisms to learn dif- cement. In classical conditioning a feature-positive
ferent modes of behavior depending on signals or occasion-setting task would involve adding a feature
cues from the environment about the imminence or to the CS. For example, if the CS is a tone, the addi-
accessibility of reinforcement. Generalization refers to tion of a light sets the occasion for reinforcement,
stimulus generalization, the capacity for signals or whereas the tone alone would not signal reinforce-
cues that are different from those used for establishment. A feature-negative task would be one in which
ing learned behavior to evoke this behavior. Stimulus the light, instead of signaling reinforcement when
generalization in classical conditioning refers to the ca- presented with the tone, would signal its absences.
pacity of a stimulus other than the conditioned stimu- If the occasion setter overlaps the CS and signals
lus to evoke a CR. In operant conditioning, one set the reinforcement, it can result in a more robust
of stimuli, an occasion setter (OS), might evoke the CR than would otherwise be the case. By contrast,
behavior controlled by another OS, depending on if OS overlaps the CS and signals the absence of
their shared features or similarity. Skinner coined the the US, it can inhibit the CR. This feature-negative
term occasion setter to refer to signals or cues that pre- discrimination recognizing the absence of an OS as a
dict reinforcement. In recent years, the term occasion signal of the absence of the US is a relatively difficult
setting has been extended to encompass both classical discrimination for animals to master. In fact, a classi-
and operant forms of behavioral learning. As a conse- cal conditioning situation like this is called a condi-
quence of this mixing, the terminology and para- tioned inhibition paradigm, and research has shown that
digms used in occasion setting research borrow from this is relatively difficult discrimination for animals to
the two domains. The mixture of the two domains has master.
led to a healthy integration of methods and ideas
(Schmajuk and Holland, 1998).
The convergence of ideas about discriminations Pattern Learning
and generalization from classical and operant condi- The relative difficulty in learning feature-
tioning began during the late 1960s, when the princi- negative discrimination as compared with feature-
ples of stimulus control enunciated by operant- positive discrimination learning extends also to the
conditioning studies involving pigeons were found to paradigms of positive and negative patterning. Dem-
extend to eyeblink conditioning in rabbits (Moore, onstrations of patterning in both classical and oper-
1972). Specifically, generalization along an auditory ant conditioning tasks involve two signaling stimuli,
frequency dimension shares many of the same charac- A and B, which could be occasion setters, and three
teristics as visual wavelength generalization in pi- types of trials. In positive patterning, the trial types
geons. Conditioned stimulus preexposure (latent in- are AB+, A, and B. Animals readily learn to re-
hibition) also generalizes along an auditory frequency spond more vigorously to the reinforced stimulus
114 DISCRIMINATION AND GENERALIZATION
Figure 1 Perceptual Learning

Exposure to discriminative stimuli enhances sub-
sequent discrimination learning, an effect known as
perceptual learning (Hall, 1991). The phenomenon of
perceptual learning might seem to contradict the idea
that exposure to stimuli retards later learning, as in
latent inhibition. But latent inhibition can actually ac-
count for perceptual learning, if we assume that pre-
exposed discriminatory stimuli share features in com-
mon. Latent inhibition develops to shared and
unshared features alike, but the common features lose
associability more rapidly than the unshared features.
Animals learn to ignore the shared features, allowing
the unshared features to dominate subsequent dis-
crimination learning and thereby facilitating perfor-
mance (McLaren and Mackintosh, 2000).
Not only do the shared features undergo more la-
tent inhibition than the unshared features, the un-
shared features also can become mutually inhibitory
through mechanisms of Pavlovian conditioned inhibi-
tion, provided the two stimuli are alternated during
the preexposure phase (Dwyer, Bennett, and Mackin-
tosh, 2001). The unique aspect of one preexposed
stimulus can actively suppress associations with the
Percentage of conditioned eyeblink responses in rabbits to the unique aspect of the other preexposed stimulus, and
target CS (X) in the presence of feature positive (A) and feature vice versa. Thus, any tendency for one stimulus to
negative (B) occasion setters as a function of the AX and BX elicit a representation of the other is reduced, and
intervals (0, 5, 15, 45 seconds) used in training. this promotes perceptual learning.
Acquired Distinctiveness
compound, AB, and to inhibit responses to A and Acquired distinctiveness refers to enhanced discrim-
B, where (+) denotes a reinforced trial and () de- ination learning to stimuli or stimulus dimensions
notes a nonreinforced trial. In negative patterning, that had been used in a prior discrimination task. Un-
the trial types are AB, A+, and B+. In order to be- like perceptual learning, however, acquired distinc-
have appropriately in a negative-patterning task, the tiveness appears to entail more than latent inhibition
animal must somehow learn that responding to the of shared stimulus features and mutual inhibition of
stimulus compound AB must be suppressed. This can unique features. The additional ingredient is correla-
happen only if the compound stimulus has unique tion with reinforcement. The nature of this correla-
features that are subsets of neither A nor B (Pearce tion remains in question. George and Pearce (1999)
and Bouton, 2001). That is, the compound has a have argued that acquired distinctiveness stems from
configural component that is shared by neither A nor prior learning about the relevance of the stimuli for
B. solving problems based on the same class of reinforc-
ers. The relevance-to-reinforcement account of ac-
The general difficulty in learning a negative- quired distinctiveness does not imply that prior learn-
patterning discrimination suggests that this configu- ing about the lack of correlation of the stimuli for
ral component is often overshadowed by the noncon- reinforcement impedes discrimination learning.
figural aspects of A and B, each of which in isolation
drives a tendency to respond by virtue of their associ-
ation with reinforcement. At the same time, some Time Discrimination
negative-patterning tasks can be mastered compara- Temporal control of behavior is one of the signa-
tively easily, suggesting that the compound AB forms ture features of operant conditioning methods, and
a unique pattern that, although similar to A and B, instances of temporal discrimination and generaliza-
nevertheless is treated as a whole, thereby allowing tion for duration has been well documented in oper-
the animal to master the discrimination. ant conditioning tasks in animals and humans (e.g.,
DISTRIBUTED PRACTICE EFFECTS 115
Wearden and Bray, 2001). Similar instances of tem- Schmajuk, N. A., and Holland, P. C., eds. (1998). Occasion setting.
poral control of behavior occur in classical condition- Washington, DC: American Psychological Association.
Siegel, S. (1972). Latent inhibition and eyelid conditioning. In A.
ing (Gallistel and Gibbon, 2000; Kehoe and Macrae, H. Black and W. P. Prokasy, eds., Classical conditioning II: Cur-
2002). Conditioned-response timing depends on the rent research and theory, pp. 231247. New York: Appleton-
CS-US interval(s) used in training. Conditioned re- Century-Crofts.
sponses are timed so that they achieve maximal am- Weardon, J. H., and Bray, S. (2001). Scalar timing without refer-
plitude at the anticipated time(s) of the US, and this ence memory? Episodic temporal generalization and bisec-
tion in humans. Quarterly Journal of Experimental Psychology
occurs whether the CS is defined operationally as the 54B, 289309.
onset or the offset of a stimulus (e.g., Kehoe, White, N. E., Kehoe, E. J., Choi, J-S, and Moore, J. W. (2000). Coef-
Schreurs, Macrae, and Gormezano, 1995). In classical ficient of variation in timing of the classically conditioned eye-
conditioning the timing of conditioned responses be- blink in rabbits. Animal Learning and Behavior 28, 520524.
comes more variable as the CS-US interval increases David R. Thomas
(Gallistel and Gibbon, 2000; White, Kehoe, Choi, and Revised by John W. Moore
Moore, 2000). Temporal discriminative control of
classically conditioned responses also occur in occa-
sion setting (Kehoe, Palmer, Weiderman, and
Macrae, 2000; see Figure 1). DISTRIBUTED PRACTICE EFFECTS
See also: CLASSICAL CONDITIONING: BEHAVIORAL Learning and memory are generally improved by
PHENOMENA repetition. However, not all repetitions are equally
beneficial. The effectiveness of repetitions depends in
Bibliography
part on their temporal distribution. A piece of infor-
Dwyer, D. W., Bennett, C. H., and Mackintosh, N. J. (2001). Evi-
dence for inhibitory associations between the unique elements mation studied on several occasions widely spaced
of two compound flavors. Quarterly Journal of Experimental Psy- apart in time will be remembered better than a simi-
chology 54B, 97107. lar fact studied on several occasions close in time.
Gallistel, C. R., and Gibbon, J. (2000). Time, rate, and condition-
ing. Psychological Review 107, 289344. The advantage of distributed repetitions over
George, D. N., and Pearce, J. M. (1999). Acquired distinctiveness spaced repetitions has long been known. Hermann
is controlled by stimulus relevance and not correlation with Ebbinghaus discussed distributed practice effects in
reward. Journal of Experimental Psychology: Animal Behavior Pro- his classic 1885 monograph on memory. He noted
cesses 25, 363373. that with any considerable number of repetitions a
Hall, G. (1991). Perceptual and associative learning. Oxford: Claren-
don Press. suitable distribution of them over a space of time is
Hearst, E. (1969). Excitation, inhibition, and discrimination learn- decidedly more advantageous than the massing of
ing. In N. J. Mackintosh and W. K. Honig, eds., Fundamental them at a single time (p. 89). Similarly, Jost (1897)
issues in associative learning, pp. 141. Halifax: Dalhousie Uni- formulated the advantage of distributed over massed
versity Press. repetitions as one of his laws of memory. In subse-
Kehoe, E. J., and Macrae, M. (2002). Fundamental behavioral and
findings in classical conditioning. In J. W. Moore, ed., A
quent decades, distribution of repetition became an
neuroscientists guide to classical conditioning, pp. 171231. New important manipulation in the study of learning and
York: Springer. memory. Because many different procedures were
Kehoe, E. J., Palmer, N., Weiderman, G., and Macrae, M. (2000). used and many conflicting results were found, the
The effect of feature target intervals in conditioned discrimi- overall pattern was long unclear, and investigators,
nations on acquisition and expression of conditioned nictitat-
ing membrane and heart-rate responses in the rabbit. Animal
such as Underwood (1961), sometimes despaired of
Learning and Behavior 28, 8091. being able to find consistent advantages for distribut-
Kehoe, E. J., Schreurs, B. G., Macrae, M, and Gormezano, I. ed practice over massed practice.
(1995). Effects of modulating tone frequency, intensity, and
duration on the classically conditioned nictitating membrane
Among researchers on human memory, an im-
response. Psychobiology 23, 103115. portant breakthrough was the research of Arthur Mel-
McLaren, I. P. L., and Mackintosh, N. J. (2000). An elemental ton (1967), who used a procedure that became stan-
model of associative learning: Latent inhibition and perceptu- dard for many subsequent investigators. Participants
al learning. Animal Learning and Behavior 28, 211246. saw a list of words presented one at a time. Some
Mis, F. W., Lumia, A. W., and Moore, J. W. (1972). Inhibitory con-
trol of the classically conditioned nictitating membrane re- words were shown once on the list, while others were
sponse of the rabbit. Behavior Research Methods and Instrumen- shown twice. Of the words that were shown twice,
tation 4, 297299. some were repeated in massed fashion; that is, they
Moore, J. W. (1972). Stimulus control: Studies of auditory general- were presented twice in a row. Other words were re-
ization in rabbits. In A. H. Black and W. P. Prokasy, eds., Clas-
peated in spaced or distributed fashion; that is, they
sical conditioning II: Current research and theory, pp. 206230.
New York: Appleton-Century-Crofts. had their occurrences separated by one or more inter-
Pearce, J. M., and Bouton, M. E. (2001). Theories of associative vening words. After presentation of the list was com-
learning in animals. Annual Review of Psychology 52, 111139. plete, participants were asked to recall the items in
116 DISTRIBUTED PRACTICE EFFECTS
any order. Melton found that the probability of recall Study-Phase Retrieval Accounts
for repeated items increased as a positive function of
Many different theories have been used to ex-
the number of intervening items between presenta-
plain distributed-practice effects. Although the de-
tions. This advantage for distributed repetitions over
tails of these theories vary, contemporary accounts
massed repetitions is often called the spacing effect;
can generally be grouped into three separate ap-
the fact that memory for spaced items may improve
proaches. One such approach emphasizes the impor-
somewhat as the distribution between repetitions is
tance of study-phase retrieval (Braun and Rubin,
increased further is sometimes called the lag effect.
1998). This approach assumes that one way in which
Are massed presentations ever more effective repetition helps to improve memory is by reminding
than distributed repetitions? Glenberg (1976) pres- the learner of earlier encounters with the studied in-
ented evidence that massed presentations may be su- formation. For example, if a particular word occurs
perior if memory is only required for a very short in- twice on a list, the second presentation may remind
terval. When the memory test is given almost the learner of the first presentation. The learner may
immediately after the last presentation, massed repe- then think about the first presentation again, and this
titions may lead to superior memory than spaced rep- retrieval and added rehearsal would make the first
etitions. Evidently, the benefit one obtains from hav- presentation particularly memorable. To explain why
ing multiple study episodes right before the test can the distribution of repetitions affects memory, one
sometimes be greater than the advantage usually must add on the assumption that this reminding pro-
gained from distributed rehearsal. Therefore, cram- cess benefits memory only if sufficient time has passed
ming may be a reasonably effective study strategy if to guarantee that the first presentation is not already
one is sure that the test will occur as soon as the cram- being consciously rehearsed.
ming is finished. However, distributed practice con-
It is reasonable that an encounter with a stimulus
sistently results in better long-term retention.
may serve as a reminder of earlier encounters. How-
ever, at least one critical claim of this approach is un-
Importance of Distributed-Practice Effects supported. A study-phase retrieval account predicts
One reason why the effects of distributed practice that memory for the first occurrence of a repeated
on memory are seen by many researchers as impor- item is affected by distribution of practice. Although
tant is that they have wide generality. One can find it is difficult to distinguish between memory for the
these effects in many different subject populations, first occurrence and memory for the second occur-
including nonhuman animals (Davis, 1970), human rence of a repeated item, what little evidence re-
infants (Cornell, 1980), children (Toppino, 1991), searchers have suggests that it is memory for the sec-
and the elderly (Balota, Duchek, and Paullin, 1989; ond occurrence, not the first, that is affected by
Benjamin and Craik, 2001). These effects are found distribution of practice (Hintzman, Block, and Sum-
on all sorts of memory tests and on a wide variety of mers, 1973). Although a study-phase retrieval process
different materials. may contribute to effects of distributed practice, there
is little direct evidence in favor of this approach.
Distributed practice can also be used to improve
learning and retention of meaningful material. For
example, Rea and Modigliani (1985) showed that dis- Deficient-Processing Accounts
tributed practice facilitates memory for spelling and
An alternative approach to explaining distribut-
multiplication facts. Reder and Anderson (1982)
ed-practice effects is to claim that learners do not pro-
found that distributed practice leads to improved
cess the second occurrence of an item as fully when
memory for information from textbooks, relative to
it is repeated in massed fashion as when it is repeated
massed practice. Dempster (1987) showed that learn-
in spaced fashion. That is, the second occurrence of
ing of the definitions of uncommon English words
a massed repetition is not given as much attention or
benefited from distributed practice. Bahrick and
rehearsal as the second occurrence of a spaced repeti-
Phelps (1987) demonstrated that retention of Spanish
tion.
vocabulary words over an eight-year period was great-
ly affected by distribution of practice. In short, distri- Zechmeister and Shaughnessy (1980) suggested
bution of practice seems to be as important in the that deficient processing of massed items may reflect
mastery of classroom material as it is in the memori- rehearsal strategies on the part of learners. For exam-
zation of lists in a laboratory. Dempster (1988) called ple, in the case of people memorizing a list of words,
for educators to be more sensitive to the importance the participants do not just rehearse each item as it
of this variable for classroom instruction, as distribut- is presented; rather, they presumably divide their re-
ing practice may improve retention without requiring hearsal between the current item and previous items.
additional time and resources. The amount of rehearsal that a person devotes to an
DISTRIBUTED PRACTICE EFFECTS 117
item presumably depends on how difficult the item each occurrence and that distributed rehearsal in-
appears to be. Zechmeister and Shaughnessy found creases the probability that variability in study will
that participants overestimate the extent to which occur.
they would remember massed repetitions. Because
people may mistakenly believe that massed items will
be easy to remember, they may not devote as much Multiprocess Accounts
rehearsal and attention to them as they do to spaced It seems increasingly likely that there will be no
items. Indeed, when participants are asked to re- single explanation for why distributed practice im-
hearse aloud or to pace themselves through a list, proves memory. Rather, there are several reasons why
massed repetitions seem to receive less rehearsal than distributed repetitions are more effective than
distributed repetitions. massed repetitions, and all three of the major ap-
proaches (study-phase retrieval, deficient processing,
Another version of this approach (advocated by
encoding variability) may apply under some circum-
Hintzman, 1974) attributes distributed-practice ef-
stances. Greene (1989) and Russo, Parkin, Taylor,
fects to deficient processing of the second occurrence
and Wilks (1998) have offered multiprocess accounts
of massed repetitions but views this as a result of an
that combine these approaches. The nature of the
automatic, unconscious process, not as the result of a
subject population, the stimulus material, and the ex-
deliberate strategy. This type of account has the ad-
perimental procedure all seem to determine the na-
vantage of being applicable to subject populations,
ture of the processes that underlie the advantage for
such as human infants, that show distributed-practice
distributed practice.
effects but that presumably do not employ conscious
rehearsal strategies. Although the theoretical explanation for the ef-
fects of distributed practice is likely to be complex,
there is no question that these effects are powerful
Encoding-Variability Accounts and of wide generality. The fact that distributed prac-
Encoding-variability accounts of distributed- tice leads to superior retention than massed practice
processing effects assume that spacing between repe- needs to be taken into account by both experimenters
titions facilitates memory by increasing the likelihood and educators.
that each occurrence of a repeated item is stored in
a very different way in memory. This sort of approach See also: REPETITION AND LEARNING
may best be explained by way of an analogy. Imagine
that a business office is very disorganized, and that Bibliography
workers often have difficulty finding papers they Bahrick, H. P., and Phelps, E. (1987). Retention of Spanish vocabu-
need. If they have an important document that they lary over 8 years. Journal of Experimental Psychology: Learning,
want to be able to find at a later time, it would make Memory, and Cognition 13, 344349.
sense for them to make multiple copies of it. Howev- Balota, D. A., Duchek, J. M., and Paullin, R. (1989). Age-related
differences in the impact of spacing, lag, and retention inter-
er, it would be wise for them to keep each of the cop-
val. Psychology and Aging 4, 39.
ies in different places and filed in different ways; this Benjamin, A. S., and Craik, F. I. M. (2001). Parallel effects of aging
would increase the probability that they would find at and time pressure on memory for source: Evidence from the
least one copy when they need it. Encoding-variability spacing effect. Memory & Cognition 29, 691697.
accounts assume that distributed practice increases Braun, K., and Rubin, D. C. (1998). The spacing effect depends on
an encoding deficit, retrieval, and time in working memory:
the probability that different occurrences of a repeat-
Evidence from once-presented words. Memory 6, 3765.
ed item will be stored in memory in different ways, Cornell, E. H. (1980). Distributed study facilitates infants delayed
thereby increasing the probability that a person recognition accuracy. Memory & Cognition 8, 539542.
would be able to retrieve at least one occurrence. Davis, M. (1970). Effects of interstimulus interval length and vari-
ability on startle-response habituation in the rat. Journal of
Different theorists have approached encoding Comparative and Physiological Psychology 78, 260267.
variability in different ways. Landauer (1976) argued Dempster, F. N. (1987). Effects of variable encoding and spaced
that distribution of practice directly influenced the lo- presentations on vocabulary learning. Journal of Educational
Psychology 79, 162170.
cation of memories of specific occurrences. Other (1988). The spacing effect: A case study in the failure to
theorists, such as Gartman and Johnson (1972) and apply the results of psychological research. American Psycholo-
Glenberg (1979), have emphasized that distribution gist 43, 627634.
of practice may increase the probability that a repeat- Ebbinghaus, H. (1885; reprint 1964). Memory: A contribution to ex-
ed item will be interpreted or analyzed differently at perimental psychology. New York: Dover.
Gartman, L. F., and Johnson, N. F. (1972). Massed versus distribut-
each occurrence. What these approaches have in comed repetition of homographs: A test of the differential-
mon is the assumption that repeated items are re- encoding hypothesis. Journal of Verbal Learning and Verbal Be-
membered better if they are studied differently at havior 11, 801808.
118 DRUGS
Glenberg, A. M. (1976). Monotonic and nonmonotonic lag effects effects in the central nervous system, allowing subjects
in paired-associate and recognition memory paradigms. Jour- to be used as their own controls. The ability to evalu-
nal of Verbal Learning and Verbal Behavior 15, 115.
(1979). Component-levels theory of the effects of spacing
ate memory performance in both a drugged and an
of repetitions on recall and recognition. Memory & Cognition undrugged state is particularly useful in human
7, 95112. studies, where the number of subjects can be limited.
Greene, R. L. (1989). Spacing effects in memory: Evidence for a An additional advantage of this approach is that using
two-process account. Journal of Experimental Psychology: Learn- the same subjects as both the control and experimen-
ing, Memory, and Cognition 15, 371377.
Hintzman, D. L. (1974). Theoretical implications of the spacing ef-
tal groups (within-subjects design) is a much more
fect. In R. L. Solso, ed., Theories in cognitive psychology: The Loy- powerful way to detect small drug effects. Using a
ola symposium. Hillsdale, NJ: Erlbaum. within-subjects experimental design is particularly
Hintzman, D. L., Block, R. A., and Summers, J. J. (1973). Modality useful for evaluating memory, because there can be
tags and memory for repetitions: Locus of the spacing effect. substantial differences in baseline performance
Journal of Verbal Learning and Verbal Behavior 12, 229238.
Jost, A. (1897). Die Assoziationsfestigkeit in Abhangigkeit von der
among individuals.
Verteilung der Wiederholungen. Zeitschrift fur Psychologie 14, Studies of the effects of drugs on memory are
436472. valuable for several reasons. First, it is important to
Landauer, T. K. (1976). Memory without organization: Properties
of a model with random storage and undirected retrieval.
know whether a drug that is being used to relieve a
Cognitive Psychology 7, 495531. particular condition can affect (usually impair) mem-
Melton, A. W. (1967). Repetition and retrieval from memory. Sci- ory as a side effect (see Table 1). For example, agents
ence 158, 532. that block the effect of acetylcholine, a neurotran-
Rea, C. P., and Modigliani, V. (1985). The effect of expanded ver- smitter long known to modulate memory, are fre-
sus massed practice on the retention of multiplication facts
and spelling lists. Human Learning 4, 1118.
quently used as part of the treatment for the move-
Reder, L. M., and Anderson, J. R. (1982). Effects of spacing and ment disorders of Parkinsons disease. At high doses,
embellishment on memory for the main points of a text. Mem- anticholinergic drugs produce profound memory im-
ory & Cognition 10, 97102. pairments. Second, experimental studies using drugs
Russo, R., Parkin, A. J., Taylor, S. R., and Wilks, J. (1998). Revising with defined mechanisms of action have helped to de-
current two-process accounts of spacing effects in memory.
Journal of Experimental Psychology: Learning, Memory, and Cogni-
fine the involvement of specific neurochemical sys-
tion 24, 161172. tems in memory. One example is the class of drugs
Toppino, T. C. (1991). The spacing effect in young childrens free called benzodiazepines. These drugs facilitate the ef-
recall: Support for automatic-process explanations. Memory & fects of the inhibitory neurotransmitter GABA
Cognition 19, 159167. (gamma-aminobutyric acid) and impair memory. Un-
Underwood, B. J. (1961). Ten years of massed practice on distrib-
uted practice. Psychological Review 4, 229247.
fortunately, not all agents that affect memory act dis-
Zechmeister, E. B., and Shaughnessy, J. J. (1980). When you know cretely. Ethanol (alcohol) impairs memory by an un-
that you know and when you think that you know but you known mechanism that appears to involve actions on
dont. Bulletin of the Psychonomic Society 15, 4144. several neurotransmitter systems and on nerve-cell
Arthur M. Glenberg membranes. Third, studies with drugs offer the possi-
Revised by Robert L. Greene bility of helping to better define different types of
memory. Fourth, drug studies offer the hope of devel-
oping treatments for diseases with symptoms of mem-
ory impairment, such as Alzheimers disease.
DRUGS
See: COGNITIVE ENHANCERS; DRUGS AND Characteristics of Memory
MEMORY; ELECTROCONVULSIVE THERAPY
There are several distinct memory systems in the
AND MEMORY LOSS; PHARMACOLOGICAL
brain. One useful classification scheme divides long-
TREATMENT OF MEMORY DEFICITS
term memory into declarative (explicit) and non-
declarative (implicit) categories. Declarative memory
involves the conscious recall of facts or events (e.g., re-
membering a name), whereas nondeclarative memo-
DRUGS AND MEMORY ry is not conscious and is usually expressed through
The psychopharmacological approach to the study of performance (e.g., riding a bicycle). Subtypes of
memory involves a systematic examination of the be- memory within each of these systems rely on different
havioral changes that occur following the administra- brain regions and therefore involve different sets of
tion of psychoactive drugs. This approach comple- neural connections and combinations of chemical
ments neuropsychological research in its neuro- neurotransmitters and second-messenger systems. In
biological approach to the study of learning and addition to distinctions based on neuroanatomy,
memory. Most psychoactive drugs produce reversible there are separate processes governing short-term
DRUGS AND MEMORY 119
and long-term memory. This diversity implies that Table 1

there are a number of sites where drugs can modulate
memory and that different drugs should influence
different types of memory.
There are at least three components of memory:
acquisition, consolidation, and retention or retrieval.
Each can be affected by drugs. The most frequently
documented effects of drugs on learning and memory
pertain to acquisition. In such studies researchers ad-
minister the drug before training begins. Arousal and
attention influence acquisition, so sedatives usually
impair memory, whereas stimulants usually enhance
it. Mood and motivation also affect acquisition;
agents that reduce either usually impair memory. At
the most basic level, drugs can impair acquisition by
interfering with sensory perceptionblurring vision,
for example. Possible drug effects on peripheral func-
tions must be controlled for in experimental studies,
especially when animals are used as subjects.
Administering the agent after training helps to
disentangle the memory-altering properties of drugs
from their effects on sensory or motor functions. This
is the protocol used to study the effects of drugs on
memory consolidation. If the drug has been metaboli-
cally eliminated by the time memory is tested, then
posttraining treatments affect only consolidation, not
acquisition or retention. Studies of consolidation in
animals have provided valuable information about
the biochemical mechanisms of memory formation
and the intervals during which learning activates
these mechanisms.
The research on drug effects on retrieval is not as
extensive as that on acquisition or consolidation. Per- structures used only in scientific research. Most com-
haps the best-known example of drug-induced alter- pounds mentioned here either have a role as medi-
ations in retrieval comes from the literature on state- cines or are well known among scientists who study
dependency. State-dependent retrieval refers to the abil- the neuropharmacology of memory.
ity of subjects to retrieve information better when The most pernicious drug-induced impairment
they are in the same state as they were when the mate- of memory is the destruction of neurons in brain re-
rial was acquired. For example, if material was gions involved in memory formation. For example,
learned under the influence of a drug such as alcohol, domoic acid destroys neurons in the hippocampus,
retrieval under sober conditions is often worse than an area that is critical for the formation of long-term
if the subject is again intoxicated. Some such effects declarative memories. Other agents act by severely
hinge on the nature of the retrieval test, and not all
depleting levels of neurotransmitters in pathways that
results appear to depend on the state of the subject.
modulate memory. One example of this phenome-
non is the recreational drug Ecstasy (MDMA), which
Drugs That Impair Memory impairs memory by reducing serotonin levels. While
the effects of these types of treatments are long last-
Many chemical compounds are known to impair ing, the memory impacts of most drugs are reversible.
memory; only a general overview of these agents is
possible here. The initial question is whether a given Many drugs that impair memory act as neuro-
compound can be considered a drug. At one end of transmitter receptors. Usually they reduce the func-
the spectrum are commonly available agents that nor- tion of a particular neurotransmitter. For example,
mally have another use. For example, sniffing the or- scopolamine blocks the receptor for acetylcholine,
ganic solvents (e.g., toluene) in glue will impair mem- propranol blocks a class of receptors for norepineph-
ory. At the other extreme are arcane chemical rine, and dizoclipine (MK-801) blocks a subtype of
120 DYSLEXIA
glutamate receptors. Sedative drugs are the excep- For example, picrotoxin reduces the inhibitory effect
tion: Benzodiazepines (e.g., diazepam [Valium]) and of GABA receptors, while amphetamine causes the re-
barbiturates enhance the action of GABA, the prima- lease of norepinephrine. Acetylcholinesterase inhibi-
ry inhibitory neurotransmitter in the brain, thus dis- tors increase the effect of acetylcholine and are used
rupting memory. to treat the memory deficits of Alzheimers disease.
All of these drugs are effective memory enhancers,
Long-term memory formation requires activation
but they all have have serious side effects.
of certain enzymes such as protein kinases (e.g, PKC,
PKA, CaMKII). This activation is controlled by several Drugs can influence many aspects of brain func-
different neurotransmitters and can be affected by tion, including learning and memory. Most drugs ad-
drugs that act at any of their receptors. Also, specific versely affect memory, although some can enhance it.
inhibitors of protein kinases impair memory. An im- There is an increasing need to develop safe and effec-
portant consequence of protein kinase activation is tive drugs to treat memory problems caused by aging
the initiation of new protein synthesis. Inhibitors of or disease.
protein synthesis (e.g., cycloheximide or anisomycin) Bibliography
can disrupt long-term memory formation. Castellano, C., Cabib, S., and Puglisi-Allegra, S. (1996). Psycho-
A given drug seldom uniformly impairs all types pharmacology of memory modulation: Evidence for multiple
interaction among neurotransmitters and hormones. Behav-
of memory. For example, ethanol and the ben- ioral Brain Research 77, 121.
zodiazepine drugs affect mainly long-term memory, Castellano, C., Cestari, V., and Ciamei, A. (2001). NMDA receptors
leaving short-term memory relatively unaffected. and learning and memory processes. Current Drug Targets 2,
These drugs also preferentially disrupt declarative 273283.
memory while leaving nondeclarative memory rela- Davis, H. P., and Squire, L. R. (1984). Protein synthesis and memo-
ry: A review. Psychology Bulletin 96, 518559.
tively intact. Duka, T. T., Curran, H. V. H., Rusted, J. M. J., and Weingartner,
There has been considerable interest in the possi- H. J. H. (1996). Perspectives on cognitive psychopharma-
cology research. Behavioral Pharmacology 7, 401410.
bility of using drugs in normal subjects to model clini- Ellis, K. A., and Nathan, P. J. (2001). The pharmacology of human
cal amnesias such as Korsakoffs syndrome or Alz- working memory. International Journal of Neuropsychopharma-
heimers disease. Knowledge of the neurophar- cology 4, 299313.
macology of a drug that would model a particular am- Farr, S. A., Flood, J. F., and Morley, J. E. (2000). The effect of cho-
nesic syndrome might suggest useful treatment strat- linergic, GABAergic, serotonergic, and glutamatergic recep-
tor modulation on posttrial memory processing in the hippo-
egies. The specific impairments seen in normal sub- campus. Neurobiology of Learning and Memory 73, 150167.
jects treated with alcohol resemble those of patients Lister, R. G., and Weingartner, H. J. (1987). Neuropharmacologi-
with Korsakoffs syndrome, and it has been suggested cal strategies for understanding psychobiological determi-
that scopolamine mimics some of the features of the nants of cognition. Human Neurobiology 6, 119127.
cognitive impairments seen in dementia patients. McGaugh, J. L., and Izquierdo, I. (2000). The contribution of
pharmacology to research on the mechanisms of memory for-
However, no drug treatment has been described that mation. Trends in Pharmacological Science 21, 208210.
completely mimics the pattern of memory disruption Squire, L. R., and Zola, S. M. (1996). Structure and function of de-
induced by a specific disease. clarative and nondeclarative memory systems. Proceedings of
the National Academy of Sciences of the United States of America 93,
13,51513,522.
Drugs That Enhance Memory Sutherland, R. J., Hoesing, J. M., and Whishaw, I. Q. (1990). Do-
moic acid, an environmental toxin, produces hippocampal
More drugs hinder memory than help it, so there damage and severe memory impairment. Neuroscience Letters
is a lively interest in discovering memory-enhancing 120, 221223.
drugs, especially because many neurological diseases Richard G. Lister
impair memory. Also, some memory loss occurs in old Herbert J. Weingartner
age even in healthy people. Finally, nearly everyone Revised by Gregory M. Rose
would like to have a better memory. An important un-
answered question is whether enhancing normal
memory is a reasonable possibility.
DYSLEXIA
In general, the agents that enhance memory act
in the opposite way from drugs that impair memory. See: LEARNING DISABILITIES
E
EARLY EXPERIENCE AND LEARNING The purpose of this entry is to illustrate the ef-
fects of alterations in early environments on learning
The brain of the developing organism is a unique and capacity in the adult organism. Several areas of inves-
dynamic system. During the prenatal and postnatal
tigation have demonstrated the importance of early
periods the brain differs dramatically from that of the
experience for later behavior, using a number of dif-
adult. For example, it contains more synapses early
ferent models to examine this issue. Thus, depriva-
in development than it does at any other stage in life
tion of visual experience early in development has
(Purves and Lichtman, 1980). Receptors for a number
been shown to markedly affect adult vision; this also
of neuropeptides (e.g., oxytocin) are found in higher
affects the animals behavior (Hyvarinen and Hyva-
concentrations early in development than later in life
rinen, 1979). There is an extensive research on the ef-
(Shapiro and Insel, 1989). In certain brain areas (e.g.,
fects of enriched environments on subsequent
the suprachiasmatic nucleus of the hypothalamus),
learning ability (Rosenzweig, 1984). This entry exam-
glucocorticoid receptors are found in high concentra-
ines the role of early handling as a model of infantile
tions only during early ontogeny (Van Eekelen et al.,
1987). These are but a few examples that attest to the stimulation on learning in the adult.
differences in the brain during development. For the In 1956 S. Levine and colleagues reported that
most part, scientists have not determined the func- neonatal manipulations have profound effects on
tional significance of these neuronal features of the later behavior. In essence, their study showed that
newborn brain. neonatal handling (i.e., removing rat pups from their
One of the critical aspects of the developing brain mother for a few minutes and returning them to their
is its plasticity. Both physiological and environmental mother), or electric shock during the same period,
stimuli have been shown to profoundly, and often markedly improved the animals capacity to learn a
permanently, influence the functional capacities of conditioned avoidance response when tested as
the organism (Levine and Mullins, 1966). Neonatal adults. This procedure involves placing the animal in
disturbance of the normal hormonal milieu leads to a two-compartment chamber. Both sides of the cham-
some of the best-known cases of permanent alter- ber contain grid floors that can be electrified. The an-
ations of function induced by early manipulations of imal is presented with a signal, the conditioned stimu-
physiological events. Neonatal exposure to hetero- lus, which is followed after a brief time by an electric
typical gonadal hormones, for example, results in a shock. The animal is required to cross from the elec-
permanently altered reproductive physiology and be- trified side of the chamber to the safe side. If it crosses
havior. Both over- and underexposure to thyroid hor- within the interval between the onset of the signal and
mone during early development cause changes in the the onset of shock, it avoids the shock (conditioned
brain that produce learning disabilities. avoidance). However, if the animal fails to cross dur-
121
122 EARLY EXPERIENCE AND LEARNING
ing this interval and the shock is delivered, this re- In order to address the issue of the effects of early
sponse is considered an escape response. handling on learning, V. H. Denenberg and J. R. C.
Researchers conducted a number of studies fol- Morton (1962) studied the effect of early handling on
lowing their initial observations that demonstrated the ability of rats to learn tasks that did not involve
these marked differences between handled and non- noxious stimuli. These investigators examined the in-
handled pups. In order to verify whether the effects teraction between early handling and environmental
of early handling were age-dependent, early-handled enrichment on the ability of adult rats to solve a
pups were compared with nonhandled pups and with Hebb-Williams maze. After weaning handled and
animals that were handled after weaning. Pups ma- nonhandled rats were reared in a neutral, restricted,
nipulated as infants were found to show avoidance or enriched environment. The animals were then re-
learning superior to that of nonhandled pups. Post- quired to solve a sequence of twelve test problems.
weaning-handled animals were more similar to non- The results indicated that the animals ability to learn
handled rats. the maze was affected by the postweaning, but not by
the pre-weaning, manipulation. Thus, rats reared in
Although experiments showed that early han- an enriched environment made significantly fewer er-
dling improved avoidance learning, the underlying rors than those reared in neutral or restricted envi-
causes of this improvement remained to be deter-
ronments. Based on these results, it was concluded
mined. One possible explanation for these findings is
that pre-weaning handling affected emotional pro-
that emotional reactivity was modified as a function
cesses but not learning ability.
of these early experiences. The more efficient avoid-
ance learning may therefore be a consequence of re- In 1972 R. Wong attempted to clarify the issue of
duced arousal levels. Increasing the shock levels dur- whether early handling directly affects associative
ing avoidance conditioning results in an inverted U- learning or whether the improvement in learning is
shaped function with regard to acquisition of the due mainly to differences in emotional reactivity. An
conditioned avoidance response. Thus, acquisition is experiment was conducted that presumably could dis-
improved by very low levels of shock but impaired by criminate these two processes. Thus, subjects were
very high levels. trained to reach a criterion of stable performance on
There are numerous reports that early handling a positive reinforcement task (food) and then pun-
reduces emotional reactivity (Whimbey and Denen- ished with an aversive stimulus (shock) for making the
berg, 1967). This reduced reactivity is demonstrable reinforced response. Comparisons could then be
using both behavioral and physiological indices. Ac- made in terms of the degree of response suppression
tivity levels and defecation in the open field differ be- following the presentation of the aversive event; and
tween handled and nonhandled subjects. Handled the rate of recovery after removal of the aversive stim-
animals explore more actively (i.e., show less freez- ulus. Wong argued that if handled animals showed
ing), defecate less in the open field, are less neopho- greater response suppression and a slower recovery
bic (Weinberg, Smotherman, and Levine, 1978), and than nonhandled animals, this would indicate a direct
are less reactive to human handling as adults (Ader, effect of handling on learning. If the contrary oc-
1965). One of the more sensitive physiological indices curred, it could be assumed that the primary influ-
of arousal is the activation of the hypothalamic- ence on the behavior was attributable to differences
pituitary-adrenal (HPA) system (Hennessy and Le- in emotionality. Animals were trained to alternate
vine, 1978). Following stress, nonhandled animals goal boxes in a T-maze to obtain food. Once the crite-
show plasma corticosterone (the primary glucocorti- rion had been reached, they were given a shock in the
coid secreted by the rodent) elevations that are both goal box where they had obtained positive reinforce-
larger and more persistent than those found in ani- ment (S+ box). Testing began one day after the pun-
mals handled during infancy (Levine et al., 1967). ished trial. The animals were placed in the maze and
received food only in the S+ box. During the acquisi-
These differences in emotionality are consistently
tion phase the handled group made more correct re-
found using a variety of different testing paradigms.
sponses (alternations) than the nonhandled group,
However, the effects of early handling on avoidance
indicating superior learning on a positive reinforce-
conditioning appear to vary when the parameters are
ment task. Following the shock exposure handled rats
different from those used in the original studies. Both
made fewer choices to the food reward box (S+) than
R. Ader (1965) and J. Weinberg and S. Levine (1977)
nonhandled animals, suggesting that handled ani-
failed to find differences in conditioned avoidance re-
mals had superior associative learning.
sponse, although differences in emotional reactivity
were clearly present. The question of whether early Decades later A. Tang (2001) investigated the
handling influences associative learning therefore role of exposure to novelty on hippocampal depen-
still remained unanswered. dent learning. Neonatal rats were exposed for three
EARLY EXPERIENCE AND LEARNING 123
minutes daily to a nonhome environment and com- shock; and 3. testing, during which latent inhibition
pared with littermate controls that remained in the was indexed by the animals suppression of licking
home cage for the period with the mother removed. during tone presentation. As in the previous experi-
The novelty exposed rats showed more rapid acquisi- ment, latent inhibition was observed in the handled
tion as juveniles and greater retention in adulthood males and females and in nonhandled females, but
in a spatial learning task. Further, these animals also not in the nonhandled males. Based on both of these
showed greater retention in an odor discrimination studies, the conclusion was that early handling exerts
task. a beneficial influence on learning capacity in the
adult animal.
In many of the studies described above, learning
was investigated using behavioral situations in which Only a very limited literature has attempted to
an aversive motivational component was present, examine the neural substrates of the early handling
making it difficult to dismiss entirely the effects of phenomenon. Michael Meaney and his colleagues
emotional reactivity on learning. Latent inhibition, a (1988) studied the long-term influence of early han-
behavioral paradigm that avoids some of these prob- dling on the neuroendocrine regulation of the HPA
lems, has been used to examine the relationship be- system. These researchers reported a long-term down
tween handling and learning. This paradigm was de- regulation of glucocorticoid receptors (GR) in the
scribed in the context of classical conditioning. Ivan hippocampus of nonhandled, but not of handled,
Pavlov was the first to demonstrate that repeated ex- aged animals. They further reported that spatial
posure to a conditioned stimulus (CS), prior to pair- learning, a hippocampus-dependent process, is sig-
ing this stimulus with an unconditioned stimulus nificantly improved in aged animals that had under-
(UCS), impairs the rate of conditioning that subse- gone early handling. However, the aged nonhandled
quently occurs (Pavlov, 1927). Thus, repeated expo- animals appeared to have suffered hippocampal cell
sure to a stimulus that is not followed by meaningful loss. The differences in spatial learning may thus be
consequences renders this stimulus ineffective for due to the prevention of this cell loss by early han-
subsequent learning. dling. However, given the results presented by Tang,
it would appear that the influence of early experi-
I. Weiner, Schnabel, Lubor, and Feldon (1985) ences on hippocampal dependent learning is evident
employed the latent inhibition paradigm to study the throughout the life span. These studies do demon-
question of early handling and learning. The experi- strate the long-term consequences of early handling
ment was conducted in two phases. During the first for at least one aspect of neural regulation. Although
phase (preexposure) members of one group were the implications of this down regulation of GRs for
placed in a shuttle box and presented with sixty five- learning have not been extensively investigated, there
second tones. Members of the second group were is some evidence that administering specific GR an-
placed in the shuttle box for an equivalent time with- tagonists interferes with the acquisition of a spatial
out exposure to the tones. In the second phase the an- learning task (Oitzl, and de Kloet, 1992). Other as-
imals were presented with one hundred tone (CS) and pects of the HPA system also seem to be involved in
shock (UCS) pairings and the number of conditioned learning and memory (van Wimersma Greidanus,
avoidance responses was recorded. The results 1982).
showed that nonpreexposed handled animals exhib-
ited better avoidance learning than nonhandled rats,
thus replicating earlier findings. However, pre- Conclusion
exposed handled animals performed more poorly There are many examples in the literature of
than non-pre-exposed handled rats. The findings long-term consequences of manipulations during in-
were to some extent sex-dependent: Whereas preex- fancy that affect learning and memory. However,
posed handled males and females and preexposed most of these studies have utilized toxic agents result-
nonhandled females exhibited the latent inhibition ing in permanent and irreversible morphological and
(i.e., performed more poorly than nonpreexposed physiological changes in the central nervous system
animals), the nonhandled males did not show any ef- that are later reflected as impairments in the adult or-
fect of preexposure on the conditioned avoidance re- ganisms ability to learn (Grimm, 1987). Environmen-
sponse. tal enrichment and early handling constitute exam-
In a further study, latent inhibition was investi- ples of subtle environmental manipulations that
gated using a conditioned emotional response (CER) cause permanent alterations in adult function that fa-
to test the influence of early handling (Weiner, Fel- cilitate rather than impair adult learning abilities.
don, and Ziv-Harris,1987). The CER procedure was
conducted in three phases: 1. pre-exposure; 2. acqui- See also: CHILDREN, DEVELOPMENT OF MEMORY IN;
sition, in which the pre-exposed tone was paired with EMOTION, MOOD, AND MEMORY
124 EBBINGHAUS, HERMANN
Bibliography Weiner, I., Schnabel, I., Lubow, R. E., and Feldon, J. (1985). The
Ader, R. (1965). Effects of early experience and differential hous- effects of early handling on latent inhibition in male and fe-
ing on behavior and susceptibility to gastric erosions in the male rats. Developmental Psychobiology 18, 291297.
rat. Journal of Comparative and Physiological Psychology 60, 233 Whimbey, A. E., and Denenberg, V. H. (1967). Experimental pro-
238. gramming of life histories: The factor structure underlying
experimentally created individual differences. Behavior 29,
Denenberg, V. H. (1969). Open-field behavior in the rat: What
296314.
does it mean? Annals of the New York Academy of Sciences 159,
Wong, R. (1972). Infantile handling and associative processes of
852859.
rats. British Journal of Psychology 63, 101108.
Denenberg, V. H., and Morton, J. R. C. (1962). Effects of prewean-
ing and postweaning manipulations upon problem-solving
Seymour Levine
behavior. Journal of Comparative and Physiological Psychology 55,
1,0961,098.
Deborah Suchecki
Grimm, V. E. (1987). Effects of teratogenic exposure on the devel-
oping brain: Research strategies and possible mechanisms.
Developmental and Pharmacology Therapeutics 10, 328345.
Hennessy, M. B., and Levine, S. (1978). Sensitive pituitary-adrenal
responsiveness to varying intensities of psychological stimula-
tion. Physiology and Behavior 21, 295297.
EBBINGHAUS, HERMANN (18501909)
Hyvarinen, J., and Hyvarinen, L. (1979). Blindness and modifica- Hermann Ebbinghaus was the founder of the experi-
tion of association by early binocular deprivation in monkeys.
mental psychology of memory. He laid the founda-
Child Care and Health Development 5, 385387.
Levine, S., Chevalier, J. A., and Korchin, S. (1956). The effects of tion for the scientific study of memory in a mono-
early shock and handling on later avoidance. Journal of Person- graph titled ber das Gedchtnis (1885), translated into
ality 24, 475493. English in 1913 under the title Memory: A Contribution
Levine, S., Haltmeyer, G. C., Karas, G. G., and Denenberg, V. H. to Experimental Psychology.
(1967). Physiological and behavioral effects of infantile stimu-
lation. Physiology and Behavior 2, 5559.
Levine, S., and Mullins, R. F. (1966). Hormonal influences on
brain organization in infant rats. Science 152, 1,5851,592.
Life
Meaney, M. J., Aitken, D. H., van Berkel, C., Bhatnagar, S., and Ebbinghaus was born on January 23, 1850, at
Sapolsky, R. M. (1988). Effects of neonatal handling on age- Barmen, near Bonn, Germany. His father was a well-
related impairments associated with the hippocampus. Science
to-do merchant. He studied languages and philoso-
239, 766768.
Oitzl, M. S., and de Kloet, E. R. (1992). Selective corticosteroid an- phy at the University of Bonn. He served in the army
tagonists modulate specific aspects of spatial orientation during the Franco-Prussian War of 18701871, and
learning. Behavior Neuroscience 106, 6271. upon returning to the university completed his doc-
Pavlov, I. P. (1927). Conditioned reflexes, trans. G. V. Arenp. London: toral dissertation in 1873. He then spent some five
Oxford University Press. years traveling in France and England. He began his
Purves, D., and Lichtman, J. W. (1980). Elimination of synapses in
research on memory at Berlin in 1878, spending
the developing nervous system. Science 210, 153157.
Rosenzweig, M. R. (1984). Experience, memory, and the brain. more than a year on the initial set of experiments.
American Psychologist 39, 365376. Upon completing these studies he became a private
Shapiro, L. E., and Insel, T. R. (1989). Ontogeny of oxytocin recep- lecturer at the University of Berlin in 1880, and he
tors in the rat forebrain: A quantitative study. Synapse 4, 259 continued his studies of memory. He repeated many
266. of the original experiments from 18791880 in 1883
Tang, A. (2001) Neonatal exposure to a novel environment en-
hances hippocampal-dependent memory function during in-
1884 and added new ones. He published the report
fancy and adulthood. Learning and Memory 8, 257264. on both series in his 1885 monograph.
Van Eekelen, J. A. M., Rosenfeld, P., Levine, S., Westphal, H. M.,
and de Kloet, E. R. (1987). Post-natal disappearance of gluco- Ebbinghauss life after he published his epoch-
corticoid receptor immunoreactivity in the suprachiasmatic making study was active and productive. He was ap-
nucleus of the rat. Neuroscience Research Communications 1, pointed a professor at the University of Berlin in
129133. 1886, remaining there until 1894, when he moved to
van Wimersma Greidanus, T. B. (1982). Disturbed behavior and
the University of Breslau. He stayed at Breslau for
memory in the Brattleboro rat. Annals of the New York Academy
of Sciences 394, 655662.
eleven years and then accepted an appointment at the
Weinberg, J., and Levine, S. (1977). Early handling influences on University of Halle. Over the years he became a
behavioral and physiological responses during active avoid- prominent and respected member of the new scientif-
ance. Developmental Psychobiology 10, 161169. ic discipline of experimental psychology. A major
Weinberg, J., Smotherman, W. P., and Levine, S. (1978). Early han- source of his renown lay in his textbook of general
dling effects on neophobia and conditioned taste aversion.
psychology, Grundzge der Psychologie (1897), which
Physiology and Behavior 20, 589596.
Weiner, I., Feldon, J., and Ziv-Harris, D. (1987). Early handling became the most widely read psychology text in Ger-
and latent inhibition in the conditioned suppression para- many. Ebbinghaus died of pneumonia at Halle on
digm. Developmental Psychobiology 20, 233240. February 26, 1909.
EBBINGHAUS, HERMANN 125
Ebbinghauss Approach to Memory

Before Ebbinghaus, the study of memory consist-
ed of philosophical armchair speculation concerning
remembering and forgetting in everyday life, and
clinical observations of patients with memory disor-
ders. The philosophical approach of the day is reflect-
ed in William Jamess Principles of Psychology (1890);.
the clinical approach is illustrated by the work of
Thodule Ribot. Both lines of thought produced
many insights into the nature and workings of normal
and impaired memory. However, there were also cu-
rious gaps; not surprisingly, the contemporary think-
ers were unaware of many of them. One widely held
view, for instance, maintained that memory could not
be studied by strict scientific methods. Although
methods of science had been applied to the lower
mental processes, such as sensation and perception,
under the general rubric of psychophysics, the
higher mental processes such as memory were re-
garded as being beyond the pale of such methods.
Another tacit idea of the time was that remembering
and forgetting occur in an all-or-nothing fashion: A
person either does or does not remember a fact, a
thought, a name, and the like. The possibility that
nonrecoverable mental contents could exist at differ-
ent levels of strength was discussed neither by philos- Hermann Ebbinghaus (Corbis-Bettmann)
ophers nor by students of memory pathology.
Ebbinghauss work changed all that. In his now-

which the associations to be learned were initially
classic monograph he introduced the general ap-
nonexistent. To that end he invented the nonsense
proach to the study and measurement of learning and
memory by psychological means, outlined the appro- syllable as a basic idea unit to be used in experiments
priate methodology, and reported a number of exon memory. A nonsense syllable is a meaningless sin-
periments illustrating the power of his methods. gle syllable consisting of two consonants separated by
a vowel or a diphthong (e.g., WEZ, SIF). A single les-
The general strategy that Ebbinghaus adopted son to be learned and remembered consisted of a se-
can be summarized in terms of three simple princi- ries of randomly chosen syllables. It was natural to
ples for the scientific study of mental processes that imagine that no associations existed between and
are not directly observable. These principles are as among the members of the series. The learning of a
valid today as they were when Ebbinghaus first made lesson (committing the series to memory) therefore
use of them. First, it is necessary to find a way of con- would involve the formation and strengthening of as-
verting the unobservable mental processes into ob- sociations between its constituent syllables. The pro-
servable behavior. Second, it is necessary to be able to cess of learning could be captured by observing and
measure this observable behavior reliably. Third, it is measuring some behavior that could be assumed to be
necessary to show that the behavior thus quantified closely correlated with changes in the associations.
varies systematically with other variables and experi-
mental conditions.
Methods and Results
The unobservable mental processes that Eb-
binghaus wanted to study and measure were associa- In all his experiments Ebbinghaus was his sole
tions between ideas. Like almost all of his contempo- subject. In numerous studies, in which he varied the
raries, he assumed that memory reflects the existence conditions of learning and retention, he would learn
of associations between ideas. He also thought that and then test himself with a large number of different
learning consists of the acquisition of associations, series of syllables. He would learn a given series by
whereas forgetting reflects their loss. Ebbinghaus de- first reading and then repeating the sequence of sylla-
cided that the study of the acquisition and loss of asso- bles aloud to the beating of a metronome, at the rate
ciations would best be undertaken in a situation in of two and a half syllables per second, until he could
126 EBBINGHAUS, HERMANN
produce the series faultlessly. The amount of effort Influence

required to master the series provided measures of Ebbinghauss work proved to be highly influential
both original learning and subsequent retention (or for a number of reasons. Despite the pioneering na-
forgetting, the opposite of retention). Ebbinghaus ture of his work, he did just about everything right by
adopted the number of readings, or the amount of the standards of science. He replaced philosophical
time required for the learning of the series, as the discussions about memory and its phenomena with
measure of learning. Some time later he would rel- tightly controlled experimental demonstrations of
earn the same series, using the same method of read- how memory could be measured and how memory
ing and repeating the syllables. The comparison of performance could be found to be related to and de-
initial learning and relearning scores provided a mea- termined by various independent variables. He dis-
sure of what Ebbinghaus called savings. Ebbinghaus cussed the sources of error and the problems of unre-
took savings to represent a measure of retention of liability of measurement. He explained and demon-
the original learning. strated how one could measure fine gradations in
mental processes that until then were thought to be
Using these methods of measurement of memo-
scientifically intractable. He showed how the higher
ry, Ebbinghaus investigated a number of basic phe-
mental processes seemed to obey the same general
nomena of learning and retention. The results of his
kinds of laws that governed the lower processes. He
experiments, concerning things such as the relation
explicitly and forcefully pointed out the intimate con-
between the length of the series and the difficulty of nection that exists between learning and memory, a
learning it, the effects of the original overlearning of realization that has guided the study of memory ever
a series on its subsequent relearning, the advantages since. Like many other novel ideas introduced by Eb-
of distributed over massed practice, and the shape of binghaus, the connection between learning and mem-
the forgetting curve, turned out to be highly regular ory is obvious in our day, but it had been overlooked
and lawful. Ebbinghaus exercised meticulous care in by most thinkers before Ebbinghaus. Perhaps the
carrying out his experiments. Among other things, he most important innovation introduced by Eb-
went to great trouble in performing large numbers of binghaus was his adoption of the basic study-and-test
replications of individual experiments. The resulting paradigm in which a subject learns some previously
regularity and lawfulness of his findings greatly im- unknown material and is subsequently tested for re-
pressed other scientists. tention of the studied material. The study-test para-
digm contrasted sharply with the then current philo-
In one particularly ingenious set of experiments
sophical practice of discussing problems and
Ebbinghaus measured and compared three kinds of
phenomena of memory from the vantage point of ex-
associations: forward associations between adjacent
isting associations.
members of a series, backward associations, and re-
mote associations. In order to measure remote associ- Three features of Ebbinghauss groundbreaking
ations he would initially learn a series of syllables in work that are most frequently mentioned in text-
a particular order, and subsequently relearn various bookshis invention of the nonsense syllable; his se-
series systematically derived from the original one. In rial learning task; and his adoption of the savings
these derived series the originally learned syllables method as a measure of strength of associations
were separated by a certain number of other syllables. have had little direct influence on succeeding genera-
For instance, if the original series is symbolized by A tions of memory researchers, who even shortly after
B C D E F . . . (. . . designating other syllables), then 1885 rapidly adopted other methods and techniques
the derived series skipping one syllable would con- of studying and measuring memory. Nonsense sylla-
sist of A C E . . . B D F . . . , and the derived series bles turned out to vary greatly in meaningfulness and
skipping two syllables would consist of A D . . . B E thus lost the advantage of their homogeneity. The se-
. . . C F. . . Ebbinghaus found that the savings in learn- rial learning task did not allow independent manipu-
ing these derived series varied systematically with the lation or assessment of stimulus and response func-
remoteness of the members of the derived series from tions in learning and retention. And the originally
one another in the originally learned series. These ingenious savings method was replaced with more di-
data suggested that in the course of learning a series rect methods of measuring retention and forgetting.
of syllables, associations are formed not only between Ebbinghauss most momentous single achieve-
immediately adjacent syllables but also among rement consisted in his convincing demonstration that
mote ones, the strength of the remote associations be- it is possible to reliably measure aspects of complex
tween any two members of a series varying directly mental processes that are not directly observable. Al-
with the degree of their remoteness in the original se- most as important were his general orientation and
ries. approach and his attitude and spirit in the matter of
ECCLES, JOHN 127
applying the methods of science to the study of the

human mind. These were embraced by his contempo-
raries and have continued to inspire and guide the
thinking of succeeding generations of students of psy-
chology interested in learning and memory.
Ebbinghauss pioneering role in the founding of
the field of research on human learning and memory
is universally acknowledged. ber das Gedchtnis rep-
resented a remarkable achievement of a great scien-
tist, one that has left an indelible stamp on the study
of one of the most fascinating problems of the human
brain (or mind).
Bibliography
Ebbinghaus, Hermann (1885). ber das Gedchtnis: Untersuchungen
zur experimentellen Psychologie. Leipzig: Duncker and Humblot.
Trans. (1913) H. A. Ruger and C. E. Bussenius, Memory: A con-
tribution to experimental psychology. New York: Teachers Col- John Eccles (Sven Landgren)
lege, Columbia University. Reprint (1964), New York: Dover.
(1897). Grundzge der Psychologie. Leipzig: Veit.
Hoffman, R. R., Bringmann, W., Bamberg, M., and Klein, R.
(1987). Some historical observations on Ebbinghaus. In D. S.
as an undergraduate. In 1927, he began working with
Gorfein and R. R. Hoffman, eds., Memory and learning: The Eb-
binghaus centennial conference. Hillsdale, NJ: Erlbaum. Charles Sherrington, Nobelist and pioneer of neuro-
Postman, L. (1968). Hermann Ebbinghaus. American Psychologist physiology. Together they discovered the dual-fiber
23, 149157. composition of the ventral spinal nerve roots and es-
Roediger, H. L. (1985). Remembering Ebbinghaus. Contemporary
tablished the time course of the central excitatory state
Slamecka, N. J. (1985). Ebbinghaus: Some associations. Journal of and the contrasting central inhibitory state, excitability
Experimental Psychology: Learning Memory, and Cognition 11, changes induced in spinal motoneurones by an im-
414435. pulse volley arriving from the same or the antagonis-
Endel Tulving tic muscle nerve, respectively. In 1929, Eccles re-
ceived a D. Phil. degree from Oxford for a thesis on
excitation and inhibition.
ECCLES, JOHN (19031997) Eccless Main Scientific Discovery

The Australian-born scientist John Carew Eccles was Eccless major scientific achievement was the
a pioneer in neuroscience, discovering the elementa- identification of the membrane potential changes un-
ry synaptic processes of the central nervous system derlying synaptic transmission in the central nervous
known as excitatory and inhibitory postsynaptic po- system. These studies began at Oxford with analyses
tentials (EPSPs and IPSPs). He was particularly eager of simpler synaptic systems such as the cat nictitating
to understand how synaptic changes could serve
membrane, the cervical sympathetic ganglia, and the
learning and memory processes. He was awarded
neuromuscular junction and continued with spinal
the Nobel Prize in physiology or medicine in 1963
cord and supraspinal synapses. In 1937, Eccles
for his pioneering analysis of central synaptic trans-
moved to Sydney, where he set up a physiological lab-
mission.
oratory in the Sydney Hospital and where he and W.
J. OConnor (1939) discovered and named the end-
Education plate potential, the immediate electrical muscle cell
Eccles was born on January 27,1903, in Mel- response to the nerve impulse. In the early 1940s
bourne, Australia. His parents, Mary and William Ec- Eccles was joined by two distinguished scholars,
cles, were both schoolteachers and strongly sup- Bernard Katz and Stephen Kuffler, who came to
ported his academic interests. Although he was highly Australia as refugees fleeing Nazi Germany and Nazi
interested in mathematics, he eventually decided to occupation of Austria, respectively. Together they an-
study medicine and entered Melbourne University at alyzed the properties of the end-plate potential (Ec-
the age of seventeen. After finishing his medical cles, Katz, and Kuffler, 1941, 1942), and Kuffler made
course with top marks in 1925, he won a Rhodes his classical report on the effect of curare (Kuffler,
Scholarship and entered Magdalen College, Oxford, 1942).
128 ECCLES, JOHN
The Electrical-Chemical Synaptic In 1951, Eccles suddenly came to the conclusion

Transmission Controversy that his former position regarding synaptic transmis-
sion in the spinal cord was untenable and that the sit-
Between 1933 and 1938, Eccles made a set of
uation was mediated by a chemical transmitter, just as
studies on synaptic transmission through the cervical
at peripheral synapses. Two convincing observations
sympathetic ganglion and identified a fast and a
resulted. First, there was invariably a distinct time
slower type of transmission. A pharmacological analy-
delay between the arrival of the afferent nerve im-
sis led him to conclude that, although the slower vari-
pulse and the onset of the postsynaptic response; the
ety of synaptic transmissions were chemically mediat-
electrical hypothesis required simultaneity. Second,
ed by acetylcholine (ACh), faster transmissions were
the polarity of responses at excitatory and inhibitory
likely to be electrically mediated. This interpretation
synapses was reversed in spite of similar presynaptic
challenged the view of Henry Dale and colleagues
action potentials in the two cases. Eccles suggested
(1936), who had hypothetized that all phases were
that the EPSP and IPSP were mediated by two differ-
due to ACh. Dale won the 1936 Nobel Prize for the
ent chemical mediators. The new results were dis-
codiscovery with Otto Loewi of acetylcholine as trans-
cussed in detail in the first five chapters of his Wayn-
mitter in the nervous system. Although the scientific
flete Lectures, The Neurophysiological Basis of Mind
debate was intense and lasted for several years, it
(1953), delivered at Magdalen College, Oxford, in
never damaged the lifelong friendship between Ec-
cles and Dale. Prompted by his friend the philoso- 1952. (A further three chapters discussed synaptic
pher Karl Popper, whom he first met in New Zealand plasticity, the cerebral cortex, learning, memory, and
in 1946, and influenced by his demand for testability the mind-brain problem.)
of the ideas and with falsification as an admirable sci- During his term at Otago University, Eccles suc-
entific goal, Eccles in 1949 restated his hypothesis cessfully reorganized and modernized the physiology
that neuromuscular and autonomous synapses are and biochemistry program, but his teaching burden
operated chemically, but transmission in the central was considerable. In 1953, he accepted a newly estab-
nervous system is likely to be electrically mediated. lished chair of physiology at the John Curtin School
of Medical Research in Canberra, Australia, and there
Discovery of EPSPs and IPSPs in he proceeded to set up a modern and efficient neuro-
Motoneurones physiological laboratory. Joined by scientific col-
leagues from Australia and abroad, Eccles continued
The introduction of new technology allowed re- to use intracellular recording to analyze reflex cir-
searchers (Fatt and Katz, 1951) to clarify the situation cuits. He also made a set of significant discoveries on
at the neuromuscular synapse, and the issue of the synaptic plasticityfor example, that posttetanic po-
spinal cord mechanism was resolved when Eccles and tentiation (PTP) is associated with an enhanced EPSP
two younger colleagues, Lawrence Brock and Jack amplitude to a standard test stimulus and, with Ar-
Coombs, working at Otago University in Dunedin, thur Buller and Rosamund Eccles (his daughter), that
New Zealand (where Eccles had moved in 1946), suc- motoneuronal type controls motor unit contraction
ceeded in impaling cat motoneurones with glass speed, a remarkable case of neuronal plasticity. In
micropipettes and recording the transmembrane re- 1955 he was invited to give the Herter lectures at
sponses to stimulation of various nerves (Brock, Johns Hopkins University, printed as The Physiology of
Coombs, and Eccles, 1952). A cathode follower and Nerve Cells (1957), a small book that had tremendous
amplifier designed by Coombs allowed the use of influence.
high-impedance electrodes that helped identify these
processes and were fundamental to the researchers Eccles and his colleagues analyzed a newly report-
success. When activating the afferent nerve, the mo- ed form of spinal inhibition reported by Frank and
toneurone response was a depolarizing signal (see Fuortes in 1957 and found that it was due to reduced
Figure 1), which Eccles called the excitatory postsynaptic transmitter release from the presynaptic terminals of
potential (EPSP). Conversely, stimulation of a nerve the test fibers, a finding they described as presynaptic
from an antagonistic muscle inhibited the motoneu- inhibition. Their pharmacological analysis pointed to
ron, signaled by a hyperpolarizing response, the in- gamma-amino-butyric acid (GABA) as the mediating
hibitory postsynaptic potential (IPSP). Injection of ions agent. In other collaborations, Eccles made a series of
via the impaling microelectrode showed that the IPSP valuable discoveries about pre- and postsynaptic inhi-
depended upon the chloride and potassium ion gra- bition in the dorsal column nuclei, recurrent inhibi-
dients across the membrane. For these discoveries, tion in the thalamus, hippocampal basket cell inhibi-
Eccles was awarded the Nobel Prize in physiology or tion, and cerebellar organization. The largest impact
medicine for 1963, shared with Alan Hodgkin and of Eccless studies of supraspinal neurons came from
Andrew Huxley. his work on the cerebellum, supported by the neuro-
ECCLES, JOHN 129
Figure 1
Properties of EPSP and IPSPs. A. superimposed records of excitatory postsynaptic potentials (EPSPs) from a biceps-semitendinosus
motoneuron in response to increasing number of afferent fiber impulses from muscle spindles of its own muscle. Insets (stars) show the
signals of the impulses in the afferent nerve as they enter the spinal cord. B. an EPSP recorded at various artificially set levels of the
membrane potential (small numbers). The reversal around zero indicates that several ion currents (mostly sodium currents) are
responsible. C. inhibitory postsynaptic potentials (IPSPs, lower traces in each pair) in response to increasingly strong impulse volleys
from an antagonistic muscle nerve (upper traces in each pair). D. the reversal of the IPSP at potentials negative to about -80 mV,
suggesting that chloride ions are involved since the chloride equilibrium potential is at this level.
histologist Jnos Szentgothai of Budapest and by brain, concentrating on models for perception and
findings made by Masao Ito of Tokyo, who found will in motor tasks. From his adolescence, Eccles had
that cerebellar Purkinje cells were monosynaptic in- been strongly interested in the mind-brain problem,
hibitory on target neurons in intracerebellar and ves- and he returned to the subject throughout his life, in
tibular nuclei. In the influential book Cerebellum as a particular after 1975. Although he met considerable
Neuronal Machine (1967), Eccles, Ito, and Szent- opposition from other neuroscientists, he proposed a
gothai characterized all neuronal elements and their set of models for illustrating an interaction between
properties and produced a detailed chart of the cere- the mind and cortical neuronal activity. His writings
bellum and. reveal him as a dualist, with ideas similar but not iden-
tical to those of Ren Descartes, maintaining that our
mental processes are not identical to the associated
The Mind and the Brain physical nervous activity. He incessantly strived to im-
In the last chapter of The Neurophysiological Basis prove and clarify his mind-brain proposals; above all,
of Mind (1953), Eccles gave a first account, largely in he wanted to design situations where a dualistic influ-
pictorial form, of a possible site of interaction be- ence on the brain, a separate effect of the mind, could
tween the mind and the physical machinery of the be experimentally tested. His book The Self and Its
130 ECOLOGICAL MEMORY
Brain (1977), coauthored with Karl Popper, achieved Frank, K., and Fuortes, M. G. F. (1957). Presynaptic and postsyn-
a wide readership and considerable influence. aptic inhibition of monosynaptic reflexes. Federeation Proceed-
ings 16, 3940.
Kuffler, S. W. (1942). Electrical potential changes at an isolated
nerve-muscle junction. Journal of Neurophysiology 5, 1826.
A Magnificent Leader Popper, K. R., and Eccles, J. C. (1977) The self and its brain. Berlin:
A major factor in Eccless success was his ability as Springer.
a team builder and a leader. He demanded much of Per O. Andersen
his colleagues, but he was generous and rewarding to
those who engaged themselves fully in the research.
Although his enthusiasm and excitement were infec-
tious, he insisted upon controlled experiments and
repeatability of observations to avoid false results. His
ECOLOGICAL MEMORY
main ambition was to acquire insight in the problems See: NATURAL SETTINGS, MEMORY IN
and phenomena under investigation and not merely
a description. He combined wide knowledge and a
tremendous working capacity (more than 550 scien-
tific articles) with a keen demand for further under- EIDETIC IMAGERY
standing, the whole coupled to a humility toward the
magnitude of the challenge of understanding the Nothing captures better the popular belief in photo-
brain. graphic memory than the term eidetic imagery, al-
though the latter hardly supports the exaggerated
See also: GUIDE TO THE ANATOMY OF THE BRAIN: claims made for the former capacity. Photographic
SYNAPSE; NEUROTRANSMITTER SYSTEMS AND memory is the general claim that people can still see
MEMORY in front of them things that were experienced in the
past. Eidetic imagery, on the other hand, is more
Bibliography closely tied to objective experimental criteria.
Brock, L. G., Coombs, J. S., and Eccles, J. C. (1952). The nature
of the monosynaptic excitatory and inhibitory processes in the A generation of German investigations of eidetic
spinal cord. Proceedings of the Royal Society of London, Series B, imagery in the early years of the twentieth century
Biological Sciences 140, 170176. (Woodworth, 1938, p. 45) was largely ignored at mid-
Coombs, J. S., Eccles, J. C., and Fatt, P. (1955a). Excitatory synaptic century when American psychology was dominated by
action in motoneurones. Journal of Physiology (London) 130,
374395. theoretical behaviorism and had, at best, no use for
(1955b). The inhibitory suppression of reflex discharges the concept of imagery. The silence was broken in
from motoneurones. Journal of Physiology (London) 130, 396 1964 by publication of a paper by R. N. Haber and
413. R. B. Haber (see also later summaries in Haber, 1979,
(1955c). The specific ionic conductances and the ionic and accompanying commentaries). Their report
movements across the motoneuronal membrane that produce
the inhibitory post-synaptic potential. Journal of Physiology
launched modern research on eidetic imagery and
130, 326373. largely sustained conclusions from the continental
Dale, H. H., Feldberg, W. and Vogt, M. (1936). Release of acetyl- work of a generation earlier.
choline at voluntary motor nerve endings. Journal of Physiology
(London) 86, 353380. Haber and Haber (1964) studied 150 elementary-
Eccles, J. C. (1949). A review and restatement of the electrical hy- school children in a standardized testing situation.
potheses of synaptic excitatory and inhibitory action. Archives The children were shown a set of four coherent pic-
des sciences physiologiques (Paris) 3, 567584. tures for thirty seconds apiece and interviewed imme-
(1953). The neurophysiological basis of mind: The principles of diately after each as to what they saw on a blank
neurophysiology. Oxford: Clarendon Press.
(1957). The physiology of nerve cells. Baltimore: Johns Hop-
easel in the same location as the picture had been.
kins Press. Eight measures were collected, such as whether they
Eccles, J. C., Ito, M., and Szenthgothai, J. (1967). The cerebellum saw an image, how long it lasted, whether the image
as a neuronal machine. Berlin: Springer. description used positive coloration (rather than
Eccles, J. C., Katz, B., and Kuffler, S. W. (1941). Nature of the end- complementary colors, as in afterimages), and wheth-
plate potential in curarized muscle. Journal of Neurophysiology
4, 362387. er it was described in the present tense. Although
(1942). Effect of eserine on neuro-muscular transmission. more than half the children (84/150) reported at least
Journal of Neurophysiology 5, 211230. some kind of imagery for the presented picture, there
Eccles, J. C., and OConnor, W. J. (1939) Responses which nerve was considerable variability in scores on these eight
impulses evoke in mammalian striated muscle. Journal of Phys-
measures: In particular, a group of twelve children
iology (London) 97, 44102.
Fatt, P., and Katz, B. (1951). An analysis of the end-plate potential was easily distinguished from the other seventy-two
recorded with an intracellular electrode. Journal of Physiology who had indicated some imagery. These twelve chil-
115, 320369. dren were discontinuous with their classmates in the
EIDETIC IMAGERY 131
presence of positive coloration, duration of the im- been unimpressed with the evidence favoring such a
ages, use of the present tense to describe images, and view, the many peer commentaries following his arti-
visual scanning (of the blank surface) during the in- cle demonstrate that the point is at least controversial.
terview after each picture. For example, positive col- Leask and colleagues (1969) showed no evidence for
oration was an average of 90 percent in the group of mental retardation among their eidetic children, but
twelve but an average of only 34 percent in the re- they did observe more minor visual defects (wearing
maining seventy-two children. In visual scanning (eye glasses) among the eidetic than among the noneidetic
movements across the blank easel where the eidetic children. Siipola and Hayden (1965) tested mentally
image was projected), the difference was even larg- retarded children and found incidence rates of about
er (100 percent versus 2 percent). Thus, the incidence fifty percent, a strikingly higher figure than among
of eidetic skill in the original survey was 8 percent (12/ the comparable nonretarded population. Further-
150). A survey using similar criteria by Paivio and more, eidetic imagery was a marker, among these
Cohen (1979), on 242 second- and third-grade chil- children, for organic as opposed to familial diag-
dren, gave excellent agreement on the incidence of noses of retardation. Other investigators (Gummer-
eidetic imagery in normal schoolchildren8.6 per- man et al., 1972; Richardson and Cant, 1970; Sym-
cent (21/242). mes, 1971) have, however, reported many fewer cases
In subsequent work (Leask et al., 1969) the fu- among retardates of all types. But Giray and col-
sion method was used to identify eidetic imagers. leagues (1976) found even a higher incidence (78.5
This method presents two individually meaningless percent of fourteen cases) in hydrocephalic children
pictures successively. After display of the first, the sec- and only a normal rate (5 percent to 10 percent)
ond picture is presented on the same surface, with the among children with other forms of mental retarda-
subject instructed to superimpose his or her image of tion. The force of these tantalizing observations is not
the first picture upon the second. The pictures are de- yet clear.
signed so that this superposition yields a meaningful
picture. Children identified by the criteria above to be
possessors of eidetic imagery could perform this task,
Cross-Cultural Approaches
whereas normal children could not. In the context of eidetic imagery as a develop-
mentally primitive information storage mode, Doob
(1966) supposed that primitive, illiterate societies
Age might show a higher incidence of it, even among
Giray and colleagues (1976) examined 280 chil- adults. An initial report did, as expected from this
dren, twenty at every age from five- to eighteen-years reasoning, show a high incidence using the Habers
old, using the Haberss (Haber and Haber, 1964) criteria among the Ibo of Nigeria. However, Doobs
criteria. Fifteen children (5.6 percent) were identified further research was disappointing: He had expected
as eidetic, but a clear relation to age emerged: Nine that within the Ibo population, the incidence of eidet-
of these fifteen were either five- or six-years old, and ic imagery would be greater in truly remote, agrarian
only a single subject was over ten. The decline in ei- settlements than in more modern, urbanized popula-
detic skills with age is well documented (Haber, 1979; tion centers, but this was not the case.
Leask et al., 1969; Richardson and Harris, 1986);
they are apparently virtually absent among adults.
Recent evidence (Giray et al., 1985; Zelhart et al.,
Conclusion
1985) suggests that the eidetic skill increases in geri- The most important conclusion about eidetic im-
atric populations and that the true relation between agery is that it is a genuine phenomenon, capable of
age and the incidence of eidetic imagery should be U- objective measurement and study. Moreover, the skill
shaped. This possibility places the interpretation of has been distinguished from such related phenomena
high eidetic skill among young children in a different as iconic memory, sensory afterimages, and extreme-
light: These young individuals might be especially ly accurate memory (but see Gray and Gummerman,
likely to show eidetic imagery not because of their age 1975, for reservations on this last point). Eidetic im-
but because of the functional level of their brains. agery is characteristic of a minority of young children
and is probably related to some forms of brain disor-
ders. It is rather certainly not the agency for good
Brain Damage memory of detail, as the Habers (1964) showed origi-
The suggestion that eidetic imagery varies in- nally. Thus, systematic work on eidetic imagery indi-
versely with age (up to the college years) indicated cates that it shares practically nothing with the popu-
that it might be a marker for retarded development lar concept of photographic memory. The apparent
within any age group. Although Haber (1979) had absence of eidetic imagery among adults and the fact
132 ELECTROCONVULSIVE THERAPY AND MEMORY LOSS
that it is not predictive of particularly good memory case one electrode is applied to each side of the head,
for detail make it a poor basis for claims of photo- or unilateral, in which case two electrodes are applied
graphic memory. to the right side of the head. The benefit of ECT is
evaluated by considering both its effectiveness for
See also: CODING PROCESSES: IMAGERY treating depression and the adverse effects of treat-
Bibliography ment. The most prominent of the adverse effects is
Doob, L. W. (1966). Eidetic imagery: A crosscultural will-o-the- impaired memory. The extent of the memory impair-
wisp? Journal of Psychology 63, 1334. ment varies depending on how ECT is administered.
Giray, E. F., Altkin, W. M., and Barclay, A. G. (1976). Frequency Memory impairment is greater after bilateral ECT
of eidetic imagery among hydrocephalic children. Perceptual than after unilateral ECT, and it is greater when ECT
and Motor Skills 43, 187194.
is administered using machines that deliver sine-wave
Giray, E. F., Altkin, W. M., Vaught, G. M., and Roodin, P. A. (1976).
The incidence of eidetic imagery as a function of age. Child current rather than brief pulses of current.
Development 47, 1,1071,210. Studies of the memory impairment associated
(1985). A life span approach to the study of eidetic imagery.
Journal of Mental Imagery 9, 2132.
with ECT suggest that memory is affected only tem-
Gray, C. R., and Gummerman, K. (1975). The enigmatic eidetic porarily. After a course of treatment, which typically
image: A critical examination of methods, data, and theories. involves six to twelve treatments given over a period
Psychological Bulletin 82, 383407. of two to four weeks, the ability to learn new material
Gummerman, K., Gray, C. R., and Wilson, J. M. (1972). An attempt is reduced and access to some memories that were
to assess eidetic imagery objectively. Bulletin of the Psychonomic
Society 28, 115118.
formed prior to ECT is lost. Anterograde amnesia refers
Haber, R. N. (1979). Twenty years of haunting eidetic imagery: to the difficulty that patients have in remembering
Wheres the ghost? Behavioral and Brain Sciences 2, 583629. events that occur after treatment begins. This difficul-
Haber, R. N., and Haber, R. B. (1964). Eidetic imagery: I. Frequen- ty persists for many weeks after treatment, gradually
cy. Perceptual and Motor Skills 19, 131138. resolving as the capacity for new learning recovers.
Leask, J., Haber, R. N., and Haber, R. B. (1969). Eidetic imagery
in children: II. Longitudinal and experimental results. Psycho-
Retrograde amnesia, the loss of memories acquired
logical Monograph Supplements 3 (3) (whole no. 35). prior to treatment, can initially involve memories ac-
Paivio, A., and Cohen, M. (1979). Eidetic imagery and cognitive quired many years earlier. Access to these memories
abilities. Journal of Mental Imagery 3, 5364. gradually recovers as time passes after treatment.
Richardson, A., and Cant, R. (1970). Eidetic imagery and brain
damage. Australian Journal of Psychology 22, 4754. It should be emphasized that memory for the pe-
Richardson, A., and Harris, L. J. (1986). Age trends in eidetikers. riod surrounding the treatment does not recover after
Journal of Genetic Psychology 147, 303308. ECT. For example, when patients were asked three
Siipola, E. M., and Hayden, S. D. (1965). Exploring eidetic imagery
among the retarded. Perceptual and Motor Skills 21, 275286.
years after treatment to identify what past time peri-
Symmes, J. S. (1971). Visual imagery in brain injured children. Per- ods they had difficulty remembering, the average pa-
ceptual and Motor Skills 21, 507514. tient reported difficulty remembering the time dur-
Woodworth, R. S. (1938). Experimental psychology. New York: Holt. ing ECT, the two months after treatment, and the six
Zelhart, P. F., Markley, R. B., and Bieker, L. (1985). Eidetic imag- months prior to treatment. Thus, except for this lacu-
ery in elderly persons. Perceptual and Motor Skills 60, 445446.
na around the time of ECT, formal memory testing
Robert G. Crowder suggests that patients eventually recover their capaci-
ty for learning and memory. At the same time, ab-
sence of evidence for a lasting memory problem is not
the same as proving that no such problem exists. It
ELECTROCONVULSIVE THERAPY is possible that more sensitive tests could be devel-
AND MEMORY LOSS oped that would detect persisting impairment. It is al-
ways difficult to prove that something does not exist.
Electroconvulsive therapy (ECT) was developed in
However, memory tests sensitive enough to show dif-
the 1930s as an alternative to psychiatric treatments
ferences between the memory abilities of healthy
that depended on inducing a convulsion. More re-
forty-year-olds and healthy fifty-year-olds (some de-
cently, ECT has been reviewed and evaluated by sci-
cline in memory ability does occur with normal aging)
entific groups in several countries, and has been
do not detect lasting memory problems in patients
found to be a safe and effective treatment for severe
who have received ECT.
and disabling depression. The therapeutic effect is
caused by a brain seizure, not a convulsion visible in In contrast with the findings from memory tests,
the limbs. In contemporary practice, ECT is adminis- it is noteworthy that some patients do report, even
tered in conjunction with a short-acting general anes- long after ECT, that their memory is not as good as
thetic and a muscle relaxant. As a result, the seizure it used to be. Although it is possible that the patients
is most easily detected by recording brain waves dur- have a degree of sensitivity about their own memory
ing treatment. ECT can be either bilateral, in which problems beyond what can be detected by memory
EMOTION, MOOD, AND MEMORY 133
tests, there are a number of other possibilities. One perform cognitive tasks. These studies usually focus
possibility is that, having recovered gradually from a on unpleasant emotions and moods, such as depres-
period of rather severe and easily documented mem- sion and anxiety. In other studies, psychologists at-
ory impairment, it is difficult for a person to know tempt to induce either unpleasant or pleasant moods
when memory abilities have recovered to what they in the participants (perhaps by having them listen to
should be. People who lead active lives use their mem- different types of music) and then examine how per-
ories many times each day to recall past events and formance is affected by these manipulations. Both
previously acquired knowledge. It is commonplace types of research have tried to answer two major ques-
for recall to be incomplete or inaccurate, especially tions about the interaction of mood and memory: Do
for information that lies at the fringes of our stored people remember events that are emotionally consis-
knowledge, such as information that was encountered tent with their moods better than other events? Do
only once or material that was not fully attended to depressed and anxious moods hinder performance
when it was first encountered. Sometimes memory on neutral cognitive tasks?
fails altogether. If someone has had ECT, how can he
or she know whether any particular failure of memory
is normal or whether it might be due to ECT? To the Mood-Congruent Memory
extent that ECT does lead many patients to doubt the People remember episodes and materials that are
integrity of their own memories, it is possible that this consistent with their moods more often than they re-
effect of treatment could be attenuated or eliminated member other occurrences; this phenomenon is
by sympathetic and informed counseling during the known as mood-congruent memory (MCM). MCM
period immediately following ECT. can sometimes be attributed to the ways that people
initially interpret the events to be remembered be-
Bibliography cause interpretations tend to be mood-congruent. A
American Psychiatric Association (1990). The practice of ECT: Recom-
clear example of mood-congruent interpretation can
mendations for treatment, training and privileging. Washington,
DC: American Psychiatric Association. be seen in research conducted by Michael Eysenck
Consensus Conference (1985). Electroconvulsive therapy. Journal and his colleagues (1987): Anxious participants, more
of the American Medical Association 251, 2,1032,108. often than nonanxious participants, spelled spoken
DElia, G., Ottosson, J. O., and Stromgren, L. S. (1983). Present homophones (such as die and dye) to coincide with the
practice of electroconvulsive therapy in Scandinavia. Archives
more threatening concept. Andrew Mathews and his
of General Psychiatry 40, 577581.
Fink, M. (1979). Convulsive therapy: Theory and practice, pp. 203 associates (1989) found similar mood-congruent in-
204. New York: Raven Press. terpretation on a test of implicit memory. In this test,
Malitz, S., and Sackeim, H., eds. (1986). Electroconvulsive therapy: the participants were shown the first three letters of
Clinical and basic research issues. Annals of the New York Academy words and were asked to complete them to form the
of Sciences 462.
first word that came to mind. The anxious subjects
Royal College of Psychiatrists. (1989). The practical administration of
electroconvulsive therapy (ECT). London: Gaskell. completed with threat-related words that they had
seen in an earlier task more often than other types of
Larry R. Squire
previously seen and unseen words; nonanxious sub-
jects did not show this bias. As they occur in both ini-
tial encounters and later indirect tests of memory,
these biased interpretations occur automatically or
EMOTIONAL MEMORY without any intent to focus on mood related meaning.
See: NEURAL SUBSTRATES OF EMOTIONAL On more direct tests of memory, such as tests of
MEMORY
explicit recall, anxious people do not always remem-
ber anxiety-related material better than other materi-
al, perhaps because they turn their attention away
from anxiety-provoking stimuli, once they are con-
EMOTION, MOOD, AND MEMORY ceived. Research concerned with depressed and sad
The ways in which we attend, learn, and remember states, however, shows more consistent evidence of
are related to our transitory moods and to our endur- MCM for both autobiographical and experimentally
ing emotional states. Intuitively appealing to the self- controlled events. A thorough review of MCM is pro-
reflective person, this claim has been verified by ex- vided in Cognitive Psychology and Emotional Disorders
perimental and clinical psychologists in both labora- (Williams, Watts, MacLeod, and Mathews, 1997),
tory and naturalistic studies. In some studies, which also describes variations and exceptions to the
psychologists measure differences in emotional states basic findings. One variation is that depressed moods
and determine whether those differences are associat- are often associated with a reduction in the recall of
ed with differences in the ways that the participants positively toned events instead of an increase in the
134 EMOTION, MOOD, AND MEMORY
recall of negative events. More exceptional is the find- Weingartner and his colleagues (1981), for example,
ing that temporarily sad students have shown evi- discovered that clinically depressed patients could
dence of mood-incongruent recall (better recall for learn lists of words organized into simple categories
positive material), which Parrott and Sabini (1990) as well as could other people, but when the same word
have interpreted as an outcome of the students elab- lists were disorganized, the depressed patients
orative processing of positive material, in attempt to learned less well. When people approach these types
improve their moods. of learning tasks by providing their own organization
or by using other elaborative strategies, they later
To the extent that MCM is observed in depres-
enjoy benefits on tests of deliberate memory. In con-
sion, it reflects enduring concerns with negative
trast, depressed people use fewer self-initiated proce-
events and at the same time plays an important role
dures and suffer the memory-related consequences.
in maintaining depressed mood. Sonja Lyubormirsky
and her colleagues (1998) found that ruminations by Similar conclusions can be reached in examining
students in depressed moods increased their recall of different types of memory tests. Hertel and Hardin
negative events. Moreover, just as they tend to re- (1990), for example, found that depressed college
member fewer positive episodes from the past, de- students performed as well as other students when the
pressed people also tend to expect fewer positive test of memory did not explicitly focus attention on
events to occur in the future (MacLeod, 1999). the past event (i.e., the spelling of homophones that
indirectly revealed memory for their prior exposure).
When the test required such explicit focus, however,
Mood-Related Impairments in Memory for depressed students did not spontaneously use strate-
Emotionally Neutral Events gies for recognition that characterized the perfor-
Because depressed and anxious people ruminate mance of the other students. Similarly, in tests of au-
about self concerns, it not hard to imagine that they tobiographical memory, depressed patients recall is
devote less attention to the emotionally neutral events often inappropriately general. They respond to in-
of everyday experience. Some routine cognitive acts structions to recall prior specific events when given
require little attention for successful performance; cues (e.g., happy) by citing categories of events instead
the cognitive processes involved are relatively auto- of particular episodes (e.g., going to baseball games, in-
matic, which means that they are well practiced and stead of the time my dad took me to see the Yankees for my
can occur simultaneously with other cognitive pro- birthday; Healy and Williams, 1999).
cesses, even rumination about personal concerns. Why do mood-impaired people experience im-
Other tasks require a more laborious and deliberate paired attention and corresponding memory deficits?
focus of attention if good performance is to be Some researchers (e.g., Weingartner et al., 1981) have
achieved. In short, cognitive tasks vary in the degree proposed that the deficits result from a fundamental
to which focused attention is required for good per- depletion of resources, possibly associated with bio-
formance. This is true of procedures that are per- chemical changes. Alternatively or in addition, the
formed during initial exposure or learning and tasks difficulties might indirectly reflect mood-impaired
that are devised to reveal memory for past events. peoples enduring and ruminative concern with
At the time of initial learning, for example, read- mood-related aspects of their experienceaspects
ing a long list of unrelated words requires little effort that are often irrelevant to the task at hand. These
by fluent readers, but organizing them in ways that task-irrelevant thoughts can distract attention when
will be useful during later attempts to remember participants are left to their own devices (e.g., when
them clearly requires more deliberate focus. Similar- they are told to learn a list of words). Yet when learn-
ly, tests of memory for those words vary in the degree ing or memory tasks are devised in ways that con-
of focused attention that they require. Rereading the strain attention to task materials, specify appropriate
same words is one index of memory (in that the previ- strategies, or distract from personal concerns, mood-
ously read words can be read faster than new words). impaired people may perform as well as others (Her-
This type of implicit-memory test involves procedures tel, 2000).
that are relatively automatic. In contrast, trying to re-
Memory for emotionally neutral events might
call the words on the list is a deliberate task that can
also benefit from the correspondence of mood at the
benefit from a great deal of attention and the use of
time of the test to the mood at the time of initial learn-
special strategies.
ing. The advantage of a similar mood state on both
The learning and memory tasks that benefit from occasions has been called mood-dependent memory
focused attention are the tasks that present difficulties (MDM) and is similar to other state-dependent mem-
to depressed and anxious people; they perform less ory effects, such as those obtained with alcohol. MDM
well than do people who are not mood-impaired. also has much in common with MCM: Remembering
EPISODIC MEMORY 135
mood-congruent events is often a matter of being in Williams, J. M. G., Watts, F. N., MacLeod, C., and Mathews, A.
the same mood at the test as when those events were (1997). Cognitive psychology and emotional disorders. New York:
Wiley.
encountered previously. However, MDM is properly
demonstrated with emotionally neutral materials to Paula Hertel
be learned and recalled. When the materials are not
inherently related to moodwhen MCM is not in-
volvedattempts to demonstrate MDM often do not
succeed. A consistent emotional state by itself is not ENGRAM
a strong basis for retrieving memories, particularly See: LOCALIZATION OF MEMORY TRACES
when other more obvious cues are available. Howev-
er, Eric Eich (1995) provides evidence that MDM is
a robust phenomenon when, like mood, the events to
be remembered and the cues for retrieving them arise EPISODIC MEMORY
from the individual rememberer, instead of being
provided by others. Psychologists have been studying memory experi-
mentally since Hermann Ebbinghauss (1885)
In conclusion, the most inclusive framework for groundbreaking work more than a hundred years
describing mood, emotion, and memory is one that ago, but only in the late twentieth century were ques-
emphasizes the nature and content of ongoing tions raised about exactly what has been and is being
thoughts. In all but the most transitory mood states, studied in memory experiments. As a result of the
the extent to which people think about their personal pursuit of these questions it became widely if not uni-
concerns is also the extent to which they remember versally accepted that there exist different kinds of
emotionally consistent experiences and fail to re- memory. Episodic memory is one of these kinds.
member experiences unrelated to their mood states. The term episodic memory is used in several senses.
One of these has to do with episodic memory as a par-
See also: ATTENTION AND MEMORY; IMPLICIT ticular class of laboratory tasks or experiments (Lock-
MEMORY hart, 2000); another concerns episodic memory as a
kind of mental capacity, or a neurocognitive system
(Schacter and Tulving, 1994), that allows people to
Bibliography remember past experiences. Although closely related,
Eich, E. (1995). Searching for mood dependent memory. Psycholog- the two senses (episodic tasks and the episodic system)
ical Science 6, 6775.
should not be confused. This entry will consider the
Eysenck, M. W., MacLeod, C., and Mathews, A. (1987). Cognitive
functioning in anxiety. Psychological Research 49, 189195. two senses in turn.
Healy, H., and Williams, J. M. G. (1999). Autobiographical memo-
ry. In T. Dalgleish and M. J. Power, eds., Handbook of cognition
and emotion. New York: Wiley.
Episodic Memory Tasks
Hertel, P. T. (2000). The cognitive-initiative account of depression- Episodic memory in the first sense manifests itself
related impairments in memory. In D. Medin, ed., The psychol- when a person remembers some information ac-
ogy of learning and motivation, Vol. 39. New York: Academic
quired on a particular occasion. Such situations occur
Press.
Hertel, P. T., and Hardin, T. S. (1990). Remembering with and frequently in real life where something happens at
without awareness in a depressed mood: Evidence for deficits one time (Time 1) and the individual who witnessed
in initiative. Journal of Experimental Psychology: General 119, the happening remembers it at a later time (Time 2).
4559. In the laboratory these situations are formalized as
Lyubormirsky, S., Caldwell, N. D., and Nolen-Hoeksema, S. tasks. A prototypical laboratory task of episodic
(1998). Effects of ruminative and distracting responses to de-
pressed mood on retrieval of autobiographical memories.
memory consists of 1. an original study experience
Journal of Personality and Social Psychology 75, 166177. during which individual items, such as words, are en-
MacLeod, A. K. (1999). Prospective cognitions. In T. Dalgleish and coded and stored by the learner (Time 1), and 2. a
M. J. Power, eds., Handbook of cognition and emotion. New York: subsequent test during which some aspect of the ex-
Wiley. perience is retrieved (Time 2). Episodic memory tasks
Mathews, A., Mogg, K., May, J., and Eysenck, M. (1989). Implicit
are sometimes also referred to as explicit memory
and explicit memory bias in anxiety. Journal of Abnormal Psy-
chology 98, 236240. tasks.
Parrott, W. G., and Sabini, J. (1990). Mood and memory under nat- Many variables affect performance on episodic
ural conditions: Evidence for mood incongruent recall. Jour-
memory tasks. They include ability differences
nal of Personality and Social Psychology 59, 321336.
Weingartner, H., Cohen, R. M., Murphy, D. L., Martello, J., and among subjects, the type of information presented for
Gerdt, C. (1981). Cognitive processes in depression. Archives study, the amount of time and effort devoted to learn-
of General Psychiatry 38, 4247. ing, subjects previous knowledge of the to-be-learned
136 EPISODIC MEMORY
material, the length of the retention interval between ined as existing beyond immediate perception, de-
study and test, and other such obvious factors. One pends on the integrity of the semantic memory system
important determinant of the rememberers perfor- and does not require episodic memory (Tulving,
mance in episodic tasks includes the way he or she 2001).
thinks about the material to be remembered as it is
studied, the so-called encoding operations. or coding The defining features of episodic memory (sys-
processes. Also important are the conditions under tem) are self, subjective time, and a special phenome-
which retrieval (recovery of stored information) oc- nal awareness of remembering, familiar to all, that is
curs or is attempted, and especially critical is the rela- referred to as autonoetic consciousness (Tulving,
tion between encoding and retrieval conditions (Tulv- 2001). Episodic memory is unique among other
ing, 1983). Performance on episodic memory tasks memory systems in that it alone allows the individual
can be measured in a variety of waysfree or non- to mentally travel through time, to remember the
cued recall, cued recall, free choice or forced choice past and to think about the future. The evidence in
recognition, frequency judgment, and recency judg- support of a separable episodic system steadily in-
ment, among others (Lockhart, 2000). Episodic mem- creases. One source of relevant observations is the
ory tasks in which both study and test (encoding and study of brain-damaged patients suffering from am-
retrieval) occur under fixed constant conditions are nesia. Some brain-damaged patients who suffer from
referred to as episodic memory tests and are widely a severe memory disorder and are severely impaired
used for the purpose of psychometric assessment of in or completely lack episodic memory are neverthe-
individuals episodic memory abilities. The informa- less capable of acquiring, even if laboriously, new fac-
tion that women outstrip men in episodic memory tual (semantic) knowledge (Hayman, MacDonald,
(Herlitz, Nilsson, and Bckman, 1997) and that the and Tulving, 1992; Kitchener, Hodges, and McCar-
earliest cognitive impairment in the functional devel- thy, 1998). This kind of dissociation between failure
opment of Alzheimers disease is episodic memory of remembering personally experienced events and
(Hodges, 2000), among other findings, is based on success of learning new facts implies that the neural
observations gleaned from episodic memory tests. substrate of episodic and other kinds of memories are
at least partially distinct. A related category of rele-
vant evidence is source amnesia: Individuals with im-
Episodic Memory System paired or frail memories, such as amnesic patients
and elderly people, can recall recently learned facts
The second sense of the term episodic memory is
better than they can recollect the episode in the
that of a hypothetical neurocognitive system that dif-
course of which they learned these facts (Shimamura
fers from the other major memory systems for which
and Squire, 1987).
evidence exists (Schacter and Tulving, 1994). These
other systems include semantic memory, procedural A second major source of evidence is electrophy-
memory, short-term memory (also known as working siological recording and functional neuroimaging of
memory), and the perceptual representation system brain activity that is correlated with memory process-
that subserves perceptual priming (also known as im- es. When episodic and semantic memory retrieval are
plicit memory). Episodic memory is most closely related compared, the findings show not only similarities in
to semantic memory, and the two are usually regard- brain activity but also differences (Dalla Barba et al.,
ed as subcategories of declarative memory. (Another 1998; Nyberg, 1999; Tulving et al., 1994).
closely related concept is autobiographical memory,
which refers to recollection and knowledge of signifi-
cant events from and facts about ones life.) The pos- Relation Between Tasks and System
tulation of the existence of separable memory systems The two senses of episodic memory (task and sys-
is part of the enterprise of the classification of natural tem) are related but they cannot be equated. The
phenomena of memory. Classification is a necessary main reason for this assertion lies in the fact that epi-
prerequisite for the study of memory mechanisms sodic tasks do not usually tap an individuals (au-
and processes. tonoetic) awareness of self-centered experiences of
The ability of an individual to consciously recol- the past but rather require only that the learner re-
lect personally experienced past events, that is, to be- produce or otherwise indicate his or her knowledge
come aware again at Time 2 of some aspect or some of the semantic contents of the learning episode. Epi-
part of a previous experience at Time 1, is possible sodic memory of course can, and usually does, greatly
only by virtue of an intact brain system specialized for contribute to this knowledge but it is not necessary for
that purpose, namely the episodic memory system. it. Putting it differently, episodic task performance
On the other hand, the ability to think about the does not only depend on the episodic system but can
world, and everything in it, which exists or is imag- be supported by nonepisodic systems as well. In the
EVOLUTION AND LEARNING 137
extreme case, individuals without or with severely im- Tulving, E. (1983). Elements of episodic memory. New York: Oxford
paired episodic memory system are capable of deal- University Press.
(2001). Episodic memory and common sense: How far
ing satisfactorily with many episodic memory tasks. In apart? Philosophical Transactions of the Royal Society of London,
the laboratory, when the rememberer has been ex- ser. B, 356, 1,5051,515.
posed to a set of study items at Time 1 and is then Tulving, E., Kapur, S., Craik, F. I. M., Moscovitsch, M., and Houle,
given a recognition test at Time 2, the rememberers S. (1994) Hemispheric encoding/retrieval asymmetry in epi-
performance is influenced both by recollection of sodic memory: Positron emission tomography findings. Pro-
ceedings of the National Academy of Sciences of the United States of
what happened at Time 1 (episodic system) and by America 91, 2,0162,020.
feelings that certain test items are familiar (other sys- Vargha-Khadem, F., Gadian, D. G., Watkins, K. E., Connelly, A.,
tems). An extensive and rapidly expanding literature Van Paesschen, W., Mishkin, M. (1997). Differential effects of
exists on the distinction between (episodic) remem- early hippocampal pathology on episodic and semantic mem-
bering and (nonepisodic) knowing as indicants of ory. Science 277, 376380.
processes involved in episodic tasks (Gardiner and Endel Tulving
Richardson-Klavehn, 2000). The feelings of familiari-
ty enable neurological patients and other organisms
whose episodic memory system is absent or severely
impaired to make correct discriminations between EVERYDAY MEMORY
previously encountered and previously nonencoun- See: NATURAL SETTINGS, MEMORY IN
tered test items in an episodic recognition task (Var-
gha-Khadem et al., 1997).
See also: AMNESIA, ORGANIC; CODING PROCESSES: EVOLUTION AND LEARNING
IMAGERY; CODING PROCESSES: LEVELS OF
PROCESSING; CODING PROCESSES: Learning is a biological adaptation, and like any
ORGANIZATION OF MEMORY; DECLARATIVE other adaptation is the outcome of evolution by natu-
MEMORY; FRONTAL LOBES AND EPISODIC ral selection. Because it is acted on by natural selec-
MEMORY; IMPLICIT MEMORY; WORKING tion, learning in different species of animals exhibits
MEMORY: HUMANS both descent with modification and specialized adap-
Bibliography tations. Many properties of learning, like the forma-
Dalla Barba, G., Parlato, V., Jobert, A., Samson, Y, Pappata, S.
tion of associations, are widely shared among ani-
(1998). Cortical networks implicated in semantic and episodic mals. The molecular mechanisms of learning are also
memory: Common or unique? Cortex 34, 547561. remarkably similar among animals as different as sea
Ebbinghaus, H. (1885). ber das Gedchtnis. Leipzig: Duncker and slugs, honeybees, and rats. But, in addition, learning
Humblot. exhibits specialized adaptations, modifications of
Gardiner, J. M., and Richardson-Klavehn, A. (2000). Remember-
ing and knowing. In E. Tulving and F. I. M. Craik, eds., The
learning which differ between species. Evolutionary
Oxford handbook of memory, pp. 229244. New York: Oxford adaptation in learning is usually investigated using
University Press. comparative methods to examine similarities or dif-
Hayman, C. A. G., MacDonald, C. A., and Tulving, E. (1993). The ferences among animals in how or what they learn.
role of repetition and associative interference in new semantic Learning can also have a reciprocal effect on the pro-
learning in amnesia. Journal of Cognitive Neuroscience 5, 375
389.
cess of evolution. Animals can learn to exploit new
Herlitz, A., Nilsson, L.-G., and Bckman, L. (1997). Gender differ- habitats and new resources within their habitat and
ences in episodic memory. Memory & Cognition 25, 801811. thus change the selective pressures they are exposed
Hodges, J. R. (2000). Memory in the dementias. In E. Tulving and to. Learning can even have an evolutionary impact
F. I. M. Craik, eds., The Oxford handbook of memory, pp. 441 that extends beyond the animal itself and affects
459. New York: Oxford University Press.
Kitchener, E. G., Hodges, J. R., and McCarthy, R. (1998). Acquisi-
other animals and plants it interacts with.
tion of post-morbid vocabulary and semantic facts in the ab- There are two basic requirements for evolution of
sence of episodic memory. Brain 121, 1,3131,327. a trait like learning by natural selection. First, the trait
Lockhart. R. S. (2000). Methods of memory research. In E. Tulving
and F. I. M. Craik, eds., The Oxford handbook of memory, pp. 45
must be at least partly heritable: Genotypic variation
57. New York: Oxford University Press. must produce phenotypic variation in the trait. Sec-
Nyberg, L. (1999). Imaging episodic memory: Implications for ond, variation in the trait must have an effect on re-
cognitive theories and phenomena. Memory 7, 585597. productive success. Learning meets both of these re-
Schacter, D. L., and Tulving, E. (1994). What are the memory sys- quirements.
tems of 1994? In D. L. Schacter and E. Tulving, eds., Memory
systems 1994, pp. 138. Cambridge, MA: MIT Press.
Shimamura, A. P., and Squire, L. R. (1987). A neuropsychological Genetic Variation in Learning
study of fact memory and source amnesia. Journal of Experi-
mental Psychology: Learning, Memory, and Cognition 13, 464 Some of the clearest evidence for genetic varia-
473. tion in learning comes from studies of learning muta-
138 EVOLUTION AND LEARNING
tions in the fruit fly Drosophila (Waddell and Quinn, discourage predation. Animals can clearly benefit
2001). Fruit flies are good learners, and can readily from learning which foods are edible and which are
learn to avoid an odor that has been associated with not, but the natural situation presents them with a
electric shock. Mutation of single genes in the Dro- problem. Toxic food may not have its effect for sever-
sophila genome can be induced with chemicals, and al hours after it was eaten. Animals usually have great
some of these mutations have dramatic effects on difficulty learning that two events are related if more
learning. Mutations in genes with whimsical names than a few minutes separates them in time. Experi-
like rutabaga and dunce make fruit flies unable to learn ments in the laboratory show that rats can associate
an association between odor and shock, or cause them illness with food even if the food was eaten several
to quickly forget what they have learned. These par- hours previously. Rats may require only a single expe-
ticular genes code for components of the intracellular rience with the food to form a strong aversion to it.
signaling system that transforms neural activity into Furthermore, rats associate the taste and odor of
more long-lasting changes in neurons that record ex- food, not its appearance, with illness. Selectivity in
perience. The rutabaga and dunce genes code for en- what is learned and the ability to associate events sep-
zymes that increase and decrease, respectively, the in- arated in time are distinctive features of taste-
tracellular concentration of cAMP (cyclic AMP), an aversion learning.
intracellular second messenger that responds to Songbirds exhibit specialized learning in the way
neurotransmitter signals received by a neuron. Other they acquire their songs. The songs that male Passer-
learning mutations have been discovered that affect ine birds use to advertise territory ownership and to
different aspects of the learning process, like volado, attract a mate are learned. The song-learning system
a gene that affects cell adhesion and influences com- possesses a number of unusual features. Most birds
munication between neurons. These mutations can learn only the songs of their own species, even if they
reveal a great deal about the molecular mechanisms are experimentally exposed to songs of other species
of learning and help unravel the neuroanatomy of for an equal period. In addition, the young of many
learning by pinpointing sites in the brain where ex- birds have a sensitive period during which they
pression of the gene makes a difference in learning. learn songs most readily. Songs are not learned by
But learning mutations are also important discoveries singing them. Instead, songs heard during the sensi-
for understanding evolution. They show that changes tive period are remembered until they are first sung
in genes (in some cases substitution of a single nucleo- many months later, when the breeding season begins.
tide) can affect the properties of learning. The phe- Finally, there are specialized nuclei in the avian brain
notypic effects of genetic variation of this kind pro- that are responsible for acquisition and production of
vide the raw material on which natural selection can song. Restrictions on what is learned, a sensitive peri-
act. od, separation in time of learning and performance,
and specialized neural structures make the song
learning system different from other kinds of learn-
Learning and Reproductive Success ing, but effective for memorizing and performing
Learning contributes directly and indirectly to re- species typical songs.
productive success in many ways. Bumblebees learn
how to obtain nectar from flowers. Colonial swallows
learn to recognize their young, and young herring Comparative Methods
gulls learn to recognize their parents. Most animals Comparative methods that have been used to ex-
must learn what is edible and what is not; others learn amine adaptation and evolutionary change in animal
migration routes, how to identify predators, how to physiology and anatomy can also be used to examine
defend a territory, and how to attract a mate. In all the evolution of learning. The clearest way to see the
of these cases, any heritable variant in learning that effect of evolution on the process of learning is by
makes the animal slightly more successful at the task comparing the ways in which different species of ani-
will make it more successful at reproducing itself, and mals learn. Closely related animals that share most of
hence more likely to pass on the variant in learning their evolutionary history but differ in some aspects
to its offspring. As a consequence of the action of nat- of their current behavior or ecology may differ in how
ural selection on learning, learning can differ be- they learn. Comparisons of learning in closely related
tween species and possess specialized adaptive prop- animals can thus reveal adaptive modifications of
erties that make learning more effective. Learned learning that are the result of recent selection. Anoth-
food aversions illustrate this kind of specialization. er approach is to compare animals that are not closely
Animals sometimes eat food that makes them ill, ei- related but share some current aspect of behavior or
ther because the food is contaminated or because the ecology. Similarities in learning between such animals
plant or animal they have eaten produces toxins to occur not because of a shared evolutionary history but
EVOLUTION AND LEARNING 139
because they have been exposed to similar selective Cricetidae) are polygynous: One male has several
pressures. If such animals show similarities in how mates. Male home ranges are larger than female
they learn or what they can learn it is likely to be the home ranges, and the home range of a polygynous
result of convergent evolution. These animals have male may encompass the home ranges of several fe-
independently evolved similar learning capacities be- males. Some species of vole, however, are monoga-
cause they have been exposed to similar selective mous, and male and female home ranges are of equal
pressures. Both kinds of comparison have been used size in these species. Laboratory experiments have
to identify the selective pressures that can influence found that males of polygynous species perform bet-
learning and to illuminate how evolutionary change ter on spatial memory problems than do females, but
in learning can occur. in monogamous species males and females perform
equally well. The sex difference in spatial ability in
Learning in food-storing birds illustrates how polygynous species is an adaptation to their breeding
these comparative methods can be used. Some spe- system and to the sex difference in home range size.
cies of birds, notably chickadees, nuthatches, and jays, As with food-storing birds, the consequences of natu-
store food. They make hundreds to thousands of food ral selection for learning ability can be seen in the
caches and return between several days and many brain. Lucia Jacobs and colleagues (1990) found that
months later to collect and consume the food they the hippocampus of male polygynous voles is larger
have hidden. Caches are widely scattered and contain than that of females, whereas in monogamous voles
only a few food items each. Remarkably, these birds there is no sex difference in the size of the hippocam-
remember precisely where they have placed each pus.
cache. Some birds in the jay family, the Corvidae, store
a great deal of food and some store little or none.
Comparisons among these closely related species of The Effect of Learning on Evolution
birds have shown that reliance on stored food is asso-
Evolutionary change occurs in learning, but
ciated with the level of performance on spatial tasks.
learning can, in turn, affect the course of evolution.
Ecological dependence on food storing has selected
Many species of animals do things that are culturally
for enhanced spatial ability. Food storing evolved in-
determined. Behaviors that are traditional within a
dependently in jays and in the chickadee family, the
population of animals are learned from other mem-
Paridae. Food-storing chickadees and food-storing
bers of the group, either directly or simply by associ-
jays both possess enhanced spatial abilities as an evo-
ating with other group members. Migration routes,
lutionary consequence of their shared reliance on
learned songs, and food preferences can all be trans-
stored food.
mitted culturally. Naive birds can learn to recognize
predators by participating in mobbing attacks on ani-
Learning and Evolution of the Brain mals that other members of the social group treat as
predators. The effect of such culturally transmitted
Evolutionary change in learning requires evolu- behavior on biological evolution is not fully under-
tionary change in the neural apparatus of learning. stood, though it is clear that learned behavior can
Evidence for evolutionary modification of learning consistently expose animals to new environments and
comes not only from observing differences between new sources of natural selection.
species in learning itself, but also from examining dif-
ferences between species in brain areas that are im- Learning can also have evolutionary effects that
portant for learning. The song control nuclei of spe- extend beyond the animal that does the learning.
cies of birds with large song repertoires are larger There is awe-inspiring diversity in protective mimicry
than the nuclei of birds with small repertoires. An in- among insects. The monarch butterfly contains toxins
crease in the size of neural structures that participate that make predators ill. The viceroy is a palatable but-
in learning has been found for a number of other terfly that mimics the monarch butterfly in appear-
kinds of learning. ance so closely that birds that have tasted a monarch
avoid both monarchs and viceroys. Many similar
The avian hippocampus plays an important role model and mimic systems are found in insects. Some
in memory, as it does in mammals, and the hippo- syrphid flies have evolved to closely resemble bees
campus of food-storing birds is over twice the size of and wasps in their appearance, posture, and behav-
the hippocampus of closely related birds that do not ior. They possess the distinctive black and yellow
store food (Sherry, 1998). Comparative studies of this banding pattern found on many bees and wasps, at
kind show that adaptive evolutionary change occurs rest they hold their forelegs in front of their head to
in brain regions involved in learning. A further exam- resemble wasp antennae, and their seasonal period of
ple illustrates that such adaptive change can occur activity coincides with that of their bee and wasp mod-
within a species. Most voles (rodents in the family els. These mimicry systems have evolved because the
140 EXPERTS MEMORIES
animals that would normally prey on the mimetic in- PARAHIPPOCAMPAL REGION; PROTEIN
sectsprimarily birdslearn to avoid the toxic, bad SYNTHESIS IN LONG-TERM MEMORY IN
tasting, or stinging model and because of its similar VERTEBRATES; SEX DIFFERENCES IN LEARNING;
appearance, also avoid the mimic. Without learning SPATIAL LEARNING: ANIMALS; TASTE AVERSION
by potential predators, there would be no mimicry. AND PREFERENCE LEARNING IN ANIMALS
As they gather nectar, insects such as honeybees, Bibliography

bumblebees, flies, and wasps carry pollen from one Jacobs, L. F., Gaulin, S. J. C., Sherry, D. F., and Hoffman, G. E.
flower to another and serve as the sole means of fertil- (1990). Evolution of spatial cognition: Sex-specific patterns of
spatial behavior predict hippocampal size. Proceedings of the
ization for many flowering plants. Among some polli-
National Academy of Sciences of the United States of America 87,
nators, such as bumblebees, different individuals 6,3496,352.
from the same colony learn a preference to visit one Jones, K. N. (2001). Pollinator-induced assortative mating: causes
kind of flower over others, a phenomenon remarked and consequences. In L. Chittka and J. D. Thomson, eds.,
on by Charles Darwin and known as constancy. Polli- Cognitive Ecology of Pollination. Cambridge, UK: Cambridge
University Press.
nators are probably constant because visiting the
Sherry, D. F. (1998). The ecology and neurobiology of spatial
same kind of flower makes it easier for them to recog- memory. In R. Dukas, ed., Cognitive Ecology. Chicago: Univer-
nize the flower and extract its nectar. This behavior sity of Chicago Press.
also affects the flowers, and indeed flowers have prob- Waddell, S., and Quinn, W. G. (2001). Flies, genes, and learning.
ably evolved to promote constancy because it in- Annual Review of Neuroscience 24, 1,2831,309.
creases the likelihood that pollen will be transferred David F. Sherry
to another flower of the same species. Constancy can
also have a further evolutionary effect (Jones, 2001).
Constancy by pollinators may promote speciation in
flowering plants by increasing assortative mating, the EXPERTS MEMORIES
tendency of similar members of a population to mate
with each other. If a population of flowering plants An expert is one who has acquired special skill in or
exhibits variation in the appearance or structure of its knowledge about a particular subject through profes-
flowers, constancy by pollinators will result in prefer- sional training and practical experience (Websters,
ential mating between flowers with the same structure 1976, p. 800). By that definition, experts will have a
and appearance, leading ultimately to the formation greater body of knowledge about their domain of ex-
of new species. pertise than other individuals. More remarkable is
the experts accurate memory for new experiences in
their domain. Some athletes can discuss minute de-
Conclusion tails of individual plays from games played years ago.
Because genetic variation can produce variation Expert chess players can readily recall chess positions
in the mechanisms of learning, and because learning from their matches in recent tournaments.
makes important contributions to the ability of ani- Early in the twentieth century many believed that
mals to reproduce themselves, learning evolves by experts were innately gifted with a superior memory.
natural selection. Many properties of learning are Numerous anecdotes attested to such amazing powers
shared among animals by virtue of their common de- of recollection. For example, Mozart was supposed to
scent. The formation of associations and some molec- be able to reproduce a presented piece of music after
ular mechanisms of learning are remarkably similar hearing it a single time. Later research, however, cast
across a wide range of animals. Evolutionary change doubt on the hypothesis of superior innate memory
in learning has also produced specialized adaptations in experts and has demonstrated that experts re-
of learning. Food-aversion learning, song learning, markable recall is limited to their specialties and
and cache retrieval in food-storing birds provide ex- arises from acquired skills and knowledge.
amples of such adaptive specialization. The effects of
evolutionary change in learning can also be observed
in brain areas that play important parts in learning. The Specificity of Experts Superior
Not only has evolution affected learning, but learning Memory
can affect evolution, both by exposing animals to se- The most influential research on experts memo-
lective pressures they would not otherwise encounter ries focused initially on chess masters superior recall
and by causing evolutionary change in the animals of board positions (Chase and Simon, 1973). Chess
and plants with which they interact. players ranging from beginners to international mas-
ters were shown a position from an actual chess game
See also: BIRDSONG LEARNING; GUIDE TO THE (such as the one illustrated in panel A of Figure 1) for
ANATOMY OF THE BRAIN: HIPPOCAMPUS AND a brief time (normally five seconds) and then asked to
EXPERTS MEMORIES 141
Figure 1
Standard diagrams of an actual chess position (Panel A) and a chessboard with randomly arranged pieces (Panel B). A nonstandard
representation of the same information using the first letter of the names of the pieces is shown in Panels C and D.
recall the location of all the chess pieces. The ability pieces (as illustrated in Panel B of Figure 1). With
to recall increased as a function of chess skill. Begin- briefly presented random chessboards, players at all
ners at chess were able to recall the correct location levels of skill had roughly the same poor recall perfor-
of about four pieces, whereas international-level play- mance and were able to recall the correct location of
ers recalled nearly all of the more than twenty pieces. only three to five pieces on the averagea perfor-
mance comparable to that of chess beginners for actu-
To rule out the innateness of the chess masters al positions from chess games. Further, Chase and
superior visual memory, Chase and Simon had chess Simon showed that when an actual chess position was
players recall chessboards with randomly placed shown using an unfamiliar notation (see Panel C in
142 EXPERTS MEMORIES
Figure 1), the chess expert was able to display a simi- and 1945 being the year of the end of World War II.
lar level of superior memory performance after a Through extended training individuals can acquire
brief period of adjustment. This result implies that memory skills allowing them to increase their memo-
the superior memory of experts is not innate but rath- ry of briefly presented lists of numbers from an initial
er a function of learned skills. Since Chase and level of seven digits to over eighty random digits.
Simons classic study, other investigators have arrived Hence, it is possible for regular college students to at-
at similar findings for experts in other fields such as tain exceptional memory performance after 50 to 200
computer programming, basketball, and dance (Er- hours of practice. Laboratory studies of individuals
icsson, Patel, and Kintsch, 2000). with exceptional memory performance for numbers,
names, and pictures reveal that they rely on acquired
memory skills that often involve some kind of mne-
The Role of Meaningful Relations in monics (Ericsson and Lehmann, 1996; Wilding and
Superior Memory Performance Valentine, 1997).
Without the expertise of a master, it is nearly im-
possible to grasp the meaningful relations between How Superior Memory of Experts Mediates
chess pieces perceived by the expert in panels A and Their Superior Performance on
C of Figure 1. If, on the other hand, the availability
Representative Tasks
of knowledge providing meaning to a stimulus is criti-
cal to superior memory, it should be possible to dem- The primary goal for all experts is to excel at the
onstrate the same effect in a domain where all adults demands of their fields. For example, chess experts
are proficient, such as language. Human adults are need to find the best moves to win chess matches, and
able to recall verbatim meaningful sentences of twen- medical experts have to diagnose sick patients in
ty or more words after a brief presentation (Chase and order to give them the best treatment. Unlike the
Ericsson, 1982). An example of such a sentence would memory experts who attempt to improve their mem-
be, The woman in front of him was eating peanuts ory performance by acquiring mnemonic techniques
that smelled so good that he could barely contain his through extended practice, chess experts and medi-
hunger. If the words of the sentence are randomly cal doctors do not deliberately train their memory.
rearranged in a manner analogous to that used in Their superior memory ability must thus be a bypro-
Chase and Simons procedure for generating random duct of their improved performance on representa-
chessboards, accurate verbatim recall drops to tive tasks (Vicente and Wang, 1988). Furthermore,
around six words. An example of a random re- experts appear to store task-relevant information in
arrangement of the above sentence would be, Was memory when they normally perform representative
smelled front that that his the peanuts he good hun- tasks in their domain, because, if they are unexpect-
ger eating barely woman of so in could that him con- edly asked to recall information about a performed
tain. task, their memory is typically far superior to that of
less skilled individuals.
For random lists of words, the recall of subjects
In fact, experts incidental memory of the rele-
is limited by the small number of words they can keep
vant information is frequently so good that instruct-
rehearsing, and once they stop rehearsal, the words
ing them to intentionally memorize the information
are quickly forgotten. In contrast, once meaningful
does not reliably improve their memory. For exam-
sentences are understood, their meaning is well re-
ple, when chess experts analyze a position to find the
tained in long-term memory. For example, during
best move, their memory of the position is just as
normal comprehension of a text, the essential infor-
good whether they were informed about an upcoming
mation in each sentence is efficiently stored in memo-
memory test or not. As part of performing the repre-
ry so it can be integrated with related information
sentative task of selecting the best move, the experts
presented later in the text (Ericsson and Kintsch,
encode the important features of the presented infor-
1995).
mation and store them in an accessible form in mem-
Stimuli from an unfamiliar domain of expertise, ory. In contrast, when subjects, after training based
such as diagrams of chess positions and medical on mnemonics and knowledge unrelated to chess, at-
terms, are about as meaningless to most adults as ran- tain a recall performance comparable with that of the
dom lists of words and digits. Recent studies have chess experts, they still lack the ability to extract the
shown that memory for meaningless information can information important for selecting the best move.
be dramatically improved through training by active- Hence, the remarkable characteristic of expert mem-
ly seeking out meaningful associations for the mean- ory is not just the amount recalled, which can often
ingless material. For example, the sequence 671945 be matched by training, but the rapid extraction and
can be remembered as 67 being the retirement age storage of important patterns and relevant informa-
EXPLICIT MEMORY 143
tion that allows superior execution of the representa- ciated memory skills. These skills are not attained au-
tive task (Ericsson, Patel, and Kintsch, 2000). tomatically with experience but require engagement
in deliberate practice that is often designed by teach-
An analysis of expert performance shows that it
ers. Even the most talented individuals have spent
is not sufficient to have merely stored the knowledge
around ten years of intense preparation before attain-
in memory; it is also critical that the relevant knowl-
ing a world-class level of performance in many do-
edge be well organized and easily retrievable. In fact,
mains such as sports, chess, and arts (Ericsson, 1996).
the principal challenge of expertise is to acquire and
organize the vast body of domain knowledge (Chi, See also: NATURAL SETTINGS, MEMORY IN
Feltovich, and Glaser, 1981) such that all relevant
prior knowledge can be immediately accessed to Bibliography
guide action in encountered situations. For example, Charness, N. (1989). Expertise in chess and bridge. In D. Klahr
and K. Kotovsky, eds., Complex information processing: The im-
with his or her superior organization of knowledge, pact of Herbert A. Simon. Hillsdale, NJ: Erlbaum.
a chess expert can rapidly perceive a promising move, Chase, W. G., and Ericsson, K. A. (1982). Skill and working memo-
or a medical expert can rapidly notice an inconsisten- ry. In G. H. Bower, ed., The psychology of learning and motiva-
cy in a suggested diagnosis. tion, Vol. 16. New York: Academic Press.
Chase, W. G., and Simon, H. A. (1973). The minds eye in chess.
Efficient and reliable storage of relevant informa- In W. G. Chase, ed., Visual information processing. New York:
tion in memory is especially important to experts Academic Press.
when they engage in planning and complex reason- Chi, M. T. H., Feltovich, P. J., and Glaser, R. (1981). Categoriza-
tion and representation of physics problems by experts and
ing that mediate their superior performance. During novices. Cognitive Science 5, 121152.
planning experts have to mentally compare many al- Ericsson, K. A., ed. (1996). The road to excellence: The acquisition of
ternative sequences of actions, storing a great deal of expert performance in the arts and sciences, sports, and games. Mah-
information in working memory. Consequently, be- wah, NJ: Erlbaum.
ginning chess players do not generate long plans, and Ericsson, K. A., and Kintsch, W. (1995). Long-term working memo-
ry. Psychological Review 102, 211245.
it takes years of chess study before chess experts are Ericsson, K. A., and Lehmann, A. C. (1996). Expert and exception-
able to plan long sequences of future moves reliably al performance: Evidence on maximal adaptations on task
(Charness, 1989). Chess masters eventually improve constraints. Annual Review of Psychology 47, 273305.
their memory skills for planning so much that they Ericsson, K. A., Patel, V. L., and Kintsch, W. (2000). How experts
are even able to play chess without seeing the chess- adaptations to representative task demands account for the
expertise effect in memory recall: Comment on Vicente and
board (blindfold chess). Analyses of the superior abili- Wang (1998). Psychological Review 107, 578592.
ty to plan suggest that experts acquire memory skills Websters Third New International Dictionary (1976). Springfield, MA:
that allow them to rely on long-term memory for stor- Merriam.
age of generated information (Ericsson and Kintsch, Wilding, J., and Valentine, E. (1997). Superior memory. Hove, UK:
1995). Research on expertise is making it increasingly Psychology Press.
clear that the vast knowledge of experts has to be well K. Anders Ericsson
organized and supplemented with special memory
skills so as to support memory-demanding planning,
design, and reasoning.
EXPLICIT MEMORY
Further research has revealed the complex and
intricate structure of expert performance and its asso- See: DECLARATIVE MEMORY
F
FALSE MEMORIES sociations, or inferences cause them to misremember
the past, whereas in the latter case, the false memories
False memories may be full-blown memories of events arise from someone elses overt suggestion or mis-
that were never experienced or (perhaps more com- leading statements.
monly) memories that are distorted (i.e., the event
one is remembering actually occurred, but it did not
occur in the way that is being recalled). Even though False Memories Arising from Internal
memory can foster an illusion of reliving an experi- Processes
ence, it is actually a reconstruction and hence subject
to departures from objective facts. This entry focuses In everyday conversation, listeners often make in-
on false episodic memories, or inaccurate memories ferences that stretch the meaning of the speakers ex-
of episodes in ones past, which can be distinguished plicit words. For example, if a colleague told you that
from false semantic memories, which include inaccu- his infant had stayed awake all night, you might
rate knowledge (e.g., erroneously believing that the infer that the baby had cried. Such inferences often
capital of Russia is St. Petersburg). insinuate themselves into memory. Indeed, when
asked later, one might be likely to recall the statement
For example, when conveying anecdotes in casual as having been the infant cried all night (Brewer,
social interactions, people sometimes embellish them 1977; Bransford and Franks, 1971).
to make them more interesting, often spicing them The literature on the role of schemas (or general
with fresh details in subsequent retellings to assure world knowledge) on memory also sheds light on the
the desired pungency. Although innocent in intent, influence that inferences can have on memory. This
such embellishments can actually alter the tellers own work is rooted in studies by Bartlett (1932), who dem-
memory of the event. Even though the raconteur onstrated that when English students were presented
might be fully aware of the fictional enhancements at with an American Indian folktale that they found dif-
the time, he or she may in time come to think of them ficult to comprehend, the flaws in their memories of
as actual components of the original event (Tversky the folktale often betrayed British cultural influences.
and Marsh, 2000).
More recent experimental investigations into in-
We distinguish here between two broad classes of ternally generated false memories include studies in
episodic false memories: those that arise from inter- which people are given short lists of about fifteen re-
nal processes (e.g., the example regarding embellish- lated words to remember (e.g., bed, rest, awake, tired,
ments) and those that arise from external events (e.g., dream). When given an immediate free recall test after
from hearing other peoples erroneous accounts of an such a list (and told to recall every word that is re-
event). In the former case, peoples own thoughts, as- membered in any order but without guessing), people
145
146 FALSE MEMORIES
often recall sleep, a related (but not presented) word markedly as a function of the verb used; when the
(Roediger and McDermott, 1995). This approach, more dramatic verb smashed was invoked, the average
which allows the rapid implanting of numerous mini estimated speed was forty-one miles per hour, where-
false memories, enables researchers to manipulate as the verb contacted elicited an average estimate of
various independent and subject variables in order to only thirty-two miles per hour. Even more amazing
observe their effects on false recall (and false recogni- was that the wording of this single question influ-
tion) probabilities (Roediger and McDermott, 2000). enced peoples memories even a week later when they
This work shows not only that people recall and rec- were asked, Did you see any broken glass? Subjects
ognize the nonpresented, related words but that they were more likely to erroneously recollect broken glass
also claim to remember the precise moment of pre- if they had encountered the verb smashed a week earli-
sentation of these (nonpresented) words. In addition, er (relative to the verb hit).
the forgetting function for the related, nonpresented Work within this tradition is often referred to as
words is less steep than the forgetting function for the misleading-information paradigm or sometimes
studied words. the eyewitness-memory paradigm. Similar findings
Another recent line of research has investigated with respect to the role of intervening suggestions on
the role that imagination can play in distorting mem- peoples memories have been demonstrated for po-
ory. The mere act of imagining an event can inflate lice lineups among other domains. Elizabeth Loftus
the probability that a person will come to have a full- combined this procedure with the imagination-
blown recollection of the (nonexistent) prior event. inflation procedures in a case study in which she cre-
This phenomenon has been dubbed imagination in- ated a full-blown memory of being lost in a shopping
flation (Garry, Manning, and Loftus, 1996; Goff and mall in a teenage boy (Chris) who was never actually
Roediger, 1998). lost in a mall (Loftus, 1993). Loftus prompted the
false memory by having Chriss brother suggest the
Not only can imagining an event that did not pre- incident to Chris, complete with specific details. Two
viously occur create memories for that event, but also weeks after the initial description of the nonevent,
describing an event that did indeed occur can color Chris was able to recall minute details from this in-
memory for that event. For example, if expert wine cident, including the balding head and the kind of
tasters describe a wine they just sampled, it does not eyeglasses worn by the man who rescued him. Ira
change their memory for the wine; if, however, inter- Hyman and his colleagues have performed systematic
mediate-level wine tasters attempt to describe the studies of this type and explored individual differ-
wine just enjoyed, the descriptions skew their later ences among people that influence the likelihood of
memory of the wine (Melcher and Schooler, 1996). such false memories (Hyman and Billings, 1998).
This interference from attempting to verbalize an ex-
perience that is not readily amenable to accurate ver-
bal description has been termed verbal overshadow- Processes That Give Rise to False
ing by Jonathan Schooler. Memories
Many theoretical perspectives have been applied
to the study of false memories. We focus here primari-
False Memories Arising from External ly on the Source Monitoring Framework, which has
Factors been espoused by Marcia Johnson and her colleagues
Some of the best-known false memory work can (Johnson, Hastroudi, and Lindsay, 1993; Johnson
be considered adaptations of the classic studies of re- and Raye, 1981). Accurate memory requires disentan-
troactive interference, in which a subsequent event gling recollection of events from speculations, infer-
can interfere with memory for a similar, prior event ences, and imaginings. Achieving the seemingly sim-
(McGeoch, 1932). In its more modern manifestation, ple goal is easier in theory than in practice. Simply
subjects might be presented with a slide show or vid- asking people to focus carefully on whether some-
eotape depicting a car crash and later be exposed to thing was experienced or only imagined or thought
misleading information about this event either and to be sure to recall only what overtly occurred
through a narrative, suggestive questions, or both. In (and not what they inferred or thought) is not suffi-
a classic study by Loftus and Palmer (1974), such a cient to avoid false memories and may even exacer-
crash was followed by a questionnaire asking people bate them in some situations (Hicks and Marsh,
a series of questions about the crash. The critical ma- 2001). Telling people before an encoding phase that
nipulation was a single verb in one of the questions: they might be misled and that they should encode the
contacted, hit, bumped, collided, or smashed. That is, peo- information carefully so as not to confuse their
ple were asked, How fast were the cars going when thoughts with the overt event may aid them somewhat
they ___ into each other? Speed estimates varied but is by no means sufficient to eliminate later false
FOOD AVERSION AND PREFERENCE LEARNING IN HUMANS 147
memories. Some research has shown that, relative to Loftus, E. F. (1993). The reality of repressed memories. American
young adults, old adults have more difficulties in Psychologist 48, 518537.
Loftus, E. F., and Palmer, J. C. (1974). Reconstruction of automo-
monitoring the retrieval process in order to avoid bile destruction: An example of the interaction between lan-
false memories. guage and memory. Journal of Verbal Learning and Verbal Be-
havior 13, 585589.
McGeoch, J. A. (1932). Forgetting and the law of disuse. Psychologi-
Practical Implications cal Review 39, 352370.
The fallibility of memory has become a conten- Melcher, J. M., and Schooler, J. W. (1996). The misremembrance
tious subject not only in psychological theory, but also of wines past: Verbal and perceptual expertise differentially
mediate verbal overshadowing of taste memory. Journal of
as a result of its practical implications in the legal sys- Memory and Language 35, 231245.
tem, where the reliability of eyewitness accounts has Roediger, H. L., and McDermott, K. B. (1995). Creating false
come increasingly into question. Several conclusions memories: Remembering words not presented in lists. Journal
can be safely reached from this research, however. of Experimental Psychology: Learning, Memory, and Cognition 21,
Perhaps the most important point is that a full-blown, 803814.
(2000). Tricks of memory. Current Directions in Psychological
vivid recollection of a prior event is not diagnostic of Science 9, 123127.
its prior occurrence; it is perfectly possible to vividly Tversky, B., and Marsh, E. (2000). Biased retellings of events yield
recollect an event that was only previously imagined biased memories. Cognitive Psychology 40, 138.
or thought about. The stories we tell ourselves and Kathleen B. McDermott
others color our memory for the object of the story. Jason C. K. Chan
In this vein, retrieval has been described as a double-
edged sword: It helps us remember what occurred
previously (the testing effect), but it also can distort
memory. Simple instructions to try to avoid false
memories are often insufficient to do so. FOOD AVERSION AND PREFERENCE
Finally, memorys reconstructive nature might be LEARNING IN HUMANS
considered a cognitive asset rather than a drawback.
Most of our misguided recollections are fairly harm- To survive, animals must select, from among myriad
less, and many inferences about anothers conversa- nonnutritive and toxic items they could ingest, those
tional intent are probably correctin the foregoing few that are both nutritious and relatively toxin-free.
example, the baby probably was crying all night. Only Humans are, of course, animals, and many of the be-
in the high-stakes atmosphere of, say, the courtroom havioral processes that guide the food choices of
or the police lineup does it become critical to disen- other animals influence humans food choices as well.
tangle the wheat of accurate memory from the chaff However, diet selection by humans is unusual in at
of imagination, inference, conjecture, and embellish- least two ways. First, most human knowledge about
ment. foods comes secondhand, either directly or indirectly
from others. Second, the feeding environment of hu-
See also: RECONSTRUCTIVE MEMORY mans living today in the developed world is dramati-
Bibliography cally different from that in which humans evolved
Bartlett, F. C. (1932). Remembering: A study in experimental and social their abilities to choose foods. We experience food ex-
psychology. New York: Macmillan. cess, rather than food shortage, extraordinary variety
Bransford, J. D., and Franks, J. J. (1971). The abstraction of lin- in available foods, rather than restricted food choices,
guistic ideas. Cognitive Psychology 2, 331350. and we are exposed to foods with artificially enhanced
Brewer, W. F. (1977). Memory for the pragmatic implications of palatability. Consequently, our evolved mechanisms
sentences. Memory & Cognition 5, 673678.
Garry, M., Manning, C. G., and Loftus, E. F. (1996). Imagination of food choice, selected for in widely different circum-
inflation: Imagining a childhood event inflates confidences stances, may sometimes prove maladaptive in the
that it occurred. Psychonomic Bulletin & Review 3, 208214. modern world.
Goff, L. M., and Roediger, H. L., III. (1998). Imagination inflation
for action events: Repeated imaginings lead to illusory recol-
lections. Memory & Cognition 26, 2033. Dietary Specialists and Dietary Generalists
Hicks, J. L., and Marsh, R. L. (2001). False recognition occurs
more frequently during source identification than during old- Solving the problem of diet selection is relatively
new recognition. Journal of Experimental Psychology: Learning, simple for animals that eat only one food. Such ani-
Memory, and Cognition 27, 375383. mals tend to evolve sense organs that identify the
Hyman, I. E., and Billings, F. J. (1998). Individual differences and chemical signature of whatever species they find edi-
the creation of false childhood memories. Memory 6, 120. ble. For the tobacco hornworm, as its name implies,
Johnson, M. K., Hashtroudi, S., and Lindsay, D. S. (1993). Source
monitoring. Psychological Bulletin 114, 328. leaves of the tobacco plant are food, and the worms
Johnson, M. K., and Raye, C. L. (1981). Reality monitoring. Psycho- taste receptors are particularly sensitive to chemicals
logical Review 88, 6785. found in tobacco leaves.
148 FOOD AVERSION AND PREFERENCE LEARNING IN HUMANS
For dietary generalistsanimals that, like hu- affective response to a stimulus (a conditioned stimu-
mans, compose a diet consisting of many different lus, or CS) is changed as a result of pairing with either
foodsthere is no chemical signature that allows dis- a liked or disliked stimulus (an unconditioned stimu-
crimination of food from nonfood items. Dietary gen- lus, or UCS). In the case of taste-aversion learning, if
eralist have inherent sensory-affective systems biasing an animal such as a rat or a human eats a relatively
them to ingest substances with certain tastes or smells; unfamiliar food and, within a few hours, becomes
at birth, human infants like the taste of sugar and re- nauseous, the sick individual will develop a distaste
ject the bitter of quinine and the sour of lemons. How- for the smell and taste of the food ingestion of which
ever, dietary generalists still must learn which specific preceded illness. Such taste-aversion learning, re-
items to ingest and which to avoid eating. flecting a change in affective response to a food, can
seem irrational, occurring even if the sick individual
Learning to eat nutritious foods while avoiding
knows that the food did not cause the nausea (for
toxic or worthless potential foods is especially difficult
example, the nausea might clearly be a symptom of
because effects of toxins and nutrients often occur
the flu).
long after their ingestion. Consequently, many ani-
mals, humans included, have evolved a special type of Taste-aversion learning differs from situations in
conditioning, called taste-aversion learning (discussed which ingestion of a food is followed by negative ef-
below as a type of evaluative conditioning), allowing fects other than nausea: for example, hives or respira-
them to bridge the temporal gap between ingesting tory distress. In the latter case, people can learn that
an item and experiencing consequences of its inges- a potential food is dangerous and should not be
tion. eaten. However, the taste of the food does not become
unpleasant, and the victim of an allergic reaction may
continue to want to eat the food causing distress, but
Human Food Rejection and Acceptance will avoid doing so from fear of the consequences. For
Humans reject potential foods for one or more of example, a person who eats shrimp and becomes nau-
four reasons. They may find a food distasteful, reject- seated tends thereafter to dislike shrimp and may
ing it because it has undesirable sensory properties. find even the smell of shrimp distasteful, whereas
Alternatively, a food may be rejected because it is per- someone who experiences respiratory distress after
ceived as dangerous, for instance, as causing illness eating shrimp will avoid eating shrimp but may still
such as allergic reaction. A potential food may also be like their taste and smell. Nausea serves as a special
rejected because it is viewed as inappropriate, as, for UCS that, when paired with a food, even once and
example, is dirt. Last, some foods may not be eaten with a lengthy delay between CS and US, often pro-
because they seem disgusting, as with rotting meat, duces distaste.
which is viewed as disgusting by members of some cul- In humans as in other animals, enhancement of
tures, but not others. liking for a neutral flavor can occur if it is paired with
a desirable flavor, and pairing a neutral flavor with in-
There are only two categories of accepted items:
troduction of nutrients into the stomach also can in-
those that taste good and, like diet soda, are con-
crease liking for the previously neutral flavor. Howev-
sumed because of their sensory properties, and those
er, pairings of flavors with calories or good tastes
consumed because they are believed to produce posi-
usually has modest effects in comparison with pair-
tive consequences, as are health foods, and medi-
ings of flavors with nausea.
cines. Many accepted items have both properties.
Social Influences on Human Food Choices

Effects of Exposure and Conditioning
In humans, social forces account for many food
Generally, previous exposure of either humans or preferences and aversions. Approval or disapproval
other animals to a food without obvious positive or of foods by respected others seems to influence ones
negative consequences (mere exposure) tends to own response to those foods, though the way in which
increase liking for that food. On the other hand, a such change occurs is not well understood. The pro-
great deal of exposure to a food in a brief period can cess of change could be cognitive, could involve social
produce a temporary decline in liking labeled sensory- learning, or could be a form of as-yet unexplored so-
specific satiety. Too much of even a good thing can pro- cial, classical conditioning. In the last case, displays of
duce temporary avoidance of it. pleasure, displeasure, or disgust by another could be-
A further means of changing response to food come associated with a flavor, changing affective re-
preference involves a form of classical conditioning sponse to it.
called evaluative conditioning, of which taste-aversion We do know that children show increased liking
learning is one example. In evaluative conditioning, for foods associated with positive displays by signifi-
FORAGING 149
cant others and that such socially induced changes in teaching, social learning, and cuisine may be in shap-
food preference can last for months. Changes in af- ing human food choices, other factors, as yet poorly
fective response to foods seem to occur when children understood, play a major role in shaping the dietary
do not feel forced or bribed to consume a food: repertoires of humans.
when children are rewarded for eating a food, they do
not tend to like it more. However, when the same See also: TASTE AVERSION AND PREFERENCE
LEARNING IN ANIMALS
food is either used as a reward or is seen to be enjoyed
by others, it does become liked more. Social factors Bibliography
are almost surely also involved in the development of Birch, L. L., Fisher, J. O., and Grimm-Thomas, K. (1996). The de-
disgust responses, as when children observe negative velopment of childrens eating habits. In H. J. H. Macfie and
responses of their parents to finding half a worm in H. L. Meiselman, eds., Food choice, acceptance and consumption,
a half-eaten apple or to body wastes. However, this pp. 161206. Glasgow: Blackie Academic and Professional.
Booth, D. A. (1994). Psychology of nutrition. London: Taylor and
process has not been investigated. Francis.
A distinctive feature of the human diet is that Capaldi, E. D., ed. (1996). Why we eat what we eat: The psychology of
eating. Washington, DC: American Psychological Association.
many foods that are liked by some humans have sen-
De Hower, J., Thomas, S., and Baeyens, F. (2001). Associative
sory properties that are inherently aversive to both learning of likes and dislikes: A review of twenty-five years of
other humans and other animals. Members of many research on human evaluative conditioning. Psychological Bul-
cultures like bitter substances such as coffee, quinine letin 127, 853869.
water, and tobacco as well as irritants such as chili Galef, B. G., Jr. (1996). Food selection: Problems in understanding
how we choose foods to eat. Neuroscience and Biobehavioral Re-
pepper and horseradish, substances that animals, in-
views 20, 6773.
fant humans, and adults from some other cultures (1996). Social enhancement of food preferences in Norway
find aversive. Such preferences may be learned in so- rats: A brief review. In C. M. Heyes and B. G. Galef Jr., eds.,
cial settings, although we do not really know how. Social learning in animals: The roots of culture, pp. 4964. San
Diego: Academic Press.
Humans clearly differ from all other animals in Garcia, J., Hankins, W. G., and Rusiniak, K. W. (1974). Behavioral
the importance of cuisine (defined here as a system for regulation of the milieu internal in man and rat. Science 185,
selecting, processing, combining, and flavoring foods 824831.
that incorporates the nutritional wisdom of past gen- Macfie, H. J. H., and Meiselman, H. L. (1996). Food choice, accep-
tance, and consumption. Glasgow: Blackie Academic and Profes-
erations) in their food selection. Although social sional.
learning affects food choices of nonhuman animals Rozin, P. (1999). Food is fundamental, fun, frightening, and far
(for example, after an observer rat interacts with a reaching. Social Research 66, 930.
demonstrator rat that has eaten a food, the observ- Rozin, P., and Shulkin, J. (1990). Food selection. In E. M. Stricker,
er prefers the food its demonstrator ate), social influ- ed., Handbook of behavioral neurobiology, Vol. 10: Neurobiology of
food and fluid intake, pp. 297328. New York: Plenum Press.
ences on food choice are neither as pervasive nor as
long lasting in nonhuman animals as in humans. Paul Rozin
Revised by Bennett G. Galef Jr. and Paul Rozin
Young humans are also probably the only animals
explicitly taught what to eat and what to avoid eating,
although some evidence suggests that chimpanzees
may have rudimentary abilities to instruct their young
about foods. Still, only humans learn socially to give FORAGING
foods emotional, social, and moral values, and only
humans learn about nutritive values, appropriate Foraging, the search for food, is a fundamental part
times for ingestion, and means of preparation of of behavior. All animals, from the simplest inverte-
foods in a manner perhaps best described as social- brates to primates, have to take in food. Because ap-
cognitive, just as they learn about other aspects of the propriate food may be more abundant at some times
physical environment. and places than others, an animal that can learn
about the characteristics of its food supply is likely to
be able to forage more efficiently than one that can-
Conclusion not learn. Indeed, the need for efficient foraging
Unfortunately, not enough is yet known about the creates a strong selection pressure for the evolution
development of food preferences in the human spe- of learning and memory.
cies to give much helpful advice to parents wishing to Since the late twentieth century, the study of for-
modify a childs food choices. Indeed, one of the most aging behavior has been guided by optimal foraging
surprising facts known about human food choices is theory, a body of mathematical models specifying how
that there is little similarity between food preferences animals should behave so as to maximize foraging ef-
of parents and their mature children. Important as ficiency. After briefly introducing this framework, this
150 FORAGING
entry describes some of the ways in which animals use On each trip it had to remember where it was in rela-
learning and memory in foraging. tion to the nest. The birds also had to learn how often
worms were available at the experimental feeder and
how valuable they were.
Optimal Foraging Theory
Optimal foraging theory is a topic in behavioral
ecology, the field of biology dealing with how behav- Prey Selection
ior contributes to an animals reproductive success or An animal encountering a potential prey item
fitness. Many aspects of foraging can be understood may accept it or go on searching for alternative prey.
by assuming that animals have evolved to maximize What it should do to maximize its rate of energy in-
the rate at which they take in energy while foraging. take can be understood intuitively. If it can expect to
An animal that can forage efficiently will have more find a bigger or more quickly consumed item soon
time for other important activities like finding a mate enough, the forager should reject the item at hand
or defending a territory. If the economics of a particu- and go on searching; otherwise it should take the en-
lar foraging situation are understood well enough, it countered item. To select prey as efficiently as they
is possible to make a mathematical model that speci- do, animals must learn about their value and their
fies what the animal should do in order to maximize abundance and adjust their behavior as the environ-
its energy intake while foraging. Stephens and Krebs ment changes. Many of the studies of prey selection
(1986) describe this approach in detail. However, and other aspects of foraging reviewed by Shettle-
some examples are easy to understand intuitively worth (1998) have emphasized how the learning
without any mathematics. mechanisms animals use in foraging are the same as
Consider a small bird in the spring collecting those revealed in experiments on operant condition-
food to bring back to the young in its nest. To feed ing.
the hungry nestlings it must spend a good part of the
day searching for food and carrying it back home.
Learning to Find Cryptic Prey
How far should it travel and how much should it col-
lect on each trip? It might seem obvious that the bird Many animals that are potential prey for other
should load up as much as it can each time, but this animals have evolved to look like their surroundings
suggestion overlooks the fact that as the bird loads its so they are hard for predators to see. For example,
beak with food like grubs or caterpillars, increasing moths may be speckled black and gray like the bark
the load becomes harder and harder. In addition, of the trees where they rest, and caterpillars that live
more energy is needed to fly back to the nest with on green plants may be green. In turn, predators
more prey items. On the other hand, if the bird has have evolved the ability to learn how to discriminate
had to fly some distance from the nest in order to find such cryptic prey from their backgrounds. Laboratory
suitable prey, it is worth its while to collect as many studies using bluejays, chicks, and pigeons searching
items as possible. This informal argument suggests for grains or for images on slides under controlled
that there should be a direct relationship between the conditions have provided evidence that predators
size of the birds load and the distance it has traveled: may form a specific search image for, or learn to
Bigger loads should be collected when the bird is far- see, cryptic prey. When a bird encounters several
ther from the nest. cryptic prey items in a row, it becomes better at de-
tecting them. It may be paying more attention to sub-
Kacelnik and Cuthill (1987) studied this problem tle details that differentiate the prey items from their
of central-place foraging with starlings nesting background, or it may be learning to search more
around a farm. They trained the birds to visit a feeder slowly when prey are difficult to detect. Both kinds of
and collect mealworms that the experimenter learning probably contribute to improving foraging
dropped down a pipe. By placing the feeder at differ- efficiency. Using a computerized virtual ecology in
ent distances from the starlings nests while keeping which bluejays search for moths, Kamil and Bond
constant the rate of dropping mealworms, Kacelnik (2001) have shown how the birds learning can con-
and Cuthill were able to obtain clear evidence of the tribute to the evolution of one prey type rather than
predicted relationship. With further experiments in another.
both the field and the laboratory they were able to ac-
count for many details of what the birds learn and re-
member. Learning about Patches of Food
Implicit in this example are a number of uses of Not only do animals have to detect and capture
learning and memory. To return straight home with prey efficiently, they have to learn where prey can be
its prey, a starling had to learn the location of its nest. found. Food generally occurs in patches. For exam-
FORAGING 151
ple, a freshly watered lawn is a good place for a robin site sensitivity to times of day and time intervals be-
to look for worms, but lawns may be separated by tween foraging opportunities is evident in the forag-
roads and sidewalks that do not provide good foraging behavior of many other species.
ing for a robin. Clearly, it is best to be in the patch
with the most abundant preyforaging in the freshly Another specialized foraging problem requiring
watered lawn with worms close to the surface is prefer- memory is faced by some birds that spend the winter
able to foraging in the dried-up lawn next door. An in a harsh climate. To have enough to eat at such
efficient forager needs good spatial learning abilities times, birds such as chickadees, nuthatches, and jays
so it can navigate from one part of the environment store food when it is abundant. The Clarks nutcrack-
to another. Information about location of suitable for- er, a bird of the American Southwest, buries thou-
aging areas and the density of prey in each has to be sands of pine seeds in the late summer and recovers
constantly updated as the environment changes. them up to six months later. Because each cache is in
Thus animals should sample the environment, some- a different place, the birds must use memory to recov-
times exploring new patches or patches that were not er the food. Experiments in the laboratory with nut-
good the last time they were tried, in order to discover crackers and chickadees, described by VanderWall
whether they have changed for the better. (1990), have shown that these birds can indeed re-
member the locations of their stores and do not need
One aspect of patch choice that has been studied to use other cues. Some research reviewed by Shettle-
extensively is how animals should respond to deple- worth (1998) suggests that they have evolved a better
tion of foraging patches. In our example, as the robin spatial memory than birds that do not store food.
hops around the lawn finding worms, its own foraging This suggestion is supported by evidence that relative
activity (and perhaps that of other birds) reduces the to body size, food-storing birds have a larger hippo-
density of worms in the patch. Some are eaten and campus (the brain area necessary for spatial memory)
others burrow down into the soil at the birds ap- than other birds.
proach. When should the robin leave this patch and
look for another? The foraging theorists answer to
this question takes into account several factors other Conclusion
than the average density of prey in the current patch.
These factors include the density of prey in other Considering what and how animals must learn
patches, how far away the patches are, and whether and remember to forage efficiently is one of the best
prey are constant or variable in size or frequency. Ex- illustrations of how observations and theories about
periments reviewed by Shettleworth (1998) have behavior in the wild have been integrated with the
shown how animals learn about and respond to all study of animals cognitive processes. This analysis is
these variables. at the core of cognitive ecology, a growing interdisci-
plinary area of animal behavior research discussed by
Healy and Braithwaite (2001).
Some Special Problems for Foragers:
Nectar-Feeding and Food-Storing
See also: EVOLUTION AND LEARNING; OPERANT
Learning where to search for prey and detecting BEHAVIOR; SPATIAL LEARNING: ANIMALS
and selecting it once a suitable patch is found are
problems for virtually any forager. Some animals also
face special problems that may require specialized Bibliography
learning and memory abilities. One set of problems, Gould, J. L. (1982). Ethology. New York: W. W. Norton.
Healy, S. D., and Braithwaite, V. (2001). Cognitive ecology: A field
related to the patch-depletion problem just discussed,
of substance? Trends in Ecology and Evolution 15, 2226.
is faced by bees, bats, and hummingbirds that suck Kacelnik, A., and Cuthill, I. C. (1987). Starlings and optimal forag-
nectar from flowers. A flower that has been depleted ing theory: Modelling in a fractal world. In A. C. Kamil, J. R.
of nectar produces more nectar at a rate that depends Krebs, and H. R. Pulliam, eds., Foraging behavior, pp. 303
on factors like what kind of flower it is. The efficient 333. New York: Plenum Press.
Kamil, A. C., and Bond, A. B. (2001). The evolution of virtual ecol-
forager will time its visits so as to return at long
ogy. In L. A. Dugatkin, ed., Model systems in behavioral ecology,
enough intervals to find the flower full, but not so pp. 288310. Princeton, NJ: Princeton University Press.
long that some other forager will have depleted the Shettleworth, S. J. (1998). Cognition, evolution, and behavior. New
flower. One way to ensure this is to travel a fixed route York: Oxford University Press.
among a number of plants. Some nectar-feeding ani- Stephens, D. W., and Krebs, J. R. (1986 ). Foraging theory. Prince-
ton, NJ: Princeton University Press.
mals do forage in this way. For example, Gould
VanderWall, S. B. (1990). Food hoarding in animals. Chicago: Uni-
(1982) describes how bees learn the features of flow- versity of Chicago Press.
ers and the times and places at which nectar can be
found. As reviewed by Shettleworth (1998), an exqui- Sara J. Shettleworth
152 FORGETTING
Table 1
FORGETTING although forgetting takes place over time, it is proba-

bly not because of some inexorable, autonomous fad-
It is a common experience to forget what one has ing of a memory trace.
learned. Usually, forgetfulness increases with with the
length of the retention interval, the time elapsed Forgetting is attributed to decay in only two areas
since the material was last studied or thought about. of memory research: pro forma decay parameters in
A graph of the amount remembered (as measured by abstract mathematical models of memory and the in-
tests of recall or recognition or relearning) as a func- vocation of an unspecified decay process in certain ac-
tion of increasing retention intervals produces a for- counts of short-term memory. The former can be
getting curve, the slope of which represents the over- seen as mathematical conveniences rather than as
all rate of forgetting. The first forgetting curve was strong theoretical statements. The latter suffer from
published in 1885 by Hermann Ebbinghaus, the pio- two major problems: ambiguity in the explication of
neer in the scientific study of memory. His curve the decay process (i.e., it is unclear exactly what aspect
showed the now-familiar monotonic and negatively of memory is decaying, what parts of the memory re-
accelerated form, where the momentary rate of for- main) and the reevaluation of the data supporting
getting decreases over time. such assertions. In principle, decay can be empirically
demonstrated if one rules out alternative, better-
established causes of forgetting, such as interference.
Trace Decay When this is done, the evidence for decay is not ap-
Perhaps the earliest and simplest attempt to ac- preciable; in the words of Cowan, Saults, and Nugent
count for forgetting was the idea of trace decay, which (2001), no clear evidence of decay has emerged in,
postulated that memorizing something lays down a lo, these many years.
neurochemical imprint or record in the brain, called
a memory trace or engram, whose later reactivation Interference
is responsible for remembering. This trace was as-
sumed to fade away spontaneously over time if it was One theoretical approach to forgetting that has
not refreshed by some reacquaintance with the inspired an enormous amount of experimental effort
learned material; hence forgetting would ensue. With over several decades is interference theory. As the
very few exceptions (noted below), trace-decay theory name implies, it focuses upon forgetting caused by in-
has been abandoned. This is not only because the hy- terference. Among its beginnings was the influential
pothetical memory trace has never been identified demonstration by Jenkins and Dallenbach (1924) that
but also because of subsequent research that under- the forgetting of a list of verbal items was markedly
mined the thesis: for example, the findings that for- reduced if subjects passed the retention interval in
getting is influenced by other activities taking place sleep rather than in their usual waking activities. This
both before and after the original learning and that result suggested that the experiences of daily life
forgetting is also greatly affected by the kinds of cues somehow interfered with the memory of the original
given at the time of test. Providing different retrieval material. Eventually, interest narrowed upon other
cues in situations where the trace was supposed to learning experiences as the sources of that interfer-
have decayed often results in successful recall of the ence.
supposedly forgotten material (Capaldi and Neath, The empirical cornerstones of interference theo-
1995); in many instances memory performance can ry are to be found in two kinds of laboratory-
improve with the passage of time (Bjork, 2001). Thus, produced forgetting. These are schematized in Table
FORGETTING 153
1, and each defines a source of interference by com- Table 2

paring the memory performance of one experimental
group, which acquires two lists in succession, with
that of a control group, which acquires only one.
Retroactive inhibition is the forgetting of the first
set of materials, which is caused by the subsequent
learning of a second set during the retention interval.
Proactive inhibition is the forgetting of the second
set, which is caused by the prior learning of a first
set. It may seem strange that a preceding list would
reduce the memory for a subsequently acquired
one, but it can, especially when a retention interval
precedes testing. Interference theory construes retro-
active and proactive inhibition to be the basic
models for its approach to forgetting, inside or
outside of the laboratory. It is the task of the theory Cue-Dependency Theory
to devise an experimentally testable description Cue-dependency theory is an alternative ap-
of the processes of retroactive and proactive inhibi- proach to forgetting. Although quite different from
tion. interference theory, it supplements rather supplants
it. Cue-dependency stresses the importance of the re-
Whenever two sets of materials are acquired in minders, or retrieval cues, that operate at the time of
succession, two processes can impair memory of test. It emphasizes that the act of remembering re-
them. One is the dynamic competition of responses quires not only the stored products of original learn-
for emergence at the time of the test. To the extent ing but also an appropriate testing environment in
that both sets of materials are activated, a response which to make contact with that learning. The suc-
from the first list may be blocked from consciousness cessful interaction of stored information and retrieval
by a stronger competing response from the second information produces recollection. Forgetting is pro-
list, or vice versa. Two competing responses of equal portional to the inadquate accessibility of retrieval
strength may even block each other. This process is cues. Remembering depends on the interaction be-
set into motion by the demands of the memory test, tween the conditions at encoding and the conditions
and it reduces performances on both lists compared at retrieval.
with the single-list control condition. The other An experiment reported by Thomson and Tulv-
process, which takes place during the learning of ing (1970) emphasizes the importance of the encod-
the second list, consists of the temporary suppres- ing/retrieval. The researchers asked subjects to recall
sion of the contents of the first list, to the extent that the target words from a list they were shown. At en-
they conflict with new response requirements. For coding the target word was sometimes presented by
example, if the first set of materials contains an itself, and sometimes with a weak cue. At retrieval
A-B association and the second set requires an A-C there were three cue conditions: Sometimes, no re-
association, then A-B may be suppressed. With both trieval cue was given, sometimes a weak cue was given,
factors taken into consideration, the memory situa- and sometimes a strong cue was given. The terms
tion immediately after second-list acquisition is that strong cue and weak cue refer to a cues ability to elicit
of a suppressed first list competing against a domi- a target. For example, if you were asked to respond
nant second list. This correctly predicts that retroac- with the first word that popped into your head when
tive inhibition will be much stronger than proactive you heard bloom, you are very likely to respond with
inhibition. Another important observation to be ex- flower. So bloom is a strong cue for flower. In contrast,
plained is that with a delayed memory test, pro- you are likely to respond with flower only about 1 per-
active inhibition increases in magnitude. This fact cent of the time to the weak cue bloom. The results,
is incorporated by postulating a gradual dissipation shown as the probability of recalling the correct target
of first-list suppression, wherein the list regains its word, are displayed in Table 2.
strength and competes more effectively, producing When there was no cue at encoding, the weak and
increased interference. This implies that a sup- strong cues at retrieval functioned as expected: More
pressed set of materials should be better recalled after target items were reported when given a strong cue
some time has passed, which is the opposite of for- (0.68) than when given a weak cue (0.43). However,
getting. Such an effect has been experimentally veri- when there was a weak cue at encoding, the weak cue
fied. at retrieval elicited the target word more than 80 per-
154 FREUD, SIGMUND
cent of the time, whereas the strong cue elicited the Postman, L., and Underwood, B. J. (1973). Critical issues in inter-
target word barely 20 percent of the time. ference theory. Memory & Cognition 1, 1940.
Slamecka, N. J. (1985). Ebbinghaus: Some associations. Journal of
What matters is the extent to which the condi- Experimental Psychology: Learning, Memory, and Cognition 11,
tions at retrieval uniquely specify the target informa- 414435.
tion. Thus, if you study in the presence of a particular Slamecka, N. J., and McElree, B. (1983). Normal forgetting of ver-
bal lists as a function of their degree of learning. Journal of Ex-
odor (say, freshly baked chocolate-chip cookies) but perimental Psychology: Learning, Memory, and Cognition 9, 384
are tested in the absence of that odor, your ability to 397.
remember is impaired. A change in mood, environ- Thomson, D. M., and Tulving, E. (1970). Associative encoding and
ment, or pharmacological state can result in similar retrieval: Weak and strong cues.Journal of Experimental Psychol-
decrements. Most of the things you have memorized ogy 86, 255262.
Underwood, B.J. (1957). Interference and forgetting. Psychological
were not accompanied by the smell of chocolate-chip Review 64, 4960.
cookies; having the odor at retrieval provides a cue
that greatly narrows down the possible items to be re- Norman J. Slamecka
called. According to this view, the way to overcome Revised by Ian Neath
forgetting is to envision what cues will be uniquely
available at the time the information is going to be
needed. Once those cues are identified, then the ma-
terial can be encoded with reference to those cues. FREUD, SIGMUND (18561939)
Sigmund Freud was the founder of psychoanalysis, a
Normal Forgetting system of psychological therapy and personality theo-
ry that remains one of the most influential and con-
Are there any conditions of original learning that troversial in psychology. Born in 1856 in Freiberg,
influence the rate of forgetting of a single set of mate- Moravia (then a province of the Austro-Hungarian
rials learned in the laboratory? Curiously, normal for- Empire and today a part of Czechoslovakia), he was
getting rates appear to be independent of the materi- the first son of Jakob Freud, a wool merchant, and of
als memorized. They also seem impervious to the Amalie Nathansohn Freud, Jakobs third wife.
levels of processing (superficial versus meaningful)
employed at study. There is little evidence that they Freuds background was Jewish, a fact that fig-
yield to deliberate mnemonic strategies. There is ured importantly in his life and work, although he
even serious doubt about whether the degree of origi- himself was an atheistin his words, a Godless
nal learning has any effect upon the rate of loss. The Jewand was to write withering critiques of religion,
present picture suggests a remarkable resistance to which he considered a psychological narcotic (see,
experimental manipulation. Surely no theory or e.g., Freud, 1927, 1930).
model of memory can be expected to account satisfac- Notwithstanding his birthplace, Freud for all in-
torily for normal forgetting short of resolution of this tents and purposes was a Viennese. His family moved
question. to Vienna when he was four, and he remained there
until about a year before his death, when he fled Nazi
Bibliography Austria to settle with his family in London. His four
Bjork, R. A. (2001). Recency and recovery in human memory. In sisters, who stayed behind, perished in the Holocaust.
H. L. Roediger III, J. S. Nairne, I. Neath, and A. M. Surpre-
nant, eds., The nature of remembering: Essays in honor of Robert In 1886 Freud married Martha Bernays. They had
G. Crowder. Washington, DC: American Psychological Associa- three sons and three daughters, the youngest of
tion. whom, Anna, became a prominent figure in the psy-
Capaldi, E. J., and Neath, I. (1995). Remembering and forgetting choanalytic movement.
as context discrimination. Learning and Memory 2, 107132.
Cowan, N., Saults, S., and Nugent, L. (2001). The ravages of abso- Freud, a heavy cigar smoker, tried several times
lute and relative amounts of time on memory. In H. L. to give up his vice but found that he could not write
Roediger III, J. S. Nairne, I. Neath, and A. M. Surprenant, without smoking. He was found to have cancer of the
eds., The nature of remembering: Essays in honor of Robert G. mouth in 1923 and endured more than thirty opera-
Crowder. Washington, DC: American Psychological Associa-
tion. tions before his death in 1939. He remained a prolific
Ebbinghaus, H. (1885/1913; 1964). Memory: A contribution to experi- writer until the end, leaving some two dozen volumes
mental psychology, trans. H. A. Ruger and C. E. Bussenius. New of psychological works.
York: Dover.
Jenkins, J. G., and Dallenbach, K. M. (1924). Obliviscence during
sleep and waking. American Journal of Psychology 35, 605612. Education and Early Career
McGeoch, J. A. (1932). Forgetting and the law of disuse. Psychologi-
cal Review 39, 352370. For the last six of his eight years in gymnasium
Neath, I. (1998). Human memory: An introduction to research, data, and (the European version of academic high school),
theory. Pacific Grove, CA: Brooks/Cole. Freud was at the top of his class. He claimed to have
FREUD, SIGMUND 155
a photographic memory, and he possessed a power-

ful, brilliant writing style. In 1930 he received the
Goethe prize, a literary award. He had less of a bent
for the hard sciences and mathematics. After an initial
inclination toward a career in law or politics, Freud
decided upon the natural sciences and at seventeen
enrolled in the medical school of the University of Vi-
enna, where he conducted research in anatomy and
physiology.
After some dawdling, Freud obtained his M.D.
degree in 1881, at the age of twenty-five. He hoped
to remain a research scientist at the university, but, re-
alizing that he was not apt to receive a permanent uni-
versity appointment (there were no likely openings
soon, and his Jewish background also worked against
him), he reluctantly opted for medical practice, train-
ing at the Vienna General Hospital (18821885) and,
after a stint as an army doctor, assuming the director-
ship of the neurological department of a hospital for
children while he established a private practice. In
this period Freud published some papers on neurolo-
gy, among them articles on childrens paralyses, and
a book on aphasia (1891). He also experimented on
himself with the then exotic drug cocaine. Enthusias-
tic over its enhancing effects on energy and creativity,
he was soon urging the drug upon his fiance and
family members. This cocaine phase came to an
abrupt end when a friend of his (whom he was trying Sigmund Freud (World Health Organization/Washington, DC)
to wean from heroin addiction with cocaine) over-
dosed and died. The episode may have hurt Freuds
medical reputation in Vienna although it did, indi- tients were not actually afflicted with organic disor-
rectly, lead to the successful use of cocaine as a local ders but were suffering from psychological conditions
anesthetic in ophthalmology. that mimicked neurological symptoms. Such patients
In 1885, while at the Vienna General Hospital, were often referred to as neurotics or hysterics.
Freud won a four-month grant to study with the Psychoanalysis was born in the effort to devise psycho-
worlds preeminent neurologist, Jean Charcot, in logical treatment for these psychological afflictions.
Paris, who at the time was pursuing research on hyste-
An important partner in Freuds early psycho-
ria and hypnosis. Four years later Freud also visited
the great center of hypnotic research and therapy, therapeutic research was his older friend and mentor,
headed by Hippolyte Bernheim, at Nancy, France. Josef Breuer, an eminent Viennese physician. Breuer
Freud was much taken by demonstrations there of had experimented with hypnotic suggestion therapy,
posthypnotic suggestion with amnesia, in which sub- a technique in which the hysteric patient, under hyp-
jects carried out hypnotic suggestions after awaken- nosis, is given the suggestion that his or her symp-
ing from their trance without being aware of the rea- toms will disappear. With his famous patient Anna O.,
son for their actions. Beyond their therapeutic Breuer varied the procedure and stumbled on a strik-
import, Freud saw such demonstrations as laboratory ing phenomenon: When Anna was able to recollect
confirmations of the power of the unconscious to af- and report the painful events that had brought about
fect behavior. her symptoms, the symptoms permanently disap-
peared. The procedure, which Anna O. referred to as
the talking cure, was more formally designated by
The Birth of Psychoanalysis Breuer as the cathartic technique. Freud experiment-
Because of Freuds background in neurology, his ed with and extended the hypnotic and cathartic tech-
colleagues tended to refer patients to him who suf- niques, and in 1895 he and Breuer published what
fered from neurological symptoms such as paralysis, may be considered the founding work in psychoanaly-
tic, or amnesia. Clearly, however, many of these pa- sis, Studies on Hysteria.
156 FREUD, SIGMUND
The central idea of the Studies was that hysterics learning and memory is the selective facilitation and
suffer mainly from reminiscencesmemories of disinhibition of the flow of excitation across the contact
turbing events from the past that are not accessible to barrierssynapses, in modern parlancebetween
their consciousness but are expressed in the recondite neurons, resulting in the gradual differentiation of
dialect of body language as hysterical symptoms. The functional neural units or systems within broader
cathartic strategy was to recover the pathogenic idea neural networks. Freuds speculative neuropsy-
and its associated effect into awareness. Freud gradu- chology foreshadowed many key notions in contem-
ally dropped hypnosis as his hypermnesic (memory- porary neuroscience, including todays influential
enhancing) agent, though he retained the couch on parallel distributed processing models of learning,
which hypnotized patients reclined. Eventually he perception, and memory.
adopted psychoanalysiss free-association technique,
Freud is well known, and much criticized, for his
in which the patient is committed to the basic rule
theory of infantile sexuality, which holds that children
of saying anything that comes to mind, without cen-
are sexual creatures who, in the first years of life, go
sorship. Freud, an ardent determinist, believed that
through psychosexual developmental stages (oral,
the superficially haphazard approach would willy-
anal, and phallic) that presage adult sexuality (the
nilly lead back to the pathogenic trauma or conflict.
genital stage), reached at puberty after a period of
He found that the problem-causing mental contents
sexual latency. Problems during the early sexual
often originated in childhood sexual and aggressive
stages (e.g., traumas, deprivation) result in develop-
feelings, especially those arising from the Oedipus
mental arrestsfixationsthat produce distinctive
complex around the age of five or six, when, Freud
adult character types: the oral (associated with
claimed, the child develops overtly erotic desires for
traits such as dependency, passivity, propensity for
the parent of the opposite sex and hateful feelings to-
smoking), anal (e.g., fussy orderliness, stinginess, sa-
ward the parent of the same sex.
dism), and phallic (e.g., seductiveness, impotence). It
A prominent yield of the free-association proce- might be noted that Freud had no direct clinical ex-
dure, considered within psychoanalysis a micro- perience with children nor any research involvement
scope of the mind, was that the basic rule was psy- with them. Modern scientific evaluations (e.g., Fisher
chologically impossible to honor. Freud came to and Greenberg, 1996) suggest that the cluster of traits
realize that certain thoughts and wishes were too dis- defining oral or anal character do exist but that
turbing to communicate even to oneself. This inward the notion that developmental events cause these
dishonesty led to the fundamental concept of defense types has little support. Thus, the typology may have
mechanisms, psychological processes deployed to dis- some validity but the etiology does not.
tort or exclude from consciousness thoughts and de-
sires that are too anxiety producing. Ultimately (The In 1900, following a self-analysis initiated in
Ego and the Id, 1923), Freud viewed the mind as divid- 1895, Freud published his seminal Interpretation of
ed into three often-conflicting subsystems: an execu- Dreams. If free associations were a microscope of the
tive, reality-oriented component, the ego; a primitive, mind, dreams to Freud were the royal road to the
passion-controlled subsystem, the id; and a moral unconscious. During sleep, defenses are weakened
subsystem, the superego. and the primitive drives and primitive modes of cog-
nitionprimary process thinkingcome to the fore.
Like any complex mental product, dreams have mul-
Attempting a Scientific Understanding of tiple levels of meaning. The surface meaning, which
Psychological Processes Freud called the manifest content, often makes no
sense; the deep or latent content, which requires in-
In addition to the publication of Studies on Hyste-
terpretation, carries the important meaning. In
ria, the year 1895 was a watershed in another way.
dreams, the latent content is often unconscious to the
Freud feverishly attempted to create a neurological
subject. Thus, like a hysteric symptom, a dream in-
model of basic psychological processesPsychology
volves a meaningful communication of which the pa-
for Neurologistsbut by the end of the year con-
tient may not be aware.
cluded that the effort was failure and withheld the
text from publication. In effect he abandoned neurol- Dream analysis, along with the free-association
ogy and became a full-fledged psychologist. His fol- method, became an integral feature of psychoanalytic
lowers published the manuscript posthumously under therapy. To these, later on, was added the analysis of
the title Project for a Scientific Psychology. Convoluted transference. Freud found that in the course of thera-
and often impenetrable, the one-hundred-page py, patients developed passionate feelings for the
monograph does contain some remarkably prescient therapist of both a loving and a hating character. Be-
neurological intuitions about learning and memory, cause these were not realistically warranted, Freud
among them the notion that the neurological basis of took the transference manifestations to be reenact-
FRONTAL LOBES AND EPISODIC MEMORY 157
ments of past significant relations, especially between ere and J. Strachey, trans. In J. Strachey, ed., The standard edi-
the patient and his or her parents. The transference tion of the complete psychological works of Sigmund Freud, Vol. 12.
neurosis, then, was a profound form of remember- London: Hogarth Press.
(1917). A general introduction to psychoanalysis, J. Riviere,
ing, involving not conscious recollection but reliving trans. New York: Liveright.
of problematic themes from the past. The goal of psy- (1923; reprint 1961). The ego and the id, J. Riviere and J.
choanalysis was always in some sense the education Strachey, trans. In J. Strachey, ed., The standard edition of the
of consciousness. Through self-knowledge, the pa- complete psychological works of Sigmund Freud, Vol. 19. London:
tient was thought to gain mastery over the irrational Hogarth Press.
(1926; reprint 1959). Inhibitions, symptoms, and anxiety, A.
determinants of his or her pathological symptoms Strachey and J. Strachey, trans. In J. Strachey, ed., The stan-
and behavior. dard edition of the complete psychological works of Sigmund Freud,
Freud always insisted on the scientific status of his Vol. 20. London: Hogarth Press.
(1927; reprint 1961). The future of an illusion, W. D. Robson-
creation, although his practices were often unscientif- Scott and J. Strachey, trans. In J. Strachey, ed., The standard
ic. Disagreements with his colleaguesfigures such as edition of the complete psychological works of Sigmund Freud, Vol.
Josef Breuer, Carl Gustav Jung, Alfred Adlerled to 21. London: Hogarth Press.
personal and professional ruptures. Freud did not en- (1930; reprint 1961). Civilization and its discontents, J. Riviere
gage in any true experimental research, and the em- and J. Strachey, trans. In J. Strachey, ed., The standard edition
of the complete psychological works of Sigmund Freud, Vol. 21. Lon-
pirical anchoring of his science was clinical observa- don: Hogarth Press.
tion, which often amounted to all-too-fallible
interpretation. Probably for this reason, more than Matthew Hugh Erdelyi
resistance to supposedly unpalatable claims (such as
the doctrine of infantile sexuality), mainstream scien-
tific psychology has maintained a suspicious, ambiva-
lent stance toward psychoanalysis. Nevertheless, for FRONTAL LOBES AND
all his great flaws, Freud had his great insights, EPISODIC MEMORY
among them the concept of unconscious mental activ- The idea that the frontal lobes are implicated in
ity; the notion of remembering without awareness memory has a long and controversial history (Luria,
through behavior and other indirect channels; the 1980; Teuber, 1964). Damage to the frontal lobes can
meaningfulness of dreams; the therapeutic value of produce memory impairment and sometimes even
talking and of the interpersonal relation in therapy; severe memory loss, but it has proved difficult to spec-
and the pervasiveness of biased, often defensive, re- ify the nature of the disorder. The scholarly consen-
constructions of memory. In the century after they sus now holds that frontal-lobe damage does not lead
were proposed, these themes emerged as mainstream to memory deficits in consolidation, storage, and re-
notions in modern psychology. tention of newly acquired information (Petrides,
Freud is a towering but transitional figure in sci- 2000; Moscovitch and Winocur, 2002). Such disor-
entific psychologya sort of Moses, who achieved ders, which in their most extreme form lead to a pro-
great things but was not himself destined to reach the found global amnesia, are associated with damage to
promised land. the medial temporal lobes, particularly the hippo-
campus and related structures, and to midline tha-
See also: GUIDE TO THE ANATOMY OF THE BRAIN; lamic nuclei (Milner, 1966).
PARALLEL DISTRIBUTED PROCESSING MODELS
OF MEMORY
Working-with-Memory
Bibliography
Memory loss following frontal-lobe lesions, on
Breuer, J., and Freud, S. (1895; reprint 1955). Studies on hysteria,
trans. A. Strachey and J. Strachey. In J. Strachey, ed., The stan- the other hand, involves organizational or strategic
dard edition of the complete psychological works of Sigmund Freud, aspects of memory that are necessary for devising
Vol. 2. London: Hogarth Press. strategies for encoding, for guiding search at retriev-
Fisher, S., and Greenberg, R. P. (1996). Freud scientifically reappr- al, for monitoring and verifying memory output, for
aised: Testing the theories and therapy. New York: Wiley. placing retrieved memories in their proper spatial
Freud, S. (1891). On aphasia. New York: International Universities
Press, 1953. and temporal contexts, and for using mnemonic in-
(1895; reprint 1966). Project for a scientific psychology. In J. A. formation to direct thought and plan future actions.
Strachey, ed., The standard edition of the complete psychological In other words, the frontal lobes memory functions
works of Sigmund Freud, Vol. 1. London: Hogarth Press. are consistent with its functions in other domains. It
(1900; reprint 1953). The interpretation of dreams. In J. A. helps organize the raw material that is made available
Strachey, ed., The standard edition of the complete psychological
works of Sigmund Freud, Vols. 4 and 5. London: Hogarth Press. by other structures so that thought and behavior can
(1914; reprint 1958). Remembering, repeating, and working be goal-directed. If the hippocampus and its related
through: Further recommendations . . . of psychoanalysis II, J. Rivi- structures can be considered raw memory structures,
158 FRONTAL LOBES AND EPISODIC MEMORY
Figure 1
The sequence of frontal and medial temporal lobe activation during retrieval of episodic memories to a direct or indirect cue.
then the frontal lobes are working-with-memory of damage to the hippocampus and midline thalamic
structures that operate on the input and output of the nuclei. Recognition and recall of isolated random
hippocampal circuit (see Figure 1). We prefer the de- words or pictures are typically normal in patients with
scriptive term working-with-memory to the more theo- frontal lesions but impaired in patients with hippo-
retically loaded terms working memory or executive func- campal or diencephalic damage (Mayes, 1988; Mil-
tion favored by others (Baddeley, 1986; Goldman- ner, Petrides, and Smith, 1985). Recall of categorized
Rakic, 1987; Smith and Jonides, 1998; Stuss and lists or of logical stories, however, is impaired in fron-
Knight, 2002); this term captures the essence of frontal patients, presumably because they cannot take ad-
tal-lobe contribution to memory but does not commit vantage of the organizational structure inherent in
the user to endorse a working-memory theory that that material (Incissa della Rochetta, 1986). Deficits
may be flawed or inappropriate (Moscovitch and may also occur in free recall if normal performance
Winocur, 1992). depends on strategic search or retrieval (Janowsky,
Shimamura, Kritchevsky, and Squire, 1989; Wheeler,
Review of Lesions Studies and Memory Stuss, and Tulving, 1995; Moscovitch and Winocur,
1995, 2002).
Tasks
One can best appreciate the contribution of the In contrast with memory for target items or facts
frontal lobes to memory by comparing the effects of that can be elicited by cues directly associated with
frontal damage on various memory tests with effects them, memory for spatiotemporal context often re-
quires strategic search. Consider introspectively the each subsequent trial, new sheets are presented with
difference in the processes involved in answering the same items arranged differently, and the subject
these two questions: Have you seen Gone with the is required to point to a different item each time.
Wind? What did you do two weekends ago? The first elic- There are as many trials as there are items. Apart
its an immediate, automatic reply; the second typical- from remembering the items and keeping spatiotem-
ly initiates a labored, strategic search. As expected, poral context distinct, subjects performing this test
memory for spatiotemporal context, but not for tar- need to monitor their responses and use their memo-
gets, is impaired after frontal lesions, whereas the re- ry of them to plan future actions. Monitoring and
verse is true after hippocampal or midline dience- planning are both prototypical frontal functions that
phalic lesions. are applied to memory and may be impaired by fron-
tal lesions. Another way to interpret these results is to
Memory for temporal order is poor in frontal pa-
say that the test requires monitoring information that
tients when it is tested by asking patients to judge the
is held in working memory (Petrides, 2000).
relative recency of a pair of items or to arrange a set
of items in the order in which they were presented Impaired estimation of frequency of occurrence
(Milner, Petrides, and Smith, 1985; Shimamura, Ja- is also associated with frontal lesions (Smith and Mil-
nowsky, and Squire, 1990). The deficit also extends ner, 1988). In one test of this association, items are re-
to remote memories that were acquired long before peated a number of times and the subjects task is to
the lesion occured. Defective memory for sources of estimate the number of repetitions. It is not clear
facts they had learned has also been observed in pa- whether the deficit on this test is a symptom of a gen-
tients with frontal lesions or frontal dysfunction. They eral deficit in cognitive estimation that accompanies
erroneously ascribe the factual information to an in- frontal damage (Shallice and Evans, 1978) or whether
correct source on tests of source recall and source rec- it also results from a failure to search memory strate-
ognition (Schacter, Harbluk, and McLachlan, 1984; gically.
Janowsky, Shimamura, and Squire, 1989b). On the
Although the formation and retrieval of new asso-
other hand, hippocampal or diencephalic damage
ciations are not dependent on the frontal lobes, learn-
leads to deficient memory for targets or facts at long
ing conditional associations is (Petrides and Milner,
delays but not for their temporal order or for sources
1982). The difference between the two tests high-
at delays in which the facts can be remembered. In
lights the distinction between memories elicited di-
other words, their memory for temporal order is no
rectly by cues associated with them, a process that in-
more impaired than, and may be superior to, their
volves the hippocampus, and memories for which
memory for facts (Milner, Petrides, and Smith, 1985;
additional extra-cue, strategic processes are necessary
Shimamura, Janowsky, and Squire, 1990).
and involve the frontal lobes. In learning new associa-
Poor memory for spatiotemporal context that re- tions, a single cue such as a light is paired with a
sults from impaired strategic processes may also un- unique response such as an arm movement, which it
derlie the frontal patients deficits on a variety of eventually elicits. In conditional associative learning,
other tests, such as delayed alternation, delayed re- all cues and potential responses are present in the sit-
sponse, and delayed match-to-sample with a small, uation and typically resemble one another. For exam-
repeated set of items (Freedman and Oscar-Berman, ple, a set of six lights is presented, each of which
1986; Prisko, 1963; Milner, Petrides, and Smith, needs to be associated with one of six movements that
1985). On delayed response, after a short delay the the subject has mastered. The cue, one of the lights,
subject must choose one of the items that had been does not elicit the response, the designated move-
designated as the target. On delayed alternation, the ment, but only provides the occasion for the subject
designated target alternates on every trial. In delayed to select from among potential responses the one that
match-to-sample, the subject chooses from a small set is appropriate for one particular cue, another one for
of items the one that matches a target that was in- another cue, and so on. That is, the subject must de-
spected earlier. On all these tests, frontal patients fail termine which light is associated with which move-
not because they cannot remember the target but be- ment. Response selection and monitoring, both stra-
cause they cannot segregate the current trial (keep tegic frontal functions, are key elements of this task.
the spatiotemporal context distinct) from preceding Patients with frontal lesions have difficulty learning
ones. only conditional associations, whereas patients with
hippocampal lesions have difficulty forming new as-
Impaired performance by frontal patients on self-
sociations.
ordered pointing tests may have a similar cause (Pe-
trides and Milner, 1982). In these tests, subjects are Less consistent effects of frontal lesions are found
required to point to one of a set of words, line draw- on other memory tests, such as release from proactive
ings, or designs that appear on a sheet of paper. On inhibition (PI) and feeling-of-knowing judgments.
In release from PI, four different lists of words from trast, the operations of the hippocampus can be run
the same semantic category are presented, followed off relatively automatically once the appropriate
by a list from a different category. Recall, which is input is received. Experiments reported in the litera-
tested after each list, declines from the first to the ture suggest that general interference at the time of
fourth list as PI builds up, but recall recovers to base- retrieval affects primarily performance on tests that
line levels on the fifth trial. Release from PI occurs at are sensitive to frontal function, such as word fluency,
retrieval. It is not surprising, therefore, that deficits recall of categorized lists, and list differentiation. A
in release from PI have been reported in patients with series of experiments designed to test this hypothesis
frontal lesions (Moscovitch, 1982); these are most re- further has confirmed that a sequential, finger-
liable, however, when a severe memory disorder ac- tapping encoding-and-retrieval task interfered with
companies frontal dysfunction (Freedman and Cer- performance on frontal-sensitive tests, such as recall
mak, 1986; Janowsky, Shimamura, Kritchevsky, and of categorized lists, release from PI, and phonemic
Squire, 1989). fluency (Moscovitch, 1994). Similar divided-attention
effects are observed on tests of source, but not item
Feeling of knowing is an aspect of metamem-
memory (Moscovitch, Fernandes, and Troyer, 2002).
ory, the knowledge about ones own memory. It refers
To affect memory on nonfrontal tests such as free re-
to a persons belief that he or she would know the cor-
call and recognition of lists of random word the inter-
rect answer to a memory question. In testing the accu-
ference at retrieval must be material-specific in the
racy of feeling-of-knowing judgments, Janowsky, Shi-
sense that the interfering task consists of material that
mamura, and Squire (1989a) gave subjects a cued
is similar to that of the target, memory task. In such
recall test for information they had learned earlier.
cases, it is believed that the locus of interference is for
For those items they failed to recall, subjects were
access to representational systems in posterior neo-
asked to rate their feeling of knowing, their likelihood
cortex (Moscovitch et al., 2002).
of recognizing the correct answer among a number of
alternatives. Because this metamemory test involves All the memory tests mentioned so far are explicit
goal-directed search and monitoring, it was expected tests that require conscious recollection of the past for
that patients with frontal lesions would perform poor- successful performance. In contrast, on implicit tests,
ly on it. Although some deficits were found, the im- memory is inferred from the effects of experience or
pairment, as in release from PI, was most reliable and practice on performance without requiring the indi-
far-reaching in patients who had severe memory vidual to refer to the past. Since the frontal lobes are
problems in addition to frontal dysfunction (Shima- working-with-memory structures, they should be im-
mura, Janowsky, and Squire, 1990). plicated on tests of implicit memory. Indeed, frontal
lesions or dysfunctions lead to impaired performance
Defective performance on memory tests sensitive
on those tests that are not simply stimulus-driven but
to frontal lesions is noted in people with neurological
require strategic search or application of organized
conditions associated with frontal dysfunctionthat
rules or procedures. Thus, patients with frontal le-
is, with signs of impaired frontal functions though
sions or dysfunction have difficulty mastering the
there is no evidence of direct frontal damage. Among
Tower of Hanoi, a cognitive puzzle whose solution de-
those are patients with Parkinsons and Huntingtons
pends on the application of a sequential, iterative rule
diseases, the neuropathology of which affects basal
(Owen et al., 1990; Shallice, 1982; St. Cyr, Taylor,
ganglia structures that are part of the complex loop
and Lang, 1988). Frontal patients may also be im-
that connects them to the frontal lobes (Brown and
paired on other implicit tests, such as learning to read
Marsden, 1990; Saint-Cyr, Taylor, and Lang, 1988)
geometrically transformed script and completing
and that may be needed for maintaining dopamine
word stems after being exposed to target words
levels in the prefrontal cortex. Declines in perfor-
(Winocur, Moscovitch, and Stuss, 1996; Nyberg,
mance on frontal-sensitive tests also occur in the el-
Winocur, and Moscovitch, 1997). The results suggest
derly, presumably because their frontal lobes deterio-
that the frontal lobes are implicated on implicit tests
rate with age (Moscovitch and Winocur, 1992), and in
that have a selection or generative component, but
patients with schizophrenia, both because of frontal
not on tests that are purely perceptually driven (Gabr-
deterioration and impaired dopamine function.
ieli et al., 1999). The extent of frontal involvement on
Even normal young adults may show deficits on implicit tests is still uncertain. A great deal of work on
frontal tests under conditions that deplete cognitive amnesic patients with hippocampal or diencephalic
resources. Because frontal functions are strategic lesions shows that their performance on a variety of
which implies that voluntary, often conscious, control implicit tests appears mostly unscathed, suggesting
is an integral part of themthey demand substantial that these structures are involved only with conscious
cognitive, attentional resources if they are to operate recollection (Moscovitch, 1982; Moscovitch and
effectively (Moscovitch and Umilta, 1990). In con- Umilta, 1991).
Many of the features of frontal-lobe memory dis- logenetic and ontogenetic histories (Pandya and
orders are observable in an especially severe and Barnes, 1986; Petrides and Pandya, 1994). Two large
striking form in confabulating patients with aneu- subdivisions of the prefrontal cortex, the orbital and
rysms or infarcts of the anterior communicating arte- dorsolateral regions, have different functions (Mil-
ry. Admittedly, the lesions that typically affect the ner, Petrides, and Smith, 1985), one more emotional
ventromedial and orbital regions of the frontal lobes and motivational and the other more cognitive. Struc-
also involve other structures in the basal forebrain tural and functional neuroimaging has allowed the
such as the anterior cingulate, the septum, and the discernment of specialized functions of even smaller
anterior hypothalamus. Nonetheless, the memory regions within these subdivisions (Goldman-Rakic,
symptoms displayed by these patients indicate frontal 1987, 2002; Petrides, 2000, 2002). Indeed, there is
rather than hippocampal circuit damage. When test- now evidence for regional specialization for many of
ed formally, their recognition on tests that do not in- the functions that contribute to performance on the
volve strategic search is relatively well preserved, a re- various tests of frontal function.
sult that distinguishes them from amnesic patients
with hippocampal damage. However, their ability to Brodmann areas 6 and 8 (premotor cortex) are
search memory and to place events in a proper spa- implicated in response selection and inhibition, a key
tiotemporal context is nearly lost. That loss likely ac- feature of conditional associative learning, and re-
counts for their tendency to confabulate or make up trieval from remote memory (Moscovitch and
patently false, often contradictory, and occasionally Winocur, 2002). Areas 46/9 (dorsolateral prefrontal
bizarre or fantastic stories. cortex) is for manipulating and operating on infor-
mation held in working memory or retrieved from
For example, one patient who had been in the long-term memory, as on tests of self-ordered point-
hospital for months claimed he was still at his office. ing, where monitoring is a crucial component. This
When asked to account for the beds in his room, he region is also important for initiating effective retriev-
suggested that they were brought in to deal with an al strategies and monitoring the outcome of imple-
epidemic. When such patients confabulate, they do menting them (i.e. retrieval output). The ventrolater-
not intentionally lie but inadvertently combine true al prefrontal cortex, area 47, is important for
memories whose spatiotemporal context they have specifying the cues needed for retrieval (Fletcher and
lost. According to Schnider and his collaborators Hensen, 2001). Perhaps it is for this reason that the
(1996; 2000), such temporal confusion is due to diffi- extent of activation of this area at encoding can pre-
culty suppressing information that no longer is rele- dict subsequent memory performance (Wagner et al.,
vant and focusing on that which is. It is as if the non- 1998) and therefore plays a role in simple recognition
frontal memory system, in response to situational
(Petrides, 2000). This region also comes into play on
cues, spews out loosely associated memories in a
tests of conceptual priming, with less activation for
quasi-ordered fashion. Lacking intact frontal lobes,
material that has been primed (Schacter and Buck-
these patients cannot evaluate this output or impose
ner, 1998). The ventromedial prefrontal cortex is im-
a sensible organization on it. Their memory deficits,
portant for inhibiting or preventing the expression or
therefore, are not restricted to recently acquired
behavioral impact of activated memories that are
memories but extend to remote, personal memories
anomalous within a given context. The region is a
and even to historical information on events that oc-
context-dependent criterion-setting device that de-
curred before they were born (Moscovitch and Melo,
termines which memories are relevantthe felt right-
1997). Their memory is intact only for events and ac-
ness of a memory. The function of area 10 in the fron-
tivities, such as their job routines, that are stored as
tal pole is much debated, with some investigators
self-organized and self-contained schemata that de-
believing that it is crucial for allowing one to experi-
pend little on supervision by the frontal lobes for their
ence memories as part of oneself and tying them in-
operation (Moscovitch, 1989).
extricably to conscious, coherent recollection of ones
past (Tulving, 2002). Others, however, believe that
Localization Within the Prefrontal Cortex: the role of area 10 is to maintain elaborate retrieval
Evidence from Lesions and Neuroimaging goals and strategies (Fletcher and Hensen, 2001).
Still others think it may operate in concert with the
Though we have paid little attention so far to lo-
verntromedial cortex to endorse memories that are
calization within the frontal lobes, we do not wish to
appropriate within a given context. The possible se-
leave the reader with the impression that the prefron-
quential interaction of the regions with each other is
tal cortex is a homogeneous structure; on the con-
displayed in Figure 1.
trary, it is a heterogeneous structure consisting of a
number of distinct areas with unique projections to In addition to regional specificity, there is also
and from other brain regions, and with different phy- lateralization of function within the frontal lobes:
Some regions show material specificity with greater See also: EPISODIC MEMORY; METACOGNITION
left-hemisphere involvement for verbal material and ABOUT MEMORY; WORKING MEMORY: ANIMALS;
right-hemisphere involvement for spatial material WORKING MEMORY: HUMANS
(Milner, Petrides, and Smith., 1985; Petrides, 2000;
Kelley et al., 1998). Over and above the material spec- Bibliography
ificity, Tulving and his colleagues (1994) noted a Baddeley, A. D. (1986). Working memory. Oxford: Oxford Universi-
hemispheric encoding/retrieval asymmetry (HERA) ty Press.
in PET studies, with the greater frontal activation on Brown, R. G., and Marden, C. D. (1990). Cognitive function in Par-
the left during encoding and on the right during re- kinsons disease: From description to theory. Trends in Neuro-
sciences 13, 2129.
trieval. This pattern has since been observed in nu- Buckner, R. L. and Wheeler, M. E. (2001). The cognitive neuro-
merous functional neuroimaging studies using differ- science of remembering. Nature Reviews: Neuroscience 2, 624
ent techniques (Cabeza and Nyberg, 2000); it cuts 634.
across material type in some regions of prefrontal Cabeza, R., Mangels, J., Nyberg, L., Habib, R., Houle, S., McIn-
tosh, A. R. et al. (1997). Brain regions differentially involved
cortex. The reason for this asymmetry is unknown;
in remembering what and when: a PET study. Neuron, 19,
nor is it clear whether such asymmetries are needed 863870.
for successful encoding and retrieval. Researchers Cabeza, R., and Nyberg, L. (2000). Imaging cognition II: an empir-
have debated which aspects of retrieval are associated ical review of 275 PET and fMRI studies. Journal of Cognitive
Neuroscience 12, 147.
with right frontal activation: retrieval mode (the es-
DeLuca, J. (2000) A cognitive neuroscience perspective on confab-
tablishment of a memory set and perhaps setting ulation. Neuro-psychoanalysis 2, 119132.
memory goals and strategies), monitoring (the deter- Fletcher, P. C., and Henson, R. N. A. (2001). Frontal lobes and
mination of whether a target item was studied), re- human memory. Insights from functional neuroimaging.
trieval success (the accuracy of recall and recognition Brain 124, 849881.
Freedman, M., and Cermak, L. S. (1986). Semantic encoding defi-
of items), retrieval task or domain (memory for source cits in frontal lobe disease and amnesia. Brain and Cognition
or for items), and retrieval effort (how difficult the re- 5, 108114.
trieval task is) (Buckner and Wheeler, 2001; Holding Freedman, M., and OscarBerman, M. (1986). Bilateral frontal
and Rugg, 2000). It seems that as memory retrieval lobe disease and selective delayed response deficits in hu-
mans. Behavioral Neuroscience 100, 337342.
requires more effort, either because the task is more
Gabrieli, J. D. E., Vaidya, C. J., Stone, M., Francis, W. S., Thomp-
difficult or reflective (Johnson, 2000) or because cog- son-Schill, S. L., Fleischman, D. A., Tinklenberg, J. R., Yesav-
nitive resources are limited (as in aging, Cabeza, age, J. A., and Wilson, R. S. (1999). Convergent behavioral
2001), both frontal regions are recruited, and the and neuropsychological evidence for a distinction between
identification and production forms of repetition priming.
asymmetry is diminished.
Journal of Experimental Psychology: General 128, 479498.
GoldmanRakic, P. S. (1987). Circuitry of primate prefrontal cor-
tex and regulation of behavior by representational memory.
Conclusion In F. Plum, ed., Handbook of physiology: The nervous system, Vol.
5. Bethesda, MD: American Physiological Society.
Although there has been some progress in identi-
(2002). The prefrontal cortex: from molecule to mind. In
fying the frontal components implicated in encoding D. T. Stuss and R. T. Knight, eds, Principles of frontal lobe func-
and retrieval and isolating them functionally and ana- tion. New York: Oxford University Press.
tomically, discrepancies and controversies persist in Incissa della Rochetta, A. I. (1986). Classification and recall of pic-
efforts to understand how the different regions inter- tures after unilateral frontal or temporal lobectomy. Cortex 22,
189211.
act with one another to yield organized, seamless per- Janowsky, J. S., Shimamura, A. P., Kritchevsky, M., and Squire, L.
formance. Future research should also help in decid- R. (1989). Cognitive impairment following frontal lobe dam-
ing between two opposing views of frontal function age and its relevance to human amnesia. Behavioral Neuro-
that have dominated recent theories. One view holds science 103, 548560.
Janowsky, J. S., Shimamura, A. P., and Squire, L. R. (1989a). Mem-
that a common function underlies the operation of all ory and metamemory: Comparison between patients with
regions in the frontal lobes but that the function ex- frontal lobe lesions and amnesic patients. Psychobiology 17,
presses itself in diverse ways determined by each re- 311.
gions unique anatomical connections. The other view (1989b). Source memory impairment in patients with fron-
tal lobe lesions. Neuropsychologia 27, 1,0431,056.
does not assume a functional link among various re-
Kelley, W. M. et al. (1998) Hemispheric specialization in human
gions but argues for true functional independence dorsal frontal cortex and medial temporal lobe for verbal and
among them along domain specific lines (Moscovitch nonverbal encoding. Neuron 20, 927-936.
and Umlita, 1990; Goldman Rakic, 2002; Petrides, Kolb, B., and Whishaw, I. Q. (1990). Fundamentals of human neurop-
2002; Stuss and Knight, 2002). The view that prevails sychology, 3rd edition. New York: Freeman.
Kopelman, M. D. (1999). Varieties of false memory. Cognitive
will determine our conception of the mechanisms un- Neuropsychology 16, 197214.
derlying memory, consciousness, and volitional be- Luria, A. R. (1980). Higher cortical functions in man. New York: Basic
havior. Books.
FUGUE STATE 163
Mayes, A. R. (1988). Human organic memory disorders. Cambridge, Petrides, M., and Milner, B. (1982). Deficits on subject-ordered
UK: Cambridge University Press. tasks after frontal and temporal-lobe lesions in man. Neuropsy-
Milner, B. (1966). Amnesia following operation on the temporal chologia 20, 249262.
lobe. In C. W. M. Whitty and O. L. Zangwill, eds., Amnesia. Petrides, M., and Pandya, D. (2002). Association pathways of the
London: Butterworth. prefrontal cortex and functional observations. In D. T. Stuss
Milner, B., Petrides, M., and Smith, M. L. (1985). Frontal lobes and and R. T. Knight, eds., Principles of frontal lobe function. New
the temporal organization of memory. Human Neurobiology 4, York: Oxford University Press.
137142. Rugg, M. D. and Wilding,. L. (2000) Retrieval processing and epi-
Moscovitch, M. (1982). Multiple dissociations of function in amne- sodic memory. Trends in Cognitive Science 4, 108115.
sia. In L. S. Cermak, ed., Human memory and amnesia. Hills- Saint-Cyr, J. A., Taylor, A., and Lang, A. (1988). Procedural learn-
dale, NJ: Erlbaum. ing and neostriatal dysfunction in man. Brain 111, 941959.
(1989). Confabulation and the frontal system: Strategic ver- Schacter, D. L., Harbluk, J. L., and McLachlan, D. R. (1984). Re-
sus associative retrieval in neuropsychological theories of trieval without recollection: An experimental analysis of
memory. In H. L. Roediger and F. I. M. Craik, eds., Varieties source amnesia. Journal of Verbal Learning and Verbal Behavior
of memory and consciousness: Essays in honor of Endel Tulving. 23, 593611.
Hillsdale, NJ: Erlbaum. Schnider, A., and Ptak, R. (1999). Spontaneous confabulators fail
(1994). Interference at retrieval from long-term memory: to suppress currently irrelevant memory traces. Nature Neuro-
The influence of frontal and temporal lobes. Neuropsychology science, 677681.
4, 525534. Schnider, A., von Daniken, C., Gutbrod, K. (1996). The mecha-
Moscovitch, M., Fernandes, M., and Troyer, A. (2002). Working- nisms of spontaneous and provoked confabulations. Brain
with-memory and Cognitive Resources: A component-process 119, 1,3651,375.
account of divided attention and memory. In M. Naveh- Shallice, T. (1982). Specific impairments of planning. Philosophical
Benjamin, M. Moscovitch, and H. L. Roediger, eds., Essays in Transactions of the Royal Society of London B298, 199209.
Honour of Fergus I. M. Craik. New York: Psychology Press. Shallice, T., and Evans, M. D. (1978). The involvement of the fron-
Moscovitch, M., and Melo, B. (1997). Strategic retrieval and the tal lobes in cognitive estimation. Cortex 14, 294303.
frontal lobes: Evidence from confabulation and amnesia. Shimamura, A. P., Janowsky, J. S., and Squire, L. R. (1990). Memo-
Neuropsychologia 35, 1,0171,034. ry for temporal order in patients with frontal lobe lesions and
Moscovitch, M., and Umilta, C. (1990). Modularity and neuropsy- patients with amnesia. Neuropsychologia 28, 803813.
chology: Implications for the organization of attention and Smith, M. L., and Milner, B. (1988). Estimation of frequency of oc-
memory in normal and brain-damaged people. In M. E. currence of abstract designs after frontal or temporal lobecto-
Schwartz, ed., Modular processes in dementia. Cambridge, MA: my. Neuropsychologia 26, 297306.
MIT/Bradford. Stuss, D. T., and Benson, D. F. (1986). The frontal lobes. New York:
(1991). Conscious and nonconscious aspects of memory: A Raven Press.
neuropsychological framework of modules and central sys- Stuss, D. T., and Knight, R. T., eds. (2002) Principles of frontal lobe
tems. In H. J. Weingartner and R. G. Lister. eds., Perspectives function. New York: Oxford University Press.
in cognitive neuroscience. Oxford: Oxford University Press. Teuber, H.-L., (1964). The riddle of frontal lobe function in man.
Moscovitch, M., and Winocur, G. (1992). The neuropsychology of In J. M. Warren and K. Akert, eds., The frontal granular cortex
memory in the elderly: The hippocampus and frontal lobes. and behavior. New York: McGraw-Hill.
In T. Salthouse and F. I. M. Craik, eds., The handbook of cogni- Tulving, E., Kapur, S., Craik, F. I. M., Moscovitch, M., and Houle,
tion and aging. New York: Academic Press. S. (1994). Hemispheric encoding/retrieval asymmetry in epi-
(1995). Frontal lobes, memory and aging. In J. Grafman sodic: Positron emission tomography findings. Proceedings of
and K. Holyoak, eds., Structure and function of human pre- the National Academy of Sciences of the United States of America 91,
frontal cortex. Annals of the New York Academy of Sciences 769, 2,0162,020.
119150. Wagner, A. D., Schacter, D. L., Rotte, M., Koutstaal, W., Maril, A.,
(2002). The frontal lobes ad working with memory. In D. Dale, A. M., Rosen, B. R., and Buckner, R. L. (1998). Building
T. Stuss and R. T. Knight, eds., Principles of frontal lobe function. memories: Remembering and forgetting of verbal experi-
New York: Oxford University Press. ences as predicted by brain activity. Science 281, 1,1881,191.
Nolde, S. F., Johnson, M. K., and Raye, C. L. (1998). The role of Wheeler, M. A., Stuss, D. T., and Tulving, E. (1995). Frontal lobe
prefrontal cortex during tests of episodic retrieval. Trends in damage produces episodic memory impairment. Journal of the
Cognitive Science 2, 399406. International Neuropsychological Society 1, 525536.
Nyberg, L., Winocur, G., and Moscovitch, M. (1997) Correlation Winocur, G., Moscovitch, M., and Stuss, D. T. (1996) A neuropsy-
between frontal-lobe functions and explicit and implicit stem chological investigation of explicit and implicit memory in in-
completion in health elderly. Neuropsychologia 11, 7076. stitutionalized and community-dwelling old people. Neuropsy-
Owen, A. M., Downes, J. J., Sahakian, B. J., Polkeg, C. E., and Rob- chology 10, 5765.
bins, T. W. (1990). Planning and spatial working memory fol-
lowing frontal lobe lesions in man. Neuropsychologia 28, 1,021 Morris Moscovitch
1,034.
Gordon Winocur
Pandya, D., and Barnes, C. I. (1986). Architecture and connections
of the frontal lobes. In E. Perleman, ed., The frontal lobes revis-
ited. New York: IRBN Press.
Petrides, M. (1989). Frontal lobes and memory. In F. Boller and
J. Grafman, eds., Handbook of neuropsychology, Vol. 3. Amster-
dam: Elsevier. FUGUE STATE
(2000). The role of the mid-dorsolateral prefrontal cortex
in working memory. Experimental Brain Research 133, 4454. See: AMNESIA, FUNCTIONAL
G
GENERALIZATION CLASSICAL CONDITIONING: FEAR CONDITIONING,
FREEZING). Induction of fear conditioning alters the ex-
See: DISCRIMINATION AND GENERALIZATION pression of immediate early genes in the amygdala, and
blocking gene expression in the amygdala prevents long-
GENES AND MEMORY term fear memory. Using viral vector gene-transfer technol-
ogy, increasing the expression of the transcriptional activa-
See: APLYSIA: MOLECULAR BASIS OF LONG-TERM
SENSITIZATION; GENETIC SUBSTRATES OF
tor CREB (Cyclic AMP Response Element Binding) in the
MEMORY; PROTEIN SYNTHESIS IN LONG- amygdala enhances fear memory, whereas decreasing CREB
TERM MEMORY IN VERTEBRATES expression in the amygdala impairs fear memory.
Genetic approaches have provided much new informa-
tion concerning the CEREBELLAR substrates of classical eye-
blink condition (see NEURAL SUBSTRATES OF CLASSICAL
GENETIC SUBSTRATES OF MEMORY
CONDITIONING: DISCRETE BEHAVIORAL RESPONSES).
[The understanding of biological substrates of learning and The Purkinje cell degeneration (pcd) mutant mouse loses all
memory has been greatly facilitated by the application of the Purkinje neurons in the cerebellar cortex within four weeks
techniques of molecular biology and genetics. Studies done after birth. These animals are able to learn eyeblink condi-
in the 1960s argued that longer-term memory processes re- tioning, but the learning is slower and to a lower degree
quired protein synthesis (i.e., required certain genes to in- than normal, and it slows extinction. Lesions of the interpo-
crease or alter their expression of proteins). Technology de- situs nucleus in the pcd mouse completely prevent learning.
veloped in the 1990s has made it possible to knock out or So both cerebellar cortex and interpositus are involved but
knock in specific genes experimentally, as well as to make the basic association seems to be formed in the interpositus.
use of naturally occurring mutants. This work has involved Mutant and knock-out mice that are deficient in cerebellar
both invertebrates (e.g., the fruit fly and C. elegans) and long-term depression (LTD) are also much impaired in
vertebrates, mostly mice. Learning is a particularly interest- learning and in adaptive timing of the CR, indicating an
ing aspect of behavior to manipulate genetically since memo- important modulatory role of the cerebellar cortex.
ry involves changes in behavior as a result of experiences in The third article, on the HIPPOCAMPUS, is written by
the environment. As with language in humans, the ability a pioneer in this new field, Alcino Silva. The first studies to
to learn and the mechanisms of memory storage are largely use genetics to manipulate plasticity and learning in mam-
determined genetically, but what is learned depends almost mals (mice) targeted kinases (calcium-calmodulin dependent
entirely on experience. kinase II) and the tyrosine kinase Fyn. These studies showed
The first article in this series focuses on the AMYGDALA that disruption of hippocampal-synaptic plasticity (i.e., long-
and fear conditioning (see NEURAL SUBSTRATES OF term potentiation, LTP) resulted in hippocampal-dependent
165
166 GENETIC SUBSTRATES OF MEMORY: Amygdala
learning deficits. Recent approaches involve targeted, selec- ioral reactions to those signals (Helmstetter and Bell-
tive, temporally controlled expression of normal or altered gowan, 1994). Changes in gene expression within
genes, an extraordinarily promising technology.] amygdala neurons may accompany both learning and
response performance in fear conditioning.
AMYGDALA
Learning Can Alter Gene Expression in the
Scientists believe that the amygdala plays a critical Amygdala
role in the ability to learn and remember certain types
A number of studies have been able to show that
of information. Converging evidence from studies
exposure to the training procedures used in Pavlov-
using humans as well as a variety of laboratory species
ian fear conditioning will selectively alter the expres-
support the idea that the amygdala, together with a
sion of messenger RNA (mRNA) or protein in the
network of closely interconnected brain regions, al-
amygdala. Immediate early genes (IEGs) tend to be
lows people to store long-term memories for events
minimally expressed in quiescent cells but are often
that are emotionally important and to respond ap-
rapidly and dramatically activated when cells are
propriately to stimuli in the environment that signal
stimulated by synaptic inputs. Many IEGs encode
threat. The ability to store long-term memories of this
transcription factors that ultimately regulate other
type is likely due to experience-dependent changes in
gene products expressed later after stimulation. In
gene expression within amygdala neurons. Many sci-
one of the first studies of gene expression during fear
entists and clinicians believe that understanding the
conditioning, Serge Campeau and colleagues (1991)
functions of the amygdala may be the key to the treat-
showed that the amount of mRNA coding c-fos in the
ment and prevention of human anxiety disorders and
amygdala was elevated by exposing rats either to foot
other psychiatric problems.
shock itself, or to an environment that had been pre-
Scientific knowledge about how the amygdala viously paired with foot shock. Therefore, the time-
works comes largely from the development and wide- dependent storage of new memory following the
spread use of Pavlovian fear-conditioning with ro- shock experience as well as the neuronal activity in
dents as a model system throughout the late 1980s the amygdala provoked by exposure to a fear-
and 1990s. In a typical fear-conditioning experiment, producing stimulus may increase production of FOS
a rat is placed in a distinctive observation chamber protein. Subsequent experiments have continued to
where a simple signal such as a flashing light or a tone shed light on this and related phenomena. Several
(the conditional stimulus, or CS) is repeatedly activat- different aspects of the fear conditioning procedure
ed prior to delivery of a mild electric shock to the ani- appear to activate c-fos transcription and the level of
mals feet. The rat will learn that the signal predicts expression of FOS protein correlates with how well
shock and will display species-appropriate fear behav- animals learn during fear conditioning (Radulovic,
iors when presented with the CS alone after training. Kammermeir, and Spiess, 1998).
The animal will also learn to fear the apparatus, or
context, in which the training took place. This rela- J. B. Rosen and colleagues (1998) reported that
tively simple form of learning is rapidly acquired, while mRNA for c-fos was elevated in the amygdala in
easy to observe in the laboratory, retained indefinite- animals that learned fear conditioning as well as in
ly, and completely disrupted in animals with signifi- control subjects that did not learn, another IEG,
cant damage to the amygdala (Davis, 2000; LeDoux, early growth response gene 1 (egr-1), showed a pat-
2000; Maren, 2001). tern of message expression in the lateral nucleus of
the amygdala that was specific to learning. Egr-1
Fear conditioning experiments have shown that
mRNA is selectively expressed in the amygdala be-
discrete anatomical regions within the amygdala may
tween fifteen and thirty minutes after learning and
play selective roles in the learning process. The later-
only in the region of the amygdala likely to undergo
al nucleus and closely related cell groups within the
critical synaptic plasticity during the formation of
basolateral region of the amygdala receive direct and
memory. Importantly, simple retrieval of fear memo-
indirect synaptic input from sensory systems that de-
ry and performance of fear responses do not alter
tect and process stimuli in the animals environment.
egr-1 expression supporting the idea that this IEG,
On the other hand, the central nucleus of the amyg-
unlike c-fos, may play a selective role in the formation
dala appears to play a key role in the expression of
of long-term memory (Mackani and Rosen, 2000).
behavioral and physiological reactions to fear-
provoking stimuli through its extensive connections Alterations in gene expression in the amygdala
with the brain stem. The amygdala is important for are not restricted to IEGs nor are they solely related
learning about new danger signals in the environ- to the initial storage of information. Cyclic AMP re-
ment as well as for producing the appropriate behav- sponsive-element binding protein (CREB) is a tran-
GENETIC SUBSTRATES OF MEMORY: Amygdala 167
Figure 1
Rats treated with a drug that reduces the formation of new mRNA in the amygdala fail to remember their experiences during fear
conditioning.
scription factor that plays a key role in memory for- D. J. Bailey and colleagues (1999) took a first step
mation in a number of species (Silva, Kogan, in this direction by asking a more general question.
Frankland, and Kida, 1998). In mice exposed to a They tested whether or not the general ability of cells
simple fear-conditioning protocol, levels of phospho- in the amygdala to synthesize new mRNA is critical
rylated CREB (pCREB) in the amygdala were elevat- for the formation of long-term memory in fear condi-
ed in trained animals relative to controls during the tioning. Actinomycin-D, a drug that selectivity pre-
memory consolidation period after conditioning vents cells from copying information from DNA in
(Stancui, Radulovic, and Speiss, 2001). Increased the nucleus onto new mRNA molecules (transcrip-
phosphorylation and thus activity of CREB is also tion), was microinjected into the amygdala of rats
seen in the amygdala when rats retrieve memories prior to training with foot shock. Twenty-four hours
learned in fear conditioning that have already been after training the animals were returned to the labo-
stored (Hall, Thomas, and Everitt, 2001). Retrieval ratory and tested for memory of what was learned the
and/or emotional responding also causes cells in the day before. The results of this experiment are sum-
marized in Figure 1. Suppressing transcription of new
central nucleus to increase production of mRNA for
mRNA in the amygdala greatly disrupted the rats
the neuropeptide enkephalin (Petrovich, Scicli,
ability to remember what happened during training.
Thompson, and Swanson, 2000).
This experiment was one of the first direct sources of
evidence that mRNA synthesis, and by implication
the synthesis of new protein, is required in the amyg-
Manipulating Gene Expression in the dala during the acquisition of fear learning.
Amygdala Can Affect Learning
Several other studies have supported the idea
Demonstrating that cells in the amygdala express that the synthesis of new proteins in the amygdala is
a different quantity or a unique pattern of gene prod- critical for memory formation. The molecular events
ucts while animals are learning is an important step occurring during the period directly after training
toward understanding the mechanisms of long-term when long-term memories are consolidated are of
memory. However, one can see that just because a particular interest. Blocking the translation of mRNA
particular protein or mRNA appears to change when into new protein in the amygdala during this post-
subjects learn, it does not necessarily follow that this training period prevents long-term retention of fear
gene product is essential for formation of the memo- conditioning, as does selective disruption of the activ-
ry. In order to know if something is required for ity of the enzyme protein kinase A (PKA) (Schafe et
memory formation, one must be able to manipulate al., 2000). In taste-aversion learning, where animals
it directly. associate a specific taste with illness, pretraining
168 GENETIC SUBSTRATES OF MEMORY: Amygdala
amygdala injections of antisense oligodeoxynucleo- Bibliography

tides (ODNs) that selectively prevent the translation Bailey, D. J., Kim, J. J., Sun, W., Thompson, R. F., and Helmstet-
of c-fos mRNA into protein prevent the formation of ter, F. J. (1999). Acquisition of fear conditioning in rats re-
taste-aversion memory (Lamprecht and Dudai, quires the synthesis of mRNA in the amygdala. Behavioral
1996).
Campeau, S., Hayward, D., Hope, B. T., Rosen, J. B., Nestler, E.
Studies in which mRNA transcription or protein J., and Davis, M. (1991). Induction of the c-fos proto-oncogene
in rat amygdala during unconditioned and conditioned fear.
translation is globally affected within amygdala neu-
Brain Research 565, 349352.
rons are somewhat limited given their lack of specific- Davis, M. (2000). The role of the amygdala in conditioned and
ity. Suppression of specific proteins with antisense unconditioned fear and anxiety. In J. P. Aggleton, ed., The
ODNs has several advantages but also suffers from amygdala: A functional analysis. New York: Oxford University
important limitations. One of the most exciting re- Press.
cent technical advances in this area involves the use Hall, J., Thomas, K. L., and Everitt, B. J. (2001). Fear memory re-
trieval induces CREB phosphorylation and Fos expression
of viral-vector gene-transfer technology to selectively
within the amygdala. European Journal of Neuroscience 13 (7),
introduce and promote the expression of target gene 1,4531,458.
products within restricted brain regions. For exam- Helmstetter, F. J., and Bellgowan, P. S. (1994). Effects of muscimol
ple, S. A. Josselyn and colleagues (2001) were able to applied to the basolateral amygdala on acquisition and ex-
cause the selective overexpression of CREB in rat pression of contextual fear conditioning in rats. Behavioral
amygdala neurons prior to training in fear condition- Neuroscience 108, 1,0051,009.
ing. Rats that had significantly more CREB being ex- Josselyn, S. A., Shi, C., Carlezon, W. A., Neve, R. L., Nestler, E. J.,
and Davis, M. (2001). Long-term memory is facilitated by
pressed actually showed better memory than control
cAMP response element-binding protein overexpression in
animals. This type of evidence, when coupled with the the amygdala. Journal of Neuroscience 21, 2,4042,412.
variety of experiments showing that down regulation Lamprecht, R., and Dudai, Y. (1996). Transient expression of c-fos
of CREB impairs memory, makes a very convincing in rat amygdala during training is required for encoding con-
case for the role of this particular transcription factor ditioned taste aversion memory. Learning and Memory 3, 31
in memory formation. 41.
LeDoux, J. E. (2000). Emotion circuits in the brain. Annual Review
of Neuroscience 23, 155184.
Malkani, S., and Rosen, J. B. (2000). Specific induction of early
Conclusion growth response gene 1 in the lateral nucleus of the amygdala
Neuroscientists are developing a clear picture of following contextual fear conditioning in rats. Neuroscience 97,
693702.
how the amygdala contributes to memory at the cir-
Maren, S. (2001). Neurobiology of Pavlovian fear conditioning. An-
cuit and neural systems level. Comparatively little is nual Review of Neuroscience 24, 897931.
known about the molecular events taking place within Petrovich, G. D, Scicli, A. P., Thompson, R. F., and Swanson, L. W.
amygdala neurons as an organism learns. Several (2000). Associative fear conditioning of enkephalin mRNA le-
gene products have been identified that undergo vels in central amygdalar neurons. Behavioral Neuroscience
changes in expression that appear to be related to 114, 681686.
Radulovic, J., Kammermeier, J., and Spiess, J. (1998). Relationship
learning. In some cases, direct manipulation of these
between fos production and classical fear conditioning:
proteins can have an effect on the formation of mem- Effects of novelty, latent inhibition, and unconditioned
ory. As interest in this area increases and the technol- stimulus preexposure. Journal of Neuroscience 18, 7,452
ogy available to scientists improves, new genes that 7,461.
are involved in amygdala-dependent memory phe- Rosen, J. B., Fanselow, M. S., Young, S. L., Sitcoske, M., and
nomena will continue to be described. Maren, S. (1998). Immediate-early gene expression in the
amygdala following footshock stress and contextual fear con-
ditioning. Brain Research 796, 132142.
See also: EMOTION, MOOD, AND MEMORY; GENETIC Schafe, G. E., Atkins, C. M., Swank, M. W., Bauer, E. P., Sweatt, J.
SUBSTRATES OF MEMORY: CEREBELLUM; D., and LeDoux, J. E. (2000). Activation of ERK/MAP kinase
GENETIC SUBSTRATES OF MEMORY: in the amygdala is required for memory consolidation of Pav-
lovian fear conditioning. Journal of Neuroscience 20 (21),
HIPPOCAMPUS; GUIDE TO THE ANATOMY OF
8,1778,187.
THE BRAIN: AMYGDALA; HORMONES AND
Silva, A. J., Kogan, J. H., Frankland, P. W., and Kida, S. (1998).
MEMORY; LONG-TERM POTENTIATION:
CREB and memory. Annual Review of Neuroscience 21,127
AMYGDALA; NEURAL SUBSTRATES OF CLASSICAL 148.
CONDITIONING: FEAR CONDITIONING, Stanciu, M., Radulovic, J., and Spiess, J. (2001). Phosphorylated
FREEZING; NEURAL SUBSTRATES OF CLASSICAL cAMP response element binding protein in the mouse brain
CONDITIONING: FEAR-POTENTIATED STARTLE; after fear conditioning: relationship to Fos production. Molec-
NEURAL SUBSTRATES OF EMOTIONAL MEMORY; ular Brain Research 94, 1524.
PROTEIN SYNTHESIS IN LONG-TERM MEMORY IN
VERTEBRATES; TASTE AVERSION AND
PREFERENCE LEARNING IN ANIMALS Fred J. Helmstetter
GENETIC SUBSTRATES OF MEMORY: Cerebellum 169
CEREBELLUM the entire cerebellar cortex without damaging the

deep nuclei. Lu Chen and colleagues (1996) exam-
The cerebellum has been traditionally regarded as a ined classical eyeblink conditioning in Purkinje-cell
part of the motor system, mainly because of clinical degeneration (pcd) mutant mice to dissect the compo-
and experimental observations that humans and ani- nents of the cerebellar circuitry that are crucial for
mals with cerebellar damage are impaired in coordi- learning. These pcd mutant mice are born with Pur-
nated movement. Early theoretical consideration in- kinje cells, but gradually lose all of them during the
dicated that the unique and stereotyped cellular first four weeks of postnatal development. Other neu-
organization of the cerebellar cortex could also func- ral degenerations occur much later, allowing a time
tion as a learning machine (Albus, 1971; Marr, 1969). window to examine behavioral consequences associat-
The cerebellar-learning hypotheses inspired broad ed with the loss of Purkinje cells. Because the Purkinje
experimental investigations. In 1982, Masao Ito and cells are the only output neurons in the cerebellar cor-
colleagues discovered that conjoint activation of the tex, pcd mice are functionally equivalent to animals
mossy fiber and climbing fiber led to long-lasting de- with complete cortical lesions. Classical delay (i.e., the
pression of the Purkinje-cell responses to mossy fiber US follows and overlays with the CS) eyeblink condi-
activation, known as cerebellar long-term depression tioning was impaired in the pcd mice in three aspects:
(LTD). Behavioral studies also indicate that the cere- the learning was slower; the maximal level of learning
bellum is involved in several forms of motor learning, was lower; and the learned responses occurred earlier
including classical eyeblink conditioning and adapta- and were hence maladaptive. The low level of residual
tion of vestibulo-ocular reflex. In the 1990s, consider- learning in pcd mice was completely abolished by le-
able progress was made in understanding the cellular sions of the cerebellar deep nuclei, indicating the
and molecular mechanisms underlying cerebellum- learned eyeblink response is stored in both the cere-
dependent learning and memory. A number of spon- bellar cortex and the deep nuclei (Chen, Bao, and
taneous mouse mutants with relatively specific cere- Thomson, 1999).
bellar deficits proved useful in dissecting the cerebel-
lar circuitry involved in learning. Studies of gene
knockout mice provided much insight into the molec- The Conditioned-Stimulus Pathway
ular and cellular basis of learning and memory.
The stargazer is another mutant with deficient cer-
ebellar cortical circuit. The gene coding for stargazin,
Classical Eyeblink Conditioning a protein that targets a-amino-3-hydroxy-5-
Classical eyeblink conditioning is a form of motor methylisoxazoleproprionic acid (AMPA) receptors to
learning, in which animals learn to associate an innoc- their synaptic locations, is truncated in this mutant.
uous stimulus to a noxious one and to respond adap- The mutation results in a loss of AMPA receptors at
tively. For example, a naive animal will blink to a mild the cerebellar mossy fibergranule cell synapses,
eyelid electrical shock but not to a brief tone. If the thereby disrupting the conditioned-stimulus pathway
tone (serving as a conditioned stimulus, CS) is repeat- to the cerebellar cortex (Chen et al., 2000).This
edly paired with the eyelid shock (serving as an un- AMPA receptor defect is observed only in the cerebel-
conditioned stimulus, US), the animal will gradually lar granule cells, possibly because in other neurons
learn to respond by eye blink to the tone. Richard F. the loss of stargazing is compensated for by expres-
Thompson and colleagues, among others, have char- sion of functionally similar proteins. These stargazer
acterized the brain circuitry involved in this learning. mutant mice were impaired in classical eyeblink con-
In essence, conditioned stimulus information is conditioning in similar ways to the pcd mutant mice,
veyed through mossy fibers, and unconditioned stim- showing slower and reduced learning and disrupted
ulus information is conveyed through climbing fibers. response timing (Qiao et al., 1998). Further, the mu-
These converge in the cerebellum, where learning oc- tant mice performed normally in classical fear condi-
curs and memory is stored (Thompson et al., 1996). tioning with a tone CS and a foot-shock US, a task that
does not involve the cerebellum.
Cerebellar Cortex
Both the cerebellar cortex and the deep nuclei re- The Unconditioned-Stimulus Pathway
ceive mossy fiber and climbing fiber projections and In early stages of rodent postnatal development,
therefore are potentially capable of supporting learn- each Purkinje cell is innervated by several climbing fi-
ing and memory of the conditioned eyeblink re- bers. By the end of third postnatal week, all but one
sponse. The relative importance of the cortex and the of the climbing fibers are eliminated. In some mu-
deep nuclei in classical eyeblink conditioning is un- tants, a large portion of their Purkinje cells remain in-
clear, partly because it is virtually impossible to lesion nervated by more than one climbing fiber even in
170 GENETIC SUBSTRATES OF MEMORY: Cerebellum
adulthood. These mutants provide useful tools to Vestibulo-Ocular Reflex Adaptation

study roles of climbing fibers as the US pathway in
Vestibulo-ocular reflex (VOR) adaptation is an-
eyeblink conditioning.
other form of motor learning that depends on an in-
For example, in the adult mutant mice deficient tact cerebellum. An animal normally moves its eyes
in the isoform of protein kinase C (PKC), 30 per- opposite in direction to its head turn to keep visual
cent of the Purkinje cells are innervated by multiple images steady on the retina. This reflex is guided by
climbing fibers. Other cerebellar physiological prop- vestibular inputs (hence called vestibulo-ocular re-
erties examined are all normal in this mutant. Multi- flex) and is present even in complete darkness. If an
ple climbing fiber innervation presumably enhances erroneous motion of images on retina is artificially in-
US input (i.e., the teaching signal) to the cortex. In troduced, for example, by fitting spectacles on the
agreement with this view, PKC mutant mice showed eyes or by moving the visual scene, the animal learns
facilitated eyeblink learning (Chen et al., 1995). to adjust the amplitude of VOR to compensate for the
Whereas wild-type mice attained maximal level of error. Cerebellar LTD has been hypothesized as a
learning in five days, the mutant mice reached the molecular mechanism for this VOR adaptation.
same level in only two days. In addition, the learned Chris I. De Zeeuw and colleagues (1998) exam-
eyeblink responses in the PKC mutant mice were ined cerebellar LTD and VOR adaptation in L7-PKCI
more resistant to extinction by presentations of CS transgenic mice in which the protein kinase C (PKC)
alone. inhibitor is artificially expressed in the cerebellar Pur-
kinje cells under the control of L7 promoter. Protein
kinase C, required for cerebellar LTD, has several iso-
Cerebellar Long-Term Depression forms coded by separate genes. Knockout of the iso-
Conjoint activation of the mossy parallel fiber form did not impair cerebellar LTD (see discussion
and the climbing fiber induces long-term depression above on PKC knockout mice), possibly due to com-
of the parallel fiberPurkinje cell synaptic transmis- pensatory expression of other isoforms of the kinase.
sion, which is considered a cellular mechanism for In L7-PKCI mice, expression of a selective inhibitor
some forms of motor learning such as classical eye- to a broad range of PKC isoforms virtually blocked
blink conditioning. Cerebellar LTD probably reduces LTD induction. These L7-PKCI mice showed normal
Purkinje-cell responses, resulting in increased activity VOR. When subjected to visuo-vestibular training, in
in the deep cerebellar nuclei, which may code for the which the visual scene was rotated in addition to the
learned responses. Testing the role of cerebellar LTD head turn, the wild-type mice learned to adapt their
in motor learning had been hindered by the lack of VOR. In contrast, L7-PKCI exhibited no VOR adap-
pharmacological agents that specifically block cere- tation. These results suggest that cerebellar LTD is a
bellar LTD and by the difficulty in applying the mechanism for VOR adaptation.
agents to the entire cerebellar cortex. The advent of
modern molecular biology made it possible to com-
pletely remove specific proteins from a mouse by de- Complex Motor-Skill Learning
leting the corresponding genes from the mouse ge- In addition to classical eyeblink conditioning and
nome. Using this powerful technique, a number of VOR adaptation, the learning of more complex
gene knockout mice with deficient cerebellar LTD motor skills has also been examined in many mutant
were generated, including the mice lacking mGluR1 mice with what is know as the rotorod test. Typically,
(type 1 metabotropic glutamate receptor), GFAP a mouse is placed on either a stationary rod or a rod
(glial fibrillary acidic protein), GluR2 (glutamate re- that rotates at a certain speed, and the time it remains
ceptor 2 subunit) or PLC4 (phospholipase C 4 iso- on the rod is measured to assess motor-skill learning.
form). Behavioral examinations indicated that they Impaired motor-skill learning has been observed in
were all impaired in classical eyeblink conditioning various mutant mice with cerebellar deficits including
(Aiba et al., 1994; Kishimoto et al., 2001; Miyata et al., altered Purkinje-cell excitability, altered short-term
2001; Shibuki et al., 1996). synaptic plasticity, impaired cerebellar long-term po-
tentiation, impaired cerebellar LTD, and multiple
Mutation of a gene may have widespread effects
climbing fiber innervation.
in the brain, making any results difficult to interpret.
In that respect, results obtained in GluR2 knockout Mutant mice with deficient cerebellar LTD are
mice are particularly informative, because GluR2 is generally impaired in learning complex motor skills,
expressed only in the dendritic spines of the Purkinje with the exception of the GFAP knockout mice. As
cells. Impairments in both cerebellar LTD and eye- mentioned before, the GFAP knockout mice showed
blink learning suggest that cerebellar LTD is involved deficient cerebellar LTD and impaired eyeblink con-
in learning conditioned eyeblink responses. ditioning. However, they learned to stay on the rotat-
GENETIC SUBSTRATES OF MEMORY: Hippocampus 171
ing rod as well as normal mice, suggesting that there Chen, C., Kano, M., Abeliovich, A., Chen, L., Bao, S., Kim, J. J.,
may be other forms of cerebellar plasticity supporting Hashimoto, K., Thompson, R. F., and Tonegawa, S. (1995).
Impaired motor coordination correlates with persistent multi-
motor-skill learning. ple climbing fiber innervation in PKC gamma mutant mice.
Chong Chen and colleagues (1995) proposed Cell 83 (7), 1,2331,242.
that multiple climbing fiber innervation of the Pur- De Zeeuw, C. I., Hansel, C., Bian, F., Koekkoek, S. K., van Alphen,
A. M., Linden, D. J., and Oberdick, J. (1998). Expression of
kinje cells disrupts learning of component movement a protein kinase C inhibitor in Purkinje cells blocks cerebellar
and, hence, learning of complex motor skills. This LTD and adaptation of the vestibulo-ocular reflex. Neuron 20
notion is supported by numerous studies but has been (3), 495508.
challenged by De Zeeuw and colleagues (1998). These Goodlett, C. R., Hamre, K. M., and West, J. R. (1992). Dissociation
investigators reported normal rotorod performance of spatial navigation and visual guidance performance in Pur-
kinje cell degeneration (pcd) mutant mice. Behavioral Brain
and multiple climbing fiber innervations in the L7- Research 47 (2), 129141.
PKCI transgenic mice. Kishimoto, Y., Kawahara, S., Fujimichi, R., Mori, H., Mishina, M.,
and Kirino, Y. (2001). Impairment of eyeblink conditioning
in GluRdelta2-mutant mice depends on the temporal overlap
Conclusion between conditioned and unconditioned stimuli. European
Impaired motor learning in the cerebellum- Journal of Neuroscience 14 (9), 1,5151,521.
specific mutants such as the GluR2 knockout and the Marr, D. (1969). A theory of cerebellar cortex. Journal of Physiology
202 (2), 437470.
L7-PKCI transgenic mice provides convincing evi- Miyata, M., Kim, H. T., Hashimoto, K., Lee, T. K., Cho, S. Y.,
dence that the cerebellum is a learning machine. De- Jiang, H., Wu, Y., Jun, K., Wu, D., Kano, M., and Shin, H. S.
ficient cerebellar LTD and motor learning have been (2001). Deficient long-term synaptic depression in the rostral
observed in vastly different mutants, suggesting a cerebellum correlated with impaired motor learning in pho-
causal link between cerebellar LTD and motor learn- spholipase C beta4 mutant mice. European Journal of Neuro-
science 13 (10), 1,9451,954.
ing. In addition to cerebellar LTD, other forms of Qiao, X., Chen, L., Gao, H., Bao, S., Hefti, F., Thompson, R. F.,
synaptic plasticity may also be involved in cerebellum- and Knusel, B. (1998). Cerebellar brain-derived neurotrophic
dependent learning. The cerebellum may also be in- factor-TrkB defect associated with impairment of eyeblink
volved in spatial learning (Goodlett, Hamre, and conditioning in Stargazer mutant mice. Journal of Neuroscience
West, 1992). Further studies using cerebellum- 18 (17), 6,9906,999.
Shibuki, K., Gomi, H., Chen, L., Bao, S., Kim, J. J., Wakatsuki, H.,
specific mutant mice should help illuminate some Fujisaki, T., Fujimoto, K., Katoh, A., Ikeda, T., Chen, C.,
other aspects of this question. Thompson, R. F., and Itohara, S. (1996). Deficient cerebellar
long-term depression, impaired eyeblink conditioning, and
See also: GENETIC SUBSTRATES OF MEMORY: normal motor coordination in GFAP mutant mice. Neuron 16
AMYGDALA; GENETIC SUBSTRATES OF MEMORY: (3), 587599.
HIPPOCAMPUS; GLUTAMATE RECEPTORS AND Thompson, R. F., Bao, S., Chen, L., Cipriano, B., Grethe, J. S.,
THEIR CHARACTERIZATION; LOCALIZATION OF Kim, J. J., Thompson, J. K., Tracy, J., Weninger, M. S., and
MEMORY TRACES; LONG-TERM DEPRESSION IN Krupa, D. J. (1996). In J. D. Schmahmann, ed., Associative
THE CEREBELLUM, HIPPOCAMPUS, AND learning. The Cerebellum and Cognition,151189. Academic
NEOCORTEX; NEURAL COMPUTATION: Press, New York.
CEREBELLUM; NEURAL SUBSTRATES OF Shaowen Bao
CLASSICAL CONDITIONING: DISCRETE Lu Chen
BEHAVIORAL RESPONSES; REINFORCEMENT;
VESTIBULO-OCULAR REFLEX (VOR) PLASTICITY
Bibliography
Aiba, A., Kano, M., Chen, C., Stanton, M. E., Fox, G. D., Herrup, HIPPOCAMPUS
K., Zwingman, T. A., and Tonegawa, S. (1994). Deficient cere- The identification and manipulation of genes (genet-
bellar long-term depression and impaired motor learning in
mGluR1 mutant mice. Cell 79 (2), 377388. ics) has influenced the study of most biological phe-
Albus, J. S. (1971). A theory of cerebellar function. Mathematical nomena, including hippocampal learning and mem-
Bioscience 10, 2561. ory. The first studies to use genetics to manipulate
Chen, L., Bao, S., Lockard, J. M., Kim, J. K., and Thompson, R. plasticity and learning in mammals targeted the -
F. (1996). Impaired classical eyeblink conditioning in cerebel- calcium calmodulin-dependent kinase II (CaMKII)
lar-lesioned and Purkinje cell degeneration (pcd) mutant
mice. Journal of Neuroscience 16 (8), 2,8292,838. (Silva, Paylor, Wehner, and Tonegawa, 1992a; Silva,
Chen, L., Bao, S., and Thompson, R. F. (1999). Bilateral lesions of Stevens, Tonegawa, and Tang, 1992b) and the tyro-
the interpositus nucleus completely prevent eyeblink condi- sine kinase Fyn (Grant et al., 1992). The results
tioning in Purkinje cell-degeneration mutant mice. Behavioral showed that genetic manipulations that disrupt hip-
Neuroscience 113 (1), 204210. pocampal synaptic plasticity result in hippocampal-
Chen, L., Chetkovich, D. M., Petralia, R. S., Sweeney, N. T., Kawa-
saki, Y., Wenthold, R. J., Bredt, D. S., and Nicoll, R. A. (2000). dependent learning deficits. These studies were also
Stargazin regulates synaptic targeting of AMPA receptors by influential in starting the new field of molecular and
two distinct mechanisms. Nature 408 (6,815), 936943. cellular cognition, which combines genetics, neuro-
172 GENETIC SUBSTRATES OF MEMORY: Hippocampus
physiology, and behavioral neuroscience. What have with calcium. The CaMKII is expressed in postnatal
we learned so far about the hippocampus from this forebrain structures such as the hippocampus and
new field, and how successful has this general ap- cortex. Studies with pharmacological inhibitors dem-
proach been? onstrated that the CaMKII family of kinases was in-
volved in the induction of LTP. These kinases, partic-
The hippocampus plays a key role in learning
and memory. Patients with hippocampal lesions are ularly the and the , can potentiate synaptic
unable to store new conscious memories, with more transmission by a variety of mechanisms, including
severe impairments in recent memories than in re- the phosphorylation of glutamate receptors (Lisman
mote ones. Hippocampal lesions spare other cogni- and McIntyre, 2001). Transgenetic studies showed
tive functions, including intelligence, attention, and that both a null mutation of CaMKII (Silva, Stevens,
emotion. Together with parallel animal experiments, Tonegawa, and Tang, 1992b; Silva, Paylor, Wehner,
studies of the hippocampus established that memory and Tonegawa, 1992a) and a transgenic overexpres-
is separate from other cognitive abilities and that the sion of a constitutively active form of this kinase
hippocampus has a prominent role in early stages of (CaMKIIT286D) (Bach et al., 1995; Mayford, Wang,
memory consolidation (Eichenbaum, 2001). Kandel, and ODell, 1995) altered hippocampal LTP
and hippocampal-dependent learning. Later studies
Nearly twenty years after the publication of the showed that a point-mutation that substituted threo-
first hippocampal lesion studies, Bliss, Lomo and col- nine at position 286 for alanine (T286A) and prevent-
leagues discovered long-term potentiation or LTP in ed the autophosphorylation of CaMKII at threonine
the hippocampus, a form of synaptic plasticity that 286 impaired hippocampal LTP and learning (Giese,
strongly suggests a role in learning and memory. Fedorov, Filipkowski, and Silva, 1998). CaMKII,
Moreover, computer simulations (parallel neuronet- however, affects activity-dependent structural plastic-
works) implementing abstractions of LTP-like mecha- ity; hence the learning abnormalities described for
nisms process information in ways that are reminis- these three mutants could be due to deficits in later
cent of human and animal learning, suggesting that stages of hippocampal development.
synaptic phenomena like LTP could be mediating
learning and memory. But is LTP a necessary compo- To address this problem, the tetracycline-
nent of learning and memory mediated by the hippo- controlled transactivator system (tTA) was used to
campus? regulate the expression of a constitutive active form
of CaMKII that disrupts LTP and learning (CaM-
Pharmacological inhibitors of the N-Methyl-D- KIIT286D) (Mayford et al., 1996). With this system it
Aspartate receptor (NMDAR) were used in the earli- was possible to repress the mutant CaMKIIT286D
est attempts to answer this question. The NMDAR transgene during development and simply lift the re-
seems to be a coincidence detector. The receptor reg- pression at appropriate experimental times. These
ulates a calcium channel that is normally blocked by
inducible studies confirmed the importance of C-
magnesium; binding of glutamate and postsynaptic
aMKII function for hippocampal LTP and learning
depolarization remove the magnesium block of the
(Mayford et al., 1996). However, CaMKII has a wide
channel, allowing calcium to enter the synapse and
substrate specificity that is normally restricted by the
induce LTP. NMDAR antagonists impair hippocam-
localization of the enzyme. Thus, it is possible that the
pal LTP and cause severe deficits in hippocampal-
higher levels of constitutively active kinase in the in-
dependent spatial learning tasks (Morris, Anderson,
ducible transgenic mice may have led to the phospho-
Lynch, and Baudry, 1986). However, NMDAR block-
rylation of proteins that are not normally phosphoryl-
ers have a number of other behavioral effects, such as
ated by this kinase. Consequently, the learning
sensimotor deficits, that complicated the interpreta-
deficits in these tTA-CaMKIIT286D mutants could
tion of those studies. When the first NMDAR studies
be an artifact of the overexpression of this constitu-
were published, there were few other pharmacologi-
tively active kinase.
cal manipulations that could be used to investigate
the possible role of LTP in learning. Fortunately, To circumvent this problem, researchers used a
transgenetic manipulations (transgenics, knock outs, new approach that combines pharmacology and ge-
knock ins, and so on) provided the means to modify netics to test the role of CaMKII in LTP and learn-
or delete any one of the molecular components ing (Ohno et al., 2001). This pharmacogenetic strate-
known to be involved in LTP (i.e. CaMKII). gy takes advantage of synergisms between pharma-
cological and genetic manipulations. For example,
unlike the homozygous T286A mutation of CaMKII
CaMKII, TP, and Learning described above, the heterozygous mutation (T286A
CaMKII is a member of a family of serine/ hets) does not affect hippocampal-dependent learn-
threonine kinases activated by calmodulin loaded ing. Similarly, although 10 mg/kg of CPP (NMDAR
GENETIC SUBSTRATES OF MEMORY: Hippocampus 173
blocker) injected intraperitoneally thirty minutes be- The cAMP-signaling pathway is also involved in
fore training blocks contextual learning (hippocam- hippocampal learning and memory. A number of
pal-dependent), 5 mg/kg of this drug does not. This studies indicated that a balance between the activities
same low dose, however, can reveal a contextual of cAMP-dependent protein kinase A (PKA) and the
learning deficit in the T286A heterozygotes, thus phosphatases PP1 and calcineurin gate the stability of
making a compelling connection between kinase ac- both synaptic changes and memory. For example,
tivity and contextual learning (Ohno et al., 2001). Al- transgenic mice expressing R(AB), an inhibitory form
though each of the experiments presented above suf- of the regulatory subunit of PKA, show unstable LTP
fers from specific technical limitations, taken together and memory (Abel et al., 1997). Transgenic expres-
the results presented demonstrate that the activation sion of a constitutively active form of calcineurin, a
of CaMKII is critical for LTP and for hippocampal- Ca/CaM-dependent Ser/Thr phosphatase, also results
dependent learning. This convergence of informa- in unstable LTP and memory. Importantly, repres-
tion is critical for every major finding described here. sion of this calcineurin transgene, under the control
of the tTA system, reverses the long-term memory
deficits of the mutants, demonstrating that these ef-
Detecting Coincidences with the NMDAR fects are not due to the developmental expression of
Learning is heavily dependent on the generation the transgene (Mansuy et al., 1998a).
of associations between previously unrelated phe-
nomena. As described above, the molecular proper- Inducible overexpression of calcineurin with a
ties of the NMDAR suggest that this receptor may modified tetracycline system (rtTa) also resulted in
have a role in these associations. A number of trans- unstable LTP and memory (Mansuy et al., 1998b).
genetic NMDAR manipulations affect both hippo- The rtTA system was also used to overexpress a
campal synaptic plasticity and learning. A deletion of COOH-terminal autoinhibitory domain that repress-
the NMDAR epsilon subunit and a point mutation of es calcineurin function (rtTA-CN inhibitor mice). In-
the glycine site in the NMDAR1 subunit (glycine ducible overexpression of this inhibitory domain en-
binding potentiates receptor function) disrupt hippo- hances early and late phases of CA1 LTP as well as
campal LTP and learning. Remarkably, a Cre- short- and long-term memory (Malleret et al., 2001).
mediated deletion of the NMDAR1 subunit restricted The studies described above indicate that calcineurin
to hippocampal CA1 pyramidal neurons also disrupts and PKA play a critical role in one of the mechanisms
CA1 LTP and hippocampal-dependent learning that gate the stability of plasticity and memory. But
(Tsien et al., 1996), indicating that NMDAR function what are the molecular mechanisms responsible for
in this hippocampal subregion is critical for learning. the stability of synaptic changes and memory?
The bacterial phage Cre-recombinase can delete any
genomic segment flanked by its recognition sites
(LoxP). Thus, any gene flanked by LoxP sites can be Transcription, Translation, and Memory
deleted from any region or cell type expressing Cre- A large body of work in a number of organisms
recombinase. This strategy could be used to delete and memory systems have demonstrated a universal
any gene from anywhere in the brain. Specific brain requirement for transcription and translation in long-
regions or cell types can even be targeted by using term memory (Davis and Squire, 1984)). Moreover,
virus vectors expressing Cre-recombinase. pioneering studies in Aplysia showed the involvement
of the transcription factor cAMP Responsive Element
Binding protein (CREB) in the stability of synaptic
The Many Roads of Plasticity and Learning changes. This transcription factor is activated by a
Aside from NMDAR-dependent activation of number of signaling pathways, including ERK and
CaMKII, there are many other signaling pathways the cAMP cascades (see above). Interestingly, a null
that are critical for both synaptic plasticity and learn- mutation of the CREB and isoforms (CREB-),
ing. For example, the ERK signaling pathways are which disrupted the stability of hippocampal LTP,
also involved in plasticity and learning (Sweatt, also impaired memory (but not learning) tested in a
2001)). Inhibitors of MEK, a kinase that activates wide range of tasks, suggesting that CREB-dependent
ERK, disrupt LTP and learning; both the induction transcription was required for long-term memory. In-
of LTP and training activate ERK. Mutations of genes jection of a herpes simplex virus carrying a CREB
thought to modulate ERK signaling, including Ras, gene into the amygdala, a structure with a well-
the neurofibromin GTPase Activating Protein, the characterized role in fear conditioning, enhanced this
guanine-nucleotide-releasing-factor (GRF), and the form of learning. These and many other studies have
B-Raf kinase disrupt hippocampal LTP and learning shown that CREB has a universal role in memory
and memory. (Silva, Kogan, Frankland, and Kida, 1998).
174 GENETIC SUBSTRATES OF MEMORY: Hippocampus
There is extensive evidence that cAMP- overexpression of TPA, the transgenic overexpres-
dependent signaling can activate CREB (Silva, sion of the NMDAR subunit 2B, which appears to
Kogan, Frankland, and Kida, 1998). Indeed, deleting lengthen the opening time of the NMDAR, also en-
both calcium/calmodulin induced adenylate cyclases hances LTP and learning (Tang et al., 1999).
(1 and 8) blocks the maintenance but not the induc-
The overexpression of the presynaptic Growth
tion of CA1 LTP and disrupts hippocampal long-
Associated Protein 43, the mutation of a telencepha-
term memory. A drug that activates adenyl cyclases
lon-specific cell adhesion molecule, the transgenic in-
turns on CREB and rescues the LTP and memory def-
hibition of Calcineurin, and the mutation of Rin 1 (a
icits of this double mutant (Wong et al., 1999). Anoth-
ras effector) all facilitate LTP and learning. Neverthe-
er protein that can activate CREB is calcium/
less, not all manipulations that enhance LTP improve
calmodulin kinase IV (CaMKIV). Experiments with a
learning or memory. For example, LTP enhance-
dominant-negative form of CaMKIV (dnCaMKIV)
ments that disrupt basic associativity mechanisms do
expressed in the postnatal forebrain showed that this
not quicken learning.
transgene affected late but not early stages of CA1
LTP. Similarly, behavioral experiments revealed spe-
cific long-term memory deficits in both spatial learn- Disconnections Between LTP and Learning
ing and contextual fear conditioning. CREB activa-
There are cases where hippocampal LTP seems
tion was impaired by the dnCaMKIV mutation,
to be disrupted with no apparent effect on learning
indicating that this kinase may activate CREB during
and memory. For example, the deletion of the Gluta-
learning (Kang et al., 2001).
mate Receptor 1 (GluR1) leads to deficits in CA1 LTP
CREB seems to regulate the expression of Zif268, but not in spatial learning. Similarly, the deletion of
a transcription factor whose expression is triggered the Thy-1 gene also spared spatial learning but dis-
by LTP and learning. Strikingly, studies of a Zif268 rupted LTP in the dentate gyrus. The interconnec-
null-mutant mouse showed that this transcription fac- tions between biological phenomena such as LTP and
tor is needed for the stability of hippocampal plastici- learning are almost always more complex than ex-
ty and memory (Jones et al., 2001). These results sug- pected. For example, in vivo recordings revealed LTP
gest that Zif268 is downstream of CREB and that this in the Thy-1 mutants, and tests with more physiologi-
transcriptional cascade is critical for memory. cal LTP-inducing protocols demonstrated that the
A key facet of the experiments summarized above GluR1 mutants can express robust levels of LTP. Nev-
is that short-term memory is normal in the mutant ertheless, it is possible that glutamate receptors are
mice, demonstrating that the processes required for not essential for the expression of all synaptic-specific
learning (sensory processing, motivation, and so on) changes underlying learning and memory. Perhaps
are unaffected in these mice, thus simplifying the in- other classes of channels (i.e., potassium channels)
terpretation of the results. can also mediate these changes and therefore support
learning in mutants with deficits in GluR LTP.
There have also been studies of the impact of
Enhancing LTP and Learning some of the mutations described above in hippocam-
Remarkably, molecular manipulations that en- pal circuit function. For example, hippocampal cir-
hance LTP often also enhance learning and memory. cuits represent spatial information. Recordings in be-
For example, the mutation of the nociceptin recep- having rats and mice have demonstrated that
tor, which mediates the inhibition of adenylyl cyclase, hippocampal pyramidal cells fire in a place-specific
facilitates hippocampal LTP and spatial learning and manner (place fields). Remarkably, a number of ma-
memory (Manabe et al., 1998). Presumably, during nipulations that disrupted hippocampal LTP did not
learning the nociceptin receptor mutants generate block these spatial representations but impaired their
higher levels of cAMP than controls. This cAMP instability. Thus, stable synaptic changes may be crucial
crease enhances LTP, resulting in faster learning. for the stability of circuit representations of informa-
tion in the brain. This possibility also implies that
A transgenic Tissue Plasminogen Activator (TPA)
mechanisms other than synaptic plasticity are respon-
also enhances LTP and learning, whereas a null mu-
sible for generating these representations. Undoubt-
tation of this gene impairs them. TPA is an extracellu-
edly, transgenetic approaches will also have a role in
lar protease that seems to be involved in synaptic re-
revealing the nature and function of these mecha-
modeling triggered by plasticity and learning.
nisms.
Therefore, loss of TPA impairs this remodeling and
disrupts hippocampal LTP and learning, whereas The molecular and cellular cognitive studies sum-
overexpressing TPA facilitates both synaptic and be- marized above provide compelling evidence that the
havioral plasticity (Madani et al., 1999). Just like the mechanisms responsible for the induction and stabili-
GLUTAMATE RECEPTORS AND THEIR CHARACTERIZATION 175
ty of synaptic changes have a critical role in the acqui- Manabe, T., Noda, Y., Mamiya, T., Katagiri, H., Houtani, T.,
sition and storage of information in the hippocam- Nishi, M., Noda, T., Takahashi, T., Sugimoto, T., Nabeshima,
T., and Takeshima, H. (1998). Facilitation of long-term po-
pus. Because many of the same molecular tentiation and memory in mice lacking nociceptin receptors.
mechanisms are present throughout the brain, they Nature 394, 577581.
might have a universal role in learning and memory Mansuy, I. M., Mayford, M., Jacob, B., Kandel, E. R., and Bach, M.
in other structures. Nevertheless, there is also evi- E. (1998a). Restricted and regulated overexpression reveals
dence for important differences between the molecu- calcineurin as a key component in the transition from short-
term to long-term memory. Cell 92, 3949.
lar mechanisms underlying learning in different Mansuy, I. M., Winder, D. G., Moallem, T. M., Osman, M., May-
brain structures (e.g., the amygdala and hippocam- ford, M., Hawkins, R. D., and Kandel, E. R. (1998b). Inducible
pus). The findings and ideas immerging from these and reversible gene expression with the rtTA system for the
studies could be the foundation stone for understand- study of memory. Neuron 21, 257265.
ing the molecular and cellular processes that underlie Mayford, M., Bach, M. E., Huang, Y. Y., Wang, L., Hawkins, R. D.,
and Kandel, E. R. (1996). Control of memory formation
our thoughts, fears, desires, and dreams. through regulated expression of a CaMKII transgene. Science
274, 1,6781,683.
See also: GLUTAMATE RECEPTORS AND THEIR Mayford, M., Wang, J., Kandel, E. R., and ODell, T. J. (1995).
CHARACTERIZATION; LONG-TERM CaMKII regulates the frequency-response function of hippo-
POTENTIATION; PROTEIN SYNTHESIS IN LONG- campal synapses for the production of both LTD and LTP.
TERM MEMORY IN VERTEBRATES; SECOND Cell 81, 891904.
MESSENGER SYSTEMS Morris, R. G. M., Anderson, E., Lynch, G. S., and Baudry, M.
(1986). Selective impairment of learning and blockade of
long-term potentiation by an N-methyl-D-asparate receptor
Bibliography
antagonist, AP5. Nature 319, 774776.
Abel, T., Nguyen, P. V., Barad, M., Deuel, T. A., Kandel, E. R., and Ohno, M., Frankland, P. W., Chen, A. P., Costa, R. M., and Silva,
Bourtchouladze, R. (1997). Genetic demonstration of a role A. J. (2001). Inducible, pharmacogenetic approaches to the
for PKA in the late phase of LTP and in hippocampus-based study of learning and memory. Nature Neuroscience 4, 1,238
long-term memory. Cell 88, 615626. 1,243.
Bach, M. E., Hawkins, R. D., Osman, M., Kandel, E. R., and May- Silva, A. J., Kogan, J. H., Frankland, P. W., and Kida, S. (1998).
ford, M. (1995). Impairment of spatial but not contextual CREB and memory. Annual Review of Neuroscience 21, 127
memory in CaMKII mutant mice with a selective loss of hip- 148.
pocampal LTP in the range of the theta frequency. Cell 81, Silva, A. J., Paylor, R., Wehner, J. M., and Tonegawa, S. (1992a).
905915. Impaired spatial learning in alpha-calcium-calmodulin kinase
Davis, H. P., and Squire, L. R. (1984). Protein synthesis and memo- II mutant mice. Science 25, 206211.
ry. Psychology Bulletin 96, 518559. Silva, A. J., Stevens, C. F., Tonegawa, S., and Wang, Y. (1992b). De-
Eichenbaum, H. (2001). The hippocampus and declarative memo- ficient hippocampal long-term potentiation in alpha-calcium-
ry: Cognitive mechanisms and neural codes. Behavioral Brain calmodulin kinase II mutant mice. Science 257, 201206.
Research 127, 199207. Sweatt, J. D. (2001). The neuronal MAP kinase cascade: A bio-
Giese, K. P., Fedorov, N. B., Filipkowski, R. K., and Silva, A. J. chemical signal integration system subserving synaptic plas-
(1998). Autophosphorylation at Thr286 of the alpha calcium- ticity and memory. Journal of Neurochemistry 76, 110.
calmodulin kinase II in LTP and learning. Science 279, 870 Tang, Y. P., Shimizu, E., Dube, G. R., Rampon, C., Kerchner, G.
873. A., Zhuo, M., Liu, G., and Tsien, J. Z. (1999). Genetic en-
Grant, S. G., ODell, T. J., Karl, K. A., Stein, P. L., Soriano, P., and hancement of learning and memory in mice. Nature 401, 63
Kandel, E. R. (1992). Impaired long-term potentiation, spa- 69.
tial learning, and hippocampal development in fyn mutant Tsien, J. Z., Chen, D. F., Gerber, D., Tom, C., Mercer, E. H., An-
mice. Science 258, 1,9031,910. derson, D. J., Mayford, M., Kandel, E. R., and Tonegawa, S.
Jones, M. W., Errington, M. L., French, P. J., Fine, A., Bliss, T. V., (1996). Subregion- and cell type-restricted gene knockout in
Garel, S., Charnay, P., Bozon, B., Laroche, S., and Davis, S. mouse brain. Cell 87, 1,3171,326.
(2001). A requirement for the immediate early gene Zif268 in Wong, S. T., Athos, J., Figueroa, X. A., Pineda, V. V., Schaefer, M.
the expression of late LTP and long-term memories. Nature L., Chavkin, C. C., Muglia, L. J., and Storm, D. R. (1999). Cal-
Neuroscience 4, 289296. cium-stimulated adenylyl cyclase activity is critical for hippo-
Kang, H., Sun, L. D., Atkins, C. M., Soderling, T. R., Wilson, M. campus-dependent long-term memory and late phase LTP.
A., and Tonegawa, S. (2001). An important role of neural ac- Neuron 23, 787798.
tivity-dependent CaMKIV signaling in the consolidation of
long-term memory. Cell 106, 771783. Alcino J. Silva
Lisman, J. E., and McIntyre, C. C. (2001). Synaptic plasticity: A mo-
lecular memory switch. Current Biology 11, R788791.
Madani, R., Hulo, S., Toni, N., Madani, H., Steimer, T., Muller,
D., and Vassalli, J. (1999). Enhanced hippocampal long-term
potentiation and learning by increased neuronal expression GLUTAMATE RECEPTORS AND
of tissue-type plasminogen activator in transgenic mice.
EMBO Journal 18, 3,0073,012. THEIR CHARACTERIZATION
Malleret, G., Haditsch, U., Genoux, D., Jones, M. W., Bliss, T. V., The amino acids L-glutamate and L-aspartate are
Vanhoose, A. M., Weitlauf, C., Kandel, E. R., Winder, D. G.,
and Mansuy, I. M. (2001). Inducible and reversible enhance- the major excitatory neurotransmitters in the mam-
ment of learning, memory, and long-term potentiation by ge- malian central nervous system (CNS). Some subtleties
netic inhibition of calcineurin. Cell 104, 675686. aside, glutamate is considered the predominant excit-
176 GLUTAMATE RECEPTORS AND THEIR CHARACTERIZATION
atory neurotransmitter and for simplicity the recep- atory postsynaptic potential (EPSP) at glutamatergic
tors that bind excitatory amino acids will be referred synapses. Binding of glutamate to these receptors
to as glutamate receptors, or GluRs. Glutamate medi- opens a channel that is permeable to Na+ and K+
ates its effects by interacting with receptors that can (and in some instances Ca2+), inducing the depolariz-
be distinguished by pharmacological, physiological, ing currents characteristic of these receptors. GluRs
anatomical, molecular, and genetic criteria. The in- can desensitize rapidly, limiting the time the channel
teraction of glutamate with its receptors underlies spends in the open state. This desensitization helps
many normal physiological processes, from rapid syn- in truncating the duration of the EPSP at excitatory
aptic signaling and information transfer to longer- synapses.
lasting changes in synaptic efficacy that are thought
The nonendogenous agonists AMPA and
to be the cellular basis of learning and memory. In ad-
quisqualate (QA), and other structurally related com-
dition, neurotransmission mediated by glutamate and
pounds, selectively activate the AMPA receptor sub-
its receptors is implicated in a variety of CNS patholo-
class. A series of quinoxaline derivatives (6-cyano-7-
gies, including epilepsy, cell death due to excitotoxi-
nitroquinoxaline-2,3-dione [CNQX]; 6,7-dinitroqui-
city and ischemia, and Alzheimers disease. This entry
noxaline-2, 3-dione [DNQX]; and 6-nitro-7-
will review the general characteristics of GluRs, in-
sulfamoyl-benzo (f)quinoxaline-2,3-dione [NBQX])
cluding pharmacological specificity, ion selectivity,
are the most selective known AMPA/QA antagonists,
and modulation by other compounds.
but they can also act as KA antagonists.
KA has been shown to be both a potent excitant
Glutamate Receptor Classes and a potent excitotoxin. The case for a KA receptor
Glutamate receptors can be divided into two that is distinct from AMPA/QA receptors initially
major classes. One class is termed ionotropic, because stemmed from pharmacological studies in which KA
glutamate binding causes a conformational change in responses were inhibited by select antagonists more
the receptor that directly opens a gate, permitting ion potently than were AMPA/QA or NMDA responses,
passage in and out of the cell. The channel and the and from anatomical studies in which KA and AMPA/
glutamate-binding site are part of a single multisub- QA receptors displayed distinct anatomical localiza-
unit macromolecule found in the plasma membrane. tions. However, even considering the quinoxaline an-
Three pharmacologically distinct glutamate receptor tagonists, accurate distinction of KA from AMPA/QA
subclasses of this ionotropic type are distinguished, receptors remains difficult because of their similar
named after their selective agonists: the N-methyl-D- pharmacological profiles. Localization of AMPA/QA
aspartate (NMDA), 2-amino-3-hydroxy-5-methyl-4- and KA receptors by autoradiographic techniques
isoxazolepropionate (AMPA), and kainate (KA) re- demonstrates a differential distribution of these two
ceptors. The AMPA and KA subclasses are frequently types. For example, [3H]AMPA-binding sites are con-
referred to as non-NMDA receptors. centrated in the hippocampal CA1 region, outer cor-
tical layers, lateral septum, and the molecular layer of
A second class of glutamate receptors is charac-
the cerebellum, whereas [3H]KA-binding sites are
terized by their capacity to couple to second-
concentrated in the hippocampal mossy fibers, deep
messenger generating systems via guanine nucleo-
cortical layers, and the granule cell layer of the cere-
tide-binding protein activation (G-proteins) and are
bellum.
termed the metabotropic glutamate receptors (mGluRs).
The binding of glutamate to these mGluRs induces Advances in molecular genetic techniques have
secondary changes that are mediated by intracellular led to the identification of a series of complementary
molecules distinct from the receptor itself. One such DNA (cDNA) clones that encode a family of GluRs.
mGluR is coupled to the phosphoinositide (PI) sec- Thus, injection of messenger RNA derived from these
ond messenger system and is named after its selective cDNA sequences into Xenopus oocytes results in elec-
agonist, trans-1-amino-1,3-cyclopentanedicarboxylic trophysiological responses to glutamate, KA, QA, and
acid (trans-ACPD). A second is named after its selec- AMPA, but not to NMDA or AP-4. The non-NMDA
tive agonist, l-2-amino-4-phosphonobutanoic acid (L- receptor family is composed of at least seven separate
AP4) and exists in the retina, whereas a distinct L- genes (GluR1GluR7) in the rat, and the proteins ex-
AP4-sensitive receptor type has been described in se- pressed from these genes each have an approximate
lect CNS pathways. molecular weight of 100 kilodaltons. GluR1GluR4
are the predominant forms expressed in the brain,
and the mature form of the receptor appears to be a
Non-NMDA Receptors tetrameric complex of approximately 600 kilodaltons
Non-NMDA receptors mediate the rapid on-off composed of different combinations of the GluR sub-
type of synaptic signaling that underlies the fast excit- units. Two genes (KA-1 and KA-2) have also been
GLUTAMATE RECEPTORS AND THEIR CHARACTERIZATION 177
identified that encode proteins that represent the other agonists. Glutamate-binding can be selectively
high-affinity KA binding site in the brain. KA-1 and and competitively antagonized by a series of com-
KA-2 do not form functional receptors themselves. pounds, including D-2-amino-5-phosphonopen-
They must combine with other GluR subunits to pro- tanoate (D-AP5) and 3-(2-carboxypiperazin-4-
duce an ionotropic glutamate receptor. Such structur- yl)propyl-1-phosphate (CPP). Second is the glycine-
al complexity helps to explain the difficulty in distin- binding site, which must be occupied in order for the
guishing AMPA/QA from KA responses using channel to be opened by glutamate binding to its rec-
pharmacological approaches. In addition, the capaci- ognition site. The pharmacology of the NMDA gly-
ty to mix different subunits (and their splice variants) cine site is distinct from that of the inhibitory glycine
provides for a large number of possible GluR pheno- receptor in the spinal cord and brain stem, in that the
types when analyzed using either pharmacological or NMDA receptor glycine site is insensitive to strych-
electrophysiological techniques. nine but is antagonized by kynurenate, 7-
chlorokynurenate, and other compounds. Although
The GluR family is evolutionarily divergent from
this site may always be saturated in vivo, glycine is
other ionotropic receptors (e.g., nicotinic AchR). Io-
often referred to as a cotransmitter. Third is a site
notropic receptors all have four membrane-spanning
within the ion channel that binds Mg++. Binding of
domains; however, the GluR family is unique in that
Mg++ at this site blocks current flow through the
the second membrane-spanning segment (TM2)
channel when the membrane is at hyperpolarized po-
forms a kink in the membrane, exiting back into the
tentials. This block is removed under depolarizing
cytoplasm instead of traversing the bilayer and exit-
conditions, providing the molecular basis for the
ing into the extracellular space. This kink in TM2 is
NMDA receptors voltage dependence. Fourth is a
analogous to a pore (P) forming element in K+ chan-
site within the ion channel that binds phencyclidine
nels and forms the structure of the GluR responsible
(PCP), dibenzocyclohepteneimine (MK-801), and
for determining ion permeation. GluRs are permeant
to Na+ and K+ although certain GluR variants can other compounds, causing blockage of ion flow
also be significantly permeable to Ca++. through the channel. Binding of these compounds is
permitted only when the ion channel is in the open
state and the compounds are, therefore, referred to
NMDA Receptors as open channel blockers. Finally, a variety of other
molecules have been shown to be modulators of the
The NMDA receptor is pharmacologically and
NMDA receptor. For example, spermine and spermi-
functionally more complex than AMPA/QA and KA
dine increase MK-801 binding to the NMDA complex
receptors. For example, binding of glutamate to the
and NMDA-evoked currents in cell culture and
NMDA receptor will open the ion channel only if the
Xenopus oocyte preparations. Additionally, Zn++ and
membrane is depolarized and if glycine is present.
H+ have been shown to modulate NMDA receptor
Because NMDA receptors open as a function of the
currents in a number of preparations.
extent of membrane depolarization, they are de-
scribed as voltage-dependent and are thus sensitive to Localization of NMDA receptors by autoradio-
postsynaptic activity. Therefore, the NMDA receptor graphic techniques demonstrates an anatomical dis-
complex provides an example of a conditional logic tribution that in general parallels the distribution of
gate where Hebb-like conditions are realized at a sin- AMPA/QA receptors, further supporting the idea that
gle synapse. Furthermore, the ion channel is signifi- these two receptors work in concert in many synapses.
cantly permeable to Ca++ in addition to Na+ and K+, Thus, NMDA receptors are found in the CA1 region
resulting in a significant influx of Ca++ that induces of the hippocampus, throughout the cortex (particu-
a variety of secondary Ca++-activated processes. larly in outer cortical layers), and in striatum. Howev-
These processes include 1. the induction of long-term er, specific localization of binding shows a different
potentiation (LTP), which is regarded as a cellular distribution of NMDA agonist- and antagonist-
model of memory analogous to Hebb-type synaptic preferring binding sites, suggesting that subpopula-
plasticity; 2. learning and memory in animal models; tions of NMDA receptors exist in different brain
3. the pathophysiology of epilepsy; and 4. some forms areas.
of excitotoxicity.
The NMDA receptor has been successfully isolat-
The NMDA receptor appears to be regulated by ed and characterized, and several distinct subunits of
a variety of endogenous and exogenous compounds this receptor complex have been cloned. NMDAR1
that act at distinct binding sites to modify the function was the original isolate, and various NMDAR2A2D
of the receptor. The first of these binding sites is the subunits coassemble with NMDAR1 to form function-
transmitter recognition site, which binds L-glutamate al receptors. Anatomical gene-mapping studies have
(or L-aspartate), the synthetic ligand NMDA, and confirmed that subpopulations of NMDA receptors
178 GLUTAMATE RECEPTORS AND THEIR CHARACTERIZATION
exist in different brain regions. As with the non- The mGluRs are widespread in the nervous sys-
NMDA receptors, numerous pharmacological and tem and are found both pre- and postsynaptically.
electrophysiological phenotypes are possible through Presynaptically, they serve as autoreceptors and ap-
the mixing of different subunits. The TM2 segment pear to participate in the inhibition of neurotransmit-
of the NMDARs is responsible for determining the ter release. Their postsynaptic roles appear to be var-
ion permeation properties of the channel and forms ied and depend on the specific G protein to which
a kink in the membrane analogous to that described they are coupled. Activation of mGluR1 has been im-
for the GluRs above. plicated in long-term synaptic plasticity at many sites
in the brain, including long-term potentiation in the
G-Protein-Coupled Glutamate Receptors hippocampus.
The G-protein coupled receptors (GPCRs) that

bind glutamate and other excitatory amino acids are Conclusion
referred to as the metabotropic glutamate receptors, or More than one receptor subclass may be present
mGluRs. They are similar in general structure to in any given excitatory synapse. The function of
other GPCRs in having seven transmembrane span- glutamatergic synapses depends on the combination
ning domains; however, they are divergent enough to of the above-mentioned receptor subtypes at a given
be considered to have originated from a separate evo- synapse (e.g., non-NMDA, NMDA, both non-NMDA
lutionary-derived receptor family. In fact, sequence and NMDA, and mGluR, among others). Working in
homology between the mGluR family and other concert these receptors yield a complex synaptic re-
GPCRs is minimal except for the GABAB receptor. sponse that depends on the number and location of
The mGluR family is heterogeneous in size, ranging individual receptors. Because these subtypes clearly
from 854 to 1,179 amino acids. The ligand-binding serve distinct functions, the capacity to manipulate in-
site for mGluRs resides in the large N-terminal extra- dividual receptor molecules is critical to understand-
cellular domain. Additionally, the mGluRs exist as ing the role of a subtype in a normal physiological
functional dimers in the membrane in contrast to the event or in the etiology of a given glutamate-linked
single subunit form of most GPCRs. These significant disease. The advent of more potent and selective
structural distinctions support the idea that the drugs for each receptor subclass will allow for more
mGluRs evolved separately from the other GPCRs. precise experimental and clinical manipulation of
The third intracellular loop, thought to be the major these receptors. Similarly, molecular genetic ap-
determinant responsible for G-protein coupling of proaches to analyzing both the structure and the
the mGluRs, is relatively small, whereas the C- function of these receptors has revealed a wealth of
terminal domain is quite large. The coupling between information, and further analysis of their genes and
mGluRs and their respective G proteins may occur gene products will have far-reaching implications for
through unique determinants that exist in the large the basic science of glutamate receptors and for a va-
C-terminal domain. riety of glutamate-linked human diseases.
Eight different mGluRs can be subdivided into
subgroups on the basis of sequence homologies and See also: GUIDE TO THE ANATOMY OF THE BRAIN:
their capacity to couple to specific enzyme systems. SYNAPSE; HEBB, DONALD; LONG-TERM
For example, both mGluR1 and mGluR5 activate a G POTENTIATION; NEUROTRANSMITTER SYSTEMS
protein coupled to phospholipase C. mGluR1 activa- AND MEMORY; SECOND MESSENGER SYSTEMS
tion can also lead to the production of cAMP and of
arachidonic acid by coupling to G proteins that acti- Bibliography
vate adenylate cyclase and phospholipase A2, respec- Boulter, J., Hollmann, M., OShea-Greenfield, A., Hartley, M., De-
neris, E., Maron, C., and Heinemann, S. (1990). Molecular
tively. The subtype mGluR5 seems more specific,
cloning and functional expression of glutamate receptor sub-
activating predominantly the G-protein-activated unit genes. Science 249, 1,0331,037.
phospholipase C. Cotman, C. W., Monaghan, D. T., Ottersen, O. P., and Storm-
Mathisen, J. (1987). Anatomical organization of excitatory
Distinctions between the mGluR subtypes can
amino acid receptors and their pathways. Trends in Neuro-
also be made pharmacologically. One group, sciences 10, 273280.
mGluR2, 3, and 8, favors the agonist trans-ACPD for Hollmann, M., and Heinemann, S. (1994). Cloned glutamate re-
activation, whereas a second group, including ceptors. Annual Review of Neuroscience 17, 31108.
mGluR4, 6, and 7, favors AP-4 for activation. The sub- Keinanen, K., Wisden, W., Sommer, B., Werner, P., Herb, A., Verd-
oorn, T. A., Sakmann, B., and Seeburg, P. H. (1990). A family
types mGluR2 and mGluR4 can be further distin-
of AMPA-selective glutamate receptors. Science 249, 556560.
guished pharmacologically by using the agonist 2- Masu, M., Tanabe, Y., Tsuchida, K., Shigemoto, R., and Nakanishi,
(carboxycyclopropyl)glycine, which is more potent at S. (1991). Sequence and expression of a metabotropic gluta-
activating mGluR2 receptors. mate receptor. Nature 349, 760765.
GUIDE TO THE ANATOMY OF THE BRAIN: Overview 179
Nakanishi, S. (1994). Metabotropic glutamate receptors: Synaptic organized, areas and pathways that subserve specific percep-
transmission, modulation, and plasticity. Neuron 13, 1,031 tual, motor, emotional, and cognitive information process-
1,037.
Nakanishi, N., Shneider, N. A., and Axel, R. (1990). A family of
ing functions. At the same time, these cortical areas are also
glutamate receptor genes: Evidence for the formation of the storehouses of memories for the specific information pro-
heteromultimeric receptors with distinct channel properties. cessed in those dedicated pathways. Therefore, understand-
Neuron 5, 569581. ing the functional organization of the cerebral cortex holds
Rosenmund, C., Stern-Bach, Y., and Stevens, C. F. (1998). The the key to characterizing the nature of the brains representa-
tetrameric structure of a glutamate receptor channel. Science
280, 1,5961,599.
tion of our accumulation of knowledge. The OLFACTORY
Sommer, B., Kohler, M., Sprengel, R., and Seeburg, P. H. (1991). CORTEX is a part of the cerebral cortex that is simpler in its
RNA editing in brain controls a determinant of ion flow in architecture than other cortical areas, and has served as a
glutamate-gated channels. Cell 67, 1119. model system for anatomical and physiological studies and
Watkins, J. C., Krogsgaard-Larsen, P., and Honore, T. (1990). for the development of computational models of combined
Structure activity relationships in the development of excit-
atory amino acid receptor agonists and competitive antago-
perceptual and memory processing.
nists. Trends in Pharmacological Sciences 11, 2533. The remaining sections consider a set of brain areas that
Waxham, M. N. (1998). Ionotropic and metabotropic receptors. In are nodal points within brain pathways that support specific
F. E. Bloom, S. L. Landis, J. L. Roberts, L. R. Squire, and M.
J. Zigmond, eds., Fundamental neuroscience, pp. 235267. New
kinds of memory, or modulate memory. The PERIRHINAL
York: Academic Press. CORTEX and HIPPOCAMPUS are major components of the
brain system that mediates declarative memory, our capac-
C. W. Cotman
ity for conscious recollection of facts and events. The
E. R. Whittemore
AMYGDALA is a major component of a system for emotional
Revised by M. N. Waxham
memory, the system that mediates assignments of affective
value to otherwise neutral stimuli and initiates automatic re-
sponses to stimuli of acquired emotional significance. The
BASAL GANGLIA and CEREBELLUM are major components
GUIDE TO THE ANATOMY
of distinct subsystems that mediate procedural memory, the
OF THE BRAIN acquisition of skills, habits, and conditioned motor responses.
The BASAL FOREBRAIN plays a major role within the modu-
[In the brain, as in other complex systems, function follows
latory pathways that regulate attentional mechanisms that
form. Therefore, an understanding of learning and memory
influence memory processing. The amygdala also plays a
is critically dependent on a complete description of the ana-
role in the modulation of memory processing in emotional
tomical organization of brain systems that support memory.
circumstances. Thus each area plays a distinctive role in
Brain organization is composed at many levels, from cellu-
memory processing. In real life, they act in parallel to medi-
lar morphology to synapse types and their patterns of connec-
ate the impact of experience on the brains information-
tion, to the cytoarchitecture of cortical and subcortical areas
processing systems. This guide provides the reader with the
and the organization of neuronal types and input/output
anatomical framework in which these brain areas make their
patterns, to major pathways and hierarchies of information
individual contributions to memory.]
processing. These levels of brain organization are surveyed
in the OVERVIEW section that follows.
The fundamental alteration in plasticity that underlies
OVERVIEW
memory occurs at the level of cellular anatomy and, in par-
ticular, synaptic structure. Therefore two sections of this The adult human brain weights about 1,400 grams,
guide provide a summary of our current knowledge about the heaviest of any species relative to total body
the structure of NEURONS and SYNAPSES. The section on weight. This perhaps most impressive human organ
neurons considers the major components of neurons and fo- appears to be responsible for our intellectual superi-
cuses on the cellular elements where plasticity occurs, specifi- ority over other speciesalthough, paradoxically,
cally in the dendrites and synapses. This section also consid- human intelligence is not strongly correlated to brain
ers different neuron types and nature of connectivity patterns size.
that form the circuitry in which plasticity exerts its effects on The modern discipline of neuroanatomy has de-
function. The section on synapses reviews the components of veloped over the past 400 years through the descrip-
synapses and functionally distinct types of synapses. This sec- tive studies of workers such as Vesalius, Willis, Retzi-
tion focuses on dendritic spines as the likely sites of plasticity us, Purkinje, Brodmann, Campbell, Ramn y Cajal,
that supports memory. and Golgi. Rather than answering a pedestrian ques-
In other sections, this guide focuses on brain areas that tionWhat does it look like?the main objective of
play prominent roles in learning and memory. One area is studying anatomy of the brain is to create physiologi-
the vast expanse of the CEREBRAL CORTEX. The cerebral cally and psychologically meaningful structural infer-
cortex includes several functionally specific, hierarchically ences. The capacity and the choice of visualization
180 GUIDE TO THE ANATOMY OF THE BRAIN: Overview
methods define the level of differentiation of the nal capsule, which interconnects the cortex with sub-
brain structure from gross features to a single neuron. cortical structures and spinal cord, and the corpus
In actuality, the anatomy of the brain proceeds on collosum, which connects two cortical hemispheres.
four levels, depending on the aim of the description. The midbrain in mammals is a much smaller
At the most superficial level, based on the evolution- structure than the forebrain. Topographically the
ary and embryonic development, the brain can be di- midbrain is a caudal continuation of the thalamus and
vided into the forebrain, the midbrain, and the hind- the hypothalamus. Two major gross features of the
brain. In development, neurons, which consitute midbrain are the upper tectum (the roof) and the
these subdivisions, multiply, develop distinct appear- lower tegmentum (the covering). (Latin terminology
ance and aggregate into groups forming nuclei, lay- is still prominent in neuronatomy alongside the
ers, or areas of the brain. This cellular level of brain Greek.) The tectum also features two swellings on
anatomy, called cytoarchitectonic, differentiates cell each side of the midline: the superior and the inferior
types and their location. Further differentiating brain colliculi, which harbor major visual and auditory
structures are the biochemical properties of neurons, pathways, respectively. Another prominent feature of
which permit classification of different cell groups the midbrain is the cerebral aqueduct, a canal con-
into chemically coded systems. At the most sophisti- necting ventricular systems of the forebrain and the
cated level of resolution development are cytoarchi- hindbrain.
tecture and chemoarchitecture, along with informa- The most caudal part of the brain, the hindbrain,
tion about cellular connections and aspects of is composed of three major parts: the pons (bridges),
neuronal functioning. These eclectic criteria of struc- the cerebellum (small brain) and the medulla oblon-
tural classification provide the most meaningful struc- gata (long brain). In neuroanatomy the medulla,
tural references within the general scope of brain and pons, and midbrain are often collectively refered to
bodily function. as the brain stem. The pons are positioned immedi-
ately caudally to the midbrain and appear as a large
protuberance (a bridge) on the ventral surface of the
Gross Anatomy
brain. Dorsally the pons are covered by the tegmen-
In early embryonic development the brain has tum pontis, which is a caudal continuation of the mes-
rather simple gross features and is composed of just encephalic tegmentum. The cerebellum is positioned
three vesicles, the hindbrain, the midbrain, and the on the posterior surface of the brain and connected
forebrain. These soon subdivide to give place to more to it by three pars of cerebellar peduncules (pillars).
complicated arrangement of four vesicles: rhomben- In gross terms the cerebellum features two lateral
cephalon (parallelogram-brain), mesencephalon hemispheres abutting the middle structure called the
(middle-brain), diencephalon (between-brain) and vermis (worm). Finally, the medulla is the most poste-
telencephalon (end-brain) (many neuroanatomists rior part of the brain, immediately succeeded by the
still use Greek to name newly discovered structures). spinal cord. Superficially, it resembles the spinal cord
Progressive development gives the brain a sophisti- with the major exception of the presence of bilateral
cated form, which is traditionally segmented into nu- inferior olives on its ventro-lateral surface and the
merous smaller structures on the basis of gross ana- floor of the fourth ventricular cavity on its dorsal sur-
tomical features. For example, the frontally face.
positioned forebrain gives rise to two voluminous
hemispheres of the cerebral cortex (brain bark)
separated from each other by the longitudinal fissure. Cytoarchitecture
The surface of the cortex is convoluted into folds The brain of humans and the head ganglion of
known as gyri, which are separated by groves known mosquitoes are both made up of two general types of
as sulci. The cortex enfolds numerous subcortical cells: neurons and glial cells. Although glial cells are
neuronal structures, such as thalamus (the main relay the most numerous cells in the brainthey outnum-
for neural input to the cortex), hypothalamus (the ber neurons by about twenty-five to oneit is neuro-
neuro-hormonal regulatory center that keeps the rest nal activity that is associated with brain function. Neu-
of the body alive), hippocampus, and amygdala rons show greater diversity in shape and size than any
(structures implicated in emotional behavior, learn- other type of cell in the body. In contrast to other cells
ing, and memory), caudate, putamen, and globus pal- in the body, neurons can generate and relay electrical
lidus (major parts of the so-called basal ganglia, which impulsesaction potentials. In order to transmit
seem to play a role in learning and control of move- action potentials over great distances, neurons have
ment). Other obvious gross features of the forebrain extended processesaxons that, for some spinal
include cerebral ventricles (a network of fluid-filled motor neurons of large mammals, may have lengths
cavities) and prominent fiber tracts such as the exter- on the order of meters.
GUIDE TO THE ANATOMY OF THE BRAIN: Overview 181
Since neurons usually do not communicate by di- nized in layers, while the tegmentum contains nuclei
rect contacteach neuron operates as an indepen- of the oculomotor and trochlear cranial nerves re-
dent unitthey have evolved highly specialized sponsible for eye movement and pupil constriction.
points of interaction: synapses. Synapses are essen- The periaqueductal gray, another distinct cellular
tially clefts between neurons. They are asymmetrically structure of the midbrain, is composed of neurons
flanked by structures that release specialized chemi- surrounding the cerebral aqueduct and appears to
cals neurotransmitters from one presynaptic neuron play a role in the regulation of blood pressure.
in response to the presence of action potentials. The The cytoarchitecture of the pons also features
response of the recipientthe postsynaptic neuron several cranial nerve nuclei, raphe nuclei, and a dis-
is determined by highly specific detectors: receptors. persed area called the reticular formation, which is,
The structure of the synapse allows information to however, not confined to pons but is spread through-
flow in one direction only. Throughout the nervous out the entire brain stem. The cytoarchitecture of the
system different neurons utilize a multiplicity of both cerebellum resembles the citoarchitecture of the fore-
neurotransmitters. brain in also having a cortex with a layered arrange-
Staining cellular ribosomes with a Nissl substance ment of cells and several subcortical nuclei.
stain allows light microscopic visualization of the cel- In contrast to the diverse cytoarchitecture of the
lular morphology of the brain. This process reveals cerebral cortex, that of the cerebellar cortex is very
that neurons within the various parts of the brain look homogeneous. Although topographically the medulla
different and are not evenly distributedthey are, can be viewed as a dorsal extension of the spinal cord,
rather, grouped into structures, which make up the the cytoarchitecture of the medulla differs radically
cytoarchitectonic plan of the brain. The form, size, from that of the spinal cord and embodies a tightly
and position of neurons in the cerebral cortex reveal packed amalgam of dispersed areas, compact nuclei,
the layered organization of this structure. Studies and fiber tracts. The medulla contains nuclei of sever-
have demonstrated differences in cytoarchitectonic al cranial nerves, including the nucleus of the hypo-
organization of neuronal layers in different parts of glossal nerve and a complex of nuclei of the vagus
the cortical mantle. These differences lead to the dif- nerve. In fact, many medullary nuclei boast complex
ferentiation of fifty-two cortical areas in the frequent- compartmental cytoarchitecture; the nucleus of the
ly used cortical map of Brodmann. solitary tract, for example, is composed of nine cyt-
Subcortical structures demonstrate even greater oarchitectonically distinct sunuclei.
cytoarchitectonic diversity. In small areas of the thala- Glial cells are the most numerous cells in the cen-
mus and hypothalamus, neurons are grouped into tral nervous system. They are less complex than neu-
numerous nuclei, areas, and zones with distinct mor- rons, and they show less structural diversity. Unlike
phological characteristics and topographical posi- neurons, they retain the ability to divide, a facility
tions. The hypothalamus, for example, contains over they use in their participation in the reaction of ner-
forty nuclei and areas on either side of the third ven- vous tissue to injury. There are two types of glial cells
tricle with distinct cytoarchitectonic characteristics. in the CNS: oligodendrocytes and astrocytes. The pe-
Cytoarchitectonic differentiation can go even further: ripheral nervous system contains a related cell type,
The paraventricular nucleus of the hypothalamus is the Schwann cell, which is crucial to the formation
composed of ten subnuclei of specific neuronal types and maintenance of the myelin sheaths of peripheral
and topographical positions. These subnuclei also nerves. Within the brain, glial cells are involved with
differ in function. Reflecting back onto brain areas structural and metabolic maintenance of neuronal
identified by their gross anatomical features, most of function and the blood-brain barrier. During the fetal
them show complicated cytoarchitectonic organiza- and postnatal development, glial cells play a role in
tion. Several cytoarchitectonically distinct nuclei have axon guidance and the correct arrangement of neural
been identified in the amygdala, whereas the layered patterns.
organization of neurons in the hippocampus shows
An overview of the cellular anatomy of the brain
only four distinct areas. Cytoarchitecture has revealed
would be incomplete without reference to the epen-
an internal and an external component of the globus
dymal cells lining the cerebral ventricles, the menin-
pallidus and has shown that neurons in the putamen
geal membranes surrounding and physically support-
are organized in patch-matrix compartments.
ing the brain, and the network of blood vessels that
There are several neuronal groups in the mid- form the vascular supply to the brain. Unlike neurons
brain, the best known of which is the substantia nigra and glial cells, these elements are not exclusive to the
(black substance), a complex, compartmentalized nervous system. They share many common structural
structure whose cells die in Parkinsons disease. Neu- and functional features with other support cells found
rons of the superior and inferior colliculi are orga- throughout the body.
182 GUIDE TO THE ANATOMY OF THE BRAIN: Overview
Chemoarchitecture erwise homologous nuclei and areas. Nevertheless, in

For most of the twentieth century, the under- terms of overall value, chemoarchitectonic delinea-
tions have become a preferred method in compara-
standing of human neuroanatomy was gleaned main-
tive neuroscience. Naturally, the neuroactive profile
ly from cytoarchitectonic observations. The most
of neurons offers grounds for determining the orga-
widely used maps of the human cortex were produced
nization of neuronal groups within a species and for
by Brodmann in 1909 on the basis of Nissl substance
comparing them across species. For example, dop-
and myelin staining, while the most detailed
amine, norepinephrine, epinephrine and -
neuroanatomical description of the human hypothal-
aminobutyric acid (GABA) are neuroactive chemical
amus was published by Brockhaus in 1942 and was
compounds that can characterize neuronal sub-
also based on early cytoarchitectonic techniques. The
groups. However, the term chemoarchitecture implies
main shortcoming of early neuroanatomical tech-
the use of chemical compounds for differentiation be-
niques was their distance from the mechanisms un-
tween neuronal populations. These compounds are
derlying human brain function. One of the most ex-
not only neurotransmitters but can also be enzymes,
citing developments in neuroanatomy was the
receptors, peptides, and molecules related to neuro-
identification of chemical coding for individual neu-
nal metabolismcalcium-binding proteins, for ex-
ral pathways and the proliferation of chemoarchitec-
ample.
tonic techniques, which allow almost unlimited scope
in the classification of neuronal groups. Chemoar- Catecholamines are a family of functionally im-
chitecture establishes a bridge between structural and portant neurotransmitters. The application of the ty-
functional characteristics of neuronal populations in rosine hydroxylase enzyme immunohistochemistry
the brain. Chemical neuroanatomy has been used to allowed the identification of fifteen groups of cat-
establish the organizational plan of brain regions in echolaminergic neurons in the mammalian brain.
experimental animals and to infer their human ho- These cell groups were not confined to traditional cyt-
mologies. It has also been useful in identifying chemi- oarchitectonic boundaries and sequentially were
cally specified connections in animals. Finally, it has termed A1 to A16 (there is no A3 cell group), extend-
helped to derive hypotheses on the function of brain ing throughout the mammalian brain from the me-
pathways and nuclei. Chemical neuroanatomy has de- dulla to olfactory bulbs. In the human, as in the rat,
veloped as a branch of the structural brain mapping the majority (A1A2 and A4A10) of catecholamine-
methodology that was previously based largely on cyt- gic neuronal groups were found in the brain stem,
oarchitectonic consideration of cell shape, size, and where, for example, tyrosine hydroxylase im-
density. The insubordination of chemically specified munostaining has been used to delineate the interme-
neurons to classic cytoarchitectonic boundaries re- diate reticular zone. Four prominent tyrosine hydrox-
quired a more meaningful delineation of the brain, ylase positive catecholaminergic cell groups (A11
one that incorporates the information about the dis- A15) are located in the hypothalamus and one (A16)
tribution of neuroactive substances, connectivity, and in the substantia innominata of the ventral forebrain.
function. In this respect, chemical neuroanatomy The later cell group is thought to be homologous to
opened a new dimension in neuroscience and allowed the rats catecholaminergic cell group in the olfactory
greater precision, resolution, and reliability in differ- bulb. Subsequent work has shown that cell groups
entiating cell groups and brain areas. such as the A1 and C1 catecholaminergic neurons are
critical for autonomic control in experimental ani-
Studies using chemical neuroanatomy were first
mals and also that these cell groups are strikingly sim-
carried out in rats to facilitate logistical and technical
ilar in rats and humans. A number of studies used
applications. It was not until the 1980s that the chem-
multiple markers to confirm a high degree of conser-
oarchitectonic techniques of histochemistry and im-
vation in the chemical identity of brain-stem neurons
munohistochemistry became sensitive enough to
among rats, monkeys, and humans.
allow their full application to human brain tissue.
Thus, it became increasingly possible to reveal the Neuropeptides are largely neuron-specific chem-
distribution of some of the important neurotransmit- ical compounds that, depending on the neuropep-
ters, receptors, and enzymes of in the human brain tide, are characteristic of specific neuronal sub-
and then make cross-species comparisons. An advan- groups. For example, vasopressin is characteristic of
tage of chemoarchitecture is that each chemical sub- large cells in the lateral magnocellular subnucleus of
stance offers a different window on the organization the paraventricular hypothalamic nucleus. The corti-
of the central nervous system, with successive stains cotropin-releasing factor (or hormone CRF) is a
revealing more of the areas of interest. There are, of neuroendocrine peptide in the cortex, basal telen-
course, significant species differences, and any given cephalon, brain stem, and hypothalamus. The distri-
substance may have inconsistent distributions in oth- bution of CRF is very specific. In neuroanatomy CRF
GUIDE TO THE ANATOMY OF THE BRAIN: Amygdala 183
distribution has been used in the human brain to dis- reveal neuronal connections, including axonal de-
tinguish the subcompartmental organization of spe- generation, anterograde and retrograde tracing, vari-
cific nuclei in the medulla and hypothalamus. In the ous dyes, and even virus tracing, which enables re-
paraventricular hypothalamic nucleus, for example, searchers to trace not just one affiliation but an entire
CRF neurons are confined to the parvicellular com- functional pathway. Techniques such as the tensor
partment, whereas the neurons that contain oxytocin MRI (magnetic resonance imaging) allow the identifi-
are found primarily in the dorsal compartment. Ap- cation of projection and their direction in the living
plying these two markers to the same brain allowed brain. Combined with cytoarchitecture tracing tech-
researchers to distinguish between these subcompart- niques, such advances have enabled researchers to
ments, which otherwise appear to be amalgamated, make meaningful conclusions about the neural cir-
and also allowed the establishment of subcompart- cuitry underlying specific functions.
mental homologies between the paraventricular nu-
cleus in rats and humans.
Conclusion
As an example of a distinct receptor distribution,
Many of the pioneer neurophysiologists (e.g.,
the NK3 receptor (a component of Neuromedin K
David Ferrier, Charles Sherrington) who provided
peptide circuitry) in the human hypothalamus was
the seminal experimental observations of the func-
found in neurons of the paraventricular nucleus, spe-
tions of the nervous system appreciated the contribu-
cifically in the parvicellular and posterior subnuclei,
tions that neuroanatomy made to the interpretation
thus distinguishing these structural subcompart-
of their findings. Since the same strategy is still a pre-
ments. Another prominent population of NK3-
requisite for the neuroscience of today, a great deal
containing cells in the human hypothalamus was
of current research concentrates on describing the di-
found in the perifornical nucleus, distinguishing it
rection and the biochemical and molecular composi-
from the rest of the lateral hypothalamic area. The
tion of the neurons connecting different neural sys-
neuromedin K circuitry in experimental animals
tems.
seems to play an important role in blood-pressure
regulation; in cross-species comparison there were The complicated array of cell groups and fiber
marked similarities in the distribution of NK3 in the pathways in the brain form an incestuous web rather
human and rat hypothalamus. than a hierarchy; however, the connecting threads
can be teased apart thanks to advances in neuroana-
tomical methods that allow a cellular level of resolu-
Functional Systems tion.
The anatomy of functional systems reflects, first
of all, the neural basis of specific neural functions. At Bibliography
the systems level we can introduce a functional aspect Mai, J. K., Assheuer, J., and Paxinos, G. (1997). Atlas of the human
brain. San Diego: Academic Press.
to structural neuroanatomy. By this means we can Nolte, G. (2002). The human brain: An introduction to its functional
begin to address which regions are involved with anatomy. St. Louis: Mosby.
which function and determine those aspects of cellu- Paxinos, G. (1990). The human nervous system. San Diego: Academic
lar and regional anatomy that contribute to special- Press.
ized functions. For example, in the visual system the Alan G. Watts
sensory part is made up of a sensory transducer (the Revised by Yuri Koutcherov and George Paxinos
retina) and a sensory nerve (the optic nerve). After
some initial processing, these components transmit
visual information into two structures in the thalamus
and midbrain (the lateral geniculate nucleus and then AMYGDALA
the superior colliculus) for further processing. Visual The amygdala, a complex of several nuclei located in
information is then projected to the visual regions of the anteromedial part of the temporal lobe, was
the neocortex (in the occipital lobes) for final assess- named in the nineteenth century for its supposed re-
ment. Cellular and regional neuroanatomy can tell us semblance to an almond (Latin, amygdalum) embed-
the detailed structure of each component, but at the ded in the temporal lobe. This portion of the brain
systems level we want to know how the components is involved in a wide range of functions, including
interact: which neurotransmitters are used in which emotion, biologically based behaviors, attention,
connections, which cells receive which type of infor- memory, and learning. It exhibits pathological and
mation, and what routes are used between the various pathophysiological changes in several important neu-
structures. rological and psychiatric diseases, including temporal
It is important to study the connections between lobe epilepsy, Alzheimers disease, schizophrenia,
neuronal groups. There are various techniques that anxiety disorders, and depression.
184 GUIDE TO THE ANATOMY OF THE BRAIN: Amygdala
Figure 1
A. Drawing of a coronal section through the human brain at the level of the amygdala (only the right half of the brain is shown; the
amygdala is actually found on both sides of the brain). Note that the amygdala (gray) is located in the anteromedial part of the
temporal lobe. B. Enlargement of the amygdala at the level shown in A, illustrating the locations of the main amygdalar nuclei.
Investigations in the 1990s demonstrated that the drenaline in the amygdala is essential for the
human amygdala is critical for the recognition of the formation and recall of memories involving emotion-
emotional significance of auditory, visual, and olfac- al events.
tory stimuli, including complex stimuli such as facial
expressions, vocal intonation, and expressive body
movements. These findings came from studies of pa- Intrinsic Structure
tients who had the amygdalar region surgically re- The amygdaloid complex consists of cortical
moved to control epilepsy, patients who have a rare areas on the medial surface of the temporal lobe and
disease (Urbach-Wiethe disease) that exhibits selec- several deep nuclei (see Figure 1). Traditionally, two
tive destruction of the amygdala, and normal individ- major amygdalar nuclear groups were recognized: a
uals who were studied using function magnetic reso- superficial corticomedial group (which included
nance imaging (fMRI). The human amygdala is also the cortical, medial, and central nuclei) and a deeper
important for learning conditioned emotional re- basolateral group (which included the lateral, basal,
sponses (usually fear) to sensory stimuli and events. and accessory basal nuclei). However, studies in the
These findings agree with numerous animal studies 1980s and 1990s indicated that the central and medi-
that have shown that the amygdala is essential for al nuclei exhibit anatomical characteristics that are
classical Pavlovian fear conditioning to simple senso- different from those of the remainder of the corti-
ry cues and to complex sensory representations such comedial group. Therefore, in this entry the amygda-
as the context in which an emotional event has oc- lar nuclei will be divided into three main groups: the
curred. Additional investigations in humans and ani- basolateral nuclear group; cortical nuclear group;
mals have demonstrated that the release of nora- and centromedial nuclear group. In addition, attenu-
GUIDE TO THE ANATOMY OF THE BRAIN: Amygdala 185
ated portions of the centromedial nuclear group ex- chic blindness). As a result, monkeys with amygdalar
tend forward to become continuous with a brain re- lesions exhibit a lack of appropriate emotional and
gion called the bed nucleus of the stria terminalis, social behavior, including a loss of fear and aggres-
which is located in the septal region adjacent to the siveness. In fact, the amygdala is thought to form an
anterior commissure. The term extended amygdala has essential link between brain regions that process sen-
been used to designate the centromedial group and sory information (such as the cerebral cortex and
its forward extensions, including the bed nucleus. thalamus) and brain regions responsible for eliciting
Extensive connections operate within and be- emotional and motivational responses (such as the
tween the amygdaloid nuclei. The major axonal sys- hypothalamus, brain stem, and striatum). For this
tems between nuclei arise in the lateral, basal, and ac- reason the amygdala has been called the sensory
cessory basal nuclei and terminate in successively gateway to the emotions.
more dorsomedial parts of the amygdala. In contrast, The amygdala has extensive connections with the
only moderate projections exist in the opposite direc- olfactory cortex and with sensory association areas in
tion. the cerebral cortex (see Figure 2A). The cortical and
medial nuclei receive olfactory information from the
Neurons olfactory cortex and olfactory bulb. The amygdala
also receives visual and auditory information from the
The cell types in the basolateral and cortical nu- temporal lobe, somatosensory and viscerosensory in-
clear groups are similar to each other. Most of the formation from the insular lobe, and polysensory in-
neurons in both groups are termed pyramidal cells be- formation from the prefrontal cortex and hippocam-
cause they resemble the pyramidal neurons in the ce- pal region. These nonolfactory inputs primarily
rebral cortex. They have large pyramidal-shaped cell target the basolateral and, to a lesser extent, the cen-
bodies and dendrites that exhibit a dense covering of tromedial amygdala. The basolateral but not the cen-
dendritic spines. The pyramidal cells are the main tromedial, amygdalar nuclei have reciprocal projec-
projection neurons of these nuclear groups (i.e., their tions back to these same cortical regions. The latter
axons project out of the amygdala and allow the projections may be important for attention to emo-
amygdala to activate other brain regions). Pyramidal tionally and behaviorally significant stimuli and for
cells are thought to utilize the amino acid glutamate the storage of emotional memories.
as an excitatory neurotransmitter. The remaining cell
types in the basolateral and cortical nuclear groups The amygdala also has connections with a variety
are nonpyramidal neurons. These cells are morpho- of subcortical regions (see Figure 2B). Projections
logically heterogeneous and have dendrites that lack from the thalamus to the amygdala, which arise main-
spines. The axons of these cells establish synaptic con- ly from the midline and intralaminar thalamic nuclei
tacts with neighboring amygdalar neurons but do not and terminate primarily in the basolateral and central
extend beyond the amygdala (i.e., they are interneu- amygdalar nuclei, convey auditory, somatosensory,
rons). They utilize gamma-aminobutyric acid (GABA) viscerosensory, and visual information to the amygda-
as an inhibitory neurotransmitter. la. Amygdalothalamic projections are more limited
and consist of projections from the central nucleus to
Unlike the nuclei of the basolateral and cortical the midline thalamic nuclei and from the basolateral
nuclear groups, the cell types of the centromedial amygdala to the mediodorsal thalamic nucleus.
group do not resemble those of the cerebral cortex.
Neurons in the lateral portions of the centromedial Extensive reciprocal connections operate be-
amygdala resemble the medium-sized spiny neurons tween the medial portions of the preoptic-
of the adjacent caudate and putamen. Most of the hypothalamic region and the amygdala, particularly
neurons in the centromedial nuclear group contain the medial amygdalar nucleus, the cortical nuclei,
neuropeptides, GABA, or both. Neurons in the more and medial portions of the basolateral amygdala.
rostral parts of the extended amygdala (e.g., the bed Consistent with these connections, stimulation and le-
nucleus of the stria terminalis) are similar to the cell sion studies in experimental animals have shown that
types found in the central and medial amygdalar nu- the amygdala is involved in behavior related to bio-
clei. logical drives and motivation, including arousal, ori-
enting, and sleep; fight or flight; feeding and drink-
ing; and social, reproductive, and maternal behavior.
Functional Anatomy In humans these behaviors are often associated with
Bilateral damage to the amygdaloid nuclei in emotional feelings (e.g., fear with flight, anger and
monkeys produces the Kluver-Bucy syndrome, in rage with fighting and defensive behavior). In each of
which the animal is unable to appreciate the emotion- these emotional states the amygdala generates a coor-
al and behavioral significance of sensory stimuli (psy- dinated response consisting of autonomic, endocrine,
186 GUIDE TO THE ANATOMY OF THE BRAIN: Amygdala
Figure 2
A. Drawing of a lateral view of the human brain illustrating the anatomy of the main cortical pathways conveying
sensory information to the amygdala. Note that somatosensory, auditory, and visual information is transmitted to the
amygdala over polysynaptic cortical pathways; only higher order cortical areas involved in processing the most complex
sensory information in these modalities have projections to the amygdala. B. Drawing of a medial view of the human
brain illustrating the connections of the amygdala with subcortical brain regions. All connections are reciprocal except
those to the caudate and nucleus accumbens, which do not have projections back to the amygdala.
GUIDE TO THE ANATOMY OF THE BRAIN: Basal Forebrain 187
and behavioral components by way of its projections REINFORCEMENT OR REWARD IN LEARNING:

to various subcortical regions, especially the hypo- STRIATUM
thalamus. The endocrine responses produced by
Bibliography
amygdalar stimulation are due to its activation of hy-
Aggleton, J. P., ed. (1992). The amygdala: Neurobiological aspects of
pothalamic areas that control the secretion of hor- emotion, memory, and mental dysfunction. NewYork: Wiley-Liss.
mones by the pituitary gland. (2000). The amygdala: A functional analysis. Oxford: Oxford
Another significant subcortical target of the University Press.
Alheid, G. F., de Olmos, J. S., and Beltramino, C. A. (1995). Amyg-
amygdala that is important for producing behavioral dala and extended amygdala. In George Paxinos, ed., The rat
responses is the striatum (caudate, putamen, and nu- nervous system, pp. 495578. Orlando, FL: Academic Press.
cleus accumbens). This projection originates mainly Gloor, P. (1997). The temporal lobe and limbic system. New York: Ox-
in the basolateral nuclear group and terminates pri- ford University Press.
marily in the ventral and medial portions of the stria- LeDoux, J. (1996). The emotional brain. New York: Simon and Schu-
ster.
tum, including the nucleus accumbens. Lesion studies McDonald, A. J. (1998). Cortical pathways to the mammalian
indicate that the projections of the basolateral amyg- amygdala. Progress in Neurobiology 55, 257331.
dala to the striatum are important for controlling be- Price, J. L., Russchen, F. T., and Amaral, D. G. (1987). The limbic
havior related to the reinforcing properties of sensory region. Part 2: The amygdaloid complex. In A. Bjrklund, T.
stimuli. Hkfelt, and L.W Swanson, eds., Handbook of chemical
neuroanatomy, Vol. 5, pp. 279388. Amsterdam: Elsevier.
The central nucleus is the main amygdalar region
Joseph L. Price
exhibiting connections with the brain stem and basal
Revised by Alexander J. McDonald
forebrain. Among these targets are several areas in-
volved in visceral function, including the parabrachial
nucleus, dorsal vagal nucleus, and nucleus solitarius.
In addition, the central nucleus innervates several BASAL FOREBRAIN
brain regions that give rise to neurotransmitter-
The septal area, the diagonal band nuclei, and the
specific fiber systems that target the amygdala and
nucleus basalis of the substantia innominata are com-
other forebrain areas. These regions include the locus
ponents of the basal forebrain. These structures lack
ceruleus (noradrenergic), substantia nigra and ven-
a true cortical organization but can be said to have a
tral tegmental area (dopaminergic), raphe nuclei
corticoid architecture because of their location on
(serotonergic), and the nucleus basalis (cholinergic).
the surface of the cerebral hemispheres (Mesulam,
The latter region is also innervated by portions of the
2000). The basal forebrain projects to many different
basolateral nuclear group. These transmitter-specific regions, using excitatory amino acids, GABA, and
systems are activated in particular behavioral states, acetylcholine as the transmitters. Its single most im-
particularly during stress, and can modulate amygda- portant output is a cholinergic projection directed to
lar activities related to emotion, attention, and mem- the hippocampus, amygdala, and all other parts of
ory. the cerebral cortex (see Figure 1). The projections
from the septal and diagonal band nuclei to the hip-
Conclusion pocampus travel in the fornix. In the human brain,
the massive projections from the nucleus basalis to
The amygdala is an anatomically complex region the cerebral cortex travel in the external and extreme
of the brain that contains many distinct nuclei. Each capsules, the uncinate fasciculus, and the cingulum
nucleus is characterized by distinctive connections (Selden et al., 1998).
with specific brain areas. By way of these connections
the amygdala generates a variety of behaviors related The medial septal nucleus of the primate brain is
to basic biological drives and motivation. The amyg- an inconspicuous structure containing relatively small
dala also appears to be critical for the formation of neurons. Less than half of the medial septal neurons
emotional memories. are cholinergic and correspond to the Ch1 sector of
the basal forebrain. A second group of somewhat larg-
er cholinergic neurons is embedded within the verti-
See also: EMOTION, MOOD, AND MEMORY; GENETIC cal nucleus of the diagonal band of Broca, a nucleus
SUBSTRATES OF MEMORY: AMYGDALA; LONG-
that is usually considered a component of the septal
TERM POTENTIATION: AMYGDALA; NEURAL
SUBSTRATES OF CLASSICAL CONDITIONING:
complex. Approximately three-quarters of the neu-
FEAR CONDITIONING, FREEZING; NEURAL rons of the vertical limb nucleus are cholinergic and
SUBSTRATES OF CLASSICAL CONDITIONING: constitute the Ch2 sector. The substantia innominata
FEAR-POTENTIATED STARTLE; NEURAL (or the subcommissural gray) is a complex region
SUBSTRATES OF EMOTIONAL MEMORY; PASSIVE composed of the ventral globus pallidus, the nucleus
(INHIBITORY) AVOIDANCE, FEAR LEARNING; basalis of Meynert, and the horizontal nucleus of the
188 GUIDE TO THE ANATOMY OF THE BRAIN: Basal Forebrain
diagonal band. A small percentage (a tenth or less) of Gorry (1963) pointed out that the nucleus basalis
the neurons in the horizontal limb nucleus are cholin- displays a progressive evolutionary trend, becoming
ergic and constitute the Ch3 sector. These neurons more and more extensive and differentiated in more
tend to be hypochromic on Nissl stains and are highly evolved species, especially in primates and ce-
shaped like a spindle. taceans. Observations of the brains of turtles, mice,
rats, squirrel monkeys, rhesus monkeys, and humans
The most conspicuous group of cholinergic neu- are consistent with this general view. There are con-
rons in the primate brain is found within the nucleus siderable interspecies differences in the organization
basalis of Meynert. Approximately 90 percent of the of these cholinergic pathways. The rodent cerebral
larger neurons in the nucleus basalis of the monkey cortex contains intrinsic cholinergic interneurons
and human brain are cholinergic and constitute the whereas the primate does not, making it entirely de-
Ch4 sector (see Figure 2). In the human brain, each pendent on afferents from the basal forebrain for its
hemisphere may contain approximately 200,000 nu- cholinergic innervation. Furthermore, the Ch4 neu-
cleus basalis neurons, about 90 percent of which be- rons of the primate express calbindin, whereas those
long to Ch4 (Arendt, Bigl, Tennstedt, and Arendt, of the rodent do not.
1985). The anteroposterior extent of the human Ch4 The Ch4nucleus basalis complex displays a par-
complex is 1.5 to 2 centimeters. In addition to the tial overlap with surrounding cell groups of the pre-
cholinergic neurons within the cytoarchitectonic con- optic area, hypothalamus, striatum, diagonal band of
fines of the nucleus basalis, there are interstitial cho- Broca, amygdala, and globus pallidus. There is no
linergic neurons embedded within the anterior com- strict delineation between nuclear aggregates and
missure, the ansa peduncularis, the ansa lenticularis, passing fiber tracts. As noted above, many Ch4 neu-
and the internal capsule (Mesulam and Geula, 1988). rons are embedded within the internal capsule, the
These neurons can be considered part of the Ch4 diagonal bands of Broca, the anterior commissure,
complex on the basis of morphological, cytochemical, the ansa peduncularis (inferior thalamic peduncle),
and hodological criteria. and the ansa lenticularis. In fact, previous designa-
GUIDE TO THE ANATOMY OF THE BRAIN: Basal Forebrain 189
tions for the nucleus basalis included nucleus of the dala and the frontoparietal operculum; the Ch4i sub-
ansa peduncularis and nucleus of the ansa lenticu- sectors provide the major cholinergic innervation for
laris. The physiological implication of this intimate peristriate, inferotemporal, and lateral frontoparietal
association with fiber bundles is unknown. In addition cortex; and the Ch4p subsector provides the major
to this open nuclear structure, the neurons of Ch4 are source of cholinergic innervation for the superior
of variable shape and have an isodendritic morpholo- temporal gyrus and the temporopolar area. Not all
gy with overlapping dendritic fields. These character- cortical areas receive an equal density of cholinergic
istics, also present in the nuclei of the brain-stem re- fibers. There is a much more intense cholinergic in-
ticular formation, have led to the suggestion that the nervation in limbic and paralimbic areas than in asso-
Ch4 complex could be conceptualized as a telence- ciation and primary sensory-motor areas. Cholinergic
phalic extension of the brain-stem reticular core drugs may therefore be expected to have their great-
(Ramon-Moliner and Nauta, 1966). est impact on limbic and paralimbic areas.
Studies using monkeys (based on the concurrent
Limbic and cortical areas of the primate brain
demonstration of perikaryal cholinergic markers and
retrograde transport) have shown that each group of contain several different types of postsynaptic musca-
cholinergic cells projects widely but also with some rinic and nicotinic receptors. The dominant species
degree of topographical specificity. According to is the pirenzepine-sensitive m1 subtype of muscarinic
these studies, Ch1 and Ch2 collectively provide the receptor. The regional distribution of this receptor
major source of cholinergic input for the hippocam- subtype shows a relatively good agreement with the
pal formation, Ch3 provides the major cholinergic regional distribution of presynaptic cholinergic fibers
input to the olfactory bulb, and Ch4 provides the (Mash, White, and Mesulam, 1988). Electron micro-
major cholinergic innervation for the amygdala and scopic examinations of immmunostained tissue shows
all neocortical regions. The primate Ch4 can be divid- that incoming cholinergic fibers make traditional syn-
ed into anteromedial (Ch4am), anterolateral (Ch4al), aptic contact predominantly on dendritic spines of
intermediate (Ch4i), and posterior (Ch4p) subsectors. cortical pyramidal neurons (Smiley, Morrell, and Me-
Each cortical area receives its cholinergic input pri- sulam, 1997). At m1 muscarinic receptor sites, the
marily (but not exclusively) from a specific subsector acetylcholine that is released by cortical cholinergic fi-
of Ch4. For example, Ch4am is the major source of bers reduces the potassium conductance of the post-
cholinergic innervation for the cingulate gyrus and synaptic membrane and promotes the activation of
adjacent medial cortical areas; the Ch4al subsector is cholinoceptive neurons by other excitatory inputs
the major source of cortical innervation for the amyg- (Krnjevic, 1981). These physiological properties have
190 GUIDE TO THE ANATOMY OF THE BRAIN: Basal Forebrain
led to the designation of acetylcholine as an excitato- See also: ALZHEIMERS DISEASE: HUMAN DISEASE
ry neuromodulator. AND THE GENETICALLY ENGINEERED ANIMAL
MODELS; COGNITIVE ENHANCERS;
The cholinergic Ch4 neurons receive cholinergic, NEOCORTICAL PLASTICITY: AUDITORY CORTEX;
glutamatergic, GABAergic, noradrenergic, serotoner- PHARMACOLOGICAL TREATMENT OF MEMORY
gic, and dopaminergic inputs (Zborszky, Cullinan, DEFICITS
and Luine, 1993; Smiley, Morrell, and Mesulam,
1999). All cortical areas receive cholinergic input, but Bibliography
only limbic and paralimbic areas send substantial Arendt, T., Bigl, V., Tennstedt, A., and Arendt, A. (1985). Neuro-
neural projections back to the nucleus basalisCh4 nal loss in different parts of the nucleus basalis is related to
complex (Mesulam and Mufson, 1984). This anatomi- neuritic plaque formation in cortical target areas in Al-
cal arrangement indicates that most cortical areas zheimers disease. Neuroscience 14, 114.
Berger-Sweeney, J., Heckers, S., Mesulam, M-M, Wiley, R. G.,
have no direct feedback control over the cholinergic Lappi, D. A., and Sharma, M. (1994). Differential effects upon
innervation that they receive, whereas limbic and par- spatial navigation of immunotoxin-induced cholinergic le-
alimbic areas have powerful feedback control over the sions of the medial septal area and nucleus basalis magnocel-
cholinergic input that they receive and over the cho- lularis. Journal of Neuroscience 14, 4,5074,519.
Gorry, J. D. (1963). Studies on the comparative anatomy of the
linergic input directed to other parts of cortex. The
ganglion basale of Meynert. Acta Anatomica 55, 51104.
Ch4 complex may thus act as a cholinergic relay for Heckers, S., and Mesulam M-M. (1994). Two types of cholinergic
rapidly shifting cortical activation in a way that re- projections to the rat amygdala. Neuroscience 60, 383397.
flects the emotional-motivational state encoded by Kilgard, M. P., and Merzenich, M. M. (1998). Cortical map reorga-
the limbic system. A restricted corticofugal control of nization enabled by nucleus basalis activity. Science 279,
1,7141,718.
widely distributed corticopetal pathways appears to Krnjevic, K. (1981). Cellular mechanisms of cholinergic arousal.
be a feature common to other transmitter-specific sys- Behavioral and Brain Sciences 4, 484485.
tems (e.g., monoaminergic) that are also implicated Mash, D. C., White, W. F., and Mesulam, M-M. (1988). Distribution
in setting global behavioral states. of muscarinic receptor subtypes within architectonic subre-
gions of the primate cerebral cortex. Journal of Comparative
Single-unit studies indicate that neurons in the Neurology 278, 265274.
nucleus basalis of the rhesus monkey are sensitive to Mesulam, M-M. (2000). Behavioral neuroanatomy: Large-scale
networks, association cortex, frontal syndromes, the limbic
sensory information that signals the delivery of re- system, and hemispheric specialization. In M-M Mesulam,
ward (Wilson and Rolls, 1990). Cholinergic projec- ed., Principles of behavioral and cognitive neurology, 1120. New
tions may thus help to enhance the cortical impact of York: Oxford University Press.
motivationally relevant events. Such an effect would Mesulam, M-M., and Geula, C. (1988). Nucleus basalis (Ch4) and
cortical cholinergic innervation in the human brain: Observa-
influence both selective attention and learning. Ex-
tions based on the distribution of acetylcholinesterase and
periments in various animal models are consistent choline acetyltransferase. Journal of Comparative Neurology 275,
with this dual role of cortical cholinergic innervation 216240.
(Berger-Sweeney et al., 1994; Sarter and Bruno, Mesulam, M-M., and Mufson, E. J. (1984). Neural inputs into the
2000; Voytko et al., 1994). Cholinergic projections nucleus basalis of the substantia innominata (Ch4) in the rhe-
sus monkey. Brain 107, 253274.
from the basal forebrain also influence learning- Ramon-Moliner, E., and Nauta, W. J. H. (1966). The isodendritic
induced physiological and structural plasticity within core of the brain. Journal of Comparative Neurology 126, 311
the cerebral cortex (Kilgard and Merzenich, 1998). 336.
Sarter, M., and Bruno, J. P. (2000). Cortical cholinergic inputs me-
The vast majority of Ch4 neurons express the p75 diating arousal, attentional processing, and dreaming: Differ-
NGF receptor (NGFr). These neurons are dependent ential afferent regulation of the basal forebrain by telence-
on NGF retrogradely transported from the cerebral phalic and brainstem afferents. Neuroscience 95, 933952.
Selden, N. R., Gitelman, D. R., Salamon-Murayama, N., Parrish,
cortex for survival. This makes them particularly vul-
T. B., and Mesulam, M-M. (1998). Trajectories of cholinergic
nerable to cortical diseases such as Alzheimers dis- pathways within the cerebral hemispheres of the human
ease (AD). The few NGFr-negative Ch4 neurons se- brain. Brain 121, 2,2492,257.
lectively project to the amygdala and adjacent Smiley, J. F., Morrell, F., and Mesulam, M-M. (1997). Cholinergic
structures (Heckers and Mesulam, 1994). There is an synapses in human cerebral cortex: An ultrastructural study
in serial sections. Experimental Neurology 144, 361368.
age-related loss of calbindin in human Ch4 neurons Smiley, J. F., Subramanian, M., and Mesulam, M-M. (1999) Mono-
(Wu, Mesulam, and Geula, 1997). This may contrib- aminergic-cholinergic interactions in the primate basal fore-
ute to their vulnerability to neurofibrillary degenera- brain. Neuroscience 93, 817829.
tion in AD. Neurofibrillary tangles in Ch4 and a se- Voytko, M.L., Olton, D. S., Richardson, R. T., Gorman, L. K.,
vere loss of cortical cholinergic innervation are Tobin, J. R., and Price, D. L. (1994). Basal forebrain lesions
in monkeys disrupt attention but not learning and memory.
hallmarks of AD neuropathology. Cholinesterase- Journal of Neuroscience 14, 167186.
inhibiting drugs are used to treat AD, with the aim of Wilson, F. A. W., and Rolls, E. T. (1990). Neuronal responses relat-
reversing some of this cholinergic depletion. ed to novelty and familiarity of visual stimuli in the substantia
GUIDE TO THE ANATOMY OF THE BRAIN: Basal Ganglia 191
innominata, diagonal band of Broca and periventricular re- Figure 1

gion of the primate basal forebrain. Experimental Brain Re-
search 80, 104120.
Wu, C.-K, Mesulam, M-M., and Geula, C. (1997). Age-related loss
of calbindin from basal forebrain cholinergic neurons. Neu-
roReport 8, 2,2092,213.
Zborszky, L., Cullinan, W. E., and Luine, V. N. (1993). Catechol-
aminergic-cholinergic interaction in the basal forebrain. Prog-
ress in Brain Research 98, 3149.
M-Marsel Mesulam
BASAL GANGLIA
The basal ganglia are subcortical nuclei that are high-
ly developed in primates and are strongly intercon-
nected with the neocortex. They are major compo-
nents of the telencephalon (endbrain) in mammals
and lie beneath the cerebral cortex, which forms the
outer sheet of the endbrain. The basal ganglia include
Schematic information flow diagram showing the basic release
two well-known parts of the extrapyramidal motor circuit of the basal ganglia. Darker shading, main elements of
system (the striatum and the pallidum). Technically, the basal ganglia. Many neocortical areas project to the striatum,
the amygdala is also part of the basal ganglia, but, as which also receives input from the amygdala (Amyg), the
a functional system quite different from the striato- intralaminar thalamus (CM-Pf), and the substantia nigra pars
pallidal complex, falls outside the scope of this article. compacta (SNc), which provides dopamine (DA) to the striatum.
The basal ganglia are involved in motor and psy- There is a double inhibitory pathway from the striatum to the
internal pallidum (GPi) and substantia nigra pars reticulata
chomotor control and are therefore a central focus for
(SNr) and then on to the thalamus. Thus excitatory input
studies of Parkinsons disease and Huntingtons dis-
coming from the cortex excites the thalamocortical projection of
ease, motor disorders involving reduced movement the system, releasing it from chronic inhibition by the GPi/SNr.
capacity (Parkinsons disease) or too much movement Other outputs of the system are shown leading downstream to
(Huntingtons disease). It is recognized that the basal the superior colliculus (Supp. Coll.) and the pendunculopontine
ganglia can affect neuropsychiatric and cognitive nucleus (PPN). Note in Figure 2 how this basic circuit, the direct
functions and that basal ganglia dysfunction thus may pathway of the basal ganglia, is paired with the indirect pathway.
contribute to neuropsychiatric disorders. Examples of
neuropsychiatric disorders involving the basal gan-
glia include obsessive-compulsive disorder and Tou- tions, probably including cognitive actions. An im-
rettes syndrome. portant aspect of the circuit is that most of the striatal
The basal ganglia are part of cortico-basal gan- neurons that project to the pallidum (projection neu-
glia circuits receiving inputs from the neocortex (and rons) fire action potentials phasically at low rates. But
thalamus), processing that information, interacting pallidal neurons fire tonically (nearly continuously) at
with modulatory loop-circuits, and passing the pro- high rates. These physiological findings suggest that
cessed information on to the frontal neocortex (via the basal ganglia tonically inhibit their targets (e.g.,
the thalamus) and to brainstem targets such as the su- the thalamus) but that when the striatum becomes ac-
perior colliculus and the reticular formation (see Fig- tive, this tonic inhibition is briefly (phasically) re-
ure 1). Nearly the entire cortex projects to the stria- leased. Because the inputs to the striatum are excit-
tum, which is considered the main input site of the atory, the general circuit plan involves cortical
basal ganglia. The striatum, in turn, projects to the excitation of the striatum, leading through the double
pallidum (globus pallidus), which gives rise to the inhibition to release of the thalamus. This release
main basal ganglia outflow to the thalamus and other from inhibition (known as disinhibition) is considered
sites. A remarkable characteristic of these basal gan- to underlie the movement release-inhibition func-
glia connections is that they are inhibitory. The stria- tions of the basal ganglia.
tum inhibits the pallidum, and the pallidum inhibits The striatum and pallidum act in close coopera-
the thalamus. This means that activation of the stria- tion with two other nuclei, the substantia nigra and
tum can release the thalamus and other output tar- the subthalamic nucleus. The substantia nigra lies in
gets (see Figure 1). the midbrain and attains a very large size in the
This double-inhibition circuit is considered to be human brain. The substantia nigra has two parts, one
critical to the release of movements and complex ac- of which is called the pars reticulata and is very much
192 GUIDE TO THE ANATOMY OF THE BRAIN: Basal Ganglia
like the pallidum. The nigral pars reticulata is, judg- Figure 2
ing by its anatomy, likely to be a differentiated extra
part of the pallidum displaced caudally into the mid-
brain. Like the pallidum, the pars reticulata of the
substantia nigra receives input from the striatum and
projects strongly to the thalamus. An important dif-
ference from the pallidum is that the nigral pars re-
ticulata also projects to the superior colliculus, a struc-
ture involved in controlling eye movements
(especially saccadic eye movements). The second part
of the substantia nigra is the pars compacta. Its neu-
rons synthesize the neurotransmitter dopamine, and
they give rise to the dopamine-containing nigrostria-
tal tract, which innervates the striatum and releases
dopamine there. In Parkinsons disease, these neu-
rons degenerate, leading to a loss of dopamine in the
striatum.
The second nucleus closely associated with the
striatum and pallidum is the subthalamic nucleus. Schematic flow diagram showing the balance between the direct
and indirect pathways of the basal ganglia. The movement
This nucleus (named for the fact that it lies in the ter-
release pathway (shown as black in Figure 1) is called the direct
ritory underneath the thalamus) is a key regulator of
pathway, and the opposing pathway is called the indirect
the release-inhibition functions of the basal ganglia pathway (shown here). When the two are in balance, the system
(see Figure 2). It receives inhibitory input from the is at equilibrium. When the direct pathway is more active, the
pallidums so-called external segment and sends ex- thalamus is disinhibited, but when the indirect pathway is more
citatory output back to the pallidum. It thus is disinhi- active, the thalamus is inhibited. Asterisks indicate focal sites for
bited when the striatum is phasically activated, and it neurosurgical intervention in Parkinsons disease therapy.
excites the pallidums internal segment. This sub- Dagger indicates site for dopamine replenishment or dopamine
thalamic loop or indirect pathway opposes the action receptor activation used in Parkinsons disease therapy and site
of the direct pathway from striatum to internal pal- of primary neurodegeneration in Huntingtons disease. Plus
lidum to thalamus (see Figure 2). signs indicate excitatory (glutamatergic) connections; minus
signs indicate inhibitory (GABAergi) connections.
The striatum, as the input side of the basal gan-
glia and as the first stage of the main pathways lead-
ing out from the basal ganglia, sets up important projections from the midbrain to the striatum. For
functional subdivisions of the basal ganglia and its cir- example, medial to the nigra substantia pars compac-
cuits. The striatum has three anatomical subdivisions ta, in and near the midline of the midbrain, there are
that roughly correspond to functional parts: the cau- other dopamine-containing neurons that form the
date nucleus, the putamen, and the ventral striatum. ventral tegmental area. This region innervates the
The caudate nucleus makes up the largest part of the ventral striatum and is part of reward circuits of the
striatum at anterior levels and receives strong inputs
brain, mediating reward-based behaviors and some
from the frontal cortex and some other areas of asso-
forms of drug addiction (e.g., to cocaine and amphet-
ciation cortex. The putamen is the large, laterally
amine).
placed nucleus of the striatum and receives most of
the input from sensorimotor and association cortex. These functional subdivisions are differentially
The ventral striatumwhich, as its name implies, lies affected in the major disorders associated with basal
at the base of the striatumreceives inputs related to ganglia dysfunction. The dorsal striatum (caudate nu-
the limbic system (including inputs from the hippo- cleus and putamen) are abnormal in Parkinsons dis-
campal formation and amygdala). ease and Huntingtons disease. The most severe
All three of these large subdivisions of the stria- motor disturbances in these disorders are associated
tum project to corresponding parts of the pallidum with loss of dopamine (Parkinsons disease) or neu-
and substantia nigra. There is considerable evidence rons (Huntingtons disease) in the putamen, which re-
that these pathways are fairly distinct from one anoth- ceives inputs from sensory and motor cortex. Cogni-
er, so that the functional channels set up in the stria- tive-affective disturbances in these disorders are
tum are maintained in the pallidum and substantia associated with loss of function in the caudate nucle-
nigra and, in their outflow pathways, to the rest of the us, which receives inputs from frontal areas of the as-
brain. This is true also for the dopamine-containing sociation cortex. Neuropsychiatric disorders ranging
GUIDE TO THE ANATOMY OF THE BRAIN: Cerebellum 193
from obsessive-compulsive disorder to depression are Bibliography

also associated with abnormal function of the caudate Graybiel. A. M. (1997). The basal ganglia and cognitive pattern
nucleus. Dysfunction of the ventral striatum is sus- generators. Schizophrenia Bulletin 23, 459469.
Graybiel, A. M., Aosaki, T., Flaherty, A. W., and Kimura, M. (1994).
pected in some psychiatric disorders, including schiz-
The basal ganglia and adaptive motor control. Science 265,
ophrenia. The ventral tegmental area, which projects 1,8261,831.
to the ventral striatum, is largely spared in Parkin- Graybiel A. M., and Penney J. B. (1999). Chemical architecture of
sons disease. The anatomical organization of basal- the basal ganglia. In F. E. Bloom, A. Bjrklund, and T. Hk-
ganglia pathways suggests that the cortically directed felt, eds., Handbook of chemical neuroanatomy. Amsterdam: Else-
vier.
outflow of the basal ganglia mainly targets executive Graybiel, A. M., and Rauch, S. L. (2000). Toward a neurobiology
areas of the neocortex, including motor, premotor, or of obsessive-compulsive disorder. Neuron 28, 343347.
prefrontal cortex. The breadth of frontal cortex af- Hikosaka O., Miyashita K., Miyachi S., Sakai K., and Lu, X. (1998).
fected may also help to account for the broad func- Differential roles of the frontal cortex, basal ganglia, and cer-
ebellum in visuomotor sequence learning. Neurobiology of
tional influences of the basal ganglia suggested by
Learning and Memory 70 137149.
clinical evidence. Leckman, J. F., and Riddle, M. A. (2000). Tourettes syndrome:
When habit-forming systems form habits of their own? Neuron
Some disorders affecting the basal ganglia are as- 28, 349354.
sociated with major changes in neurotransmitter sys- Mink, J. W. (1996). The basal ganglia: Focused selection and inhi-
tems in basal-ganglia circuits. Drugs affecting these bition of competing motor programs. Progressive Neurobiology
systems include not only levodopa, given to Parkin- 50, 381425.
Salmon, D. P., and Butters, N. (1995). Neurobiology of skill and
sons patients as a replacement therapy for the lost habit learning. Current Opinion in Neurobiology 5, 184190.
dopamine, but also agents with powerful effects on White, N. M. (1997). Mnemonic functions of the basal ganglia. Cur-
mental activity and behavior, including antipsychotics rent Opinion in Neurobiology 7, 164169.
such as haloperidol (a dopamine receptor antagonist) Wichmann T., and DeLong, M. R. (1996). Functional and patho-
and psychoactive drugs such as marijuana. This large physiological models of the basal ganglia. Current Opinion in
Neurobiology 6, 751758.
range also conforms to evidence that the basal ganglia
mediate cognitive-affective as well as motor functions. Ann M. Graybiel
The basal ganglia are implicated not only in the

continuing control of action but also in the learning
mechanisms that underlie the development of the CEREBELLUM
near-automatic behaviors that we think of as habits That cerebellums importance as a component of the
and rituals. Studies of these learning functions of the motor system is clearly indicated by its anatomy and
basal ganglia suggest that the anatomical circuits by the effects of lesions or pathology. Abnormal de-
summarized here are actually highly dynamic net- velopment of or injury to the cerebellum results in se-
works. Recordings from behaving animals suggest vere impairment of the ability to produce accurate
that the activity patterns of neurons in the striatum and coordinated movements. Normal, everyday
undergo major changes as the animals become condi- movements become uncoordinated and clumsy, hob-
tioned or learn new procedures. A teaching signal for bled by errors in direction, rate, amplitude, sequence,
the striatal neurons is thought to come from the dop- and precision. Depending on which portions of the
amine-containing neurons of the substantia nigra. cerebellum are damaged, these impairments can in-
This new evidence suggests that the very nuclei dis- clude abnormal voluntary movements of the limbs
abled in disorders such as Parkinsons disease and (e.g., reaching movements that miss the target), ab-
Huntingtons disease normally help to modify corti- normal eye movements (e.g., saccades that miss the
co-basal ganglia circuits as a result of experience, so target), or abnormal vestibular (balance) reflexes and
that habits and procedures can be learned and pro- postural adjustments (e.g., patients struggling simply
duced as whole sequences or chunks. Such automa- to stand without falling over).
tized behaviors are fundamentally important in free-
How the neurons and synapses of the cerebellum
ing the brain to react to new events in the contribute to accurate and coordinated movements
environment and to carry out many cognitive func- has been the subject of much research and debate.
tions. The basal ganglia, then, may provide a base for The anatomy of the cerebellum indicates that some
cognitive activity as well as for motor activity. sort of sensory-motor integration is taking place.
There is also evidence that the cerebellum contributes
See also: REINFORCEMENT OR REWARD IN not only to the learning of and adaptation to move-
LEARNING: ANATOMICAL SUBSTRATES; ments based on experience, but also to the proper
REINFORCEMENT OR REWARD IN LEARNING: timing of movements. These various ideas about cere-
STRIATUM bellar function are all related to a relatively simple
194 GUIDE TO THE ANATOMY OF THE BRAIN: Cerebellum
Figure 1
The cerebellum (in gray) varies widely in relative size, shape, and folding pattern in different animals: A. alligator; B.
pigeon; C. opossum; D. domestic cat; E. chimpanzee. All brains are viewed from the left side and are drawn to the same
scale. In the larger animals (cat and chimp), the cerebral cortex extends backward over the top of the cerebellum,
concealing its greater size and complexity.
Figure 2. Stained 30-micron section through the whole cerebellum (A) and through a single fold or folium (B) of a domestic cat. In
A the front of the brain is to the right, where cerebral cortex (Cor) overlies a portion of the cerebellum and an auditory center
(inferior colliculus, IC) protrudes up in front of the cerebellum at the anterior part of the medulla (Med) of the brainstem. The
asterisk (*) in A indicates the folium that is enlarged in B. In B, the molecular layer consists mostly of parallel fibers as well as the
dendrites (not stained) of Purkinje, Golgi, basket, and stellate cells. The large Purkinje cells lie in a row just above the densely packed
tiny granule cells. Magnification of each photograph is indicated by the scale bars. Abbreviations: DN, deep cerebellar nuclei; GC,
granule cells; Mol, molecular layer; PC, Purkinje cells; W, white matter (axons). (T. P. Stewart and Carol L. Dizack)
function that the cerebellum performs. It uses sensory sists of numerous axons, providing connections be-
information in a particular way (feedforward) to con- tween the cortex, deep nuclei, and other parts of the
tribute to the accuracy of movements. Understanding central nervous system.
what this means and how the cerebellum accomplish-
All ideas about cerebellar function stem from its
es this task requires a basic appreciation of the anato-
unique connectivity or wiring diagram. Although the
my and connectivity of the cerebellum.
cerebellum contains a great many neurons, there are
All vertebrates have a cerebellum (see Figure 1). only seven types of neurons in the cerebellar cortex
It lies behind the forebrain and on top of the hind- that are interconnected in highly organized and spe-
brain. If you grab the back of your neck (fairly high), cific ways (see Figure 3). Thus, the wiring diagram of
your hand surrounds the cerebellum. It is a small, the cerebellum lends itself to accessible depiction of
smooth protuberance in most fish, amphibians, and the sort shown in Figure 4. Neurons of the deep nu-
reptiles, but in birds and mammals the surface of the clei provide the output of the cerebellum; they project
cerebellum is folded into a complex arrangement of to the motor nuclei such as the red nucleus and vestib-
thin, elongated folia. The size, shape, and complexity ular nuclei, and to the VL region of thalamus, which
of the cerebellum vary widely in different mammals; projects mostly to motor regions of cerebral cortex.
those that have more complex behavioral repertoires The Purkinje cells are the sole output of the cerebel-
that are modifiable by learning are likely to have a lar cortex; they inhibit neurons of the deep nuclei.
larger and more convoluted cerebellum. Mossy fibers represent one type of input to the cere-
bellum. They project directly to the deep nucleus
Figure 2 illustrates some of the unique anatomi- neurons and to the cerebellar cortex. In the cortex,
cal features of the cerebellum. A stained section just they branch diffusely to contact numerous granule
lateral to the midline reveals the highly folded cere- and Golgi cells. In turn, the granule cell axons, the
bellar cortex (each fold is a folia), which consists of a parallel fibers, make excitatory connections onto
three-layer sheet of neurons. The cerebellar cortex is many Purkinje cells. Despite its many subtleties, the
about one millimeter thick in all mammals. Near the main properties of the cerebellums wiring diagram
base of the cerebellum are several clusters of neurons are relatively simple. Output via the deep nucleus
that make up the cerebellar deep nuclei, which pro- neurons is influenced by two parallel pathways: the
vide the output of the cerebellum. The white matter direct excitatory pathway from mossy fibers and the
beneath the cortex and around the deep nuclei con- more complex inhibitory pathway from the Purkinje
Figure 3
Schematic drawing of the major cell and axon types and their connections in a portion of cerebellar cortex. Connections to the deep
nuclei (Nuclei) are also shown.
cells. A simple numeric ratio illustrates the relative weak collateral projections to the deep nuclei. Where-
complexity of these pathways. For every deep-nucleus as each Purkinje cell has around 100,000 inputs from
output neuron there are about 2 million parallel fiber different granule cells, it gets input from only one
synapses onto the Purkinje cells. climbing fiber. The climbing fiber branches profuse-
ly, climbs over the dendrites of the Purkinje cell,
A second type of input to the cerebellum, the and makes numerous synapses. This spatially distrib-
climbing fibers, are quite different from the mossy fi- uted input is so powerful that each action potential in
bers. Climbing fibers project to the cerebellar cortex the climbing fiber produces an all-or-none response
and contact only a few Purkinje cells. There are also in the Purkinje cell, which, in part, involves the wide-
Figure 4
A schematic representation of the synaptic organization or wiring diagram of the cerebellum. Cells in the deep nuclei provide the
output of the cerebellum. These neurons are controlled by excitatory mossy fiber inputs and inhibitory input from the Purkinje cells of
the cerebellar cortex. The mossy fiber inputs to the cerebellar cortex excite both granule and Golgi cells. The granule cells provide
excitatory input to the Purkinje cells via their axons, which are called parallel fibers. Climbing fiber inputs to the cerebellum project
primarily to the Purkinje cells.
spread influx of calcium into the Purkinje cell. This to Purkinje synapses that were active at that time.
calcium influx is important for the synaptic changes In this way the movement could be different next
that mediate the learning in the cerebellum. time, and this learning would be specific for the
In 1969 David Marr published a theory of cere- combination of circumstances encoded by that
bellum that proposed that the wiring diagram de- particular pattern of granule-cell activity.
scribed above is well suited to adapt movements via
The basic tenets of this theory have been sup-
learning. There are three main components of Marrs
ported by a wide variety of experimental evidence.
theory:
There are several well-characterized examples of
The mossy fiber/granule cell inputs encode what motor learning that require the cerebellum, and plas-
is happening for the Purkinje cells. Mossy fibers ticity in the cerebellum controlled by climbing fibers
convey all sorts of sensory information, especially has been firmly established. Results from eyelid-
the position of the body, and information about conditioning experiments provide a straightforward
motor commands from the cerebral cortex. Thus, example. This simple example of motor learning in-
each combination of motor commands and senso- volves training an animal by repeatedly presenting a
ry input would produce a unique pattern of relatively neutral stimulus like a tone paired with a re-
mossy-fiber activity and an even more complex inforcing stimulus like a puff of air in the eye (see Fig-
pattern of granule-cell activity. These different ure 5A). Initially, the eyelids do not move during the
patterns would allow Purkinje cells to respond tone, but with repeated training, the tone elicits a
differently under various circumstances. learned closure of the eyelids. Lesion, stimulation,
Climbing fibers, in contrast, seem to convey sig- and recording experiments have shown that the tone-
nals indicating that a movement should be differ- activated mossy-fiber inputs to the cerebellum, the
ent from each other. puff activates climbing-fiber inputs, and output from
Marr then suggested that these climbing-fiber a cerebellar deep nucleus is responsible for the ex-
signals would alter the strength of the granule cell pression of the learned responses (see Figure 5B).
Figure 5 Eyelid-conditioning experiments also reveal the tem-

poral specificity or timing component of this cerebel-
lar learning. The eyelids do not close at the onset of
the tone (as they do for the puff); instead, the re-
sponses are delayed to peak at the time the puff is
presented, optimizing the adaptive, protective nature
of the response.
Figure 6 illustrates the sequence of events be-
lieved to occur during cerebellar learningfor exam-
ple, eyelid conditioning. The tone activates a certain
subset of mossy fibers, which, in turn, activate a subset
of the granule cells. All these tone-activated mossy fi-
bers may excite the deep nucleus neurons; eyelid re-
sponses are absent in part because the Purkinje cells
are spontaneously active, and their inhibition of the
nucleus cells prevents a response. Since the puff acti-
vates climbing fibers, the synapses made by the tone-
activated granule cells onto the Purkinje cells are
modifiedthey are made weaker. Eventually, the
Purkinje cells learn to decrease activity during the
tone, which releases the deep nucleus cells from inhi-
bition and contributes to the expression of the condi-
tioned responses. Learning seems to involve changes
in the deep nuclei as well, which also contribute to the
ability of the tone to excite the deep-nucleus neurons
and produce a learned conditioned response.
Although such results are a clear illustration, the
cerebellum did not evolve to mediate eyelid condi-
tioning. Rather, such experiments reveal the basic ca-
pacity of the cerebellum for learning and reveal the
basic properties of this learning. From such revela-
tions the purpose of the cerebellums capacity to learn
The relationship of Pavlovian eyelid conditioning and the
cerebellum. A1. A schematic representation of a typical eyelid
becomes clearer. It is exactly the type of learning re-
conditioning training trial. A neutral stimulus such as a tone is quired to permit sensory input to improve the accura-
paired with a reinforcing stimulus such as a puff of air directed cy of movements. Making accurate movement neces-
at the eye. A2. Example eyelid responses from a conditioning sarily requires input from sensory systems; this input
experiment. For each trace, eyelid closure is indicated by is revealed by the severe motor impairments that re-
upward deflection. Before training presentation of the tone has sult from sensory impairments like large-fiber neu-
no consistent effect on the eyelids. After training the tone elicits ropathies, wherein information about the position of
a robust and well-timed conditioned eyelid closure. A3. Example the body does not reach the brain.
learning curves for eyelid conditioning. Training with the tone
paired with the puff gradually increases the likelihood that the In principle, sensory input can be used in two
tone will elicit a conditioned eyelid response. This process is ways to guide movements. One is via feedback, where-
called acquisition. Extinction of the conditioned responses by a movement command is initiated; then sensory
occurs when training consists of presentation of the tone alone, input is used during the execution of the movement
without the puff. In extinction training, the tone gradually loses
to make the corrections required for proper perfor-
its ability to elicit the conditioned eyelid response. B. A
mance. Feedback can be very effective and requires
schematic representation of the way in which eyelid conditioning
engages the cerebellum. Conditioned stimuli such as a tone are no learning, but it is slow. For this reason feedback
conveyed to the cerebellum via activation of mossy fibers, cannot be used for most of our movements. The other
whereas presentation of the reinforcing unconditioned stimulus alternative is feedforward. In this mode, sensory
is conveyed to the cerebellum via activation of climbing fibers. input is used at the beginning of a movement (when
Output of the cerebellum via one of the deep nuclei (anterior the motor command is issued) to guess what muscle
interpositus in the case of eyelid responses) drives the forces will produce the proper movement. It can be
expression of the conditioned response. schematized in this way: Given this command and the
present sensory input, and based on previous experi-
Figure 6
A schematic representation of the cerebellar learning thought to occur during eyelid conditioning. Each cartoon depicts typical activity
of a Purkinje cell, a deep nucleus cell, and the corresponding eyelid response that would be expected. Also, the strength of the granule
to Purkinje synapses and the mossy fiber to deep nucleus synapses (those activated by the tone) are depicted by the relative size of the
triangles. Before training (left panel), the granule-to-Purkinje synapses are relatively strong, and the mossy fiber-to-nucleus synapses
are relatively weak. Thus, presenting a tone has little effect on the spontaneous activity of the Purkinje cell and provides little
excitatory input to the deep nucleus cells. The result is little change in the activity of the nucleus cell. After acquisition training (center
panel), the tone-activated granule-to-Purkinje synapses are weaker so that presenting the tone causes a decrease or cessation of
Purkinje cell activity. The mossy fiber-to-nucleus synapses may also be strong. The combination of decreased inhibition and increased
excitation increases the deep nucleus cell activity and elicits a conditioned response. After tone-alone extinction training (right panel)
the granule-Purkinje synapses have strengthened, and the spontaneous Purkinje cell activity does not decrease during the tone. Since
the deep nucleus neurons are strongly inhibited during the tone by the Purkinje cells, the conditioned responses are suppressed.
ence, here are the forces that are likely to produce an Chan-Palay, V. (1977). Cerebellar dentate nucleus: Organization, cytolo-
accurate movement. gy and transmitters. New York: Springer-Verlag.
Glickstein, M., Yeo, C., and Stein, J., eds. (1987). Cerebellum and
Notice that based on previous experience neuronal plasticity. New York: Plenum Press.
means learning. Indeed, it means the type of learning Ito, M. (1984). The cerebellum and neural control. New York: Raven
where errors in performance (like those that the Press.
Linden D. J., and Connor, J. A. (1995). Long-term synaptic depres-
climbing fibers convey) make changes so that later,
sion. Annual Review of Neuroscience 18, 31957.
given a similar situation (like those that the mossy fi- Marr, D. (1969). A theory of cerebellar cortex. Journal of Physiology
bers convey), performance will improve. If you reach 202, 437470.
for an object and miss, cerebellar learning makes ad- Mauk, M. D., and Donegan, N. H. (1997). A model of Pavlovian
justments so that the next time your hand naturally eyelid conditioning based on the synaptic organization of the
arrives at the right spot. The cerebellum is constantly cerebellum. Learning and Memory 4, 130158.
Mauk, M. D., Medina, J. F, Nores, W. L. and Ohyama, T. (2000).
using this type of learning to make small adjustments Cerebellar function: Coordination, learning or timing? Cur-
so that our movements remain accurate. We can see, rent Biology 10, R522R525.
then, that the inaccurate and uncoordinated move- Medina, J. F., Garcia, K. S., Nores, W. L., Taylor, N. M., and Mauk,
ments arising from damage to the cerebellum reflect M. D. (2000). Timing mechanisms in the cerebellum: Testing
performance without the benefit of the cerebellums predictions of a large scale computer simulation. Journal of
reservoir of previous experience. Medina, J. F., and Mauk, M. D. (2000). Computer simulation of
cerebellar information processing. Nature Neuroscience 3,
See also: GENETIC SUBSTRATES OF MEMORY: 1,2051,211.
CEREBELLUM; LONG-TERM DEPRESSION IN THE Palay, S. L., and Chan-Palay, V. (1974). Cerebellar cortex: Cytology
CEREBELLUM, HIPPOCAMPUS, AND NEOCORTEX; and organization. New York: Springer-Verlag.
NEURAL COMPUTATION: CEREBELLUM; NEURAL Raymond, J., Lisberger, S.G., and Mauk, M. D. (1996). The cere-
SUBSTRATES OF CLASSICAL CONDITIONING: bellum: A neuronal learning machine? Science 272, 1,126
DISCRETE BEHAVIORAL RESPONSES; VESTIBULO- 1,132.
Thach, W. T., Goodkin, J. P., and Keating, J. G. (1992). Cerebel-
OCULAR REFLEX (VOR) PLASTICITY
lum and the adaptive coordination of movement. Annual Re-
Bibliography
Bloedel, J. R., Dichgans, J., and Precht, W., eds. (1985). Cerebellar W. I. Welker
functions. New York: Springer-Verlag. Revised by Michael Mauk
200 GUIDE TO THE ANATOMY OF THE BRAIN: Cerebral Cortex
Figure 1 ry nor motor, called association regions, which fur-

ther analyze sensory information or combine
information from two or more senses or from several
brain regions. For example, in the frontal lobe there
is a region called the prefrontal cortex, which is im-
portant for planning and keeping track of sequential
tasks such as cooking an elaborate meal or remember-
ing driving directions long enough to get to a destina-
tion. Specialized parts of the prefrontal cortex behind
the eyes are associated with control of social behavior,
sorting appropriate and inappropriate responses de-
manded by various situations (Damasio, 1994; Fuster,
1997).
The cerebral cortex of the human is highly spe-
cialized. For example, one region at the junction of
the frontal and temporal lobes, Brocas area, coordi-
nates the movements of the mouth and tongue to pro-
duce speech (motor speech), while a second region
at the junction of the parietal and temporal lobes,
Wernickes area, is necessary for understanding the
meaning of spoken and written words (sensory
The division of the cortex of the human brain into lobes (thick speech). Humans with damage in Brocas area are
dashed lines) and into the primary sensory and motor areas unable to articulate grammatically correct speech,
(thin dotted lines). The gray lines show the numerous sulci of while those with damage in Wernickes area are capa-
the cortex. ble of fluent, though nonsensical, speech (Geschwind,
1979).
All cortical lobes are further divided into ana-
CEREBRAL CORTEX tomically and physiologically distinct areas. Each cor-
The cerebral cortex is the large overgrowth of the tical area is defined by the unique organization and
mammalian forebrain. It is best developed in pri- physiology of the cells within it and by the set of con-
mates and especially in humans, where it makes up a nections the area has with other parts of the brain.
thin sheet, about 3 mm thick and 1600 cm2 in area, There are nearly one hundred known cortical areas,
folded into intricate convolutions to fit in the skull. which are usually designated by a numbering scheme,
Most of the cortex is buried in the banks and depths or code. For example, the primary area of cortex de-
of elongated crevices called sulci. At a gross level the voted to vision is designated area 17, or V1 (visual
area 1), and the primary area for movement is area
cerebral cortex is divided into four lobes. In the
4, or M1 (see Figure 1).
human brain the frontal lobe is located behind the
forehead and above the eyes; the occipital lobe is The cells of the cortex make up the outer gray
found at the back of the head. Between them are the matter of the brain, and their axons leave the cortex
parietal lobe, near the top of the head, and the tem- and connect with other cortical areas and other parts
poral lobe, along the sides of the head, behind the of the brain. The cortex also receives information
ears (see Figure 1). These lobes are easiest to locate from other brain regions. The white matter just be-
in the brains of humans and other primates, where neath the cortex is made up of axons entering or leav-
the sulci divide the cortex into lobes, and the lobes ing the cortex. These connections render the cortex
into smaller units, called lobules. a massive communication system. For example, there
is two-way communication between the cortex and a
Even at the gross level of lobes, each region of region beneath the cortex, the thalamus (Steriade,
anatomy in the cerebral cortex entails a distinct physi- Jones, and McCormick, 1997). This reciprocal link is
ology. Thus, within the frontal lobes there are motor very precise, so that one area of the cortex communi-
regions devoted to the planning and execution of cates mainly with one or two of the many groups of
movements; within the parietal lobe, regions devoted cells, or nuclei, in the thalamus. In addition, each cor-
to the sense of touch; within the occipital lobe, retical area on one side of the brain has some connec-
gions devoted to vision; and, within the temporal tions with its reciprocal area and a few neighboring
lobe, regions devoted to the sense of hearing (see Fig- areas on the opposite side. The connections to the op-
ure 1). Each lobe also includes regions, neither senso- posite side are made by axons of cortical cells, which
GUIDE TO THE ANATOMY OF THE BRAIN: Cerebral Cortex 201
form bundles called the corpus callosum and the an- Figure 2
terior commissure. The corpus callosum is a major
highway linking most of the cortical areas between the
two brain halves, or hemispheres. The anterior com-
missure is a smaller pathway, connecting mostly tem-
poral areas across the hemispheres.
There are other connections between cortical
areas on the same side of the brain. Each of the pre-
cise connections with the thalamus, opposite cortex,
and cortex on the same side contributes to the unique
physiological characteristics of a cortical area. More
widespread connections that go to all areas of cortex
come from cells beneath the cortex and use the chem-
icals dopamine, norepinephrine, serotonin, acetyl-
choline, or histamine as neurotransmitters. These
widespread connections appear responsible for set-
ting the overall level of activity in the cortex (Foote
and Morrison, 1987; Kandel et al., 2000).
A typical area of the cerebral cortex is divided
horizontally into six layers. The cells in each layer
have similar structure and connections and are segre-
gated from cells in other layers with different proper-
ties (see Figure 2, top left panel). Superimposed on
the horizontal organization is a vertical grouping of
cells across layers, forming cortical columns, or mod-
ules. The key to the cortical column is that the cells
within a single column are interconnected such that Layers and cell types of the cerebral cortex. (Top, left)
all its cells exhibit common physiological properties, Photograph of a thin section through the prefrontal cortex of a
while cells in neighboring columns exhibit different nonhuman primate showing the horizontal division of the cortex
properties. Columns or modules of cells with similar into six layers. This section was stained to show mostly
properties are the fundamental units of organization pyramidal cells in layers 3, 5 and 6; the rest of the neurons in
and function in the cerebral cortex and are found this section are unstained and consequently invisible. (Top,
right) Pyramidal cells from the top left section are magnified to
throughout the cortex, including areas involved in vi-
show their triangular shape. (Bottom panels) Nonpyramidal cells
sion, hearing, movement, or high-order cognitive
in the cortex are a heterogeneous group. The panels show
processes and memory (Mountcastle, 1998). For ex- photographs of nonpyramidal inhibitory neurons in the cortex
ample, in the visual area of the primate cortex, cells (arrows). The scale for the top left panel is 0.5 millimeters (500
that analyze signals from one eye make up one set of micrometers), and at the bottom the scales are 30 micrometers.
columns, each 0.5 to 1.0 millimeter wide, and these
interdigitate with a second set of columns containing
cells that analyze signals from the other eye. Thus, ent shapes and colorful names, such as chandelier
cells in different columns are dominated by one eye cells, basket cells, and double-bouquet cells (see Fig-
or the other (Hubel and Wiesel, 1977). ure 2, bottom panels). Inhibitory cells, which use
The correlation of function with structure pre- gamma-aminobutyric acid (GABA) as a neurotran-
vails among single cells as well. There are two general smitter, are important in controlling the activity of
categories of cells in the cortex. One, called a pyrami- other neurons in the cortex (Somogyi et al, 1998). Ex-
dal cell, has a triangular cell body and is excitatory citatory cells in the cortex use glutamate as a neuro-
(see Figure 2, top panels). Pyramidal cells are the transmitter. From these excitatory and inhibitory
principal source of axons that leave an area of cortex, cells and the interconnections they give rise to physio-
carrying information to other cortical areas or to re- logical properties that are unique to the cerebral cor-
gions outside the cortex. The second category in- tex (Peters and Jones, 1984).
cludes cells with a rounded cell body; called nonpy-
ramidal, they play a role in local communication. Although extraordinarily complex, the cerebral
Nonpyramidal cells are a heterogeneous group cortex has a general plan of organization. The great
and include one or two types of excitatory cells and diversity of function of the cerebral cortex is built up
many types of inhibitory cells. The latter have differ- from the physiology of single cells to progressively
202 GUIDE TO THE ANATOMY OF THE BRAIN: Hippocampus, Parahippocampal Region
larger compartments that link neurons within mod- essay differences in numbers of principal layers have
ules, layers, areas, and lobes. been selected as the major criterion. All fields of the
hippocampus exhibit a characteristic three-layered
Bibliography appearance. The parahippocampal region, in con-
Damasio, A. (1994). Descartess error: Emotion, reason, and the human
trast to the hippocampus, comprises cortical regions,
brain. New York: G. P. Putnams Sons.
Foote, S. L., and Morrison, J. H. (1987). Extrathalamic modulation which show more than three, generally five or six, lay-
of cortical function. Annual Review of Neuroscience 10, 6795. ers. The hippocampus consists of three major subdivi-
Fuster, J. (1997). The prefrontal cortex: Anatomy, physiology, and sions: the dentate gyrus, the Ammons horn (fields
neuropsychology of the frontal lobe. 3rd edition. New York: Lip- CA1, CA2, and CA3), and the subiculum. Note that
pincott, Williams, and Wilkins.
some authors distinguish a prosubiculum from the
Geschwind, N. (1979). Specializations of the human brain. Scientific
American 241, 180199. subiculum proper. Also, readers may find texts in
Hubel, D. H., and Wiesel, T. N. (1977). Ferrier Lecture: Functional which reference is made to the subicular complex,
architecture of macaque monkey visual cortex. Proceedings of consisting of (pro)subiculum, presubiculum, and
the Royal Society of London B 198, 159. parasubiculum. Based on connectional arguments as
Kandel, E. R., Schwartz, J. H., and Jessel, T. M. (2000). Principles
well as the criterion chosen for this discussion (num-
of neural science. New York: McGraw-Hill.
Mountcastle, V. B. (1998). Perceptual neuroscience: The cerebral cortex. ber of layers), it is preferred to consider the presu-
Cambridge: Harvard University Press. biculum and parasubiculum as functionally different
Peters, A., and Jones, E. G., eds. (1984). Cerebral cortex, Vol. 1: Cellu- from the subiculum.
lar components of the cerebral cortex. New York: Plenum.
Somogyi, P., Tamas, G., Lujan, R. and Buhl, E. (1998). Salient fea- In all species the parahippocampal region com-
tures of synaptic organization in the cerebral cortex. Brain Re- prises five main regions: the pre- and parasubiculum,
search Reviews 26, 113135. the perirhinal and entorhinal cortices, and a fifth area
Steriade, M., Jones, E. G., and McCormick, D. A. (1997). Thala-
musorganization and function, Vol. 1: Organization and func-
that in primates is commonly referred to as the par-
tion. Oxford: Elsevier. ahippocampal cortex. In nonprimates, the most likely
homologue for the latter area is the so-called postrhi-
Stewart Hendry
nal cortex. Different researchers have subdivided
Revised by Helen Barbas and Stewart Hendry
each of these five regions in a variety of different ways.
All fields of the (para)hippocampus can be easily
appreciated as longitudinal strips of cortex, neatly
HIPPOCAMPUS AND aligned one next to the other, beginning at the den-
PARAHIPPOCAMPAL REGION tate gyrus on one end and the most outside portion
The hippocampus is part of the cerebral cortex. It is of the parahippocampal region, in effect the perirhi-
closely linked to a restricted number of related corti- nal field, at the other end. The precise orientation
cal areas, which are collectively referred to as the par- and curvature of this entire cortical structure and thus
ahippocampal region. Both the hippocampus and its overall position in the brain may vary between dif-
parahippocampal region are reciprocally connected ferent species. In species with a clearly developed
with a wide variety of higher order association cortices temporal lobe (humans and monkeys, for example),
representing all sensory domains. From this recently the hippocampus is more ventrally and anteriorly lo-
emerged perspective it comes as no surprise that sci- cated; in contrast, the rats hippocampus looks more
entific conjectures about the functions of the like a c-shaped structure positioned in the caudal
(para)hippocampal network emphasize an impor- third of the hemisphere. However, such difference in
tant role in higher order cognitive functions, spe- position does not alter the major characteristics and
cifically learning and memory processes. The the topological relations between the hippocampus
(para)hippocampal structures also have strong func- and the parahippocampal structures. Nor does it in-
tional links with subcortical structures that research- fluence the fact that the most anterior/ventral portion
ers believe play a role in the selection of behavior in of the hippocampus has a close spatial relationship
favor of either survival of the individual or of the spe- with the amygdala.
cies.
Wiring of the (Para)Hippocampus
General Features of the Most detailed information about the connectivity
(Para)Hippocampus of the system comes from anatomical tracing studies
In order to describe any feature of a structure, ei- in the rat, although for some pathways relevant de-
ther morphological or connectional, one needs to tailed information has been collected in other species
have a nomenclature. It is inevitable that opinions as well. Overall, the connectivity is rather conserva-
about nomenclature differ, so for the purposes of this tive, such that a general non-species-specific descrip-
GUIDE TO THE ANATOMY OF THE BRAIN: Hippocampus, Parahippocampal Region 203
tion may suffice. Moreover, this conserved connec- nally described, but also to CA3. Cells in one particu-
tional organization does allow making inferences lar subdivision of the entorhinal cortex, generally re-
concerning the overall organization of the human ferred to as lateral entorhinal cortex, distribute their
(para)hippocampus. The hippocampus has two major axons exclusively to the most distal portions of the
pathways through which is connected to the rest of dendrites of dentate and CA3 cells. The other subdi-
the brain. The first pathway makes use of the parahip- vision, referred to as medial entorhinal cortex, sends
pocampal connectivity and predominantly mediates its projection to the middle portions of the apical
the connections with the cortex. The second pathway, dendrite. This routing results in a specific laminar
which mainly, but not exclusively, links the hippo- termination, such that each neuron receives both
campus to subcortical structures, makes use of the for- pathways but on different segment of its apical den-
nix. drite. The second pathway, which attracted more ex-
perimental attention only in the late twentieth centu-
The Parahippocampal-Cortical System ry, originates from layer III cells and distributes to
The famous Spanish anatomist Ramn y Cajal, CA1 and the subiculum. In contrast to the laminar
who provided one of the most detailed and earliest pattern as described for the lateral and medial layer
descriptions of the (para)hippocampal system, em- II components, axons of layer III cells target only re-
phasized the so-called trisynaptic circuit, comprising stricted groups of the available neurons in CA1 and
an exclusive unidirectional pathway from the entorhi- the subiculum. This targeting results in an organiza-
nal cortex to the distal dendrites of the granular (sim- tion such that the lateral entorhinal cortex dissemi-
ple pyramidal) cells in the dentate gyrus, which, in nates its output only to the neurons clustered around
turn, give rise to the mossy fiber pathway to the proxi- the CA1-subiculum border, whereas the medial en-
mal part of the apical dendrites of CA3 pyramidal torhinal cortex selectively innervates CA1 neurons
cells. These CA3 neurons finally convey their output close to the border with CA2/CA3 and subiculum neu-
by way of the Schaffer axon collaterals to the apical rons, which are close to the border with the presu-
and basal dendrites of pyramidal cells in CA1. The biculum.
trisynatic circuit was once thought to be organized in
restricted planes perpendicular to the longitudinal CA1 neurons project to the subiculum and, by
axis of the hippocampus, such that the structure as a doing so, add a fourth synapse to the originally de-
whole was considered as a series of repetitive circuits, fined trisynaptic circuit. These projections, which are
so-called lamellae, stacked together along the long again almost entirely unidirectional, show an interest-
axis. This proposal has stimulated functional analysis ing topographical organization along the transverse
of the network using isolated brain slices containing axis, which is reminiscent of that of the entorhinal ter-
these lamellar circuits. However, it also emphasized minal distribution. Neurons in CA1, close to the bor-
the relative isolation of the trisynaptic circuitry, which der with the subiculum, are connected to subicular
contrast to the notion that the hippocampal trisynap- neurons, which are similarly close to that border;
tic circuit has to be part of a larger neural system in these two connected populations are thus most likely
order to function. Added details emphasize the over- innervated by entorhinal inputs from the lateral en-
all longitudinal connectivity as an integral feature of torhinal cortex. In contrast, CA1 neurons closer to
the trisynaptic organization. Moreover, the entorhi- the border with CA3 will distribute their axons to sub-
nal cortex is a complex cortical structure in itself that, icular neurons close to the border with the presu-
in addition, has strong reciprocal connections with biculum; these two connected populations thus most
widespread association cortices, largely mediated by likely receive medial entorhinal inputs.
its neighboring parahippocampal fields. Finally, the The lateral and medial entorhinal cortex may
subiculum has been added on a crucial position with-
process functionally different types of information.
in this circuit.
The main cortical input of the lateral entorhinal cor-
Neurons in entorhinal layers II and III, which are tex originates from the perirhinal cortex and olfacto-
the main recipients of the extensive cortical inputs, ry cortices, whereas the medial entorhinal cortex re-
give rise to the already mentioned input to the hippo- ceives strong inputs from the presubiculum as well as
campus, the so-called perforant pathway. This name from the parahippocampal or postrhinal cortex. In
is derived from the traditional descriptions by Ramn view of the aforementioned anatomical organization
y Cajal, who noted that fibers from the entorhinal cor- researchers have suggested that in the hippocampus
tex perforate the pyramidal-cell layer of the sub- these two functionally different input streams con-
iculum to gain access to the dentate gyrus molecular verge at the level of the dentate gyrus and CA3,
layer. The perforant pathway harbors two different whereas they are kept more or less separate at the
projection systems. Layer II cells distribute their level of CA1 and the subiculum. This connectional
axons to most, if not all of the dentate gyrus, as origi- differentiation taken in conjunction with the overall
204 GUIDE TO THE ANATOMY OF THE BRAIN: Neuron
differences in intrinsic wiring between, for example, Conclusion

CA3 and CA1 suggests that the (para)hippocampal The (para)hippocampus may be viewed as a corti-
system harbors two systems that may be related to dif- cal system with bidirectional connections with almost
ferent memory processes. all of the multimodal associational domains of the
cerebral cortex as well as a number of subcortical
Both CA1 and the subiculum constitute the major
structures reportedly involved in motivation and se-
output structures of the hippocampus. They distrib-
lection of appropriate behaviors. Therefore, the
ute strong projections back to the entorhinal cortex,
(para)hippocampus may be at the crossroad of the
mainly to its deep layers. The overall topographical
cognitive and the affective side of behavior. Most
organization of this projection is in register with that functionally oriented research unfortunately address-
of the perforant pathway, such that a particular por- es either of those sides, whereas a combined analysis
tion of the entorhinal cortex, projecting to restricted of both is rare.
populations of neurons in CA1 and the subiculum, re-
ceive a return projection originating from these same
Bibliography
neuronal groups in CA1 and the subiculum. This
Amaral, D. G., and Witter, M. P. (1995). Hippocampal formation.
striking reciprocity, taken in conjunction with the In G. Paxinos, ed., The rat nervous system. San Diego, CA: Aca-
aforementioned overall intrinsic hippocampal orga- demic Press.
nization, may have important functional implications Burwell, R. D. (2000). The parahippocampal region: Corticocorti-
that researchers do not yet fully understand. cal connectivity. Annals of the New York Academy of Sciences 911,
2542.
Ramn y Cajal, S. (1911). Histologie du Systme Nerveux de lHomme
et des Vertebrs. Maloine, Paris
The Fornix Witter, M. P., Wouterlood, F. G., Naber, P. A., and van Haeften,
T. (2000). Anatomical organization of the parahippocampal-
The fornix is a major fiber bundle that connects hippocampal network. Annals of the New York Academy of Sci-
the hippocampus to the hypothalamus, in particular ences 911, 124.
the mammillary bodies. The fornix originates from
Larry W. Swanson
CA1 and subiculum, although parahippocampal cor- Revised by Menno P. Witter
tices, in particular the pre- and parasubiculum and,
to a much lesser extent, the entorhinal cortex, con-
tribute fibers. On its way to the mammillary bodies,
the fornix also issues fibers to the septal complex, the NEURON
ventral striatum, and the amygdala. The fornix also
Cells of the central nervous system are divided into
carries the fibers from CA1 and subiculum targeting
two categories, neurons and glial cells. The following
parts of the prefrontal cortex. The fornix is not a pure entry deals with the characteristics of neurons in the
output pathway since projections from the septal vertebrate central nervous system. Neurons are inde-
complex to the hippocampus and in part to the en- pendent morphological, trophic, and functional enti-
torhinal cortex travel by way of the fornix. These sep- ties; they develop from the neural plate of the ecto-
tal afferents provide the (para)hippocampus with derm. They differ from glial cells in their ability to
most of its cholinergic inputs. The fornix also forms generate propagated action potentials (spikes), in the
one of the input routes for the noradrenergic, sero- release of neuroactive substances called neurotransmit-
tonergic, and dopaminergic innervation. Additional ters, and in their ability to communicate with other
aminergic fibers enter the parahippocampal region cells through specialized membrane junctions called
through a ventral route. Finally, the commissural con- synapses. A long debate in the first half of the twentieth
nections between the left and right hippocampi also century pitted those who maintained that the brain
partially travel by way of the fornix. was a continuous reticulum of fibers against those who
proposed that elements of the nervous system were
The hippocampal connection to the mammillary discrete cells (for a historical account see Peters,
bodies is part of the traditionally described limbic or Palay, and Webster, 1991). The first electron micro-
Papez circuit, which includes the mammillo-thalamic scopic studies decisively resolved the issue by showing
tract connecting the mammillary bodies to the anteri- that each neuron is delineated by a continuous plasma
or complex of the thalamus, which in turn project to membrane and is separated from other cells by a gap.
large portions of the limbic cortex, including anterior However, like other cells of the body, neurons in
and posterior cingulate cortex, and pre- and parasu- some parts of the nervous system are interconnected
biculum. All these structures, in turn, provide input through continuous cytoplasmic bridges organized
to the hippocampus, either directly, or indirectly by into gap junctions that are permeable to ions and small
way of the entorhinal cortex. molecules.
GUIDE TO THE ANATOMY OF THE BRAIN: Neuron 205
Neurons are polarized cells receiving informa- meters); and microfilaments (polymers of actin, diame-
tion at certain locations on their plasma membrane ter five to seven nanometers).
and releasing neurotransmitters to other cells, usually In addition to rough ER, many neurons are rich
from other sites (Kandel and Schwartz, 1985; Peters, in smooth ER that is involved in intracellular Ca2 +
Palay, and Webster, 1991; Shepherd, 1990). They storage and release. Cysternae of the ER are often
emit several processes originating from the cell body or closely aligned with the plasma membrane of both the
soma. One (occasionally two or three) of the processes soma and dendrites forming subsurface cysternae.
is an axon propagating the action potential to the
Lysosomes are found in all neurons. Secondary ly-
transmitter-releasing nerve terminals. The other pro-
sosomes accumulate throughout the life of the cell
cesses are called dendrites and are usually shorter and
and coalesce into lipofuscin granules showing charac-
branch less frequently than the axon. The shape and
teristic distribution for each neuronal type.
three-dimensional distribution of the processes are
characteristic of each category of neuron and reflect
their connections with other cells and the neurons The Dendrites
place in the neuronal network. The general arrange-
The dendrites are rarely longer than one milli-
ment is that information arrives through afferent (i.e.,
meter and can be as short as ten to fifty microns with
inward-transmitting) synapses on the dendritic pro-
a diameter of three to 0.05 microns, tapering toward
cesses and the cell body, and is transmitted to other
their tip and decreasing in diameter at branching
cells through axonal enlargements, also called bou-
points. The main difference between them and the
tons, present on the axonal arborization. However,
axon is that dendrites lack the morphologically dis-
significant exceptions to this rule occur in some parts tinct initial segment at their origin from the soma.
of the brain. Neurons in invertebrates usually have With the exception of peripheral sensory neurons, all
only one process originating from the cell body that neurons have dendrites. They are generally postsyn-
gives rise to branches, all of which both receive and aptic to axon terminals; from scores to tens of thou-
give information and are involved in different opera- sands of synapses converge on the dendritic tree of a
tions. single neuron.
In some parts of the nervous system, most promi-
The Soma nently in the retina (amacrine cells), the olfactory
bulb, the thalamus, the substantia gelatinosa of the
The soma has a diameter of five to fifty microns
brain stem and spinal cord, and the superior col-
and contains the nucleus and the usual cell organelles
liculus, dendrites of some classes of cells can be both
present in most cells, with great similarity to those
pre- and postsynaptic. In these cases the dendrite at
present in secretory cells. This is in line with the ob-
the presynaptic site contains synaptic vesicles and
servation that most neurons secrete proteins and pep- presynaptic membrane specialization as well as near-
tides in addition to small transmitter molecules. For by postsynaptic membrane specializations at synapses
example, the rough endoplasmic reticulum (ER), the received by the neuron. Often these combined pre-
site of protein synthesis, is often highly developed and postsynaptic sites are located on protrusions,
and is organized into parallel lamellae forming large grapelike clusters or gemmules, isolating the forma-
Nissl bodies. The Golgi apparatus is similarly highly de- tions from each other on the same cell, and providing
veloped and often extends into the proximal dendrit- a basis for independent action. Synapses between two
ic processes, which also contain ER and ribosomes. dendrites can be reciprocal, each partner receiving as
The axons are usually devoid of ribosomes and, to- well as giving synapses to the other at closely located
gether with the nerve terminals, lack the ability for sites.
significant protein synthesis. Thus the neuron is also
a polarized biochemical machine where protein and Dendrites have various short postsynaptic extru-
other components synthesized in the cell body are sions, the best-known of them being dendritic spines
transported through the axon to the nerve terminals. (see Figure 1). Spines are particularly prominent and
numerous on cortical pyramidal and spiny stellate
The transport of molecules and organelles be- cells, on Purkinje cells, and on the spiny neurons of
tween the processes and the soma is bidirectional and the neostriatum. Spines frequently contain a special-
is supported by the cytoskeleton, which also main- ized organelle, the spine apparatus, consisting of par-
tains the shape of the processes (Kandel and allel membrane saccules and continuous with the
Schwartz, 1985). The cytoskeleton consists of microtu- smooth ER of the dendritic shaft. The spine appara-
bules (polymers of tubulin dimers, external diameter tus is thought to be involved in Ca2+ sequestration.
twenty-five to twenty-eight nanometers); neurofila- Spines usually receive excitatory synaptic input and
ments (polymers of cytokeratins, diameter ten nano- occasionally an additional inhibitory input. Numer-
Figure 1. Parts of different neurons in the rat hippocampus shown in light microscopic
photographs. The hippocampal formation is involved in memory formation. All cells were marked
by intracellular injection of a marker molecule through a fine glass micropipette. The marker was
transported to all the processes of the living cells. A. Pyramidal cell in the CA1 region emitting
dendrites into two different layers toward the top and bottom of the picture; the two sets of
dendrites sample different inputs coming from different sources. The light band in the middle
contains the cell bodies of many more unmarked pyramidal cells. The rightmost dendrite is shown
at higher magnification (rotated) in B, illustrating the large number of dendritic spines typical of
these cells. C. Another type of cell, the basket cell, which has its cell body and dendrites in the
same layers as the pyramidal cells, has smooth dendrites. Such dendrites have integrative
properties different from spiny dendrites. D-G. Terminal axonal segments of four different
neuronal types, showing the differences in transmitter-releasing terminals (arrows) reflecting the
specializations in synaptic connections. D. Axon collateral of a CA1 pyramidal cell similar to that
shown in A. E. Two terminals of a mossy fiber originating from a granule cell in the dentate gyrus
and making synapses with the apical dendrites of hippocampal pyramidal cells (not marked). F.
Terminals of a cell (P), seen as a pale silhouette. G. The vertically aligned boutons of a chandelier
cell form multiple synaptic contacts with the axon initial segment (not marked) of a pyramidal cell.
Basket and chandelier cells release the inhibitory neurotransmitter GABA, but to different parts of
the same postsynaptic cell. The pyramidal and granule cell terminals release the excitatory amino
acid glutamate. Scales: A, 50 microns; B and C, 10 microns; D-G, 20 microns. (Peter Somogyi)
ous theories have been put forward for the role of other dendrites and nerve terminals through small
spines. Of these, the formation of a biochemical com- cytoplasmic bridges forming gap junctions. Gap junc-
partment, semi-independent from the dendritic shaft tions are sites of electrotonic transmission because
and from other spines, seems the most attractive. The they are permeable to ions and can facilitate the syn-
integrative properties of dendrites are determined by chronization of neurons.
1. their shape; 2. their intrinsic membrane proper-
ties, underlined by the presence and distribution of
different ion channels; and 3. the location of synaptic The Axon
inputs and their relationship to other inputs (for Most neurons have axons; the few exceptions are
more detail, see Shepherd, 1990). retinal amacrine cells and granule cells of the olfacto-
In addition to receiving and sometimes giving ry bulb (Shepherd, 1990). Axons usually originate
synaptic junctions, dendrites can also be connected to from the soma, rarely from a major dendrite, and
Figure 2
Tracing of the processes of a local circuit inhibitory neuron, the chandelier cell, from the hippocampus of the cat. The
terminals of this cell make synapses exclusively with the axon initial segment of pyramidal cells; therefore, the axon is
localized mainly to the layer of pyramidal cell bodies enclosed by broken lines. Each vertically oriented terminal axon
segment targets one initial segment; thus from this partial reconstruction it can be established that this single chandelier cell
makes synapses with at least 320 pyramidal cell dendrites, so the cell has access to all the information pryamidal cells receive.
The cell was visualized by Golgi impregnation.
Figure 3
Tracing of the processes of a pyramidal cell in the visual cortex of the cat. The cell body is in the superficial layer and emits dendrites
reaching a few hundred microns. The main axon descends to the white matter (lower broken line) and proceeds to innervate other
cortical areas. On its way through the gray matter, it gives rise to several axon collaterals traveling for several millimeters within the
same cortical area and addressing groups of cells with local ramifications while ignoring other groups of cells. The axon also richly
supplies neurons in the vicinity of the cell body. Such selective connections enable effective coordination of neurons with similar
properties. Having both a local axon arborization and a distant one gives the neuron both a local circuit role and a role in connecting
different areas of the brain. The call was visualized by the intracellular injection of a marker molecule.
begin with the axon hillock. A specialized trilaminar ment is the presence of interconnected microtubules
inner coat of the membrane, recognizable with the organized into fascicles. In some regions of the brain
electron microscope, identifies the axon initial segment, the axon initial segment can receive numerous syn-
which has the highest density of voltage-sensitive apses. Such synapses are provided by the chandelier
sodium channels. It is thought to be the site of the cell, a specialized inhibitory neuron unique to the cor-
generation of the propagated-action potential. A tex, which makes synapses exclusively on the axon ini-
similar membrane undercoating is found in the tial segment of pyramidal and spiny stellate neurons
axon at nodes of Ranvier between myelin seg- (see Figure 2). Most axons emit several collaterals
ments. The other unique feature of the initial seg- along their course, addressing particular brain areas
or groups of cells in the same brain area (see Figure mon families of vesicles are the small clear vesicles with
3). a diameter of thirty to fifty nanometers and the large
granulated vesicles with a fine electron dense core and
a diameter of about eighty to two hundred nano-
Myelin Sheath meters. Boutons are also rich in mitochondria.
The axons of neurons in the brain can be
The synaptic vesicle-containing varicosities of
myelinated for part or for the whole of their course,
some neurons do not establish morphologically rec-
or can be completely unmyelinated (Peters, Palay,
ognizable synaptic junctions at all of their boutons.
and Webster, 1991). Some types of neurons, such as
This applies in particular to neurons that use mono-
corticospinal cells and Purkinje cells, always have
amines as transmitters.
myelinated axons. The myelin is segmented, and
each segment is formed by the plasma membrane of
an oligodendroglial cell. Segments are interrupted by
Analysis of Neuronal Circuits
nodes of Ranvier, where axon collaterals often origi-
nate. The axons may acquire myelin for part of their Connectivity patterns of morphologically identi-
course as they traverse a particularly heavily myelinat- fied neurons can be traced via the transport of marker
ed part of the brain. Axons may contain synaptic vesi- molecules through the processes (Heimer and Zabor-
cles at nodes of Ranvier, and they may be presynaptic szky, 1989). The active transport in the living cell can
to neighboring dendrites. be exploited by introducing suitable tracers into the
neuron that are carried to the dendritic and axonal
processes (see Figures 1 and 3). Tracer molecules can
Nerve Terminals be introduced directly into the cell or into the sur-
The terminal axon arborizations are characteris- rounding extracellular space from which the cell can
tic of each cell type. The transmitter-releasing sites take them up by an active process. It is also possible
are bulbs or varicose enlargements having a diameter to label the processes of neurons that have been fixed
usually of 0.5 to three microns; they may have a posi- with chemical agents (see Figure 2). The visualization
tion on the end of axon branches as boutons terminaux, of processes makes it possible to identify the connec-
or may be varicosities along the axon forming boutons tions of particular types of neurons in the same area
en passant (see Figure 1DG). Many terminal boutons of the brain or among different brain regions. The
sit on the ends of short stalks branching off from main morphological appearance of neurons reflects their
axon collaterals (see Figure 1D, vertical arrow). Spe- patterns of connections. In many cases synapses from
cialized formations of nerve terminals evolved, such a given source terminate on certain parts of the neu-
as large mossy fiber terminals providing multiple lo- ron because the operation that the given input pro-
calized input to the same target (see Figure 1E), and vides is best carried out in that part of the cell (see
climbing fibers providing multiple synapses distribut- Figures 1EG, and 2). Homologous parts of numer-
ed over the postsynaptic dendritic tree of the same ous postsynaptic cells in a given area of the brain tend
target dendrite or cell. Boutons form synaptic junc- to align with inputs arriving at that part of the cell,
tions. Boutons are usually only presynaptic to other and this leads to the development of laminated struc-
cells, but terminals of a few cell types, most promi- tures. For example, axons originating from the CA3
nently the primary sensory afterents in the brain stem region of the hippocampus terminate mainly on the
and spinal cord, may receive synapses and are post- main apical dendrites of pyramidal cells in the CA1 re-
synaptic to inhibitory terminals. One bouton may gion, and the local recurrent collaterals of pyramidal
make synaptic junctions with only one postsynaptic cells in the CA1 sector address the basal dendritic re-
element or may provide input to up to about ten dif- gion of the pyramidal cells (see Figure 1A). The sepa-
ferent postsynaptic targets originating from about the ration of inputs and the minimum amount of axon
same number of individual neurons. In some cases al- necessary to achieve addressing is ensured by the
most every bouton of the same axon establishes syn- alignment of pyramidal cells.
apses with a different cell, providing a large degree The terminals of some neurons are all localized
of divergence in information transfer. Cortical cells, in the same brain area where the cell body is located.
for example, may make synapses with thousands of These cells play a role in the local processing of infor-
other cortical neurons in a given area (see Figure 3). mation and are called local circuit neurons (see Figure
The boutons contain synaptic vesicles, which are 2). Other cells connect different brain regions or sup-
membrane-delineated discrete structures. The mor- ply the periphery with their axons; these are called
phology of the vesicles is characteristic of cell types projection neurons. Many projection neurons also have
and to some degree correlates with the chemistry of axon collaterals within the same brain area where the
their neurotransmitter content. The two most com- cell is located, and thus play both a local circuit and
210 GUIDE TO THE ANATOMY OF THE BRAIN: Olfactory Cortex
a projection role (see Figure 3). Accurate knowledge drites of relay neurons, the large mitral cells and
of the connectivity, especially in quantitative terms, is smaller tufted cells. These cells interact with interneu-
a prerequisite of establishing the operations taking rons in the olfactory bulb and send the processed in-
place in real neural networks. formation by means of impulse discharges in their
axons in the lateral olfactory tract (LOT) on the ven-
See also: NEUROTRANSMITTER SYSTEMS AND trolateral surface of the brain. The axons give rise to
MEMORY
numerous collaterals, which terminate in the third
Bibliography main region, the olfactory cortex, to make synapses
Heimer, L., and Zaborszky, L., eds. (1989). Neuroanatomical tract- on the dendrites of cortical pyramidal neurons.
tracing methods, Vol. 2: Recent progress. New York: Plenum
Press.
The olfactory cortex is the area of cortex in the
Kandel, E. R., and Schwartz, J. H., eds. (1985). Principles of neural vertebrate forebrain that receives direct input form
science, 2nd edition, New York: Elsevier. the olfactory bulb. In most mammals there are several
Kisvarday, Z. F., Martin, K. A. C., Freund, T. F., Magloczky, Z., areas of this cortex: The main region is the piriform
Whitteridge, D., and Somogyi, P. (1986). Synaptic targets of
cortex (PC) (meaning pear-shapedit was origi-
HRP-filled layer III pyramidal cells in the cat striate cortex.
Experimental Brain Research 64, 541552. nally termed prepyriform and sometimes spelled
Peters, A., Palay, S. L., and Webster, H. deF., eds. (1991). The fine pyriform). In most mammalian brains it extends
structure of the nervous system, neurons and their supporting cells, over much of the ventrolateral surface of the brain
3rd edition. New York: Oxford University Press. dorsal to the LOT, as shown in Figure 2. The piriform
Shepherd, G. M., ed. (1990). The synaptic organization of the brain,
cortex sends its output axons to several areas. One
3rd edition. New York: Oxford University Press.
Somogyi, P., Freund, T. F., Hodgson, A. J., Somogyi, J., Beroukas, target area is the mediodorsal nucleus of the thala-
D., and Chubb, I. W. (1985). Identified axo-axonic cells are mus, which has connections to the prefrontal areas of
immunoreactive for GABA in the hippocampus and visual the neocortex. This pathway is believed to be involved
cortex of the cat. Brain Research 332, 143149. in conscious perception of odors. Depending on the
Peter Somogyi species, the piriform cortex also sends fibers to other
cortical areas, such as the insula, where odor informa-
tion may be combined with taste information to give
OLFACTORY CORTEX the overall perception of flavor. Subcortical connec-
tions are made to parts of the limbic system, including
The olfactory cortex is the only part of the vertebrate the hypothalamus, hippocampus, and basal ganglia.
forebrain to receive a direct sensory input. Present in
even the most primitive fish, it retains its place and The other olfactory cortical areas include the fol-
form throughout the vertebrate series, suggesting lowing: The anterior olfactory nucleus (AON) is a
that it is a core element in the basic plan of the verte- major station for integrating activity of the olfactory
brate forebrain. Since olfaction is the dominant sen- bulb with that of olfactory cortical areas on both sides
sory modality in most vertebrate species, an under- of the brain via connections through the anterior
standing of olfactory cortical mechanisms can yield commissure. The olfactory tubercle (OT) in rodents
insight into basic behavioral patterns underlying lies medial to the LOT; it is notable for receiving a
much of mammalian and primate behavior. The ol- heavy input of dopaminergic axons from the mid-
factory system is also one of the first sensory systems brain. Because of the implication of dopamine sys-
to differentiate and become functional during fetal tems in various types of mental disorders (depression,
life. sleep disturbances, schizophrenia), the olfactory tu-
bercle has been studied intensively in rodents for its
possible role in these types of disorders. Another tar-
Overall Structure get for olfactory input is the amygdala. The accessory
An understanding of the olfactory cortex requires olfactory bulb (AOB) receives input from the
a clear view of its place in the olfactory pathway. This vomeronasal organ in the nose about chemical signals
pathway and its constituent neurons were first re- involved in mating in many vertebrate species and
vealed by the use of the Golgi stain in the later part projects to the corticomedial nuclei within the amyg-
of the nineteenth century. The pathway consists of dalar complex; these, in turn, project to the hypothal-
three main parts (see Figure 1). First is the olfactory amus, where they presumably activate some of the be-
sensory epithelium in the nose, containing the olfac- havioral patterns in mating. Finally, there is the
tory sensory neurons, which transduce the stimulat- lateral entorhinal cortex (LEC), a major region for
ing odor molecules into impulses that are sent over multimodal integration of olfactory, visual, auditory,
their axons in the olfactory nerve to the olfactory and somatosensory inputs; the output of this region
bulb, the second main structure of the olfactory path- is carried in fibers of the perforant pathway to the
way. Here the axons extend synapses onto the den- hippocampus, which is a critical region for storage
GUIDE TO THE ANATOMY OF THE BRAIN: Olfactory Cortex 211
Figure 1
The main types of neurons and their connections in the olfactory pathway in the mammal.
and retrieval of information of behavioral signifi- ters the brain. The lateral root goes to a cortical area
cance. at the junction of the frontal and temporal cortexes,
which seems to be the homologue of the piriform cor-
tex. From here there are connections to the prefron-
Human Olfactory Cortical Areas tal cortex, called the orbital cortex because it is on the
The regions of the rodent brain described above surface of the brain facing the orbit of the eye. The
are similar in the brains of other mammals and some medial root dives into a small area that, being pock-
lower primates; in higher primates, including hu- marked with many penetrating blood vessels, is called
mans, there are a number of differences. In the mon- the perforated substance, a homologue of the olfacto-
key brain, the olfactory pathway appears much re- ry tubercle. The medial root includes fibers that proj-
duced in relative terms compared with the large ect toward or into the hypothalamus, from which
forebrain, seemingly reflecting the reduced impor- there is a projection to the orbital cortex complemen-
tance of the sense of smell. However, absolute num- tary to that from the piriform cortex. There are thus
bers of neurons or fibers are not a reliable guide to several routes over which information can be pro-
the behavioral importance of a particular system; for cessed, through different olfactory cortical regions, to
example, the LOT contains many more fibers than both cortical and subcortical regions.
the auditory nerve, yet no one would say that hearing
is unimportant in human life. The Basic Cortical Circuit
As a result of the overgrowth of the forebrain in The main regions and pathways described above
primates, the olfactory pathway is limited to the ven- are like cities and motor routes on a map. An under-
tral surface of the brain. The olfactory bulbs give rise standing of the basis of information processing re-
to a long LOT, which divides into three roots as it en- quires an analysis of the structures within each region.
212 GUIDE TO THE ANATOMY OF THE BRAIN: Olfactory Cortex
Figure 2 functions of a given region. For the olfactory cortex,

the main input elements are the fibers from the LOT,
which make excitatory synapses on the spines of the
distal dendrites of the pyramidal neurons. The main
output elements are the pyramidal neurons, each
consisting of apical and basal dendrites, receiving ex-
citatory synapses on their dendritic spines and inhibi-
tory synapses on their dendritic shafts and cell bodies.
The main intrinsic elements are two types of inter-
neurons that make the inhibitory connections onto
the pyramidal neurons. There are two main types of
intrinsic circuits: a reexcitatory feedback circuit
through long axon collaterals of the pyramidal neu-
rons and inhibitory circuits for feedforward and feed-
back inhibition of the pyramidal neurons.
This basic circuit not only summarizes the main
anatomical circuits within the piriform cortex but
also, with minor variations, applies to the other olfac-
tory cortical areas; moreover, it is similar to the other
circuits: those of the hippocampus; the neocortex,
particularly its superficial layers; and the canonical
circuit proposed by some researchers for the neo-
cortex. Thus, the basic circuit for the olfactory cortex
is a useful model for correlating the general proper-
ties of all types of forebrain cortices.
Researchers seek to characterize the functional
properties of each type of cortical element and synap-
Schematic drawing of the ventral surface of the rat brain tic circuit and to identify the stages of expression of
showing components of the olfactory system. Receptor cells
different circuit components and types of excitatory,
activated in the olfactory epithelium (1) project into the
inhibitory, and modulatory neurotransmitters.
olfactory bulb (OB, 2). Principal neurons of the olfactory bulb
project along the lateral olfactory tract (lot) into the piriform
Through such studies they seek correlations with the
cortex (PC, 3). Pyramidal cells of the piriform cortex project to developing behavior of the fetus and the neonate.
entorhinal cortex (EC, 4). Other structures receiving olfactory Studies of plasticity have indicated that the properties
projections include the anterior olfactory nucleus (AON) and the of the excitatory synapses made by the sensory input
olfactory tubercle (OT). Cholinergic innervation of these fibers from the LOT are different from those of the
structures arises from the horizontal limb of the diagonal band intrinsic reexcitatory collateral system; the latter give
(HDB). Rs = rhinal sulcus. evidence of n-methyl-D-aspartate (NMDA) receptors
and are thus strong candidates for mediating the
long-term potentiation in the piriform cortex. Com-
A traditional way to characterize the organization of putational networks based on the basic circuit model
the olfactory cortex is by its layers. The olfactory cor- suggest that sensory inputs and reexcitatory feedback
tex is the prototypical three-layer cortex, consisting interact over extensive areas of the olfactory cortex to
of an outer molecular layer of incoming fibers and provide a distributed parallel system; such analyses
apical dendrites; a middle layer of pyramidal cell bo- shed light on the mechanisms of odor discrimination
dies; and an inner plexiform or polymorphic layer of and may constitute a model for related types of pat-
basal dendrites, fibers, and interneurons. It shares tern recognition by other cortical systems. In particu-
this three-layer organization with the hippocampus lar, the excitatory feedback between pyramidal cells
and contrasts with the six-layer construction of the within the piriform cortex may mediate autoassocia-
neocortex. tive memory function. During initial encoding, syn-
aptic modification of these recurrent connections in
Researchers have clarified the synaptic circuits of the piriform cortex could store associations between
these cortical regions. A basic circuit is made up of the neurons activated by a given odor stimulus. Subse-
minimum types of input fibers, output neurons, in- quent encounters with similar stimuli could activate a
trinsic neurons, and their connections sufficient to subset of these neurons, and activity could spread
represent the most important information-processing across the modified connections to complete the pre-
GUIDE TO THE ANATOMY OF THE BRAIN: Perirhinal Cortex 213
viously encoded pattern of activity. This pattern com- parasubiculum. This basic circuitry, however, is aug-
pletion could serve as a retrieval of the previously en- mented by additional pathways. Each parahippocam-
coded memory for the odor. Further knowledge of pal structure, including the perirhinal cortex, is inter-
the structural organization of the olfactory cortex connected with at least two hippocampal-formation
should provide a better basis for understanding these structures. Moreover, all parts of the parahippocam-
and other important aspects of olfactory function and pal region are highly interconnected with cortical and
behavior. subcortical regions.
Bibliography
Haberly, L. B. (1990). Olfactory cortex. In G. M. Shepherd, ed., Cortical Connections
The synaptic organization of the brain. New York: Oxford Univer-
sity Press. The cortical afferentation of the perirhinal cortex
Ramn y Cajal, S. (1894). Les nouvelles ides sur la structure du systme is consistent with a role in higher-level visual percep-
nerveux chez lhomme et chez les vertbrs. Paris: Reinwald. tion, such as the identification of objects based on
Shepherd, G. M. (1989). A basic circuit for cortical organization.
complex features such as size, shape, color, and pat-
In M. C. Gazzaniga, ed., Perspectives on memory research. Cam-
bridge, MA: MIT Press. tern. Most of its input arises from sensory association-
Shepherd, G. M., and Stewart, W. B. (1985). The chemical senses: al regions. In primates the input is weighted toward
Taste and smell. In M. Swash and C. Kennard, eds., Scientific higher-order visual areas. The perirhinal cortex also
basis of clinical neurology. London: Churchill Livingstone. receives substantial input from polymodal associa-
Gordon M. Shepherd tional regions; a large proportion arises from its
Revised by Michael E. Hasselmo neighbor, the parahippocampal cortex, and other
temporal regions. The rest arises roughly equally
from prefrontal, frontal, and insular regions. In the
PERIRHINAL CORTEX AND ASSOCIATED primate brain there is no input to perirhinal cortex
CORTICAL AREAS from the retrosplenial cortex, a structure that may
play a role in visuospatial functions. The connectivity
Research on learning and memory has confirmed of the parahippocampal cortex suggests a role in
that the structures within the hippocampal formation, visuo-spatial functions. A large input to the parahip-
including the dentate gyrus, Ammons horn, and the pocampal cortex arises in the cingulate and retros-
subiculum play an important role in certain kinds of plenial cortices, and the projections are reciprocal.
memory. Memory functions are also subserved by There is also substantial input from the posterior pa-
nearby cortical regions that are closely intercon- rietal cortex, a region that seems to contribute to vi-
nected with the hippocampal formation. These re- suospatial attention. The parahippocampal cortex is
gions, called the parahippocampal region, include not solely a visual region, however; one of its subdivi-
the perirhinal, parahippocampal, and entorhinal cor- sions is connected with unimodal auditory regions.
tices together with the parasubiculum and the presu- The region is also connected with the orbital frontal
biculum (The postrhinal cortex in the rat brain is cortex.
analogous to the primate parahippocampal cortex
[Burwell and Amaral, 1998a]). Of these regions, the The perirhinal and parahippocampal cortices
perirhinal cortex has received the most emphasis provide the major cortical input to the entorhinal cor-
from researchers on memory. However, the individu- tex. The connections are organized such that the per-
al structures within the parahippocampal region are irhinal cortex projects to rostral portions of the en-
so strongly interconnected with one another and with torhinal cortex and the parahippocampal cortex
the hippocampal structures that it is difficult to dis- projects to caudal portions. The entorhinal cortex
cuss one without addressing the others. With interest also receives direct cortical input arising in the piri-
in all of these regions on the rise, it is likely that future form cortex, frontal and prefrontal regions, the insu-
research will identify roles for each in memory func- lar cortex, the inferior and temporal gyrus, and the
tions. retrosplenial cortex.
In the basic scheme of cortico-hippocampal cir- The presubiculum and parasubiculum are recip-
cuitry, highly processed cortical input reaches the en- rocally connected with the entorhinal cortex, but they
torhinal cortex though projections from the perirhi- are also interconnected with several neocortical re-
nal and parahippocampal cortices. The entorhinal gions. Input to the presubiculum includes the dorso-
cortex projects primarily to the dentate gyrus, the lateral prefrontal cortex, superior and inferior tem-
input structure of the hippocampal formation. The poral gyrus, cingulate and retrosplenial cortex, and
output structure of the hippocampal formation, the parietal cortex. The parasubiculum appears to re-
subiculum, projects back to the entorhinal cortex di- ceive input from the inferior temporal gyrus and dor-
rectly and indirectly through the presubiculum and solateral prefrontal cortex.
214 GUIDE TO THE ANATOMY OF THE BRAIN: Perirhinal Cortex
Subcortical Connections Hippocampal Connections

With the advent of modern neuroanatomical tract The entorhinal cortex provides the major input
tracing methods, subcortical connectivity has become to the hippocampal formation through the perforant
particularly useful in the attempt to understand brain pathway. To summarize the projection, all regions of
functions. Perirhinal subcortical connections are ex- the entorhinal cortex project to the outer two-thirds
tensive and include structures within the basal gan- of the molecular layer of the dentate gyrus. The re-
glia, claustrum, and thalamus, including the pulvinar gion also projects to fields CA3, CA2, and CA1 of the
and mediaodorsal nucleus, basal forebrain, and hippocampus proper (Ammons horn) and to the sub-
amygdala. There are extensive and sometimes recip- iculum. The projections to the dentate gyrus, CA3,
rocal perirhinal conections with basal-forebrain struc- and CA2 originate in layer II of the entorhinal cortex,
tures, including the medial septum, the diagonal whereas the projections to CA1 and the subiculum
band of Broca, and the substantia innominata. originate in layer III. The entorhinal projections to
The perirhinal connections with the amygdala the CA1 and subiculum exhibit a certain topography
have received much interest because of a possible role such that the rostral and caudal entorhinal cortex
in emotional memory. The perirhinal cortex is most project to different parts of CA1 and the subiculum.
strongly interconnected with the deep amygdaloid
nuclei, including the lateral, basal, and accessory The dentate gyrus and field CA3 do not receive
basal nuclei. The perirhinal cortex connections with direct cortical input from regions other than the en-
subdivisions of the amygdala are stronger than those torhinal cortex. However, the situation is different for
with any other region in temporal cortex, including CA1 and the subiculum. Both receive input from en-
the parahippocampal cortex. torhinal cortex, but they also receive direct input
The parahippocampal cortex, like the perirhinal from the perirhinal and parahippocampal cortices.
cortex, is widely connected with subcortical struc- The perirhinal and parahippocampal cortices do not
tures. These include the basal ganglia, particularly provide any direct input to the dentate gyrus. The
the head, body, and tail of the caudate nucleus, the subiculum projects to the entorhinal cortex directly
nucleus accumbens, and the ventral putamen; septal and indirectly through the presubiculum and the
regions, including the medial septum, the diagonal parasubiculum. It appears that at least the presu-
band of Broca, and the substantia innominata; and biculum-entorhinal projection is reciprocal, but there
several thalamic nuclei, including the medial pulvi- is no conclusive evidence for that hypothesis.
nar, the medial dorsal, and the lateral dorsal. The
parahippocampal cortex is also connected with the
amydala, particularly the basal nucleus. Conclusion
Not to be overshadowed, the entorhinal cortex Highly processed sensory and sensory associa-
also receives direct input from multiple subcortical tional input terminates in the perirhinal and parahip-
regions, including the basal forebrain, claustrum, pocampal cortices. Further processing occurs in these
amygdala, thalamus, hypothalamus, and brain stem. regions, and they, in turn, provide the primary corti-
The input from the claustrum, especially the ventral, cal input to the entorhinal cortex. Via the perforant
visual portion is relatively strong. Most of the subcor- pathway, the entorhinal cortex projects mainly to the
tical structures that project to the entorhinal cortex dentate gyrus of the hippocampal formation. The
also project to fields of the hippocampal formation. output structure of the hippocampus, the subiculum,
The entorhinal cortex projects to the lateral, is reciprocally connected with the entorhinal cortex.
basal, and accessory basal nuclei of the amygdala and The subiculum also projects indirectly to the entorhi-
the periramygdaloid cortex. All parts of the entorhinal cortex through the presubiculum and parasu-
nal cortex receive amygdalar input, but the rostral biculum. Although a primary source of cortical input
subfields, those most strongly interconnected with the to the hippocampal formation is from the perirhinal
perirhinal cortex, are more strongly interconnected cortex through an entorhinal cortical projection, an
with the amygdala than the caudal subfields. The important principle of cortico-hippocampal circuitry
major thalamic input arises in midline nuclei, particu- is that all other parahippocampal regions also project
larly the nucleus reunions. directly to one or more structures in the hippocampal
The presubiculum and parasubiculum also re- formation. Moreover, all structures within the par-
ceive direct input from the lateral, basal, and ba- ahippocampal have extensive cortical and subcortical
somedial nuclei of the amygdala. Other subcortical connections. Thus the connections between the par-
connections include the laterodorsal thalamic nucle- ahippocampal region and the hippocampal forma-
us, the midline thalamic nuclei, the ventral tegmental tion include multiple parallel and redundant path-
area, and the dorsal raphe nucleus. ways.
GUIDE TO THE ANATOMY OF THE BRAIN: Synapse 215
See also: GUIDE TO THE ANATOMY OF THE BRAIN: and subicular complex. Journal of Comparative Neurology 307,
CEREBRAL CORTEX; GUIDE TO THE ANATOMY 437459.
OF THE BRAIN: HIPPOCAMPUS AND Witter, M. P., Groenewegen, H. J., Lopes de Silva, F. H., and Loh-
PARAHIPPOCAMPAL REGION man, A. H. M. (1989). Functional organization of the extrinsic
and intrinsic circuitry of the parahippocampal region. Pro-
Bibliography gressive Neurobiology 33, 161253.
Witter, M. P., Naber, P. A., and Lopes da Silva, F. (1999). Perirhi-
Aggleton, J. P., Vann, S. D., Oswald, C. J., and Good, M. (2000).
nal cortex does not project to the dentate gyrus. Hippocampus
Identifying cortical inputs to the rat hippocampus that sub-
9, 605606.
serve allocentric spatial processes: A simple problem with a
Witter, M. P., Naber, P. A., van Haeften, T., Machielsen, W. C.,
complex answer. Hippocampus 10 (4), 466474.
Rombouts, S. A., Barkhof, F., Scheltens, P., and Lopes da
Amaral, D. G., and Witter, M. P. (1995). Hippocampal formation.
Silva, F. H. (2000). Cortico-hippocampal communication by
In G. Paxinos, ed., The rat nervous system,. 2nd edition. San
way of parallel parahippocampal-subicular pathways. Hippo-
Diego: Academic Press.
campus 10, 398410.
Baleydier, C., and Mauguiere, F. (1985). Anatomical evidence for
medial pulvinar connections with the posterior cingulate cor- Rebecca D. Burwell
tex, the retrosplenial area, and the posterior parahippocam-
pal gyrus in monkeys. Journal of Comparative Neurology 232,
219228.
Burwell, R. D. (2001). The perirhinal and postrhinal cortices of the SYNAPSE
rat: Borders and cytoarchitecture. Journal of Comparative Neu-
rology 437, 1741. Neurons give rise to processes known as dendrites
Burwell, R. D., and Amaral, D. G. (1998a). Cortical afferents of the and axons that form a complex network of intercon-
perirhinal, postrhinal, and entorhinal cortices. Journal of Com- nections throughout the brain at specialized sites
parative Neurology 398, 179205.
(1998b). The perirhinal and postrhinal cortices of the rat: called synapses. Around the turn of the nineteenth
Interconnectivity and connections with the entorhinal cortex. century, the Italian physician Camillo Golgi (1873)
Journal of Comparative Neurology 391, 293321. developed a silver staining technique that revealed
Insausti, R., Amaral, D. G., and Cowan, M. W. (1987a). The en- the full extent of dendritic and axonal arbors. From
torhinal cortex of the monkey: II. Cortical afferents. Journal these images, he proposed the reticular theory sug-
of Comparative Neurology 264, 356395.
(1987b). The entorhinal cortex of the monkey: III. Subcor- gesting that the neurons are not discrete cells, but in-
tical afferents. Journal of Comparative Neurology 264, 396408. stead are continuous with each other and form a syn-
Naber, P. A., Witter, M. P., and Lopez da Silva, F. H. (1999). Per- cytium. The renowned Spanish neuroscientist
irhinal cortex input to the hippocampus in the rat: Evidence Santiago Ramn y Cajal (1891) used Golgis method
for parallel pathways, both direct and indirect. A combined to reveal individual neurons in the brain and spinal
physiological and anatomical study. European Journal of Neuro-
science 11, 4,1194,133. cord during development and after a variety of exper-
(2001). Evidence for a direct projection from the postrhinal imental manipulations. These experiments led Cajal
cortex to the subiculum in the rat. Hippocampus 11, 105117. to conclude that neurons were discrete cells with
Pikkarainen, M., and Pitkanen, A. (2001). Projections from the lat- axons ending on dendrites of different cells, a theory
eral, basal, and accessory basal nuclei of the amygdala to the referred to as the neuron doctrine. The term syn-
perirhinal and postrhinal cortices in rat. Cerebral Cortex 11,
1,0641,082.
apse is derived from the Greek word meaning to
Saunders, R. C., and Rosene, D. L. (1988). A comparison of the ef- clasp and was first used by Sir Charles Sherrington
ferents of the amygdala and the hippocampal formation in (1897) in reference to Cajals findings to indicate the
the rhesus monkey: I. Convergence in the entorhinal, prorhi- point of contact between the axons and dendrites.
nal, and perirhinal cortices. Journal of Comparative Neurology Electron microscopy resolved the controversy in the
271, 153184.
Shi, C. J., and Cassell, M. D. (1999). Perirhinal cortex projections
1950s by showing that a gap, the synaptic cleft, which
to the amygdaloid complex and hippocampal formation in is about twenty nanometers wide, separates the pre-
the rat. Journal of Comparative Neurology 406, 299328. synaptic axon from the postsynaptic dendrite, in
Stefanacci, L., Suzuki, W. A., and Amaral, D. G. (1996). Organiza- favor of the neuron doctrine (Peters et al., 1991;
tion of connections between the amygdaloid complex and the Cowan et al., 2001). Scientists widely acknowledge
perirhinal and parahippocampal cortices in macaque mon-
keys. Journal of Comparative Neurology 375, 552582. that communication between neurons via signals at
Suzuki, W. A., and Amaral, D. G. (1990). Cortical inputs to the CA1 synapses provides an important cellular basis for per-
field of the monkey hippocampus originate from the perirhi- ception, behavior, learning, and memory.
nal and parahippocampal cortex but not from area TE. Neuro-
science Letters 115, 4348. Synapses are characterized presynaptically by an
van Groen, T., and Wyss, J. M. (1990). The connections of presu- axon that contains membrane bound vesicles and
biculum and parasubiculum in the rat. Brain Research 518, postsynaptically by a thickening in the membrane re-
227243. ferred to as the postsynaptic density (see Figure 1).
van Hoesen, G. W., Rosene, D. L., and Mesulam, M. M. (1979).
Excitatory synapses contain round vesicles in the pre-
Subicular input from temporal cortex in the rhesus monkey.
Science 205, 608610. synaptic bouton (swelling in the axon) and an asym-
Witter, M. P., and Amaral, D. G. (1991). Entorhinal cortex of the metric postsynaptic density along the inside of the
monkey: V. Projections to the dentate gyrus, hippocampas, postsynaptic membrane (see Figure 1a, b). Depend-
216 GUIDE TO THE ANATOMY OF THE BRAIN: Synapse
ing on the particular synapse, the presynaptic bouton neuron through the channels associated with the
may contain as few as three or as many as 5,000 or NMDA receptor. A variety of calcium-dependent sig-
more of the synaptic vesicles each having diameters naling molecules are tethered in the postsynaptic
of approximately twenty to sixty nanometers. The membrane and comprise the heavily stained postsyn-
other primary organelles in the presynaptic bouton aptic density. Depending on the strength of synaptic
include mitochondria (for production of ATP); activation and the amount of calcium entering the
smooth endoplasmic reticulum (to store and release dendrite, the signaling molecules will trigger second
calcium); and organelles involved in the recycling of messenger cascades that result in temporary changes
the vesicular membrane called endosomes. The post- in synaptic strength through activation of phospho-
synaptic dendrite also contains these organelles in the rylation events, or more permanent changes in synap-
vicinity of synapses. The internal structure is held to- tic strength via the synthesis and insertion of more
gether by cytoskeletal elements composed primarily synaptic proteins thereby enlarging the synapse.
of actin filaments and their linking proteins, and a Inhibitory synapses are distinguished morpho-
few microtubules that provide transport of molecules logically from excitatory synapses by the presence of
and organelles to and from the cell body. The pre- symmetric pre- and postsynaptic thickenings that are
and postsynaptic elements are held together by adhe- not as wide as the postsynaptic density of excitatory
sion molecules that span the synaptic cleft. synapses (see Figure 1c). The presynaptic vesicles of
The presynaptic vesicles of excitatory synapses inhibitory synapses are typically smaller and are less
contain the neurotransmitter glutamate, visualized by uniform in shape, often appearing somewhat flat-
antibody labeling of the molecule in the vesicles, or tened. These synapses contain glycine or GABA as
special electrodes that detect glutamate released from their primary neurotransmitter, and when they are
the synapses. When an action potential arrives at the activated they typically reduce the level of activity at
synapse, one or more vesicles fuse with the presynap- the neuron. These synapses also have a specific set of
tic membrane, and the contents of the vesicle are re- signaling molecules tethered to their densities, such
leased into the synaptic cleft. Glutamate released in that depending on the degree of activation their
this way activates receptors located in the postsynap- strength can also be modified.
tic membrane, resulting in a variety of signals to the In addition to the excitatory and inhibitory syn-
postsynaptic neuron. Activation of the AMPAergic apses some synapses are modulatory, containing
glutamate receptors causes the postsynaptic mem- neuropeptides and hormones that act to modify the
brane to become depolarized. This relieves the mag- strength of the excitatory or inhibitory synaptic input.
nesium blockade of the calcium channel associated These neuromodulatory substances, along with
with a second important glutamate receptor, the growth factors, are often found co-localized in the
NMDA receptor. When the NMDA receptor is both same presynaptic boutons that form excitatory or in-
activated by glutamate and the postsynaptic mem- hibitory synapses. Their release also depends on the
brane is depolarized, calcium enters the postsynaptic rate or intensity of presynaptic activation. In addition,
GUIDE TO THE ANATOMY OF THE BRAIN: Synapse 217
there are electrical synapses (often called gap junc- at a specialized site called the axonal hillock and trav-
tions) that form very close appositions between neu- els millimeters to meters to make synapses with differ-
rons such that channels in the adjacent membranes ent neurons. Therefore, inhibitory synapses located
are aligned and current passes directly between the between the excitatory inputs on the dendrites and
cells, without the need for a chemical neurotransmit- the output axon can dampen the total amount of exci-
ter. If one of the two neurons that are coupled via tation and prevent the neuron from firing too rapidly.
electrical synapses is depolarized by chemical synap- In this way, the inhibitory synapses serve to put the
tic transmission at other synapses along its dendrites, brakes on rapidly firing neurons before their activi-
then the current generated will also depolarize the ty can spread to other cells and cause a seizure.
coupled cell. Neuromodulatory synapses are also widely distribut-
ed along the dendrites and cell bodies, but usually at
The third important partner at a synapse is the
a lower frequency than the excitatory and inhibitory
glial cell process (see Figure 1b). Astrocytic glia form
inputs. Gap junctions also occur at a lower density
long processes that end in numerous tiny projections. and it remains to be determined whether all neurons
Some of these processes form end feet on the blood are coupled in this way to one or more other neurons
capillaries of the brain and through these end feet or the surrounding glia.
pass glucose and other substances from the blood.
The glia store the glucose in the form of glycogen and The remainder of this section focuses on the
provide the neuron with glucose or lactate as a form structure and function of dendritic spines because
of energy to make ATP. Neurons usually do not store many scientists agree that the dendritic spines are
their own glucose, and hence are dependent on the crucial elements for learning and memory. Dendritic
glia for this metabolic energy. Glia also extend tiny spines vary greatly in their structure. Some spines are
processes to surround or partially surround synapses. short and stubby. Other spines, called thin or sessile
In some brain regions, such as the cerebellar cortex, spines, are short or long with a constricted neck and
nearly all of the synapses are completely ensheathed a slightly enlarged head. Yet other spines have a con-
with the astrocytic glia. In other brain regions, such stricted neck with a very large head and are referred
as hippocampus and cortex, the glial processes end to as mushroom spines. Spines range in length
at the edge of the synaptic cleft, but rarely completely from about 0.3 to 2 m; volume from 0.01 to 0.6 m3;
surround the synapse. Glia are critical for regulating and in synapse area from 0.03 to 0.5 m2, both within
ionic conditions in the extracellular milieu. If, for ex- and across brain regions. There is a near perfect cor-
ample, a glutamatergic neuron fires rapidly, it is the relation between the number of presynaptic vesicles
glia that remove the potassium that is extruded into and the size of the synapse on a spine head, as well
the extracellular space, and also the glia that take up as the spines volume, suggesting that larger spines
the excess glutamate that is released during neuronal have greater synaptic activity at them. The full range
firing. The glutamate stimulates the glia to break in spine diversity can be found along a single short
down glycogen and provide glucose or lactate to the segment of dendrite suggesting that each spine has a
neuron to replenish its energy. Removal of glutamate unique history of activation.
and potassium returns the neurons environment to Like other excitatory synapses, spines contain a
its normal resting state, thereby readying the neuron postsynaptic density. Some spines contain smooth en-
to respond to the next activation. doplasmic reticulum (SER) in proportion with their
Synapses occur in many different locations on the volume, presumably to regulate internal calcium. Due
neurons. The majority of excitatory synapses are dis- to the very small volume of some spines, they appear
tributed along the dendrites and are located up to not to need SER and cytoplasmic calcium buffers are
one millimeter away from the cell body. More than 90 sufficient. Spine cytoskeleton is primarily made up of
percent of the neurons in the brain have tiny protru- actin, and microtubules are not found in the spine
sions called dendritic spines emanating from the compartment, though they run extensively in the
postsynaptic surface (see Figure 1b). Most of the ex- neighboring dendritic shaft beneath the spines.
citatory synapses are located on these dendritic spines The constricted spine neck imparts several im-
and the neurons are referred to as spiny or sparsely portant properties (Harris and Kater, 1994). In most
spiny neurons depending on the number of dendritic brain regions the spine necks are just long and thin
spines they have. Nonspiny neurons comprise about enough to elevate the postsynaptic response to activa-
1 percent to 10 percent of the neurons in a particular tion of the glutamatergic synapse during synaptic
brain region and these often have excitatory synapses transmission without choking off the signal to the
distributed directly onto the dendritic shaft. Most in- parent dendrite. This elevation in postsynaptic po-
hibitory synapses are located on the neuronal cell tential facilitates the opening of the voltage-
body. The axon leaves the neuron from the cell body dependent channels on the spine head, such as the
218 GUTHRIE, EDWIN R.
calcium channel associated with the NMDA gluta- of Washington, where he served as an instructor in
matergic receptor discussed above. The spine necks philosophy from 1914 to 1918 and as an assistant
are also constricted enough to concentrate and com- professor for a year before he joined the department
partmentalize the calcium and possibly other signal- of psychology as an assistant professor. He was pro-
ing molecules in the head near to the specific syn- moted to associate professor in 1925 and was made
apses that were activated. In this way, only those a professor in 1928. During World War II he served
synapses will be modified during particular patterns in Washington, DC, as chief (civilian) consultant to
of activity that lead to elevated or depressed synaptic the overseas branch of the general staff in 1941, and
responses underlying different forms of learning and as chief psychologist of the overseas branch of the Of-
memory. In addition, by concentrating the calcium in fice of War Information in 1942. Upon his return to
the spine head, spines protect the neuron from exci- the University of Washington, he was named dean of
totoxicity by isolating the calcium near the synapse the graduate school from 1943 until he reached re-
and away from the rest of the neuron where high con- tirement age in 1951. He was honored by having a
centrations would damage key structural elements, campus building named for him while he was still
such as microtubules, in the dendrite. alive. Among his other honors was an honorary LLD.
Neurons are born without spines. As an animal from his alma mater, the University of Nebraska,
matures, more dendritic spines are acquired. Den- where he had received his A.B. in 1907, his A.M. in
dritic spine number is affected by experience, such 1910, and his Ph.D. in 1912 (under Edgar A. Singer,
that an enriched environment results in more den- a philosopher whom he much admired and whose
dritic spines along the dendrite, compared to animals views continued to influence his thinking). In 1945,
raised in an impoverished environment. Animals the year he received the LLD, Guthrie was elected
given extensive training also have more dendritic president of the American Psychological Association.
spines. An open question among scientists is whether In 1958, the year before his death, he received the
learning, memory, and other experiences induce the Gold Medal of the American Psychological Founda-
formation of new spines and synapses, or if instead tion, awarded for outstanding lifetime contribution
experience preserves spines from an ongoing pro- to psychology.
duction and loss of spines and synapses. Guthrie was born December 9, 1886, in Lincoln,
Nebraska, the eldest of five children of Edwin R.
See also: APLYSIA: MOLECULAR BASIS OF LONG-TERM
Guthrie and Harriett Pickett Guthrie. His father, the
SENSITIZATION; GLUTAMATE RECEPTORS AND
THEIR CHARACTERIZATION; GUIDE TO THE
son of a clergyman, managed a piano store; his moth-
ANATOMY OF THE BRAIN: NEURON; LONG-TERM er, the daughter of a newspaperman, taught elemen-
POTENTIATION: SIGNAL TRANSDUCTION tary school prior to her marriage. In 1920 Edwin mar-
MECHANISMS AND EARLY EVENTS; ried Helen Macdonald, who helped him translate
MORPHOLOGICAL BASIS OF LEARNING AND Pierre Janets Principles of Psychotherapy (1924). Their
MEMORY: INVERTEBRATES; MORPHOLOGICAL son, Peter M. Guthrie (b. 1926), followed in his fa-
BASIS OF LEARNING AND MEMORY: thers footsteps and became a professor of psychology
VERTEBRATES; NEUROTRANSMITTER SYSTEMS (and department head) at Carleton College.
AND MEMORY
Although Guthrie had already published two
Bibliography short philosophical papers in the Midwest Quarterly,
Cowan, W. M., Sudhof, T. C., and Stevens, C. F. (2001). Synapses. his first major publication was his Ph.D. dissertation
Baltimore, MD: The Johns Hopkins University Press. on Bertrand Russell, which appeared as a monograph
Harris, K. M., and Kater, S. B. (1994). Dendritic spines: Cellular
specializations imparting both stability and flexibility to syn-
(Guthrie, 1915). Once he became a psychologist, he
aptic function. Annual Review of Neuroscience 17, 341371. published chiefly in outlets for psychology, although
Peters, A., Palay, S. L., and Webster, H. deF. (1991). The fine struc- some aspects of his psychological work appeared in
ture of the nervous system: The neurons and supporting cells, 3rd Journal of Philosophy (e.g., Guthrie, 1924) and in a col-
edition. New York: Oxford University Press. lection of philosophical essays (Guthrie, 1942). In
Thomas H. McNeill other words, he never forgot his affiliation with phi-
Jonathan R. Day losophy. Because of his primary interest in learning
Revised by Kristen M. Harris and motivation, he had a scientific interest in educa-
tion (Guthrie, 1945, 1949, 1959a; Guthrie and Pow-
ers, 1950).
Guthries philosophy mentor, Edgar Singer, had
GUTHRIE, EDWIN R. (18861959)
written an article titled Mind as an Observable Ob-
Edwin Ray Guthrie, a distinguished psychologist, ject (Singer, 1911). His point of view prepared Guth-
spent most of his professional career at the University rie for a behaviorist orientation prior to American
GUTHRIE, EDWIN R. 219
psychologist John B. Watsons announcement of be-

haviorism (Watson, 1913). It was not too surprising,
therefore, that his early text General Psychology in
Terms of Behavior, written in collaboration with Ste-
venson Smith, his senior colleague and friend at the
University of Washington (Smith and Guthrie, 1921),
should have stressed behavior and introduced to a
wider audience the type of conditioning as an inter-
pretation of learning that was carried on particularly
by Guthrie (e.g., Guthrie, 1930, 1960). At least among
later students at the University of Washington, the
theory was known as Guthries, whatever role Smith
may have played earlier.
Guthries students became imprinted with this
theory, and its influence upon them is recalled to this
day. Guthrie developed his theory through a charm-
ing writing style, with an emphasis upon convincing
examples rather than experimental demonstration.
The most convincing experimental demonstration
took place many years after the theory was first an-
nounced (Guthrie and Horton, 1946), and eleven
years after the first edition of his Psychology of Learning
(1935).
Much of American psychology had long empha-
sized learning as a fundamental psychological pro-
Edwin Guthrie (Psychology Archives, University of Akron)
cess. What humans are competent to do or learn may
be considered a combination of what one is born with
and the capacities that develop as a matter of natural
growth processes. Children learn to crawl before they row, and the first lamb observed may be thought to
learn to walk, and healthy children require little in- be a different kind of dog. But associations get to be
struction to achieve this skill. Other things they have corrected, both by discrimination and by generaliza-
to learn with assistance. While there is doubtless some tion over a wider range. A fish and a duck do not at
natural tendency to babble and hence to speak, in first seem to be like other animalsat least they do
order to speak the artificial language that their home not have four feetbut they are alive and breathe and
environment demands, children obviously have to move about, and similarities such as these permit
learn the words that they hear and gradually to un- them eventually to be included as animals.
derstand and talk in sentences. So the problem for The law of contrast operates in a corresponding
the psychology of learning is how they acquire way, and the child soon learns that day contrasts with
speech, reading, writing, and arithmetic. These can- night, up with down, and large with small. Thus the
not be entirely inborn, although one child may have contrasting pairs become associated, and one may
more potentialthat is, may be brighterthan an- suggest the other: day-night, up-down, large-small,
other and hence a more accomplished learner. and many other such pairs.
One of the early and persistent theories of learn- The third law of association in this scheme is that
ing was known as the doctrine of association: that one things that occur togetherthat is, in contiguity
word (or one idea) became associated with another, become associated. Hence a color name may suggest
just as the name becomes associated with the object an object often of that colora red apple or a red to-
named or a skilled action becomes associated with the mato, or the contiguity may be more emphasized, as
task performed, as when a hammer is used to pound in an orange orange, where name and object are kept
a nail. There developed theories of how such associa- together.
tions came about. One widespread theory assumed
that there were three laws of association: the laws of Another idea that entered into early learning the-
similarity, contrast, and contiguity. These are familiar ory was that the consequences of an act, either plea-
enough. Many animals are four-footed, so the child surable or painful, would affect the associations
easily learns to group a cat, a dog, and a cow as ani- formed, as in The burned child fears the fire or
mals through their similarities. At first the class is nar- Nothing succeeds like success.
220 GUTHRIE, EDWIN R.
This kind of theory was further extended by as- associative strength on its first pairing with a re-
suming that a child would develop certain prefer- sponse. This is a remarkably parsimonious basic the-
ences (hence activities to be enjoyed) and also some ory. Guthrie elaborated it to explain many puzzling
annoyances (to be avoided). These were elaborated problems of learning and forgetting.
into theories of motivationhow ones goals affect Learning with repetition leads to the usual
what one does, and how they are tied to theories of learning curve because a skilled response becomes
learning based on the sensible notions that humans conditioned to a variety of cues, so that as learning
learn what affords pleasure by satisfying their own proceeds, the proportion of cuesinternal and exter-
wants, and that one hesitates to learn tasks that lead nalthat have become conditioned, each on a single
to results that one dislikes. trial, increases the probability that the intended re-
The Russian scientist Ivan Pavlov introduced the sponses will occur, but a limit is approached as nearly
idea of a conditioned reflex (CR). The standard experi- all the cues become conditioned.
ment was to present a conditioned stimulus (CS) to a Guthrie gave a cogent analogy of how the proba-
dogsay the turning on of a light. This was indiffer- bility model works. He told of an artist whose meager
ent to the dog in that it did not start the flow of saliva, income derived sporadically from the sale of his pic-
which was to become the conditioned reflex (CR). Fol- tures. In order not to live beyond his means, he con-
lowing the CS, the flow of saliva was produced as an verted his cash to dimes that he scattered about the
unconditioned reflex (UR) by a natural unconditioned messy floor of his studio. When he needed money for
stimulus (US), such as some meat powder in a dish ac- a meal, he could at first easily find enough dimes; but
cessible to the dog and attached to an apparatus de- as the dimes became used up, it was harder and
signed to measure the flow of saliva. After the two harder to find more than enough for a very modest
stimuli, CS and US, were repeated several times, the meal. The difficulty became greater and greater as
CS would yield the original UR as a newly learned CR. the dimes were used up, but by diligent searching he
It is easy to see that the CR can also be described as could find enough to keep him going until he made
a form of associative learning, but the experimental another sale.
setting in which it was achieved gave rather good con-
trols over some circumstances favorable to learning. Much later, Estes (1950) developed a stimulus-
sampling mathematical theory of learning that fol-
It soon became clear that all that gets conditioned lowed much the same logic of one-trial learning lead-
is not simple reflexes, such as the salivary reflex to ing to higher achievement as stimuli were assimilated
food in the mouth, and many psychologists began to response, but again approaching an asymptote
talking about conditioned responses instead of condi- after most available stimuli had been used up.
tioned reflexes (Hilgard and Marquis, 1940). This is
where Guthrie came in. He had some disagreements Guthries theory was given a more formal state-
with Pavlov, and offered his own interpretation of ment by his student V. W. Voeks (1950), and in the
conditioning. area of learning theory and experimentation others
of his students have shown the direct influence of
Guthrie proposed to reduce the laws of associa- what they learned from him (e.g., Sheffield, 1961).
tion (and the other theories of learning) to a single
law of association by contiguity. He pointed out that
similarity and contrast were effective only because of See also: REPETITION AND LEARNING
their occurrence together, and then interpreted the
other theories of motivated learning and of condi- Bibliography
tioned response by his one principle appropriate to Estes, W. K. (1950). Toward a statistical theory of learning. Psycho-
all instances of learning: A combination of stimuli logical Review 57, 94107.
which has accompanied a movement will on its recur- Guthrie, E. R. (1915). The paradoxes of Mr. Russell: With a brief ac-
count of their history. Lancaster, PA: New Era Printing.
rence tend to be followed by that movement. (Note (1924). Purpose and mechanism in psychology. Journal of
that Guthrie does not limit the stimulating condition Philosophy 21, 673682.
to a single stimulus [S], nor the movement to a precise (1930). Conditioning as a principle of learning. Psychological
reflex as a response [R].) A slight revision was intro- Review 37, 412428.
duced later, to give some credit to attentive processes. (1942). Conditioning: A theory of learning in terms of stim-
ulus, response, and association. Yearbook, National Society for
Guthrie suggested that his basic law might alterna- the Study of Education 41 (part 2), 1760.
tively be stated as follows: What is being noticed be- (1945). The evaluation of faculty service. Bulletin, American
comes a signal for what is being done (Guthrie, Association of University Professors 31, 255262.
1959b). (1949). The evaluation of teaching. Educational Record 30,
109115.
A second statement is needed before his theory (1959a). The state university: Its function and its future. Seattle:
can be understood: A stimulus pattern gains its full University of Washington Press.
GUTHRIE, EDWIN R. 221
(1959b). Association by contiguity. In S. Koch, ed., Psycholo- Sheffield, F. D. (1961). Theoretical considerations in the learning
gy: A study of a science, Vol. 2, pp. 158195. New York: of complex sequential tasks from demonstration and practice.
McGraw-Hill. In A. A. Lumsdaine, ed., Student expectations in programmed in-
(1935; reprint 1960). The psychology of learning, rev. edition. struction, pp. 1332. Washington, DC: National Academy of
Gloucester, MA: Smith. Sciences/National Research Council.
Guthrie, E. R., and Horton, G. P. (1946). Cats in a puzzle box. New Singer, E. A. (1911). Mind as an observable object. Journal of Philos-
York: Rinehart. ophy, Psychology and Scientific Method 8, 180186.
Guthrie, E. R., and Powers, F. F. (1950). Educational psychology. New Smith, S., and Guthrie, E. R. (1921). General psychology in terms of
behavior. New York: Appleton.
York: Ronald Press.
Voeks, V. W. (1950). Formalization and clarification of a theory of
Hilgard, E. R., and Marquis, D. G. (1940). Conditioning and learning.
learning. Journal of Psychology 30, 341362.
New York: Appleton-Century.
Janet, P. (1924). Principles of psychotherapy, trans. H. M. Guthrie and
E. R. Guthrie. New York: Macmillan.
Ernest R. Hilgard
H
HABITUATION AND SENSITIZATION 1819) enumerated nine parametric features of habit-
IN VERTEBRATES uation and dishabituation that can be seen in a variety
of organisms:
When a ringing bell is presented to a cat, it may evoke
a turning of the head toward the sound source. If that 1. Given that a particular stimulus elicits a response,
same stimulus is repeated over and over again, the repeated applications of the stimulus result in de-
probability and magnitude of this orienting response creased response (habituation). The decrease is
decrease. This phenomenon is called habituation. If a usually a negative exponential function of the
mouse now runs in front of the cat and then the bell number of stimulus presentations.
is rung again, the cat may reorient to the bell. This 2. If the stimulus is withheld, the response tends to
phenomenon is called dishabituation. By recording recover over time (spontaneous recovery).
electrical activity in the first central synapse in the au- 3. If repeated series of habituation training and
ditory system or using another stimulus that elicits an spontaneous recovery are given, habituation be-
orienting response of the same size, it can be shown comes successively more rapid (this might be
that habituation cannot be explained by either senso- called potentiation of habituation).
ry adaptation or muscle fatigue (Thompson and
4. Other things being equal, the more rapid the fre-
Spencer, 1966). Thus, even though the original re-
quency of stimulation, the more rapid and/or
sponse no longer occurs, the stimulus still evokes the
more pronounced is habituation.
same electrical activity in early auditory structures as
it did before and the original response can be fully 5. The weaker the stimulus, the more rapid and/or
elicited by a different stimulus or the same stimulus more pronounced is habituation. Strong stimuli
following dishabituation (e.g., the bell after the may yield no significant habituation.
mouse ran by). Hence, the decrement in response 6. The effects of habituation training may proceed
strength must result from a synaptic change some- beyond the zero or asymptotic response level (i.e.,
where within the nervous system, and this change is additional habituation training given after the re-
specific to the stimulus that was presented repetitive- sponse has disappeared or reached asymptote
ly. will result in slower recovery).
Habituation has been the subject of a great deal 7. Habituation of response to a given stimulus ex-
of empirical investigation because practically every hibits stimulus generalization to other stimuli.
organism displays habituation, even those with very 8. Presentation of another (usually strong) stimulus
primitive nervous systems (Harris, 1943). In review- results in recovery of the habituated response
ing this literature, Thompson and Spencer (1966, pp. (dishabituation).
223
224 HABITUATION AND SENSITIZATION IN VERTEBRATES
9. Upon repeated application of the dishabituatory The Dual-Process Theory of Habituation

stimulus, the amount of dishabituation produced
On the basis of the observation that dishabitua-
habituates (this might be called habituation of
tion appeared to result from a facilitating effect su-
dishabituation).
perimposed on the habituation process (Humphrey,
Thompson and Spencers extremely influential 1933; Sharpless and Jasper, 1956; Thompson and
review gave investigators working in diverse areas an Spencer, 1966) and of some unusual results when
explicit operational definition of habituation against stimulus intensity was used to study habituation
which to test plasticity (change in response output (Davis and Wagner, 1969), Groves and Thompson
with experience) in their particular preparations. In (1970) proposed that novel and especially intense
addition, it led to the general belief that habituation stimuli activate two hypothetical processes: habitua-
might be mediated by a single, fundamental mecha- tion, which decreases response strength; and sensiti-
nism inherent to most organisms across the phyloge- zation, which increases response strength. The final
netic scale. response output is then the net result of these two op-
posing influences. With strong stimuli, the underly-
Other experiments indicated, however, that the
ing habituation process may be masked by a compet-
way in which one interrogates an animal can deter-
ing sensitization process that tends to decrease the
mine these parametric relationships. For example,
rate of response decrement during stimulus repeti-
the probability or amplitude of response is generally
tion. However, because sensitization may not last as
larger, the higher the intensity of the eliciting stimu-
long as habituation, subsequent test sessions may be
lus. It is not surprising, therefore, that it takes longer
used to evaluate the effects of prior habituation,
to reach a low level of response with intense, as op-
somewhat less contaminated by sensitization
posed to weak, stimulus intensities. However, if the ef-
(e.g., Davis, 1972). The Groves and Thomp-
fects of prior exposure to strong and weak stimuli are
son dual-process theory received wide empirical
subsequently evaluated, when all animals are tested
support and provided a fundamental theoretical
with a common stimulus intensity, the magnitude of
base upon which to study the neural mechanisms of
response change is actually greater following strong,
response change during iterative stimulus presenta-
as opposed to weak, stimuli (Davis and Wagner,
tion, in both invertebrates and simple mammalian
1968). Similarly, the probability of response is gener-
systems.
ally lower the shorter the interval from an immediate-
ly prior stimulus. This leads to a rapid rate of re-
sponse decrement when stimuli are presented at Mechanisms of Habituation and
short, rather than long, interstimulus intervals. How-
Sensitization in Vertebrates
ever, when animals are subsequently tested under
conditions where the interstimulus interval is identical Because of the ubiquity of habituation, many be-
for all animals, prior exposure with long intersti- lieve it is the simplest form of learning. The most de-
mulus intervals actually produces a greater decrease finitive analysis of the cellular mechanisms of habitua-
in response strength (Davis, 1970). On the one hand, tion and sensitization has been done in invertebrates.
habituation (i.e., the change in response during stim- In vertebrates, the cellular mechanisms of habituation
ulus repetition) seems to be greater with weak stimuli and sensitization are poorly understood. In broad
presented at short intervals (Thompson and Spencer, terms, habituation could be mediated by some neural
1966); but on the other hand, habituation (i.e., the process intrinsic to the neural pathway in the reflex
change in response strength following stimulus repe- circuit under study or by activation of other neural
tition) seems to be greater with strong stimuli pres- circuits extrinsic to, but impinging on, the reflex
ented at long intervals (Davis, 1970; Davis and Wag- pathway. Much of the literature on humans has
ner, 1968). assumed the latter mechanism, probably because
of the influential theory of E. N. Sokolov (1960),
These disparities illustrate how the term habitua- who proposed a brain comparator process where-
tion has been used to denote the empirical observa- by higher brain centers form a neuronal model
tion of response decrement with stimulus repetition of incoming stimuli and actively inhibit response
as well as a theoretical construct to describe the un- output when subsequent stimuli match the neuronal
derlying process that accounts for the observed remodel.
sponse decrement. However, the two terms may not
be isomorphic, so that it is just as necessary to apply Many animal experiments have attempted to pre-
a distinction between performance and learning with- vent habituation by making lesions of brain areas ex-
in the study of habituation as it is with other forms of trinsic to the reflex circuits under study, or by giving
learning such as classical and instrumental condition- drugs that might prevent these systems from inhibit-
ing. ing the reflex pathway. On balance, however, there
HABITUATION AND SENSITIZATION IN VERTEBRATES 225
are very few behavioral experiments that clearly show Habituation and Sensitization of the Startle
that habituation within a single test session is actually Reflex
prevented by lesions or drugs. When effects are re-
In complex mammalian systems habituation and
ported, they generally result from a change in overall
sensitization of the acoustic startle reflex have been
response levels that does not affect the slope of re- studied by eliciting startlelike responses at different
sponse decrement (provided the manipulations do points along the neural pathway believed to mediate
not push the initial response levels to the ceiling or the very-short-latency (eight milliseconds) startle re-
floor of the response scale and that measures such as flex in rats, with a high level of background noise used
percent decrement or trials to criterion, which de- to sensitize startle (Davis et al., 1982). Startle elicited
pend heavily on changes in overall response level, are by electrical stimulation in the early part of the path-
not used). Moreover, when effects on the slope of the way was increased by the noise but then decreased
response decrement curve are found, it is not clear with repeated elicitations. Startle elicited by stimula-
whether this is due to a change in the underlying pro- tion of the part of the pathway that projected directly
cess of habituation or of sensitization. It has thus been to the spinal cord was also increased by noise but did
extremely difficult to study in whole organisms how not decrease with stimulus repetition. In humans, the
various manipulations affect the process of habitua- R1 component of the eyeblink reflex (latency = ten
tion, since the change in behavioral output may well milliseconds), elicited by electrical stimulation of the
be the product of two underlying processes, which facial nerve, which is mediated by a disynaptic circuit,
cannot be distinguished with a single measure. shows a net increase in response amplitude with stim-
ulus repetition. However, the R2 component (latency
= twenty-five to forty milliseconds) elicited by the
Habituation and Sensitization in the Spinal same stimulus, which involves a polysynaptic pathway,
Cord shows a net decrease in response strength (Sanes and
Ison, 1983). Taken together, the data suggest that
As illustrated by the landmark studies of Spencer, sensitization tends to act on the motor side of reflex
Thompson, and Neilson (1966a, 1966b), the most de- arcs, whereas habituation tends to act on earlier parts
finitive work on the mechanism of habituation and of the circuitry. This suggestion is consistent with data
sensitization in vertebrates has been done in the spi- from Thompson and Spencer (1966) on spinal prepa-
nal cord, because this is one of the few places where rations.
the underlying neural circuitry of the reflexes being
studied is reasonably well understood. Habituation The best evidence for extrinsic control of habitua-
and sensitization have been investigated in centrally tion and sensitization has been gathered by looking
projecting sensory fibers and in the interneurons and at between-session or long-term habituation of the
motoneurons to which they project. The reflexes startle reflex. Leaton and Supple (1986) have shown
that lesions of the cerebellar vermis, which is not part
most studied include the monosynaptic stretch reflex,
of the acoustic startle pathway, significantly attenuate
the oligosynaptic plantar cushion reflex (Egger,
the decrease in startle amplitude seen across daily test
1978; Egger, Bishop, and Cone, 1976), the poly-
sessions without affecting the rate of response decre-
synaptic flexion reflex in the cat (Mendell, 1984); and
ment within test sessions. This blockade of long-term
the lateral column-motoneuron pathway in the frog
habituation was observed with two different stimulus
(Farel, Glanzman, and Thompson, 1973). Mamma-
intensities and cannot be explained by ceiling or floor
lian monosynaptic stretch reflexes (activated by pri- effects caused by the cerebellar lesions. This effect has
mary afferents from muscle spindles that project di- been replicated when the lesion was made before ha-
rectly to motoneurons) typically do not demonstrate bituation training, but not when the lesion was made
marked habituation or sensitization, in contrast to re- afterward (Lopiano, DeSperati, and Montarolo,
flexes involving interneurons, which typically do. 1990). Hence, the cerebellar vermis appears to be
Current evidence indicates that habituation and sen- necessary for the acquisition but not the retention of
sitization are mediated by synaptic changes intrinsic long-term habituation. In contrast, lesions of the mes-
to interneurons within the reflex pathways being encephalic reticular formation, which again is not it-
studied (depression or facilitation of transmitter re- self part of the acoustic startle pathway, block both the
lease). To date, there is no direct evidence that inter- acquisition (Jordan and Leaton, 1982) and the ex-
neuronal networks extrinsic to the reflex pathway ac- pression (Jordan, 1989) of long-term habituation. In
count for habituation and sensitization by actively other studies Borszcz, Cranney, and Leaton (1989)
inhibiting or facilitating transmission (Mendell, have shown that loud startle stimuli produce sensiti-
1984), although such mechanisms cannot be entirely zation that can best be described as fear of the experi-
ruled out. mental context in which startle is measured. Reintro-
226 HABITUATION AND SENSITIZATION IN VERTEBRATES
ducing animals into this context produces a good deal Bibliography

of freezing, a reliable index of fear. Fear of the con- Borszcz, G. S., Cranney, J., and Leaton, R. N. (1989). Influence of
text elevates startle on subsequent test sessions, lead- long-term sensitization on long-term habituation of the
acoustic startle response in rats: Central gray lesions, pre-
ing to a reduction in the amount of long-term re-
exposure, and extinction. Journal of Experimental Psychology
sponse decrement. Treatments such as lesions of the and Animal Behavior Processes 15, 5464.
ventral central gray matter (Borszcz, Cranney, and Davis, M. (1970). Effects of interstimulus interval length and vari-
Leaton, 1989) that are known to reduce freezing, and ability on startle response habituation in the rat. Journal of
treatments such as lesions of the amygdala or drugs Comparative Physiology and Psychology 72, 177192.
such as diazepam, which reduce fear in many situa- (1972). Differential rates of decay of sensitization and habit-
uation of the startle response in the rat. Journal of Comparative
tions, facilitate the degree of long-term response dec- Physiology and Psychology 78, 260267.
rement, presumably by blocking sensitization and Davis, M., Parisi, T., Gendelman, D. S., Tischler, M. D., and Kehne,
hence allowing long-term habituation to be revealed. J. H. (1982). Habituation and sensitization of startle re-
Taken together, these data provide some of the best sponses elicited electrically from the brainstem. Science 218,
evidence that extrinsic systems may be involved in 688690.
Davis, M., and Wagner, A. R. (1968). Startle responsiveness follow-
both long-term habituation and sensitization of the
ing habituation to different intensities of tone. Psychonomic
startle reflex elicited by intense auditory stimuli. Science 12, 337338.
(1969). Habituation of the startle response under an incre-
mental sequence of stimulus intensities. Journal of Comparative
Conclusion Physiology and Psychology 67, 486492.
Egger, M. D. (1978). Sensitization and habituation of dorsal horn
Because habituation could be observed at all le- cells in cats. Journal of Physiology 279, 153166.
vels of the phylogenetic scale, there was great hope Egger, M. D., Bishop, J. W., and Cone, C. H. (1976). Sensitization
that its analysis would lead to fundamental insights and habituation of the plantar cushion reflex in cats. Brain Re-
into the neural mechanisms of learning and memory. search 103, 215228.
Moreover, because habituation was such a basic mech- Farel, P. B., Glanzman, D. L., and Thompson, R. L. (1973). Habit-
uation of a monosynaptic response in vertebrate central ner-
anism, deficits in habituation might allow one to un-
vous system: Lateral column-motoneuron pathway in isolated
derstand complex cognitive disturbances such as frog spinal cord. Journal of Neurophysiology 36, 1,1171,130.
schizophrenia or mental retardation. These hopes Groves, P. M., and Thompson, R. F. (1970). Habituation: A dual
have stimulated a great deal of research. Curiously, process theory. Psychological Review 77, 419450.
however, insights gained from this experience have Harris, J. D. (1943). Habituatory response decrement in the intact
not been as profound as anticipated. In the only sys- organism. Psychological Bulletin 40, 385422.
Humphrey, G. (1933). The nature of learning. New York: Harcourt,
tems where habituation could be analyzed at the cel- Brace.
lular level (e.g., invertebrates and short-term spinal Jordan, W. P. (1989). Mesencephalic reticular formation lesions
preparations), the decrease in response output made after habituation training abolish long-term habitua-
seemed to result from a relatively short-term decrease tion of the acoustic startle response. Behavioral Neuroscience 4,
in transmitter release. However, the actual cellular 805815.
Jordan, W. P., and Leaton, R. N. (1983). Habituation of the acous-
mechanism that mediates this effect is still unknown.
tic startle response in rats after lesions in the mesencephalic
Long-term habituation seems to be a more interest- reticular formation or in the inferior colliculus. Behavioral
ing phenomenon with respect to learning and memo- Neuroscience 97, 710724.
ry, yet it has been much more difficult to study. Theo- Leaton, R. N., and Supple, W. F. (1986). Cerebellar vermis: Essen-
ries that seek to account for enduring, long-term tial for long-term habituation of the acoustic startle response.
Science 232, 513515.
habituation (Wagner, 1976) or sensitization (Borszcz,
Lopiano, L., DeSperati, C., and Montarolo, P. G. (1990). Long-
Cranney, and Leaton, 1989) inevitably appeal to an term habituation of the acoustic startle response: Role of the
associative process, that of classical conditioning. As cerebellar vermis. Neuroscience 35, 7984.
a result, research on the neural mechanisms of habit- Mendell, L. M. (1984). Modifiability of spinal synapses. Physiologi-
uation and sensitization has been largely replaced by cal Review 64, 260324.
Sanes, J. N., and Ison, J. R. (1983). Habituation and sensitization
research on the neural mechanisms of classical condi-
of components of the human eyeblink reflex. Behavioral
tioning. Neuroscience 97, 833836.
Sharpless, S., and Jasper, H. (1956). Habituation of the arousal re-
action. Brain 79, 655682.
See also: APLYSIA: DEVELOPMENT OF PROCESSES
Sokolov, E. N. (1960). Neuronal models and the orienting reflex.
UNDERLYING LEARNING; APLYSIA: MOLECULAR
In M. A. B. Brazier, ed., The central nervous system and behavior:
BASIS OF LONG-TERM SENSITIZATION; III. New York: Josiah Macy Foundation.
INVERTEBRATE LEARNING: C. ELEGANS; Spencer, W. A., Thompson, R. F., and Neilson, D. R. (1966a). Re-
INVERTEBRATE LEARNING: HABITUATION AND sponse decrement of the flexion reflex in the acute spinal cat
SENSITIZATION IN TRITONIA; ORIENTING and transient restoration by strong stimuli. Journal of Neuro-
REFLEX HABITUATION physiology 29, 221239.
HARLOW, HARRY 227
(1966b). Alterations in responsiveness of ascending and re-

flex pathways activated by iterated cutaneous afferent volleys.
Journal of Neurophysiology 29, 240252.
Thompson, R. F., and Spencer, W. A. (1966). Habituation: a model
phenomenon for the study of the neuronal substrates of be-
havior. Psychological Review 73, 1643.
Wagner, A. R. (1976). Priming in STM: An information processing
mechanism for self-generated or retrieval-generated depres-
sion. In T. N. Tighe and R. L. Leaton, eds., Habituation: Per-
spectives from child development, animal behavior, and neurophy-
siology. Hillsdale, NJ: Erlbaum.
Michael Davis
M. David Egger
HARLOW, HARRY (19051981)

Harry F. Harlow was born in Fairview, Iowa, on Octo-
Harry F. Harlow (Harlow Primate Laboratory, University of
ber 31, 1905, and died on December 6, 1981. He at-
Wisconsin)
tended Reed College in 1923 before transferring to
Stanford, where he received his Ph.D. in 1930 under
the supervision of C. P. Stone. Harlows first appoint-
ment was at the University of Wisconsin, where he ing trials animals worked largely by trial and error,
later established the Wisconsin Primate Laboratory. but at some point they began to catch on to a general
Except for the period from1949 to 195l, when he was principle that could be used to successfully solve
chief psychologist for the U.S. Army, he remained at problems in a single trial; that is, the animals had
Wisconsin until his formal retirement in 1974, when learned how to learn, a process that he called the de-
he moved to the University of Arizona. Harlows re- velopment of a learning set. This simple observa-
search, characterized by imaginative methods of tion was important for theoretical reasons, but it had
studying cognition and motivation, led to important another important impact: It could be used to study
discoveries. He was elected president of the American the organization of cortical processes related to learn-
Psychological Association and received its Distin- ing and memory. This represented another major
guished Scientific Contribution Award. He also re- breakthrough, and this type of behavioral analysis
ceived awards from the Society for Research in Child still represents a major tool in studies of neural mech-
Development and was recipient of the Kittay Interna- anisms underlying learning and memory processes.
tional Scientific Foundation Award, the Gold Medal Harlow was convinced that many of the shortcom-
Award of the American Psychological Foundation, ings of the contemporary theoretical systems were
and the U.S. National Medal of Science. due to the paucity of data on the ontogenetic develop-
Harlow was primarily interested in the cortical lo- ment of learning, perception, and motivation. In an
calization of intellectual functions such as learning effort to determine how and when different learning
and memory, and decided to work with primates be- and perceptual processes developed, he and his col-
cause of their obvious cognitive capacities. As he ap- leagues began a major program in which infant mon-
proached this problem, he was convinced that the keys were separated from their mothers at birth and
contemporary learning systems were fundamentally then studied intensively over the ensuing years. While
limited because they were based upon inadequate in- the contributions of these experiments go far beyond
formation, and he set out to collect such information the questions of learning and memory, it was clear
from his monkeys in a systematic manner. One of the from these studies that there were major maturational
most important innovations that Harlow introduced changes in the performance of monkeys on different
was the study of transfer of training. Although there learning tests. These changes could not be accounted
was a long history in the study of this general issue, for by traditional learning theories and thus led to a
previous workers had studied interproblem learning new field of inquiry in learning and memory research.
over a narrow range of problems in which subjects Finally, Harlow was a major advocate of the use
were trained to mastery on a given problem before of the comparative method in studies of learning and
being shifted to new ones. Harlow departed radically memory: Basically the problems of generalization of
from this approach by training animals on individual behavioral data between species are simpleone can-
problems for a small number of trials before shifting not generalize, but one must. If the competent do not
to new problems. He showed that on the initial train- wish to generalize, the incompetent will fill the field
228 HEAD INJURY
(Harlow, Gluck, and Suomi, 1972). The comparative studies of persons after CHI reveal that neuroaxonal
studies of Harlow and those that followed were based cellular damage to the frontal lobes may be present
upon the concepts related to learning sets, which led even within normal-appearing white matter (Garnett
in turn to major advances in the understanding of the et al., 2000), a finding consistent with the weak corre-
phylogenetic changes in learning and memory abili- lation between visible structural damage revealed by
ties. conventional imaging techniques and performance
Harlow thus made a unique and long-lasting con- on neuropsychological tests (Wilson et al., 1988).
tribution not only to the way in which learning and
memory are now studied but also to the development Posttraumatic versus Retrograde
of psychological theory of learning and memory. His
In contrast with the duration of PTA, which may
work helped shape the nature of the questions that
exceed a week after severe CHI, memory loss associat-
are now being addressed.
ed with RA typically substantially resolves during the
See also: COMPARATIVE COGNITION early stages of recovery. RA can initially extend back-
ward several months from the time of injury but even-
Bibliography tually shrinks to encompass only the minutes or sec-
Harlow, H. F. (1956). Learning set and error factor theory. In S. onds immediately before the injury. (Russell, 1971).
Koch, ed., Psychology: A study of a science, Vol. 2, pp. 492537.
During early recovery, RA frequently manifests back-
New York: McGraw-Hill.
Harlow, H. F., Gluck, J. P., and Suomi, S. J. (1972). Generalization ward displacement of temporal orientation far into
of behavioral data between nonhuman and human animals. the past (High, Levin, and Gary, 1990).
American Psychologist 27, 709716.
Sears, R. R. (1982). Harry Frederick Harlow (19051981). American
Psychologist 37, 1,2801,281. Residual Memory Disorders
Bryan E. Kolb The possible link between localized traumatic
focal lesion and memory disorder is complicated by
variation across patients in type and size of lesion.
Nonetheless, recent research relating to specific
memory deficits after CHI indicates that certain types
HEAD INJURY of memory disorders are a more likely consequence
Closed head injuries (CHI; nonpenetrating head in- than others, possibly reflecting the greater probabili-
jury produced by sudden acceleration/deceleration, ty of frontal and temporal lobe damage after CHI.
as in motor vehicle crashes) often cause memory im-
pairments. Both retrograde amnesia (RA) and diffi- Verbal and Nonverbal Learning
culty in learning and retaining new information (an-
Adults and children who have had severe CHI
terograde amnesia) can result from CHI. Post-
and are in the postacute stages of injury are likely to
traumatic amnesia (PTA) refers to the symptom com-
show deficits in immediate and delayed. Many studies
plex of anterograde amnesia, disorientation, and at-
have found that individuals with severe CHI suffer
tentional disturbance during the initial stage of recov-
from recognition impairment on tests of nonverbal
ery, but dissociations can occur in some patients. Even
memory that use abstract, nonverbalizable stimuli,
after the abatement of PTA and RA, half of severe
drawings of nameable objects, and unfamiliar faces.
CHI victims suffer from residual impairments. (Rus-
Although deficient initial learning is common, accel-
sell, 1971).
erated forgetting has also occurred in a subgroup of
The pathophysiologic contribution to residual CHI patients. Although most studies of verbal recall
memory disturbance is difficult to isolate because of report reduction of memory span or other errors of
the heterogeneity of many injuries associated with omission after CHI, persons with severe CHI have
CHI, including primary and secondary causes of frequent commission errors, including a high inci-
brain damage and confounding comorbidities such as dence of intrusions and perseverations.
alcohol abuse. However, several pathophysiologic
features of CHI, including the Glasgow Coma Scale
score (GCS) and duration of coma have been linked Long-Term Memory, Semantic Knowledge
to memory deficit. Proximity of the sphenoidal ridges Although patients with CHI may have normal se-
and bony protrusions on the base of the skull to the mantic knowledge, access to it is often slowed or diffi-
frontal cortex and the anterior temporal lobes makes cult and may be qualitatively different than that of un-
these areas particularly vulnerable to diffuse CHI, injured persons (Haut, 1991). CHI seems to cause
often with superimposed focal lesions (Adams, 1975). impairment of the ability to use semantic information
Further, magnetic resonance spectroscopy (MRS) to facilitate recall in children and adults (Levin and
HEBB, DONALD 229
Goldstein, 1986). Naming impairments, especially in Garnett, M. R., Blamire, A. M., Rajagopalan, B., Styles, P., and Ca-
relation to familiar people, sometimes coexists with doux-Hudson, T. A. D. (2000). Evidence for cellular damage
in normal-appearing white matter correlates with injury se-
accurate and specific semantic knowledge of the peo- verity in patients following traumatic brain injury. A magnetic
ple that could not be named (Sunderland, Harris, and resonance spectroscopy study. Brain 123, 1,4031,409.
Baddeley, 1983). Impaired naming of familiar objects Glisky, E. (1993). Computer-assisted instructions for patients with
is also frequently present following CHI in adults traumatic brain injury: Teaching of domain specific knowl-
(Levin, Grossman, and Kelly, 1977). For example, edge. Journal of Head Trauma Rehabilitation 7, 112.
Haut, M. W., Petros, T. V., Frank, R. G., and Haut, J. S. (1991).
Levin, Grossman, and Kelly (1977) found impaired Speed of processing within semantic memory following severe
object naming in 40 percent of the fifty patients they closed-head injury. Brain and Cognition 17, 3141.
studied in whom PTA had resolved. High, W. H. J., Levin, H. S., and Gary, H. E. J. (1990). Recovery
of orientation and memory following closed-head injury. Jour-
nal of Clinical and Experimental Neuropsychology 12, 703714.
Procedural Memory Levin, H. S. (1989). Memory deficit after closed head injury. Jour-
nal of Clinical and Experimental Neuropsychology 12, 129153.
Most procedural memory studies have focused on the
Levin, H. S., and Goldstein, F. C. (1986). Organization of verbal
acquisition of the knowledge to perform actions, rath- memory after severe closed-head injury. Journal of Clinical and
er than the systematic testing of loss or preservation Experimental Neuropsychology 8, 643656.
of knowledge. Most studies have shown intact proce- Levin, H. S., Goldstein, F. C., High, W. H. J., and Eisenberg, H.
dural learning, even when conjoined with impaired M. (1988). Automatic and effortful processing after severe
closed-head injury. Brain and Cognition 7, 283297.
verbal learning (Timmerman and Brouwer, 1999).
Levin, H. S., Grossman, R. G., and Kelly, P. J. (1977). Aphasic dis-
order in patients with closed head injury. Journal of Neurology,
Implicit Memory Neurosurgery and Psychiatry 39, 1,0621,070.
Russell, W. R. (1971). The traumatic amnesias. New York: Oxford
Although studies directly comparing conscious University Press.
learning (explicit tests of memory) to learning with- Shum, D., Valentine, M., and Cutmore, T. (1999). Performance of
out conscious awareness (implicit tests of memory) in individuals with severe long-term traumatic brain injury on
time-, event-, and activity-based prospective memory tasks.
CHI patients are relatively sparse, findings suggest Journal of Clinical and Experimental Neuropsychology 21, 4958.
that implicit memory is more resistant to impairment Sunderland, A., Harris, D., and Baddeley, A. D. (1983). Do labora-
after CHI than is explicit memory in both adults tory tests predict everyday memory? A neuropsychological
(Glisky, 1993) and children. study. Journal of Verbal Learning and Verbal Behavior 22, 341
357.
Timmerman, M. E., and Brouwer, W. H. (1999). Slow information
Prospective Memory processing after very severe closed-head injury: Impaired ac-
cess to declarative knowledge and intact application and ac-
There has been little research on prospective quisition of procedural knowledge. Neuropsychologia 37, 467
memorythe intention to remember in the future 478.
after CHI. Self-reports indicate that adult CHI pa- Wilson, J. T., Wiedmann, K. D., Hadley, D. M., Condon, B., Teas-
tients report impaired prospective memory as op- dale, G., and Brooks, D. N. (1988). Early and late magnetic
resonance imaging and neuropsychological outcome after
posed to uninjured adults. In studies using experi- head injury. Journal of Neurology, Neurosurgery and Psychiatry
mental tasks to investigate prospective memory in 51, 391396.
adults after CHI, impairments have been reported
Harvey S. Levin
after short delay (Shum, Valentine, and Cutmore,
Revised by Gerri Hanten and Harvey S. Levin
1999) but not long delay. Children with severe CHI
also suffer impairment on experimental tasks of pro-
spective memory.
HEBB, DONALD (19041985)
Conclusion
Donald Olding Hebb was born July 22, 1904, in the
Most studies of persons with CHI using verbal small community of Chester Basin on the Atlantic
and nonverbal learning paradigms have found that coast of Nova Scotia, Canada. Both his parents were
learning impairment is a likely consequence of seri- physicians; his two brothers followed in their foot-
ous head injury. The magnitude of the impairment is steps, and his sister Catherine became a well-known
related to the severity of the injury as determined by neurophysiologist after receiving her medical degree.
the GCS (Levin, Goldstein, High, and Eisenberg, Donald, however, was the familys nonconformist: he
1988). wanted to be a novelist, and in 1925 he graduated
Bibliography from Dalhousie University with a degree in English.
Adams, J. H. (1975). The neuropathology of head injury. In P. J. He spent the next two years teaching and traveling.
Binken and G. W. Bruyn, eds., Handbook of clinical neurology, Hebb never completed his novel, but he was an excel-
Vol. 23. New York: Elsevier. lent writer (his classical work on the neurological
230 HEBB, DONALD
non that later became known as the Hebb synapse, stat-

ing that the discharge of one neuron into another is
increased by the discharge of the second neuron.
Hebb further proposes that: An excited neuron
tends to decrease its discharge to inactive neurons,
an idea that he apparently considered too audacious
to include in the famous neurophysiological postulate
that appears in his 1949 book, The Organization of Be-
havior. Boris Babkin, a McGill physiologist and for-
mer student of Pavlov, read Hebbs thesis, and when
Hebb was mobile again, Babkin encouraged him to
embark on a conditioning experiment. This type of
research soon palled, however, and after his wife was
killed in a motor accident, Hebb decided to leave
Montreal. He was offered an assistantship at Yale
University, but Babkin recommended that he apply
to study with Karl Lashley in Chicago. Hebb took his
advice, a crucial step in his career, and was accepted.
Karl Lashley as Mentor and Colleague

At the University of Chicago, Hebb came under
the influence of L. L. Thurstone, C. Judson Herrick,
Nathaniel Kleitman, Wolfgang Khler, and, of
course, Lashley himself. Lashley believed that the
goal of psychology was to discover how the brain de-
termines behavior, so the questions he asked were dif-
ferent from those arising at the time in mainstream
behaviorism.
After a year Lashley moved to Harvard University
and Hebb went with him, completing his experiments
Donald Hebb (Peter M. Milner)
on the effect of dark-rearing on rat visual perception.
After receiving his doctorate, Hebb remained at Har-
vard for a year and then received a fellowship to study
bases of behavior, The Organization of Behavior, 1949,
Wilder Penfields patients at the newly established
is eminently readable), and he taught numerous stu-
Montreal Neurological Institute. A major project was
dents how to communicate clearly.
to determine the effect of frontal-lobe removal on
measures of intelligence. Hebb found that extensive
damage to that area had a negligible effect on stan-
Early Career
dard intelligence test scores, which pleased Penfield
In 1927, having become interested in the work of and posed an enduring problem for Hebb. Another
Sigmund Freud, Hebb enrolled as a part-time gradu- finding was that large right-temporal-lobe ablations
ate student in psychology at McGill University. He impaired visual perception.
supported himself by teaching and soon was appoint- At the end of his two-year fellowship, Hebb went
ed principal of a local elementary school. He experi- to Queens University, in Kingston, Ontario. There
mented with teaching methods in his school, and at he lectured and experimented with rats to follow up
McGill he developed an interest in the nature- his work at the Neurological Institute. He devised a
nurture question. rat intelligence test (Hebb and Williams, 1946) that
Around this time, newly married, Hebb almost he considered analogous to the tests used on human
decided to make a career in education, but illness and patients and, on the basis of his experiments, con-
tragedy diverted his course. He was immobilized by cluded that early experience had a significant effect
tuberculosis of the hip for more than a year, and dur- on intelligence. Hebb developed this theme and used
ing his illness he wrote a theoretical masters thesis, it to explain the lack of effect of large frontal-lobe le-
Conditioned and Unconditioned Reflexes and Inhi- sions in adults (Hebb, 1942).
bition, which his university examiners rated cum In 1942 Lashley was appointed director of the
laude. In this thesis he speculates about a phenome- Yerkes Laboratories of Primate Biology in Orange
HEBB, DONALD 231
Park, Florida, and he invited Hebb to join the staff. lating that during the initial investigations of an ob-
The opportunity to work with Lashley again easily ject, many different sensory patterns contribute in-
outweighed Hebbs doubts as to the suitability of puts that are incorporated into, or closely linked to,
chimpanzees as experimental subjects, and he seized a single cell assembly representing the object. The
the opportunity. Lashley intended to develop tests of dense intrinsic connections ensure that the whole as-
learning and problem solving while Hebb worked on sembly fires when only partial input is presented.
tests of temperament and emotionality. The effects of
Cell assemblies that are active successively ac-
brain lesions on these tests would then be studied. In
quire connections with each other, which explains ex-
fact no operations were performed until after Hebb
pectancy and the association of ideas. In effect, Hebb
left five years later, largely because Lashleys experi-
developed a neural model of the idea, rescuing it
ence with rats did not adequately prepare him for a
from the obloquy of mentalism under which it had
battle of wits with chimpanzees. In the meantime
languished for almost half a century. Many psycholo-
Hebb made some interesting discoveries about the
gists were chafing under the constraints and dogmas
causes of fear, and he accumulated a fund of chim-
of behaviorism, and they received Hebbs liberating
panzee anecdotes that enlivened his writing and lec-
ideas with enthusiasm. Hebb presented psychologists
tures for the next forty years. But it was the time he
with an alternative to the mind-body dichotomy that
spent pondering the neurophysiology of thought and
had forced behaviorists to outlaw many important
intelligence, and discussing his ideas with Lashley
psychological phenomena.
and other members of the staff, that Hebb valued
most. In 1944 a paper by Lorente de N on neuronal The establishment of cell assemblies and the con-
loops and recurrent circuits came to Hebbs attention nections between them requires neural plasticity. The
and restructured his conception of the brain. Strange- mechanism Hebb proposed was that if an axon termi-
ly, although many psychological phenomena, such as nal and its postsynaptic neuron were active at about
attention, cried out for explanations in terms of feed- the same time, the effectiveness of the synapse would
back from central to peripheral brain areas, and increase. This model of synaptic change is now known
neurophysiologists had known about recurrent cir- as the Hebb synapse, and in the years since his death
cuits for some time, the idea of one-way sensory-to- it may be that Hebb has come to be viewed as a neuro-
motor paths remained deeply ingrained among psy- physiologist. If so, this is a pity; Hebbs attempt to ex-
chologists. The only feedback route envisaged was via plain how the brain represents the outside world, in
the effects of responses on sensory input. which the Hebb synapse played a role, was a more
original contribution than the Hebb synapse itself
One of the most important questions raised by
and, by illustrating the power of neurological models,
Lashley in his monograph Brain Mechanisms and Intel-
was of more far-reaching influence.
ligence (1929) was why the connections established
during learning could not be localized by brain le- Although Hebb made no attempt to test a com-
sions. Another topic that must have been discussed at puter or mathematical model of his conception of the
length by Lashleys circle was stimulus generalization. cell assembly, a not very successful attempt to do so
Why, for example, is learned recognition not dis- was made by Rochester, Holland, Haibt, and Duda
turbed by a change in the size of a visual stimulus? (1956) at IBM shortly after The Organization of Behav-
Hebb thought that Lorente de Ns neural loops ior was published. The cell-assembly idea has influ-
would provide answers to many long-standing prob- enced all subsequent neural-network models of learn-
lems concerning the nature of the representation of ing, from the perceptron (Rosenblatt, 1962) to the
objects in the brain. parallel distributed processing models of Rumelhart and
McClelland (1986). Although Hebbs neural network
was a paper and pencil model that he showed little in-
A Neural Model of the Idea clination to computerize, its pioneering quality
The theory that finally emerged (Hebb, 1949) should make Hebb nerve nets would be a more fitting
was, in keeping with the views of the day, strongly em- memorial to his name than Hebb synapses.
piricist. Hebb postulated that initially random con-
nections between cortical cells become organized by
sensory input into densely interconnected groups Later Career
that he called cell assemblies. Hebbs answer to Lash- Hebb returned to the McGill department of psy-
leys localization problem was that the cells of each as- chology in 1947 and became chairman a year later.
sembly are dispersed over a large area of the cortex He studied the effects of rearing environment in rats
so that enough interconnected cells survive all but the and dogs, obtaining results that were influential in
largest lesions, ensuring that objects continue to be shaping programs for providing more stimulating
represented. Hebb explained generalization by stipu- surroundings for disadvantaged infants. Later he
232 HORMONES AND MEMORY
measured effects of prolonged sensory deprivation on consolidation processes. In addition to such acute ef-
human subjects and probed the cell assembly hypoth- fects, hormones may also have more sustained effects
esis more directly by studying the patterns of break- on cognition through neurotrophic and neuroprotec-
down of perception at reduced contrast, as when an tive actions.
image is held fixed on the retina.
Hebb was elected president of the Canadian Psy-
chological Association in 1952 and of the American
Epinephrine
Psychological Association in 1960. In 1966 he was The adrenergic catecholamine epinephrine is re-
made a fellow of the Royal Society, and in 1979 a for- leased into the blood from the adrenal medulla fol-
eign associate of the National Academy of Sciences lowing arousing or stressful stimulation. Systemic in-
(USA). He was chancellor of McGill University from jections of epinephrine administered after training
1970 to 1974. Hebb retired to his birthplace in 1977 produce dose-dependent effects on subsequent reten-
and died there in August 1985. tion: Low doses enhance retention, whereas high
doses impair retention. As epinephrine passes the
See also: PARALLEL DISTRIBUTED PROCESSING blood-brain barrier poorly, epinephrines effects on
MODELS OF MEMORY memory appear to be initiated by the activation of pe-
Bibliography ripheral adrenoceptors. The finding that drugs that
Hebb, D. O. (1932). Conditioned and unconditioned reflexes and
selectively block peripheral -adrenoceptors (when
inhibition. Masters thesis, McGill University. administered systemically) prevent the memory-
(1942). The effect of early and late brain injury upon test enhancing effects of epinephrine supports this con-
scores, and the nature of normal adult intelligence. Proceed- clusion. Epinephrine acts on receptors in the liver to
ings of the American Philosophical Society 85, 275292. initiate the release of glucose. Evidence that glucose
(1949). The organization of behavior. New York: Wiley.
(1980). D. O. Hebb. In G. Lindzey, ed, A history of psychology
can enhance memory suggests that glycogenolysis
in autobiography, Vol. 7, pp. 273303. San Francisco: Freeman. may play a role in mediating epinephrine influences
Hebb, D. O., and Williams, K. (1946). A method of rating animal on memory. Epinephrine also stimulates the release
intelligence. Journal of General Psychology 34, 5965. of the arousal-related neuromodulator norepineph-
Orbach, J. (1998). The neuropsychological theories of Lashley and Hebb. rine in the brain by activating -adrenoceptors locat-
Lanham, MD: University Press of America.
Rochester, N., Holland, J. H., Haibt, L. H., and Duda, W. L.
ed on vagal afferents terminating on brain-stem nora-
(1956). Tests on a cell assembly theory of action of the brain, drenergic cell groups in the nucleus of the solitary
using a large digital computer. IRE Transactions. Information tract and locus coeruleus. Noradrenergic projections
Theory IT-2, 8093. originating in these nuclei innervate forebrain struc-
Rosenblatt, F. (1962). Principles of neurodynamics: Perceptrons and the tures involved in learning and memory. Extensive ev-
theory of brain mechanisms. Washington, DC: Spartan Books.
Rumelhart, D. E., and McClelland, J. L. (1986). Parallel distributed
idence indicates that the memory-enhancing effects
processing, Vol. 1: Foundations. Cambridge, MA: MIT Press. of epinephrine involve the release of norepinephrine
and activation of -adrenoceptors within a specific
Peter M. Milner
brain region, the basolateral nucleus of the amygda-
loid complex. Microinfusions of -adrenoceptor an-
tagonists into this brain region block the memory-
modulating effects of peripherally administered epi-
HORMONES AND MEMORY nephrine and microinfusions of norepinephrine into
Hormones influence many physiological systems in- this nucleus after training enhance memory. Norepi-
volved in regulating adaptation to environmental nephrine also enhances memory consolidation when
changes. In addition to their many other influences, administered to other brain regions, including the
hormones serve an adaptive role in regulating the nucleus of the solitary tract and the hippocampus.
neurobiological processes underlying memory forma-
tion and other cognitive processes. Extensive evi-
dence indicates that, in laboratory animals, hormones Glucocorticoids
administered shortly after training influence the re- Glucocorticoids (corticosterone in the rat, cortisol
tention of recently acquired information. Although in humans), which is a class of steroid stress hormones
many hormones either enhance or impair long-term released from the adrenal cortex, also influence long-
retention when administered after training, the ef- term memory consolidation. Glucocorticoid hor-
fects of hormones normally released by mildly stress- mones freely enter the brain and bind to two intracel-
ful stimulation have been studied most extensively. lular types of adrenal steroid receptors, glucocorti-
The findings suggest that endogenous hormones re- coid receptors and mineralocorticoid receptors. The
leased by emotionally arousing training experiences low-affinity glucocorticoid receptors are involved in
influence long-term memory by modulating memory mediating glucocorticoid effects on memory consoli-
HORMONES AND MEMORY 233
dation. Glucocorticoids act through intracellular and of CRH into the hippocampus, the amygdala, or the
intranuclear receptors and can affect gene transcrip- bed nucleus of the stria terminalis affect memory con-
tion by direct binding of receptor homodimers to solidation by interacting with noradrenergic mecha-
DNA. Glucocorticoids may also affect memory consol- nisms.
idation through transactivation or protein-protein in-
teractions with other transcription factors or effector
systems. Administration of low doses of glucocorti- Vasopressin and Oxytocin
coids enhances memory consolidation and glucocor- Vasopressin and oxytocin are synthesized in the
ticoid antagonists impair consolidation. Memory- hypothalamus and are released in the brain and into
modulating effects are induced by systemic injections the cerebrospinal fluid and general circulation (via
of glucocorticoid receptor agonists and antagonists as the posterior pituitary). The possibility that vasopres-
well as by infusions administered into several brain sin may influence learning and memory was suggest-
regions, including the nucleus of the solitary tract, the ed by findings that lesions of the posterior pituitary
basolateral amygdala, or the hippocampus. Glucocor- affected the extinction of an avoidance response and
ticoid-induced memory-modulating effects require that the effect was attenuated by injections of an ex-
concurrent noradrenergic activation within the amyg- tract of pituitary tissue containing vasopressin. Subse-
dala. Glucocorticoid effects on memory consolidation quent research has provided extensive evidence that
are, thus, conditional, that is, their actions require memory is enhanced by vasopressin administered im-
coactivation of other transmitter systems. Glucocorti- mediately after training. Although some evidence
coid effects on memory are not restricted to influ- suggests that the effects of vasopressin on memory
ences on consolidation. Glucocorticoids generally im- may be mediated, at least in part, by peripheral ef-
pair short-term memory in experimental animals and fects, including alterations in blood pressure, other
human subjects, and, when administered shortly be- evidence indicates that central actions of vasopressin
fore retention testing, glucocorticoids also impair re- influence memory consolidation. Peptide metabolites
trieval of long-term memory for spatial-contextual of vasopressin that do not affect blood pressure have
and declarative information. These effects are tempo- highly potent effects on memory. Additionally, ad-
rary and dissipate within several hours after stress ex- ministration of a vasopressin antagonist into the cere-
posure or hormone injection. Prolonged exposure to bral ventricles blocks the memory-enhancing effect of
high levels of glucocorticoids due to chronic stress peripherally administered vasopressin without alter-
tends to induce prolonged impairment of cognitive ing vasopressin effects on blood pressure. Memory is
functioning. also enhanced by low doses of vasopressin adminis-
tered directly into a number of brain regions, includ-
ing the hippocampus. Evidence that lesions of the
ACTH and CRH dorsal adrenergic bundle block the memory-
The peptide hormone adrenocorticotropin enhancing effect of vasopressin indicates that, as with
(ACTH) is secreted from the anterior pituitary. Early other hormonal effects of memory, nonadrenergic ac-
studies investigating the influence of hormones on tivation may be required. Although the effects of oxy-
learning and memory reported that hypophysectomy tocin have been less extensively examined, oxytocin
produced impairment of learning and memory and may either impair or enhance memory, depending
that the impairment was attenuated by injections of on the experimental training conditions.
ACTH. Many subsequent experiments using animals
with intact pituitary glands found that memory is en-
hanced by low doses of ACTH and impaired by high Opioid Peptides
doses. Evidence that retention is affected by frag- Research findings indicating that the opiate drug
ments of the ACTH peptide that do not induce the re- morphine impairs memory when injected after train-
lease of glucocorticoids from the adrenal cortex indi- ing suggest that endogenous opioid peptides may
cates that ACTH effects on memory are not mediated play a role in the regulation of memory consolidation.
by glucocorticoids. The corticotropin-releasing hor- The endogenous opioid peptide -endorphin is re-
mone (CRH), another peptide hormone released into leased within the brain and is also released from the
the blood and brain after arousing or stressful stimu- anterior pituitary into the blood along with ACTH.
lation, enhances memory consolidation when admin- Enkephalins are released from the adrenal medulla,
istered shortly after training, an effect that does not together with epinephrine. Experimental findings in-
require release of glucocorticoids. The involvement dicating that effects of these opioid peptides on mem-
of extra-hypothalamic CRH in memory and related ory are similar to those of morphine suggest that en-
cognitive functions has been found in experiments dogenously released opioid peptides regulate
with both animal and human subjects. Local infusions memory consolidation. Studies of the effects of opiate
234 HULL, CLARK L.
antagonists provide additional support for this view: ditions also serve a highly adaptive role of influencing
Opiate antagonists enhance retention when adminis- memory by interacting with neuromodulatory sys-
tered after training. As with other hormones, opioid tems in brain regions involved in regulating memory
peptide influences on memory involve the amygdala. consolidation.
Injections of opiate agonists and antagonists adminis-
See also: STRESS AND MEMORY
tered directly into the amygdala after training pro-
duce effects on memory highly comparable to those Bibliography
produced by systemic injections. Such effects also in- Bohus, B. (1994). Humoral modulation of learning and memory
volve interactions with the noradrenergic system: - processes: Physiological significance of brain and peripheral
adrenoceptor antagonists injected into the amygdala mechanisms. In J. Delacour, ed., The memory system of the brain,
pp. 337364, Singapore: World Scientific.
block the memory-enhancing effects of the opiate an- De Wied, D. (1991). The effects of neurohypophyseal hormones
tagonist naloxone. Such findings agree with the evi- and related peptides on learning and memory processes. In
dence that opioid peptides inhibit the release of nor- R. C. A. Frederickson, J. L. McGaugh, and D. L. Felten, eds.,
epinephrine in the amygdala and other brain regions. Peripheral signaling of the brain, pp. 335350. Toronto: Hogre-
Memory can also be influenced by a number of pep- fe and Huber.
Koob, G. F. (1987). Neuropeptides and memory. In L. L. Iversen
tide hormones, including substance P, neuropeptide and S. H. Snyder, eds., Handbook of psychopharmacology, Vol.
Y, somatostatin, prolactin, and cholecystokinin (CCK) 19: Behavioral pharmacology, and update, pp. 531573. New
that are not generally thought to be stress related. York: Plenum.
McEwen, B. S. (2000). Stress, sex, and the structural and functional
plasticity of the hippocampus. In M. S. Gazzaniga, ed., The
Gonadal Hormones new cognitive neurosciences, 2nd edition, pp. 171198. Cam-
bridge, MA: MIT Press.
Gonadal hormones generally have more sus- McGaugh, J. L., Roozendaal, B., and Cahill, L. (2000). Modulation
tained effects on cognitive functioning, but acute ef- of memory storage by stress hormones and the amygdaloid
fects on memory consolidation have also been report- complex. In M. S. Gazzaniga, ed., The new cognitive neuro-
sciences, 2nd edition, pp. 1,0811,098. Cambridge, MA: MIT
ed. For example, testosterone and estradiol, major Press.
gonadal steroid hormones, have neurotrophic and Roozendaal, B. (2000). Glucocorticoids and the regulation of mem-
neuroprotective actions on the brain. A decline in ory consolidation. Psychoneuroendocrinology 25, pp. 213238.
cognitive function (particularly spatial memory) can James L. McGaugh
be found in postmenopausal women and estrogen- Revised by James L. McGaugh and Benno Roozendaal
replacement therapy may attenuate the deficits. Es-
trogens also appear to have a protective effect against
the neurodegenerative and memory-impairing ef-
fects of excessive glucocorticoids. HULL, CLARK L. (18841952)
Clark Leonard Hull was born in Akron, New York, on
Interactions Among Hormonal Systems May 24, 1884, and died in New Haven, Connecticut,
on May 10, 1952. He earned his bachelors degree
There are many interactions among hormones as from the University of Michigan in 1913, his masters
well as interactions among hormones and transmitter degree in 1915, and a Ph.D. in experimental psychol-
systems in modulating memory consolidation. For ex- ogy in 1918 from the University of Wisconsin. His
ample, the effects of epinephrine on memory are in- graduate work was done primarily under the direc-
fluenced by glucocorticoids. The memory-enhancing tion of Joseph Jastrow, Daniel Starch, and Vivian A.
effects of epinephrine are blocked by metyrapone, a C. Henmon.
drug that prevents the synthesis and release of gluco-
corticoids. Furthermore, vasopressin is ineffective in Hull recorded some of his earliest career plans in
influencing memory consolidation in adrenal de- his Idea Books. It seems, he wrote, that the great-
medullated animals. Opioid peptides modulate est need in the science at present is to create an experi-
memory consolidation through influences on central mental and a scientific knowledge of higher mental
noradrenergic and cholinergic systems. Naloxone ef- processes (Ammons, 1962, p. 814). He planned to
fects on memory are blocked by the (-adrenoceptor become the supreme authority in the psychology of
antagonist propranolol as well as by the muscarinic abstraction, concept formation, and, possibly, reason-
cholinergic antagonist atropine. Additionally, the ingto both know the literature and create the liter-
muscarinic cholinergic agonist oxytremorine and the ature on the subject. His doctoral dissertation, Quan-
acetylcholinesterase inhibitor physostigmine attenu- titative Aspects of the Evolution of Concepts, was published
ate the memory-impairing effects of -endorphin. in Psychological Monographs (1920).
Thus, several hormonal systems known to have im- Hull was afflicted with a variety of health prob-
portant roles in enabling adaptation to stressful con- lems, and his relatively late entry into the field was an-
HULL, CLARK L. 235
other threat to his long-range goal. In 1930, at the

age of forty-six, he wrote, Sometimes I have been de-
pressed and discouraged in my hope to achieve a
major contribution to the theory of knowledge by the
fact of my age. Recently, however, the examination of
the ages at which several of the great critics have pro-
duced their best works has shown that I have by no
means reason to be depressed (Ammons, 1962, p.
836). The list included English philosopher Thomas
Hobbes, sixty-three; Dutch philosopher Baruch Spi-
noza, forty-five; German philosopher and mathema-
tician Gottfried Wilhelm von Leibnitz, sixty-eight;
English philosopher John Locke, fifty-eight; and Ger-
man philosopher Immanuel Kant, fifty-seven. (He
could have added that Russian scientist Ivan Pavlov
was fifty-six years old when he first proposed the prin-
ciple of the conditioned reflex.)
Hull was elected to the American Academy of Arts
and Sciences in 1935, to the National Academy of Sci-
ences in 1936, and to the presidency of the American
Psychological Association 1935. He received the War-
ren Medal of the Society of Experimental Psycholo-
gists in 1945. A measure of his influence is that during
the decade spanning 1941 to 1950 approximately 40
percent of all experimental articles published in the
Journal of Experimental Psychology and the Journal of Clark L. Hull (Psychology Archives, University of Akron)
Comparative and Physiological Psychology included refer-
ences to his work (Spence, 1952). Today Hull is best
remembered for one theoretical publication on learn- of a machine that learned, he thought, should allow
ing theory, Principles of Behavior (1943). However, a for common reactions to stimuli with sensory similari-
case has been made that the twenty-one theoretical ty (generalization); for the learning of common reac-
articles he published in the Psychological Review be- tions to stimuli with quite different sensory qualities
tween 1929 and 1950, the earliest of which intro- (discrimination); and for the possibility of differential
duced the Pavlovian construct the fractional anticipato- responding to minute differences in stimulus pattern
ry goal response (rG-sG), represent at least as important (afferent neural interaction). These became central
a contribution to learning theory. (For an extensive principles in his papers and in his later formal writ-
account of Hulls work and influence, see Amsel and ings (Hull, 1943, 1952).
Rashotte, 1984.)
In 1929, Hull accepted the position of research
After receiving his doctorate, Hull stayed in the
professor of psychology at the Institute of Psychology
psychology department at the University of Wiscon-
(later the Institute of Human Relations) at Yale Uni-
sin, where he became director of the laboratory in
versity, where he remained for the rest of his career.
1925. During these years, his projects included the ef-
When he arrived at Yale, at the age of forty-five, he
fects of tobacco smoking on mental and motor effi-
made this entry in his Idea Books, which seems pro-
ciency; the definitive books of the time in aptitude
phetic in the light of the computer age:
testing and hypnosis; and, as forerunners of the com-
puter age, a logic machine that automatically comput- Just as the correlation machine has been inti-
ed correlations, and a robot that learned. mately associated with the testing program,
For Hull, Pavlovs conditioning identified the so it appears that the design and construction
principle by which an event in one subsystem could of automatic physical machines will be inti-
come to influence functioning in remote parts of such mately associated with my attempts to work
machines, and Edward Thorndikes law of effect, the out my program involving the higher mental
principle whereby originally random movements processes. . . . I am pretty certain to be criti-
could be selected and be linked as responses to specif- cized and called a trifle insane at the very
ic stimulus patterns (trial-and-error learning and the least. But whatever genius I have quite evi-
principle of the habit-family hierarchy). The design dently lies in this direction. I can do no less
236 HULL, CLARK L.
than make the best of itlet the tendency, cluding conscious experience, from action, i.e., habit
have free rein and go as far as it possibly (Ammons, 1962, p. 837).
can. . . . It may lead to real insight into the
To carry out this task, Hull planned a three-
higher mental processes. . . . It may [on the
volume work. The first volume was intended mainly
other hand] possibly serve as the final reductio
to present a set of formal axioms that constituted a
ad absurdum of a mechanistic psychology. If it
logical system from which hypotheses about mamma-
does, well and good. But even if this should
lian adaptive behavior could be deduced. Completed
take place, it may at the same time result in
in Hulls fifty-eighth year, this volume appeared in
such a development in psychic machines dis-
1943 and was titled Principles of Behavior: An Introduc-
playing an utterly new and different order of
tion to Behavior Theory. The orientation of the second
automaticity that mechanical engineering of
volume, based on the principles set down in the first,
automatic machines will be revolutionized to
was toward specific instances of more complex adap-
a degree similar to the introduction of steam
tive behavior, such as maze learning and problem
engines and electricity. (Ammons, 1962, pp.
solving. Completed in Hulls sixty-eighth year, it ap-
828829)
peared shortly after his death in 1952 and was titled
As research professor at Yale, Hull had no formal A Behavior System: An Introduction to Behavior Theory
teaching assignment, but he engaged in graduate in- Concerning the Individual Organism. The third volume,
struction through a weekly seminar that attracted stu- which Hull thought would be the most important, was
dents and faculty of the institute for discussion of a va- to have applied the system to some elementary phe-
riety of issues in behavior theory. nomena of social mammalian behavior. It was never
completed.
By Hulls own account, the Institute of Human
Relations was a loose organization of behavioral sci- The basic formal structure of the theorizing in
entists from various fields, mainly psychologists, soci- Hulls Principles, as revised in Essentials of Behavior
ologists, and cultural anthropologists. However, the (1951), was an intervening variable approach bor-
conceptual framework for the unified contribution it rowed from Edward Tolman (1938). It involved ante-
provided was Hulls theoretical system, and it seems cedent, manipulable environmental conditions (inde-
fairly clear, in retrospect, that this conceptual frame- pendent variables), consequent behavioral conditions
work was an amalgam of influences from English nat- (dependent variables), and a bridge of lower- and
uralist Charles Darwin, Pavlov, and Austrian neurolo- higher-level constructs connecting the two (the inter-
gist Sigmund Freud. Their influences, filtered vening variables). Hulls final version of the theory is
through Hulls own, are evident in several important summarized in the first chapter of A Behavior System
publications written by members of the institute, all as a set of postulates and corollaries, most of which
of which featured stimulus-response analyses of com- are restated in mathematical form. Examples are Pos-
plex learning. Books in this genre include Frustration tulate II, relating the strength of an intervening vari-
and Aggression (Dollard et al., 1939), Social Learning able, the molar stimulus trace S1, to the indepen-
and Imitation (Miller and Dollard, 1941), and Personal- dent variables, the physical stimulus (S), as a power
ity and Psychotherapy (Dollard and Miller, 1950). function of time since the beginning of the stimulus;
and Postulates XIV, XV, and XVI, in which the mo-
Even before going to Yale, Hull had planned to mentary effective reaction potential SER, which al-
prepare a magnum opus on the scientific (mechanis- ready takes into account the behavioral oscillation
tic) analysis of higher mental processes. The earliest SOR and the threshold for responding SLR, is related
titles he considered for his work, listed in his Idea to four response measuresresponse latency, re-
Books in 1928 (Ammons, 1962, pp. 824825), indi- sponse amplitude, response frequency, and resistance
cate this intended focus and the mechanistic empha- to experimental extinctionby negatively accelerat-
sis: Mechanisms of Thought, Mechanisms of Mind, ed functions in the first, third, and fourth cases, and
Mental Mechanisms, Mechanisms of Mental Life, Psy- by a linear function in the second. This is the nature
chology from the Standpoint of a Mechanist. of Hulls formal theorizing.
Hulls plan was to invert the direction taken by These later, more formal portions of Hulls theo-
the great philosophersDavid Hume, Locke, Kant, rizing were not taken up after his death except by
and Hobbeswho had attempted to construct a theo- Kenneth Spence (1954, 1956) and a few of Spences
ry of knowledge, thought, and reason on the basis of students (Grice, 1968; Logan, 1960). One reason was
conscious experience but who, in Hulls eyes, had the advent of the cognitive revolution in psychology,
failed. He planned to attack the same problem using which was for the most part a reaction against B. F.
the opposite strategy. He wrote, I shall start with ac- Skinners behaviorism, but also against Hulls stimu-
tionhabitand proceed to deduce all the rest, in- lus-response associationism, his hypothetico-
HUNTER, WALTER S. 237
deductive version of it in particular (Amsel, 1989). Amsel, A., and Rashotte, M. E. (1984). Mechanisms of adaptive behav-
Many thought that such a formal Newtonian treat- ior: Clark L. Hulls theoretical papers with commentary. New York:
Columbia University Press.
ment was both too general in scope and premature. Dollard, J., and Miller, N. E. (1950). Personality and psychotherapy.
However, more recent mathematical models of asso- New York: McGraw-Hill.
ciative learning, greatly influenced by Hull but more Dollard, J., Doob, L. W., Miller, N. E., Mowrer, O. H., and Sears,
restricted in explanatory scope (Rescorla and Wag- R. R. (1939). Frustration and aggression. New Haven: Yale Uni-
ner, 1972), remain influential. Still very important in versity Press.
Grice, G. R. (1968). Stimulus intensity in response evocation. Psy-
learning theory is the general approach, taken from chological Review 75, 359373.
Hulls earlier work, of employing hypothetical con- Hintzman, D. L. (1992). Twenty-five years of learning and memo-
structs derived from Pavlovian conditioning in the ex- ry: Was the cognitive revolution a mistake? In D. E. Meyer and
planation of instrumental behavior. Hull (1931) re- S. Kornblum, eds., Attention and performance XIV. Hillsdale,
garded this work as stimulus-response analyses of NJ: Erlbaum.
Hull, C. L. (1920). Quantitative aspects of the evolution of concepts. Psy-
goal attraction and directing ideas, and called chological monographs 28, whole no. 123.
the mechanism the fractional anticipatory goal response (1931). Goal attraction and directing ideas conceived as
(rG-sG). An extensive treatment of such a mechanism habit phenomena. Psychological Review 38, 487506.
in appetitive learning occurs in frustration theory (1943). Principles of behavior: An introduction to behavior theory.
(Amsel, 1962). New York: Appleton-Century-Crofts.
(1951). Essentials of behavior. New Haven: Yale University
The similarity of Hulls theorizing in the 1930s Press.
and 1940s to recent computer-generated models of (1952). A behavior system: An introduction to behavior theory con-
learning and memory has been noted by Hintzman cerning the individual organism. New Haven: Yale University
Press.
(1992) in an article with the subtitle Was the Cogni- Logan, F. A. (1960). Incentive. New Haven: Yale University Press.
tive Revolution a Mistake? Hintzman, himself a (1979). Hybrid theory of operant conditioning. Psychological
product of the cognitive revolution, asserts that Key Review 86, 507541.
ideas of the cognitive revolution, including cognitive Miller, N. E., and Dollard, J. (1941). Social learning and imitation.
organization and the computer metaphor, have been New Haven: Yale University Press.
Rescorla, R. A., and Wagner, A. R. (1972). A theory of Pavlovian
largely abandoned; and basic concepts from the era conditioning: Variations in the effectiveness of reinforcement
of behaviorism and functionalism, such as association, and nonreinforcement. In A. H. Black and W. F. Prokasy,
inhibition, similarity, unconscious learning, and eds., Classical conditioning II: Current research and theory. New
transfer, have reemerged. He points particularly to York: Appleton-Century-Crofts.
the similarity of Hullian theories of fifty years ago Spence, K. W. (1952). Clark Leonard Hull: 18841952. American
to connectionist and production-system models of (1954). The relation of response latency and speed to the
more recent cognitivists. Much the same can be said intervening variables and N in S-R theory. Psychological Review
about the approaches to theory based on neural net- 61, 209216.
works that often depend on what is called the Hebbi- (1956). Behavior theory and conditioning. New Haven: Yale
an synapse. The proposition here is that the efficiency University Press.
Tolman, E. C. (1938). The determiners of behavior at a choice
of cell A in firing cell B is increased as a function of point. Psychological Review 45, 141.
a growth process that takes place at the synapse be-
Abram Amsel
tween cell A and cell B. As a principle of association,
this is not greatly removed from Hulls reinforcement
postulate. In short, it appears that a part of Hulls last-
ing contribution will be to modern theories of the HUNTER, WALTER S. (18891954)
neurobiology of learning and memory, an ironic out-
come considering the criticisms Hull suffered for Scientists studying learning and memory know Walter
what was called his neurologizing. Samuel Hunter best for his analytical application of
a device for assessing short-term retention, the de-
layed-response task. But his contributions to psychol-
See also: HEBB, DONALD ogy were much more broad, and deep, than that. He
was an influential and moderating force in the behav-
Bibliography iorist movement; his thoughts and experiments in-
Ammons, R. B. (1962). Psychology of the scientist: IV. Passages
cluded topics that remain issues of mainstream inter-
from the Idea Books of Clark L. Hull. Perceptual and Motor est, such as the nature of an animals representation
Skills 15, 807882. for memory, consciousness, and genetic influences on
Amsel, A. (1962). Frustrative nonreward in partial reinforcement intellectual achievement; and he was a significant ap-
and discrimination learning: Some recent history and theo- plied psychologist with the military.
retical extension. Psychological Review 69, 306328.
(1989). Behaviorism, neobehaviorism, and cognitivism in learn- When Hunter was born, in 1889, it had been
ing theory. Hillsdale, NJ: Erlbaum. about a century since his Scotch-Irish ancestors had
238 HUNTER, WALTER S.
thinking as a scientist was shaped by two of his profes-

sors, James Angell and Harvey Carr, who were among
the leading functionalists of the day. Their influence
undoubtedly helped form Hunters favorable view of
behaviorism, which had close links to functionalism
(an orientation that emphasized the relationships be-
tween environmental or task variables and learning,
rather than theories of how hypothetical processes
determine the occurrence of learning). In addition,
Angell had directed the Ph.D. dissertation of John
Watson at the University of Chicago a few years earli-
er. He undoubtedly maintained sufficient contact
with Watson to transmit the latters ideas to his own
students. It is likely that in this way, Hunter was ex-
posed early to the notions set forth by Watson in his
Psychological Review paper mapping out the elements
of behaviorism just one year after Hunter received his
Ph.D.
Undeterred by a teaching load of four courses per
semester, Hunter was productive in his research at
the University of Texas, and after four years was ap-
pointed professor and head of the department of psy-
chology at the University of Kansas (at age twenty-
seven). Soon afterward, World War I required that
Hunter work in the military service. As a psychologi-
cal examiner he helped illustrate the predictive value
Walter S. Hunter (Brown University Archives) of simply administered psychological tests, and in
World War II he served as one of the militarys lead-
ing administrators of such testing. In 1925 Hunter be-
come to the United States. Born in Decatur, Illinois, came the first G. Stanley Hall Professor of Genetic
within a year of the births of U.S. president Dwight Psychology at Clark, and in 1936 he was named head
D. Eisenhower (1953 to 1961), French soldier Charles of the department at Brown University, where he re-
de Gaulle, German dictator Adolf Hitler, English mo- mained until his death in 1954 (Carmichael, 1954).
tion-picture actor Charles Chaplin, and English poet Hunter worked influentially in the editing of several
and critic T. S. Eliot, Hunter spent his early adoles- journals (Psychological Bulletin, Comparative Psychology
cence working on his fathers farm near Fort Worth, Monographs, Behavior Monographs, and Journal of Ani-
Texas. His intellectual interests emerged early. In the mal Behavior), and he created Psychological Abstracts
first stages of his adolescence he purchased and read and edited it for twenty years. His several and diverse
English naturalist Charles Darwins groundbreaking honors included membership in the National
Origin of Species and Descent of Man, and by age fifteen Research Council and the National Academy of Sci-
he had developed an active interest in a career in elec- ences, posts as president of several professional re-
trical engineering (Hunt, 1956). He attended Poly- search societies, including the American Psychologi-
technic College at Fort Worth with this objective in cal Association, and receipt of the U.S. Presidents
mind, but psychology attracted his attention through Medal for Merit for his service as chairman of the Ap-
his reading of American psychologist and philoso- plied Psychology Panel of the National Defense Re-
pher William Jamess textbooks, and he transferred search Council during World War II (Schlosberg,
to the University of Texas in 1908 to study this field 1954).
(Hunt, 1956).
Hunters scientific contributions are equally im-
After graduating from the University of Texas in pressive and diverse, including his study of visual af-
1910, he pursued his graduate work at the University terimages, auditory discriminations, and issues of so-
of Chicago. Within only two years he completed his cial psychology, consciousness, and thinking, but his
doctorate and returned as an instructor to the Univer- most lasting contributions dealt with the topic that
sity of Texas. But these two years at Chicago must pervaded his writing in all other respects: memory.
have been a concentrated experience indeed, in view This is despite the fact that Hunter himself assiduous-
of two particularly substantial outcomes. First, his ly avoided reference to memories or a memory
HUNTER, WALTER S. 239
process in animals. To describe how a horse seemed the rewardmight be represented within the animal
to remember the route home, Hunter preferred inas an intraorganic cue, probably of a kinesthetic na-
stead the concept of sensory recognition, which, he ture.
felt, avoided attributing ideation to animals. Hunters research stimulated a number of studies
Hunters Ph.D. thesis was intended to test the ani- by other scientists directed at one of two issues of phy-
mals behavior in the absence of an eliciting stimulus, logeny: which animals can perform such a delayed re-
the sort of circumstance that might be thought to response task without the use of physical orientation
quire a memory system of some sort. He chose a test and thus reveal a capacity for internal ideation of the
of retention that had been tried out by two of Carrs sensory thought type suggested by Hunter; and
students, Haugh and Reed. Whatever its origin, how animals are ordered phylogenetically in terms of
Hunter adapted the delayed-response task in a fash- their capacity for such ideation, which was presumed
ion sufficiently convincing to persuade others to use to be tapped by the degree to which correct respond-
it as the dominant task for short-term retention over ing occurred after a long delay interval. Hunter later
the next fifty years, until delayed matching-to-sample argued that correct performance of the double alter-
came to be more effectively applied through the use nation task also could reveal a capacity for ideation.
of computers. Hunter made substantial contributions to psychology
throughout his forty-two years in the field. His ideas
In his first version of this task Hunter trained in a variety of areas within psychology were influen-
hungry animals initially to obtain food by entering tial, but it can be argued that none of his subsequent
the lighted door of three possible doors (in a later ver- experiments had the impact of his Ph.D. thesis.
sion, no light was used and only the location of the
door indicated the reward site). After the animal had Some intriguing paradoxes of Hunters career
perfected this discrimination, Hunter began trials in have been described by Hunt (1956) and are para-
which the light was shown in the same way but turned phrased here. Despite Hunters success in avoiding
off before the animal was allowed to choose a door. serving in university administration, for which he was
For the animal to reach this point required hundreds frequently sought, he was obviously a successful ad-
of training trials. Yet, because responding was per- ministrator, as indicated by the strength of the de-
mitted only in the absence of the stimulus, at specified partment he built at Brown and his work in the mili-
intervals after taking it away, reliable correct respond- tary. Despite his sympathy with and support of
ing hardly could be caused by sensory recognition; behaviorism, some of his most important research
there were no sensory events to be recognized. So seemed more clearly understandable in terms of the
long as he could discount overt orienting attitudes use of symbolic processes by animals. Although he was
of the sort that a hunting dog might use when indicat- one of the most influential experimental psycholo-
ing the location of a bird in the field, accurate choice gists of the first half of the twentieth century, he spent
could apparently be attributed to control by an inter- the last ten years of his life focusing on psychologys
nal representational response, or some other form of applications. Finally, Hunt notes two paradoxes of
ideation. Hunter reasoned that the efficiency of ide- Hunters personal life: Despite a high rate of profes-
ation could be compared across various species of ani- sional productivity, Hunter always seemed relaxed
mals or even between animals and humans by increas- and with time to spare; and, despite his emphasis on
ing the interval between the offset of the light and the being impersonal and objective as a scientist, he was
opportunity to respond. warm in his relationships with others.
Hunter estimated the longest retention interval Hunters primary contribution to the field of
that each of his subjects could tolerate before reten- learning and memory was to implement two tasks, the
tion no longer was significant. For rats he estimated delayed response and the double alternation task,
this to be about ten seconds, for raccoons twenty-five which provided a metric for discussing animal memo-
seconds, for dogs five minutes, for his thirteen- ry in an objective fashion and a forum for considering
month-old daughter Thayer about twenty-four sec- the use of representation in memory by animals.
onds, for two-and-a-half-year old children fifty sec-
onds, and for a six-year-old child twenty minutes. See also: DISCRIMINATION AND GENERALIZATION
Hunter judged that the rats were solving the problem
by using the overt orienting attitudes of the pointer, Bibliography
but this did not seem necessary for the raccoons and Carmichael, H. (1954). Walter Samuel Hunter: 18891954. Ameri-
can Journal of Psychology 67, 732734.
children, so Hunter entertained the possibility that
Hunt, J. McV. (1956). Walter Samuel Hunter. Psychological Review
sensory thought might be involved for these sub- 63, 213217.
jects. He supposed that the critical stimulusthe po- Schlosberg, H. (1954). Walter S. Hunter: Pioneer objectivist in psy-
sition of the light or the location of the exit holding chology. Science 120, 441442.
240 HYPNOSIS AND MEMORY
Tinklepaugh, O. L. (1928 ). An experimental study of representa- observed in other forms of amnesia, however, the
tive factors in monkeys. Journal of Comparative Psychology 8, items in question were deeply processed at the time
197.
of encoding. Moreover, posthypnotic amnesia is the
Norman E. Spear only form of amnesia where implicit memories can be
restored to explicit recollection, following adminis-
tration of the prearranged reversibility cue.
HYPNOSIS AND MEMORY

Hypnotic Agnosia
Hypnosis is a social interaction in which one person,
In contrast to posthypnotic amnesia, which is a
called the subject, acts on suggestions from another
disruption in episodic memory, hypnotic agnosia is
person, called the hypnotist, for imaginative experi-
best construed as a disruption in semantic or proce-
ences involving alterations in cognition and voluntary
dural memorythat is, in the subjects ability to ac-
action. Among those individuals who are most highly
cess generic, context-free, declarative, and procedur-
hypnotizable, these alterations in consciousness can
al knowledge. For example, subjects who receive
be associated with subjective conviction bordering on
suggestions that a particular digit will disappear from
delusion, and an experience of involuntariness bor-
their number systems have difficulty when asked to
dering on compulsion.
perform additions in which the offending digit ap-
peared in the problem, intermediate step, or solu-
Posthypnotic Amnesia tion. Similarly, subjects who are told that certain
Upon termination of hypnosis, some subjects find words are meaningless show no priming when these
themselves unable to remember the events and expe- words are presented for free associations. In contrast
riences that transpired while they were hypnotized. to posthypnotic amnesia, hypnotic agnosia has not
This posthypnotic amnesia does not occur unless it been subject to much experimental investigation.
has been specifically suggested to the subject, and the
memories are not restored when hypnosis is merely Hypnotic Hypermnesia
reinduced. Moreover, amnesia can be suggested for A great deal of popular interest in hypnosis stems
events that occurred outside of hypnosis. Thus, post- from its reputation as a means of transcending nor-
hypnotic amnesia is not a form of state-dependent mal limits on human performance. While subjects
memory. However, it is temporary: Upon administra- who receive suggestions for performance enhance-
tion of a prearranged cue, the amnesia is reversed ment often have the impression that their perfor-
and the formerly amnesic subject is able to remember mance has in fact improved, this impression appears
the events perfectly well. Reversibility marks posthyp- to be illusory. This conclusion holds for learning and
notic amnesia as a disruption of memory retrieval, as memory as it does for strength and endurance. For
opposed to encoding or storage, somewhat like the example, the induction of hypnosis and suggestions
temporary retrograde amnesias observed in individu- for enhanced memory add little or nothing to the hy-
als who have suffered concussive blows to the head. permnesia that often occurs on repeated test trials in
The difference is that posthypnotic amnesia is a func- the normal waking state.
tional amnesiaan abnormal amount of forgetting
that is attributable to psychological factors rather Although there is little or no evidence that hyp-
than to brain insult, injury, or disease. Posthypnotic nosis enhances accurate recollection, hypnosis does
amnesia may serve as a laboratory model of the func- appear to increase false recollection, or illusory mem-
tional amnesias associated with hysteria and dissocia- ories. On recognition tests, for example, hypnosis in-
tion, such as psychogenic (dissociative) amnesia, creases the frequency of false alarms and confidence
fugue, and multiple-personality disorder (dissociative levels attached to items endorsed by subjects without
identity disorder). increasing the accuracy of recognition itself. More-
over, perhaps by virtue of their increased suggestibili-
Posthypnotic amnesia impairs conscious recollec- ty, hypnotized subjects may be more vulnerable to
tion, or explicit memory, while leaving implicit mem- postevent misinformation effects. It seems likely that
ory unimpaired. Evidence for spared implicit memo- the suggestive atmosphere of hypnosis interacts with
ry is provided by studies of savings in relearning, the reconstructive nature of memory retrieval to
proactive and retroactive interference, preserved skill create, or enhance, an illusion of remembering.
learning during hypnosis, source amnesia for factual
information learned during hypnosis, repetition
priming in word-stem completion, and semantic Hypnotic Age Regression
priming on free association and category generation The role of illusory experience in hypnosis is dra-
tasks. In contrast to the explicit-implicit dissociation matically revealed in the phenomenon of age regres-
HYPNOSIS AND MEMORY 241
sion. While age-regressed subjects may genuinely be- produced during age regression is likely to be
lieve that they are children again, and behave in a blended with a great deal of false recall, and the ulti-
childlike manner, they do not grow smaller in the mate test is whether the procedure reliably enhances
chair. In terms of psychological changes, there are at memory.
least three different facets of age regression that bear
on questions of hypnosis and memory. First is abla-
tion: To what extent does an age-regressed person
Hypnotic Recovery of Memory in the Court
lose access to the fund of knowledge and repertoire and the Clinic
of skills characteristic of his or her chronological age? Some proponents of clinical hypnosis have criti-
This is really a question about both amnesia and ag- cized studies, such as those described in this entry, on
nosia, because the loss of access extends to semantic the grounds that they test memories that are devoid
and procedural knowledge as well as episodic memo- of affect and personal meaning in the sterile confines
ry. The question of ablation is generally coupled to of the experimental laboratory. They have suggested
the conceptually distinct question of reinstatement: To that different results might be obtained with more
what extent does an age-regressed adult return to ar- lifelike materials and settings. However, this claim
chaic modes of cognitive and emotional functioning rests on an evidentiary base that is almost entirely an-
characteristic of the suggested age? Finally, there is ecdotal and uncorroborated. Testimonials are no sub-
the question of revivification: Can the imagined expe- stitute for evidence. One carefully controlled in vivo
rience of returning to childhood, perhaps coupled study staged a mock organized-crime execution in
with specific suggestions for hypermnesia, enhance front of an introductory criminal justice class (after
memory for childhood events? first insuring that none of the police officers in atten-
dance were carrying their service weapons). Although
Unfortunately for those who would like to use a standard forensic interview technique produced an
hypnosis as a shortcut in developmental research, increase in correct responses compared to controls,
studies employing a wide variety of experimental par- chiefly by reducing the incidence of response omis-
adigmsincluding the Babinski reflex, a characteris- sions, hypnosis added nothing to the experiment. In
tic of infancy, various illusions that show developmen- another controlled but lifelike study, in which subjects
tal trends, memory tests, and a host of tasks derived viewed police training films, a nonhypnotic cogni-
from the developmental theories of Heinz Werner tive interview increased the number of correct mem-
and Jean Piaget, not to mention psychoanalysis ory reports; however, when hypnosis was added to the
have yielded nothing by way of replicable evidence of experiment, performance went down somewhat. Al-
either ablation or reinstatement. Positive findings ei- though police investigators still sometimes turn to
ther have not replicated or have proved to be artifacts hypnosis in an attempt to enhance the memories of
of the demand characteristics of the testing situation. witnesses and victims, most jurisdictions in the United
Age-regressed adults may have the subjectively com- States hold that no special credence should be placed
pelling experience of being children again, and they on hypnotically refreshed memory is of uncertain re-
may appear to behave in a childlike manner, but what liability, and some states prohibit testimony based on
observers see is an imaginative reconstruction of such memories.
childhoodnot a reversion to the genuine article.
Despite the conclusions of laboratory research
Although age regression does not yield a faithful and the courts, some clinical practitioners continue to
reproduction of childlike mental functioning, the use hypnosis to recover repressed or dissociated
subjectively convincing experience of being a child memories of incest, sexual abuse, or other forms of
might produce revivification, the third issue in age- trauma, and hypnosis has even been used to recover
regression research, in a manner analogous to state- memories of prenatal experiences and of alien abduc-
or context-dependent memory. As with hypnotic hy- tions. However, there is little evidence that genuine
permnesia, however, there is no convincing evidence amnesia occurs in these situations: The chief effect of
that revivification actually occurs. The few studies that emotional arousal is to increase, rather than dimin-
have attempted to corroborate the memories report- ish, memory. Moreover, there is no reason to think
ed by age-regressed subjects did yield results favor- that hypnosis can overcome the effects of infantile or
able to hypnosis. However, these studies suffer from childhood amnesia. Most important, the memories
serious methodological flaws that render their posi- recovered in recovered-memory therapy are rarely
tive findings suspect. There may be some memory en- subjected to independent confirmation, so it cannot
hancement produced by hypnotic age regression, as be determined to what extent they are distorted or il-
would be expected with any reinstatement of context, lusory. In the clinic as in the courtroom, uncorrobo-
but age regression is a product of the imagination. As rated memory reports are useless as evidence about
with hypnotic hypermnesia, any accurate memory the historical past. In fact, there is almost no evidence
242 HYPNOSIS AND MEMORY
supporting either the validity of the trauma-memory Bibliography

argument or the efficacy of recovered memory thera- Hilgard, E. R. (1965). Hypnotic susceptibility. New York: Harcourt,
py. Brace, and World.
Kihlstrom, J. F. (1979). Hypnosis and psychopathology: Retrospect
and prospect. Journal of Abnormal Psychology 88 (5), 459473.
(1985). Posthypnotic amnesia and the dissociation of mem-
Conclusion ory. Psychology of Learning and Motivation 19, 131178.
Although hypnosis appears to be incapable of en- (1997). Hypnosis, memory and amnesia. Philosophical
Transactions of the Royal Society: Biological Sciences 352, 1,727
hancing memory, hypnotic procedures can impair
1,732.
memory in at least two different ways. First, by means (1998). Exhumed memory: In S.J. Lynn and K. M. Mc-
of suggestions for posthypnotic amnesia, hypnosis Conkey, eds., Truth in memory. New York: The Guilford Press.
can impair explicit memory for the events and experi- (1998). Hypnosis and the psychological unconscious. In H.
ences that transpired during hypnosisalthough, as S. Friedman, ed., Encyclopedia of mental health, Vol. 2. San
Diego, CA: Academic Press.
with many other forms of amnesia, it appears to spare Kihlstrom, J. F., and Barnhardt, T. M. (1993). The self-regulation
implicit memory. The mechanism for this amnesia of memory: For better and for worse, with and without hypno-
appears to be a division of consciousness, such that sis. In D. M. Wegner and J. W. Pennebaker, eds., Handbook of
the subject is unaware of events that would otherwise mental control. Englewood Cliffs, NJ: Prentice-Hall.
be memorable. Hypnosis appears incapable of ex- Kihlstrom, J. F., and Eich, E. (1994). Altering states of conscious-
ness. In D. Druckman and R. A. Bjork, eds., Learning, remem-
panding awareness, so as to enable subjects to re- bering, and believing: Enhancing performance. Washington, DC:
member things that would otherwise remain forgot- National Academy Press.
ten. However, the social context of hypnosis, Kihlstrom, J. F., and Schacter, D. L. (2000). Functional amnesia.
including widely shared (though false) beliefs about In F. Boller and J. Grafman, eds., Handbook of neuropsychology,
its capacity for memory enhancement (with or without 2nd edition, Vol. 2. Amsterdam: Elsevier.
Nash, M. (1987). What, if anything, is regressed about hypnotic age
age regression), and the suggestive context in which regression? Psychological Bulletin 102, 4252.
hypnosis occurs in the first place, renders the hypno- (2001). The truth and the hype of hypnosis. Scientific Ameri-
tized subject vulnerable to various kinds of distortions can (July), 4755.
in memory. Because the risks of distortion vastly out- Schacter, D. L. (1987). Implicit memory: History and current sta-
tus. Journal of Experimental Psychology: Learning, Memory, and
weigh the chances of obtaining any useful informa-
tion, forensic investigators and clinical practitioners
should avoid hypnosis as a technique for enhancing R. Edward Geiselman
recollection. Revised by John F. Kihlstrom
I
IMPLANTED MEMORIES to the study phase and either produce the target ma-
terials (recall) or discriminate the targets from items
See: FALSE MEMORIES; HYPNOSIS AND MEMORY that were not presented during the study phase (rec-
ognition).
The typical implicit memory experiment includes
a study phase and a retention interval that are similar
IMPLICIT MEMORY to those used in studies of explicit memory. The criti-
Psychological investigations of human memory have cal difference between implicit and explicit memory
traditionally been concerned with conscious recollec- experiments is observed during the test phase. In-
tion or explicit memory for specific facts and epi- stead of being asked to try to remember previously
sodes. Since the 1980s, however, there has been grow- studied items, people are simply instructed to per-
ing interest among the scientific community in a form a perceptual or cognitive task, such as identify-
nonconscious form of memory, referred to as implicit ing a word from a brief exposure or rating how much
memory (Roediger and McDermott, 1993; Schacter, they like a face or an object; no reference is made to
1987; Toth, 2000). Implicit memory does not require the prior study episode. However, some of the test
any explicit recollection of specific episodes. Recent items represent previously exposed target items
experimental investigations have revealed dramatic that is, words, faces, or objects that have been pres-
differences between implicit and explicit memory, ented during the study episode; other test items rep-
and these differences have had a major impact on resent novel or nonstudied words, faces, or objects
psychological theories of the processes and systems that have not been presented during the study phase.
involved in human memory. Implicit memory for a previously studied item is in-
ferred when it can be shown that subjects task perfor-
To understand the nature and significance of im-
mance is influenced by prior exposure to studied
plicit memory, it is necessary first to consider the
items.
types of experimental paradigms that are used to as-
sess both explicit and implicit memory. In a tradition- The hallmark of implicit memory is a facilitation
al explicit memory paradigm, there are three main of performance that is attributable to information ac-
phases: a study episode in which people are exposed quired during a specific episode on a test that does
to a set of target materials, such as a list of words or not require conscious recollection of the episode.
a set of pictures; a retention interval, which typically This facilitation of performance is often referred to as
lasts for several minutes or hours, during which peo- direct or repetition priming. Examples of tests used
ple perform tasks unrelated to the study phase; and to assess priming include stem or fragment comple-
a memory test in which people are asked to think back tion tasks, in which people are asked to complete
243
244 IMPLICIT MEMORY
word stems or fragments (e.g., tab) with the first word cognitive judgments can be biased by information ac-
that comes to mind (e.g., table), and priming is in- quired during specific episodes that the amnesic pa-
ferred from an enhanced tendency to complete the tient cannot remember explicitly (Mayes, 2000).
stems with previously studied words relative to non- Researchers have reported several other popula-
studied words; and perceptual identification tasks, in tion dissociations. For instance, healthy older adults
which people try to identify a word or object from a generally show worse explicit memory than younger
brief (e.g., fifty milliseconds) perceptual exposure, adults. However, older adults often produce the same
and priming is indicated by more accurate identifica- amount of priming on implicit tests (Zacks, Hasher,
tion of recently studied items than of new, nonstudied and Li, 2000). Similarly, patients with depression or
items. schizophrenia have deficits in explicit remembering
Although it has been studied quite extensively, (compared to healthy control subjects) but produce
priming is not the only type of implicit memory. For intact priming.
instance, tasks in which people learn motor or cogni-
tive skills may involve implicit memory, in the sense Pharmacological Dissociations
that skill acquisition does not require explicit recol- The administration of certain drugs (e.g., ben-
lection of a specific previous episode but does depend zodiazepines, a class of drugs used in anesthesia and
on information acquired during such episodes. Simi- to treat anxiety) produce similar dissociations be-
larly, tasks in which people are required to make vari- tween implicit and explicit memory. Scientists have
ous kinds of cognitive judgments, such as judging long known that these drugs produce a temporary
whether a name is famous or how much they like a form of amnesia that wears off as the drug is metabo-
face, can be influenced by implicit memory, in the lized. However, recent research indicates that these
sense that the cognitive judgment may be altered by effects are restricted to explicit memory. In a typical
information acquired during a study episode, even in study, one group of subjects is administered a dose of
the absence of conscious recollection of the episode. a benzodiazepine and another group is given a place-
bo, after which both groups are presented with study
materials. Some time later, memory is tested. Com-
Dissociations Between Implicit and pared to the placebo group, the group given the drug
Explicit Memory exhibits poor explicit memory for the studied materi-
One of the most fascinating aspects of implicit als. In contrast, the two group produce equivalent
memory is that it can be separated or dissociated from performance on implicit tests, such as stem and frag-
explicit memorythat is, there are various experiment completion. Thus, drug-induced amnesia pro-
mental manipulations that affect implicit and explicit duces the same type of dissociation as that produced
memory differently, and there are populations of by organic amnesia: It affects conscious recollection
people for whom explicit memory is impaired while but appears to have little or no effect on nonconscious
implicit memory is spared. influences of memory (Curran, 2000).
Population Dissociations Functional Dissociations

Perhaps the most dramatic dissociation between Experimental manipulations also produce disso-
implicit and explicit memory has been provided by ciations between implicit and explicit memory, pro-
studies of brain-damaged patients with organic amne- viding insight into functional differences between
sia. Amnesic patients are characterized by a severe im- these forms of memory. For example, researchers
pairment of explicit memory for recent events, al- have observed one important dissociation in experi-
though intelligence, perception, and language are ments that have compared the effects of semantic and
relatively normal. Lesions to either medial temporal nonsemantic study tasks on implicit and explicit
or diencephalic brain regions typically produce this memory. Research has well established that perfor-
memory deficit. In contrast, a number of studies have mance on explicit recall and recognition tests is
demonstrated that amnesic patients show intact im- higher following semantic, rather than nonsemantic,
plicit memory. For example, researchers have found study of an item; researchers have called this the le-
repeatedly that amnesic patients show just as much vels of processing effect. For example, when subjects
priming as do normal subjects on stem completion perform a semantic encoding task during the study
and similar tasks, despite their inability to remember phase of an experiment (e.g., rate the pleasantness of
explicitly the target items or the study episode in a word or provide a definition of the word), subse-
which the items were encountered (Shimamura, quent probability of explicitly remembering the word
1993). In addition, it has been demonstrated that am- is much higher than if subjects perform a nonseman-
nesic patients often show normal or near-normal tic encoding task during the study phase (e.g., count
learning of motor and perceptual skills, and that their the number of vowels and consonants in the word). In
IMPLICIT MEMORY 245
contrast, however, research has shown that the mag- served following exposure to unfamiliar materials,
nitude of priming on a stem completion task or per- such as nonsense words or novel patterns and objects,
ceptual identification task is less affected, or even un- that do not have any preexisting representation in
affected, by the same manipulation: Priming effects memory as a unit. Third, priming often shows per-
do not differ significantly following semantic and ceptual specificity (e.g., effects of study modality),
nonsemantic encoding (Roediger and McDermott, which seems incompatible with the idea that priming
1993). is attributable to abstract, amodal representations of
Conversely, the similarity between the physical words and concepts.
(or perceptual) features of the stimuli as presented at Multiple Memory Systems
the time of study and test has a strong effect on many
Neuropsychological and neuroscientific analyses
implicit memory tests but little or no effect on many
have proposed multiple memory systems for implicit
explicit tests. An example is the effect of study modal-
and explicit memory. Because amnesia is typically as-
ity. If some of words on a study list are presented visu-
sociated with damage to the hippocampus and medial
ally and some are presented aurally, later explicit
temporal lobes, scientists believe that these parts of
memory for the words is typically unaffected. Howev-
the brain are necessary for explicit memory. Initially,
er, when memory is tested with the implicit tests of
theories differentiated between two memory systems
stem and fragment completion, study modality has a
(e.g., declarative versus procedural, episodic versus
large impact; visually presented words leading to
semantic) to account for explicit and implicit memo-
more priming than the aurally presented words. Pre-
senting study items as pictures versus words produces ry, respectively. However, as evidence for dissociable
similar matching effects on implicit memory, in which forms of implicit memory mounted, the number of
words produce more priming on word-based implicit proposed systems has increased. A 2000 version of
tests (e.g., stem and fragment completion) and pic- this theory proposes four: episodic memory, semantic
tures produce more priming on implicit tests which memory, the perceptual representation system (PRS),
use pictorial cues (such as picture fragment comple- and procedural memory (Schacter, Wagner, and But-
tion). ter, 2000). Episodic memory stores information about
episodes from ones personal past, enabling the expe-
The foregoing is only a partial list of dissociations rience of recollection. Semantic memory stores gen-
between implicit and explicit memory. In addition, eral knowledge about the world, including facts, con-
similarities between implicit and explicit memory ceptual information, and vocabulary. The PRS is a
have also been observed; this is to be expected, be- perceptual memory systems that processes informa-
cause both are forms of memory and hence presum- tion about the form and structure of words and ob-
ably share some common characteristics. However, jects independently of their semantic content. Finally,
the differences between the two are most revealing procedural memory represents knowledge of cogni-
theoretically and have led to a variety of proposals tive and motor skills. Explicit memory is assumed to
concerning the nature of implicit memory. be a product of the episodic memory system, whereas
various forms of implicit memory are produced by the
other systems.
Theoretical Accounts of Implicit Memory
Some explanations of implicit memory focus Population and pharmacological dissociations
heavily on the sorts of dissociations described above argue for the existence of multiple memory systems
and emphasize differences between implicit and ex- in the brain. For example, amnesia produces deficits
plicit memory. Other explanations seek to elucidate on explicit memory but typically not on verbal implic-
common principles that may underlie both implicit it memory tests (such as stem and fragment comple-
and explicit memory phenomena. tion) or on tests of skill learning, indicating that am-
nesia damages the episodic memory system but not
Activation View the other systems. Even stronger support for the mul-
An early theoretical view held that implicit mem- tiple-memory-systems view comes from reports of
ory is attributable to the temporary activation of pre- double dissociations, dissociations in which two differ-
existing units or nodes in memory: Exposure to a ent patient groups (with damage to different parts of
word or object automatically activates a memory rep- the brain) exhibit complementary dissociative pat-
resentation of it, and this activation subsides rapidly. terns on implicit and explicit memory tests. For ex-
Although able to accommodate some experimental ample, amnesic patients (with damage to medial-
results, this general idea has difficulty accounting for temporal regions of the brain) have disrupted memo-
other important results. First, priming can be surpris- ry on explicit tests but not on implicit tests, as noted.
ingly long-lived under certain conditions lasting The opposite dissociation occurs in patients with oc-
weeks and months. Second, priming has been ob- cipital-lobe lesions, who exhibit preserved explicit
246 IMPLICIT MEMORY
memory coupled with deficits in implicit memory for perceptual) but produce normal levels of priming on
visual-perceptual information (Gabrieli et al., 1995). conceptual as well as perceptual implicit tests. Like-
This provides strong support for the view that brain wise, older and younger adults generally produce
systems mediating performance on these two types of equivalent levels of conceptual and perceptual prim-
tests differ. Similar dissociative patterns support dis- ing, even though older adults perform worse on tests
tinctions among the other imputed memory systems of explicit memory. These dissociations are difficult
(Schacter et al., 2000). to explain in terms of perceptual versus conceptual
processing.
Transfer-Appropriate Processing
Although population and pharmacological disso- Component-Processes Approach and
ciations provide strong support for distinct memory Evidence from Neuroimaging
systems, functional dissociations have often been in- The complementary successes of multiple memo-
terpreted in terms of the divergent processing re- ry systems view and the transfer-appropriate process-
quirements of different memory tests within a single ing approach has produced the view that emphasizes
memory system. The general idea is that dissociations multiple forms of priming and attempts to articulate
between implicit and explicit memory are special the component processes that mediate performance
cases of the general principles of transfer-appropriate on various memory tasks (Foster and Jelicic, 1999).
processing and encoding specificity, which hold that Neuroimaging techniques, such as positron emission
memory performance is best when encoding and re- tomography (PET) and functional magnetic reso-
trieval processes match. Specifically, it is held that nance imaging (fMRI), play a critical role in supple-
most standard explicit memory tests require a good menting traditional approaches. These techniques at-
deal of conceptual processing: semantically based, tempt to determine neural regions involved in
subject-initiated attempts to recollect the study epi- explicit and implicit memory tests. One result is the
sode. By contrast, performance on such implicit substantial degree of overlap in the active brain re-
memory tests as word completion and perceptual gions during explicit and implicit retrieval tasks. De-
identification is held to be largely dependent on per- spite this overlap, there are important differences in
ceptual, or data-driven, processingprocessing that the patterns of brain activation. In particular, explicit
is determined largely by the physical characteristics of retrieval relies heavily on anterior frontal (especially
test cues. It thus follows that explicit memorybut right prefrontal) lobe as well as the medial-temporal
not primingshould benefit from semantic study regions, including the hippocampus. Scientists be-
processing (which researchers believe support con- lieve the activity in frontal lobe reflects an explicit re-
ceptually driven processing) more than from nonse- trieval mode in which the individual intentionally
mantic study processing, whereas priming should be tries to retrieve information about past events and is
strongly dependent on matching of surface features generally oriented toward the past. They believe the
between study and test (Jacoby, 1983; Roediger and medial temporal activity contributes to the recollec-
McDermott, 1993). However, it is possible to devise tive experience itself when a memory is successfully
implicit tests that entail conceptual processing, and retrieved (Nyberg and Cabeza, 2000). Another find-
the transfer-appropriate processing view has led to ing is that priming on implicit tests is often associated
predictions and corresponding demonstrations of with decreased activity in certain regions of the brain,
dissociations between implicit tests, by contrasting im- which may reflect a reduction in processing demands
plicit tasks that draw primarily on perceptual process- when a stimulus is processed a second time (Schacter
ing with implicit tasks that draw primarily on concep- and Badgaiyan, 2001 ). The brain areas involved de-
tual processing. pend on whether the implicit test is perceptual or
conceptual. Perceptual tests, such as stem or fragment
The transfer-appropriate processing view is par- completion, produce decreased processing in visual
simonious and has enjoyed great success in account- cortex (in the posterior occipital lobe). In contrast,
ing for functional dissociations between explicit and conceptual implicit tests produce decreased activity in
implicit tests, and among implicit tests of different the inferior frontal lobe and the mid-temporal lobe.
types (i.e., perceptual and conceptual). Thus, the dis- In general, these results indicate that the same neural
tinction between perceptual and conceptual process- substrates required for the initial (perceptual or con-
ing is likely an important aspect of any explanation ceptual) processing are reengaged at the time of test
of implicit memory phenomena. However, it does not and exhibit the effects of the initial processing by
provide a complete account because it does not natu- their subsequent reduced activity.
rally account for population and pharmacological dis-
sociations. In particular, it has been generally found See also: AMNESIA, ORGANIC; CODING PROCESSES:
that both organic and drug-induced amnesia disrupt LEVELS OF PROCESSING; DECLARATIVE
performance on explicit tests (whether conceptual or MEMORY; EPISODIC MEMORY; FRONTAL LOBES
IMPRINTING 247
AND EPISODIC MEMORY; GUIDE TO THE The phenomenon of filial imprinting was described
ANATOMY OF THE BRAIN; MOTOR SKILL as early as 1518 by Sir Thomas More in his Utopia.
LEARNING; PREFRONTAL CORTEX AND MEMORY However, imprinting was investigated experimentally
IN PRIMATES; SEMANTIC MEMORY: COGNITIVE much later, by D. A. Spalding in 1873 and by Oscar
ASPECTS; SEMANTIC MEMORY: Heinroth in 1911. Konrad Lorenz, who gave the phe-
NEUROBIOLOGICAL PERSPECTIVE
nomenon its name, subsequently provided a detailed
Bibliography description of imprinting in a number of bird species
Curran, H. V. (2000). Psychopharmacological approaches to in an influential work published in 1935.
human memory. In M. S. Gazzaniga, ed., The new cognitive
neurosciences, 2nd edition. Cambridge, MA: MIT Press.
Foster, J. K., and Jelicic, M., eds. (1999). Memory: Systems, process, Filial Imprinting
or function? New York: Oxford University Press.
Gabrieli, J. D. E. et al. (1995). Double dissociation between memo-
Although filial imprinting may occur in mammals
ry systems underlying explicit and implicit memory in the (Sluckin, 1972), it has been studied mostly in preco-
human brain. Psychological Science 6, 7682. cial birds such as ducklings and chicks. These birds
Jacoby, L. L. (1983). Remembering the data: Analyzing interactive can move about shortly after hatching, and they ap-
processes in reading. Journal of Verbal Learning and Verbal Be- proach and follow an object to which they are ex-
havior 22, 485508.
Mayes, A. R. (2000). Selective memory disorders. In E. Tulving and
posed. In a natural situation the first object the young
F. I. M. Craik, eds., The Oxford handbook of memory. New York: bird encounters usually is its mother. In the absence
Oxford University Press. of the mother, inanimate mother surrogates are effec-
Nyberg, L., and Cabeza, R. (2000). Brain imaging of memory. In tive in eliciting filial behavior (Bateson, 1966; Horn,
E. Tulving and F. I. M. Craik, eds., The Oxford handbook of mem- 1985; Sluckin, 1972). When the chick or duckling is
ory. New York: Oxford University Press.
Roediger, H. L., and McDermott, K. B. (1993). Implicit memory
close to an appropriate object, it will attempt to snug-
in normal human subjects. In F. Boller and J. Grafman, eds., gle up to it, frequently emitting soft twitters. Initially
Handbook of neuropsychology. Amsterdam: Elsevier. the young bird approaches a wide range of objects,
Schacter, D. L. (1987). Implicit memory: History and current sta- though some are more attractive to it than others.
tus. Journal of Experimental Psychology: Learning, Memory, and After the bird has been exposed to one object long
Schacter, D. L., and Badgaiyan, R. D. (2001). Neuroimaging of
enough, it remains close to this object and may run
priming: New perspectives on implicit and explicit memory. away from novel ones. If the familiar object is re-
Current Directions in Psychological Science 10, 14. moved, the bird becomes restless and emits shrill
Schacter, D. L., Wagner, A. D., and Buckner, R. L. (2000). Memory calls. When given a choice between the familiar stimu-
systems of 1999. In E. Tulving and F. I. M. Craik, eds., The lus and a novel one, the bird shows a preference for
Oxford handbook of memory. New York: Oxford University Press.
Shimamura, A. P. (1993). Neuropsychological analyses of implicit
the familiar stimulus. Thus, filial imprinting refers to
memory: History, methodology, and theoretical interpreta- the acquisition of a social preference and not just an
tions. In P. Graf and M. E. J. Masson, eds., Implicit memory: increase in following (Sluckin, 1972).
New directions in cognition, development, and neuropsychology.
Hillsdale, NJ: Erlbaum.
Toth, J. P. (2000). Nonconscious forms of human memory. In E. Conditions for Imprinting
Tulving and F. I. M. Craik, eds., The Oxford handbook of memo-
ry. New York: Oxford University Press. To study visual imprinting in the laboratory,
Zacks, R. T., Hasher, L., and Li, K. Z. H. (2000). Human memory. chicks or ducklings may be hatched in darkness and
In F. I. M. Craik and T. A. Salthouse, eds., The handbook of exposed for a period of one to two hours to a conspic-
aging and cognition, 2nd edition. Mahwah, NJ: Erlbaum. uous object when they are about twenty-four hours
Daniel L. Schacter old. The animals are then returned to a dark incuba-
Revised by Neil W. Mulligan tor and kept there until their preferences are tested
by exposing them to the familiar object and a novel
object. A widely used measure of filial preference is
approach to the familiar object relative to approach
IMPRINTING to a novel object. Another measure takes advantage
Imprinting is the learning process through which the of the fact that an imprinted chick emits distress calls
social preferences of animals of certain species be- in the presence of a novel object and does not do so,
come restricted to a particular object or class of ob- or does so less frequently, in the presence of the fa-
jects. A distinction is made between filial and sexual miliar.
imprinting. Filial imprinting is involved in the forma- The effectiveness of imprinting stimuli varies. For
tion, in young animals, of an attachment to, and a example, young ducklings approach and follow ob-
preference for, the parent, parent surrogate, or sib- jects larger than a matchbox, but peck at smaller ob-
lings. Sexual imprinting is involved in the formation of jects. For chicks, red and blue objects are more effec-
mating preferences that are expressed in later life. tive imprinting stimuli than yellow and green objects.
248 IMPRINTING
Movement, brightness, contrast, and sound all en- Auditory Imprinting

hance the attractiveness of an imprinting stimulus.
In the natural context auditory stimuli play an
important role in the formation of filial preferences
Imprinting and Learning (Gottlieb, 1971). Auditory preferences may be formed
Filial imprinting has been regarded as different in the same way as visual preferences (i.e., learning as
from other forms of learning because it proceeds a result of exposure). However, preferences resulting
without any obvious reinforcement such as food or from exposure to an auditory stimulus only, whether
warmth (Bolhuis, De Vos, and Kruijt, 1990). However, before or after the birds have hatched, are relatively
an imprinting object may itself be a reinforcer, that weak and short-lived. Such preferences can be
is, a stimulus that an animal finds rewarding. Just as strengthened when the young bird is simultaneously
a rat learns to press a pedal to receive a reward of exposed to an auditory stimulus and a visual stimulus
food, so a visually naive chick learns to press a pedal during auditory training, just as exposure to auditory
to see an imprinting object. When chicks are exposed stimuli can improve visual imprinting.
to two imprinting stimuli simultaneously, they learn
more about the individual stimuli than when they are Sexual Imprinting
exposed to the stimuli sequentially, or to only one
stimulus. This so-called within-event learning has also It might seem that one of the consequences of fil-
been found in conditioning paradigms in rats and hu- ial imprinting would be the determination of adult
mans (Bolhuis and Honey, 1998). The ability to learn sexual preferences, but research suggests that filial
and remember the characteristics of objects to which imprinting and sexual imprinting are two separate
an animal is exposed may be a common form of learn- (although perhaps partially overlapping) processes.
ing. Not only is the time of expression of the preferences
different, but so is the period during which experi-
ence affects preferences. Sexual preferences continue
Reversibility and Sensitive Periods to be affected by experience up to the time of mating.
Imprinting was thought to be irreversible and to Furthermore, filial preferences may be formed after
occur during a sensitive period (or critical period). Nu- a relatively short period of exposure to an object. In
merous studies have demonstrated that filial prefer- contrast, sexual preferences develop as the result of
ences can in fact be reversed when the original object a long period of exposure to and social interaction
is removed and the animal is exposed to a novel ob- with the parents as well as the siblings. Normally, sex-
ject. Evidence suggests that there is a difference be- ual imprinting ensures that the bird will mate with a
tween the memory of the first stimulus and that of member of its own strain or species. When the young
subsequent stimuli to which the animal is exposed. bird is cross-fosteredthat is, reared with adults of a
Under certain circumstances the preference for the different speciesit develops a sexual preference for
first object may return (Bolhuis, 1991). The ability to the foster species. In Japanese quail (Coturnix coturnix
form filial attachments has been shown to depend on japonica) and domestic chickens (Gallus gallus domes-
both developmental age and time since hatching. The ticus), mating preferences are for individual members
ability to imprint is related to the development of the of the opposite sex that are different, but not too dif-
animals sensorimotor abilities. The sensitive period ferent, from individuals with which the young bird
is brought to an end by the learning experience (im- was reared (Bateson, 1978).
printing) itself: Once the bird has formed a prefer-
ence for a particular object, it avoids novel objects.
Consequently it tends not to be exposed to them for Predispositions
long and so may learn little about them. When the Research shows that filial preferences are formed
bird is left in its cage, it may form an attachment to not only as a result of learning through exposure but
features of its rearing environment. Rearing the bird that they are also influenced by a specific predisposi-
in a visually impoverished environment (in darkness tion (Bolhuis, 1991; Horn, 1985; Johnson and Bol-
or deprived of patterned light) extends the period huis, 2000). This predisposition may be measured in
during which it forms an attachment to a conspicuous the laboratory by giving chicks a choice between a ro-
object (Bateson, 1966). The sensitive period pertains tating stuffed jungle fowl and (for instance) a rotating
to filial attachment and may relate to the link formed red box. Under some conditions the two stimuli are
between the (neural) representation of the imprinted equally attractive. But if the young chick is given a
object and approach behavior. Although the forma- certain amount of nonspecific experience, such as
tion of this link may have a sensitive period, there is being handled and allowed to run, the chick prefers
no reason to suppose that the learning and recogni- the fowl to the box when tested twenty-four hours
tion processes have one. later. In order to be effective, this nonspecific experi-
IMPRINTING 249
ence must occur within a sensitive period (about twen- Figure 1

ty to forty hours after hatching). It appears that the
target stimuli of the predisposition are in the head
and neck region but are not species-specific. Once the
predisposition has developed, it does not function as
a filter that prevents the chick from learning about
objects that do not resemble conspecifics; such chicks
can learn about other objects by being exposed to
them. Thus, it is likely that the mechanisms underly-
ing the predisposition and those underlying learning
influence behavior independently.
Neural Mechanisms of Filial Imprinting

The neural basis of the recognition memory un-
derlying filial imprinting has been studied most ex-
tensively in the domestic chick (Horn, 1985, 1998,
2000). When dark-reared chicks are trained by expos-
ing them to an imprinting object for approximately
one to two hours, metabolic changes occur in the dor-
sal part of the cerebral hemispheres. Specifically,
there is an increase in the incorporation of radioac-
tively tagged uracil into RNA in this brain region of
trained chicks compared with control chicks (dark-
reared chicks or chicks that have merely been ex-
posed to overhead light). Several reasons support the
idea that the biochemical changes are related to
learning rather than to various side effects of training
(e.g., to differences in movement, excitement, senso-
ry stimulation between the trained chicks and their
controls): when visual input is restricted to one hemi-
sphere, incorporation of radioactive uracil into RNA
is higher in the trained than in the untrained hemi-
sphere; the amount incorporated is related to how
much the chicks learn and not to various other mea-
sures of behavior; the increase is not related to short-
term effects of sensory stimulation.
Outline drawing of the chick brain. The vertical lines AA above

Neural Changes Localized to Specific Brain and below the drawing of the lateral aspect a indicate the plane
Regions of the transverse section outline b of the brain. IMHV extends
Imprinting leads to changes in the incorporation approximately 2.5 millimeters from front to back of the cerebral
hemisphere.
of radioactive uracil into RNA in a restricted brain re-
gion, the intermediate and medial part of the hyper-
striatum ventrale (IMHV), a sheet of cells in the cere-
bral hemispheres (see Figure 1). Further evidence region. At the end of the period of electrical stimula-
that the region is crucially involved in learning is the tion, the chicks were shown two lights, one flashing at
following: destruction of IMHV before training pre- the rate of 4.5 per second and the other at 1.5 per sec-
vents imprinting; if the region is destroyed immedi- ond. If the IMHV region had been stimulated at 4.5
ately after training, chicks do not prefer the training trains per second, the chicks preferred the light flash-
object, though for chicks with lesions of IMHV an im- ing at this frequency. In contrast, chicks that had re-
printing object still elicits approach behavior but the ceived electrical stimulation of IMHV at the rate of
chicks appear incapable of learning its characteristics; 1.5 trains per second preferred the light flashing at
it is possible to bias chicks preferences by delivering this rate. Electrical stimulation of two visual receiving
trains of short pulses of electric current to IMHV areas of the forebrain did not influence the chicks
through electrodes that have been implanted into the preferences. Taken together, these results strongly
250 IMPRINTING
suggest that the IMHV region is involved in the rec- Cerebral Asymmetry and Imprinting
ognition memory of imprinting, probably storing in-
Studies in which the left or right IMHV region
formation.
has been surgically damaged suggest that, in accor-
dance with the data on synaptic changes, the left
Neuronal Mechanisms of Memory IMHV functions as a long-term store. During the first
day after imprinting, however, another memory sys-
Imprinting leads to changes in the structure of tem, referred to as S, is established outside IMHV. S
synapses in IMHV, in particular to an increase in the is formed under the influence of the right IMHV: If
area of thickened membrane on the postsynaptic side this region is absent, S is not formed. The right
of certain synapses. This area of membrane is known IMHV perhaps transfers information to S, a possibili-
as the postsynaptic density. The changes are restrict- ty that implies a dynamic element to memory forma-
ed to synapses on the spines of dendrites (axospinous tion. Electrophysiological evidence (Horn, Nicol, and
synapses) and are found in the left but not in the right Brown, 2001) demonstrates that imprinting leads to
IMHV. Certain spine synapses in the mammalian changes in neuronal responsiveness in the right
brain are excitatory and possess, in the postsynaptic IMHV that are similar to those in the left. These re-
density, receptors for the excitatory neurotransmitter sults are puzzling because various morphological and
l-glutamate. The increased area of the postsynaptic biochemical measures indicate differences in the ef-
density of axospinous synapses seems to imply that fect of learning on the two brain regions. To reconcile
imprinting leads to an increase in the number of re- these findings it has been suggested that both regions
ceptors for l-glutamate: After chicks have been are similarly affected by learning, but that other func-
trained, there is an increase in the number of a cer- tions of the right IMHV (e.g., in the formation of S)
tain type of receptors for l-glutamate in the left bring about additional structural and biochemical
IMHV, but not in the right. The increased number of changes that obscure these effects (Horn, 1998). The
receptors is related to the amount the chicks have right and left IMHV regions are not directly intercon-
learned about the imprinting object. One conse- nected, so it is likely that they and S are independent
quence of this change may be that, after training, the stores working in parallel. Thus, even in the case of
release of a given amount of l-glutamate from a pre- the recognition memory of imprinting, more than
synaptic ending may exert a greater excitatory action one memory system is formed. Further evidence that
on the postsynaptic cell than before training. That is, several memory systems exist in the young chick is
as many hypotheses have suggested, learning leads to that chicks with lesions of IMHV (placed before S has
an increased efficacy of synaptic transmission. been formed), while being severely impaired in their
The particular l-glutamate receptors shown to be ability to imprint, are nevertheless able to learn cer-
affected by imprinting are those of the N-methyl-D- tain other tasks. Multiple memory systems therefore
aspartate (NMDA) variety. In mammals, these recep- not only are found in mammalian brains but also may
tors are involved in other forms of synaptic plasticity. be a fundamental part of the design of the vertebrate
Thus the cellular mechanisms of synaptic plasticity brain.
may be similar in diverse systems though the func- On the basis of its connections and developmen-
tions of the synaptic change may be different: In cir- tal history IMHV has been compared to the prefron-
cuits involved in learning, the synaptic changes may tal and cingulate areas of the primate cerebral cortex
play a part in information storage. In other systems, (Horn, 1985). Evidence from humans suggests that
these changes may be a response to either physical these regions play an important role in the organiza-
damage or physiological dysfunction. Exposure to an tion of memory (Janowsky, Shimamura, and Squire,
imprinting stimulus for approximately two hours 1989; Tulving et al., 1994).
leads to a trebling in the proportion of neurons in
IMHV responding to the imprinting stimulus com- See also: BIRDSONG LEARNING; GLUTAMATE
pared with the proportion responding before train- RECEPTORS AND THEIR CHARACTERIZATION;
ing. Some neurons in IMHV respond in a highly se- REINFORCEMENT
lective way to the imprinted stimulus and so critically Bibliography
have the properties of neurons postulated by Donald Bateson, P. P. G. (1966). The characteristics and context of im-
Hebb (1949) as forming part of a memory trace. The printing. Biological Reviews 41, 177220.
increase in responsiveness does not develop linearly (1978). Sexual imprinting and optimal outbreeding. Nature
but, remarkably, develops in contrasting phases dur- 273, 659660.
Bolhuis, J. J. (1991). Mechanisms of avian imprinting: A review. Bi-
ing and in the hours that follow training. These find-
ological Reviews 66, 303345.
ings challenge models of memory based on synaptic Bolhuis, J. J., De Vos, G. J., and Kruijt, J. P. (1990). Filial imprint-
strengthening to asymptote (Horn, Nicol, and Brown, ing and associative learning. Quarterly Journal of Experimental
2001). Psychology 42B, 313329.
INDIVIDUAL DIFFERENCES IN LEARNING AND MEMORY 251
Bolhuis, J. J., and Honey, R. C. (1998). Imprinting, learning, and ent strategies for learning. Fourth, people differ in
development: From behaviour to brain and back. Trends in the retrieval efficiency with which they can summon
Neurosciences 21, 306311.
Gottlieb, G. (1971). Development of species identification in birds. Chi-
information from more permanent, long-term mem-
cago: University of Chicago Press. ory. To these can be added one negative conclu-
Heinroth, O. (1911). Beitrge zur Biologie, namentlich Ethologie sion: There do not appear to be consistent sex differ-
und Psychologie der Anatiden. In Verhandlungen des 5, 589 ences in learning and memory ability, although
702. Berlin: Internationaler Ornithologischer Kongress. women and men may choose to learn different infor-
Hebb, D. O. (1949). The organization of behavior. New York: Wiley.
Horn, G. (1985). Memory, imprinting, and the brain. Oxford: Claren-
mation (which affects their knowledge base). This
don Press. entry examines these four main conclusions in more
(1998). Visual imprinting and the neural mechanisms of detail.
recognition memory. Trends in Neurosciences 21, 300305.
(2000). In memory. In J. J. Bolhuis, ed., Brain, perception,
Memory: Advances in cognitive neuroscience. Oxford: Oxford Knowledge Differences
University Press.
Horn, G., Nicol, A. U., and Brown, M. W. (2001). Tracking memo- Standard intelligence tests (Wechsler, 1945,
rys trace. Proceedings of the National Academy of Sciences of the 1958) measure two aspects of memory, the first being
United States of America 98, 5,2825,287. general knowledge and vocabulary. People differ in
Janowsky, J. S., Shimamura, A. P., and Squire, L. R. (1989). Source
their breadth and depth of knowledge. Consider the
and impairment in patients with frontal lobe lesions. Neuro-
psychologia 27, 1,0431,056. paradox of the expert: A simple theory might claim
Johnson, M. H., and Bolhuis, J. J. (2000). Predispositions in per- that forgetting is caused by interference among relat-
ceptual and cognitive development. In J. J. Bolhuis, ed., ed concepts in memory. This would imply that some-
Brain, perception, memory. Advances in cognitive neuroscience. Ox- one who knows more about a topic should be more
ford: Oxford University Press.
subject to forgetting in that domain. But if this were
Lorenz, K. (1935). Der Kumpan in der Umwelt des Vogels. Journal
fr Ornithologie 83, 137213, 289413. true, how would one ever become an expert in a do-
Sluckin, W. (1972). Imprinting and early learning. London: Methuen. main? One resolution is that people integrate their
Spalding, D. A. (1873). Instinct, with original observations on knowledge so that related ideas are joined, support-
young animals. Macmillans Magazine 27, 282293. ing rather than competing with each other (Smith,
Tulving, E., Kapur, S., Markowitsch, H. J., Craik, F. I. M., Habib,
Adams, and Schorr, 1978).
R. and Houle, S. (1994). Neuroanatomical correlates of re-
trieval in episodic memory: Auditory sentence recognition. Indeed, those with high knowledge learn and re-
Proceedings of the National Academy of Sciences of the United States tain new facts in that domain more easily than do
of America 91, 2,0122,015.
those with low knowledge (Voss, Vesonder, and
Johan J. Bolhuis Spilich, 1980). Furthermore, experts seem especially
G. Horn superior in remembering the important information.
Popular memory metaphors of libraries or ware-
houses do not fit comfortably here. Instead, think of
memory as a scaffolding: The more memories that
INDIVIDUAL DIFFERENCES IN are attached to the structure, the more places there
LEARNING AND MEMORY are to attach new memories. The scaffolding may
even guide people to where it would be best to attach
Individual differences in learning and memory abili-
each new memory.
ties have fascinated people since they began thinking
about how their minds work. In discussing his wax People differ not only in what they know and how
metaphor of memory, Plato (1953) noted that memo- much they know, but also in how that knowledge is or-
ries made of pure and clear [wax] . . . easily learn and ganized (Coltheart and Evans, 1981). Because retriev-
easily retain, whereas those made of muddy and of al relies heavily on the association between facts and
impure wax [have] . . . a corresponding defect in the ideas, organization influences how people retrieve
mind. Plato realized that people differ in what they their knowledge. An individual who has two facts di-
learn and remember and in how well they do both. rectly connected in memory should be able to get
This is certainly true at the extremes, but how rele- from one to the other more quickly than someone
vant is it over the normal range of memory abilities? who must go through multiple way stations. Part of
coming to understand a domain better is reorganiz-
The psychological research supports four main
ing ones knowledge more appropriately.
conclusions about individual differences in learning
and memory (Bors and MacLeod, 1996). First, people Each person is unique in part because of what he
differ in what they know, their knowledge base. Sec- or she knows, both in terms of autobiographical
ond, people differ in their working memory capacity, knowledge and general world knowledge. Thus, con-
the ability to hold information in consciously accessi- tent of memories is one major source of individual
ble memory. Third, people possess and invoke differ- differences. The other three differences all relate
252 INDIVIDUAL DIFFERENCES IN LEARNING AND MEMORY
more to how people acquire, store, transform, and (Gagn, 1967). But people also differ in their learn-
use that knowledgethat is, the cognitive processes. ing styles, in which processes they use, and in when
they use them (Sternberg and Zhang, 2001).
Capacity Differences Studies of learning style suggest that, at a global
level, learners emphasize either overall comprehen-
Probably the best known individual differences
sion or specific detail, appearing to be either conclu-
dimension in learning and memory is that of the ca-
sion oriented or description oriented (Schmeck,
pacity of conscious, working memory. This is the
1983). Those who emphasize overall comprehension
other aspect of memory that is directly measured in
engage in deeper processing; those concerned with
standard intelligence tests: memory span (Dempster,
specific detail focus more on surface processing. Or-
1981). Individuals have a sharp limitation on how
dinarily such a strategic difference favors those who
much they can consciously think about and retain at
undertake deeper processing. Thus, note taking in a
one time. Even in the normal range of intelligence,
classroom induces deeper processing and better re-
not everyones span is the same. What causes these
tention of important information.
differences? There seem to be two main mechanisms
underlying working memory spanthe ability to Another illustration is individual differences in
identify the specific elements to be held and the abili- reported imagery ability (Richardson, 1999). People
ty to retain their order (Humphreys et al., 1983). vary in whether they use more language-based or pic-
Speed of scanning through the information held in ture-based strategies to learn and remember. Indeed,
working memory does not seem to differ reliably research suggests that visual memory can be quite in-
across individuals (Hunt, 1978), although speed is af- dependent of verbal memory. People who recognize
fected by age (Salthouse, 1996). faces or pictures well will not necessarily remember
It is tempting to view span differences as irrele- what they read better than will people with poorer vi-
vant to the real world. After all, how often do peo- sual memory. Yet the widely held belief that some
ple listen to a string of items and then repeat them people learn best visually and others learn best audi-
back? In fact, holding information in working memo- torily has not been strongly supported experimental-
ry is something people do constantly and rely upon ly. One can make sense of this by realizing that verbal
heavily. Meredyth Daneman and Patricia Carpenter skills need not be auditory.
(1980; Hannon and Daneman, 2001; Miyake, 2001) There are many other techniques and strategies
showed that span differences have powerful implica- for learning. Thus it seems quite reasonable to sup-
tions for how successfully people read. Those with pose that different individuals learn most effectively
larger spans in reading (or in listening) show better using different strategies. However, evidence to sup-
comprehension of what they read: They can hold pre- port this intuition has been notoriously difficult to ob-
vious sentences in mind more easily, and thus assem- tain. Usually a strategy that improves one persons
ble the overall meaning more effectively. This has led learning also improves another persons learning.
Marcel Just and Carpenter (1992) to a theory of com- What may differ, then, are the choices that people
prehension based on working memory capacity. make, a metacognitive issue as Christian Schunn and
Individual differences in working memory capac- Lynne Reder (2001) argue, relating successful choice
ity do not influence only language. These effects are to working memory capacity. How people select opti-
also apparent in spatial processing (Shah and Miyake, mal process(es) from their repertoire for a particular
1996) and in enumeration (Tuholski, Engle, and learning situation may be one of the most critical dif-
Baylis, 2001). Human ability to acquire and execute ferences of all.
skills (Ackerman, 1987) is also a function of working
memory capacity (Perlow, Jattuso, and Moore, 1997).
Individual differences in working memory and atten-
Retrieval Speed Differences
tion, in fact, appear to be intimately linked (Kane, Retrieval time for information in working memo-
Bleckley, Conway, and Engle, 2001). ry is not a reliable source of individual differences.
Given the sharp capacity limitation, the small content
of working memory is easily searched. But the same
Learning Strategy Differences is not true of long-term memory, where all of a per-
There are many different ways to learn, from rote sons knowledge is stored. If the average person
repetition to complex mnemonics. All learning in- knows upward of 10 billion facts, as some scientists
volves recoding: Information must be transformed speculate, how can he or she find any one of them
from its perceived form into a form suitable for re- quickly? Even searching at the impossibly fast rate of
membering. It is well established that people differ in 1 millisecond per fact, it would take months to find
their speed and in their efficiency of recoding any single fact.
INDIVIDUAL DIFFERENCES IN LEARNING AND MEMORY 253
Extensive research (Hunt, 1978) has shown that Daneman, M., and Carpenter, P. A. (1980). Individual differences
different people retrieve information from long-term in working memory and reading. Journal of Verbal Learning
and Verbal Behavior 19, 450466.
memory at different rates. Consider a very simple re-
Dempster, F. N. (1981). Memory span: Sources of individual and
trieval: It takes longer to determine whether two let- developmental differences. Psychological Bulletin 89, 63100.
ters have the same name for Aa than for aa or AA. Pre- Gagn, R. M., ed. (1967). Learning and individual differences. Colum-
sumably this is because long-term memory access is bus, OH: Charles E. Merrill.
required only for Aa, where the two are not physically Hannon, B., and Daneman, M. (2001). A new tool for measuring
identical. On average, this retrieval time is about 80 and understanding individual differences in the component
processes of reading comprehension. Journal of Educational
milliseconds, but high-ability individuals are faster
than low-ability individuals by about 30 milliseconds Humphreys, M. S., Lynch, M. J., Revelle, W., and Hall, J. W.
or more. A naive critic might say, But this is a tiny (1983). Individual differences in short-term memory. In R. F.
difference. Consider reading, however. If one were Dillon and R. R. Schmeck, eds., Individual differences in cogni-
to lose 30 milliseconds for every letter read in this tion, Vol. 1. New York: Academic Press.
entry, this time would quickly add up. And that is just Hunt, E. (1978). Mechanics of verbal ability. Psychological Review
85, 109130.
for such highly learned facts as the letters of the al-
Just, M. A., and Carpenter, P. A. (1992). A capacity theory of com-
phabet; the problem must be vastly greater for more prehension: Individual differences in working memory. Psy-
complex and unfamiliar types of knowledge. If the el- chological Review 99, 122149.
ementary processes are not executed as efficiently by Kane, M. J., Bleckley, M. K., Conway, A. R. A., and Engle, R. W.
one individual as by another, the cost for learning and (2001). A controlled-attention view of working-memory ca-
memory as a whole can be large. Indeed, it is quite pacity. Journal of Experimental Psychology: General 130, 169
183.
clear that working memory capacity also affects re-
Miyake, A. (2001). Individual differences in working memory: In-
trieval from long-term memory (Rosen and Engle, troduction to the special section. Journal of Experimental Psy-
1997). chology: General 130, 163168.
Perlow, R., Jattuso, M., and Moore, D. D. W. (1997). Role of verbal
working memory in complex skill acquisition. Human Perfor-
Conclusion mance 10, 283302.
The goal of this brief sketch has been to localize Plato. (1953). The dialogues of Plato, 4th edition, trans. B. Jowett.
Oxford, UK: Clarendon Press.
four of the more crucial individual differences in
Richardson, John T. E. (1999). Imagery. Hove, UK: Psychology
learning and memory and to provide some of the evi- Press.
dence for these differences. Of course, there are many Rosen, V. M., and Engle, R. W. (1997). The role of working memo-
other differences in how people learn and remember, ry capacity in retrieval. Journal of Experimental Psychology: Gen-
but these are usually more isolated and less character- eral 126, 211227.
istic of the memory system as a whole. Scientists hope Salthouse, T. A. (1996). The processing-speed theory of adult age
differences in cognition. Psychological Review 103, 403428.
to find ways of recognizing these differences in educa-
Schmeck, R. R. (1983). Learning styles of college students. In R.
tional systems around the world in order to realize the F. Dillon and R. R. Schmeck, eds., Individual differences in cog-
full potential of learning skills. nition, Vol. 1. New York: Academic Press.
Schunn, C. D., and Reder, L. M. (2001). Another source of individ-
See also: AMNESIA, FUNCTIONAL; AMNESIA, ual differences: Strategy adaptivity to changing rates of suc-
ORGANIC; CODING PROCESSES: IMAGERY; cess. Journal of Experimental Psychology: General 130, 5976.
CODING PROCESSES: LEVELS OF PROCESSING; Shah, P., and Miyake, A. (1996). The separability of working mem-
CODING PROCESSES: ORGANIZATION OF ory resources for spatial thinking and language processing:
MEMORY; EPISODIC MEMORY; EXPERTS An individual differences approach. Journal of Experimental
MEMORIES; INTERFERENCE AND FORGETTING; Psychology: General 125, 427.
Smith, E. E., Adams, N., and Schorr, D. (1978). Fact retrieval and
MEMORY SPAN; METACOGNITION ABOUT
the paradox of interference. Cognitive Psychology 10, 438464.
MEMORY; MNEMONISTS; SEMANTIC MEMORY:
Sternberg, R. J., and Zhang, L., eds. (2001). Perspectives on thinking,
COGNITIVE ASPECTS; SEMANTIC MEMORY: learning, and cognitive styles. Mahwah, NJ: Erlbaum.
NEUROBIOLOGICAL PERSPECTIVE Tuholski, S. W., Engle, R. W., and Baylis, G. C. (2001). Individual
differences in working memory capacity and enumeration.
Bibliography Memory & Cognition 29, 484492.
Ackerman, P. L. (1987). Individual differences in skill learning: An Voss, J. F., Vesonder, G. T., and Spilich, G. J. (1980). Text genera-
integration of psychometric and information processing per- tion and recall by high-knowledge and low-knowledge indi-
spectives. Psychological Bulletin 102, 327. viduals. Journal of Verbal Learning and Verbal Behavior 19, 651
Bors, D. A., and MacLeod, C. M. (1996). Individual differences in 667.
memory. In E. L. Bjork and R. A. Bjork, eds., Handbook of per- Wechsler, D. (1945). Wechsler memory scale. San Antonio, TX: The
ception and cognition, Vol. 10: Memory. San Diego, CA: Academ- Psychological Corporation.
ic Press. (1958). The measurement and appraisal of adult intelligence, 4th
Coltheart, V., and Evans, J. S. B. T. (1981). An investigation of se- edition. Baltimore, MD: Williams and Wilkins.
mantic memory in individuals. Memory & Cognition 9, 524
532. Colin M. MacLeod
254 INFANCY, MEMORY IN
INFANCY, MEMORY IN Advances in technology allow researchers new

ways of looking at the brain basis of memory in in-
According to classic developmental theory, infants fants. Event-related potentials (ERP) can be used to
operate in the present, without thoughts of the past examine the neural correlates of recognition memory
or anticipations of the future. Adults cannot remem- and attention in infants (Nelson, 1995). Typically,
ber events from infancy, a fact sometimes cited to cor- one of two approaches is used. In one case, infants are
roborate the notion that memories are not formed habituated or familiarized with a stimulus. Brain
during the preverbal period. However, experimental activity is then measured by recording electroenceph-
studies conducted in the late twentieth century dem- alogram (EEG) signals in response to a series of trials
onstrate that young infants have more robust memo- in which either the habituated stimulus or a novel
ries than heretofore believed. Indeed, modern theo- stimulus is presented. Brain measures record whether
rists focus on the different types of memory infants infants discriminate between the new and the old
might have. Infants seem to remember particular stimulus. In a second type of study, infants are shown
things under certain conditions and not others; they stimuli that are familiar: for example, their mothers
may also have privileged memory for biologically rel- face or a familiar toy, and brain activity to these re-
evant signals such as faces and speech sounds. membered stimuli is measured and compared to
novel stimuli (de Haan and Nelson, 1999). The first
approach provides brain measures of memory for
Techniques for Investigating Infant well-controlled, briefly exposed stimuli; the second
Memory approach measures brain reactions to naturally occur-
Three experimental procedures have been devel- ring stimuli that have been frequently seen.
oped to probe infant preverbal memory: visual pref- Age-Related Findings. Initially, studies of visual
erence tests; conditioning procedures; and deferred recognition memory reported no memory in infants
imitation. Each approach measures a different type of younger than about ten to twelve weeks of age. However,
nonverbal memory. A fourth technique, object per- researchers soon discovered that if the length of time in-
manence, can also be used and it is only briefly men- fants studied the to-be-remembered stimuli was in-
tioned in this entry because it is reviewed elsewhere creased (up to five minutes) and the patterns were made
in the encyclopedia. very different from one another, even newborn infants
could retain information in the visual recognition para-
Infant Visual Recognition Memory
digm, at least for delays of a few seconds. Research then
The procedures used to evaluate infant visual rec- shifted to the effect of study time on memory; and the
ognition rely on infants curiosity for exploring novel length of retention interval that can be tolerated by in-
visual patterns (Bremner, 1994). Infants are shown a fants.
visual pattern for a certain length of time. A delay is
imposed, and then they are presented with the old vi- Study Time (Length of the Encoding Phase).
sual pattern and a new one. If infants devote more Infants require shorter study times to demonstrate the
looking time to the new pattern than to the old one, novelty preference (the measure of memory) when the
this is taken as evidence that they have memory of the choice stimuli are vastly different than when the stimuli
previously exposed target. Two specific techniques are similar. This idea was illustrated in a study by Joseph
use this underlying principle: the habituation- Fagan (1990) using the paired-comparison technique.
dishabituation technique and the paired-comparison He tested five-month-old infants using pairs of patterns
technique. For habituation-dishabituation, infants that were graded in the degree to which they were dis-
are repeatedly exposed to the initial target until they criminable from one another. The results showed that
become bored with it (habituated). The new pattern when the easiest pair was used, infants needed only
is then introduced, and if looking time increases sig- about five seconds of study time to demonstrate the nov-
nificantly (dishabituation), this shows that the infants elty preference; when the pair of medium discriminabili-
recognize the pattern as being different from the one ty was used, they needed about twenty seconds; and
when the least discriminable pair was used, they re-
in memory. For the paired-comparison technique, in-
quired about thirty seconds. Thus, infants, like students
fants are initially shown two identical patterns side by
studying for an exam, seem to need relatively more time
side for a certain familiarization period. It is not re-
to study material if they are asked to remember subtle
quired that the infants habituate, only that they visu-
distinctions.
ally examine the display (usually thirty seconds to two
minutes). Then a delay is imposed, and two patterns Length of Retention Interval. Fagan showed vi-
(the old one and a new one) are again presented. The sual patterns to five-month-olds and then imposed de-
index of memory is their preference for the novel pat- lays of three hours and one, two, seven, and fourteen
tern. days. The results revealed that infants could recognize
INFANCY, MEMORY IN 255
which target they had previously seen even after the Retrieving Infant Memories That Were Once
fourteen-day delay. What makes babies forget? Results Forgotten. One remarkable discovery made using this
from a variety of studies show that young infants will for- conditioning technique is that infants can be reminded
get if they are shown highly related material during the about a past event that they have forgotten (Rovee-
retention interval. For example, if infants study photo- Collier, Hayne, and Colombo, 2001). This reminder
graphs of faces and then are shown other face photo- stirs (reactivates) a previously inaccessible memory.
graphs during the retention interval, their subsequent Rovee-Colliers classic demonstration involved three-
memory performance will be poorer. The two factors month-old infants who had forgotten the learned re-
that lead to maximum interference are stimuli that sponse after two weeks. These infants were then exposed
closely resemble the to-be-remembered material; and to a brief reminder (the mobile, which was being moved
interfering presented soon after the initial exposure. surreptitiously by the experimenter). Then the infant
This is reminiscent of adult memory, inasmuch as inter- was given another twenty-four-hour delay; finally, the
ference with remembering a telephone number is maxi- stationary mobile was reintroduced to assess memory.
mized by hearing other numbers soon after the initial in- Infants administered the reminder had their memories
formation is delivered. reactivated and kicked vigorously when the mobile
Another factor that influences the length of infant was reintroduced. Control infants who were not given
retention is the temporal spacing of the initial study- the reminder did not show any memory under the same
ing time. In one study, two groups of infants were circumstance.
given the same length of time to study a face photo- Memory Specificity and the Importance of
graph (twenty seconds). However, for one group this Context. In the first six months of life, infants are ex-
study time was massed, meaning it consisted of four tremely sensitive to the context in which a behavior is ac-
five-second intervals with only a few seconds separat- quired and show better memory if the test occurs in the
ing each interval. For the other group this experience same context as the learning episode. In one study, six-
was distributed, meaning there were much longer month-olds were given foot-kick training in a specially
pauses between the four five-second intervals. Both decorated crib (the context). As long as these infants
groups demonstrated immediate recognition memo- were tested in the same context as the original training,
ry; however, only the group that received the distrib- they remembered to kick, even after a fourteen-day
uted exposure remembered over long delays. This ef- delay. However, if the crib decoration changed, the in-
fect of distributed study is also a well-documented fants could not access their memories, even after a one-
aspect of adult memory. day delay (Rovee-Collier, 1997).
Conditioning Techniques Deferred Imitation
The second approach to evaluating infant memo- The third procedure for testing infant memory is
ry was developed by Carolyn Rovee-Collier (1997). It deferred imitation, that is, imitation after a delay
involves training infants to produce a foot-kick re- (Barr and Hayne, 2000; Meltzoff and Moore, 1998).
sponse to a mobile hanging in their cribs. The mobile This technique capitalizes on the fact that preverbal
is often attached by a ribbon to one of the infants an- infants enjoy imitating the actions of adults. To test
kles, so that the frequency and intensity of the move- memory, the infant is shown the to-be-imitated event,
ment mimics that of the infant (known as conjugate and then a delay is inserted before the infant is al-
reinforcement). Infants as young as two to three lowed to demonstrate the response. Memory is in-
months rapidly learn the contingency, doubling or dexed by accurate reproduction of the target behav-
tripling their baseline rates during the nine-minute ior after the delay. Control groups are tested to
training period. Once the infant has learned the re- ensure that the production of the target behavior
sponse, a delay period can be inserted between the would not have occurred spontaneously in infants not
initial training period and the reintroduction of the exposed to the initial modeling.
mobile into the crib. Memory for the learned re-
sponse is indexed by an increase in kicking over base- Age-Related Findings. Classic developmental
line rates, even after this delay interval. theories, such as those of the Swiss psychologist Jean Pia-
get, had supposed that imitation from memory was a
Age-Related Findings. With this technique,
cognitive achievement that first emerged at about eigh-
Rovee-Collier and colleagues discovered a steady in-
teen months of age. Empirical research revised this clas-
crease in the duration over which infants retain re-
sic view by showing that infants can perform deferred
sponses from two to eighteen months of age. Whereas
imitation as early as nine months of age (Meltzoff, 1988).
the youngest infants tested in this paradigm remem-
bered the response contingency for one day to one week, Imitative Learning and Length of Retention
eighteen-month-olds remembered for thirteen weeks or Interval. Scientists have been interested in infants
longer. learning novel material through observation. Andrew
256 INFANCY, MEMORY IN
Meltzoff (1990) showed infants a novel act that had not sured during the delay interval to pictures of toys. Brain
occurred in the baseline behavior of infants, and tested electrical responses mapped onto individual differences
their memory for this act after a one-week delay. The act in behavior: Infants who successfully performed de-
consisted of bending forward from the waist and tapping ferred imitation showed recognition memory as mea-
the top surface of a box with the top of ones forehead. sured by ERP patterns during the delay interval; infants
Even after a one-week delay, 67 percent of fourteen- who did not imitate showed no recognition memory as
month-olds bent forward from the waist and touched measured by ERP. These results, and others, suggested
their head to the panel (see Figure 1). Leslie Carver and that there are changes in the organization of brain sys-
Patricia Bauer (2001) found that ten-month-olds were tems involved in memory near the end of the first year
able to retain at least some information about to-be- of life, and that these changes can be observed in mea-
imitated events for delays of up to six months. Jean sures of individual differences in both brain and behav-
Mandler (1990) assessed memory for temporal order by ior.
presenting a sequence of behaviors that could be per-
formed in one order or another. Both the sixteen- and Real-World Implications. Infants remember
twenty-month-olds showed immediate memory for the what they see on television. In one study infants were
temporal order of these arbitrary sequences; the groups shown how to manipulate a new toy by an experimenter
differed, however, on their long-term memory perfor- who appeared on television. The infants were not al-
mance. The twenty-month-olds still showed memory for lowed to handle the real toy, but the next day the real
temporal order after a two-week delay, whereas the six- toy was presented on the table. The results showed that
teen-month-olds were at chance levels. Long-term mem- infants accurately imitated the television-presented ac-
ory for arbitrarily sequenced events may develop toward tions they had seen one day earlier. Another study exam-
the second half of the second year, and perhaps is aided ined whether fourteen-month-olds could remember ac-
by the emergence of language during this time. tions performed by other infants. A tutor infant was
taught how to perform a series of particular acts. This
Combined Brain and Imitation Measures. Car- tutor infant was then brought to day-care centers where
ver, Bauer, and Nelson (2000) tested nine-month-olds he demonstrated the acts to naive infants. After a
for deferred imitation of ordered sequences of actions. forty-eight-hour delay, the naive infants were visited at
About half of the infants recalled the events after a delay their homes by a researcher who laid out the toys on the
of one month. Event-related potentials (ERP) were mea- floor. The results showed that the infants imitated what
INFANCY, MEMORY IN 257
they had seen their peer do two days earlier. A general man, 1997). There is a (heated) social policy debate
implication of this work is that imitation and memory as to whether tests of early memory should be adver-
are robust enough to play a significant role in the social tised as infant intelligence tests. This debate should
and personality development of the preverbal infant not mask the scientific discovery that there is continu-
(Meltzoff and Moore, 1998). ity between the mental performance of infants, as
measured by their performance on tests of memory,
and childhood IQ scores.
Conclusion and a Look to the Future
Multiple Memory Systems in Infants Early and Rapid Memory Formation for
Biologically Relevant Signals
The three techniques used to explore infant
memory complement each other but do not address Newborn infants are predisposed to encode and
precisely the same aspects of infant memory. The dis- remember biologically important signals such as fa-
tinctions are important for theory. cial and speech signals. Newborn infants, with only a
few hours of exposure to the mother, look longer at
Recognition Versus Recall. In deferred imita- their own mothers face than at a strangers face. New-
tion, infants go beyond the regulation of attention; they borns also choose to listen to the voice of their own
do more than react to the newness of a pattern. They mother than to that of a strange female talker, sug-
must produce an absent act without seeing it and without
gesting that perhaps the sound of the mothers voice
having previously imitated it. Deferred imitation taps
is learned prenatally. Finally, recent discoveries show
something more than visual recognition memory and
that young infants respond differently to native- as
provides a measure of recall memory prior to the acqui-
opposed to foreign-language speech sounds. For ex-
sition of language.
ample, Patricia Kuhl and colleagues (2001) showed
Imitative Learning Versus Conditioning. Like that six-month-old Swedish infants have committed
deferred imitation, the conditioned foot-kick technique Swedish but not English speech sounds to memory,
goes beyond visual recognition memory because the in- whereas American infants have done the opposite,
fants do more than recognize the familiar mobile; they demonstrating that infants are listening to the ambi-
also retrieve from memory what to do (kick). However, ent sounds in the environment and remembering
the conditioning procedure and the deferred imitation them even before they can talk.
task differ in the type of information retained. Deferred
imitation is based not on an incrementally learned pro- Infantile Amnesia Revisited
cedure (as in the case of conditioned foot kicks) but on Research indicates that infants have far more ro-
the performance of an act that was simply perceived dur- bust and complex memories than classic theories pre-
ing a brief previous episode. The deferred imitation test dicted. The puzzling phenomenon of infantile amne-
does not involve any motor practice during acquisition sia becomes more of a mystery when considered in
of the to-be-remembered event (no immediate imitation light of this modern infancy research, because it can
is allowed). The two tests also differ because the link be-
no longer be thought that infants do not form memo-
tween the stimulus and the infants response is not
ries or that they are confined to sensorimotor skill
forged through conditioning in deferred imitation.
routines during the preverbal period (Meltzoff,
These distinctions are relevant for theories of mem-
1995). It is possible that the amnesia adults experi-
ory and its development. Cognitive scientists and
ence about their own infancy is due to the extreme
neuroscientists have made a distinction in different
mismatch between the cognitive, emotional, and
types of memory systems, particularly between proce-
physical context of the initial learning and that of the
dural and declarative memory systems. One hypothe-
adult. It is sometimes reported that adults can gain
sis is that deferred imitation provides a technique
for exploring a primitive form of declarative memory,
access to lost childhood memories by immersing
which researchers believe is mediated by the medial tem- themselves in unique situations they have not encoun-
poral lobe. tered since childhood. Alternatively, the fact that
adults use a linguistic code may shroud their memo-
Relations Between Infant Memory and ries of the preverbal period.
Childhood IQ
Do scores on infant memory tests predict later See also: AMNESIA, INFANTILE; CHILDREN,
cognitive performance? Empirical work demonstrates DEVELOPMENT OF MEMORY IN; DISTRIBUTED
that infants who perform better on tests of memory PRACTICE EFFECTS; INTERFERENCE AND
between two to nine months of age score higher on FORGETTING; LANGUAGE LEARNING: HUMANS;
IQ (intelligence quotient) tests given later in child- OBJECT CONCEPT, DEVELOPMENT OF;
hood (McCall and Carriger, 1993; Rose and Feld- REINFORCEMENT
258 INSECT LEARNING
Bibliography INSECT LEARNING

Barr, R., and Hayne, H. (2000). Age-related changes in imitation:
Implications for memory development. In C. Rovee-Collier, Why study learning in insects? What can it contribute
L. P. Lipsitt, and H. Hayne, eds., Progress in infancy research. to a general knowledge of how learning takes place
Mahwah, NJ: Ablex. in a wide variety of animals? There are many poten-
Bremner, J. G. (1994). Infancy, 2nd edition. Cambridge, MA: tial answers to these questions. This review will focus
Blackwell. on the general contribution that can be made to sys-
Carver, L. J., and Bauer, P. J. (2001). The dawning of a past: The
tematic understanding of how learning has evolved
emergence of long-term explicit memory in infancy. Journal
of Experimental Psychology: General 130, 726745.
and is controlled in a wide variety of vertebrate and
Carver, L. J., Bauer, P. J., and Nelson, C. A. (2000). Associations invertebrate species. Any systematic study must begin
between infant brain activity and recall memory. Developmen- with a well-defined phylogenetic lineage. Insects are
tal Science 3, 234246. in the phylum Arthropoda, which contains animals
de Haan, M., and Nelson, C. A. (1999). Brain activity differentiates that have jointed exoskeletons (e.g., insects, ticks,
face and object processing in 6-month-old infants. Develop- crabs, lobsters, spiders). With at least 2 million extant
mental Psychology 35, 1,1131,121.
species (some estimates range as high as 30 to 50 mil-
Fagan, J. F., III. (1990). The paired-comparison paradigm and in-
fant intelligence. In A. Diamond, ed., Annals of the New York
lion), the arthropod class Insecta comprises the most
Academy of Sciences, Vol. 608: The development and neural bases diverse group of multicellular organisms (Borror,
of higher cognitive functions. New York: The New York Academy Triplehorn, and Johnson, 1989). Insects have adapt-
of Sciences. ed to a wide array of living conditions, ranging from
Kuhl, P. K., Tsao, F. M., Liu, H. M., Zhang, Y., and de Boer, B. most terrestrial to many aquatic environments. The
(2001). Language/culture/mind/brain: Progress at the mar- diverse insect species found in these environments
gins between disciplines. In A. R. Damasio et al., eds., Unity
must solve the basic problems inherent in locating re-
of knowledge: The convergence of natural and human science. New
York: The New York Academy of Sciences.
sources such as food or mates and avoiding predatory
Mandler, J. M. (1990). Recall of events of preverbal children. In A. or environmental threats. The learning abilities ob-
Diamond, ed., Annals of the New York Academy of Sciences, Vol. served in any laboratory situation probably evolved to
608: The development and neural bases of higher cognitive func- solve these problems.
tions. New York: The New York Academy of Sciences.
McCall, R. B., and Carriger, M. S. (1993). A meta-analysis of infant
Because of this species and habitat diversity, in-
habituation and recognition memory performance as predic- sects provide an excellent means of testing patterns
tors of later IQ. Child Development 64, 5779. of phylogenetic emergence of different learning
Meltzoff, A. N. (1988). Infant imitation and memory: Nine-month- mechanisms. The appropriateness of using cognitive
olds in immediate and deferred tests. Child Development 59, explanations, the lack of generally agreed upon oper-
217225. ational definitions, the need for learning taxonomies,
(1990). Towards a developmental cognitive science: The
and a recognition of individual differences in perfor-
implications of cross-modal matching and imitation for the
development of representation and memory in infancy. In A.
mance are all relevant issues in the study of inverte-
Diamond, ed., Annals of the New York Academy of Sciences, Vol. brate learning. One must also carefully distinguish
608: The development and neural bases of higher cognitive func- phylogenetically homologous traits from analogous
tions. New York: The New York Academy of Sciences. ones. Homology refers to traits that arise from a com-
(1995). What infant memory tells us about infantile amne- mon ancestral condition. Researchers must always
sia: Long-term recall and deferred imitation. Journal of Experi- consider rigorous criteria for any proposed homology
mental Child Psychology 59, 497515.
between behaviors of different groups of inverte-
Meltzoff, A. N., and Moore, M. K. (1998). Object representation,
identity, and the paradox of early permanence: Steps toward
brates and/or vertebrates (Wenzel, 1992). A mono-
a new framework. Infant Behavior and Development 21, 201 phyletic group of animals that possesses homologous
235. traits is a clade, and the process of modification of
Nelson, C. A. (1995). The ontogeny of human memory: A cognitive those traits is cladogenesis.
neuroscience perspective. Developmental Psychology 31, 723
738. Through a comparison of traits that may appear
Rose, S. A., and Feldman, J. F. (1997). Memory and speed: Their dissimilar and thus unrelated in several closely relat-
role in the relation of infant information processing to later ed, extant species, it is possible to obtain a picture of
IQ. Child Development 68, 630641. which traits are ancestral (plesiomorphic) and which
Rovee-Collier, C. (1997). Dissociations in infant memory: Rethink- are derived. For example, are nonassociative learning
ing the development of implicit and explicit memory. Psycho- mechanisms homologous to associative mechanisms?
That is, did ancestral species possess the ability to
Rovee-Collier, C., Hayne, H., and Colombo, M. (2001). The develop-
ment of implicit and explicit memory. Philadelphia: John Benja- modify behavior through habituation and sensitiza-
mins Publishing Co. tion prior to the ability to express associative condi-
tioning? Were changes in the evolution of learning
Andrew N. Meltzoff abilities gradual, adaptive alterations or the result of
Revised by Andrew N. Meltzoff and Leslie J. Carver rapid, discontinuous changes resulting from dramatic
INSECT LEARNING 259
reorganization of neural tissue (Wyers, Peeke, and brain) but can be distributed throughout several stim-
Herz, 1973)? Through a study of insect species whose ulus and motor processing pathways in a nervous sys-
phylogenies with respect to other characteristics (e.g., tem.
morphological, physiological) are known, such hy-
potheses can be tested. The Horridge paradigm also brought into focus
the adequacy of the yoked-control design in separat-
Analogous traits are physically similar but have ing learning from nonassociative effects. In the origi-
been derived from very different ancestral conditions. nal Horridge experiment, learning was inferred from
For example, the expression of operant conditioning a difference in the number of shocks received by ex-
of leg movement in an insect might be analogous to perimental subjects and by their yoked controls. The
operant conditioning of leg movement in a verte- experimental subjects were shocked contingent on
brate. Insect legs and vertebrate legs are not homolo- leg position; control subjects were yoked in such a way
gous structures, but both enable animals to move that they received a shock whenever the experimental
throughout their environment and thus the rules for subjects did, but independent of their own behavior.
operant conditioning of each may be similar. Analo- The yoked paradigm has been extensively criticized
gous traits associated with complex learning abilities in the literature. Church (1964), for instance, pointed
arise through convergent evolution, perhaps because out that because of the nature of the yoked paradigm,
of similar environmental problems that require one random differences in inherent responsiveness will
or more ways to modify behavior based on experi- lead to artifactual learning in the population.
ence. This description of analogous learning abilities
in terms of degrees of complexity is anagenesis (De- To answer such criticisms, a new experimental
marest, 1983). A comparative study of potentially design was developed for training leg position in the
analogous learning mechanisms in such phylogeneti- locust (Forman, 1984). Rather than simply requiring
cally diverse groups as insects and vertebrates allows the animal to raise a leg to terminate a series of
the testing of hypotheses about the conditions that shocks, Forman required his locust to maintain a par-
give rise to analogous learning abilities. Thus learn- ticular range of leg movements arbitrarily selected by
ing abilities studied in insects have an important the experimenter. After a few minutes of training, the
value for deriving hypotheses that are testable in ver- animal learns to shift its leg position to an angle that
tebrates. Even if learning traits do not have a common terminates aversive heat to the head or produces ac-
phylogenetic origin, working out mechanisms in one cess to food. Locusts can also be trained to manipulate
species can generate conceptual advances in under- leg position to produce heat to the head in a cold en-
standing a similar learning ability in another species. vironment. The task can be made more complex by
narrowing or shifting the range of leg movement nec-
No single study can be designed to investigate all
essary to control the heat stimulus. The Forman ex-
of these issues. In the long run this approach must
periment is important because it represents the first
make extensive use of several species that are chosen
significant improvement in the Horridge paradigm;
to appropriately test defined phylogenetic hypothe-
both the response and the reinforcer are arbitrary,
ses (Wenzel, 1992). What follows, then, is not a com-
and learning can be identified in an individual ani-
prehensive review of the invertebrate learning litera-
mal. This procedure has the additional advantage of
ture but rather a highlighting of significant studies.
eliminating shock as the aversive stimulus.
The rationale for developing the Forman para-
Orthoptera digm is to obtain data on the cellular mechanisms un-
The order Orthoptera comprises cockroaches, derlying operant behavior. By using electromyo-
grasshoppers, and locusts. Horridge (1962) pub- grams and intracellular techniques, the motor
lished results of an experiment in which headless neurons involved in learning have been found and
cockroaches and locusts learned to keep one leg characterized. Forman and Zill (1984), for example,
raised to terminate a series of electric shocks. This ex- identified three separate motor strategies utilized by
periment generated considerable interest because it the locusts during training (kicking, changes in
was one of the first to suggest that an insect can be muscle tonus, and tonic slow excitor motor neuron ac-
used to explore the physiology of learning and mem- tivity). Each of these strategies can be selectively
ory. Subsequently, leg-position learning, or the Hor- trained. An exciting application of the technique and
ridge paradigm, has demonstrated learning in a wide a fine example of the comparative method is an anal-
variety of experimental situations, ranging from in- ysis of the similarities and differences in response
tact animals to a single ganglion. This latter informa- strategies between locusts and the weta, a primitive
tion demonstrated that learning need not be confined New Zealand insect related to the locust (Hoyle and
to a single area of the central nervous system (e.g., the Field, 1983).
260 INSECT LEARNING
Diptera associative learning mechanisms may be mechanisti-

cally related to simpler nonassociative processes. Ac-
The order Diptera comprises all flies. Fly species
cordingly, learning studies with both P. regina and D.
that have been extensively used in studies of learning
melanogaster have focused on developing associative
include Phormia regina (blowflies) and Drosophila me-
conditioning paradigms and testing for genetic corre-
lanogaster (fruit flies). Research interest in flies was
lates with nonassociative processes. By associating ei-
generated by the extreme ease of maintaining popu-
ther a saline or a water conditioned stimulus (CS) with
lations under controlled mating conditions over gen-
sucrose, P. regina can learn to extend their proboscis-
erations that cover only weeks rather than years. Con-
es to the CS (Nelson, 1971). Tully, Zawistowski, and
trolled breeding experiments have characterized the
Hirsch (1982) selected for high and low learning lines
behavioral, genetic, and biochemical bases of differ-
of blowflies. These lines showed a positive correlation
ent learning mechanisms.
between CES levels and asymptotic levels of learning
Some of the first comprehensive studies of learn- performance. Therefore, CES and associative condi-
ing behavior in flies began with the pioneering work tioning appear to have at least some common genetic
of Dethier and colleagues (1990), which worked out bases, which might include pleiotropic genetic effects.
in considerable detail the stimulus control of feeding
Aversive conditioning has been widely used to se-
reflexes in Phormia regina. They described a proce-
lect large numbers of D. melanogaster in order to rap-
dure in which the tarsal (leg) receptors that mediate
idly isolate mutant strains that show deficiencies in
sucrose taste sensation were stimulated to elicit exten-
learning and/or memory (Dudai, 1983). The proce-
sion of a subjects proboscis (the sucking mouthparts)
dure involves exposing flies sequentially to two odors
through which it feeds on the sucrose-water droplet.
while they walk across a metal grid (Tully and Quinn,
They found that prior exposure to sucrose greatly in-
1985). While they are exposed to one odor, they re-
creased the probability that a fly would extend its pro-
ceive shocks through the grid. Flies are then pres-
boscis to the presentation of water alone. The motiva-
ented with a sequence of new collection tubes that
tional state that was modulated by the sucrose
contain either the odor paired with shock (S+) or the
exposure was termed central excitatory state (CES),
odor that was not paired with shock (S-). The re-
which describes a nonassociative (sensitization) modi-
sponse measure is the number of flies that enter a new
fication of the probability of proboscis extension to a
tube that contains the S+ odor versus the number
neutral stimulus (water). CES can be characterized by
that enter a tube containing the S- odor. Over several
at least three factors: There is a decay over time be-
runs, decreased entries into the tube with the S+ odor
tween the sensitizing and test stimuli; increased su-
relative to the tube with the S- odor indicate learning.
crose concentrations lead to increases in CES; food
deprivation leads to increased levels of CES for any
given sucrose concentration. Hymenoptera
Tully and Hirschs studies (Tully, 1984) have doc- The order Hymenoptera contains a diverse
umented genetic bases for CES effects in P. regina, group of insects commonly referred to as sawflies,
and other studies extended the results on the genetic ants, wasps, and bees. Hymenopterans such as ants
basis for CES effects to D. melanogaster (Vargo and and honeybees have been widely used to document
Hirsch, 1986). Bidirectional selection for high and learning abilities related directly to learning prob-
low CES lines in P. regina has shown that the response lems in the animals natural environment. Further-
to selection is rapid and may reach asymptotic levels more, experiments with ants and bees in laboratory
in one or a few generations. Hybridization of the dif- learning paradigms have demonstrated that these
ferent lines indicated that one major gene segregated abilities conform to standard definitions of learning.
in the selected lines was responsible for producing But an ants or a bees learning ability may be less
most of the variability in CES effect. Selection for CES complex and/or less generalizable to new situations
effects in D. melanogaster has shown a slightly different than that of animals with larger, more elaborate ner-
genetic basis. Sometimes a low but not a high line was vous systems (Demarest, 1983). Indeed, the crucial
produced, or vice versa. These data indicate that sev- question is how complex these abilities are and to
eral genes may be involved in regulating CES in fruit what natural situations they can be applied.
flies. Further studies have shown that genes reside on
at least two chromosomes; heritable cytoplasmic fac- Learning in ants was first brought into the labora-
tors may also be involved (Vargo and Hirsch, 1986). tory by Fielde (1901), who reported that ants can suc-
cessfully negotiate a simple maze. Schnierla (1946)
Theories of anagenesis predict that as metazoan described the chemical, visual, and kinesthetic cues
life becomes physically more complex, more complex used by ants in solving a more complex maze. DeCar-
learning abilities will emerge (Demarest, 1983). Thus lo and Abramson (1989) used a different procedure
INSECT LEARNING 261
to extend vertebrate learning paradigms to ants. The question whether bees and ants have a cogni-
They demonstrated an ants ability to choose one tive map has generated controversy. That is, have
compartment of a two-compartment chamber based they learned a mental analogy of a topographic
on rates of stimulus delivery. map, i.e., an internal representation of the geometric
relations among noticeable points in the animals en-
The honeybee (Apis mellifera) is an ideal species vironment (Wehner and Menzel, 1990, p. 403)? Evi-
with which to research similar questions. On warm, dence to date points to a vector-based navigation sys-
sunny days worker bees regularly depart from the col- tem combined with memory matching of relative
ony on foraging trips during which they collect re- positions of landmarks (Cartwright and Collett, 1987)
sources (e.g., nectar, pollen, water) crucial for survival rather than to a more complex topographic represen-
and reproduction of the colony. A large number of tation.
studies have documented the abilities of freely flying
honeybees to learn the relationships of visual, tactile,
and olfactory cues to appetitive and aversive stimuli New Developments in Invertebrate
(Menzel, 1990; Bitterman, 1996). For example, for- Learning
ager bees learn the association of nectar, which for
Later research in insect learning has emphasized
most conditioning studies is replaced with a sucrose- one species of insect: the honeybee (Bitterman,
water mixture and floral color, shape, odor, and the 1996). This work emphasizes the type of research
time of day that floral rewards are available. Other program into learning mechanisms that ought to be
work has documented the honeybees ability to learn taken up on a much wider array of species.
compounds of stimuli. The unconditioned stimulus-
preexposure effect and latent inhibition have also Blocking
been studied. For studies of aversive learning, Abram- Kamin (1968) conditioned rats to one stimulus
son (1986) has demonstrated the ability of freely fly- (A) and followed that with conditioning to a second
ing bees to use certain stimuli as a means of avoiding stimulus that was a compound of the first with a novel
exposure to an aversive-shock stimulus. The learned stimulus (AX). When finally tested with X, subjects
avoidance ability of the honeybee may have evolved typically revealed that they had learned less about X
as a means to cope with bitter and even toxic nectars than subjects in appropriate control groups. This
found in some flowers. blocking phenomenon went on to spawn a tremen-
dous volume of both theoretical and empirical re-
Proboscis-extension conditioning of honeybees is
search. Most research on vertebrates has emphasize
a common technique for studying learning under eas-
blocking between cues from different stimulus modal-
ily controllable stimulation variables (Menzel 1990).
ities (e.g., compounds of one visual and one acoustic
Honeybees restrained individually in harnesses can
cue). Recently, blocking has been investigated in a se-
be readily conditioned to extend their proboscises
ries of studies in honeybees (Smith 1997). In contrast
upon presentation of a floral odor. After one or a few to vertebrates, blocking is more evident in com-
pairings of an odor-conditioned stimulus with a su- pounds made of the same stimulus modality (mix-
crose unconditioned stimulus, 40 to 90 percent of the tures of two olfactory or two visual cues), although in-
subjects will extend their proboscises (conditioned re- termodal blocking has been identified for some
sponse) to the odor alone. Enhancement of a back- stimulus compounds (Couvillon, Campos, Bass, and
ground rate of proboscis extension to odor is specific Bitterman, 2001). Whether this stimulus specificity
to situations in which the CS precedes the US (for- represents a fundamental difference between verte-
ward pairing) and is sensitive to latency of onset of brates and invertebrates must await further study.
odor relative to sucrose. Proboscis-extension condi-
tioning has been used to study a variety of phenome- CS Preexposure
na in honeybees, such as sensory discrimination Unreinforced exposure to a conditioned stimulus
(Smith and Menzel, 1989a); control of motor systems leads to retardation of learning about that CS when
(Smith and Menzel, 1989b); memory consolidation it is subsequently paired with reinforcement in a way
(Menzel, 1990). More recent work has extended to re- that would normally result in robust conditioned re-
strained bees Abramsons (1986) study of aversive sponding. This mechanism has also been referred to
conditioning in freely flying bees. The majority of as latent inhibition. Abramson and Bitterman
subjects that received a shock contingent upon their (1986) initially reported a CS-preexposure effect in
response to sucrose in the context of a particular odor free-flying honeybees. Studies of odor preexposure
learned to withhold proboscis extension to sucrose in have revealed the effect in laboratory studies in hon-
order to avoid shock (Smith, Abramson, and Tobin, eybees (Chandra, Hunt, and Smith, 2001). These lat-
1991). ter studies highlight the importance of recognition of
262 INSECT LEARNING
individual difference in expression of a preexposure subspecies difference between European and African-
effect. Individual honeybees differ in the extent that ized honeybees. In addition to work on basic phe-
they exhibit the effect, and both studies were success- nomena, experiments on practical applications of
ful in demonstrating that individual differences have conditioning methodology are illustrated with studies
a genetic basis. This latter finding indicates that indi- demonstrating the effects of insecticides on learning
vidual differences are most likely not an artifact of the and the reaction of Africanized bees to consumer
approach, and they may point in future studies to an products (Abramson, Aquino, Silva, and Price, 1997;
ecological meaning for individual differences. Chan- Abramson, Aquino, Ramalho, and Price, 1999;
dra, Hunt, and Smith (2001) isolated segments of Abramson, Aquino, and Stone, 1999).
chromosomal DNA that presumably house genes that
The Development of a Social Insect Model
influence this trait.
of Alcoholism
Risk Sensitivity and Choice Behavior Honeybees are also a model for studies of alco-
Two studies have revealed risk sensitivity in hon- holism in humans. Honeybees have much to recom-
eybees (Shafir, Wiegmann, Smith, and Real 1999; mend them for such studies, including a language
Shapiro, Couvillon, and Bitterman, 2001). When two and social structure. Studies have shown that bees will
different conditioned stimuli are reinforced at the self-administer ethanol and that locomotion and
same mean rate but the reinforcement differs in vari- learning is impaired in a dose-dependent manner
ance, animals may prefer the less variable option (risk (Abramson et al., 2000).
aversion), the more variable option (risk prone), or
neither (risk insensitive). Under normal conditions See also: EVOLUTION AND LEARNING;
honeybees are risk-averse. As with CS preexposure, INVERTEBRATE LEARNING: ASSOCIATIVE
individuals differ in their degree of risk sensitivity. LEARNING AND MEMORY PROCESSING IN BEES;
This behavior resembles risk sensitivity in vertebrates. INVERTEBRATE LEARNING: NEUROGENETICS OF
But a model that incorporates associationist concepts MEMORY IN DROSOPHILA; OPERANT BEHAVIOR;
can account for risk sensitivity in honeybees (Shapiro, PAVLOV, IVAN; SPATIAL LEARNING: ANIMALS
Couvillon, and Bitterman, 2001). So more cognitive
interpretations of this phenomenon need not be in- Bibliography
voked. More recently, Shafir, Waite, and Smith (2002) Abramson, C. I. (1986). Aversive conditioning in honeybees (Apis
have shown that honeybees violate basic properties of mellifera). Journal of Comparative Psychology 100, 108116.
rational choice because the choice between two op- Abramson, C. I., Aquino, I. S., Ramalho, F. S., and Price, J. M.
(1999). Effect of insecticides on learning in the Africanized
tions for reward depends on available alternatives.
honey bee (Apis mellifera L.). Archives of Environmental Contami-
Their relative preference between two options nation and Toxicology 37, 529535.
changes with the introduction of a third, relatively un- Abramson, C.I., Aquino, I. S., Silva, M. C., and Price, J. M. (1997).
attractive option. Learning in the Africanized honey bee: Apis mellifera L. Physi-
ology and Behavior 62, 657674.
Matching-to-Sample Abramson, C. I., Aquino, I. S. and Stone, S. M. (1999). Failure to
find proboscis conditioning in one-day old Africanized honey-
Honeybees master more abstract relationships
bees (Apis mellifera L.) and in adult Uruu honeybees (Melipona
among cues, such as sameness or difference (Gi- scutellaris). International Journal of Comparative Psychology 12,
urfa et al., 2001). Honeybees can be conditioned to 242262
respond to a cue based on whether or not it matched Abramson, C. I., and Bitterman, M. E. (1986). Latent inhibition in
a sample cue to which animals were recently exposed. honeybees. Animal Learning and Behavior 14, 184189.
Abramson, C. I, Stone, S. M., Ortez, R. A., Luccardi, A., Vann, K.
This rule can also be used in a more general sense:
L., Hanig, K. D. and Rice, J. (2000). The development of an
once the rule is learned in one stimulus modality. it ethanol model using social insects I: Behavior studies of the
can apply to cues from a different modality. Risk sen- Honey Bee (Apis mellifera L.). Alcoholism: Clinical and Experi-
sitivity and matching-to-sample may yet prove to be mental Research 24, 1,1531,166.
fruitful for pursuit of more cognitive manifestations Bitterman, M.E. (1996). Comparative analysis of learning in hon-
eybees. Animal Learning and Behavior 24, 123141.
of invertebrate learning.
Borror, D. J., Triplehorn, C. A., and Johnson, N. F. (1989). An in-
Comparative Analysis of Learning in troduction to the study of insects, 6th edition. Philadelphia: Saun-
ders College Publishing.
Africanized Honeybees Cartwright, B. A., and Collett, T. S. (1987). Landmark maps for
Another recent development in learning of hon- honeybees. Biological Cybernetics 57, 8593.
eybees is the study of the Africanized honeybee. Chandra, S. B. C., Hunt, G., and Smith, B. H. (2001) Quantitative
trait loci associated with reversal learning and latent inhibi-
Studies have examined conditioning to various stimu-
tion in honeybees (Apis mellifera) Behavior Genetics 31, 275
li, extinction (both unpaired and CS only), condi- 285.
tioned inhibition, color and odor discrimination, and Church, R. M. (1964). Systematic effect of random error in the
learning in day-old bees. The results suggestion a yoked control design. Psychological Bulletin 62, 122131.
INTELLIGENCE AND MEMORY 263
Couvillon, P. A., Arakaki, L., and Bitterman, M. E. (1997). Intra- lifera) during discriminative punishment. Journal of Compara-
modal blocking in honeybees. Animal Learning and Behavior tive Psychology 105, 345356.
25, 277282. Smith, B. H., and Menzel, R. (1989a). The use of electromyogram
Couvillon, P. A., Campos, A. C., Bass, T. D., and Bitterman, M. E. recordings to quantify odor discrimination in the honeybee.
(2001). Intermodal blocking in honeybees. Quarterly Journal of Journal of Insect Physiology 35, 369375.
Experimental Psychology B. 54, 369381. (1989b). An analysis of variability in the feeding motor pro-
DeCarlo, L. T., and Abramson, C. I. (1989). Time allocation in the gram of the honey bee: The role of learning in releasing a
carpenter ant (Camponotus herculeanus). Journal of Comparative modal action pattern. Ethology 82, 68 81.
Psychology 103, 389400. Tully, T. (1984). Drosophila learning: Behavior and biochemistry.
DeJianne, D., McGuire, T. R., and Pruzan-Hotchkiss, A. (1985). Behavior Genetics 14, 527557.
Conditioned suppression of proboscis extension in Drosophila Tully, T., and Quinn, W. G. (1985). Classical conditioning and re-
melanogaster. Journal of Comparative Physiology and Psychology tention in normal and mutant Drosophila melanogaster. Journal
99, 7480. of Comparative Physiology 157, 263277.
Demarest, J. (1983). The ideas of change, progress, and continuity Tully, T., Zawistowski, S., and Hirsch, J. (1982). Behavior-genetic
in the comparative psychology of learning. In D. W. Rajecki, analysis of Phormia regina: III. A phenotypic correlation be-
ed., Comparing behavior: Studying man studying animals. Hills- tween the central excitatory state (CES) and conditioning re-
dale, NJ: Erlbaum. mains in replicate F2 generations of hybrid crosses. Behavior
Dethier, V. G. (1990). Chemosensory physiology in an age of tran- Genetics 12, 181191.
sition. Annual Review of Neuroscience 13, 113. Vargo, M., and Hirsch, J (1986). Biometrical and chromosome
analyses of lines of Drosophila melanogaster selected for central
Dudai, Y. (1983). Mutations affect storage and use of memory dif-
excitation. Heredity 56, 1924.
ferentially in Drosophila. Proceedings of the National Academy of
Wehner, R., and Menzel, R. (1990). Do insects have cognitive
Sciences of the United States of America 80, 5,4455,448.
maps? Annual Review of Neuroscience 13, 403404.
Fielde, A. (1901). A further study of an ant. Proceedings of the Nation-
Wenzel, J. W. (1992). Behavioral homology and phylogeny. Annual
al Academy of Sciences of the United States of America 53, 521544.
Review Entomology 23, 361381.
Forman, R. R. (1984). Leg position learning by an insect. I. A heat
Wyers, E. J., Peeke, H. V. S., and Herz, M. J. (1973). Behavioral ha-
avoidance learning paradigm. Journal of Neurobiology 15, 127 bituation in invertebrates. In H. V. S. Peeke and M. J. Herz,
140. eds., Habituation, Vol. 1. New York: Academic Press.
Forman, R. R., and Zill, S. N. (1984). Leg position learning by an
insect. II. Motor strategies underlying learned leg extension. Brian H. Smith
Journal of Neurobiology 15, 221237. Charles I. Abramson
Giurfa, M., Zhang, S., Jenett, A., Menzel, R., and Srinivasan, M.V.
(2001). The concepts of sameness and difference in an in-
sect. Nature 410, 930933.
Horridge, G. A. (1962). Learning of leg position by headless in-
sects. Nature 193, 697698. INSTRUMENTAL CONDITIONING
Hoyle, G., and Field, L. H. (1983). Elicitation and abrupt termina- See: APLYSIA: CLASSICAL CONDITIONING AND
tion of behaviorally significant catchlike tension in a primitive
OPERANT CONDITIONING; OPERANT
insect. Journal of Neurobiology 14, 299312.
BEHAVIOR
Kamin, L. J. (1968). Attention-like processes in classical condition-
ing. In Jones, M.R., ed., Miami symposium on predictability of be-
havior: Aversive stimulation. Coral Gables, FL: Miami Universi-
ty Press.
INTELLECTUAL DISABILITIES
Menzel, R. (1990). Learning, memory, and cognition in honey- See: MENTAL RETARDATION (INTELLECTUAL
bees. In R. P. Kesner and D. S. Olton, eds., Neurobiology of com- DISABILITIES)
parative cognition. Hillsdale, NJ: Erlbaum.
Nelson, M. C. (1971). Classical conditioning in the blowfly (Phormia
regina): Associative and excitatory factors. Journal of Compara-
tive Physiology and Psychology 77, 353368.
Schnierla, T. C. (1946). Ant learning as a problem in comparative INTELLIGENCE AND MEMORY
psychology. In P. Harriman, ed., Twentieth-century psychology.
New York: Philosophical Library. Memory plays an important role in intelligent behav-
Shafir, S., Waite, T. A., and Smith, B. H. (2002). Context- ior. Modern assessments of cognitive functioning,
dependent violations of rational choice in honeybees (Apis such as the Wechsler Adult Intelligence Scale (WAIS-
mellifera) and gray jays (Perisoreus canadensis). Behavioral Ecolo- III), include memory tests among other essential
gy Sociobiology 51, 180187.
measure of mental capability. Modern information-
Shafir, S., Wiegmann, D. D., Smith, B. H., and Real, L. A. (1999).
Risk-sensitivity of harnessed honeybees to variability in vol-
processing conceptions of memory, however, repre-
ume of reward. Animal Behaviour, 57, 1,0551,061. sent a distinct change from the information-reservoir
Shapiro, M. S., Couvillon, P. A., and Bitterman, M. E. (2001). conceptions of memory put forth by earlier research-
Quantitative tests of an associative theory of risk-sensitivity in ers such as Richard Atkinson and Richard Shiffrin
honeybees. Journal of Experimental Biology 204, 565573. (1968). This change, a consequence of the shift from
Smith, B. H. (1997). An analysis of blocking in binary odorant mix-
using storage metaphors for memory to using com-
tures: An increase but not a decrease in intensity of reinforce-
ment produces unblocking. Behavioral Neuroscience 11, 5769. puter processing metaphors, has revolutionized sci-
Smith, B. H., Abramson, C. I., and Tobin, T. R. (1992). Condition- entific understanding of memory and its link to intel-
al withholding of proboscis extension in honeybees (Apis mel- ligence.
264 INTELLIGENCE AND MEMORY
Figure 1
Memory as Information Processing performance on intelligence tests (Kyllonen, 1996;

Kyllonen and Stephens, 1990). Kyllonen discovered
In their edited volume, Akira Miyake and Priti
a high correlation between measures of working
Shah (1999) integrated the work of leading memory
memory and tests of reasoning, even concluding that
theorists into a consensus definition of memory,
working memory and general intelligence amount to
which dispenses with the notion of a biochemical fil-
the same thing (Kyllonen, 2002).
ing cabinet in the brain and proposes a multifaceted,
working information-processing system whose func- Other work has been geared toward determining
tion is to aid in complex cognition. As illustrated in more precisely what accounts for the relationship be-
Figure 1, the limitations of this information- tween working memory and cognitive functioning.
processing system can be attributed to multiple fac- Researchers have focused on the encoding, process-
tors: such as the encoding of novel information, the ing, and retrieval aspects of the working memory sys-
simultaneous storage and processing of information, tem.
and the retrieval of information stored long term.
The concept of working memory is not new but was The Role of Encoding
first put forth by Alan Baddeley and Graham Hitch
in 1974. However, much hard work since the 1970s Encoding involves transforming perceptual in-
formation from the environment into initial input for
has extended Baddeley and Hitchs initial work to de-
the information-processing system (see Figure 1).
velop a more precise characterization of this system,
Limitations in receiving information from the envi-
its limitations, and its relationship to intelligent be-
ronment have implications for later information pro-
havior.
cessing and, ultimately, intelligent behavior.
Measures of working memory have shown a Christopher Jarrold, Alan Baddeley, and Alexa
strong and consistent relationship with measures of Hewes (2000) examined the relative amount of imme-
intelligence and complex cognitive processing. diate recall, near-immediate recall, and short-term
Though numerous experimental findings demon- recall in children of varying levels of retardation.
strate this link, those of Patrick Kyllonen and his col- They briefly presented the children with three spoken
leagues, Raymond Christal and Deborah Stephens, words, each of which was associated with one of three
have been especially illuminating. Seeking to provide horizontal positions on a computer screen such that
an information-processing explanation for intelli- the first word spoken was associated with the leftmost
gence, Kyllonen and his colleagues developed and position. They then highlighted one of the three posi-
tested his model of intelligent performance, which tions and asked the children to recall the word that
posited that mental functioning depends on four cog- had been associated with that position. Recalling the
nitive sources: working memory, processing speed, word associated with the leftmost position, spoken
declarative knowledge, and procedural knowledge. earliest, required short-term recall, the middle posi-
They found that of the four sources, working memory tion, near-immediate recall, and the rightmost posi-
showed the strongest relations to skill acquisition and tion, immediate recall.
These researchers were interested in determining helped to develop several working-memory tests,
whether failure to rehearse to-be-remembered infor- each with a storage and processing aspect (see Figure
mation could account for memory deficits in children 2). The critical difference between these tests is the
with Downs syndrome. They discovered, however, type of processing required in each. The reading-
that differences in short-term recall associated with span test requires verbal processing specific to read-
intellectual ability occurred despite the fact that none ing in addition to verbal storage. The individual must
of the children used rehearsal as a memory aid. Fur- read a set of sentences and, after finishing reading all
ther, these memory differences occurred only during of the sentences in the set, recall the last word of each
short-term recall, in which children with Downs syn- sentence. The speaking-span test is similar in calling
drome demonstrated memory deficits that the other for the remembering of a set of target words but calls
children did not; all of the children had equally poor for the presentation of the target words first; the indi-
near-immediate recall and equally good immediate vidual must remember the words while using each of
recall. Jarrold, Baddeley, and Hewes suggested that them to orally generate a sentence, a task that re-
one possible explanation for these results could be quires verbal fluency. The operation-span test, creat-
encoding limitationeven though their perceptual ed by Marilyn Turner and Randall Engle (1989), re-
processing was equivalent, children with Downs syn- quires mathematical processing: After completing a
drome could not transform environmental informa- set of simple mathematical operations, the individual
tion into system input (i.e., memory traces) as effi- must recall the word associated with each one.
ciently as the other children with lesser intelligence
Findings from research using these memory tests
deficits could.
suggest that the correlation between memory and
This conclusion mirrors that of Norman Ellis and task performance depends on the degree of similarity
Darlene Meador (1985), who investigated mnemonic between the type of processing required by the task
strategies and short-term recall deficits in retarded and the memory test. Daneman (1991) found, for ex-
children. They presented children with an experi- ample, that speaking-span scores showed a stronger
mental stimulus and then, after a delay, presented a relation to a measure of oral fluency than did read-
probe stimulus. They asked the children to compare ing-span scores. Conversely, reading-span scores
the probe stimulus with their memory of the experi- showed a stronger relation to oral reading skills than
mental stimulus and indicate whether they matched. did speaking-span scores. Daneman and Hannon
As expected, recognition accuracy on this task de- (2001) demonstrated that operation-span scores
creased as IQ decreased. More important, rates of showed a relatively weaker correlation to reading
forgetting across different levels of IQ were equiva- comprehension than did reading-span scores.
lent on this task, even with the precluding of mne- Not all research agrees with the conclusion the
monic strategies. Performance differences correlated processes on memory tests must be task-specific. In
with IQ differences with the simultaneous presenta- contrast, Engle his colleagues have repeatedly dem-
tion of experimental and probe stimuli, even at reten- onstrated that the relationship between memory and
tion intervals of twenty seconds. cognition is not mediated by efficient task-specific
processing. Engle, Kane, and Tuholski (1999) de-
scribed a series of studies in which task-specific pro-
The Role of Processing cessing efficiency was statistically or experimentally
As illustrated in Figure 1, after encoding, infor- controlled but in which the relationship between
mation must remain temporarily active during its working memory and performance on various tests
support of cognitive activity. Limitations in simulta- was not eliminated. In one study, the mathematical
neously storing and processing information have im- processing demands of the operation span test were
plications for complex cognition because the active equated across individuals. The mathematical ability
maintenance and processing of information in work- of each participant was determined, and the difficulty
ing memory plays an important role in intelligent be- of the operations administered in the operation span
havior. There are, however, differing views on the na- test was tailored to ability level. Controlling for indi-
ture of limitations in this aspect of the working- vidual differences in mathematical ability failed to re-
memory system, whether it is task-specific or task- duce the strong correlation between performance on
independent. the memory test and reading comprehension.
Meredyth Daneman and her colleagues, Patricia The operation span test required the allocation
Carpenter and Brenda Hannon, have tested the hy- of attention to both completing mathematical opera-
pothesis that the efficiency of task-specific processing tions and remembering target words, but equating
is the critical link between working memory and com- mathematical ability did not reduce the correlation
plex cognition. This work and related research between working memory and complex cognition.
266 INTELLIGENCE AND MEMORY
Figure 2
Engle and his colleagues argued, therefore, that do- found that performance on the reading span test cor-
main-free controlled attention is the essence of work- related substantially with verbal Scholastic Assess-
ing-memory limitations and drives the relationship ment Test (SAT) scores and reading comprehension,
between measures of working memory and measures whereas a test of verbal storagememorizing a list of
of intelligence. They claimed that controlled atten- wordsshowed only moderate correlations with SAT
tion activates information, either from the immediate scores and reading comprehension. Consistent with
environment or from long-term memory, and main- these findings, Engle and his colleagues (1999) used
tains that information in memory while it is being structural equation modeling techniques to demon-
processed, particularly in the face of distraction. strate that working memory had a stronger correla-
Tuholski, Engle, and Baylis (2001) found that perfor- tion with general reasoning than did short-term stor-
mance on the operation span test could predict dis- age. They also showed that working-memory
tractibility on a computerized counting task, indicat- measures were better predictors of verbal and quanti-
ing the role of controlled attention in reducing the tative SAT scores than was short-term storage.
impact of distraction.
Regardless of disagreement over the nature of
working-memory limitations, the work of both Dane- The Role of Knowledge and Skills
man and her colleagues and Engle and his colleagues An emerging interest in understanding everyday
clearly indicates that the storage aspect of memory cognitive functioning has turned the focus of some re-
alone does not account for its strong relationship to searchers to the role of long-term knowledge and
complex cognition. Daneman and Carpenter (1980) skills in the memory system. These researchers are
wary of experimental results based on performance understanding of the critical role of information pro-
on simple laboratory tasks such as simple mental cal- cessing in complex cognition.
culations. They argue that, in typical college student
samples, such tasks highlight working memory de-
Bibliography
mands at the expense of acquired knowledge and ex-
Atkinson, R. C., and Shiffrin, R. M. (1968). Working memory. In
perience, which is a critical aspect of everyday func- G. H. Bower, ed., The psychology of learning and motivation: Ad-
tioning. Instead, they study how highly developed vances in research and theory, Vol 2. New York: Academic Press.
skills and knowledge influence memory performance. Baddeley, A. D., and Carpenter, P. A. (1980). Individual differ-
ences in working memory and reading. Journal of Verbal Learn-
Anders Ericsson and Peter Delaney (1999) pro- ing and Verbal Behavior 19, 450466.
posed that the functioning of the working-memory Baddeley, A. D., and Hannon, B. (2001). Using working memory
system during highly skilled performance is deter- theory to investigate the construct validity of multiple-choice
mined by specialized encoding and retrieval mecha- reading comprehension tests such as the SAT. Journal of Ex-
perimental Psychology: General 130, 208223.
nisms formed over extended periods of practice.
Baddeley, A. D., and Hitch, G. (1974). Working memory. In G. H.
They claimed that individuals use these mechanisms Bower, ed., The psychology of learning and motivation: Advances
to encode information from the environment into in research and theory, Vol. 2. New York: Academic Press.
long-term memory to enable efficient and extensive Baddeley, A. D., and Logie, R. H. (1999). Working memory: The
retrieval of this information later on. These mecha- multiple-component model. In A. Miyake and P. Shah, eds.,
nisms allow the individual to demonstrate an unex- Models of working memory: Mechanisms of active maintenance and
executive control. Cambridge, UK: Cambridge University Press.
pectedly large skill-specific memory capacity because
Daneman, M. (1991). Working memory as a predictor of verbal flu-
of the ready availability of information relevant to ency. Journal of Psycholinguistic Research 20, 445464.
that skill that readily available in long-term memory. Ellis, N., and Meador, D. M. (1985). Forgetting in retarded and
nonretarded persons under conditions of minimal strategy
Ericsson and Delaney found support for their use. Intelligence 9, 8796.
theory in the verbalizations of memory experts, peo- Engle, R. W., Kane, M. J., and Tuholski, S. W. (1999). Individual
ple who, through training, have developed above- differences in working memory capacity and what they tell us
average memory skills. These verbal reports indicat- about controlled attention, general fluid intelligence, and
functions of the prefrontal cortex. In A. Miyake and P. Shah,
ed that memory experts use existing knowledge to en-
eds., Models of working memory: Mechanisms of active maintenance
code information into larger chunks for later and executive control. Cambridge, UK: Cambridge University
retrieval. For example, to memorize a very large Press.
string of digits, one expert reported encoding the Engle, R. W., Tuholski, S. W., Laughlin, J. E., and Conway, A. R.
digits as mile running times and retrieving the digit A. (1999). Working memory, short-term memory and general
string as groups of digits rather than as individual fluid intelligence: A latent variable approach. Journal of Exper-
imental Psychology: General 128, 309331.
numbers. When digits were presented in such a way
Ericsson, K. A., and Delaney, P. F. (1999). Long-term working
that they could not be associated with existing knowl- memory as an alternative to capacity models of working mem-
edge (e.g., they would make nonsensical running ory in everyday skilled performance. In A. Miyake and P.
times), there was no expansion of memory. In addi- Shah, eds., Models of working memory: Mechanisms of active main-
tion, Ericsson and Delaney reported that chess ex- tenance and executive control. Cambridge, UK: Cambridge Uni-
perts not only encoded individual chess pieces ac- versity Press.
Jarrold, C., Baddeley, A. D., and Hewes, A. K. (2000). Verbal short-
cording to meaningful configurations but also term memory deficits in Down Syndrome: A consequence of
developed retrieval structures associated with loca- problems in rehearsal? Journal of Child Psychology and Psychiatry
tions on the chessboard. and Allied Disciplines 41, 223244.
Kyllonen, P. C. (1996). Is working memory capacity Spearmans g?
Although memory assessment is still a critical as- In I. Dennis and P. Tapsfield, eds., Human abilities: Their na-
pect of commercial intelligence tests, such as the ture and measurement. Mahwah, NJ: Erlbaum.
WAIS-III and its companion Wechsler Memory Scale (2002). g: knowledge, speed, strategies, or working
(WMS), memory testing has changed. In contrast to memory capacity? A systems perspective. In R. J. Sternberg
and E. L. Grigorenko, eds., The general factor of intelligence:
the WAIS-R, the WAIS-III no longer features digit
How general is it? Mahwah, NJ: Erlbaum.
span as a mandatory test, but it does feature several Kyllonen, P. C., and Stephens, D. L. (1990). Cognitive abilities as
additional tests of working memory. Also consistent determinants of success in acquiring logic skill. Learning and
with the findings of recent research, the WMS fea- Individual Differences 2, 29160.
tures tests of immediate, delayed, and working mem- Miyake, A., and Shah, P. (1999). Toward unified theories of work-
ory. Instead of reflecting a static storage capacity, the ing memory. In A. Miyake and P. Shah, eds., Models of working
memory: Mechanisms of active maintenance and executive control.
memory aspect of modern intelligence tests now re-
Cambridge, UK: Cambridge University Press.
flects the dynamic functioning of a complex, multifa- Tuholski, S. W., Engle, R. W., and Baylis, G. C. (2001). Individual
ceted information-processing system. Such changes differences in working memory capacity and enumeration.
in memory and intelligence testing indicate growing Memory & Cognition 29, 484492.
268 INTERFERENCE AND FORGETTING
Turner, M. L., and Engle, R. W. (1989). Is working memory capaci- when two German researchers, Georg Elias Mller
ty task dependent? Journal of Memory and Language 28, 127 and A. Pilzecker, first demonstrated retroactive inter-
154.
ference under controlled conditions. The history of
Anna T. Cianciolo that research interests contemporary scientists, partly
Robert J. Sternberg because it is a case where intuition proved a poor
guide to theorizing.
Early Theories That Proved Inadequate
INTERFERENCE AND FORGETTING Consolidation. Mller and Pilzecker (1900)
found that subjects memory for a series of nonsense syl-
Human long-term memory is characterized by a near- lables (consonant-vowel-consonant nonword syllables,
ly limitless storage capacity. At any time, however, such as DAX) was impaired by subsequent activity, such
much of the information that exists in long-term as learning a new series of nonsense syllables (compared
memories (names, numbers, facts, procedures, with a condition where subjects simply rested for a simi-
events, and so forth) is not recallable. Why do people lar period of time). They put forward a perseveration-
forget information that was once recallable? Because consolidation hypothesis to explain their results. They
access to information in memory is subject to interfer- argued that the changes in the nervous system that result
ence from competing information in memory. Before in true learning are not complete by the end of train-
characterizing such interference processes in more ingthat activity in the brain perseverates after learn-
detail, it is necessary to introduce some terminology. ing, and that during that perseveration the memory
The first concept is transfer. After some early traces corresponding to learning are consolidated. A
point in life, people rarely, if ever, learn anything that subsequent activity, particularly if demanding and close
is entirely new. Rather, people bring to any new in time to the original learning task, can disrupt the per-
learning of knowledge or skills an accumulation of re- severation process, resulting in retroactive interference.
lated knowledge, skills, and habits from the past. Such
The consolidation idea seems plausible, especial-
prior learning influences the qualitative and quantita-
ly given the evidence that certain traumas, such as
tive character of the new learning process. Such trans-
electroconvulsive shock or a head injury, can produce
fer effects may be positive or negative, depending on
retrograde amnesia (loss of memory for events occur-
whether prior experiences facilitate or impair the new
ring just prior to the injury), and that a period of
learning process.
sleep after a learning session produces less forgetting
The second concept is retroactive interference. than does a comparable period of waking activity.
Whereas transfer refers to the effect of earlier learn- The consolidation hypothesis proved unsatisfactory,
ing on later learning, retroaction refers the impact of however, because it does not provide an account for
interpolated (intervening) learning experiences on a variety of empirical phenomena. Long after the per-
ones memory for something teamed earlier. Once severation-consolidation process should be complete,
again, such effects may be positive or negative (retro- for example, interpolated learning still produces sub-
active facilitation and interference, respectively), de- stantial retroactive interference. Other problematic
pending on the similarity of the original and interpo- findings are that increasing the intensity of an unre-
lated learning tasks. It is the negative casewhere lated interpolated activity results in little or no in-
retroactive interference causes forgettingthat ap- crease in forgetting, whereas increasing intertask sim-
plies to this discussion. Thus, if ones ability to recall ilarity does play an important role in forgetting. A
the maiden name of a woman friend is impaired by final blow for the theory is that it cannot explain
virtue of having learned her married name, one is suf- proactive interference.
fering from retroactive interference.
Decay. An explanation of forgetting that seems
The third concept is proactive interference. Some- particularly plausible was put forth by Edward Thorn-
thing learned earlier may also impair ones ability to dike (1914) as his so-called law of disuse. The thrust of
recall something learned more recently. If, for exam- his law is straightforward: Unless a person continues
ple, one is less able to recall a woman friends married to access and use the memory representations corre-
name by virtue of having learned her maiden name sponding to skills and information, those representa-
at an earlier time, one is suffering from proactive in- tions decay. Learning processes create memory repre-
terference. sentations; practice maintains those representations; but
they fade with disuse.
A Brief History of Research on Forgetting The decay theory seems in general agreement
Rigorous research on the possible causes of for- with the average persons introspections as to how
getting dates back to the turn of the twentieth century memories are formed and lost, but it proved entirely
INTERFERENCE AND FORGETTING 269
inadequate as a theory of forgetting. Thorndikes law list items during the relearning of the first list actually
was thoroughly discredited in a devastating critique decreased with high levels of training on the second list.
by John McGeoch (1932). Among the problems with They argued that response competition could not, there-
the theory are that forgetting is a function not simply fore, be the sole factor contributing to retroactive inter-
of disuse across some retention interval but also of the ference, because such intrusions are a straightforward
nature of the activity in that interval, particularly its measure of such competition. They proposed a second
similarity to what is being remembered; information factor: unlearning of first-list responses during second-
appears not to be lost from memory in some absolute list learning. Their idea, which must have seemed some-
sense, as implied by the theory, but, rather, becomes what bizarre, is analogous to a basic result in the animal-
nonrecallable except under special circumstances; learning literature: Learned responses are gradually ex-
and it does not account for proactive interference. tinguished when no longer reinforced by a reward of
some kind. From that perspective, intrusions of first-list
The Emergence of Interference Theory responses during second-list learning constitute unrein-
As an alternative to the consolidation and decay forced errors.
ideas, McGeoch (1932, 1942) put forth the initial ver- As if the unlearning idea were not strange enough
sion of what came to be called interference theory. by itself, it had an additional counterintuitive implica-
That theoretical framework, as modified and refined tion: If unlearning is analogous to experimental ex-
over subsequent decades, constitutes the most signifi- tinction, thenas in animal-learning researchthe
cant and systematic theoretical formulation in the unlearned responses should show spontaneous recov-
field of human learning and memory. ery over time. That is, the unlearned responses
Reproductive Inhibition. McGeoch argued that should recoverbecome more available in memo-
human memory is fundamentally associativethat recall ryas time passes following the retroactive learning
is guided by cues or stimuli to which items in memory are episode. Such an implication seems to violate a law or
associated. As a consequence of a given individuals vari- first principle of memorynamely, that items in
ous experiences, however, multiple items in memory (re- memory become less available with time. However
sponses) may become associated to the same cue. The re- unintuitive the unlearning/spontaneous recovery idea
call of a given target response to a given cue, then, can may seem, research carried out over the twenty years
suffer competition from other responses associated to or so following the Melton and Irwin (1940) paper
that cue. Such competition, according to McGeoch, pro- provided unambiguous support for the basic idea (see
duces forgetting through reproductive inhibition: Recall Barnes and Underwood, 1959; Briggs, 1954; Under-
of the target response is blocked or inhibited by the re- wood, 1957).
trieval of other responses associated to that cue. Those By the late 1960s the basic interplay of proactive
other responses may have been learned before or after and retroactive interference had become clear. The
the response in question (proactive and retroactive in- dynamics of that interplay are summarized in the
terference, respectively), and such interference should next section of this entry. More complete versions of
be a function of intertask similarity across learning epi- the history and final state of classical research on
sodes. interference and forgetting are available in Roberta
Another factor in forgetting, according to Mc- Klatzky (1980), Gordon Bower and E. Hilgard (1981),
Geoch, is that the stimulus conditions existing at the Robert Crowder (1976), and Leo Postman (1971).
time recall is tested will differ from the conditions
that existed during training. Such differences are The Dynamics of Interference and
likely to increase as the interval from training to test Forgetting
grows longer; and to the degree the stimulus condi-
Figure 1 summarizes the dynamics of interfer-
tions at test do differ they will become less effective
ence and forgetting. Assume that the original learn-
as cues for the response that was the target of train-
ing episode involves learning to associate each mem-
ing.
ber, B, of a set of responses with a particular member,
Unlearning and Spontaneous Recovery. In a A, of a set of stimuli. Assume further that the new (in-
pivotal study, Arthur W. Melton and J. M. Irwin (1940) terpolated) learning episode involves associating
took issue with McGeochs analysis of retroactive inter- each member, D, of a different set of responses with
ference. In their experiment, subjects learned two simi- a particular member, A, of a set of stimuli that may
lar lists of verbal items and then were asked to relearn vary from being only generally similar to the A stimuli
the first list. They found that the retroactive interference to being essentially identical. At the time memory is
caused by the second list was, as predicted, an increasing tested, assume that a given member of stimulus set A
function of the number of learning trials on the second or A is presented as a cue for the associated B or D
list, but that the frequency of overt intrusions of second- response.
Figure 1 (A-B) are gradually suppressed or extinguished. Such

suppression facilitates A-D learning by reducing the
negative transfer from competing B responses, but it
also impairs any subsequent efforts to recall B re-
sponses. On the basis of a considerable body of re-
search (particularly McGovern, 1964; Postman et al.,
1968), it appears thatdepending on the relation-
ship of the stimulus-response pairings in the two
learning episodesone or more of three distinct
types of unlearning may take place. Forward associa-
tions (from A to B) can be unlearned (which facilitates
A-D learning), backward associations (from B to A)
can be unlearned (which would, for example, facili-
tate C-B learning, where C denotes stimuli that are
not similar to A), and the entire set of B responses can
be suppressed (which would aid A-D or C-D learn-
ing).
Spontaneous Recovery
During the retention interval following A-D
learning (typically filled with other real-world activi-
ties on the part of the learner) the A-B associations
that were suppressed during A-D learning gradually
recover in strength. Any other preexperimental asso-
ciations to a given A or A stimulus that may have been
learned prior to A-B or A-D learning will recover in
strength as well.
Thus, at the end of A-D learning, the D re-
sponses will be highly accessible in memory and the
Summary of the processes underlying proactive and retroactive
B responses will be relatively inaccessible (the exact
interference; A-B and A-D denote associative learning tasks in
ratio of B and D strengths will depend, of course, on
which A and A are similar or identical stimuli and B and D are
different responses. the initial levels of A-B and A-D learning, and on the
overlap of the A and A stimuli). As the retention in-
terval from the end of A-D learning increases, howev-
er, the pattern changes: The D responses become less
The A-B, A-D notation is meant to be interpreted
recallable as the interval increases, and the B re-
quite broadly. A given stimulus might be a persons
sponses become relatively or absolutely more recall-
face and the response that persons name, for exam-
able. Whether the B responses themselves become
ple, and the number of A-B and A-D pairings to be
more recallable in absolute terms appears to depend
learned might vary from one of each to some large
on whether those responses are also in competition
number (as in the case of a grade-school teacher
with other (recovering) responses learned prior to the
learning the names of the students in each years
A-B episode. If a given A-B association is itself subject
class). In certain cases the stimulus might actually cor-
to proactive interference from one or more prior as-
respond to a configuration of stimuli and the re-
sociations (A-E, A-F, and so forth), recall of the B re-
sponse might be a coordinated set of verbal or motor
sponse will tend to decrease, not increase, as the re-
responses (A-B and A-D, e.g., could refer to learning
tention interval increases.
to operate two different automobiles, the first in the
United States and the second in England). The time As the B responses (and any other prior associa-
course of the A-B and A-D learning episodes might tions to a given A stimulus) recover, recall of the D
vary from brief to very extended (as would be the case responses will suffer increasing proactive interfer-
if A-B denotes learning to label objects in a first lan- ence. One implication of such recovery is that the rate
guage and A-D denotes learning to label those same of forgetting of D responses after A-D training
objects in a second language). should be a function of the number of preceding simi-
lar lists a subject has learned. In an analysis of the re-
Unlearning sults of many experiments reported in the literature,
During the new learning episode (A-D), compet- Underwood (1957) found striking support for that
ing responses from the original learning episode prediction.
INTERFERENCE AND FORGETTING 271
Response Competition ing all the items in that set that exist in memory be-
At the time a given A or A stimulus is presented cause the early items recalled impede the recall of
as a cue for recall of the appropriate B or D response subsequent items; having been recalled, the early
learned earlier, that target response will be in compe- items become more accessible in memory and block
tition with any other responses associated to that stim- access to yet-to-be-recalled items.
ulus. The impact of that competition is to inhibit ac- Similar dynamics are probably at work in the in-
cess to the target response in memory. In general, hibitory consequences of part-list cuing. When some
recall of a given target response will decrease as the members of a list or category of items are presented
number and strength of competing responses in- to subjects as cues to aid their recall of the remaining
creases. That generalization, in more modern terms, items, the recall of those remaining items is typically
is the cue-overload principle (Watkins and Watkins, hindered rather than helped. Such inhibitory effects,
1975). In the analysis of such response competition, considered an enigma in memory research (Nicker-
however, an important distinction is relevant. It is the son, 1984), are at least in part a consequence of the
functional stimulus, not the nominal stimulus, that cued items becoming too available in memory.
cues the retrieval of items in memory. Thus, A and A
may be nominally identical or highly similar stimuli, Retrieval-Induced Forgetting
but if the learning episodes involving those stimuli The negative consequences of retrieval have been
differ substantiallyin terms of the environmental, examined more explicitly via a retrieval-practice par-
temporal, or social context, or even in terms of the adigm introduced by M. C. Anderson, R. A. Bjork,
learners emotional or physical statethe functional and E. L. Bjork (1994). The procedure includes an
encoding of those stimuli may differ markedly. Thus, initial study phase, during which a number of catego-
any stimulus, together with the context in which it is ry-exemplar pairs (such as Fruit-Orange) are pre-
embedded, offers the learner a variety of aspects that sented. Typically, about forty-eight pairs are pre-
may be sampled (Estes, 1955) or attended to, and sented, which might be composed, for instance, of six
that process determines the functional encoding of a exemplars of each of eight categories. There is then
given stimulus. a retrieval-practice phase, during which participants
Consistent with the foregoing analysis, the degree are cued to recall, in response to cues such as Fruit-
to which different learning episodes result in later re- Or____, half of the exemplars of half of the studied
sponse competition depends on how discriminable categories multiple times. Finally, after a retention in-
on one basis or anotherthose episodes are from terval of twenty minutes or so, a surprise test is admin-
each other at the time of test. The more such episodes istered during which participants are presented each
are separated from each other temporally, for exam- of the category names and asked to recall all of the ex-
ple, the less they will interfere with each other (Un- emplars they can remember having been paired with
derwood and Ekstrand, 1967). that name during the study phase.
Not surprisingly, the practiced members of prac-
ticed categories are recalled better on the final test
Retrieval as a Memory Modifier
than are corresponding unpracticed members of un-
The results of research conducted in the late practiced categories. In contrast, the unpracticed
twentieth century add to the body of research. There members of practiced categories are actually recalled
is abundant evidence that the recall process alters the worse than are corresponding unpracticed members
relative accessibility of items in memory. The act of of unpracticed categories. That is, there seems to be
recall is itself a learning event in the sense that an retrieval-induced forgetting of exemplars that are in
item recalled in response to a given cue becomes the same category as the practiced exemplars, but are
more recallable in the future. One consequence of not themselves practiced.
such response-produced strengthening of future ac-
cess to recalled items, however, is that other items as- The findings obtained with various versions of
sociated with that cue may become less recallable. the retrieval-practice paradigm suggest that retrieval-
That is, the recall process can alter the pattern of rela- induced forgetting is a consequence of suppression of
tive access strengths across the set of items associated not-to-be-recalled exemplars during the retrieval-
to a given cue. practice phase. Thus, correctly recalling Orange in
response to a cue such as Fruit-Or____ requires not
Output Interference and Part-List Cuing only selecting from among the studied items the fruit
Consistent with the foregoing argument, there is that fits Or____, but also not recalling (suppressing)
evidence that recall is a self-limiting process other studied exemplars in the fruit category, such as
(Roediger, 1978). Attempts to recall the members of Banana. M. C. Anderson and B. A. Spellman (1995)
a category or list of items occasion difficulty in recall- have suggested that such selection-suppression pro-
cesses are simiar to those known to categorize human tion still exists in memoryin contrast to overwritten
attention and that retrieving targeted information items in a computers memoryit tends to be recog-
from memory requires a type of conceptually focused nizable and readily relearned should the need arise.
attention, one consequence of which is retrieval- Finally, should one stop using the new information
induced forgetting of unselected items. (e.g., how to drive in Britain), there will be some re-
covery of the old information (e.g., how to drive in the
United States), which will often be adaptive as well.
Conclusion In general, it appears that differences in accessi-
Interference and transfer are fundamental to bility across the vast number of items in memory acts
human learning, memory, and performance. After a as a kind of filter. The information and skills most
period of almost twenty years, from roughly 1970 to readily accessible in human memories will tend to be
1990, during which research on interference and for- those people have been using in the recent past. On
getting was not a dominant theme in experimental a statistical basis, those are the same skills and knowl-
psychology, there has been a resurgence of interest in edge people will tend to need in the near future
such phenomena. Several contributing factors in that
resurgence can be mentioned. First, there is renewed See also: FORGETTING; MCGEOCH, JOHN A.; MEMORY
evidence of and appreciation for the role inhibitory CONSOLIDATION: PROLONGED PROCESS OF
REORGANIZATION; MLLER, GEORG ELIAS;
processes play in human cognition. Second, in certain
RECONSTRUCTIVE MEMORY; THORNDIKE,
applied fields, such as research on memory factors in EDWARD
advertising and witness testimony, there is a need to
understand how successive inputs to memory com- Bibliography
pete and interact. In research on witness memory, for Anderson, M. C., Bjork, R. A., and Bjork, E. L. (1994). Remember-
example, an issue of intense concern is how memory ing can cause forgetting: Retrieval dynamics in long-term
representations are modified by misleading postevent memory. Journal of Experimental Psychology: Learning, Memory,
and Cognition 20, 1,0631,087.
information. Third, among researchers who are Anderson, M. C., and Spellman, B. A. (1995). On the status of in-
working to implement and test various types of math- hibitory mechanisms in cognition: Memory retrieval as a
ematical and computer models of human memory model case. Psychological Review 102, 68100.
there is a growing realization that any plausible Barnes, J. M., and Underwood, B. J. (1959). Fate of first-list associ-
model must account for the basic patterns of proac- ations in transfer theory. Journal of Experimental Psychology 58,
97105.
tive and retroactive interference (Mensink and Raaij- Bjork, R. A. (1989). Retrieval inhibition as an adaptive mechanism
makers, 1988). in human memory. In H. L. Roediger and F. I. M. Craik, eds.,
Varieties of memory and consciousness: Essays in honour of Endel
Forgetting is not simply a failure or weakness of Tulving. Hillsdale, NJ: Erlbaum.
the memory system. In terms of the ovrerall function- Bower, G. H., and Hilgard, E. R. (1981). Theories of learning, 5th
ing of the system, there must be some means of re- edition. Englewood Cliffs, NJ: Prentice-Hall.
stricting what is retrieved in response to a given cue: Briggs, G. E. (1954). Acquisition, extinction and recovery functions
in retroactive inhibition. Journal of Experimental Psychology 47,
Information that is out of date or inappropriate needs
285293.
to be suppressed, segregated, or eliminated. During Crowder, R. G. (1976). Principles of learning and memory. Hillsdale,
the attempt to recall ones home phone number, or NJ: Erlbaum.
where one left the car, for example, it is not useful to Estes, W. K. (1955). Statistical theory of spontaneous recovery and
retrieve ones prior home phone number, or where regression. Psychological Review 62, 145154.
Klatzky, R. L. (1980). Human memory: Structures and processes, 2nd
one left the car yesterday or a week ago. In short, in
edition. San Francisco: Freeman.
terms of speed and accuracy of the recall process, one McGeoch, J. A. (1932). Forgetting and the law of disuse. Psychologi-
does not want everything that exists in ones memory cal Review 39, 352370.
to be accessible, especially given the essentially unlim- (1942). The psychology of human learning. New York: Long-
ited capacity of human memory. mans, Green.
McGovern, J. B. (1964). Extinction of associations in four transfer
There are clearly some adaptive functions of the paradigms. Psychological Monographs 78 (16, whole no. 593).
Melton, A. W., and Irwin, J. M. (1940). The influence of degree of
interference mechanisms that underlie forgetting. As
interpolated learning on retroactive inhibition and the overt
a person continues to learn and continues to use new transfer of specific responses. American Journal of Psychology 53,
information, for example, access to the out-of-date in- 173203.
formation it replaces is inhibited. Such retrieval inhi- Mensink, G. J., and Raaijmakers, J. G. W. (1988). A model for in-
bition has several advantages over the kind of de- terference and forgetting. Psychological Review 93, 434455.
Mller, G. E., and Pilzecker, A. (1900). Experimentelle Beitrge
structive updating of memory characteristic of
zur Lehre von Gedchtnis. Zeitschrift fr Psychologie 1, 1300.
computers (Bjork, 1989). Because the old informa- Nickerson, R. S. (1984). Retrieval inhibition from partlist cuing: A
tion is inhibited, it tends not to interfere with the re- persisting enigma in memory research. Memory & Cognition
call of the new information, but because that informa- 12, 531552.
INVERTEBRATE LEARNING: Associative Learning and Memory Processing in Bees 273
Postman, L. (1971). Transfer, interference, and forgetting. In J. W. ASSOCIATIVE LEARNING AND MEMORY
Kling and L. A. Riggs, eds., Woodworth and Schlosbergs experi-
mental psychology, 3rd edition. New York: Holt, Rinehart and
PROCESSING IN BEES
Winston. The social life of the honeybee colony forms the eco-
Postman, L., Stark, K., and Fraser, J. (1968). Temporal changes in
logical framework for the individual animals behav-
interference. Journal of Verbal Learning and Verbal Behavior 7,
672694. ior and is crucial for each bees survival, because an
Roediger, H. L. (1978). Recall as a self-limiting process. Memory & individual bee cannot exist on its own (Frisch, 1967;
Cognition 6, 5463. Lindauer, 1967).
Thorndike, E. L. (1914). The psychology of learning. New York:
Teachers College Press.
Underwood, B. J. (1957). Interference and forgetting. Psychological
Review 64, 4960.
Associative Learning
Underwood, B. J., and Ekstrand, B. R. (1967). Studies of distribut- The study of learning and memory formation in
ed practice: XXIV: Differentiation and proactive inhibition. bees under natural conditions has focused on latent
Journal of Experimental Psychology 74, 574580.
Watkins, O. C., and Watkins, M. J. (1975). Buildup of proactive in-
learning during navigation and on operant learning
hibition as a cue-overload effect. Journal of Experimental Psy- in the context of food collection. In the laboratory it
chology: Human Learning and Memory 104, 442452. has focused on appetitive classical conditioning. Bees
departing from the hive perform observatory learn-
Robert A. Bjork ing flights (Capaldi et al., 2000), and establish a map-
like spatial memory for their colonys location relative
to landmarks within the framework of their sun com-
pass system (Menzel et al., 2000). When a searching
bee discovers a nectar- or pollen-producing flower, it
INVERTEBRATE LEARNING quickly learns to associate the surrounding visual and
[Invertebrates are particularly useful for analyzing the neu- olfactory signals with the reward. It learns olfactory
ral and molecular events underlying learning and memory. stimuli (e.g., floral odorants) and colors very quickly
The nervous systems of many invertebrates contain only sev- (within one or a few learning trials). Patterns need
eral thousand cells (compared with the billions of cells in the more learning trials. Whereas latent learning during
vertebrate nervous system). Despite the small number of cells, navigational tasks may not require a rewarding stimu-
an invertebrate ganglion can control a variety of behaviors. lus, reward learning is a forward-associative process
A given behavior may, therefore, be mediated by 100 or because signals perceived before the reward are asso-
fewer neurons, and this small size of the circuit makes com- ciated, whereas those perceived during the reward or
plete description easier. Moreover, many neurons are rela- during the departure flight are associated less effec-
tively large and can be repeatedly identified as unique indi- tively or not at all.
viduals, permitting one to examine the functional properties Research on various operant learning phenome-
to a specific behavior mediated by the cell. Changes in cellu- na (e.g. reversal and multireversal learning, over-
lar properties that occur when a behavior is modified by learning, inhibitory learning, context-dependent
learning can then be related to specific changes in behavior. learning, and reward schedule learning) has found
Molecular and biophysical events underlying the changes in performances similar to those in mammals (Couvillon
cellular properties can then be determined. This approach and Bitterman, 1988; Menzel, 1990). Multiple expe-
has been particularly successful with the bee and the mol- rience with varying signals but one constant feature
lusks Aplysia, Hermissenda, Limax, and Tritonia. (e.g., different kinds of symmetrical patterns) leads to
Invertebrates are also excellent subjects for a genetic dis- the formation of a concept (the concept of symmetry)
section of behavior and learning and memory. Two animals that allows the bee to choose new patterns with the
that have been particularly useful are the fruit fly Drosoph- same feature as learned targets (Giurfa, Eichmann,
ila and the worm C. elegans. The basic strategy is to alter and Menzel, 1996). Bees also develop a concept of
the genotype with a mutagen and to test for specific defects sameness and difference when they are trained in de-
in the ability of the animals to learn or remember. The role layed matching-to-sample tasks, in which they are re-
of individual biochemical processes and genes then can be quired to respond to a matching stimulus or a non-
related to specific aspects of learning and memory. matching stimulus (Giurfa et al., 2001). They also
transfer the learned rules to new stimuli of the same
The entries that follow discuss each of these inverte- or a different sensory modality. Thus, not only can
brates except APLYSIA, which is the subject of a separate sec- bees learn specific objects and their physical parame-
tion. For additional information on insect species, see IN- ters, but they also extract rules and apply them to
SECT LEARNING.] novel situations.
274 INVERTEBRATE LEARNING: Associative Learning and Memory Processing in Bees
Classical Conditioning introduced in theories about associative learning. It

is thus interesting from a comparative point of view
Classical conditioning of reflexes is a convenient
whether the bee with its tiny brain shows the blocking
way to study the behavioral and neural mechanisms
phenomenon. This question cannot yet be definitive-
of associative learning. In the honeybee, the probos-
ly answered. Blocking is found in some studies (Smith
cis-extension reflex (PER) to a sucrose stimulus at the
and Cobey, 1994; Thorn and Smith, 1997), but not in
antennae is a reliable reflex in the context of feeding.
others (Gerber and Ullrich, 1999). Blocking across
A hungry bee will reflexively extend its proboscis
sensory modalities was also not seen in training free-
(tongue) when the antennae are touched with a drop
flying bees (Bitterman, 1996; Funayama, Couvillon,
of sucrose solution. An odor (conditioned stimulus,
and Bitterman, 1996). Second-order conditioning is
CS) presented shortly before the sucrose reward (un-
another procedure that tests whether associative
conditioned stimulus, US) will be associated with the
learning requires contiguity between CS and US. In
reward, even under conditions in which the animal is
a positive outcome of second-order conditioning one
harnessed in a tube or is being prepared for physio-
argues that a CS can acquire the potential of a US.
logical studies (Menzel and Mller, 1996). The asso-
This has been demonstrated for olfactory PER condi-
ciative nature of PER conditioning to odors has been
tioning (Menzel, 1990).
established by demonstrating that only forward-
pairing of CS-US sequences is effective. Unpaired CS Rules of elementary associative learning assume
and US presentations or CS- or US-only presenta- that in learning a compound stimulus, animals learn
tions do not lead to learning, and in differential con- the associations between the reinforcer and the com-
ditioning (one odor CS+ paired with US, the other pound elements separately (Rescorla and Wagner,
CS- unpaired), bees respond only to the CS+ and not 1972). Contrary to this assumption, configural learn-
to CS- (Menzel, 1990). The predictive value of the CS ing theories assume that, in learning a compound, an-
depends on the reliability with which it is causally re- imals build a new entity made from the conjunction
lated to the US. In differential conditioning, the re- of compound elements and that a connection is made
versal to the initially unpaired stimulus CS- is slower between this new configuration and the reinforcer
after more frequent unreinforced preexposures than (Rudy and Sutherland, 1992). The different process-
after fewer preexposures. The same applies for US- ing strategies underlying elementary and configural
only preexposures in an otherwise reinforced con- olfactory learning were studied by the negative pat-
text, indicating that the absence of an expected US terning discrimination. In negative patterning two
leads to inhibitory learning. If naive animals are stim- single stimuli are reinforced (A+, B +), while the com-
ulated with a compound of two odors and one of them pound is not (AB-). Solving this problem
is later associated with sucrose reward, the animals responding less to the compound than to the single
will also respond the second odor of the compound, elementscan be explained only by taking configural
even though this odor was not explicitly experienced associations into account. Otherwise, summation of
the elementary associative strengths in the compound
during a learning trial (Mller et al., 2000). This form
should result in stronger response to the compound
of learning (sensory preconditioning) indicates stimu-
than to the elements. Honeybees can solve negative
lus associations between equally evaluated stimuli and
patterning discrimination in olfactory conditioning of
thus transcends the classical associative paradigm.
the PER (Deisig, Lachnit, Hellstern, and Giurfa,
The role of reinforcement in the formation of an 2001). The fact that bees solve negative patterning
association is an essential question in learning theo- discrimination in olfactory conditioning and in color/
ries: Are associations formed only by close contiguity odor tasks (Couvillon and Bitterman, 1988) shows
between the CS and US? The blocking phenomenon that linear associations between single stimuli and the
indicates that this is not the case: If a novel CS ap- reinforcer are not the only ones underlying associa-
pears together with a learned stimulus, this novel tive learning in honeybees (Giurfa et. al, 2001).
stimulus will be learned to a lesser extent or not at all
(Kamin, 1968). The blocking paradigm is central to
most current models of associative learning, and the Memory Dynamics and Memory
phenomenon is explained either by the assumption Localization
of a competition between the two CSs (the already Memory is an animals capacity to retain acquired
learned one and the novel one) for attention (Mackin- information and to use it for future behavior. In the
tosh, 1975) or for a limitation of reinforcing function context of association theory, memory is the potential
that depends on the expectation or prediction of re- of a conditioned stimulus to activate an established as-
inforcement (Rescorla and Wagner, 1972). Since at- sociative link. Some researchers, however, view learn-
tention, expectation and prediction are cognitive facing as acquiring information rather than responses,
ulties, it is argued that cognitive capacities need to be in which case memory would be a dynamic and self-
INVERTEBRATE LEARNING: Associative Learning and Memory Processing in Bees 275
Figure 1
A single learning trial leads to an early form of short-term memory (eSTM) that is accompanied by a short enhancement of PKA and
PKC activity. Consolidation to mid-term memory after a single trial (MTMs) is a time-dependent process lasting several minutes. The
molecular and cellular events related to the transition from eSTM to MTMs are unknown (see ?). MTMs decays after several hours but
retention is still significant after one day, indicating that even a single trial can induce longer-lasting forms of memory to a low
degree. Multiple learning trials lead to a succession of four memory phases that are arranged partially sequentially and partially in
parallel. Early and late STM (e/lSTM) are not separable, because consolidation is strongly facilitated by trial repetition, high retention
rates within the acquisition process, and strong resistance to extinction and reversal trials even immediately after conditioning. e/lSTM
is accompanied by stronger and longer-lasting PKA activation and an activation of NO synthase (NOS). Both cellular responses are
required for the transition to LTM, but may not be necessary for MTM formation. Transition to MTM after multiple trials (MTMm) is
accompanied by constitutive activation of PKM via a proteolytic pathway that is essential for MTMm formation. Blocking proteolysis,
however, does not inhibit the transition to the two forms of long-term memory (LTM), indicating parallel pathways from STM to
MTM and LTM. Inhibition of protein synthesis interferes only with the formation of lLTM. Massed conditioning leads predominantly
to eLTM, spaced conditioning to lLTM.
organizing process of information storage. Support formation is not identical to the process of acquisi-
for such a cognitive interpretation of memory in the tion. Memory needs time to develop and proceeds
honeybee comes from the fact that olfactory memory through phases that differ in their susceptibility to in-
276 INVERTEBRATE LEARNING: Associative Learning and Memory Processing in Bees
terfering events, their content, and their neural and aging at distributed and unreliable food sources (e.g.,
cellular substrates (Menzel, 1999; Menzel and Mller, flowers; Menzel, 1999; Giurfa et al., 2001).
1996; see Figure 1).
The memory trace for olfactory cues is distributed Conclusion
and involves two of the three convergence sites be- The honeybee provides a model system for the
tween the olfactory pathway and the reward pathway. study of neural substrates of low and intermediate le-
The reward pathway was identified by Hammer vels of cognitive faculties. Neural analysis is sup-
(1993) and assigned to a single identified neuron, the ported by robust forms of associative learning that
VUMmx1 neuron. Two of the three convergence occur even under conditions when intracellular re-
sitesantennal lobes and mushroom bodiesare, recordings or optophysiological measurements of sin-
spectively, the primary and secondary processing re- gle or multiple neuron activities are performed. The
gions in the olfactory pathway, and each of these two functional organization of the brain with a consider-
neuropils establishes its own memory trace indepen- able number of uniquely identifiable neurons is also
dently of the other (Hammer and Menzel, 1998). The advantageous for relating cognitive functions to neu-
two traces are, however, different at least with respect ral events in circumscribed circuits. Biochemical anal-
to their dynamics, and are likely to store different in- yses of the role of protein kinases (e.g., PKA, PKC)
formation. and enzymes (e.g., NO synthase) relate directly to be-
Researchers have made progress in unraveling havioral phenomena such as memory stages. Associa-
the neural correlates of memory for the antennal lobe tive processes in the bee brain are not restricted to el-
by visualizing the changes in odor coding as a conse- ementary forms but reflect configural processing and
quence of olfactory conditioning (Faber, Joerges, and context-dependent associations. Thus the bee brain
Menzel, 1999). The antennal lobe is organized into may serve as a model for the study of cognitive pro-
glomeruli; odors are coded as specific spatial- cesses at an intermediate level of complexity.
activation patterns of the glomeruli. These patterns See also: CONDITIONING, CELLULAR AND NETWORK
can be imaged using calcium-sensitive fluorescent SCHEMES FOR HIGHER-ORDER FEATURES OF;
dyes. As a result of conditioning, the neural represen- CONDITIONING, CLASSICAL AND
tation of a trained odor becomes more pronounced INSTRUMENTAL; INSECT LEARNING; KAMINS
and more distinct from nonrewarded odors, but its BLOCKING EFFECT: NEURONAL SUBSTRATES;
general features do not change, indicating that learn- SECOND MESSENGER SYSTEMS
ing at this level intensifies the neural code of the Bibliography
learned signal but does not create a new representa- Bitterman, M. E. (1996). Comparative analysis of learning in hon-
tion. It is unclear what specifies the odor-induced ac- eybees. Animal Learning Behavior 24, 123141.
tivity patterns as those of a learned odor, because a Botzer, D., Markovich, S., and Susswein, A. J. (1998). Multiple
stronger stimulus also induces a more pronounced memory processes following training that a food is inedible in
Aplysia. Learning and Memory 5, 204219.
and distinct activity pattern, but bees have no trouble
Capaldi, E. A., Smith, A. D., Osborne, J. L., Fahrbach, S. E., Farris,
distinguishing between a strong nonlearned odor and S. M., Reynolds, D. R., Edwards, A. S., Martin, A., Robinson,
a weak learned odor. G. E., Poppy, G. M., and Riley, J. R. (2000). Ontogeny of ori-
entation flight in the honeybee revealed by harmonic radar.
A stable, lifelong memory is formed even after Nature 403, 537540.
only a few learning trials as the result of sequential Couvillon, P.A., and Bitterman, M. E. (1988). Compound-
steps of memory processing. Five memory stages are component and conditional discrimination of colors and
distinguishuable on the basis of their respective tem- odors by honeybees: Further tests of a continuity model. Ani-
mal Learning and Behavior 16, 6774.
poral dynamics and their physiological and biochemi-
Deisig, N., Lachnit, H., Hellstern, F., and Giurfa, M. (2001). Con-
cal properties. The cellular and neural machinery un- figural olfactory learning in honeybees: Negative and positive
derlying the memory stages is basically similar to patterning discrimination. Learning and Memory 8, 7078.
those known for other model systems (Aplysia: Botzer, Dubnau, J., and Tully, T. (1998). Gene discovery in Drosophila, new
Markovich, and Susswien, 1998; Mller and Carew, insights for learning and memory. Annual Review of Neuro-
science 21, 407444.
1998; Drosophila: Dubnau and Tully, 1998; Chick:
Faber, T., Joerges, J., and Menzel, R. (1999). Associative learning
Rose, 1991), although each model system has its own modifies neural representations of odors in the insect brain.
temporal dynamics. This indicates that the cellular Nature Neuroscience 2, 7478.
and molecular machinery underlying the processing Funayama, E. S., Couvillon, P. A., and Bitterman, M. E. (1996).
of associative memory follows general rules but is flex- Compound conditioning in honeybees, blocking tests of the
independence assumption. Animal Learning and Behavior 23,
ible enough to adapt to the particular timing required
429437.
under natural conditions. Some researchers argue Gerber, B., and Ullrich, J. (1999). No evidence for olfactory block-
that the timing of memory stages in the honeybee re- ing in honeybee classical conditioning. Journal of Experimental
flects an adaptation to the requirements during for- Biology 202, 1,8391,854.
INVERTEBRATE LEARNING: Associative Learning in Hermissenda 277
Giurfa, M., Eichmann, B., and Menzel, R. (1996). Symmetry per- the adaptation of organisms to changing environ-
ception in an insect. Nature 382, 458461. mental demands, but also, and more important, for
Giurfa, M., Zhang, S., Jenett, A., Menzel, R., and Srinivasan, M. V.
(2001). The concepts of sameness and difference in an in- the persistence of learningthat is, long-term memo-
sect. Nature 410, 930933. rywhich provides us with a history of human experi-
Hammer, M. (1993). An identified neuron mediates the uncondi- ence. In spite of the widespread interest in learning
tioned stimulus in associative olfactory learning in honeybees. and memory, their basic mechanisms remain among
Nature 366, 5963.
Hammer, M., and Menzel, R. (1998). Multiple sites of associative
the least thoroughly understood areas of physiology.
odor learning as revealed by local brain microinjections of oc-
topamine in honeybees. Learning and Memory 5, 146156. An attractive experimental approach to this prob-
Kamin, L. J. (1968). Attentionlike processes in classical condition- lem at a fundamental level is the analysis of learning
ing. In M. R. Jones, ed., Miami symposium on predictability, be- and memory in the less-complex central nervous sys-
havior and aversive stimulation. pp. 932. Miami, FL: University tem of invertebrates. One animal that has contributed
of Miami Press.
Mackintosh, N. J. (1975). A theory of attention: Variations in the
to the physiology of learning and memory is the nudi-
associability of stimuli with reinforcement. Psychology Review branch mollusk Hermissenda crassicornis, whose behav-
82, 276298. ior can be modified by a Pavlovian conditioning pro-
Menzel, R. (1990). Learning, memory, and cognition in honey cedure. The Hermissenda central nervous system is
bees. In R. P. Kesner and D. S. Olton, eds., Neurobiology of com-
parative cognition. Hillsdale, NJ: Erlbaum.
relatively simple, consisting of many identifiable neu-
(1999). Memory dynamics in the honeybee. Journal of Com- rons that can be studied in detail using biochemical,
parative Physiology[A] 185, 323340. biophysical, and molecular techniques. An additional
Menzel, R., Brandt, R., Gumbert, A., Komischke. B., and Kunze, advantage is that the two sensory structures and their
J. (2000). Two spatial memories for honeybee navigation. Pro-
central pathways supporting the conditioned stimulus
ceedings of the Royal Society of London B 267, 961968.
Menzel, R., and Giurfa, M. (2001). Cognitive architecture of a (CS) and unconditioned stimulus (US) are totally in-
mini-brain: The honeybee. Trends in Cognitive Sciences 5, 62 tact in the isolated central nervous system. This at-
71. tractive feature facilitates the search for cellular corre-
Menzel, R., and Mller, U. (1996). Learning and memory in hon-
lates of learning that have been identified and have
eybees: From behavior to neural substrates. Annual Review of
Neuroscience 19, 379404. been the focus of biochemical and molecular analyses.
Mller, D., Gerber, B., Hellstern, F., Hammer, M., and Menzel, R.
(2000). Sensory preconditioning in honeybees. Journal of Ex-
perimental Biology 203, 1,3511,364. Pavlovian Conditioning
Mller, U., and Carew, T. J. (1998). Serotonin induces temporally
and mechanistically distinct phases of persistent PKA activity Pavlovian conditioning of Hermissenda involves
in Aplysia sensory neurons. Neuron 21, 1,4231,434. changes in light-elicited locomotion and foot length
Rescorla, R.A., and Wagner, A. R. (1972). A theory of classical con- (conditioned responses, CRs) produced by stimula-
ditioning: Variations in the effectiveness of reinforcement
and non-reinforcement. In A. H. Black A.H. and W. F.
tion of the visual and vestibular systems with their ad-
Prokasy, eds., Classical conditioning II: Current research and theo- equate stimuli (Crow and Alkon, 1978; Lederhendler,
ry. New York: Appleton-Century-Crofts. Gart, and Alkon, 1986). The Pavlovian conditioning
Rose, S. P. R. (1991). How chicks make memories: The cellular cas- procedure consists of pairing light, the conditioned
cade from c-fos to dendritic remodelling. Trends in Neuro-
stimulus (CS), with high-speed rotation, the uncondi-
sciences 14, 390397.
Rudy, J. W., and Sutherland, R. J. (1992). Configural and elemen- tioned stimulus (US). As shown in Figure 1, after con-
tal associations and the memory coherence problem. Journal ditioning, the CS suppresses normal light-elicited lo-
of Cognitive Neuroscience 4, 208216. comotion and elicits foot shortening. Retention of
Smith, B. H., and Cobey, S. (1994). The olfactory memory of the
conditioned behavior persists for several days to
honey bee, Apis mellifera. II: Blocking between odorants in bi-
nary mixtures. Journal of Experimental Biology 195, 91108. weeks depending upon the number of conditioning
Thorn, R. S., and Smith, B. H. (1997). The olfactory memory of the trials used in initial acquisition (Alkon, 1989; Crow
honeybee Apis mellifera, III. Bilateral sensory input is neces- and Alkon, 1978). Pavlovian conditioning in Her-
sary for induction and expression of olfactory blocking. Jour- missenda exhibits CS specificity and is dependent
nal of Experimental Biology 200, 2,0452,055.
upon the association of the two sensory stimuli involv-
Randolf Menzel ing both contiguity and contingency. Crow and Of-
fenbach (1983) showed that conditioned animals ex-
hibit suppressed locomotor behavior in the presence
of the CS; however, their locomotor behavior in the
ASSOCIATIVE LEARNING IN HERMISSENDA dark was not significantly changed. Nonassociative
Few features of conscious experience have captured contributions to behavior are expressed in the initial
the human imagination more than the proclivity of trials of the conditioning session and the decrement
animals to learn and to retain the consequences of ex- rapidly following the termination of multitrial condi-
perience in memory. Learning not only provides for tioning.
278 INVERTEBRATE LEARNING: Associative Learning in Hermissenda
In addition to multiple-trial conditioning of sup- of conditioned animals (see Figure 1). Facilitation of
pression of light-elicited locomotion and foot con- type-Ie interneuron EPSPs (excitatory postsynaptic
traction, one-trial conditioning also modifies light- potentials) elicited by lateral type-B spikes is also ob-
elicited locomotion (Crow and Forrester, 1986). Pair- served following conditioning. Studies of the signal
ing the CS with the direct application of one of the transduction pathways responsible for the modifica-
transmitters of the US pathway (serotonin 5-HT, tion of diverse K+ currents of type-B photoreceptors
nominal US) to the exposed nervous system of other- of conditioned animals have identified several second
wise intact Hermissenda produces suppression of light- messenger systems. Both protein kinase C (PKC)
elicited locomotion when the animals are tested twen- (Farley and Auerbach, 1986; Crow et al., 1991) and
ty-four hours following the one conditioning trial. extracellular signal-regulated protein kinase (ERK)
One-trial conditioning also produces enhanced excit- (Crow et al., 1998) contribute to modifications of ex-
ability of type-B photoreceptors (see Figure 1), a com- citability and synaptic efficacy of conditioned Her-
ponent of the CS pathway that expresses cellular plas- missenda. A second site of cellular plasticity in condi-
ticity produced by multitrial Pavlovian conditioning tioned animals is the type-A photoreceptor. Lateral
(discussed further below). type-A photoreceptors of conditioned animals exhibit
an increase in CS-elicited spike frequency, a decrease
in generator potential amplitude, and enhanced ex-
Cellular and Synaptic Plasticity Associated citability and decreased spike frequency accommoda-
with Pavlovian Conditioning tion to extrinsic current (Frysztak and Crow, 1993).
An essential step in the analysis of Pavlovian con- Taken collectively, the evidence for localization of cel-
ditioning is the search for the loci in the animals ner- lular changes in the CS pathway indicates that multi-
vous system where memories of the associative expe- ple sites of cellular and synaptic plasticity involving
rience are stored. Crow and Alkon (1980) identified changes in both excitability and synaptic strength
the primary sensory neurons (photoreceptors) of the exist in the photoreceptors and interneurons of con-
pathway mediating the CS as one site for memory ditioned animals (see Figure 1). Anatomical studies of
storage. Intrinsic modifications of cellular and synap- type-B photoreceptors indicate the existence of spa-
tic plasticity in classically conditioned animals involve tially segregated compartments (Alkon, 1989). Photo-
both enhanced excitability and synaptic facilitation of transduction occurs in the soma-rhabdomeric com-
connections between sensory neurons in the pathway partment, spike generation in the distal axon, and
mediating the CS (Alkon, 1989; Crow and Alkon, synaptic interactions in the axon terminal regions
1980; Frysztak and Crow, 1994). Enhanced excitabili- within the cerebropleural neuropil. Therefore, a de-
ty in identified photoreceptors of conditioned Her- crease in K+ conductances of type-B photoreceptors
missenda is expressed by a significant increase in spike could contribute both directly and indirectly to en-
activity elicited by the CS or extrinsic current, an in- hanced excitability by increasing the amplitude of CS-
crease in the input resistance, an alteration in the am- elicited generator potentials and increasing CS-
plitude of light-elicited generator potentials, de- elicited spike activity in the spike-generating zone by
creased spike frequency accommodation, and a modification of conductance that influence the inter-
reduction in the peak amplitude of voltage- spike interval.
dependent (IA, ICa) and Ca2+-dependent (IK,Ca) cur-
rents (Alkon et al., 1985; for reviews, see Alkon, 1989;
Crow, 1988; Sahley and Crow, 1998). Enhanced excit- Mechanisms of Memory Consolidation
ability, expressed by an increase in both the ampli- Underlying Pavlovian Conditioning
tude of CS-elicited generator potentials and the num- Studies of memory have identified components of
ber of action potentials elicited by the CS, may be a memory consolidation that can be differentiated
major contributor to changes in the duration and am- based upon the contribution of signal transduction
plitude of CS-elicited complex postsynaptic poten- pathways, protein synthesis, and gene induction (for
tials (PSPs) and enhanced CS-elicited spike activity review, see DeZazzo and Tully, 1995). An analysis of
observed in postsynaptic targets. However, changes one-trial conditioning in Hermissenda has provided in-
in the strength of synaptic connections between iden- sights into the mechanisms of different stages of
tified type-B photoreceptors and other components memory consolidation. One-trial conditioning pro-
of the CS pathway have also been detected following duces long-term suppression of light-elicited locomo-
conditioning (Frysztak and Crow, 1994). tion (Crow and Forrester, 1986) and short-term and
Facilitation of the amplitude of unitary inhibitory long-term enhancement of excitability in sensory
postsynaptic potentials (IPSPs) elicited by single neurons of the CS pathway (Crow and Forrester,
spikes in identified type-B photoreceptors are detect- 1991, 1993). Short-term and long-term enhanced ex-
ed in type-A photoreceptors and type-Iii interneurons citability appear to be independent, parallel process-
INVERTEBRATE LEARNING: Associative Learning in Hermissenda 279
Figure 1
I. Pavlovian-conditioned foot-shortening and conditioned suppression of light-elicited locomotion of Hermissenda. A. Foot length in
the dark before presentation of the unconditioned stimulus (US). B. The unconditioned response (UCR) elicited by rotation (US) of
the animal in the dark. C. Foot length in the dark after Pavlovian conditioning before presentation of the light (CS). D. Conditioned
response (CR), foot-shortening elicited by presentation of the CS. The area indicated between the dashed lines in D represents the
magnitude of foot-shortening elicited by the CS after conditioning. E. Light-elicited locomotion towards a light source assessed before
conditioning. F. Suppression of light-elicited locomotion detected after Pavlovian conditioning. Pseudorandom or random
presentations of the CS and US do not result in the development of suppression of either light-elicited locomotion or CS-elicited foot
shortening. II. Components of the CS pathway that express plasticity in conditioned Hermissenda. A. The CS elicits a larger amplitude
generator potential (upper trace) recorded from type B photoreceptors as compared to pseudorandom controls (lower trace). B. An
extrinsic current pulse elicits more action potentials in type B photoreceptors from conditioned preparations as compared to
pseudorandom controls. C. Conditioning results in facilitation of the synaptic connections between type B photoreceptors and type A
photoreceptors and type B photoreceptors and type Ii and Ie interneurons D. E. as compared to control animals that received
pseudorandom presentations of the CS and US.
280 INVERTEBRATE LEARNING: Associative Learning in Hermissenda
es, because long-term enhanced excitability can be Conclusion

expressed in the absence of prior short-term en- One-trial and multi-trial Pavlovian conditioning
hanced excitability (Crow and Forrester, 1993). produce changes in both synaptic efficacy and cellular
Moreover, enhanced excitability in type-B photore- excitability in several identified neurons of the path-
ceptors follows a biphasic pattern in its development way supporting the CS. Activation of PKC and ERK
following one-trial conditioning. Excitability reaches is produced by both one-trial and multi-trial Pavlov-
a peak three hours after one-trial conditioning, de- ian conditioning of Hermissenda. Enhanced cellular
clines toward baseline control levels five to six hours excitability expressed by CS-elicited spike activity or
after conditioning, and is followed by an increase to extrinsic current results from the reduction in several
a stable plateau at sixteen to twenty-four hours post- K+ conductances in type-B photoreceptors of condi-
conditioning. In addition, one-trial conditioning pro- tioned animals. Studies of memory formation in Her-
duces an intermediate stage of memory that depends missenda have shown that intermediate and long-term
on translation but not transcription, whereas long- memory produced by one-trial conditioning involves
term memory for enhanced excitability depends the synthesis and phosphorylation of cellular pro-
upon both translation and transcription (Crow and teins. One protein, CSP24, a -thymosin-like protein
Forrester, 1990; Crow, Xue-Bian, and Siddiqi, 1999). associated with intermediate memory, is regulated by
conditioning and may contribute to reorganization of
Associated with intermediate memory is the
the actin cytoskeleton underlying structural remodel-
phosphorylation of a 24 kDa protein (CSP24). A one-
ing supporting long-term memory.
trial conditioning procedure that only produces
short-term memory does not result in the increased
phosphorylation of CSP24 (Crow and Xue-Bian, See also: SECOND MESSENGER SYSTEMS
2000). Therefore, the regulation of this phosphopro-
tein by one-trial conditioning is associated with ex- Bibliography
perimental conditions that produce intermediate and Alkon, D. L. (1989). Memory storage and neural systems. Scientific
American 261 (1), 4250.
long-term memory. The protein-synthesis inhibitor Alkon, D. L., Sakakibara, M, Forman, R., Harrigan. J., Leder-
anisomycin, present during the intermediate phase of hendler, I., and Farley, J. (1985). Reduction of two voltage-
memory, blocked the increased phosphorylation of dependent K+ currents mediates retention of a learned asso-
CSP24 but did not block the increased phosphoryl- ciation. Behavioral and Neural Biology 44, 278300.
ation of other proteins associated with one-trial con- Crow, T. (1988). Cellular and molecular analysis of associative
learning and memory in Hermissenda. Trends in Neurosciences
ditioning (Crow, Xue-Bian, and Siddiqi, 1999). Ex- 11, 136142.
periments examining 35S-methionine labeling of Crow, T., and Alkon, D. L. (1978). Retention of an associative be-
CSP24 during the intermediate phase of memory havioral change in Hermissenda. Science 201, 1,2391,241.
showed similar labeling of CSP24 for the conditioned (1980). Associative behavioral modification in Hermissenda:
Cellular correlates. Science 209, 412414
group and unpaired controls. However, both condi- Crow, T., and Forrester, J. (1986). Light paired with serotonin
tioned and unpaired groups were greater than unsti- mimics the effects of conditioning on phototactic behavior in
mulated controls with respect to 35S-methionine la- Hermissenda. Proceedings of the National Academy of Sciences of the
beling of CSP24. Therefore the requirement for United States of America 83, 7,9757,978.
protein synthesis is necessary but not sufficient for (1990). Inhibition of protein synthesis blocks long-term en-
hancement of generator potentials produced by one-trial in
long-term associative memory, or it may reflect an in- vivo conditioning in Hermissenda. Proceedings of the National
direct involvement in phosphorylation due to Academy of Sciences of the United States of America 87, 4,490
changes in the synthesis of protein kinases in signal 4,494.
transduction pathways or phosphatase inhibitors. Ex- (1991). Light paired with serotonin in vivo produces both
short- and long-term enhancement of generator potentials of
periments where CSP24 was excised from multiple
identified B-photoreceptors in Hermissenda. Journal of Neuro-
two-dimensional gels and subjected to reverse phase science 11, 608617.
HPLC (high pressure liquid chromatography) and (1993). Down-regulation of protein kinase C and kinase in-
automated sequence analysis showed that the se- hibitors dissociate short- and long-term enhancement pro-
quenced peptides exhibited a homology to the - duced by one-trial conditioning of Hermissenda. Journal of
Neurophysiology 69, 636641.
thymosin family of actin-binding proteins (Crow and Crow, T., Forrester, J., Williams, M., Waxham, N., and Neary, J.
Xue-Bian, 2000). All known vertebrate and inverte- (1991). Down regulation of protein kinase C blocks 5-HT-
brate -thymosins bind actin monomers (Nachimias, induced enhancement in Hermissenda B photoreceptors.
1993). Cytoskeletal-related proteins such as CSP24 Neuroscience Letters 121, 107110.
thus may contribute to long-term structural remodel- Crow, T., and Offenbach, N. (1983). Modification of the initiation
of locomotion in Hermissenda: Behavioral analysis. Brain Re-
ing in the CS pathway by regulating the turnover of search 271, 301310.
actin filaments during the intermediate-term transi- Crow, T. J., and Xue-Bian, J. J. (2000). Identification of a 24 kDa
tion period between short- and long-term memory. phosphorylation associated with an intermediate stage of
INVERTEBRATE LEARNING: Associative Learning in Limax 281
memory in Hermissenda. Journal of Neuroscience 20 (10), RC74, sential components of olfactory information process-
15. ing from mollusks to mammals (Hudson, 1999). Plas-
Crow, T., Xue-Bian, J. J., and Siddiqi, V. (1999). A protein synthe-
ticity can be induced by passive exposure to odors or
sis-dependent and mRNA synthesis-independent intermedi-
ate phase of memory in Hermissenda. Journal of Neurophysiology by arranging stimulus contingencies in which odors
82 (1), 495500. predict the presence or absence of other stimuli or re-
Crow, T., Xue-Bian, J. J., Siddiqi, V., Kang, Y., and Neary, J. T. wards (Sahley, 1990; Eichenbaum, 1998; Slotnick et
(1998). Phosphorylation of mitogen-activated protein kinase al., 2000). The ample evidence of plasticity in olfacto-
by one-trial and multi-trial classical conditioning. Journal of
ry processing systems over wide phyletic boundaries
Neuroscience 18 (9), 3,4803,487.
DeZazzo, J., and Tully, T. (1995). Dissection of memory formation suggested that Limax would be a good candidate for
from behavioral pharmacology to molecular genetics. Trends behavioral assessment of odor learning. I will empha-
in Neuroscience 18, 212218. size here recent work in the Limax odor-learning sys-
Farley, J., and Auerbach, S. (1986). Protein kinase C activation in- tem and related work on olfactory learning in closely
duces conductance changes in Hermissenda photoreceptors
related species Helix, for examplewhich has ap-
like those seen in associative learning. Nature 319, 220223.
Frysztak, R. J., and Crow, T. (1993). Differential expression of cor- peared since my previous review (Gelperin, 1992).
relates of classical conditioning in identified medial and later-
al type-A photoreceptors of Hermissenda. Journal of Neuro-
Limax is an odor-dominated species dependent
science 13 (7), 2,8892,897. on olfaction for finding food, mates, and homesites
(1994). Enhancement of type-B- and type-A photoreceptor (Gelperin, 1974). Associative learning about odor
inhibitory synaptic connections in conditioned Hermissenda. cues is rapid (Gelperin, 1975; Teyke, 1995) and shows
Journal of Neuroscience 14 (3), 1,2451,250. several higher-order contingencies such as com-
Lederhendler, I., Gart, S., and Alkon, D. L. (1986). Classical condi-
tioning of Hermissenda: Origin of a new response. Journal of
pound conditioning, second-order conditioning, and
Neuroscience 6, 1,3251,331. blocking (Sahley, 1990; Sekiguchi et al., 1991; Yama-
Nachmias, V. T. (1993). Small actin-binding proteins: the - da et al., 1992; Sekiguchi et al., 1994; Suzuki et al.,
thymosin family. Current Opinion in Cell Biology 5, 5662. 1994; Sekiguchi et al., 1997). Appetitive conditioning
Sahley, C., and Crow, T. (1998). Invertebrate learning: Current can modify feeding behavior so that previously aver-
perspectives. In J. Martinez and R. Kesner, eds., Neurobiology
of Learning and Memory, pp. 171209. New York: Academic
sive odors become attractive (Sahley et al., 1992; Gel-
Press. perin, 1999). The reliable and robust nature of odor
conditioning in Limax, combined with the complex
Terry J. Crow
nature of the logic operations performed during odor
conditioning (Gelperin et al., 1986) prompted an ex-
ploration of the central circuits for odor processing.
ASSOCIATIVE LEARNING IN LIMAX
Given the goal of understanding the cellular basis of Central Olfactory Centers
associative learning (Kandel, 2001), comparative
physiology seeks to identify a brain donor with highly Primary and second-order input from olfactory
developed associative learning and a nervous system receptors projects to a distinctive integrative center,
well suited to biophysical analysis. Early work made the procerebral (PC) lobe of the cerebral ganglion
clear that the central neurons and networks of gastro- (Chase, 2000), where some 105 interneurons process
pod mollusks were particularly favorable for the cellu- olfactory inputs (Ratt and Chase, 1997, 2000) and
lar analysis of synaptic plasticity and behavioral con- may store odor memories (Kimura et al., 1998a;
trol. With this background we chose to explore the Nakaya et al., 2001). Our initial search for the central
odor-learning ability of the terrestrial gastropod mol- site of odor learning therefore focused on the PC
lusk Limax maximus, which has large central neurons lobe, which has oscillatory dynamics of its local field
that are useful for studies of calcium transients during potential (LFP) (Gelperin and Tank, 1990; Kawahara
single-action potentials (Chang et al., 1974), neural et al., 1997) and propagates activity waves along its
control of heart rate (MacKay and Gelperin, 1972), apical-basal axis (Delaney et al., 1994).
cellular analysis of [3H]-2-deoxyglucose uptake (Sej- Because odor-memory coding in the PC lobe de-
nowski et al., 1980), neural control of feeding motor pends on the dynamics of wave propagation, it is im-
programs (Prior and Gelperin, 1977; Delaney and portant to understand the mechanism of wave initia-
Gelperin, 1990c), and serotonergic modulation of tion and propagation. The PC lobe contains a small
feeding motor programs (Gelperin, 1981). (12 percent) population of bursting inhibitory neu-
Learning cued by olfactory stimuli was chosen be- rons (Watanabe, 1998) that seem to couple to each
cause olfactory systems in a wide variety of species can other by electrical and excitatory chemical synapses
modify their input-output functions, linking odors to (Ermentrout et al., 2001). Two-photon laser-scanning
behaviors in ways that depend on previous olfactory microscopy has discerned two populations of inhibi-
experience. Synaptic plasticity and learning are es- tory-bursting neurons, differing in the speed and di-
282 INVERTEBRATE LEARNING: Associative Learning in Limax
rection of propagation of calcium-based action poten- then, twenty minutes after the conditioning trial, the
tials in their neurites (Wang et al., 2001). The animal receives an injection of with Lucifer yellow
bursting inhibitory neurons produce chloride- (LY) into the blood space. The hour-long interval that
mediated inhibitory synaptic potentials in the major begins twenty minutes after the conditioning trial is
population of nonbursting neurons (Watanabe et al., when the short-term memory of odor conditioning is
1999). The population of bursting inhibitory neurons converted to a long-term form (Sekiguchi et al., 1991;
shows a gradient of excitability from the apex to the Sekiguchi et al., 1994; Sekiguchi et al., 1997). The LY
base of the PC, such that bursting occurs first in the in the PC neuron somata after conditioning is con-
most apical bursting neurons and then, because of ex- tained in membrane-bound vesicles, as in the original
citatory coupling between bursting neurons, propa- reports of activity-dependent LY labeling in fly reti-
gates along the apical-basal axis to the base of the PC na (Wilcox and Franceschini, 1984), perhaps because
lobe, where the activity wave ends. The gradient of
of an activity-dependent pinocytotic process. The
excitability is shown by taking a series of transverse
causal coupling between electrical, synaptic, or bio-
slices of the PC lobe along the apical-basal axis and
chemical events in the labeled neurons and pinocyto-
measuring the frequency of spontaneous oscillations
tic uptake of LY is unknown.
of the LFP in each slice. The most apical slice oscil-
lates fastest, the most basal slice oscillates slowest, and The striking feature of the learning-dependent
the intermediate slices have intermediate oscillation labeling in the Limax PC lobe is that various control
frequencies, depending on their apical-basal posi-
procedures, such as odor presentation alone, which
tion.
do not allow learning to occur, do not lead to LY la-
There is evidence of the apical-basal activity wave beling (Kimura et al., 1998a). If two odors are used
in two-site LFP measurements made simultaneously as separate conditioned stimuli during training in the
in apical and basal recording sites (Ermentrout et al., sequential aversive training trials, two LY-labeled
1998; Gelperin et al., 2001) or in optical recordings bands appear in one PC lobe (Kimura et al., 1998a).
where voltage-sensitive or calcium-sensitive dyes stain The unilateral nature of the learning-dependent la-
PC neurons. Sequential images of the PC lobe based beling was completely unexpected, but is also occurs
on the calcium or voltage signals show the initiation
in the replication and extension of the original work.
of the activity wave at the apex and its propagation to
(Gelperin, 1999) The existence of crossed inhibition
the base (Kleinfeld et al., 1994; Toda et al., 2000;
between the right and left odor-processing circuits,
Nikitin and Balaban, 2001b). The LFP signal at a par-
demonstrated in an in vitro nose-brain preparation
ticular site along the apical-basal axis is produced by
(Teyke et al., 2000), may explain why only one PC
the simultaneous generation of a 57-mV inhibitory
synaptic potential in a large number of nonbursting lobe is the dominant site of odor-memory storage.
neurons by synaptic divergence from the local burst- The second major indication that the PC lobe is
ing neurons. The oscillatory nature of the LFP derives the likely site of odor-memory storage is that one-trial
from the periodic nature of the activity wave in the
odor conditioning selectively activates a small set of
bursting neurons propagating past a particular re-
genes in PC lobe neurons (Nakaya et al., 2001). Brain
cording site along the apical-basal axis. Oscillatory
tissue from 200 Limax given one-trial odor condition-
dynamics is a universal feature of olfactory analyzers
ing was obtained, and differential mRNA expression
in mollusks, arthropods, and vertebrates (Tank et al.,
was compared between this collection of learned
1994; Gelperin, 1999; Laurent, 1999) and may con-
tribute an essential temporal component to the odor brain and tissue from 200 brains of control animals
code. given odor stimulation (CS) and an aversive uncondi-
tioned stimulus (US) with a CS-US delay too long to
permit learning to occur (Gelperin, 1975). A gene
Odor-Memory Storage in the Procerebral coding for a twenty-three amino-acid peptide was
Lobe identified, cloned, and sequenced and the deduced
Learning-dependent labeling of a band of non- peptide was constructed for antibody production.
bursting PC neurons after one-trial odor condition- The level of expression of the learning-activated gene
ing provides some of the most direct evidence that was clearly enhanced selectively in neurons of the
odor memories are stored in the PC lobe (Kimura et PC lobe in trained slugs relative to control slugs
al., 1998a; Gelperin, 1999). The learning-dependent (Nakaya et al., 2001). The peptide is secreted to ex-
labeling has been demonstrated as a consequence of tracellular space and may play a role in stabilizing
both aversive and appetitive one-trial odor condition- synapses, as suggested by recent work on cellular con-
ing. The slug is given a single training trial with odor sequences of learning in Drosophila (Connolly and
as a positive or negative conditioned stimulus, and Tully, 1998).
Modulation of Procerebral Lobe Dynamics bursting neurons cannot receive direct synaptic input
from olfactory afferents. The demonstration that
The PC lobe has both intrinsic and extrinsic syn-
odor inputs can alter the frequency of the PC LFP
aptic modulation using twenty-one known and puta-
(Gervais et al., 1996) must be due to odor inputs onto
tive neurotransmitters (Gelperin, 1999), notably ni-
nonbursting neurons having indirect excitatory ef-
tric oxide (NO) (Gelperin, 1999), carbon monoxide
fects on bursting neurons.
(Gelperin et al., 2000), acetylcholine (Watanabe et al.,
2001), dopamine (Gelperin et al., 1993; Rhines et al.,
1993), serotonin (Yamane, 1989; Inoue, 2001), and Model of Odor Memory Formation
glutamate (Watanabe et al., 1999), along with numer-
ous small peptides such as FMRFamide and small car- The current model of odor-memory band forma-
dioactive peptide B (Yamane and Gelperin, 1987; tion in the Limax PC lobe is summarized in Figure 1.
Cooke and Gelperin, 1988). It is therefore not sur- The odor-memory bands have a bandlike shape be-
prising that recordings of PC lobe LFP in vivo using cause of the interaction of the wave front of the prop-
implanted fine-wire electrodes reveal a much richer agating-activity wave with the region of the PC most
range of activity than the isolated brain in vitro strongly driven by the odor used as the conditioned
(Cooke and Gelperin, 2001). There are periods in stimulus. The confluence of the activity wave-front
vivo during which the 0.51.0 Hz oscillation of the PC and sensory drive from the odor produces a short-
LFP in vitro is recorded, but during other periods the term memory of the odor. Behavioral evidence for
diversity of LFP waveforms makes clear that the PC the existence of a short-term odor memory comes
lobe has modes of activity in vivo not predicted from from the finding that a brief odor stimulus can be
recordings in vitro. given, and then, after tens of minutes, the delivery of
unconditioned stimulus can still result in CS-US pair-
The maintenance of LFP oscillations and there- ing (Gelperin, 1975). This long-delay or trace-
fore wave propagation is dependent on synthesis of conditioning aspect of odor conditioning is typical.
NO in the Limax PC lobe. If NO synthesis is blocked The short-term odor memory is subsequently con-
in the terrestrial snail Helix by injecting a substituted verted to a long-term odor memory because of the ac-
arginine that blocks the activity of NO synthase, odor tion of a modulatory transmitter liberated on the PC
learning is blocked (Teyke, 1996). It is tempting to neurons storing the short-term memory as a conse-
speculate that the dose of NO synthase blocker that quence of application of the US.
blocks oscillatory dynamics in the PC lobe is also the
We have elaborated the model to suggest that if
dose that produces the odor-learning deficit.The
two odors are learned, then the spacing of the two
Limax PC lobe stains intensely for nitric oxide syn-
bands representing the odors will depend on the sim-
thase (Cooke et al., 1994; Gelperin et al., 2001). NO
ilarity of the two odors (see Figure 1). If the two odors
may set the level of bursting by the bursting neurons
are very similar (odor A and odor A), then the odor-
in the PC and hence determine the frequency of LFP
memory bands representing odor A and odor A will
oscillation and perhaps the rate of wave propagation
be at the minimum interband spacing consistent with
(Inoue, 2001).
the ability to access the two memory bands individual-
The use of olfactory nerve (ON) shock as a substi- ly and uniquely. If odor A is learned, if its odor mem-
tute for odor stimulation makes it is clear that odor ory band is formed, and then if wave propagation is
inputs to the PC lobe have a phase-dependent effect blocked by blocking NO, then the brain will report
on the bursting neurons (Inoue et al., 2000). Excitato- that odor A is the same as odor A (see Figure 1C).
ry postsynaptic potentials (EPSPs) are recorded with This result has been obtained with the isolated Limax
short latency after ON shock in nonbursting neurons, nose-brain preparation (Teyke and Gelperin, 1999),
while EPSPs with longer and variable latency are re- whereas in honeybee blocking the oscillatory dynam-
corded in bursting neurons after ON shock. This pro- ics of the antennal lobe with picrotoxin produced the
cedure suggests a monosynaptic connection from fi- behavioral result that the bee could not discriminate
bers in the ON onto nonbursting PC neurons and an between two similar odors but could discriminate be-
excitatory connection from nonbursting neurons tween two very different odors (Stopfer et al., 1997).
onto bursting neurons. This direct excitation of non- Blockade of NO synthase activity in honeybee anten-
bursting neurons by ON input is consistent with the nal lobe impairs olfactory discrmination (Hosler et
known anatomy of the ON projection into the PC lobe al., 2000).
as the ON fibers fill the neuropil of the PC lobe but
do not project to the layer of neuronal somata (Gel-
perin and Flores, 1997; Kawahara et al., 1997). The PC Lobe Inputs and Outputs
neurites of the bursting PC neurons are confined to Researchers have identified the nature of the be-
the layer of neuronal somata (Watanabe, 1998), so the havioral modifications occurring during Helix food-
Figure 1
Diagram of a possible relationship between waves, odor memory bands, and odor discrimination. A. Forming odor memories for two
dissimilar odors leads to two well-separated odor memory bands. B. Blocking of oscillations and waves in the procerebral lobe blocks
odor learning. C. Learning two similar odors (A and A1) leads to the formation of two closely spaced odor memory bands. After
learning odor A, if oscillations and waves are blocked, odor A1 is not discriminated from odor A. Arrows signify the propagated activity
wave. L-NAME, N-nitro-L-arginine methyl ester.
odor conditioning and some of the motor pathways the cerebral ganglion, such as feeding command neu-
expessing the learned alteration in odor response rons.
(Peschel et al., 1996; Friedrich and Teyke, 1998). The
same motor-output pathway that expresses odor The PC lobe receives inputs from other regions
learning in the Helix, the peritentacular nerves con- of the cerebral ganglion and from the tentacular gan-
trolling the superior tentacle muscles, also expresses glia adjacent to the olfactory receptor epithelia of the
the modified motor output attributable to odor learn- superior and inferior noses. The tentacular ganglia
ing in the Limax (Teyke and Gelperin, 1999). Re- appear to have an oscillatory component to their LFP
searchers have identified a neuron in the metacere- that is modulated by odor stimulation (Ito et al.,
bral lobe of the Limax PC lobe with a neurite in the 2001). The PC lobes contain neurites of neurons lo-
PC lobe (Shimozone et al., 2001) that shows mem- cated in other parts of the cerebral ganglia (Ratt and
brane potentials oscillations arising from input in the Chase, 1997, 2000; Shimozone et al., 2001) and in the
PC lobe. This neuron may convey PC-processed olfac- buccal (Gelperin and Flores, 1997) and pedal (Chase
tory information to other chemointegrative sites in and Tolloczko, 1989) ganglia that provide the path-
ways for integrating olfactory information with other governing the transition from short-term memory to
inputs to make behavioral decisions. Activation of the long-term memory (Burrell and Sahley, 2001; Sutton
PC-connected buccal neurons can reset the PC LFP et al., 2001) are modified (Yin and Tully, 1996; Dub-
oscillation, whereas the pedal cells are activated by PC nau and Tully, 1998) to produce an odor-memory sys-
stimulation but cannot reset the LFP oscillation by tem wherein one-trial learning is the normal condi-
their activity. The PC LFP oscillation can also be mod- tion. Also, the learning-activated gene in the PC lobe
ulated by electrical stimulation of the digits of the ten- has significant homology with gene sequences in ze-
tacular ganglia (Ito et al., 1999), a result that may be brafish, mouse, and human, providing further sup-
due to the FMRFamide contained in some of the pri- port for the idea that molluscan model systems imple-
mary sensory neurons projecting through the tentac- ment synaptic plasticity mechanisms of relevance to
ular ganglia to the PC lobe (Suzuki et al., 1997). mammalian systems.
Imaging Odor Memories Bibliography

Burrell, B. D., and Sahley, C. L. (2001). Learning in simple systems.
A number of imaging studies have attempted to
Current Opinion in Neurobiology 11,757764.
clarify the nature of odor responses in the PC lobe Chang J. J., Gelperin, A., and Johnson, F. H. (1974). Intracellularly
(Kleinfeld et al., 1994; Inoue et al., 1998; Toda et al., injected aequorin detects trans-membrane calcium flux dur-
2000; Nikitin and Balaban, 2001a), particularly after ing action potentials in an identified neuron from the terres-
odor conditioning (Kimura et al., 1998c; Nikitin and trial slug, Limax maximus. Brain Research 77, 431442.
Balaban, 2000). Initial studies on nave preparations Chase R (2000). Structure and function in the cerebral ganglion.
Microscopic Research Techniques 49, 511520.
showed that odor stimulation led to a collapse of the Chase, R., and Rieling, J. (1986). Autoradiographic evidence for
apical-basal phase gradient for a few cycles after odor receptor cell renewal in the olfactory epithelium of a snail.
stimulation (Kleinfeld et al., 1994). Imaging studies Brain Research 384, 232239.
of PC lobe dynamics after one-trial odor conditioning Chase R., and Tolloczko, B. (1986). Synaptic glomeruli in the olfac-
have provided evidence for regional localization of tory system of a snail. Cell Tissue Research 246, 567573.
(1989). Interganglionic dendrites constitute an output
odor excitation (Kimura et al., 1998c). Recordings of
pathway from the procerebrum of the snail Achatina fulica.
PC lobe LFP oscillations after odor training also indi- Journal of Comparative Neurology 283, 143152.
cate regional localization of baseline shifts in re- Connolly, J.B., and Tully, T. (1998). Integrins, a role for adhesion
sponse to application of conditioned odors (Kimura molecules in olfactory memory. Current Biology 8, R386R389.
et al., 1998b). It would be interesting to know if the Cooke, I., and Gelperin, A. (1988). Distribution of FMRFamide-
region of localized excitation to a conditioned odor like immunoreactivity in the nervous system of the slug Limax
maximus. Cell Tissue Research 253, 6976.
corresponds to the region of learning-dependent LY
Cooke, I. R. C., Edwards, S.L., and Anderson, C. R. (1994). The
labeling in the same slug. distribution of NADPH-diaphorase activity and immunoreac-
tivity to nitric oxide synthase in the nervous system of the pul-
monate mollusc Helix aspersa. Cell Tissue Research 277, 565
Summary 572.
Cooke, I. R. C., and Gelperin, A. (2001). In vivo recordings of
The olfactory-processing system of the Limax dis- spontaneous and odor-modulated dynamics in the Limax ol-
plays many of the design features of other species, in- factory lobe. Journal of Neurobiology 46, 126141.
cluding receptor turnover (Chase and Rieling, 1986), Delaney, K., and Gelperin, A. (1986). Post-ingestive food-aversion
central neurogenesis (Zakharov et al., 1998), oscilla- learning to amino acid deficient diets by the terrestrial slug
tory dynamics of central-processing centers (Gel- Limax maximus. Journal of Comparative Physiology A 159, 281
295.
perin, 1999), rapid and long-lasting learning (De-
(1990). Cerebral interneurons controlling fictive feeding in
laney and Gelperin, 1986), extensive local-feedback Limax maximus. III. Integration of sensory inputs. Journal of
synapses (Zs.-Nagy and Sakharov, 1970; Ratt and Comparative Physiology A 166, 327343.
Chase, 2000), glomerular processing of some inputs Delaney, K. R., Gelperin, A., Fee, M. S., Flores, J. A., Gervais, R.,
(Chase and Tolloczko, 1986), nitric-oxide-dependent Tank, D.W., and Kleinfeld, D. (1994). Waves and stimulus-
synaptic plasticity (Teyke, 1996), memory-dependent modulated dynamics in an oscillating olfactory network. Pro-
ceedings of the National Academy of Sciences of the United States of
alterations in early-processing stages (Nakaya et al., America 91, 669673.
2001), and several dozen neurotransmitters and Dubnau J., and Tully, T. (1998). Gene discovery in Drosophila: New
neuromodulators involved in experience-dependent insights for learning and memory. Annual Review of Neuro-
circuit reconfiguration (Gelperin, 1999). The com- science 21, 407444.
parative approach has produced evidence that mol- Eichenbaum, H. (1998). Using olfaction to study memory. Annual of the
New York Academy of Science 855, 657669.
luscan model systems can yield principles of synaptic
Ermentrout, B., Flores, J., and Gelperin, A. (1998). Minimal model
plasticity and associative learning (Kandel, 2001). Fu- of oscillations and waves in the Limax olfactory lobe with tests
ture work in the Limax will extend this foundation, for of the models predictive power. Journal of Neurophysiology 79,
example, by clarifying how the biochemical substrates 2,6772,689.
(2001). Model for olfactory discrimination and learning in Kawahara S., Toda, S., Suzuki, Y., Watanabe, S., and Kirino, Y.
Limax procerebrum incorporating oscillatory dynamics and (1997). Comparative study on neural oscillation in the
wave propagation. Journal of Neurophysiology 85, 1,4441,452. procerebrum of the terrestrial slugs Incilaria bilineata and
Friedrich, A., and Teyke, T. (1998). Identification of stimuli and Limax marginatus. Journal of Experimental Biology 200, 1,851
input pathways mediating foodattraction conditioning in the 1,861.
snail, Helix. Journal of Comparative Physiology A 183, 247254. Kimura, T., Suzuki, H., Kono, E., and Sekiguchi, T. (1998a). Map-
Gelperin, A. (1974). Olfactory basis of homing behavior in the ping of interneurons that contribute to food aversion condi-
giant garden slug, Limax maximus. Proceedings of the National tioning in the slug brain. Learning and Memory 4, 376388.
Academy of Science of the United States of America 71, 966970. Kimura T., Toda, S., Sekiguchi, T., Kawahara, S., and Kirino, Y.
(1975). Rapid food-aversion learning by a terrestrial mol- (1998c). Optical recording analysis of olfactory response of
lusk. Science 189, 567570. the procerebral lobe in the slug brain. Learning and Memory 4,
(1981). Synaptic modulation by identified serotonergic 289400.
neurons. In B. Jacobs and A. Gelperin, eds., Serotonin Neuro- Kimura, T., Toda, S., Sekiguchi, T., and Kirino, Y. (1998b). Behav-
transmission and Behavior. Cambridge, MA: MIT Press. ioral modulation induced by food odor aversive conditioning
(1992). Associative learning in Limax. In L. Squire, ed., En- and its influence on the olfactory responses of an oscillatory
cyclopedia of learning and memory. New York: Macmillan. brain network in the slug Limax marginatus. Learning and Mem-
(1999). Oscillatory dynamics and information processing in ory 4, 365375.
olfactory systems. Journal of Experimental Biology 202, 1,855 Kleinfeld, D., Delaney, K. R, Fee, M. S., Flores, J. A., Tank, D. W.,
1,864. and Gelperin, A. (1994). Dynamics of propagating waves in
Gelperin, A., and Flores, J. (1997). Vital staining from dye-coated the olfactory network of a terrestrial mollusc: An electrical
microprobes identifies new olfactory interneurons for optical
and optical study. Journal of Neurophysiology 72, 1,402 1,419.
and electrical recording. Journal of Neuroscience Methods 72,
Laurent, G. (1999). A systems perspective on early olfactory cod-
97108.
ing. Science 286, 723728.
Gelperin, A., Flores, J., Raccuia-Behling, F., and Cooke, I. R. C.
MacKay, A., and Gelperin, A. (1972). Pharmacology and reflex re-
(2000). Nitric oxide and carbon monoxide modulate oscilla-
sponsiveness of the heart of the giant garden slug, Limax maxi-
tions of olfactory interneurons in a terrestrial mollusc. Journal
mus. Comparative Biochemistry and Physiology 43A, 877896.
of Neurophysiology 83, 116127.
Nakaya, T., Kawahara, S., Watanabe, S., Lee, D.-S., Suzuki, T.,
Gelperin, A., Hopfield, J. J., and Tank, D. W. (1986). The logic of
Kirino, Y. (2001). Identification and expression of a novel
Limax learning. In A. I. Selverston, ed., Model neural networks
gene in odour-taste associative learning in the terrestrial slug.
and behavior. New York: Plenum Press.
Genes to Cells 6, 4356.
Gelperin, A., Kao, J. P. Y., and Cooke, I. R. C. (2001). Gaseous ox-
Nikitin, E. S., and Balaban, P. M (2000). Optical recording of odor-
ides and olfactory computation. American Zoology 41, 332345.
Gelperin, A., Rhines, L., Flores, J., and Tank, D. (1993). Coherent evoked responses in the olfactory brain of the nave and aver-
network oscillations by olfactory interneurons: Modulation by sively trained terrestrial snails. Learning and Memory 7, 422
endogenous amines. Journal of Neurophysiology 69, 1,930 432.
1,939. (2001a). Optical recording of responses to odor in olfactory
Gelperin, A., and Tank, D. W. (1990). Odor-modulated collective structures of the nervolus system in the terrestrial mollusk
network oscillations of olfactory interneurons in a terrestrial Helix. Neuroscience and Behavioral Physiology 31, 2130.
mollusc. Nature 345, 437440. (2001b). Optical recording of responses to odor in olfactory
Gervais, R., Kleinfeld, D., Delaney, K. R., and Gelperin, A. (1996). structures of the nervous system in the terrestrial mollusk
Central and reflex neuronal responses elicited by odor in a Helix. Neuroscience and Behavioral Physiology 31, 2130.
terrestrial mollusc. Journal of Neurophysiology 76, 1,3271,339. Peschel, M., Straub, V., and Teyke, T. (1996). Consequences of
Hosler, J. S., Buxton, K. L, and Smith, B. H. (2000). Impairment foodattraction conditioning in Helix, A behavioral and elec-
of olfactory discrimination by blockade of GABA and nitric trophysiological study. Journal of Comparative Physiology A 178,
oxide activity in the honeybee antennal lobes. Behavioral 317327.
Neuroscience 114, 514525. Prior, D., and Gelperin, A. (1977). Autoactive molluscan neuron,
Hudson, R. (1999). From molecule to mind, the role of experience Reflex function and synaptic modulation during feeding in
in shaping olfactory function. Journal of Comparative Physiology the terrestrial slug Limax maximus. Journal of Comparative Physi-
A 185, 297304. ology 114, 217232.
Inoue, T., Kawahara, S., Toda, S., Watanabe, S., and Kirino, Y. Ratt, S., and Chase, R. (1997). Morphology of interneurons in the
(1998). Selective optical recording of the neural activity in the procerebrum of the snail Helix aspersa. Journal of Compartive
olfactory center of land slug using a calcium indicator dye. Neurology 384, 359372.
Bioimages 6, 5967. (2000). Synapse distribution of olfactory interneurons in
Inoue, T., Watanabe, S., Kawahara, S., and Kirino, Y. (2000). the procerebrum of the snail Helix aspersa. Journal of Compara-
Phase-dependent filtering of sensory information in the oscil- tive Neurology 417, 366384.
latory olfactory center of a terrestrial mollusk. Journal of Rhines, L., Sokolove, P., Flores, J., Tank, D. W., and Gelperin, A.
Neurophysiology 84, 1,1121,115. (1993). Cultured olfactory interneurons from Limax maximus,
Ito, I., Kimura, T., and Ito, E. (2001). Odor responses and sponta- Optical and electrophysiological studies of transmitter-
neous oscillatory activity in tentacular nerves of the terrestrial evoked responses. Journal of Neurophysiology 69, 1,9401,947.
slug, Limax marginatus. Neuroscience Letters 304, 1,4551,148. Sahley, C. L. (1990). The behavioral analysis of associative learning
Ito, I., Kimura, T., Suzuki, H., Sekiguchi, T., and Ito, E. (1999). Ef- in the terrestrial mollusc Limax maximus, The importance of
fects of electrical stimulation of the tentacular digit of a slug interevent relationships. In S. Hanson and C. Olson, eds.,
upon the frequency of electrical oscillations in the procerebral Connectionist modeling and brain function: The developing inter-
lobe. Brain Research 815, 121125. face. Cambridge, MA: MIT Press.
Kandel, E. R. (2001). The molecular biology of memory storage, Sahley, C. L., Martin, K. A., and Gelperin, A. (1992). Odors can in-
a dialogue between genes and synapses. Science 294, 1,030 duce feeding motor responses in the terrestrial mollusc Limax
1,038. maximus. Behavioral Neuroscience 106, 563568.
INVERTEBRATE LEARNING: C. elegans 287
Sejnowski, T. J., Reingold, S. C., Kelley, D. B., and Gelperin, A. induced retrograde amnesia in Limax flavus. Journal of Neuro-
(1980). Localization of [3H]-2-deoxyglucose in single mollus- science 12, 729735.
can neurones. Nature 287, 449451. Yamane, T., and Gelperin, A. (1987). Aminergic and peptidergic
Sekiguchi, T., Yamada, A., and Suzuki, H. (1997). Reactivation- amplification of intracellular cyclic AMP levels in a molluscan
dependent changes in memory states in the terrestrial slug neural network. Cell Molecular Neurobiology 7, 291301.
Limax flavus. Learning and Memory 4, 356364. Yin, J., and Tully, T. (1996). CREB and the formation of long-term
Sekiguchi, T., Yamada, A., Suzuki, H., and Mizukami, A. (1991). memory. Current Opinion in Neurobiology 6, 264268.
Temporal analysis of the retention of a food-aversion condi- Zakharov, I. S., Hayes, N. L., Ierusalimsky, V. N., Nowakowski, R.
tioning in Limax flavus. Zoological Science (Tokyo) 8, 103111. S., and Balaban, P. M. (1998). Postembryonic neuronogenesis
Sekiguchi, T., Suzuki, H., Yamada, A., and Mizukami, A. (1994). in the procerebrum of the terrestrial snail, Helix lucorum. Jour-
Cooling-induced retrograde amnesia reflexes Pavlovian con- nal of Neurobiology 35, 271276.
ditioning associations in Limax flavus. Neuroscience Research 18, Zs.-Nagy, I., and Sakharov, D. A. (1970). The fine structure of the
267275. procerebrum of pulmonate molluscs, Helix and Limax. Tissue
Shimozone, S., Watanabe, S., Inoue, T., and Kirino, Y. (2001). Cell 2, 399411.
Identification and characterization of an output neuron from
the oscillatory molluscan olfactory network. Brain Research
Alan Gelperin
921, 98105.
Slotnick, B., Hanford, L., Hodos, W. (2000). Can rats acquire an
olfactory learning set? Journal of the Experimental Psychology of
Animal Behavioral Process 26, 399415. C. ELEGANS
Stopfer, M., Bhagavan, S., Smith, B. H., and Laurent, G. (1997).
Simple invertebrate organisms have provided a
Impaired odour discrimination on desynchronization of
odour-encoding neural assemblies. Nature 390, 7074. wealth of information concerning the molecular and
Suttona, M.A., Masters, S. E., Bagnall, M. W., and Carew, T. J. cellular basis of learning and memory. Research con-
(2001). Molecular mechanisms underlying a unique interme- ducted in one such simple system, the nematode
diate phase of memory in Aplysia. Neuron 31, 143154. Caenorhabditis elegans (C. elegans), has made several
Suzuki, H., Kimura,T., Sekiguchi. T., and Mizukami, A. (1997).
important contributions to the field.
FMRFamide-like-immunoreactive primary sensory neurons
in the olfactory system of the terrestrial mollusc, Limax mar-
ginatus. Cell Tissue Research 289, 339345. Advantages of Caenorhabditis elegans as a
Suzuki, H., Sekiguchi, T., Yamada, A., and Mizukami, A. (1994).
Sensory preconditioning in the terrestrial mollusk, Limax Model System
flavus. Zoological Science 11, 121125. C. elegans is a small (about one millimeter), her-
Tank, D. W., Gelperin, A., and Kleinfeld, D. (1994). Odors, oscilla-
maphroditic, soil-dwelling nematode (see Figure 1).
tions, and waves, Does it all compute? Science 265, 1,819
1,820. In the laboratory, C. elegans spends its short life
Teyke, T. (1995). Food-attraction conditioning in the Roman snail, (about twenty days) swimming on agar-filled petri
Helix pomatia. Journal of Comparative Physiology A 177, 409414. dishes laying eggs and feeding on E. coli bacteria. Its
(1996). Nitric oxide, but not serotonin, is involved in acqui- hermaphroditic mode of reproduction makes the
sition of food-attraction conditioning in the snail Helix poma-
maintenance of strains easy, thus allowing for large
tia. Neuroscience Letters 206, 2932.
Teyke, T., and Gelperin, A. (1999). Olfactory oscillations augment quantities of inexpensive, readily available animals
odor discrimination not odor identification by Limax CNS. for testing. Its nervous system is simple, consisting of
NeuroReport 10, 1,0611,068. 302 neurons with 5,000 electrical and 10,000 chemi-
Teyke, T., Wang, J. W., and Gelperin, A. (2000). Lateralized mem- cal synapses, all of which have been identified and
ory storage and crossed inhibition during odor processing by mapped at the electron microscope level (White,
Limax. Journal of Comparative Physiology A 186, 269278.
Toda, S., Kawahara, S., and Kirino, Y. (2000). Image analysis of ol-
Southgate, and Durbin, 1988; White, Southgate,
factory responses in the procerebrum of the terrestrial slug Thomson, and Brenner, 1986). Complete develop-
Limax marginatus. Journal of Experimental Biology 203, 2,895 mental lineages of all cells are also known. Further-
2,905. more, the worm is transparent, so that with the appro-
Wang, J. W., Flores, J., Gelperin, A., and Denk, W. (2001). Initia- priate microscopic power, its simple nervous system
tion and propagation of calcium-dependent action potentials
in a coupled network of olfactory interneurons. Journal of
is easily visible. Single neurons can be easily ablated
Neurophysiology 85, 977985. with a laser, while leaving the rest of the nervous sys-
Watanabe S., Kawahara, S., and Kirino, Y. (1999). Glutamate in- tem intact. Such an approach can elucidate the role
duces Cl- and K+ currents in the olfactory interneurons of a of specific neurons in behavioral processes.
terrestrial slug. Journal of Comparative Physiology A 184, 553
562. Consistent with the simplicity of its nervous sys-
Watanabe, S., Inoue, T., Murakami, M., Inokuma, Y., Kawahara, tem, the C. elegans genome is small and has been fully
S., and Kirino, Y. (2001). Modulation of oscillatory neural ac- sequenced, with 8 x 107 nucleotide pairs arranged on
tivities by cholinergic activation of interneurons in the olfacto- six haploid chromosomes. Both classic genetic tech-
ry center of a terrestrial slug. Brain Research 896, 3035.
niques and modern genetic engineering have pro-
Wilcox, M., and Franceschini, N. (1984). Illumination induces dye
incorporation in photoreceptor cells. Science 225, 851854.
duced a large number of mutant strains, which have
Yamada, A., Sekiguchi, T., Suzuki, H., and Mizukami, A. (1992). provided the opportunity for investigating the role of
Behavioral analysis of internal memory states using cooling single genes in behavioral processes. Worms are able
288 INVERTEBRATE LEARNING: C. elegans
Figure 1
Schematic drawing of the nematode Caenorhabditis elegans. The adult worm is approximately one millimeter long and forty micrometers
wide.
to survive freezing, and thus mutant strains, once ac- mented or habituated response following presenta-
quired, can be maintained indefinitely. tion of a novel, usually aversive, stimulus. Sensitization
The behavioral repertoire of C. elegans is complex is the facilitation of a nondecremented response re-
enough to offer a number of interesting behaviors to sulting from presentation of an aversive stimulus. The
study. In the laboratory worms move forward along TWR of C. elegans shows each of these three nonasso-
the surface of agar-filled petri dishes, using rhythmic, ciative types of learning (Rankin, Beck, and Chiba,
coordinated contractions of dorsal and ventral muscle 1990).
groups, resulting in smooth sinusoidal waves of for- The response habituation observed in the TWR
ward locomotion. Worms will respond to a variety of with repeated presentation of the tap stimulus is not
stimuli by changing direction and by swimming back- due to fatigue of the system, because presentation of
ward. Stimuli that produce reversals include touch, a novel stimulus to a habituated animal immediately
heat probes, some chemical compounds, and vibra- causes a return of the behavior to prehabituation le-
tions caused by the force of a mechanical tapper ap- velsthat is, dishabituation (Rankin, Beck, and
plied to the side of the dish. The response to a me- Chiba, 1990). Consistent with the rules of habituation
chanical tap has been termed the tap withdrawal outlined by Groves and Thompson (1970), the rate of
response (TWR) by Rankin, Beck, and Chiba (1990) habituation of the TWR in C. elegans is sensitive to the
and has proved to be important for studies of both frequency of stimulation. In wild-type worms habitua-
short- and long-term memory. Using laser ablation tion occurs more rapidly at short interstimulus inter-
techniques, Wicks and Rankin (1995) determined vals (ISIs) (i.e., intervals of two or ten seconds) com-
that the TWR consists of six sensory neurons, ten in- pared to long ISIs (i.e., intervals of sixty seconds);
terneurons, and approximately sixty-nine motor neu- subsequent recovery from habituation is also affected
rons. by ISI, with animals recovering more rapidly from ha-
bituation with short ISIs than with long ISIs (Rankin
Short-Term Memory and Broster, 1992; see Figure 2). Thus short-term
memory for habituation training lasts less than fifteen
The simplest forms of learning are habituation, minutes when trained with short ISIs and can last for
dishabituation, and sensitization. These are nonasso- one to two hours when trained with long ISIs.
ciative forms of learning. Habituation is a decrease in
the rate or amplitude or both, of responding due to As C. elegans follows the generally agreed upon
repeated stimulus presentation (Groves and Thomp- rules of habituation, it provides an effective model in
son, 1970). Dishabituation is the facilitation of a decre- which to study the role of genes underlying habitua-
INVERTEBRATE LEARNING: C. elegans 289
tion. There are mutant strains of worms that show dif- dependent habituation, eat-4 worms had more rapid
ferences in short-term habituation, one of which is overall rates of habituation with depressed asymptotic
eat-4. Studies of eat-4 and its mammalian homologues levels compared to wild-type worms. Furthermore,
suggest that it regulates the amount of glutamate in their recovery from habituation was slower compared
neuron terminals (Bellocchio, Reimer, Fremeau, and to wild-type worms. Rankin and Wicks (2000) hypoth-
Edwards, 2000; Lee et al., 1999). The eat-4 gene esized that with repeated stimulation glutamate be-
product is expressed on the sensory neurons of the comes rapidly depleted, resulting in more rapid re-
TWR circuit (Lee et al., 1999). Rankin and Wicks sponse decrements in eat-4. However, the fact that the
(2000) showed that while responding normally to a usual effects of ISI on the rate of habituation were
single tap, and displaying the usual pattern of ISI- preserved in eat-4 indicates that it is more than simple
290 INVERTEBRATE LEARNING: C. elegans
neurotransmitter depletion underlying habituation. TWR (three to four blocks of twenty stimuli each at a
The eat-4 strain did not show dishabituation. Since sixty-second ISI, separated by one-hour periods) re-
dishabituation does not occur in the eat-4 strain, dis- sulted in retention of the habituated response twenty-
habituation does require the presence of an intact four hours later when tested with a series of twenty
eat-4 gene product. The cellular processes underlying taps (Beck and Rankin, 1997; Rose, Chen, Kaun, and
habituation and dishabituation are likely not the Rankin, 2001). This effect was observed only when a
same, and each involves eat-4 to a different extent. sixty-second ISI was used during training (i.e., it was
not seen with a ten-second ISI) and when a distribut-
ed training procedure was used. Consistent with what
Associative Learning has been observed in other species, such as Aplysia and
Although for many years researchers believed Drosophila, massed training (one block of sixty taps),
that the response decrements seen in habituation with either a ten-second or a sixty-second ISI, did not
were due solely to repeated stimulus presentation, it produce long-term habituation in C. elegans. In addi-
has now been demonstrated that organisms can make tion, Beck and Rankin (1995) showed that long-term
associations during habituation training that affect fu- habituation was protein-synthesis dependent. When
ture performance. The most common of such associa- worms were exposed to as few as fifteen minutes of
tions is context conditioning, where some aspect of heat shock, which disrupts protein synthesis, during
the training environment is encoded by the organism, the early part of the rest period in a distributed-
which then influences future responses to the original training paradigm, long-term memory was signifi-
training stimulus. Thus the long-held view that habit- cantly reduced, suggesting that the time immediately
uation is a purely nonassociative form of learning may after training is most vulnerable to heat shock and
not be warranted. Rankin (2000) showed that C. ele- thus may be a critical phase in the formation of long-
gans is capable of context conditioning during habitu- term memory. Heat shock applied during the one-
ation training. Worms were habituated to thirty-tap hour rest period did not, however, have any effect on
stimuli in the presence or absence of a distinctive che- the accumulation of short-term habituation seen over
mosensory cue (sodium acetate). When tested an hour successive blocks of training, suggesting that short-
later, worms trained in the presence of sodium ace- and long-term memory recruit different cellular and
tate showed greater retention of training when tested molecular processes.
in the presence of sodium acetate than when tested on As with short-term memory, C. elegans also pro-
plain agar plates. Placing worms on sodium acetate vides an effective model with which to study the role
plates for the hour between training and testing proof genes underlying long-term memory for habitua-
duced extinction of the context effect. tion training. For example, glr-1, a gene coding for
Classical conditioning has also been demonstrat- a homologue of mammalian kainate/AMPA-type glu-
ed in C. elegans using both appetitive and aversive tamate receptor, has been identified and cloned in C.
conditioning paradigms. Conditioned worms showed elegans. Worms missing functional glr-1 showed no
clear postconditioning preferences for distinctive long-term retention of habituation training twenty-
tastes that were paired with food during training four hours later (Rose, Chen, Kaun, and Rankin,
(Wen et al., 1997). Similarly, in an aversive- 2001). These results suggest that stimulation of kai-
conditioning paradigm, worms learned to avoid tastes nate/AMPA-type glutamate receptors on the inter-
that were paired with an aversive stimulus during neurons is required for long-term habituation to tap.
training. Similar results have been obtained using ol- The same mutation has also been shown to prevent
factory rather than taste cues, where worms learned olfactory associative learning (Morrison and van der
to avoid a previously attractive odor after it was paired Kooy, 2001).
with an aversive acetic acid solution (Morrison, Wen,
Runciman, and van der Kooy, 1999). Two mutant
strains of worms, lrn-1 and lrn-2, were generated that Conclusion
were not able to form taste or olfactory associations. C. elegans has proved to be a useful model system
for the study of learning and memory processes. Un-
like other animals used in this area of research, C. ele-
Long-Term Memory gans is an inexpensive, easily obtained and main-
C. elegans is also capable of retention of habitua- tained organism, with a large amount of information
tion training for at least twenty-four hours. This phe- available about its nervous system and genome.
nomenon, long-term habituation, has been used as an Knowledge of the neural circuitry underlying behav-
effective model for the study of long-term memory ior combined with knowledge of the genome has al-
(defined as retention for twenty-four hours or more). lowed for the investigation of genetic factors involved
Distributed or spaced habituation training of the in both short-term and long-term memory. That is,
INVERTEBRATE LEARNING: Habituation and Sensitization in Tritonia 291
if a certain gene is expressed on a component of a out affecting the response itself. Journal of Neuroscience 20,
neural circuit known to underlie a behavior, then that 4,3374,344.
Rose, J., Chen, S., Kaun, K., and Rankin, C. H. (2001). Glutamate
gene is a prime candidate for playing a role in either
and AMPA-receptor function are necessary for long-term
the behavior itself or its plasticity. This kind of an ap- memory in C. elegans. Society for Neuroscience Abstracts 27, 45.
proach has proved fruitful in C. elegans by elucidating Wen, J. Y. M., Kumar, N., Morrison, G., Rambaldini, G., Runci-
the importance in learning and memory of gluta- man, S., Rousseau, J., and van der Kooy, D. (1997). Mutations
mate, a neurotransmitter that has also been deter- that prevent associative learning in C. elegans. Behavioral
Neuroscience 111 (2), 354368.
mined to play a major role in mammalian learning White, J. E., Southgate, E., and Durbin, R. (1988). Appendix 2:
and memory processes. The fact that the same results Neuroanatomy. In W. B. Wood, ed., The nematode Caenorhab-
are obtained in C. elegans attests to its validity as a ditis elegans, pp. 433455. Cold Spring Harbor, NY: Cold
model system. The rules of habituation outlined by Spring Harbor Laboratory Press.
White, J. E., Southgate, E., Thompson, J. N., and Brenner, S.
Groves and Thompson (1970) apply to C. elegans,
(1986). The structure of the nervous system of the nematode
making it an appropriate model system for the study Caenorhabditis elegans. Philosophical Transactions of the Royal So-
of short-term memory as well. Research using C. ele- ciety of London. Series B: Biological Sciences 314, 1340.
gans has made it clear that habituation is not a unitary Wicks, S. R., and Rankin, C. H. (1995). Integration of mechanosen-
process but is rather a set of processes, differentially sory stimuli in Caenorhabditis elegans. Journal of Neuroscience 15,
2,4342,444.
recruited by short and long ISIs that cannot be neatly
tucked away into the category of simple, nonasso- Stephan Steidl
ciative learning. Catharine H. Rankin
Bibliography
Beck, C. D. O., and Rankin, C. H. (1992). Caenorhabditis elegans: A HABITUATION AND SENSITIZATION
simple systems approach to the genetics of behavior. In D. IN TRITONIA
Goldowitz, D. Wahlstein, and R. E. Wimer, eds., Techniques for
the genetic analysis of brain and behavior: Focus on the mouse, 445 Studies of learning in both vertebrates and inverte-
463. Amsterdam: Elsevier. brates indicate that individual memories are stored in
(1995). Heat-shock disrupts long-term memory consolida- the brain as distributed sets of cellular and synaptic
tion in Caenorhabditis elegans. Learning and Memory 2, 161177. modifications. Most research in this area has focused
(1997). Long-term habituation is produced by distributed
training at long ISIs and not by massed training or short ISIs
on characterizing the detailed mechanisms underly-
in Caenorhabditis elegans. Animal Learning and Behavior 25 (4), ing specific sites of learning-related plasticity, such as
446457. presynaptic facilitation at sensory to motor neuron
Bellocchio, E. E., Reimer, R. J., Fremeau, R. T., and Edwards, R. synapses in the marine mollusc Aplysia and long-term
H. (2000). Uptake of glutamate into synaptic vesicles by an in- potentiation at synapses in the vertebrate hippocam-
organic phosphate transporter. Science 289, 957960.
pus. However, the fragmented nature of memory
Groves, P. M., and Thompson, R. F. (1970). Habituation: A dual-
process theory. Psychological Reviews 77, 419450. storage raises additional issues at a network, rather
Lee, R. Y. N., Sawin, E. R., Chalfie, M., Horvitz, H. R., and Avery, than synaptic, level. For example, how is the informa-
L. (1999). Eat-4, a homolog of a mammalian sodium- tion represented by a given memory organized across
dependent inorganic phosphate co-transporter, is necessary the different sites of plasticity encoding itdo differ-
for glutamatergic neurotransmission in Caenorhabditis elegans.
ent sites store the same information redundantly, or
Morrison, G. E., and van der Kooy, D. (2001). A mutation in the
does each encode a unique component of the total ac-
AMPA-type glutamate receptor, glr-1, blocks olfactory asso- quired information? If different memories overlap in
ciative and nonassociative learning in Caenorhabditis elegans. the brain, as seems likely, how do they avoid interfer-
Behavioral Neuroscience 115 (3), 640649. ing with one another? Such network-level issues of
Morrison, G. E., Wen, J. Y. M., Runciman, S., and van der Kooy, memory storage have been investigated in the marine
D. (1999). Olfactory associative learning in Caenorhabditis ele-
gans is impaired in lrn-1 and lrn-2 mutants. Behavioral Neuro-
mollusc Tritonia diomedea, an organism well suited to
science 113 (2), 358367. both cellular and network studies of learning and
Rankin, C. H. (2000). Context conditioning in habituation in the memory.
nematode Caenorhabditis elegans. Behavioral Neuroscience 114,
When touched by the tube feet of its highly mo-
496505.
Rankin, C. H., Beck, C. D. O., and Chiba, C. M. (1990). Caenorhab- bile seastar predators, Tritonia responds with vigorous
ditis elegans: A new model system for the study of learning and alternating ventral and dorsal body flexions that pro-
memory. Behavioral Brain Research 37, 8992. pel it away to safety. The neural circuit underlying
Rankin, C. H., and Broster, B. S. (1992). Factors affecting habitua- this response is known in detail and consists of identi-
tion in the nematode Caenorhabditis elegans. Behavioral Neuro-
fied afferent neurons, command interneurons, cen-
science 106, 239242.
Rankin, C. H., and Wicks, S. R. (2000). Mutations of the Caenorhab- tral pattern generator (CPG) interneurons, and effer-
ditis elegans brain-specific inorganic phosphate transporter ent flexion neurons. A key advantage for cellular
eat-4 affect habituation of the tap-withdrawal response with- studies of learning is that the neural program under-
292 INVERTEBRATE LEARNING: Neurogenetics of Memory in Drosophila
lying the swim behavior can be readily elicited in the learning simultaneouslyhabituation of swim cycle
isolated brain preparation, where the neurons and number and sensitization of swim onset latency (Mon-
synapses storing memory can be easily identified and geluzi and Frost, 2000). That these two components
studied. of the different memories can coexist without conflict
Studies of behavioral plasticity in Tritonia have fo- suggests that they involve different anatomical loci in
cused on two universal forms of nonassociative learn- the swim circuit. Other swim features modified in sen-
ing: habituation and sensitization. Sensitization refers sitizationcycle number and thresholdreverse di-
to the increase in responsiveness that follows a single, rection during habituation training, suggesting that
unexpected, and therefore potentially dangerous the underlying circuit modifications may be located
stimulus. Habituation refers to the gradual decrease in at the same sites for both forms of learning. Studies
responsiveness that occurs in response to repetitive since the 1990s have sought to determine the behav-
innocuous stimuli. In Tritonia, a single swim stimulus ioral role (information content) of each site of learn-
produces a period of sensitization, during which test ing-related plasticity.
swims have a lower threshold, faster onset latency, See also: APLYSIA: MOLECULAR BASIS OF LONG-TERM
and higher cycle number (Frost, Brandon, and Mon- SENSITIZATION; INVERTEBRATE LEARNING: C.
geluzi, 1998). On the other hand, repeatedly eliciting ELEGANS; ORIENTING REFLEX HABITUATION;
the escape swim leads to habituation, characterized by VESTIBULO-OCULAR REFLEX (VOR) PLASTICITY
swims with fewer cycles, a higher threshold, and a lon-
ger cycle period (Frost, Brown, and Getting, 1996). Bibliography
Frost, W. N., Brandon, C. L., Mongeluzi, D. L. (1998). Sensitization
Evidence suggests that the multiple, habituation- of the Tritonia escape swim. Neurobiology of Learning and Memo-
related behavioral changes are encoded by distribut- ry 69, 126135.
ed sites of plasticity in the swim circuit. First, habitua- Frost, W. N., Brown, G., Getting, P. A. (1996). Parametric features
tion is accompanied by a progressive drop in the of habituation of swim cycle number in the marine mollusc
Tritonia diomedea. Neurobiology of Learning and Memory 65, 125
number of incoming afferent neuron action poten- 134.
tials per stimulus. In addition, the synaptic connec- Katz, P. S., Frost, W. N. (1997). Removal of spike frequency adapta-
tions made by the afferent neurons progressively de- tion via neuromodulation intrinsic to the Tritonia escape swim
crease in strength with repeated stimulation. central pattern generator. Journal of Neuroscience 17, 7,703
Bypassing the afferent neurons with repeated intra- 7,713.
Katz, P. S., Getting, P. A., Frost, W. N. (1994). Dynamic neuro-
cellular stimulation of the swim-command neurons modulation of synaptic strength intrinsic to a central pattern
still results in a progressive decrement of the number generator circuit. Nature 367, 729731.
of cycles per swim motor program, implicating the in- Mongeluzi, D. L., Frost, W. N. (2000). Dishabituation of the Tri-
volvement of interneuronal sites of plasticity. This is tonia escape swim. Learning and Memory 7, 4347.
further supported by the finding that habituation William N. Frost
produced by stimulating one sensory neuron popula-
tion also leads to habituation of swims elicited by pre-
viously unstimulated sensory populations (Frost et al.,
1996). NEUROGENETICS OF MEMORY
The memory for sensitization also appears to be IN DROSOPHILA
encoded by multiple circuit modifications. Sensitizing Neurogenetic analysis of memory seeks to identify
stimuli produce a long-lasting enhancement of the genes involved in behavioral plasticity, to characterize
excitability and synaptic strength of CPG neuron C2. the cells and neuroanatomies in which they are ex-
An important effect of this enhancement is to pressed, and to define the biochemistries and cell bi-
strengthen a positive feedback connection from C2 to ologies in which they participate. From the study of
the command neurons that drive the swim motor pro- Drosophila (fruit flies) beginning in the 1970s, two fun-
gram. Sensitizing stimuli also cause prolonged tonic damental notions became apparent. First, a complex
firing of the CPGs Dorsal Swim Interneurons. These circuitry, consisting of sensory inputs, central process-
serotonergic cells participate in generating the swim ing, and motor outputs from thousands of neurons,
motor program and also appear responsible for the likely is required. In addition, hundreds to thousands
neuromodulatory enhancement of C2 excitability and of genes likely participate in the neuronal-synaptic
synaptic strength observed in sensitization (Katz, Get- plasticity underlying behavioral plasticity. Hence, a
ting, and Frost, 1994; Katz and Frost, 1997). vertical integration of function, from gene to behav-
Scientists are just beginning to understand how ior, will entail five primary levels of experimental
the information represented by habituation and sen- analysis: molecular-genetic, gene network computa-
sitization is organized in the Tritonia circuitry. After tion, neurophysiological, neural network computa-
ten stimulus trials animals often display both forms of tion, and behavioral. Second, molecular biological
INVERTEBRATE LEARNING: Neurogenetics of Memory in Drosophila 293
work across the animal kingdom has revealed a re- California Institute of Technology, yielded dunce, the
markable degree of evolutionary conservation, from first experimentally induced mutant gene that pro-
basic phenomenology of cell biology during develop- duced deficient learning. The second screen, in W. G.
ment to behavioral biology of sleep and senescence. Quinns laboratory at Princeton University, produced
Thus, a horizontal integration of gene action from the vegetable mutants, rutabaga, radish, turnip, and
lower organisms to humans appears the rule rather cabbage, which also showed no odor-avoidance learn-
than the exception. Given the economy of scale (short ing (Aceves-Pina et al., 1983). A modification of this
generation time, inexpensive rearing costs), extant behavioral screen also yielded the amnesiac mutant,
experimental tools, and a knowledge base of Drosophi- which showed normal learning but defective memory
la genetics that goes back to the nineteenth century, retention thereafter (Quinn, Sziber, and Booker,
this model system will continue to yield valuable in- 1979). The third screen, in T. Tullys laboratory at
formation on the neurobiological basis of plasticity. Brandeis University, looked for performance defects
Moreover, functional insights gained from flies will three hours after Pavlovian training and yielded the
inform the mammalian condition directly, thereby al- mutants latheo, linotte, nalyot, and golovan (Tully et al.,
lowing discovery of genetic and pharmacological 1996).
therapies for various forms of cognitive dysfunction Anatomical experiments (see below) have re-
in humans. vealed that olfactory memory formation involves the
mushroom body structures, neuropillar structures in
the central brain thought to integrate sensory input.
Behavioral Plasticity Accordingly, another screen for memory mutants was
In the final quarter of the twentieth century, a accomplished by looking first for enhancer trap
plethora of behavioral learning tasks were developed strains that expressed a beta-galactosidase reporter
for Drosophila (Connolly and Tully, 1998). Nonasso- gene preferentially in mushroom bodies and then by
ciative tasks exist for the landing response, cleaning screening for behavioral defects in olfactory memory.
reflex, proboscis extension response, and odor avoid- Mutants of dunce and rutabaga were reisolated with
ance response. Associative tasks include conditioned this approach, which also yielded the mutants PKA-
task aversion, courtship conditioning, operant condi- C1, leonardo, and Volado (Roman and Davis, 2001).
tioning to visual cues in a flight simulator, and olfac-
tory discriminative conditioning. The last task, in
particular, has proved valuable. Because initial condi- Biochemistry
tioned avoidance levels are robust and memory reten- Early biochemical experiments (see Aceves-Pina
tion can last more than a week, this learning task has et al., 1983) and subsequent molecular cloning (see
been used extensively to identify and characterize sin- Dubnau and Tully, 1998) have established that dunce
gle-gene mutations (see below). Careful analyses of encodes a cAMP (Cyclic adenosine monophosphate)-
the learning/memory defects in these mutants has specific phosphodiesterase (PDE), rutabaga encodes a
suggested that olfactory memory formation occurs in calcium-sensitive adenylyl cyclase (AC), and amnesiac
five genetically distinct phases: acquisition or learn- encodes a neuropeptide similar to vertebrate PACAP
ing (LRN), short-term memory (STM), middle-term (pituitary adenylyl cyclase activating peptide). Three
memory (MTM), anesthesia-resistant memory reverse-genetic experiments, focusing on additional
(ARM), and long-term memory (LTM) (Tully et al., components of this pathway, have strengthened this
1996; Tully, Preat, Boynton, and Del Vecchio, 1994). notion: 1. Mutations targeted to the regulatory sub-
Processing through these memory phases seems to unit of PKA (PKA-RI) produced flies with olfactory
occur sequentially from LRN to STM to MTM, at learning defects. 2. Overexpression of a dominant-
which point memory is processed independently (in negative mutation of the stimulatory G protein sub-
parallel) into ARM and LTM. These temporal stages unit (Gs) in transgenic flies produced olfactory learn-
of memory processing likely reflect a combination of ing defects. 3. Overexpression of a repressor isoform
neural activity in different anatomic sites and bio- of the CREB (cAMP response element binding pro-
chemical activity within each. tein) transcription factor in transgenic flies blocked
olfactory long-term memory. Together, these behav-
ior-genetic experiments strongly support the notion
Genetics that cAMP signaling is central to olfactory memory
Three mutageneses have been conducted to formation in Drosophila. These data also are consis-
screen for behavioral mutants with learning/memory tent with behavioral, electrophysiological, and cellu-
defects. All have used an odor-avoidance procedure lar experiments in Aplysia (Bailey, Bartsch, and Kan-
(Boynton and Tully, 1992; Dudai et al., 1976). The del, 1996) and in mice (Wong et al., 1999), suggesting
first screen, in Seymour Benzers laboratory at the that cAMP signaling is part of an evolutionarily con-
served molecular mechanism underlying synaptic neuromuscular junction (NMJ). In short, all memory
and behavioral plasticity. mutants known to date produce distinct defects in
The MAP (mitogen-activated protein) kinase sig- synaptic structure or function, or both, at the NMJ
naling pathway also may be involved. Molecular clon- (Saitoe and Tully, 2000). Increases in neural activity
ing of leonardo revealed it to be a mutation of the 14- or in cAMP levels lead to increases in the number of
3-3 gene, which, among other tasks, regulates the synaptic boutons onto muscles (structure) and to in-
function of Ras/Raf, two of several proteins involved creased excitability of synaptic transmission (func-
in GTP (guanine triphosphate) exchange. The Volado tion). Adf1 (nalyot) or fasII (another cell adhesion mol-
mutation has been shown to reside in a gene encoding ecule) appear involved in changes of synaptic
-integrin, a cell surface adhesion molecule. This structure but not function, while dCREB2 regulates
general class of molecule often activates the MAP ki- synaptic function but not structure (Davis, Schuster,
nase pathway via interactions with receptor tyrosine and Goodman, 1996; DeZazzo et al., 2000). Thus, a
kinases, though such a connection has not yet been genetic dissection of this form of developmental plas-
made for Volado. ticity is under way.
Mysteries remain and puzzles present them-
selves. The mutants radish, turnip, and cabbage have Anatomy
yet to be cloned. The linotte mutation resides either in
the receptor tyrosine kinase, derailed, or in a novel Early electrophysiological and lesion experi-
neighboring transcript. The nalyot mutation lies in ments in bees identified a distinct anatomical region
the Adf1 transcription factor; golovan is in the neuro- of the insect brain, the mushroom body, to be in-
genetic locus, extra machrochaetea. Exactly how these volved in associative processes (Menzel et al., 1991).
molecular-genetic components fit into the cell biolo- In Drosophila, mutants with structural defects in spe-
gy of olfactory memory is not yet clear. Curiously, the cific adult brain anatomies were identified by M. Hei-
latheo mutation has been shown to disrupt ORC3, senbergs laboratory in Wurzburg, and those with ab-
which encodes a protein subunit of the Origin Recog- normal mushroom bodies also displayed olfactory
nition Complex involved in DNA replication during learning defects (Heisenberg, Borst, Wagner, and
cell proliferation. Surprisingly, the protein LAT also Byers, 1985). Four subsequent experiments clearly es-
is expressed in synapses of terminally differentiated tablished mushroom bodies as a neural substrate of
neurons, thereby suggesting a completely novel func- adult olfactory memory:
tion for this protein outside of the nucleus of dividing 1. Chemical ablation of mushroom body neurons
cells. Although such genetic pleiotropy is not unusual, completely abolishes olfactory learning (de Belle
this observation nevertheless emphasizes how genetic and Heisenberg, 1994);
screens can break the bondage of hypothesis-driven
research. These dangling mutants presage an ulti- 2. Overexpression of a dominant-negative Gs pro-
mate understanding of a more complicated and com- tein specifically in mushroom bodies completely
plete genetic basis of memory. abolishes olfactory learning (Connolly et al.,
1996);
3. Transgenic expression of RUTABAGA (the pro-
Physiology
tein encoded by the rutabaga gene) in mushroom
Gaining electrophysiological access to neurons in bodies specifically rescues the learning defect of
the adult central nervous system that subserve olfacto- rutabaga mutants (Zars, Wolf, Davis, and Heisen-
ry memory has been challenging. Thus, initial physio- berg, 2000); and
logical characterizations of memory mutants relied
on neural circuitries underlying other behaviors. Cor- 4. Structural mutants with lesions restricted to the
fas and Dudai (1990), for instance, found that habitu- alpha lobes (axonal projections from intrinsic
ation of the (bristle) cleaning reflex in dunce and ruta- mushroom body neurons) specifically abolish
baga mutants was abnormal, but this behavioral effect long-term memory (Pascual and Preat, 2001).
resulted from opposite physiological defects in the Together, these observations suggest that olfacto-
underlying sensory neurons. Sensory fatigue was ac- ry learning and memory depend, at least in part, on
celerated in dunce mutants, while it was retarded in ru- the activity of mushroom body neurons. Though
tabaga mutants. Research by Engel and Wu (1996) these neurons remain largely inaccessible to classic
began to characterize normal and mutant electrophy- electrophysiological investigations (but see Wright
siological responses in the giant-fiber neurons, which and Zhong, 1995), less invasive imaging techniques
contribute to habituation of the jump response. are revealing some of their cellular properties in re-
Much insight on a role for memory genes in syn- sponse to experience (Rosay, Armstrong, Wang, and
aptic plasticity has come from studies of the larval Kaiser, 2001; Wang et al., 2001).
INVERTEBRATE LEARNING: Neurogenetics of Memory in Drosophila 295
The primary strength of neurogenetic analysis of proach, 2nd edition, pp. 265318. Oxford, UK: Oxford Uni-
memory in Drosophila lies in the discovery of genes. versity Press.
Behavioral screens for memory mutants enable this Corfas, G., and Dudai, Y. (1990). Adaptation and fatigue of a me-
chanosensory neuron in wild-type Drosophila and in memory
discovery process without any preconceived (and
mutants. Journal of Neuroscience 10, 491499.
naive) assumptions about the underlying biochemical Davis, G. W., Schuster, C. M., and Goodman, C. S. (1996). Genetic
or anatomical substrates. To this end, yet another be- dissection of structural and functional components of synaptic
havioral screen by Tully and coworkers at Cold plasticity, Part 3: CREB is necessary for presynaptic functional
Spring Harbor laboratory has discovered fifty-seven plasticity. Neuron 17 (4), 669679.
new mutants, defining forty-seven new genes. DNA de Belle, J. S., and Heisenberg, M. (1994). Associative odor learn-
ing in Drosophila abolished by chemical ablation of mushroom
microarrays also have been used to identify more than
bodies. Science 263, 692695.
one thousand candidate memory genes (CMGs), DeZazzo, J., Sandstrom, D., deBelle, S., Velinzon, K., Smith, P.,
which are transcriptionally regulated in normal flies Grady, L., DelVecchio, M., Ramaswami, M., and Tully, T.
during olfactory long-term memory formation. Sig- (2000). nalyot, a mutation of the Drosophila myb-related Adf1
nificant genetic overlap exists between these two ex- transcription factor, disrupts synapse formation and olfactory
perimental approaches. As outlined above for the first memory. Neuron 27, 145158.
Dubnau, J., and Tully, T. (1998). Gene discovery in Drosophila: New
few (dunce, rutabaga, latheo), these genes become an
insights for learning and memory. Annual Review of Neuro-
experimental common currency with which to in- science 21, 407444.
vestigate mechanisms of plasticity at several biological Dudai, Y., Jan, Y.-N., Byers, D., Quinn, W., and Benzer, S. (1976).
levels of organization (biochemical, physiological, an- dunce: A mutant of Drosophila melanogaster deficient in learn-
atomical, and so forth) and across many animal mod- ing. Proceedings of the National Academy of Sciences of the United
els. Combined, these data will reveal in more detail States of America 73, 1,6841,688.
Engel, J. E., and Wu, C. F. (1996). Altered habituation of an identi-
the molecular and cellular mechanisms by which fied escape circuit in Drosophila memory mutants. Journal of
memories form. In humans, various types of heritable Neuroscience 16, 3,4863,499.
mental retardation will become associated with ho- Heisenberg, M., Borst, A., Wagner, S., and Byers, D. (1985). Dro-
mologues of these memory genes and the neuro- sophila mushroom body mutants are deficient in olfactory
biological pathways in which they participate (Oike et learning. Journal of Neurogenetics 2, 130.
Menzel, R., Hammer, M., Braun, G., Mauelshagen, J., and Sugawa,
al., 1999; Petrij et al., 1995; Zhang et al., 2001). Ulti-
M. (1991). Neurobiology of learning and memory in honey-
mately, these mammalian homologues will become bees. In R. C. Fisher, ed., The behavior and physiology of bees, pp.
targets of drug discovery, thereby yielding viable 323353. London: CAB International.
therapies for those who suffer from cognitive dysfunc- Oike, Y., Hata, A., Mamiya, T., Kaname, T., Noda, Y., Suzuki, M.,
tion (Scott et al., 2002). Yasue, H., Nabeshima, T., Araki, K., and Yamamura, K.
(1999). Truncated CBP protein leads to classical Rubinstein-
See also: APLYSIA: CLASSICAL CONDITIONING AND Taybi syndrome phenotypes in mice: Implications for a domi-
OPERANT CONDITIONING; GENETIC nant-negative mechanism. Human Molecular Genetics 8, 387
SUBSTRATES OF MEMORY: HIPPOCAMPUS; 396.
INSECT LEARNING; INVERTEBRATE LEARNING: Pascual, A., and Preat, T. (2001). Localization of long-term memo-
ASSOCIATIVE LEARNING AND MEMORY ry within the Drosophila mushroom body. Science 294, 1,115
PROCESSING IN BEES; INVERTEBRATE 1,117.
LEARNING: C. ELEGANS; SECOND MESSENGER Petrij, F., Giles, R., Dauwerse, H., Saris, J., Hennekam, R., Masuno,
M., Tommerup, N., van Ommen, G.-J. B., Goodman, R., Pe-
SYSTEMS
ters, D., and Breuning, M. (1995). Rubinstein-Taybi syn-
Bibliography drome caused by mutations in the transcriptional co-activator
Aceves-Pina, E. O., Booker, R., Duerr, J. S., Livingstone, M. S., CBP. Nature 376, 348351.
Quinn, W. G., Smith, R. F., Sziber, P. P., Tempel, B. L., and Quinn, W., Sziber, P. P., and Booker, R. (1979). The Drosophila
Tully, T. P. (1983). Learning and memory in Drosophila, stud- memory mutant amnesiac. Nature 277, 212214.
ied with mutants. Cold Spring Harbor Symposia on Quantitative Roman, G., and Davis, R. L. (2001). Molecular biology and anato-
Biology 48, 831839. my of Drosophila olfactory associative learning. Bioessays 23,
Bailey, C. H., Bartsch, D., and Kandel, E. R. (1996). Toward a mo- 571581.
lecular definition of long-term memory storage. Proceedings of Rosay, P., Armstrong, J. D., Wang, Z., and Kaiser, K. (2001). Syn-
the National Academy of Sciences of the United States of America 93, chronized neural activity in the Drosophila memory centers
13,44513,452. and its modulation by amnesiac. Neuron 30, 759770.
Boynton, S., and Tully, T. (1992). latheo, a new gene involved in as- Saitoe, M., and Tully, T. (2000). Making connections between syn-
sociative learning and memory in Drosophila melanogaster aptic and behavioral plasticity in Drosophila. In J. McEachern
identified from P element mutagenesis. Genetics 131, 655 and C. Shaw, eds., Toward a theory of neuroplasticity, pp. 193-
672. 220. New York: Psychology Press.
Connolly, J. B., Roberts, I. J., Armstrong, J. D., Kaiser, K. M. F., Scott, R., Bourtchuladze, R., Gossweiler, S., Dubnau, J., and Tully,
Tully, T., and OKane, C. J. (1996). Associative learning dis- T. (2002). CREB and the discovery of cognitive enhancers.
rupted by impaired Gs signaling in Drosophila mushroom bo- Journal of Molecular Neuroscience.
dies. Science 274, 2,1042,107. Tully, T., Bolwig, G., Christensen, J., Connolly, J., DelVecchio, M.,
Connolly, J. B., and Tully, T. (1998). Behaviour, learning, and DeZazzo, J., Dubnau, J., Pinto, S., Regulski, M., Svedberg, B.,
memory. In D. B. Roberts, ed., Drosophila: A practical ap- and Velinzon, K. (1996). A return to genetic dissection of
memory in Drosophila, Cold Spring Harbor Symposium on Quanti- Wright, N. J. D., and Zhong, Y. (1995). Characterization of K+ cur-
tative Biology 61, 207218. rents and the cAMP-dependent modulation in cultured Dro-
Tully, T., Preat, T., Boynton, S. C., and Del Vecchio, M. (1994). Ge- sophila mushroom body neurons identified by lacZ expression.
netic dissection of consolidated memory in Drosophila. Cell 79, Journal of Neuroscience 15, 1,0251,034.
3547. Zars, T., Wolf, R., Davis, R., and Heisenberg, M. (2000). Tissue-
Wang, Y., Wright, N. J., Guo, H., Xie, Z., Svoboda, K., Malinow, specific expression of a type I adenylyl cyclase rescues the ru-
tabaga mutant memory defect: in search of the engram.
R., Smith, D. P., and Zhong, Y. (2001). Genetic manipulation
Learning and Memory 7, 1831.
of the odor-evoked distributed neural activity in the Drosophila
Zhang, Y. Q., Bailey, A. M., Matthies, H. J., Renden, R. B., Smith,
mushroom body. Neuron 29, 267276.
M. A., Speese, S. D., Rubin, G. M., and Broadie, K. (2001).
Wong, S., Athos, J., Figueroa, X., Pineda, V., Schaefer, M., Chav-
Drosophila fragile X-related gene regulates the MAP1B homo-
kin, C., Muglia, L., and Storm, D. (1999). Calcium-stimulated log Futsch to control synaptic structure and function. Cell 107,
adenylyl cyclase activity is critical for hippocampus- 591603.
dependent long-term memory and late phase LTP. Neuron
23, 787798. Yadin Dudai
Revised by Tim Tully
J
JAMES, WILLIAM (18421910) Like most people of his time, James saw psycholo-
gy as part of philosophy, an empirical approach to-
The American philosopher and psychologist William ward philosophical questions. Findings in science and
James was one of the most important American intel- medicine stimulated a vision of a physiological psy-
lectuals of his era, making key contributions to the de- chology in James, spurred by the writings of people
velopment of both philosophical pragmatism and like Wilhelm Wundt, Ivan Sechenov, John Hughlings
psychological theory. Jackson, Henry Maudsley, and Theodor Meynert. His
varied academic appointments reflected his early ef-
James was born in New York City in 1842, the
forts to establish a new psychology in philosophy
oldest of five children (his brother Henry became a fa-
and then, skeptical about much of the normal sci-
mous novelist). His father was a man of leisure who
ence that began growing up around him, to seek a
gave his children an unusual and rich education
more meaningful inquiry into psychic life through
based on large amounts of travel and instruction in spiritualism and the study of exceptional mental
Europe. Young William set out to be an artist, ap- states.
prenticing with William Morris Hunt for a year, but
then turned toward science. He entered Harvards The culmination of Jamess first was his 1890
Lawrence Scientific School in 1861 and worked with Principles of Psychology, an enormously popular and in-
Charles William Eliot in chemistry and Jeffries fluential work that brought some people into new
Wyman in comparative anatomy. He then entered psychology and induced many others to accept it as
Harvard Medical School, took a year off to go with a possibility. The two volumes are a delight to read
today, brimming with life, ideas, and insights.
Louis Agassiz on an expedition to the Amazon, and
eventually received his M.D. in 1869. James defines psychology as the Science of Men-
tal Life, both of phenomena and their conditions (p.
James never practiced medicine but pursued an 1). His book explores the psyche using a mixture of
academic career at Harvard. In 1872 he was appoint- introspective, physiological, medical, comparative,
ed instructor in physiology and anatomy. He became and experimental observations. The first six chapters
assistant professor of philosophy in 1880 and full pro- are given to the physiological preliminaries, a psy-
fessor in 1885. In 1889 he was named professor of chobiology of mind. Both the anatomy and the de-
psychology, returning to philosophy in 1892 when tailed physiology of the brain are achievements of the
Hugo Mnsterberg came from Germany to take present generation, or rather we may say (beginning
charge of Harvards psychological laboratory. James with Meynert) the last twenty years (p. 14). James
retired in 1907 and spent his last years writing out his sets forth a picture of a hierarchically organized brain
systematic philosophy. and mind, one that remains today, with amplifica-
297
298 JAMES, WILLIAM
tions of the behavior of animals, mental patients, chil-

dren, and people of other cultures, and inferences
about the mind drawn from artifacts such as human
languages, customs, and social and political institu-
tions.
Jamess discussions of learning and memory are
not well integrated. His discussion of learning rests
largely on comparative observations, whereas he ap-
proaches memory introspectively. His book appeared
only five years after Hermann Ebbinghaus launched
the systematic study of human memory. James dis-
cusses Ebbinghauss work in his chapter on memory,
but the discussion is clearly an appendage. An inte-
grated, fully elaborated discussion of human learning
and memory began to emerge five years after James
published his book.
A Parliamentary Theory of Learning

Like all evolutionary writers, James addresses the
problem of the instinctual and the learned in the or-
ganization of human activity. Contrary to most writ-
ers, he argues that humans have many instincts and
that they play a large role in the behavior of higher
organisms.
Nature [in the lower wild animals] has made
them act always in the manner which would
be oftenest right. There are more worms unat-
tached to hooks than impaled upon them;
therefore, on the whole, says Nature to her
William James (Archive Photos, Inc.)
fishy children, bite at every worm and take
your chances. But as her children get higher,
and their lives more precious, she reduces the
tions and amendments, a framing conception of con-
risks. . . . Nature implants contrary impulses to act
temporary neuropsychology. There are levels of be-
on many classes of things, and leaves it to slight
havioral organization: first reflexes and automatisms,
alterations in the conditions of the individual
then instincts, then habits, and then, finally, volun- case to decide which impulse shall carry the
tary and planned activity. Lower levels are more me- day. (pp. 1,0121,013)
chanical and sense-driven; higher centers are the seat
of spontaneity and intellectual control. In all ages The proliferation of instincts in the higher ani-
the man whose determinations are swayed by refer- mals leads to the beginnings of deliberation and
ence to the most distant ends has been held to possess choice. In higher animals, instincts do not remain
the highest intelligence (p. 23). fixed action patterns. Once an instinct is exercised, it
produces consequences, and anticipations of those
In Chapter 7, James turns to psychological inqui- consequences henceforth accompany instinctive im-
ry, which, he says, uses three methods: introspection, pulses. The experienced organism faces ever more
experiment, and the comparative approach. Intro- new situations, armed with a number of impulses; the
spection is the foundation, what we have to remem- organism is not a slave to any one instinct; and there
ber first and foremost and always (p. 185). The ex- are rational anticipations of the consequences various
perimental method had been developed in Germany. actions might bring. Jamess parliamentary theory of
[It] taxes patience to the utmost, and could hardly the inner competition among impulses and their
have arisen in a country whose natives could be bored anticipated consequences foreshadows twentieth-
(p. 192). Experimental psychology explores psycho- century behaviorism. Edward L. Thorndike, Jamess
physics, sensation and perception, attention span, student, was the bridging figure. His connectionism
and rote memory. The comparative method is loos- built the parliamentary theory into a logic of response
est, wild work, and includes psychological observa- choice in problem-solving situations; that connection-
JAMES, WILLIAM 299
ism, in turn, has influenced all subsequent learning Conception, Discrimination and Comparison, Memo-
theories. ry, and so on.
Optimal human development, James argues, im- In older writings, memories are referred to as re-
plies the fruition of as many instinctive tendencies as turns or reinstatements of experiences of the past;
possible. After attributing to humans a large set of in- James, looking at mental life with a fresh eye, argues
stincts (more, he says, than any animal), James writes, there must be more to the experience of a memory
than that. Memory, he says, is the knowledge of an
In a perfectly-rounded development every event, or fact, of which meantime we have not been
one of these instincts would start a habit to- thinking, with the additional consciousness that we have
wards certain objects and inhibit a habit to- thought or experienced it before (p. 610). It is not
wards others. Usually this is the case; but, in enough to see or hear something that one has seen or
the one-sided development of civilized life, it heard before. There must be some aspect of the expe-
happens that the timely age goes by in a sort rience that says the event has occurred before, not in
of starvation of objects, and the individual just any past but in the persons past. It must have
then grows up with gaps in his psychic consti- that warmth and intimacy which were so often spo-
tution which future experiences can never ken of in the chapter on the self, as characterizing all
fill. Compare the accomplished gentleman experiences appropriated by the thinker as his
with the poor artisan or tradesman of a city: own. James concludes that a memory is far more
during the adolescence of the former, objects than an image or copy of a fact in the mind, that it is
appropriate to his growing interests, bodily in fact a very complex representation with objective,
and mental, were offered as fast as his interpersonal, and metacognitive components. What we
ests awoke, and, as a consequence, he is usually refer to as memory, James argues, is a form
armed and equipped at every angle to meet of belief.
the world. . . . Over the city poor boys youth Jamess analyses of habit, as we have seen, had an
no such golden opportunities were hung, and immediate and large influence on the development
in his manhood no desires for most of them of psychology in subsequent decades. His discussions
exist. Fortunate it is for him if gaps are the of memory and other cognitive phenomena were out
only anomalies his instinctive life presents; of step with the elementistic introspectionism of his
perversions are too often the fruit of his un- time. It seems very likely that contemporary cognitive
natural bringing-up. (pp. 1,0561,057) psychologyin particular, the study of personal, nar-
rative memoriesis picking up the thread of William
Jamess thought.
Memories as Beliefs
Bibliography
James approaches the phenomena of human
Bjork, D. W. (1983). The compromised scientist: William James in the de-
memory introspectively, though his conception of the velopment of American psychology. New York: Columbia Univer-
presenting phenomena of mental life differs radically sity Press.
from that of previous psychological writings. Begin- Feinstein, H. M. (1984). Becoming William James. Ithaca: Cornell
ning with John Locke, and continuing through the re- University Press.
James, W. (1890). The principles of psychology. 2 vols. Reprinted
search programs of nineteenth-century brass-
(1981) in F. H. Burkhardt, F. Bowers, and I. K. Skrupskelis,
instruments laboratories, psychologists repeatedly as- eds,. The works of William James. Cambridge, MA: Harvard
serted that human experience begins with simple University Press.
sensations and ideas. James argues that this is false to (1899). Talks to teachers on psychology: And to students on some
experience. Simple sensations and ideas are, in of lifes ideals. New York: Holt.
(1902). The varieties of religious experience: A study in human
fact, abstractions contrived from experience. What
nature. New York: Longmans.
presents itself to the mind is an always dynamic, often (1983). Essays in psychology. Vol. 13 in F. H. Burkhardt, F.
inchoate, moving flowa fluidity that in the Principles Bowers, and I. K. Skrupskelis, eds., The works of William James.
he calls the stream of consciousness and in later Cambridge, MA: Harvard University Press.
writings he calls pure experience. Redefining the Myers, Gerald E. (1986). William James: His life and thought. New
Haven: Yale University Press.
mental life with which the psychologist must deal,
Perry, R. B. (1935). The thought and character of William James, as re-
James at once sets aside the traditional denizens of vealed in unpublished correspondence and notes, together with his
that mental lifesensations, ideas, faculties. In re- published writings. 2 vols. Boston: Little, Brown.
turn, he has the obligation and the opportunity to ask Taylor, E. (1983). William James on exceptional mental states: The 1896
in a very wide-open way where in the fluidity one may Lowell lectures. New York: Scribners.
locate the conventional chapter headings: Attention, Sheldon H. White
K
KAMINS BLOCKING EFFECT: than temporal, relationship between the CS and US
NEURONAL SUBSTRATES is the essential determinant of classical conditioning
(Kamin, 1968; Rescorla, 1968; Wagner et al., 1968)
Blocking is a classical conditioning phenomenon that and profoundly shaped subsequent thinking about as-
has profoundly influenced thinking about associative sociative learning. One of the three findings was the
learning. This article will discuss the key characteris- phenomenon of blocking, discovered by Leon Kamin.
tics of blocking and the role it may play in several
In a typical blocking experiment (see Table 1), one
mammalian brain systems in regulating particular
CS (denoted A) is first extensively paired with a US
types of learning.
(AUS). Then a second CS (denoted B) undergoes
compound conditioning with A and the same US (AB
Introduction and Significance US). Later, when B is tested, almost no (or very little)
conditioning has accrued to B. However, if A was pre-
Classical or Pavlovian conditioning is an elemen- viously not (or weakly) conditioned with the US, then
tary form of associative learningsystematically de-
B (as well as A) accrues substantial associative strength
scribed by a Russian physiologist Ivan Pavlov in the
during the compound conditioning phase. Thus, con-
early twentieth centurythat is considered an essen-
ditioning to B during the compound (ABUS) condi-
tial building block for complex learning. Typically,
tioning is inversely proportional to the magnitude of
classical conditioning occurs when an initially neutral
previous conditioning to A, and it is not Bs temporal
stimulus (conditional stimulus, CS) is paired in close
relationship with the US that determines whether or
temporal proximity (or contiguously) with a biologi-
not conditioning will develop to B.
cally significant stimulus (unconditional stimulus, US)
that elicits an unlearned, reflexive behavior (uncondi- It was originally suggested that if a US is already
tional response, UR). Through CS-US association for- fully predicted by one stimulus and if the addition of
mation, the animal acquires a behavior (conditional a new stimulus provides no new significant informa-
response, CR) to the CS that typically resembles the tion about the US, then the US will not activate or
UR (but not always), precedes the US in onset time,
support the learning process responsible for estab-
and reaches a maximum magnitude at about the time
lishing a new CS-US association (Kamin, 1968, 1969).
of US onset.
Since its discovery, blocking has become the corner-
Although the temporal arrangement between the stone of all modern learning theories (see Rescorla
CS and the US was thought to be the critical feature and Wagner, 1972; Mackintosh, 1975; Pearce and
of classical conditioning, three separate studies in the Hall 1980; Wagner, 1981; Sutton and Barto, 1990)
late 1960s revealed that the informational, rather and has been considered as an instance of cognitive
301
302 KAMINS BLOCKING EFFECT: Neuronal Substrates
Table 1 well-predicted reinforcer) reduces the associability

(the ability to enter into new associations) of CSs pres-
ent on that trial (Mackintosh, 1975; Pearce and Hall,
1980). In this view, blocking occurs because stimulus
B rapidly loses associability because no surprising
events occur after the compound CS presentation.
Conditioning to B is blocked because its associability
is reduced (rapidly to zero) and thus it does not enter
into associations with the US, not because of any loss
of effectiveness of the US as a reinforcer.
The US and CS processing theories, however,
make somewhat different predictions regarding
blocking; for example, whereas the Rescorla-Wagner
model predicts one trial blocking, theories based on
processing in classical conditioning (Thompson et al., CS processing do not; conversely, CS-processing
1998). models can account for the disruption of blocking
(unblocking) when the magnitude of the reinforcer
is decreased simultaneously with the addition of an el-
Theoretical Aspects
ement to the CS (Dickinson et al., 1976; Holland and
Robert Rescorla and Allan Wagner (1972) pro- Gallagher, 1993), a phenomenon that causes difficul-
posed a simple learning equation based on US pro- ty for the Rescorla-Wagner model.
cessing Progress in identifying neural circuits subserving
particular learning phenomena has yielded insight
Vn = ( Vi) into the neural basis of this phenomenon. We will
briefly discuss the involvement of the cerebellum,
that elegantly describes the blocking effect; where amygdala, midbrain dopamine neurons, and sept-
is a learning constant; is the maximum associative ohippocampal systems in blocking.
strength conditionable with a given US; Vi is the
sum of associative strengths between all CS elements
present and the US; and Vn is the change in the asso- Cerebellum and Eyeblink Conditioning
ciative strength of a particular CS on trial n. In es- In eyeblink conditioning the animals learn to re-
sence, the associative strength between the CS-US is spond with eyelid closure to a CS (tone or light) that
driven by errors between the expected US and the has been contingently paired with a US (e.g., airpuff
actual US. According to this equation, blocking occurs to the eye). The anatomically rooted neural circuit
when the associative strength acquired by a CS (A) underlying eyeblink conditioning (Kim and Thomp-
that was paired with a US (Phase I; A-US) reaches the son 1997) is remarkably similar to the US processing
value. Then, when a new CS (B) is introduced dur- view of the Rescorla-Wagner model (see Figure 1).
ing the compound conditioning (Phase II; AB-US), Briefly, the CS pathway consists of excitatory
stimulus A already fully predicts the US and has ac- mossy-fiber projections from the pontine nuclei to the
quired all of its associative strength, because cerebellum, whereas the US pathway consists of excit-
atory climbing-fiber projections from the inferior
VA = . olive to the cerebellum. The cerebellum, in turn,
sends monosynaptic -aminobutyric acid (GABA)-
Hence, the stimulus B will not accrue any associative containing projections to the inferior olive. As GABA
strength, because neurotransmitters generally exert inhibitory influ-
ences, it is conceivable that this cerebello-olivary
VB = (VA+VB) = 0 pathway serves a negative feedback function and thus
gates the inferior olive activity. Consistent with this
because view, neurons in the inferior olive show evoked neural
activity to the airpuff US during the initial stage of
VA = . CS-US training (before the animal exhibits any CRs),
but not when the animals perform CRs during CS-US
Other learning theories based on CS processing trials (Sears and Steinmetz, 1991; Hesslow and Ivars-
rather than US processing can also effectively accom- son, 1996).
modate blocking by postulating that the absence of If the cerebello-olivary GABAergic projection
surprising events on a trial (e.g., in the presence of a serves a negative-feedback function in regulating the
KAMINS BLOCKING EFFECT: Neuronal Substrates 303
US information from reaching the cerebellum (where Figure 1

the CS-US association is thought to take place), then
this can explain the blocking effect in eyeblink condi-
tioning in the following manner: According to Figure
1, blocking will occur when a CS (e.g., auditory CSA)
acquires sufficient associative strength to activate the
cerebellum, which then inhibits the inferior olive (via
the GABA-containing cerebello-olivary pathway)
from US activation. Because the input representing
the US can no longer reach the cerebellum, it cannot
support conditioning to a new CS (e.g., visual CSB).
Consistent with this view, a study found that pharma-
cological blockade of the cerebello-olivary activity (by
infusing a GABA antagonist directly into the inferior
olive) during the compound tone/light-airpuff condi-
tioningthereby disinhibiting inferior olive neu-
ronsprevented blocking (Kim et al., 1998).
Amygdala and Fear Conditioning

The Rescorla-Wagner equation has been mapped onto a
Fear conditioning typically involves pairing a
simplified eyeblink conditioning circuit to illustrate how one
tone (or light) CS with a nociceptive shock US. After form of blocking might occur in the mammalian brain.
few CS-US pairings, the CS not only becomes capable
of activating fear responses but also acquires the abili-
ty to inhibit/decrease sensitivity to nociceptive stimuli
al., 1994; Schultz et al., 1993). For example, some re-
(e.g., foot shock) via a conditioned analgesic response
searchers report that many dopamine neurons in the
(involving partly endogenous opioids) (Chance,
substantia nigra (SN) and the ventral tegmental area
1980; MacLennan et al., 1980). Because the amygdala
(VTA) show phasic responses to the delivery of liquid
has been implicated as the locus of fear conditioning
reward in monkeys undergoing a spatial delayed-
and sends projections to hypothalamus and brain-
response task (Schultz et al., 1993). However, once
stem nuclei that mediate various fear responses (Le-
learning is established (i.e., the animal learns that a
Doux, 1997), some researchers believe that as fear
light cue predicts the reward), the delivery of the re-
conditioning proceeds, the ability of the nociceptive
ward no longer elicits phasic responses in dopaminer-
US to support fear conditioning diminishes as a func-
gic neurons. Such negative-feedback circuits in the
tion of the CS ability to elicit conditioned analgesia
brain may well provide the neuronal instantiation of
via the amygdala (see Figure 2). Consistent with this
notion, systemic administration of the opioid antago- behavioral phenomena of blocking.
nist naloxone during compound fear conditioning Indeed, when examined in a blocking paradigm,
(phase II, blocking procedure) attenuated blocking dopamine neurons did not fire in response to the
(Fanselow and Bolles, 1979; Fanselow, 1998). It is blocked element of a compound stimulus but did fire
possible that naloxone, by opposing the conditioned to one element of a compound control stimulus
opioid analgesia response to the first CS, prevented (Waelti et al., 2001). These neurons seem to enable
the CS-associated decline in the nociceptive USs abil- blocking phenomena in more complex forms of asso-
ity to support fear conditioning to the second CS ciative learning that involve appetitive stimuli and
(during compound conditioning). However, the locus more complex motor responses and that depend on
of naloxones effects on blocking is not known, and forebrain structures by conveying a prediction
further research is necessary to determine whether error signal that enables the formation of associa-
the US-evoked responses in the amygdala decrease as tions (Schultz and Dickinson, 2000). On a neural
a function of fear conditioning. level, this system is analogous to that in the cerebel-
lum (see Figure 1), replacing the inferior olive with
the SN/VTA and the cerebellum with the basal gan-
Striatum and Dopaminergic Midbrain glia and neocortex (which receive dopaminergic
Neurons input from the SN/VTA). These findings are also con-
Other brain structures may employ a negative- sistent with reports of disrupted blocking after ma-
feedback mechanism similar to that of the cerebellum nipulations of the dopamine system (Crider et al.,
to regulate the US or reinforcing input (Graybiel et 1982, 1986).
304 KAMINS BLOCKING EFFECT: Neuronal Substrates
Figure 2 nisms. Baxter et al. (1999) examined blocking in rats

with selective removal of medial septal cholinergic
neurons. These lesions remove hippocampal cholin-
ergic input and impair reductions in CS associability
in several different learning paradigms (Baxter et al.,
1997). Although lesioned rats showed a normal block-
ing effect when tested in an extinction test with the
added cue B, learning about B was facilitated relative
to controls in subsequent savings tests, in which B
alone was paired with reinforcement or served as a
conditioned inhibitor. These findings suggested that
although B did undergo a loss of associability as a
consequence of the blocking procedure, this loss of
associability was not necessary for blocking of learn-
ing about B to occur.
A simple model of fear conditioning illustrating how the
regulation of US processing may account for blocking. A CS-
evoked conditioned analgesia response is hypothesize to dampen Conclusion
sensitivity/reactivity to nociceptive US, thereby inhibiting or Functionally, blocking (or other similar process-
decreasing the reinforcers ability to support further es) may play an important role in how animals process
conditioning. and attend to information in their environments. Be-
cause animals are constantly bombarded by numer-
ous stimuli, it benefits them respond selectively to
Septohippocampal System those stimuli that reliably predict biologically signifi-
The hippocampus appears to play a role in the cant events. Other stimuli that provide no new useful
capacity to reduce the associability of a CS (Han, Gal- information should be disregarded (or filtered), oth-
lagher, and Holland, 1995; Kaye and Pearce, 1987; erwise animals would be constantly forming unneces-
Solomon and Moore, 1975; Honey and Good, 1993; sary associations with various stimuli in their sur-
Gallo and Cndido, 1995; Reilly, Harley, and Re- roundings, thus inviting information overload.
vusky, 1993); moreover, this function of the hippo- Indeed, malfunctioning selective attention mecha-
campus appears to depend on the integrity of its cho- nisms in the brain may contribute to psychopatholog-
ical conditions such as schizophrenia (Bender et al.,
linergic input (Baxter et al., 1997). Based on the view
2001). The behavioral phenomenon of blocking,
that the blocking effect results from variations in pro-
which appears to use heuristic negative-feedback at-
cessing of the US, one would not expect the hippo-
tentional processes, may circumvent such redundant
campal damage to affect blocking, and some experi-
learning.
ments have found just that (Garrud et al., 1984). In
contrast, CS-processing theories of blocking propose
See also: CONDITIONING, CELLULAR AND NETWORK
that the associability of the added element of the com- SCHEMES FOR HIGHER-ORDER FEATURES OF;
pound CS is reduced because the reinforcer is already LEARNING THEORY: CURRENT STATUS;
well predicted. On this view, damage to the hippo- REINFORCEMENT OR REWARD IN LEARNING:
campus (or its cholinergic input) should eliminate CEREBELLUM; REINFORCEMENT OR REWARD IN
this reduction in associability and, by extension, the LEARNING: STRIATUM
blocking effect. In support of this view, several investi-
gators have reported that lesions of the hippocampus Bibliography
reduced or eliminated blocking (Gallo and Cndido, Baxter, M. G., Gallagher, M., and Holland, P. C. (1999). Blocking
1995; Rickert et al., 1978; Rickert et al., 1981; Solo- can occur without losses in attention in rats with selective le-
sions of hippocampal cholinergic input. Behavioral Neuro-
mon 1977). These variations in the effectiveness of science 113, 881890.
hippocampal system damage on blocking are consis- Baxter, M. G., Holland, P. C., and Gallagher, M. (1997). Disrup-
tent with the view that blocking has diverse origins tion of decrements in conditioned stimulus processing by se-
(Holland, 1988) and may reflect different contribu- lective removal of hippocampal cholinergic input. Journal of
tions of CS and US processing systems in different
Bender, S., Muller, B., Oades, R. D., and Sartory, G. (2001). Condi-
conditioning situations. tioned blocking and schizophrenia: A replication and study of
the role of symptoms, age, onset-age of psychosis and illness-
Experiments in an appetitive Pavlovian condi- duration. Schizophrenia Research 49, 157170.
tioning paradigm suggested that a blocking para- Chance, W. T. (1980). Autoanalgesia: opiate and nonopiate mecha-
digm produces both CS- and US-processing mechanisms.Neuroscience and Biobehavioral Reviews 4, 5567.
KAMINS BLOCKING EFFECT: Neuronal Substrates 305
Crider, A., Blockel, L., and Solomon, P. R. (1986). A selective at- Pavlov, I. P. (1927). Conditioned Reflexes. London: Oxford Universi-
tention deficit in the rat following induced dopamine recep- ty Press.
tor supersensitivity.Behavioral Neuroscience 100, 315319. Pearce, J. M., and Hall, G. (1980). A model for Pavlovian learning:
Crider, A., Solomon, P. R., and McMahon, M. A. (1982). Disrup- Variations in the effectiveness of conditioned but not of un-
tion of selective attention in the rat following chronic d- conditioned stimuli. Psychological Review 87, 532552.
amphetamine administration: relationship to schizophrenic Reilly, S., Harley, C., and Revusky, S. (1993). Ibotenate lesions of
attention disorder. Biological Psychiatry 17, 351361. the hippocampus enhance latent inhibition in conditioned
Dickinson, A., Hall, G., and Mackintosh, N. J. (1976). Surprise and taste aversion and increase resistance to extinction in condi-
the attenuation of blocking.Journal of Experimental Psychology: tioned taste preference. Behavioral Neuroscience 107, 996
Animal Behavior Processes 2, 313322. 1,004.
Fanselow, M. S. (1998). Pavlovian conditioning, negative feedback, Rescorla, R. A. (1968). Probability of shock in the presence and ab-
and blocking: Mechanisms that regulate associative forma- sence of CS in fear conditioning. Journal of Comparative and
tion. Neuron 20, 625627. Physiological Psychology 66, 15.
Fanselow, M. S., and Bolles, R. C. (1979). Naloxone and shock Rescorla, R. A., and Wagner, A. R. (1972). A theory of Pavlovian
elicited freezing in the rat. Journal of Comparative and Physiolog- conditioning: Variations in the effectiveness of reinforcement
ical Psychology 94, 736744. and nonreinforcement. In A. H. Black and W. F. Prokasy,
Gallo, M., and Cndido, A. (1995). Dorsal hippocampal lesions im- eds., Classical conditioning: II. Current research and theory. New
pair blocking but not latent inhibition of taste aversion learn- York: Appleton-Century-Crofts.
ing in rats. Behavioral Neuroscience 109, 413425. Rickert, E. J., Bennett, T. L., Lane, P., and French, J. (1978). Hip-
Garrud, P., Rawlins, J. N. P., Mackintosh, N. J., Goodall, G., Cot- pocampectomy and the attenuation of blocking. Behavioral Bi-
ton, M. M., and Feldon, J. (1984). Successful overshadowing
ology 22, 147160.
and blocking in hippocampectomized rats. Behavioural Brain
Rickert, E. J., Lorden, J. F., Dawson Jr., R., and Smyly, E. (1981).
Research 12, 3953.
Limbic lesions and the blocking effect. Physiology and Behavior
Graybiel, A. M., Aosaki, T., Flaherty, A. W., and Kimura, M. (1994).
26, 601606.
The basal ganglia and adaptive motor control. Science 265,
Schultz, W., Apicella, P., and Ljungberg, T. (1993). Responses of
1,8261,831.
monkey dopamine neurons to reward and conditioned stimu-
Han, J.S., Gallagher, M., and Holland, P. C. (1995). Hippocampal
li during successive steps of learning a delayed response task.
lesions disrupt decrements but not increments in condi-
tioned stimulus processing. Journal of Neuroscience 15, 7,323
Schultz, W., and Dickinson, A. (2000). Neuronal coding of predic-
7,329.
tion errors. Annual Review of Neuroscience 23, 473500.
Hesslow, G., and Ivarsson, M. (1996). Inhibition of the inferior
Sears, L. L., and Steinmetz, J. E. (1991). Dorsal accessory inferior
olive during conditioned responses in the decerebrate ferret.
olive activity diminishes during acquisition of the rabbit classi-
Experimental Brain Research 110, 3646.
Holland, P. C. (1988). Excitation and inhibition in unblocking. cally conditioned eyelid response. Brain Research 545, 114
Journal of Experimental Psychology: Animal Behavior Processes 14, 122.
261279. Solomon, P. R. (1977). Role of the hippocampus in blocking and
Holland, P. C., and Gallagher, M. (1993). Effects of amygdala cen- conditioned inhibition of the rabbits nictitating membrane
tral nucleus lesions on blocking and unblocking. Behavioral response. Journal of Comparative and Physiological Psychology 91,
Neuroscience 107, 235245. 407417.
Honey, R. C., and Good, M. (1993). Selective hippocampal lesions Solomon, P. R., and Moore, J. W. (1975). Latent inhibition and
abolish the contextual specificity of latent inhibition and con- stimulus generalization of the classically conditioned nictitat-
ditioning.Behavioral Neuroscience 107, 2333. ing membrane response in rabbits (Oryctolagus cuniculus) fol-
Kamin, L. J. (1968). Attention-like processes in classical condition- lowing dorsal hippocampal ablation. Journal of Comparative
ing. In M. R. Jones, ed., Miami Symposium. Predictability, and Physiological Psychology 89, 1,1921,203.
Behavior and Aversive Stimulation. Miami: University of Miami Sutton, R. S., and Barto, A. G. (1990). Time-derivative models of
Press. Pavlovian reinforcement. In M. Gabriel and J. Moore, eds.,
(1969). Predictability, surprise, attention, and conditioning. Learning and computational neuroscience. Cambridge, MA: MIT
In. B. A. Campbell and R. M. Church, eds., Punishment and Press.
Aversive Behavior. New York: Appleton-Century-Crofts. Thompson, R. F., Thompson, J. K., Kim, J. J., Krupa, D. J., and
Kaye, H., and Pearce, J. M. (1987). Hippocampal lesions attenuate Shinkman, P. G. (1998). The nature of reinforcement in cere-
latent inhibition of a CS and of a neutral stimulus. Psychobiolo- bellar learning. Neurobiology of Learning and Memory 70, 150
gy 15, 293299. 176.
Kim, J. J., Krupa, D. J., and Thompson, R. F. (1998). Inhibitory Wagner, A. R. (1981). SOP: a model of automatic memory process-
cerebelloolivary projections and blocking effect in classical ing in animal behavior. In N. E. Spear and R. R. Miller, eds.,
conditioning. Science 279, 570573. Information processing in animals: Memory mechanisms. Hillsdale,
Kim, J. J., and Thompson, R. F. (1997). Cerebellar circuits and syn- NJ: Erlbaum.
aptic mechanisms involved in classical eyeblink conditioning. Wagner, A. R., Logan, F. A., Haberlandt, K., and Price, T. (1968).
Trends in Neuroscience 4, 177181. Stimulus selection in animal discrimination learning. Journal
LeDoux, J. E. (1997).The Emotional Brain. New York: Simon and of Experimental Psychology 76, 171180.
Schuster. Waelti, P., Dickinson, A., and Schultz, W. (2001). Dopamine re-
Mackintosh, N. J. (1975). A theory of attention: Variations in the
sponses comply with basic assumptions of formal learning
associability of stimuli with reinforcement. Psychological Review
theory. Nature 412, 4348.
82, 276298.
MacLennan, A. J., Jackson, R. L., and Maier, S. F. (1980). Condi-
tioned analgesia in the rat. Bulletin of the Psychonomic Society 15,
Jeansok J. Kim
387390. Mark G. Baxter
306 KNOWLEDGE SYSTEMS AND MATERIAL-SPECIFIC MEMORY DEFICITS
KNOWLEDGE SYSTEMS AND tinuous variations in the degree of engram accessibili-

MATERIAL-SPECIFIC MEMORY ty.
DEFICITS Recent studies using PET (positron emission to-
The face of cognitive neuroscience has changed dras- mography) and fMRI (functional magnetic imaging)
tically since the midtwentieth century. In the past, le- methodology have further elucidated the neuroana-
sions were the only basis for inference regarding the tomical distinctiveness of these systems. There is evi-
functional neuroanatomy of normal cognition. Today dence to indicate that many aspects of semantic mem-
the tools of cognitive neuroscience include various ory processing involve regions of the left lateral
methods of neuroimaging, both structural and func- prefrontal cortex and the anterior temporal cortex.
tional, in normal subjects. Some findings converge to suggest that this system is
organized hierarchically along the posterior to anteri-
or axis (Martin and Chao, 2001). Functional imaging
Classifications of Memory Systems studies of episodic memory have not provided sup-
port for the view that these processes are mediated
The structure of knowledge representation in the
primarily by the medial temporal lobe memory sys-
brain is elucidated by the studies of specific dissocia-
tem (Bookheimer, 1996). Researchers have suggested
tions of knowledge loss in brain disease. The question
that encoding and retrieval of episodic information
whether there is one memory store or several is of the
might be related more to the functions of the left and
foremost interest. Studies of amnesias are particularly
right prefrontal cortex respectively (Tulving et al.,
illuminating in this respect. In most amnesic syn-
1994), although this view is not consistent with the
dromes, skills are better preserved than facts. Within
bulk of information obtained from lesion studies and
the declarative domain, context-free information is
may be too simplistic to account for the complex na-
usually better preserved than context-dependent in-
ture of these processes.
formation. Generic information is better preserved
than singular information. Generally, the patients
ability to give a conscious account of previously ac- Modality-Specificity of Knowledge Systems
quired knowledge is more likely to be impaired than
the ability to benefit from this knowledge in various Two additional types of knowledge-base dissocia-
behavioral situations. These observations have been tions have been described: sensory modalities and by
interpreted to indicate the neuropsychological reality semantic categories. Modality-specific knowledge loss
of the distinctions between procedural knowledge is exemplified by associative agnosias and modality-
(skills) and declarative knowledge (facts) (Cohen and specific aphasias. In associative agnosias, the subject
Squire, 1980), semantic memory (for general facts) loses the ability to identify objects as members of ge-
and episodic memory (for personal facts) (Tulving, neric categories (Warrington, 1975; Goldberg, 1990).
1983; Kinsbourne and Wood, 1975), generic knowl- The deficit may be isolated, in that it may be present
edge (referring to large classes of equivalent objects) without sensory or language impairment, and without
and singular knowledge (referring to unique entities), dementia. Most important in the context of this anal-
and explicit knowledge (demonstrated through con- ysis, the deficit is modality-specific, and at least three
scious reports) and implicit knowledge (demonstrat- types of associative agnosias have been identified: vi-
ed through behavioral gains) (Schacter, 1987; Tulv- sual object agnosia (McCarthy and Warrington,
ing and Schacter, 1990). 1986), pure astereognosis (Hecaen and Albert, 1978),
and auditory associative agnosia (Vignolo, 1982).
While the phenomenal distinctness of these
knowledge categories is widely accepted, consensus is In each of these agnosias, the ability to interpret
lacking as to whether these knowledge types are me- object meaning is impaired with respect to a distinct
diated by neurally distinct stores, or by different pro- input modality. A patient can see that a watch is round
cessing demands. Robustness and uniformity of the and flat but does not recognize it as a watch in visual
dissociations are often mentioned as the arguments in object agnosia; a patient can feel that it has a smooth,
favor of separate stores (Schacter, 1985). An alterna- glassy surface and a small bump on the side, but does
tive hypothesis is that the difference between proce- not recognize it as a watch in pure astereognosis; and
dural and declarative, semantic and episodic, generic a patient can hear it tick but does not recognize it as
and singular, and explicit and implicit knowledge a watch in auditory associative agnosia. The knowl-
types reflects different degrees of accessibility of en- edge-base impairment in associative agnosias is evi-
grams that are part of the same store. This position dent both in patients inability to correctly name the
suggests that the differences between components of object and in their inability to signal its correct mean-
the above dichotomies are quantitative rather than ing through nonverbal means, such as pantomime.
qualitative, and are discrete approximations of con- However, successful object identification, both verbal
KNOWLEDGE SYSTEMS AND MATERIAL-SPECIFIC MEMORY DEFICITS 307
and nonverbal, will be accomplished in each of these and action names (Goodglass, Klein, Carey, and
three conditions with reliance on other sensory mo- Jones, 1966; Miceli, Silveri, Villa, and Caramazza,
dalities. The existence of modality-specific associative 1984). Further fractionation of noun loss has also
agnosias has led to the hypothesis that the nonverbal been reported (Warrington and Shallice, 1990; Hart,
knowledge base is dimensionalized at least in part by Berndt, and Caramazza, 1985; McKenna and War-
sensory modalities (Warrington, 1975; Goldberg, rington, 1980).
1990; Damasio, 1990; Shallice, 1987). This hypothe- Category-specific knowledge loss also may mani-
sis is strengthened by the presence of double dissocia- fest itself as a selective inability to describe objects or
tions between any two types of associative agnosia. elicit their mental images (Warrington and Shallice,
In modality-specific aphasias, the patient can cor- 1984), or as selective agnosia for certain categories of
rectly identify the object meaning through nonverbal objects but not for others (Nielsen, 1946). The most
means (e.g., pantomime) but cannot come up with a common and consistent observation of category-
correct name (Beauvois, 1982). However, the name is specific knowledge loss is that the knowledge of living
easily retrieved when the patient is allowed to resort objects or foods is more impaired than the knowledge
to other sensory input modalities. The existence of of inanimate objects (Vignolo, 1982; Goldberg, 1989;
sense-specific aphasias further supports the notion of Hart, Berndt, and Caramazza, 1985; Goodglass et al.,
modality-specific knowledge stores, by suggesting 1986). However, researchers have also reported the
that each of them has a separate access to an amodal opposite pattern (Warrington and McCarthy, 1983,
lexical store (Beauvois, 1982). 1987).
Modality-specific associative agnosias are distinct To account for the overwhelming unidirectionali-
not only phenomenally but also neuroanatomically. ty of dissociation, with most studies reporting greater
Each type of agnosia has a distinct cortical territory preservation of knowledge about inanimate than liv-
that is consistent across patients. This observation ing things, and very few reporting the opposite pat-
lends further support to the notion of multiple, neu- tern, it has been proposed that the difference may re-
rally distinct, modality-specific knowledge stores. It flect inherently greater perceptual similarities, and
has been suggested that the modality-specific associa- therefore confusability, within the living domain than
tive agnosias are all linked predominantly to the left within the inanimate domain (Riddoch, Humphreys,
hemisphere (Warrington, 1975; Goldberg, 1990). If Coltheart, and Funnell, 1988). Alternatively, it has
this assertion is true, then the left hemisphere been proposed that the category-specific knowledge
emerges as the repository of multiple knowledge sys- loss may reflect different patterns of relative salience
tems, verbal and nonverbal alike. It has been further of different sensory modalities for different categories
suggested that the neocortical functional organiza- (Goldberg, 1989; Warrington and McCarthy, 1987).
tion within the posterior portion of a hemisphere is The latter helps to explain category-specific double
characterized by continuous, gradiental distributions dissociations. It also interrelates category- and sense-
of cognitive functions. The geometry of the cognitive specific aspects of mental representations.
gradients is determined by the sensory cortices (Gold- The inanimate objects used in most studies are in
berg, 1989). This position is consistent with the no- fact human-made objects or tools. Therefore, it is dif-
tion that representations are dimensionalized in ficult to know which of the two distinctions, living ver-
terms of sensory modalities. Functional neuroimag- sus inanimate or human-made versus natural, best
ing studies in normal subjects have further supported captures the observed differences. The latter distinc-
the notion of the distributed nature of mental repre- tion emphasizes the secondary nature of category-
sentations. Information from both PET and fMRI specific aspects relative to modality-specific aspects of
have demonstrated that information about objects knowledge representations. This is because human-
and their features may be represented in the same made tools have mandatory somatosensory and
neural systems that are active during their perception motor representations in the brain that are absent for
(Martin, 2001). most natural objects or foods. Therefore, tools are en-
coded with reliance on more sensory dimensions
compared with most natural objects, which would
Category-Specific Knowledge make the corresponding engrams more robust.
Representation In considering the more esoteric types of catego-
Category-specific knowledge loss has also been ry-specific knowledge loss or knowledge preservation
reported (Damasio, 1990; Warrington and Shallice, (Hart, Berndt, and Caramazza, 1985; Yamadori and
1990; Hart, Berndt, and Caramazza, 1985). In the Albert, 1973; McKenna and Warrington, 1978), one
lexical domain, this pertains to the double dissocia- must also take into account the possible premorbid
tion of comprehension and naming of object names idiosyncrasies of individual lexical strengths and
308 KNOWLEDGE SYSTEMS AND MATERIAL-SPECIFIC MEMORY DEFICITS
weaknesses. This may be a potent source of artifact in Bibliography

analyzing postmorbid performance. Beauvois, M. F. (1982). Optic aphasia: A process of interaction be-
tween vision and language. Philosophical Transactions of the
Finally, combined category- and modality- Royal Society (London) B298, 3547.
specific knowledge loss has been reported in a patient Bookheimer, S. Y. (1996). Functional MRI applications in clinical
who had a selective loss of living things but not objects epilepsy. Neuroimage 4, S139146.
in the verbal but not visual domain (McCarthy and Caramazza, A., and Shelton, J. R. (1998). Domain-specific knowl-
Warrington, 1988). Elizabeth Warrington and Tim edge systems in the brain: The animate inanimate distinction.
Shallice (1984) conclude that knowledge is organized Journal of Cognitive Neuroscience 10, 134.
Cermak, L. S., and OConnor, M. (1983). The retrieval capacity of
along both sensory and category dimensions.
a patient with amnesia due to encephalitis. Neuropsychologia
Knowledge of the objects superordinate category 21, 213234.
is well preserved in modality-specific, category- Chao, L. L., Haxby, J. V., and Martin, A. (1999). Attribute-based
specific, and combined knowledge loss (Warrington, neural substrates in temporal cortex for perceiving and know-
ing about objects. Nature Neuroscience 2, 913919.
1975; McCarthy and Warrington, 1988). This perva-
Cohen, N. J., and Squire, L. R. (1980). Preserved learning and re-
sive observation has lent support to the hypothesis tention of pattern-analyzing skill in amnesia: Dissociation of
that knowledge about things is hierarchic. Research- knowing how and knowing that. Science 210, 207209.
ers have proposed that the access to a specific catego- Damasio, A. R. (1990). Category related recognition defects as a
ry member invariably begins with accessing a superor- clue to the neural substrates of knowledge. Trends in Neuro-
dinate category (Warrington, 1975). While this may sciences 13, 9598.
Devlin, J. T., Russell, R. P. Davis, R. H., Price, C. J., Moss, H. E.,
be true in some cases, the observation of the relative
Fadili, M. J., and Tyler, L. K. (2002). Is there an anatomical
preservation of superordinate knowledge does not in basis for category-specificity? Semantic memory studies in
itself necessitate this conclusion. In fact, a different PET and fMRI. Neuropsychologia 40, 5475.
route of object identification has also been proposed, Goldberg, E. (1989). Gradiental approach to the neocortical func-
from the basic category to superordinate and subordi- tional organization. Journal of Clinical and Experimental Neurop-
nate categories (Rosch, 1978). sychology 11, 489517.
(1990). Associative agnosias and the functions of the left
Researchers have evoked both degraded-store hemisphere. Journal of Clinical and Experimental Neuropsy-
(Warrington and Shallice, 1984) and impaired-access chology 12, 467484.
(Humphreys, Riddoch, and Quinlan, 1988) hypothe- Goodglass, H., Klein, B., Carey, P., and Jones, K. (1966). Specific
ses to account for category- and modality-specific semantic word categories in aphasia. Cortex 2, 7489.
memory loss. They have suggested that a degraded Goodglass, H., Wingfield, A., Hyde, M. R., and Theurkauf, J.
(1986). Category-specific dissociation in naming and recogni-
store is characterized by the uniformity of responses
tion by aphasic patients. Cortex 22, 87102.
across recall trials, and impaired access by their vari- Hart, J., Berndt, R. S., and Caramazza, A. (1985). Category-specific
ability (Warrington and Shallice, 1984; Cermak and naming deficit following cerebral infarction. Nature 316, 439
OConnor, 1983; Shallice, 1988). The possible 440.
neuroanatomical basis for this distinction may be re- Hecaen, H., and Albert, M. L. (1978). Human neuropsychology. New
lated to whether the critical lesion affects neocortical York: Wiley.
Humphreys, G. W., Riddoch, M. J., and Quinlan, P. T. (1988). Cas-
sites where representations are distributed, thus re-
cade processes in picture identification. Cognitive Neuropsy-
sulting in degraded store, or subcortical structures in- chology 5, 67103.
volved in various aspects of activation and arousal, Kinsbourne, M., and Wood, F. (1975). Short-term memory pro-
thus resulting in impaired access. cesses and the amnestic syndrome. In D. Deutsch and J. A.
Deutsch, eds., Short-term memory. New York: Academic Press.
Functional neuroimaging studies in normal sub-
Martin, A. (2001). Functional neuroimaging of semantic memory.
jects also point to the segregation of the neural sys- In R. Cabaza and A. Kingstone, eds., Functional imaging of se-
tems involved in category-specific knowledge. Investi- mantic memory. Cambridge, MA: MIT Press.
gators have shown that specific regions of the ventral McCarthy, R. A., and Warrington, E. K. (1986). Visual associative
temporal cortex respond differentially to processing agnosia: A clinico-anatomical study of a single case. Journal of
of various categories (Chao, Haxby, and Martin, Neurology, Neurosurgery and Psychiatry 49, 1,2331,240.
(1988). Evidence for modality-specific meaning systems in
1999). There have been some indications that dissoci-
the brain. Nature 334, 428430.
ations in the pattern of activation follow the distinc- McKenna, P., and Warrington, E. K. (1978). Category-specific
tion between animate and inanimate categories (Car- naming preservation: A single case study. Journal of Neurology,
amazza and Shelton, 1998); however research Neurosurgery and Psychiatry 41, 571574.
findings in the late 1990s and early 2000s have failed (1980). Testing for nominal dysphasia. Journal of Neurology,
to find evidence that these category distinctions exist Neurosurgery and Psychiatry 43, 781788.
Miceli, G., Silveri, M. C., Villa, G., and Caramazza, A. (1984). On
at the neural level (Devlin et al., 2002).
the basis for the agrammatics difficulty in producing main
verbs. Cortex 20, 207220.
See also: CONCEPTS AND CATEGORIES, LEARNING Nielsen, J. M. (1946). Agnosia, apraxia, aphasia: Their value in cerebral
OF; MODALITY EFFECTS localization, 2nd edition. New York: Hoeber.
KONORSKI, JERZY 309
Riddoch, M. J., Humphreys, G. W., Coltheart, M., and Funnell, E.

(1988). Semantic systems or system? Neuropsychological evi-
dence re-examined. Cognitive Neuropsychology 5, 325.
Rosch, E. (1978). Principles of categorization. In E. Rosch and B.
B. Lloyd, eds., Principles of categorization. Hillsdale, NJ: Erl-
baum.
Schacter, D. L. (1985). Multiple forms of memory in humans and
animals. In N. M. Weinberger, J. L. McGaugh, and G. Lynch,
eds., Memory systems of the brain. New York: Guilford Press.
(1987). Implicit memory: History and current status. Jour-
13, 501518.
Shallice, T. (1987). Impairments of semantic processing: Multiple
dissociations. In M. Coltheart, G. Santori, and R. J. Job, eds.,
The cognitive neuropsychology of language. Hillsdale, NJ: Erl-
baum.
(1988). From neuropsychology to mental structure. Cambridge,
MA: Cambridge University Press.
Tulving, E. (1983). Elements of episodic memory. Oxford: Oxford Uni-
versity Press.
Tulving, E., Kapur, S., Craik, F. I. M., Moscovitch, M., and Houle,
S. (1994). Hemispheric encoding/retrieval asymmetry in epi-
sodic memory: Positron emission tomography findings. Pro-
ceedings of the National Academy of Sciences of the United States of Jerzy Konorski (Nencki Institute)
America 91, 2,0162,020.
Tulving, E., and Schacter, D. L. (1990). Priming and human mem-
ory systems. Science 247, 301306.
Vignolo, L. A. (1982). Auditory agnosia. Philosophical Transactions of acquired behavior, especially instrumental or oper-
of the Royal Society (London) B298, 1633. ant conditioning. Along with a fellow student, Ste-
Warrington, E. K. (1975). The selective impairment of semantic phan Miller, he set up a makeshift conditioning labo-
memory. Quarterly Journal of Experimental Psychology 27, 635 ratory to investigate whether instrumental and
657.
Warrington, E. K., and McCarthy, R. A. (1983). Category specific
classical (Pavlovian) conditioning obey the same prin-
access dysphasia. Brain 106, 859878. ciples of reinforcement. On the basis of this work,
(1987). Categories of knowledge: Further fractionations Miller and Konorski (1969) published the first state-
and an attempted integration. Brain 110, 1,2731,296. ment of the distinction between the two forms of con-
Warrington, E. K., and Shallice, T. (1984). Category specific se- ditioning in 1928, while they were still students. The
mantic impairments. Brain 107, 829853.
Yamadori, A., and Albert, M. L. (1973). Word category aphasia.
existence and importance of this distinction was not
Cortex 9, 112125. realized in the West until Konorski and Miller en-
tered into a published debate with B. F. Skinner on
Elkhonon Goldberg
the matter in the next decade.
William B. Barr
Their subsequent work on instrumental condi-
tioning, conducted in Pavlovs laboratory between
1931 and 1933, and subsequently at the Nencki Insti-
KONORSKI, JERZY (19031973) tute of Experimental Biology in Warsaw, was un-
Although Jerzy Konorski always regarded himself as known in the West before World War II. Their studies
a neurophysiologist, his empirical and theoretical leg- of both modulatory and conditioned reinforcing ef-
acy has been in psychology. Like those of the Russian fects of Pavlovian stimuli on instrumental condition-
scientist Ivan Pavlov, Konorskis theories, although ing led them to a two-process theory that was not ap-
expressed in terms of speculative physiology, were proximated in the West until the 1960s. Their
largely based on behavioral experiments and are analysis of avoidance conditioning still stands.
readily recast into psychological concepts. And it is in Ignorance of Konorski and Millers work before
this form that his ideas have come to exert a preemi- World War II is understandable, for the presentation
nent influence over the contemporary study of asso- of this research in English had to await the publica-
ciative learning through conditioning. tion of Konorskis monograph Conditioned Reflexes and
Konorski was born in the Polish city of Ldz. Neuron Organization (1948). The neglect of this vol-
From his earliest student days, he was fascinated by ume is less comprehensible, however, for in it Konor-
brain function, and while studying medicine at War- ski presented the first detailed connectionist account
saw University, he came into contact with Pavlovs of conditioning within the framework of a Sherring-
work. Although inspired by Pavlovs ideas, he doubt- tonian conception of the central nervous system.
ed that Pavlovian mechanisms could explain all forms Aside from its pioneering a connectionist approach to
310 KONORSKI, JERZY
learning, two aspects of Konorskis 1948 theory de- Stalins death eventually loosened these intellec-
serve special mention. The first is his treatment of tual shackles and allowed Konorski to establish con-
conditioned inhibition. Although inhibitory processes tacts with American researchers. These contacts had
were studied intensively in Pavlovs laboratory, their a profound influence on him and culminated in the
importance was not recognized in the West until the publication of Integrative Activity of the Brain (1967).
1960s. When research on this topic finally got under The scope of this book was ambitious, attempting to
way in the West, Konorskis conception of an inhibito- describe the brain mechanisms subserving not just
ry connection was incorporated into current theories learning and conditioning but also perception, cogni-
of conditioning. tion, motivation, and emotionindeed, the overall
integration of these functions. Although replete with
The second notable feature of Konorskis theory, many interesting and perceptive ideas, this volume is
which has not yet received due recognition, is the less satisfactory than the first book. Its theoretical sub-
rules that he outlined for changing connection stance is too dependent upon the neuroscientific the-
weights. Konorski (1948, p. 106) suggested that a pos- ories and claims of the time, many of which have suf-
itive increment in an excitatory connection weight oc- fered at the hands of subsequent research. Moreover,
curs when activity in an input element is paired with its scope precluded one of the most elegant and im-
a rise in activity in the receptor element. Correspond- pressive features of the earlier book, the attempt to
ingly, inhibitory connections are strengthened when achieve a detailed concordance between theory and
input element activity is paired with a fall in the acti- data.
vation of the receptor element. The importance of
When Konorski died in 1973, Western psycholo-
these suggestions is that they provide a way of imple- gy had just begun to appreciate his legacy. His influ-
menting in connectionist terms the error-correcting ence on research in conditioning and associative
learning rules (e.g., the Rescorla-Wagner rule) that learning rivals that of Pavlov and the neobehaviorists
have come to dominate current theories of associative (see Dickinson and Boakes, 1979).
learning.
See also: CONDITIONING, CLASSICAL AND
When Konorski returned to Poland immediately INSTRUMENTAL
after World War II, he was instrumental in reestab-
lishing the Nencki Institute, first as the head of the Bibliography
Dickinson, A., and Boakes, R. A., eds. (1979). Mechanisms of learning
Department of Neurophysiology and later as its direc-
and motivation: A memorial volume to Jerzy Konorski. Hillsdale,
tor. His enthusiasm and dedication to behavioral NJ: Erlbaum.
neuroscience are clear from the recollections of some Konorski, J. (1948). Conditioned reflexes and neuron organization.
of his students, published in his memorial issue of the Cambridge, MA: Cambridge University Press.
institutes journal, Acta neurobiologiae experimentalis (1967). Integrative activity of the brain: An interdisciplinary ap-
proach. Chicago: University of Chicago Press.
(Zernicki, 1974), which contains a full bibliography. Miller, S., and Konorski, J. (1969). On a particular form of condi-
Unfortunately, his research activity was constrained at tioned reflex. Journal of the experimental analysis of behavior 12,
that time by the promulgation of Pavlovian orthodoxy 187189. English translation by B. F. Skinner from the origi-
in the later years of Russian dictator Joseph Stalins nal French publication (1928).
reign, because his 1948 monograph was regarded as Zernicki, B., ed. (1974). Acta neurobiologiae experimentalis 34 (6).
a revisionist text. Anthony Dickinson

L
LANGUAGE type of animal is something that we must learn
through experience. The totality of this knowledge
See: LANGUAGE LEARNING: HUMANS; LANGUAGE constitutes the lexicon of a language.
LEARNING: NONHUMAN PRIMATES
Pragmatics concerns the relationships among the
language, the speaker, and the speakers knowledge
of the real world. People employ knowledge of the
world to infer the communicative intentions and ex-
LANGUAGE LEARNING: HUMANS pectations of other speakers and hearers (Bach and
An understanding of language learning presupposes Harnish, 1979). For example, when someone asks,
clarification of what knowledge of language consists What are you eating? we might use our knowledge
of, what mechanisms are available for language learn- of the world to infer that they might be asking to
ing, and what the course of language development is. share what we are eating. Our language may also en-
There are two broad types of knowledge of language: code aspects of the social relationships between
grammatical knowledge, the nature of which is largely speakers, knowledge of which is acquired through ex-
biologically determined; and experiential knowledge, perience. For example, both an imperative sentence
which arises mostly from the learners encounter with (e.g., Give me a piece of pizza) and a question (e.g.,
the world (OGrady, Archibald, Aronoff, and Rees- Could you give me a piece of pizza?) can be used
Miller, 2001). to ask the hearer to do something. But it is part of the
knowledge of how English is used that the imperative
Grammatical knowledge bears on rule-governed
is much more direct and in many circumstances will
phenomena susceptible of a precise formal descrip-
be viewed as insufficiently polite.
tion. A formal description of a language is called a
grammar of the language. Typically, grammar con- Paralleling the distinction between grammatical
sists of phonology (sound patterns), morphology and experiential knowledge of language are two
(structure of words), syntax (structure of phrases and broad areas of study: development of knowledge of
sentences), and semantics (meanings of words, formal structure of language (development of gram-
phrases, and sentences). Experiential knowledge of mar) and language development (development of
language consists in part of the specific meanings and word meanings and of the use of language to commu-
pronunciations of the individual words of a language. nicate ideas and intentions and to interact socially in
Such knowledge is more or less idiosyncratic; it must other ways). This article focuses on the study of the re-
be acquired item by item and cannot be predicted by lationship between the theory of grammar of natural
general rules. For example, the fact that the word pig language and the development of grammar, since it
is pronounced a certain way and refers to a certain is this aspect of language learning that is least likely
311
312 LANGUAGE LEARNING: Humans
to prove to be merely a special case of a more general Let us turn now to an example of a universal
theory of knowledge acquisition. grammatical principle. Consider first the following
sentences:
Universal Grammar 1a.[S [NP George] [VP loves himself]].

1b. [S [NP *Georges mother] [VP loves himself]].
The most influential (and controversial) view of
the development of grammar is called universal 1c. [S [NP *Himself] [VP loves George]].
grammar, or UG (Chomsky, 1975). The UG theory If two constituents, A and B, are sisters, then A is
claims that linguistic knowledge consists of an inborn, said to c-command B and all of the constituents of B.
universal, skeletal protolanguage, the details of which The subject NP George in (1a) is a sister of VP, and
are elaborated in the course of learning. (A similar hence the NP George c-commands loves and himself.
view has been proposed for the development of cer- The c-command relation is illustrated in the tree dia-
tain birdsong systems.). On this view, innate knowl- gram in Figure 1, corresponding to the sentence (1a),
edge of language has two basic components: a catalog where the NP George c-commands the circled constit-
of the fixed, universal set of possible grammatical cat- uents. Sisters are represented as branches from the
egories along with a set of universal constraints on same node.
how these categories may combine to form phrases Sentences (1a) to (1c) show that a reflexive must
and sentences, and a fixed set of specific, universal be c-commanded by its antecedent. (This principle is
grammatical principles that determine in detail the part of what is called the binding theory.) In (1a), the
shapeliness of linguistic phrasal structures. Typically, antecedent George and the VP loves himself are sisters
each constraint and principle may allow for some very in S; thus George c-commands himself. In contrast,
limited range of variation. George does not c-command himself in (1b) or (1c).
The grammatical categories are assumed to be Again, the UG claim is that the knowledge of this
the familiar lexical categories of noun (e.g., pig), verb principle is not acquired through experience but is
(e.g., imagine), adjective (e.g., tall), and preposition part of the linguistic endowment of the language
(e.g., about), and so on, and a set of functional catego- learner.
ries whose members have formal grammatical func-
tions and limited meaning (such as that in the sen- Support for the Theory of Universal
tence, I think that it is raining) and grammatical Grammar
inflection such as tense, case, and agreement.
Why is the UG view plausible? There are several
Here is an example of a constrained phrasal reasons. Language apparently is learned almost en-
structure that allows for limited variation between tirely through exposure to examples of speech in a
languages. It is known that in natural languages there natural setting. That is, there is little or no explicit
is a privileged relationship between the verb and its language instruction, and what instruction does take
direct object such that the two form, or are constitu- place is not in general perceived as such by learners.
ents of, a verb phrase (VP) (e.g., read the book). The spoken language to which the learner is exposed
The verb and the direct object can thus be said to be is not structured in the form of organized language
sisters within VP. A native speaker of English knows lessons designed to reveal certain important proper-
that in a VP, the verb (V) normally precedes its sister, ties of the language. Hence the learner is not system-
the direct object noun phrase (NP). On the basis of atically provided with information about what is
this knowledge, the native speaker judges that *the grammatical and what is ungrammatical. In fact,
book read is not a valid VP of English. (The asterisk while much of the speech encountered by learners is
indicates that a string of words does not constitute a grammatical, much is not, yet all learners acquire the
grammatical sequence in the language.) But in a lan- correct grammar of the language or languages that
guage such as Japanese, the proper order is the re- they are exposed to. Many utterances are incomplete,
verse of that in English: [VP=NP V]. (Constituents which also constitutes a potentially confounding
within square brackets are sisters.) input to the learner (Newport, Gleitman, and Gleit-
On the UG approach to this phenomenon, lan- man, 1977). There is little if any explicit correction by
guage learners do not have to learn what the sisters adults of learners errors, especially grammatical er-
in VP may be; they already know that the sisters in VP rors (Brown and Hanlon, 1970; Wexler and Culi-
are V and NP. What learners do learn about their lan- cover, 1980). In fact, it is not clear that the learner
guage is the proper order of the sisters inside VP. would know what the correction is about, even if such
Learners of English learn that V precedes NP in VP, correction occurred (Pinker, 1989).
while learners of Japanese learn that V follows NP in In spite of all this, children make remarkably few
VP. errors in their acquisition of a grammar, and they
LANGUAGE LEARNING: Humans 313
have essentially completed the task within the first Figure 1

three years of life. The implication is that in order for
learning to be accomplished so rapidly and under
such unfavorable circumstances, much of what the
learner knows in the end must be known from the
outset. There is also experimental evidence that cer-
tain aspects of childrens knowledge of language are
in place at the earliest stages of language use, before
experience could have determined the form of this
knowledge. For example, Crain and McKee (1985)
have shown that children exhibit correct knowledge
of aspects of the binding theory as early as age two
(Crain, 1991).
Further, there is the argument from the poverty
of the stimulus, which says that much of our linguis- The c-command relation illustrated in this tree diagram,
tic knowledge could not be based on experience, since corresponding to the sentence (1a) (see text), where the NP
the crucial evidence for acquiring this knowledge George c-commands the circled constituents. Sisters are
does not exist in the discourse around us. One exam- represented as branches from the same node.
ple is the binding theory outlined above. A very dif-
ferent example involves sentences that contain rela-
tive clauses such as Finally, there are formal learnability consider-
2. John knows [NP the man [S that loves Susan]]. ations that support the UG view. Without severe con-
straints on rules, formal systems that have the essen-
When a noun phrase is questioned in English, it tial properties of natural languages are not learnable
must be located at the beginning of the sentence. under plausible definitions of what the input to the
Thus, if we replace Susan in John said that the man learner is, and of what constitutes learning (Wexler
loves Susan with a question word like who, we have and Culicover, 1980). It is posited that the learner is
a sentence like Who does John say that the man exposed to examples of utterances in context from
loves. The dash indicates the original position of some language L and must acquire the grammar of
the moved who. L, call it G(L). Only if the learner has available a high-
Strikingly, it is impossible to perform the same ly restricted set of possible hypotheses consistent with
sort of replacement on a noun phrase when the noun experience can the learners hypothesis about the
phrase is situated within a relative clause, as in grammar to be learned converge rapidly to G(L). On
this approach, universal principles such as the bind-
3. *Who does John know [NP the man [S that
ing theory and subjacency exist precisely because they
loves]].
restrict the range of hypotheses available to the learn-
The question naturally arises as to how we know er so that such rapid convergence is possible.
that this sentence is ungrammatical. The literature on
language development provides no evidence that At the same time, there is substantial evidence
there is any stage at which children learning English that not all grammatical knowledge is available to the
treat (3) as grammatical. There is no evidence that learner at the earliest stages of development. Wheth-
children produce systematic errors of the form given er this is because the learner simply lacks this knowl-
in (3), and hence no evidence that children are explic- edge and acquires it gradually (in contrast to the
itly corrected for producing such ungrammatical sen- strong UG view) or whether it is wired into the learner
tences. There is no evidence that children are in- but not fully accessible is an open question. This issue
structed as to the ungrammaticality of (3). Thus, the is often referred to as continuity (Pinker, 1984). In
claim of the UG approach is that the environment of contrast to continuity, Tomasello (2000), among oth-
the learner simply does not provide evidence, in the ers, has argued that learners first acquire specific
form of either instruction or correction, on the basis forms and sequences and generalize to rules relatively
of which the ungrammaticality of (3) could be deter- late in their linguistic development.
mined. The UG approach claims that what is going Let us turn briefly to the problem of acquiring
on here is that there is a universal, inborn principle knowledge of the lexicon. For every word, the learner
of language that prevents interrogatives (and other must learn at least the following:
types of phrases) from moving out of a relative clause,
among other structures. This principle is called subja- The meaning of the wordFor a noun (e.g., book,
cency (Chomsky, 1973). grandfather, happiness) the meaning is tied inti-
314 LANGUAGE LEARNING: Nonhuman Primates
mately to its reference; for a verb (e.g., read, eat, vast array of knowledge of the semantic and phono-
think, see) the meaning is tied to a class of events logical properties of individual words and expres-
or states of affairs (Dowty, 1979; Jackendoff, sions, and learns how to use language to convey com-
1990); plex meanings.
The category of the word (e.g., noun, verb, adjec- Bibliography
tive); Bach, K., and Harnish, R. M. (1979). Linguistic communication and
The conceptual structure of the wordWord speech acts. Cambridge, MA: MIT Press.
Bates, E., Bretherton, I., and Snyder, L. (1988). From first words to
meanings are composed of subtle and effectively grammar. Cambridge. UK: Cambridge University Press.
arbitrary components (e.g., the difference be- Brown, R., and Hanlon, C. (1970). Derivational complexity and the
tween hop and skip, between kill, murder, and assas- order of acquisition in child speech. In J. R. Hayes, ed., Cogni-
sinate, or the difference between broil and roast). tion and the development of language. New York: Wiley.
But some meaning components appear to have a Carey, S. (1977). The child as word learner. In M. Halle, J. Bres-
nan, and G. A. Miller, eds., Linguistic theory and psychological re-
privileged status in being intimately tied to the ality. Cambridge, MA: MIT Press.
structure of sentences in which they appear. Chomsky, N. (1973). Conditions on transformations. In S. Ander-
Whether the subject of a verb is a causal agent son and P. Kiparsky, eds., Festschrift for Morris Halle. New
(John in John read the book versus the bomb in York: Holt, Rinehart, and Winston.
The bomb exploded) is a more fundamental (1975). Reflections on language. New York: Pantheon.
Crain, S. R. (1991). Language acquisition in the absence of experi-
component of its meaning than is the manner of ence. Brain and Behavioral Sciences 14, 597612.
locomotion (e.g., hop and skip) (Jackendoff, Crain, S. R., and McKee, C. (1985). Acquisition of structural restric-
1990). tions on anaphora. In Proceedings of the Northeastern Linguistics
Society 16, 94110.
In acquiring the category and the conceptual De Villiers, J. G., and De Villiers, P. A. (1978). Language acquisition.
structure of a word, the learner learns to link the word Cambridge, MA: Harvard University Press.
appropriately to the syntactic structures in which the Dowty, D. R. (1979). Word meaning and Montague grammar. Dor-
drecht, The Netherlands: D. Reidel.
word is used. Experiments on lexical development Gleitman, L. R. (1990). The structural sources of verb meanings.
suggest that acquisition of the linguistic properties of Language Acquisition 1, 355.
a word proceeds very rapidly; between the ages of one Jackendoff, R. S. (1990). Semantic structures. Cambridge, MA: MIT
and a half and six years, a learner acquires about Press.
14,000 words, which works out to about nine words a Marler, P., and Sherman, V. (1983). Song structure without audito-
ry feedback: Emendations of the auditory template hypothe-
day (Carey, 1977). Learning is bootstrapped by sis. Journal of Neuroscience 3, 517531.
generalization of the conceptual structure of the word Newport, E. L., Gleitman, H., and Gleitman, L. R. (1977). Mother
with that of words with similar meanings and gram- Id rather do it myself: Some effects and non-effects of mater-
matical functions (Pinker, 1989; Gleitman, 1990). nal speech style. In C. E. Snow and C. A. Ferguson, eds., Talk-
ing to children. Cambridge, UK: Cambridge University Press.
On the other hand, to acquire the meaning of a OGrady, W., Archibald, J., Aronoff, M., and Rees-Miller, J. (2001).
word, the learner must be able to link the word with Contemporary linguistics. Boston/New York: Bedford/St. Mar-
the world. Full understanding of the word in a strict tins.
Pinker, S. (1984). Language learnability and language development.
sense thus depends crucially on the learners ability Cambridge, MA.: Harvard University Press
to interpret the fine detail of perceptual, cognitive, (1989). Learnability and cognition: The acquisition of argument
and social experience correctly. Experiments on the structure. Cambridge, MA: MIT Press.
acquisition of word meanings show that this aspect of Slobin, D. I., ed. (1985). The crosslinguistic study of language acquisi-
lexical development, not unexpectedly, is tightly tion. Hillsdale, NJ: Erlbaum.
Tomasello, M. J. (2000). Do young children have adult syntax. Cog-
yoked to perceptual, cognitive, and social develop- nition 74, 209253.
ment (Slobin, 1985; De Villiers and De Villiers, 1978; Wexler, K., and Culicover, P. W. (1980). Formal principles of lan-
Bates, Bretherton, and Snyder, 1988). guage acquisition. Cambridge, MA: MIT Press.
In summary, then, two major types of linguistic Peter W. Culicover

knowledge are acquired in the course of language
learning. The grammar of the language is acquired
through a process in which the learner starts with a
universal set of structures and constraints, and LANGUAGE LEARNING: NONHUMAN
through experience fixes aspects of the grammar that
the theory of grammar allows to vary from language
PRIMATES
to language. There is considerable debate as to how Guided by modern evolutionary theory, scientists
much knowledge of language the learner begins with have been able to explicate the origin of the morphol-
and how much is acquired through experience. In ad- ogy of the human body. However, the evolution of the
dition to linguistic structure, the learner acquires a human mind has yet to be so satisfactorily accounted
LANGUAGE LEARNING: Nonhuman Primates 315
for. Comparative psychology has already succeeded and music. A cucumber was innovatively called ba-
in providing evidence to help clarify the biobehavior- nana which-is green, an overly ripe banana was
al origins of human language and symbolic compe- called banana which-is black, and an orange was
tence through studies of our nearest living relatives called apple which-is orange (colored). Compre-
the Pongidae (chimpanzee, Pan; orangutan, Pongo; hensive analyses clearly indicated that Lanas produc-
and gorilla, Gorilla). These great apes are substantial- tions could not be satisfactorily attributed either to
ly more like humans than are the lesser apes, mon- rotely learned sequences or to imitation. Project
keys, or any other mammals. The similarities between LANA contributed what was perhaps the first success-
human and chimpanzee DNA exceed 98 percent. Ge- ful application of a computer to a system for commu-
netic relatedness enhances the probability that life nicative research. Romski, Sevcik, and Pate (1988)
forms will have similar psychology as well as appear- have demonstrated that keyboards similar to the one
ance. Consequently, we might reasonably expect to
developed for LANA augment the communicative ef-
find linguistic and cognitive competencies in the ape
fectiveness of children with language deficits. Project
that are similar to those observed in humans.
LANA also affirmed the ability of the chimpanzee to
learn large numbers of symbols, to use them in pre-
Early Studies scribed sequences, to alter use of those sequences cre-
Speculation that apes might be capable of lan- atively, and to use symbols so as to facilitate percep-
guage began in the nineteenth century. Early in the tions of sameness and difference between items when
twentieth century, occasional studies were un- one was presented visually and the other by touch.
dertaken to determine whether or not this might be
true, but none succeeded. Efforts were renewed in the
mid-1960s, with studies by the Gardners (Gardner Other Language-Learning Projects with
and Gardner, 1969; Gardner, Gardner, and van Chimpanzees
Cantfort, 1989) and by Premack (1971, 1976). The Terrace (1979) began Project Nim in the early
Gardners used American Sign Language, and Pre- 1970s. The chimpanzee Nim learned signs, but Ter-
mack used plastic tokens of different shapes and col- race concluded that Nims signsand also those of
ors to function as words, with the aim of establishing
the Gardners Washoewere predominantly partial
two-way communication and an experimental analy-
or complete imitations of the human researchers
sis of language functions, respectively. Their chim-
working with these apeshad recently signed. (As
panzees, Washoe and Sarah, learned relatively large
noted, imitation was not an issue in the LANA Project
numbers of signs and tokens and, seemingly, their ap-
propriate use. because the project was carried out using a computer
to monitor and record Lanas lexigram utterances.)
Project LANA Also in the 1970s, Miles (1990) began Project

Chantek. Chantek, an orangutan, was taught manual
In 1971, the LANA Project was initiated by Rum-
signs; Miles concluded that Chanteks signing was not
baugh and his colleagues. A computer-based research
confounded by or due to his imitation of human sign-
system provided Lana, a chimpanzee, with a keyboard
that held 125 keys, each embossed with a distinctive ers, as Terrace had claimed to be the case with Nim
geometric symbol called a lexigram (see Figure 1). and Washoe.
Lexigrams functioned as words, and their concatena-
Research of the mid- and late 1970s by Savage-
tion was computer-monitored for correctness. The
Rumbaugh and Rumbaugh with two chimpanzees
computer could activate a bank of devices, some of
named Sherman and Austin used a variation of Lanas
which delivered various foods and drinks, movies,
slides, and music. Lana readily learned her lexigrams keyboard. These chimpanzees demonstrated their
in order to request things and to give the names and abilities to classify lexigrams, each of which represent-
colors of objects. Tests revealed that she saw Munsel ed a specific food or tool, through the use of two other
color chips in a manner that resembles the way hu- lexigrams, glossed as food and tool, embedded among
mans see them, an observation reaffirmed by Mat- more than one hundred other keys on their keyboard.
suzawa (1985) in Japan with the chimpanzee Ai. This ability to categorize lexigrams was strong evi-
dence that each symbol did, in fact, represent an item
Within limits, Lana demonstrated her ability to
to the apes and that they could use these representa-
modify her sentences so as to achieve ends other than
the specific ones for which they were designated, such tions, on trial-1 tests in controlled conditions, to clas-
as using them to attract attention to malfunctioning sify almost without error the symbols of their working
food vendors and to the devices that produced slides vocabulary.
316 LANGUAGE LEARNING: Nonhuman Primates
Studies with Bonobos does the bonobo (Brakke and Savage-Rumbaugh,

1995a; Brakke and Savage-Rumbaugh, 1995b).
Research by Savage-Rumbaugh with the bonobo
(Pan paniscus), a species apart from the common One bonobo, Kanzi, responded appropriately to
chimpanzee (P. troglodytes), was the first to demon- about three-fourths of more than seven hundred
strate the chimpanzees ability to learn the meanings novel requests presented to him verbally under con-
of symbols spontaneously. The bonobo came to com- trolled test conditions to preclude cuing. His perfor-
prehend the meanings of lexigrams, and even human mance was generally comparable with that of a two-
speech, prior to developing competence in use of the year-old girl whose mental age was two and a half
keyboard. All previous apes had been able to talk years. Neither Kanzi nor the girl had the benefit of
(e.g., produce signs, use tokens, use lexigrams) as a people modeling the requests, nor had they been
result of specific training programs, but their com- trained to do what was requested of them. The con-
prehension skills were either deferred or never as- clusion is, then, that they comprehended the syntax
sessed. For the first time, a model of language learn- of the novel requests conveyed by normal human
ing in the ape tracked or paralleled the course of speech.
language acquisition in the normal child. Early expe- Kanzi not only comprehends single spoken utter-
rience from birth has been critical for such observa- ances, he also participates in linguistically mediated
tional learning of language to occur in the bonobo. communicative interactions with humans in which
The common chimpanzee also benefits similarly from both Kanzi and his human counterparts employ a va-
such rearing, though possibly to a lesser degree than riety of communicative modalities concomitantly
LASHLEY, KARL 317
(e.g., lexigram use, gesture, vocalization, speech).

These exchanges share characteristics with human
conversations, as Kanzi participates in turn taking, re-
sponds to commands, and makes requests of his own.
Taglialatela and Savage-Rumbaugh (2001) have re-
ported that during these conversations, Kanzi pro-
duces acoustically distinct vocal utterances that vary
with the semantic context in which they are produced.
Conclusion
Language competence rests fundamentally in an
individuals ability to comprehend, process, and pro-
duce meaningful utterances. Chimpanzees can come
to acquire and use human language if raised in an en-
vironment that permits the emergence of these abili-
ties. It is no longer reasonable to insist that the poten-
tial for language is unique to the human speciesand
reason, according to Descartes, should be humans
strongest suit.
See also: COMPARATIVE COGNITION; LANGUAGE

LEARNING: HUMANS
Bibliography
Brakke, K. E., and Savage-Rumbaugh, E. S. (1995a). The develop-
ment of language skills in bonobo and chimpanzee, Part 1: Karl Lashley (UPI/Corbis-Bettman)
Comprehension. Language and Communication 15, 121148.
(1995b). The development of language skills in bonobo and
chimpanzee, Part 2: Production. Language and Communication
16, 361380. Taglialatela, J. P., and Savage-Rumbaugh, S. (2001). Bonobo cog-
Gardner, B. T., and Gardner, R. A. (1969). Teaching sign language nition: Expectations, explications, and conversations. Paper
to a chimpanzee. Science 162, 664672. presented at the 109th convention of the American Psycho-
Gardner, R. A., Gardner, B. T., and van Cantfort, T. E. (1989). logical Association. San Francisco, CA.
Teaching sign language to chimpanzees. Albany: State University Terrace, H. S. (1979). Nim. New York: Knopf.
of New York Press. Duane M. Rumbaugh
Greenfield, P. M., and Savage-Rumbaugh, E. S. (1990). Grammati-
E. Sue Savage-Rumbaugh
cal combination in Pan paniscus: Processes of learning and in-
vention in the evolution and development of language. In S. Revised by Duane M. Rumbaugh, E. Sue Savage-Rumbaugh,
T. Parker and K. R. Gibson, eds., Language and intelligence and Jared P. Taglialatela
in monkeys and apes: Comparative developmental perspectives, pp.
540578. New York: Cambridge University Press.
Matsuzawa, T. (1985). Color naming and classification in a chim-
panzee (Pan troglodytes). Journal of Human Evolution 14, 283
291.
Miles, H. L. W. (1990). The cognitive foundations for reference in LASHLEY, KARL (18901958)
a signing orangutan. In S. T. Parker and K. R. Gibson, eds., Karl Spencer Lashley pioneered the study of brain
Language and intelligence in monkeys and apes: Comparative de-
velopmental perspectives, pp. 511539. New York: Cambridge mechanisms of learning and memory. He was born in
University Press. 1890 in Davis, West Virginia, and entered the Univer-
Premack, D. (1971). On the assessment of language competence in sity of West Virginia at the age of fifteen. As a fresh-
the chimpanzee. In A. M. Schrier and F. Stollnitz, eds., Behav- man he signed up for a class in zoology under the dis-
ior of nonhuman primates, Vol. 4, 186228. New York: Academ- tinguished neurologist John Black Johnston, and
ic Press.
(1976). Language and intelligence in ape and man. Hillsdale, within a few weeks he knew that I had found my life
NJ: Erlbaum. work. After graduation in 1910, he obtained a teach-
Romski, M. A., Sevcik, R. A., and Pate, J. L. (1988). The establishing fellowship at the University of Pittsburgh and re-
ment of symbolic communication in persons with mental received his M.S. degree there. Lashley then went to
tardation. Journal of Speech and Hearing Disorders 53, 94107. Johns Hopkins University to study for his doctorate
Rumbaugh, D. M. (1977). Language learning by a chimpanzee: The
Lana Project. New York: Academic Press. in zoology under Herbert S. Jennings. He elected a
Savage-Rumbaugh, E. S. (1986). Ape language: From conditioned re- minor in psychology with John B. Watson, the found-
sponse to symbol. NewYork: Columbia University Press. er of behaviorism. His work with Watson convinced
318 LASHLEY, KARL
him to make his career in psychology. This was the from the work with Franz. During this period, Lash-
critical time in Lashleys development as a scientist. leys theoretical view of learning was heavily influ-
In his own words: enced by two widely held notions: localization of func-
tion in neurology and behaviorism in psychology.
In 1914, John Watson called attention to a
seminar in the French edition of Bechterevs Localization of function, the notion that each psy-
book on the conditioned reflex. In that win- chological trait or function has a specific locus of
ter the seminar was devoted to the translation representation, a particular place, in the brain, was
and discussion of the book. In the spring I perhaps the major intellectual issue concerning brain
served as an unpaid assistant and we con- organization at the beginning of the twentieth centu-
structed apparatus and did experiments, re- ry. An extreme form of the idea of localization, known
peating a number of Pavlovs experiments. as phrenology, was popular early in the nineteenth cen-
Our whole program was then disrupted by tury. The field of neurology then moved away from
the move to the lab in Meyers Clinic. There that position, though the discovery of a speech center
were no adequate animal quarters there. Wat- began to move the pendulum back. Work in the last
son started work with infants as the next best three decades of the nineteenth century identified the
material available. I tagged along for awhile general locations of the motor, visual, and auditory
but disliked the babies and found me a rat lab regions of the cerebral cortex. Localization of func-
in another building. We accumulated a con- tion appeared to be winning the day.
siderable amount of experimental material In Watsons behaviorism, the learning of a partic-
on the conditioned reflex that was never pub- ular response was held to involve the formation of a
lished. Watson sought the basis of a systemat- particular set of connections, a series set. Consequent-
ic psychology and was not greatly concerned ly, Lashley argued, it should be possible to localize the
with the reaction itself (letter from K. S. Lash- place in the cerebral cortex where that learned
ley to Ernest Hilgard, ca. 1930; reprinted change in brain organization, the engram (memory
with the permission of Professor Hilgard). trace), was stored. It was generally believed at the
The conditioned reflex formed the basis of Wat- time, consistent with Ivan Pavlovs view, that learning
sons behaviorism. Lashley, on the other hand, be- was coded in the cerebral cortex. Thus, behaviorism
came interested in the physiology of the reaction and and localization of function were consistent; they sup-
the attempt to trace conditioned reflex paths through ported the notion of an elaborate and complex
the central nervous system. Over the next several switchboard on which specific and localized changes
years Lashley collaborated with Shepherd Franz, who occurred in the cerebral cortex when specific habits
was then at St. Elizabeths Hospital in Washington, were learned.
DC. Their research, examining effects of lesions of Lashley systematically set about finding the
the frontal cortex on learning abilities in the rat, was places in the cerebral cortex where learning oc-
the foundation of his major work, Brain Mechanisms curredthe engramsin an extensive series of
and Intelligence (1929). In 1920 Lashley accepted an studies culminating in his 1929 monograph. In this
assistant professorship at the University of Minnesota research he used mazes of differing difficulty and
and began in earnest his search for the memory trace. made lesions of varying sizes in different regions of
He was made full professor at Minnesota in 1924, and the cerebral cortex of the rat. The results profoundly
in 1926 moved to the University of Chicago, where he altered Lashleys view of brain organization and had
became a professor in 1929. In this same year he was an extraordinary impact on the young field of physio-
president of the American Psychological Association logical psychology: the locus of the lesion was unim-
and published his monograph Brain Mechanisms and portant; the size was critically important, particularly
Intelligence. In 1935 Lashley accepted a professorship for the difficult mazes. These findings led to Lashleys
at Harvard University, and in 1937 he was appointed notions of equipotentiality (locus not important) and
research professor of neuropsychology with nominal mass action (size critical).
teaching duties, which made it possible for him to ac- Subsequently, Lashley focused on a detailed anal-
cept the directorship of the Yerkes Laboratories of ysis of the role of the cerebral cortex in vision and in
Primate Biology in Orange Park, Florida, in 1942. He visual discrimination learning and memory. But his
held both these positions, spending only a few weeks interests and research also included classic work on
a year at Harvard (he was not fond of formal teach- the cytoarchitectonics (microscopic structure) of the
ing), until his death in 1958. cerebral cortex; a brilliant analysis of the problem of
Lashley devoted many years to an analysis of serial order in human language and thought; and a
brain mechanisms of learning, using the lesion- penetrating analysis of the biological substrates of
behavior method that he developed and elaborated motivation. Lashley, more than any other scientist,
LEARNED HELPLESSNESS 319
shaped and developed the field of physiological psy- paws, terminates each of the shocks. This rat thus has
chology. control over the termination of each shock. A second
It is fitting to close this biographical entry with rat is placed in a similar apparatus, but this rat does
Lashleys own summing up of his search for the mem- not have control. Each shock begins for this rat at the
ory trace, written in 1950: same instant as it does for the rat with control, but for
this second rat pushing the lever has no consequence.
This series of experiments has yielded a good Each shock terminates whenever the rat with control
bit of information about what and where the presses the lever. Thus both rats receive identical
memory trace is not. It has discovered noth- shocks, but one has behavioral control and the other
ing directly of the real nature of the memory does not. A third rat is merely placed in the apparatus
trace. I sometimes feel, in reviewing the evi- and receives no shock. Subsequent behavior and
dence of the localization of the memory trace, physiological functioning can be examined, and it is
that the necessary conclusion is that learning possible to determine which changes are caused by
is just not possible. It is difficult to conceive the stressor per se (here the animals with and without
of a mechanism that can satisfy the conditions control would be identical and differ from the non-
set for it. Nevertheless, in spite of such evi- shocked controls), and which are a function of the
dence against it, learning sometimes does controllability of the stressor (here the animals with
occur. and without control would differ).
See also: LOCALIZATION OF MEMORY TRACES Use of this sort of experimental design has re-
vealed that many of the consequences that are nor-
Bibliography
mally thought to be produced by stressors actually are
Lashley, K. S. (1929). Brain mechanisms and intelligence: A quantitative
study of injuries to the brain. Chicago: University of Chicago
determined by the controllability/uncontrollability of
Press. the stressor rather than by mere exposure to the stres-
(1935). The mechanism of vision, Part 12: Nervous struc- sor. Three kinds of behavioral changes follow expo-
tures concerned in the acquisition and retention of habits sure to stressors, but only if they are uncontrollable.
based on reactions to light. Comparative Psychology Monographs
11, 4379. The first type is cognitive changes. A particularly
(1950). In search of the engram. Society of Experimental Biolo- important consequence of exposure to uncontrollable
gy, Symposium 4, 454482.
aversive events has come to be called the learned help-
Orbach, J., ed. (1982). Neuropsychology after Lashley. Hillsdale, NJ:
Erlbaum. lessness effect. This refers to the fact that organisms
ranging from fish to humans fail to learn to escape
Richard F. Thompson
and avoid aversive events such as electric shocks, loud
noises, and cold water after an initial exposure to
aversive events that are uncontrollable (Overmier and
Seligman, 1967). A great deal of research has been
LEARNED HELPLESSNESS conducted to determine why this learning deficit oc-
curs, and it has been found that at least part of the
The term learned helplessness is used to refer to any be- reason is cognitive. Uncontrollable aversive events in-
havioral or physiological consequence of exposure to terfere with some of the information-processing steps
an aversive event that is produced not by the event it- required to learn relationships between behavior and
self but by the organisms lack of behavioral control over outcomes (Maier, 1989).
the event. By behavioral control is meant the organisms
ability to alter the onset, termination, duration, inten- The second type is motivational changes. The moti-
sity, or temporal pattern of the event. If the event vation to obtain many of the reinforcers that are nor-
(e.g., a loud noise, a painful electric shock, an attack mally important for that organism is undermined.
by another animal or person) can be altered by some For example, a rat that is exposed to uncontrollable
behavioral response, then the organism has some shock (but not to equal amounts of controllable
control over the event. If there is nothing that the or- shock) does not later compete for food, becomes inac-
ganism can do to change the event, then the event is tive, and shows decreased sexual and maternal behav-
uncontrollable. ior.
This concept has been studied using an experi- The third type is emotional changes. Uncontrolla-
mental paradigm called the triadic design. Here one ble aversive events lead to increases in aspects of emo-
subject, say a rat, is given control over the event, say tionality such as fear and anxiety. Physiological indi-
a mild electric shock, delivered to the rats tail. The cants of stress such as ulcer formation and blood
rat is exposed to a number of shocks, and performing pressure increases are similarly influenced by the con-
some behavioral response, say pushing a lever with its trollability/uncontrollability dimension.
320 LEARNING CURVE
A considerable amount of research has been de- guage impairment, or SLI), even before entering
voted to uncovering the behavioral and physiological school, but typically a diagnosis emerges in the con-
mechanisms that produce these learned helplessness text of an academic setting, for example when a child
effects. This fact is at least in part attributable to the has unexpected difficulties with reading (called spe-
resemblance between the consequences of uncontrol- cific reading impairment, or dyslexia). There is sig-
lable stressors in animals and human depression. In- nificant overlap between preschool SLI and school-
deed, what is now known about the substrates of age dyslexic populations (Stark and Tallal, 1988).
learned helplessness is remarkably similar to what is
thought to underlie human depression (Weiss and
Simson, 1986). Learning Disabilities from a Behavioral
Perspective
Bibliography The long-standing assumption that learning dis-
Maier, S. F. (1989). Learned helplessness: Event co-variation and
abilities reflect subtle anomalies in the brain is reflect-
cognitive changes. In S. B. Klein and R. R. Mowrer, eds., Con-
temporary learning theories, pp. 73109. Hillsdale, NJ: Erlbaum. ed in early terms such as minimal brain damage and
Overmier, J. B., and Seligman, M. E. P. (1967). Effects of inescap- minimal brain dysfunction. Although an emergent body
able shock upon subsequent escape and avoidance behavior. of research has shown that learning-disabled popula-
Journal of Comparative Physiology and Psychology 63, 2333. tions exhibit neural anomalies, the field remains at a
Weiss, J. M., and Simson, P. G. (1986). Depression in an animal
point where behavioral data are used to make an eval-
model: Focus on the locus coeruleus. In R. Porter, G. Bock,
and S. Clark, eds., Antidepressants and receptor function, pp. uation or clinical diagnosis, and neural features of
191216. New York: Wiley. these behaviorally identified populations are then
studied to discern differences from normal controls.
Steven F. Maier
A diagnosis of specific reading impairment (or dys-
lexia) may someday be confirmed or even routinely
screened (like vision testing) via neuroimaging.
LEARNING CURVE Moreover, neurological diagnosis of learning disor-
ders may alter their very definitions (e.g., it may be-
See: REPETITION AND LEARNING come possible to diagnose neural anomalies associat-
ed with dyslexia prior to behavioral expression of the
disorder, possibly even in infancy; see Leppnen and
Lyytinen, 1997; Leppnen, Pihko, Eklund, and Lyy-
LEARNING DISABILITIES tinen, 1999; Molfese and Molfese, 1997; Pihko et al.,
An individual is said to have a learning disability if he 1999). This dependence on behavioral criteria is re-
or she has difficulty using or understanding written flected in debate over whether traditional health in-
or spoken language (including academic subjects surance should cover treatment for learning disor-
such as reading, writing, and mathematics) in a way ders (Tallal, 1988).
that is inconsistent with overall intelligence quotient Diagnostic dependence on behavioral criteria is
(IQ). Individuals with low IQ in all areas (for example, associated with other intrinsic pitfalls, including those
in cases of mental retardation) are excluded from the inherent to the definition of abnormal behavior. This
diagnosis. Environmental factors such as deprivation pivotal issue forms the focus of controversy surround-
are also excluded (because learning disabilities are ing learning disabilities (e.g., assertions of gender
defined as congenital rather than acquired prob- and ethnic bias in school curricula, and over-referral
lems), as are physical disabilities that might cause sec- of boys for special education due to boys more ac-
ondary learning problems (for example, hearing and tive behavior). One objective way that behavior can
vision deficits, epilepsy, or attention deficit and hyp- be quantified is by comparison to a norm derived
eractivity disorder [ADHD]). Although learning dis- from sample population data, allowing clinicians to
abilities can be concurrent with other physical disabil- use a standard curve to set cutoff criteria to define
ities (i.e., blind or deaf children can also be learning when an individual is significantly impaired relative
disabled), it is generally established that the physical to their peers for a specific function. Academic per-
disability is not inherently the cause of the learning formance in a school setting may also be compared to
disability. ADHD has a high concurrence with learn- age-appropriate norms in a more general way by
ing disabilities, but clinicians usually try to establish schoolteachers who refer potentially learning dis-
that attention deficits (not typically considered under abled children for further testing. Learning disabili-
the scope of learning disabilities) are not the direct ties are also defined by comparison of a childs own
cause of the learning problems. A child who exhibits scores or performance in a specific area relative to
difficulties with language may be diagnosed with a their overall intelligence, entailing use of a discrep-
learning disability (for example, with specific lan- ancy criteria.
LEARNING DISABILITIES 321
Even with the use of such criteria, clinicians and 1. Heterogeneity of subject criteria. If the popula-
educators may disagree over the definition of what tion under study is not homogeneous, biological
constitutes disabledfor example, whether a child data from this sample will be difficult to interpret.
is simply part of the lower tail of a normal curve (e.g., Also, when different researchers use different be-
is a poor reader) or whether that child functions dis- havioral criteria, cross-study comparisons will
tinctly differently from the norm and thus has a clini- show inconsistencies.
cally defined disorder (e.g., specific reading impair- 2. Heterogeneity of disorders. Even where careful
ment or dyslexia; see Shaywitz et al., 1992). attention to consistent behavioral criteria is ap-
Moreover, individual cases are sometimes difficult to plied, diverse underlying etiologies can produce
capture within the parameters of standard behavioral the same behavioral profile. Conversely, behav-
definitions. For example, if a child demonstrates gen- ioral profiles (even within a relatively specific sub-
erally superior academic performance but performs type of a disorder) can show large individual dif-
only at an average level in one specific area such as ferences.
reading, is he or she reading impaired? Such a child
is unlikely to be referred, at any rate, for special edu- 3. Difficulties inherent to the study of children. Al-
cation services. Similarly, if a child demonstrates gen- though modern magnetic resonance imaging
eral intelligence well below the mean, but performs (MRI) techniques are considered noninvasive,
even worse in one specific area (e.g., reading) than they are stressful and time consuming, and par-
would be predicted by nonverbal IQ, is this child ents will not always consent to participation of
mentally retarded, or reading impaired, or both? their affected children. Although studies can be
Confusion surrounding such unresolved issues is fu- conducted using adults affected since childhood,
eled by educational dictates that define the aid and the data obtained from such studies may not ac-
curricula a child may receive depending on clinical curately reflect anomalies that characterize early
diagnosis, and also fueled by social controversies sur- disruption of brain development. Moreover, re-
rounding the labeling of children. (These issues are trospective diagnosis of childhood disorders fre-
not addressed in this entry, but see Wong, 1986, or quently relies on memory (e.g., a patient is asked,
Woody, 1989, for review.) Finally, the definition of Did you have difficulty reading as a young
learning disabilities is weighted to some degree by a child?) and hence can be unreliable.
standard academic curricula that does not tend to
Despite these hindrances, research has uncov-
focus on deficiencies in art, music, or sports but, rath-
ered important neurological and genetic features that
er, only in reading, writing, and math.
seem to be associated with specific developmental
The populations of children who receive special learning disorders. Such studies have shown that
education services at school are distinct but overlap- learning disabled children do not exhibit focal dam-
ping: Some children are in SE due to lagging academ- age to specific regions of the brain, with all other sys-
ic performance, whereas others in SE classes have tems and functions spared intact. Although localized
been clinically diagnosed as learning disabled. SE en- damage can indeed occur via localized hemorrhage,
compasses a heterogeneous population of students trauma, or tumor (and can lead to subsequent learn-
exhibiting general cognitive delays and deficits, so- ing disabilities) these types of injuries comprise ac-
cial-environmental deprivation, psychological prob- quired rather than congenital disabilities. For exam-
lems, and concurrent disabilities such as ADHD. This ple, language deficits caused by an acquired lesion of
entry focuses on children identified as learning dis- the left hemisphere are classified as acquired aphasia,
abled through relatively rigorous clinical diagnosis rather than developmental language impairment
that is, language, reading, or mathematically im- (Tallal, 1987), which is considered congenital. This
paired children performing below levels predicted by distinction relies on the notion of inherent versus
their overall intelligence in one of these specific aca- imposed neural damage (although most diagnoses
demic areas. accept brain damage of prematurity within the scope
of congenital disabilities). Research suggests that
some learning disabled children may, in fact, exhibit
Neurobiological Correlates of Learning evidence of relatively focal (i.e., narrowly circum-
Disabilities scribed) brain damage of unknown cause from very
Advances in neuroimaging and gene linkage early (i.e., prenatal) periods of neurodevelopment
technology have expanded research possibilities for (e.g., see Galaburda, 1992; Galaburda et al., 1985).
noninvasive study of clinical populations, but charac- This early focal injuryparticularly if it occurs during
teristics of various learning disabilities have yet to be critical neural events such as neuromigration (cortical
isolated and defined. Reasons for these limitations in- neurons migrating to cortical layer locations)may
clude: disrupt neurodevelopment, with deleterious behav-
322 LEARNING DISABILITIES
ioral consequences. Unlike damage to a relatively measurements taken from the scalp, the electrophy-
static adult brain, focal damage that occurs during the siologic response to visual stimuli designed to activate
formation of neural circuitry exerts widely distributed the magnocellular subsystem was found to be signifi-
effects on brain organization (see Aicardi, 1994; Gala- cantly slower in dyslexic as compared to control sub-
burda, 1992; Kolb and Fantie, 1989). Some children jects. Consistent with this functional data, postmor-
with learning disabilities appear to have experienced tem studies of several dyslexic brains demonstrated
such subtle injuries to the brain or interference with that the magnocellular neurons of the lateral genicu-
development of the brain during critical periods of late nucleus (the visual thalamic nucleus, which relays
growth, resulting in pervasive reorganization of brain visual information to the cortex) were smaller in dys-
systems. lexic than in control brains. The authors speculate
that electrophysiological and anatomical evidence of
a visual processing disturbance could reflect interfer-
Specific Language Impairment (SLI) and ence with normal reading (see also Lovegrove, Garz-
Dyslexia ia, and Nicholson, 1990; Slaghuis, Lovegrove, and
SLI and dyslexia are the two most frequently Freestun, 1992). A subsequent postmortem study of
studied forms of learning disability. Longitudinal the same dyslexic brains studied by Livingstone, Gala-
studies suggest that a large majority of children with burda, Menard, and Rosen (1994) found that there
SLI overcome the most noticeable aspects of language were more small neurons and fewer large neurons in
delays of early childhood (e.g., they do eventually the left medial geniculate nucleus (MGN) of dyslexics
learn to talk) but that underlying components of their as compared to controls. This finding may relate to
disorder appear to be expressed more visibly in other the behavioral and electrophysiological evidence of
areas, for example as the child attempts to use phono- disrupted auditory processing in both SLI and dyslex-
logical systems to learn to read (see Leonard, 1998, ic individuals, which may in turn give rise to the pho-
for review). Consequently, studies on the neurophy- nological processing deficits observed in these pa-
siology of SLI (which typically comprises a childhood tients (see Tallal, Miller, and Fitch,, 1993, for review).
population) are difficult to separate from studies of These findings may further relate to evidence of
dyslexia (which frequently is studied in adults). neuropathologic lesions in dyslexic brains (Benton,
Some functional deficits in children with SLI may 1964; Cohen, Campbell, and Yaghamai, 1989; Gala-
arise from basic sensory integration deficits, reflect- burda and Kemper, 1979; Galaburda et al., 1985;
ing in turn dysfunctional neural processing of rapidly Humphreys et al., 1991; reviewed in Galaburda,
changing sensory information. For example, research 1992). Specifically, researchers have used postmor-
has shown that language-impaired children exhibit tem analysis to show evidence of cerebrocortical
severe deficits in the ability to perform auditory dis- microdysgenesis consisting of neuronal ectopias in
criminations of information that changes rapidly in neocortical layer I, subjacent laminar dysplasia, focal
time. Although this deficit is profoundly evident when microgyria, and microvascular anomalies. These ab-
children with SLI are asked to discriminate speech normalities tend to be located in prefrontal and peri-
stimuli characterized by brief and rapidly changing sylvian regions, and they usually involve the left more
acoustic spectra (e.g., consonant-vowel syllables), def- than the right hemisphere. Evidence from animal
icits are also observed on nonlingual tasks such as models suggests that focal cortical damage during the
tone-sequence discrimination (Tallal and Piercy, period of cortical neuromigration induces the forma-
1973). Such rapid auditory processing deficits may tion of cortical cellular anomalies, as well as structural
disrupt phonological processing and speech percep- anomalies at the thalamic level, much like those seen
tion, leading to developmental impairment of lan- in dyslexic brains. Moreover, these anomalies are as-
guage skills acquisition (Tallal, Miller, and Fitch, sociated with auditory processing deficits strikingly
1993). Although a neurophysiological basis for senso- similar those seen in language-impaired children
ry processing deficits in language-impaired children (Clark, Rosen, Tallal, and Fitch, 2000a, 2000b; Fitch
has not been identified, evidence has shown cellular et al., 1994; Fitch, Brown, Tallal, and Rosen, 1997;
anomalies in the visual and auditory divisions of the Herman et al., 1997).
thalamus (lateral geniculate and medial geniculate
On a more global level, studies of both language
nucleus; LGN and MGN) of dyslexic brains.
impairment and dyslexia have focused heavily on re-
Specifically, Livingstone, Rosen, Drislane, and gions of the brain known to be involved in language
Galaburda (1991) found that the magnocellular com- processing, particularly left-hemisphere regions of
ponent of the visual system, which is responsible for temporal, parietal, and frontal cortex such as Wern-
processing fast, low-contrast information, appears to ickes area and Brocas area. In most normal subjects,
be impaired in dyslexics. Using evoked potential the planum temporale (which lies on the superior sur-
LEARNING DISABILITIES 323
face of the temporal lobe and encompasses a portion pon, 1994; Byring and Jrvilehto, 1985; Dawson, Fin-
of Wernickes area) is larger on the left as compared ley, Phillips, and Lewy, 1989; Jirsa and Clontz, 1990;
to the right side of the brain (Geschwind and Levitsky, Kraus et al., 1996; Lincoln, Courchesne, Harms, and
1968; Kulynch, Vladar, Jones, and Weinberger, 1994; Allen, 1995; Neville, Coffey, Holcomb, and Talall,
Teszner, Tzavaras, Gruner, and Hcaen, 1972; Wada, 1993; Pinkerton, Watson, and McClelland, 1989;
Clarke, and Hamm, 1975; Witelson and Pallie, 1973). Tonnquist-Uhlen, Borg, Persson, and Spens, 1996;
Such findings are consistent with left-hemisphere Wood, Flowers, Buchsbaum, and Tallal, 1991), sup-
specialization for language processing found in be- porting the existence of an auditory processing deficit
havioral studies, neuroimaging studies, and studies of in these groups as discussed above.
the behavioral effects of lateralized lesions. In one of
the few studies of neuropathology underlying specific Conclusion
language impairment, Jernigan, Hesselink, Sowell,
and Tallal (1991) used MRI techniques to examine a Learning disabilities are characterized by unex-
sample of SLI children and showed that the volume pected difficulties with language, reading, or math
of the posterior perisylvian region (which includes the that are inconsistent with IQ,or ability in other areas,
planum temporale) was reduced bilaterally in lan- or both. These problems remain after other known
guage-impaired children (markedly so in the left causes for learning problems have been excluded.
hemisphere; see also Plante, Swisher, Vance, and The neurobiological basis for these disorders remains
Rapcsak, 1991). Asymmetries in the inferoanterior elusive but appears to involve very early neurode-
and superoposterior cerebral regions showed signifi- velopmental disruption leading to distributed pro-
cant differences between control and impaired chil- cessing deficits in the brain. These disruptions appear
dren. Neurological studies of dyslexia have also con- to leave some higher processing systems intact while
sistently reported anomalous asymmetry of the causing severe deficits in other systems (e.g., rapid
planum temporale (Dalby, Elbro, and Stodkilde- sensory processing). Nevertheless, learning disabili-
Jorgensen, 1998; Galaburda et al., 1985; Haslam, ties encompass a wide range of subtypes of disorders,
Dalby, Johns, and Rademaker, 1981; Hier, LeMay, with large individual variability in behavioral expres-
Rosenberger, and Perlo, 1978; Humphreys et al., sion, and they appear to reflect a heterogeneity of
1991; Hynd and Semrud-Clikeman, 1989; Hynd et neural anomalies of varied origin, ranging from ge-
al., 1990; Hynde, Marshall, and Semrud-Clikeman, netic factors to early focal injury. Improved neuroi-
1991; Jancke, Schlaug, Huang, and Steinmetz, 1994; maging technology should allow scientists to create
Kushch et al., 1993; Larsen, Hoien, Lundberg, and more precise neurobehavioral templates of disrup-
Odegaard, 1990; Leonard et al., 1993; Rosenberger tion characterizing various subtypes of learning dis-
and Hier, 1980; Rumsey et al., 1986; Schultz et al., ability, providing in turn more reliable and consistent
1994). These anatomical anomalies are consistent diagnostic tools for the assessment and remediation
with evidence of activational differences in the cortex of learning disabilities.
of dyslexics during the performance of reading- Bibliography
related tasks (e.g., Demb, Boynton, and Heeger, Aicardi, J. (1994). The place of neuronal migration abnormalities
1998; Shaywitz et al., 1998). Atypical asymmetries in child neurology. Canadian Journal of Neurological Sciences 21,
(along with language difficulties) also occur in the 185193.
parents and siblings of children with language dis- Benton, A. L. (1964). Developmental aphasia and brain damage.
Cortex 1, 4052.
abilities (Plante, 1991). Brunswick, N., and Rippon, G. (1994). Auditory event-related po-
Evidence of structural left hemisphere anomalies tentials, dichotic listening performance, and handedness as
indices of lateralisation in dyslexic and normal readers. Inter-
in dyslexics is further consistent with data showing
national Journal of Psychophysiology 18 (3), 265275.
anomalous EEG recording from left-hemisphere lan- Byring, R., and Jrvilehto, T. (1985). Auditory and visual evoked
guage regions of dyslexics, as well as anomalous re- potentials of schoolboys with spelling disabilities. Developmen-
cordings from anterior frontal regions (Duffy and tal Medicine and Child Neurology 27, 141148.
McAnulty, 1990), anomalies in regional cerebral ac- Clark, M., Rosen, G., Tallal, P., and Fitch, R. H. (2000a). Impaired
processing of complex auditory stimuli in rats with induced
tivity in frontal and occipital areas of dyslexics as mea-
cerebrocortical microgyria. Journal of Cognitive Neuroscience 12
sured during reading tasks (Gross-Glenn et al., 1991; (5), 828839.
Shaywitz et al., 1998), and electrophysiological ano- (2000b). Impaired two-tone processing at rapid rates in
malies of cerebral lateralization in learning-disabled male rats with induced microgyria. Brain Research 871, 9497.
children (Mattson, Sheer, and Fletcher, 1992). Signif- Cohen, M., Campbell, R., and Yaghamai, F. (1989). Neuro-
pathological anomalies in developmental dysphasia. Annals of
icant differences in electrophysiological responses to
Neurology 25, 567570.
acoustic stimuli (including tones and speech) have Dalby, M. A., Elbro, C., and Stodkilde-Jorgensen, H. (1998). Tem-
also been reported in studies of both dyslexic and lan- poral lobe asymmetry and dyslexia: An in vivo study using
guage-impaired individuals (e.g., Brunswick and Rip- MRI. Brain and Language 62, 5169.
324 LEARNING DISABILITIES
Dawson, G., Finley, C., Phillips, S., and Lewy, A. (1989). A compari- guage- and learning-impaired children. Archives of Neurology
son of hemispheric asymmetries in speech-related brain po- 48, 539545.
tentials of autistic and dysphasic children. Brain and Language Jirsa, R. E., and Clontz, K. B. (1990). Long latency auditory event-
37 (1), 2641. related potentials from children with auditory processing dis-
Demb, J. B., Boynton, G. M., and Heeger, D. J. (1998). Functional orders. Ear and Hearing 11 (3), 222232.
magnetic resonance imaging of early visual pathways in dys- Kolb, B., and Fantie, B. (1989). Development of childrens brain
lexia. Journal of Neuroscience 18, 6,9396,951. and behavior. In C. R. Reynolds and J. Fletcher-Janzen, eds.,
Duffy, F. H., and McAnulty, G. (1990). Neuropsychological hetero- Handbook of child clinical neuropsychology, pp. 1740. New York:
geneity and the definition of dyslexia: Preliminary evidence Plenum Press.
for plasticity. Neuropsychologica 28, 555571. Kraus, N., McGee, T. J., Carrell, T. D., Zecker, S. G., Nicol, T. G.,
Fitch, R. H., Brown, C., Tallal, P., and Rosen, G. (1997). Effects of and Koch, D. B. (1996). Auditory neurophysiologic responses
sex and MK-801 on auditory processing deficits associated and discrimination deficits in children with learning prob-
with developmental microgyric lesions in rats. Behavioral lems. Science 273, 971973.
Neuroscience 111, 404412. Kulynych, J. J., Vladar, K., Jones, D. W., and Weinberger, D. R.
Fitch, R. H., Tallal, P., Brown, C., Galaburda, A., and Rosen, G. (1994). Gender differences in the normal lateralization of the
(1994). Induced microgyria and auditory temporal processing supratemporal cortex: MRI surface-rendering morphometry
in rats: A model for language impairment? Cerebral Cortex 4, of Heschls gyrus and the planum temporale. Cerebral Cortex
260270. 4, 107118.
Galaburda, A. M. (1992). Neurology of developmental dyslexia. Kushch, A., Gross-Glenn, K., Jallad, B., Lubs, H., Rabin, M., Feld-
Current Opinion in Neurology and Neurosurgery 5 (1), 7176. man, E., and Duara, R. (1993). Temporal lobe surface area
Galaburda, A. M., and Kemper, T. L. (1979). Cytoarchitectonic ab- measurements on MRI in normal and dyslexic readers.
normalities in developmental dyslexia: A case study. Annals of Neuropsychologia 31, 811821.
Neurology 6, 94100. Larsen, J., Hoien, T., Lundberg, I., and Odegaard, H. (1990). MRI
Galaburda, A. M., Menard, M. T., and Rosen, G. D. (1994). Evi- evaluation of the size and symmetry of the planum temporale
dence for aberrant auditory anatomy in developmental dys- in adolescents with developmental dyslexia. Brain and Lan-
lexia. Proceedings of the National Academy of Sciences of the United guage 39, 289301.
States of America 91, 8,0108,013. Leonard, C., Voeller, K. K. S., Lombardino, L. J., Morris, M. K.,
Galaburda, A. M., Sherman, G. F., Rosen, G. D., Aboitiz, F., and Hynd, G. W., Alexander, A. W., Andersen, H. G., Garofalakis,
M., Honeyman, J. C., Mao, J., Agee, O. F., and Staab, E. V.
Geschwind, N. (1985). Developmental dyslexia: Four consec-
(1993). Anomalous cerebral structure in dyslexia revealed
utive cases with cortical anomalies. Annals of Neurology 18,
with magnetic resonance imaging. Archives of Neurology 50,
222233.
461469.
Geschwind, N., and Levitsky, W. (1968). Human brain: Left-right
Leonard, L. B. (1998). Children with specific language impairment.
asymmetries in temporal speech region. Science 161, 186187.
Gross-Glenn, K, Duara, R, Barker, W. W., Lowenstein, D, Chang,
Leppnen, P. H. T., and Lyytinen, H. (1997). Auditory event-
J. Y., Yoshii, F., Apicella, A. M., Pascal, S., Boothe, T., Sevush,
related potentials in the study of developmental language-
S., Jallad, B. J., Novoa, L., and Lubs, H. A. (1991). Positron
related disorders. Audiology and Neuro-Otology 2, 308340.
emission tomographic studies during serial word reading by
Leppnen, P. H. T., Pihko, E., Eklund, K. M., and Lyytinen, H.
normal and dyslexic adults. Journal of Clinical and Experimental
(1999). Cortical responses of infants with and without a genet-
Neuropsychology 13, 531544.
ic risk for dyslexia, Part 2: Group effects. NeuroReport 10, 969
Haslam, R. H. A., Dalby, J. T., Johns, R. D., and Rademaker, A. W.
973.
(1981). Cerebral asymmetry in developmental dyslexia. Ar-
Lincoln, A. J., Courchesne, E., Harms, L., and Allen, M. (1995).
chives of Neurology 38, 679682.
Sensory modulation of auditory stimuli in children with au-
Herman, A., Galaburda, A., Fitch, R. H., Carter, A. R., and Rosen, tism and receptive developmental language disorder: Event-
G. (1997). Cerebral microgyria, thalamic cell size, and audito- related brain potential evidence. Journal of Autism and Develop-
ry temporal processing in male and female rats. Cerebral Cor- mental Disorders 25 (5), 521539.
tex 7, 453464. Livingstone, M. S., Rosen, G. D., Drislane, F. W., and Galaburda,
Hier, D. B., LeMay, M., Rosenberger, P. B., and Perlo, V. P. (1978). A. M. (1991). Physiological and anatomical evidence for a
Developmental dyslexia. Archives of Neurology 35, 9092. magnocellular defect in developmental dyslexia. Proceedings
Humphreys, P., Rosen, G. D., Press, D. M., Sherman, G. F., and of the National Academy of Sciences of the United States of America
Galaburda, A. M. (1991). Freezing lesions of the newborn rat 88, 7,9437,947.
brain: A model for cerebrocortical microgyria. Journal of Lovegrove, W., Garzia, R., and Nicholson, S. (1990). Experimental
Neuropathology and Experimental Neuroloy 50, 145160. evidence for a transient system deficit in specific reading dis-
Hynd, G. W., Marshall, R. M., and Semrud-Clikeman, M. (1991). ability. Journal of the American Optometric Association 2, 137146.
Developmental dyslexia, neurolinguistic theory, and devia- Mattson, A. J., Sheer, D. E., and Fletcher, J. M. (1992). Electrophy-
tions in brain morphology. Reading and Writing 3, 345362. siological evidence of lateralized disturbances in children with
Hynd, G. W., and Semrud-Clikeman, M. (1989). Dyslexia and learning disabilities. Journal of Clinical and Experimental
brain morphology. Psychological Bulletin 106, 447482. Neuropsychology 14, 707716.
Hynd, G. W., Semrud-Clikeman, M., Lorys, A. R., Novey, E. S., and Molfese, D. L., and Molfese, V. J. (1997). Discrimination of lan-
Eliopulos, D. (1990). Brain morphology in developmental guage skills at five years of age using event-related potentials
dyslexia and attention deficit disorder/hyperactivity. Archives recorded at birth. Developmental Neuropsychology 13, 135156.
of Neurology 47, 919926. Neville, H., Coffey, S., Holcomb, P., and Tallal, P. (1993). The
Jancke, L., Schlaug, G., Huang, Y., and Steinmetz, H. (1994). neurobiology of sensory and language processing in language
Asymmetry of the planum temporale. NeuroReport 5, 1,161 impaired children. Journal of Cognitive Neuroscience 5, 235
1,163. 253.
Jernigan, T. L., Hesselink, J. R., Sowell, E., and Tallal, P. (1991). Pihko, E., Leppnen, P. H. T., Eklund, K. M., Cheour, M., Gut-
Cerebral structure on magnetic resonance imaging in lan- torm, T. K., and Lyytinen, H. (1999). Cortical responses of in-
LEARNING THEORY: A History 325
fants with and without a genetic risk for dyslexia, Part 1: Age Witelson, S. F., and Pallie, W. (1973). Left hemisphere specializa-
effects. NeuroReport 10, 901905. tion for language in the newborn: Neuroanatomical evidence
Pinkerton, F., Watson, D. R., and McClelland, R. J. (1989). A of asymmetry. Brain 96, 641646.
neurophysiological study of children with reading, writing, Wong, B. Y. L. (1986). Problems and issues in the definition of
and spelling difficulties. Developmental Medicine and Child Neu- learning disabilities. In J. R. Torgesen and B. Y. L. Wong,
rology 31, 569581. eds., Psychological and educational perpectives on learning disabili-
Plante, E. (1991). MRI findings in the parents and siblings of spe- ties, 326. Orlando, FL: Academic Press.
Wood, F., Flowers, L., Buchsbaum, M., and Tallal, P. (1991). Inves-
cifically language-impaired boys. Brain and Language 41, 67
tigation of abnormal left temporal functioning in dyslexia
80.
through rCBF, auditory evoked potentials, and positron emis-
Plante, E., Swisher, L., Vance, R., and Rapcsak, S. (1991). MRI
sion tomography. Reading and Writing 3 (34), 379393.
findings in boys with specific language impairment. Brain and
Woody, R. H. (1989). Public policy and legal issues for clinical child
Language 41, 5266.
psychology. In C. R. Reynolds and J. Fletcher-Janzen, eds.,
Rosenberger, P. B., and Hier, D. B. (1980). Cerebral asymmetry Handbook of child clinical neuropsychology, 573584. New York:
and verbal intellectual deficits. Annals of Neurology 8, 300304. Plenum Press.
Rumsey, J. M., Dorwart, R., Vermess, M., Denckla, M. B., Kruesi,
M. J. P., and Rapoport, J. L. (1986). Magnetic resonance im- Chris Chase
aging of brain asymmetry in severe developmental dyslexia. Paula A. Tallal
Archives of Neurology 43, 1,0451,046. Elena Plante
Schultz, R. T., Cho, N. K., Staib, L. H., Kier, L. E., Fletcher, J. M., Revised by R. Holly Fitch
Shaywitz, S. E., Shankweiler, D. P., Katz, L., Gore, J. C., Dun-
can, J. S., and Shaywitz, B. A. (1994). Brain morphology in
normal and dyslexic children: The influence of sex and age.
Annals of Neurology 35, 732742.
Shaywitz, S. E., Escobar, M. D., Shaywitz, B. A., Fletcher, J. M., and
Makuch, R. (1992). Evidence that dyslexia may represent the LEARNING THEORY: A HISTORY
lower tail of a normal distribution of reading ability. New
Even before psychology became an experimental sci-
England Journal of Medicine 326, 145150.
Shaywitz, S. E., Shaywitz, B. A., Pugh, K., Fulbright, R., Constable, ence in the 1890s, learning was an important part of
R., Mencl, W., Shankweiler, D., Liberman, A., Skudlarski, P., it. But there came a time in the 1910s when psycholo-
Fletcher, J., Katz, L., Marchione, K., Lacadie, C., Gatenby, C., gists started to become fascinated by learning con-
and Gore, J. (1998). Functional disruption in the organization cepts and learning theories. The 1930s and 1940s are
of the brain for reading in dyslexia. Proceedings of the National sometimes called the golden age of learning theory;
Academy of Sciences of the United States of America 95, 2,636
that was when learning was the heart and soul of psy-
2,641.
Slaghuis, W. L., Lovegrove, W. J., and Freestun, J. (1992). Letter
chology. And then gradually the gold began to lose
recognition in peripheral vision and metacontrast masking in its glitter. The theorists did not seem able to settle
dyslexic and normal readers. Clinical Vision Sciences 7, 5365. their differences of opinion, psychologists began to
Stark, R. E., and Tallal, P. (1988). The children with specific lan- think that the differences were only a matter of opin-
guage impairment. In R. J. McCauley, ed., Children: Neuropsy- ion with little empirical significance, and there
chological Studies. Boston: College Hill Press. emerged a growing distaste for the great debates over
Tallal, P. (1987). Dysphasia, developmental. In G. Adelman, ed.,
fundamental issues. In the 1960s new procedures and
Encyclopedia of Neuroscience, Vol. 1, 351353. Boston:
Birkhauser Press. new phenomena were discovered that led psycholo-
(1988). Speech before the US Congress on clinical defini- gists away from the basic issues that the learning theo-
tions of developmental disability. rists had debated. Learning remains an important
Tallal, P., Miller, S., and Fitch, R. H. (1993). Neurobiological basis part of psychology, but the issues are quite different
of speech: A case for the preeminence of temporal processing. from the classical ones, and there is little theorizing
In P. Tallal, A. M. Galaburda, R. Llinas, and C. von Euler,
in the grand style that characterized the golden age.
eds., Temporal information processing in the nervous system, with
special reference to dyslexia and dysphasia, pp. 2747. New York: British empiricism culminated with the fourth
New York Academy of Sciences. edition of Alexander Bains The Senses and the Intellect
Tallal, P., and Piercy, M. (1973). Defects of non-verbal auditory (1894). Everything psychological about humans is
perception in children with developmental aphasia. Nature
based on experience and is due to learning, he said.
241, 468469.
Teszner, D., Tzavaras, A., Gruner, J., and Hcaen, H. (1972).
Bain argued that through association, sensations are
Lasymmtrie droite-gauche du planum temporale: A propos linked with each other and with responses. He argued
de ltude anatomique de 100 cervaeux. Revue Neurologique further that sensations can arouse ideas, and that
126, 444449. when ones ideas of pleasure and pain are aroused,
Tonnquist-Uhlen, I., Borg, E., Persson, H. E., and Spens, K. E. they are particularly likely to produce responses. At
(1996). Topography of auditory evoked cortical potentials in the same time, Morgan (1894) reported a number of
children with severe language impairment: The N1 compo-
rather casual learning experiments, interpreting his
nent. Electroencephalography and Clinical Neurophysiology 100,
250260. results very much as Bain had. Morgan had the same
Wada, J. A., Clarke, R., and Hamm, A. (1975). Cerebral hemisphe- reliance upon associationist theory, empiricist philos-
ric asymmetry in humans. Archives of Neurology 32, 239246. ophy, and the pleasure-pain principle. The difference
326 LEARNING THEORY: A History
was that Bain was a philosopher who thought about Thus began two long lines of theoriststhe contiguity
human knowledge, whereas Morgan was a naturalist people and the reinforcement people. Watson also
who conducted research with animals. Looking back, discarded all motivation concepts; hunger became
it appears that Morgans orientation was compelling just an internal stimulus, and emotion just a set of
because learning theory turned to the study of ani- fixed responses to certain kinds of stimuli. Watsons
mals rather than of human learning, and to experi- behaviorism was appealing because it was so mecha-
mental studies rather than philosophical speculation. nistic and so conceptually simple. At the same time,
Watson suggested that to understand anything and
everything psychological, one had to start with learn-
Edward Thorndike ing. The description of the fear conditioning of Little
The first systematic experimental study with ani- Albert (Watson and Rayner, 1920) stated that that is
mals was Edward Thorndikes (1898) puzzle-box ex- how our personalities take shape. The message, which
periment. Thorndike simply measured the time it was believed by many, if not all, psychologists, was
took a cat to pull a string that opened the door of the that wherever one wanted to go in the field, one had
box so that it could go out and eat. He was struck by to start with learning theory.
the fact that the time scores decreased steadily and
smoothly over trials; he never found a sudden im- There was, however, one problem that would be
provement in performance. He therefore concluded significant historically: In order to deal with certain
that the cat was not learning anything about ideas but cognitive-looking phenomena, Watson introduced
must be acquiring some sort of direct connection be- some miniature responses, responses so small that
tween the stimulus (S) that was present and the re- they were basically unobservable. When individuals
sponse (R) of pulling the string. It must be a direct first learn to read, they read out loud. As they become
neural connection between S and R. Thus, at the out- practiced enough to read silently, they still move our
set of learning experiments and learning theory, lips. Ultimately nothing seems to move. But no, Wat-
there was a strong commitment to the S-R concept of son asserted, there are still tiny responses in the
learning. One attraction of this approach was that it mouth and throat. And it is the feedback from these
minimized all mentalist concepts; it took the mind small responses that mediates and controls what looks
out of the picture. It was scientific. like intelligent or speech-related behavior. So Wat-
Thorndike (1911) introduced what he called the sons learning theory, which was so elegantly simple,
Law of Effect, what we now call the law of Reinforce- objective, and scientific, was obliged to hypothesize
ment. Whether an S-R connection is strengthened on unobservable little responses.
a particular trial depends, he argued, upon the envi-
ronmental effect of the response. If the effect is posi-
tive, such as providing the hungry cat with food, then
Edwin Guthrie
the connection gets stronger. Bain had a similar prin- Edwin Guthrie (1935) was a contiguity theorist
ciple; he said the association will get stronger if the who followed Watson in rejecting motivation and in
response produces pleasure. But pleasure is a mental other ways. He explained the great complexity and
concept, and so it had to go. Thorndikes version pre- unpredictability of behavior in terms of the complexi-
served the reinforcement mechanism but got rid of ty of the stimulus situation. At any moment there are
the mind and everything in it. It also took the control potentially millions of stimulus elements that one
of the organisms behavior away from the organism might respond to or have ones behavior conditioned
and put it in the environment. The cats behavior was to. Conditioning itself is simple and sudden, but the
totally controlled by the stimuli in the situation and effective stimulus situation is impossible to control. So
the food. Guthries learning theory was forced to hypothesize
unobservable little stimuli. The same problem in time
caused the demise of Clark L. Hulls theory, which, in
John B. Watson order to account for what looked like cognitive behav-
John B. Watson (1914) called Thorndikes ap- ior, had to hypothesize unobservable little motivation
proach behaviorism; it was the ultimate mechanistic terms, entities called rG. The Skinnerians are no bet-
psychology. Everything remotely related to the mind ter off, for all their claims of objectivity and freedom
was discarded. Even Thorndikes reinforcement from theory. They talk about self-reinforcement when
mechanism was tainted because a positive effect the organism does something it is not supposed to,
looked too much like pleasure. In Watsons psycholo- and they talk about conditioned (acquired) reinforce-
gy just the stimulus and response occurring together ment when it does something that looks cognitive.
would create a connection between them. Everything Thus they are hypothesizing unobservable little rein-
was habit, or what was called learning by contiguity. forcers.
LEARNING THEORY: A History 327
Ivan Pavlov and Edward Tolman stimuli, sequences of events (what leads to what), and
where things are located in space (a map of a maze).
When Russian scientist Ivan Pavlovs work finally
In the 1940s Tolman developed the theme that ani-
became available in English translation in 1927, it
mals learn places rather than responses (see, e.g.,
seemed vaguely familiar. It reminded readers of Wat-
Tolman, Ritchie, and Kalish, 1946).
sons. They shared the same view of how important
learning is and how it should be studied. The theory Those were exciting times. There were two para-
included no motivation, no reinforcement, no mind, digms, Pavlovian and Thorndikian, to be organized.
and it was all very scientific. Pavlov emphasized inhi- One could explain all learning with this one, or that
bition, something that American psychologists had one, or with some of each (Mowrer, 1947). Mowrer at-
largely ignored but in time found fascinating. What tributed emotional and motivational learning to Pav-
was new was the procedure, the pairing of two stimuli; lovian mechanisms, and most other learned behavior
the bell and the food had to occur together. The criti- to reinforcement. There were contiguity theorists and
cal contingency the experimenter had to control was reinforcement theorists. Some people studied moti-
the timing of the stimuli. With Thorndikes proce- vation and others ignored it. Some were mechanists,
dure the critical contingency was the relationship be- and others appeared very cognitive. Some believed in
tween the response and its effect. The procedural tiny stimuli or responses, twinkles and twitches, and
contrast was called by different people Pavlovian ver- others looked at behavior globally. And new behavior-
sus trial and error, or classical conditioning versus inal phenomena were being discovered at an accelerat-
strumental, or respondent versus operant. There was ed rate.
always the uneasy feeling that while the two proce-
dures were easily distinguished by their defining con-
tingencies, perhaps there were not two separate un- Clark L. Hull
derlying processes involved. Could anyone put it all together? It seemed that
If there is a variable that one never varies, then Clark L. Hull and his dedicated followers might do it;
one will never see its significance. Pavlov knew that they certainly tried. The great theory (Hull, 1943) was
his dogs had to be hungry or they would not salivate, based on the reinforcement of S-R habits, but habits
so he always worked with hungry dogs. And so he were only indirectly expressed in behavior. To be
never saw the significance of motivation. The first manifest, a habit had to be motivated by drive and/or
learning theorist to stress motivation was Edward Tol- incentive, and had to overcome the different kinds of
man (1932). He described a study by his student Tink- inhibition that might be present. It was a very com-
lepaugh, who was studying monkeys and reinforcing plex theory, but its virtue was that its complexity
their correct responses with pieces of banana. Occa- promised to match that of the empirical world. It was
sionally Tinklepaugh would substitute lettuce for the also a very explicit theory; everything was spelled out
banana; when this happened, the animals threw tan- in detail. The theory even appeared to be able to ex-
trums and became emotionally upset. Monkeys usual- plain away some of the mysterious things Tolman had
ly like lettuce and it can certainly be used as a rein- reported. It was full of promise, and it gathered an
forcer, so what had Tinklepaugh encountered here? enormous amount of attention.
First, he had the trivial finding that monkeys like ba-
Hull was fortunate to have a number of brilliant,
nanas better than lettuce. Second, he had discovered
energetic young associates who all agreed that this
that monkeys can anticipate receiving, or expect to
was the right kind of theory. Their disagreements
receive, a particular kind of food. Thus, he had dis-
were over details, and those differences called for fur-
covered what we call incentive motivation, motivation
ther experiments to get everything straightened out.
that depends upon the expected value of the out-
One could fuss over details, but all the Hullians en-
come. Tolmans students also demonstrated effects of
dorsed the basic program. Miller and Dollard (1941)
drive motivation, motivation that depends on the
proposed a simpler model that anticipated many fea-
physiological state of the animal.
tures of the great theory. Mowrer (1939) anticipated
Thus, Tolman suddenly introduced two kinds of the all-important mechanism of reinforcement; he
motivation, one psychological and one physiological, said a response is reinforced when it results in the re-
and he had abundant evidence for both kinds. He duction of some source of drive, such as fear. Kenneth
also challenged other conventional parts of Watsoni- Spence was another early associate of Hulls, and he
an behaviorism, such as its mechanistic commitment. had a multitude of graduate students who were proud
He introduced a cognitive language (e.g., expectancies) to call themselves neo-Hullians and to work out dif-
in place of connections and neurons. Tolman main- ferent aspects of the theory. For them, the 1950s
tained that animals learn not S-R connections but the looked like the golden age because it was the time of
predictive significance and value of environmental awakening, the time of promise, and the time of pay-
328 LEARNING THEORY: A History
off. Many of them moved away from animal learning behind and move into other areas of psychology and
and into human experimental, social, developmental, into applied problems. Others began to understand
and clinical psychology. Learning was the center, but that there was something fundamentally flawed in the
the time had come to apply the principles of behavior whole enterprise begun by Thorndike and Watson.
far and wide. Watson and Thorndike, the first learn- Psychology did not become a science because it exor-
ing theorists, had promised to build a better world cised the mind and analyzed everything into atomic
with learning theory, and the neo-Hullians felt that S-R units; it became a science as it looked systemati-
the time had come to make good on that promise. cally at psychological phenomena. If one wants to un-
Two things went wrong. One, which should have derstand a social phenomenon, then one does not
been only a minor tactical setback, was that the drive- need a basic learning theory; one needs to look at so-
reduction hypothesis of reinforcement was wrong. cial situations, social motivation, and social behavior
That was discovered early on (Sheffield and Roby, strategies. And that is the sort of thing psychologists
1950), and in his last written work Hull (1952) ac- do now.
knowledged the problem and said the hypothesis
might have to be altered. The whole point of theory,
according to Hull, is to use it to generate research and See also: GUTHRIE, EDWIN R.; HULL, CLARK L.;
then use the research to modify the theory. So the loss MATHEMATICAL LEARNING THEORY; PAVLOV,
of this particular hypothesis should not have hurt the IVAN; THORNDIKE, EDWARD; TOLMAN, EDWARD
basic Hullian program. But the neo-Hullians were se- C.; WATSON, JOHN B.
verely wounded and badly discouraged. Further-
more, by about 1970, new difficulties had arisen with
Bibliography
the concept of reinforcement itself (Bolles, 1975).
Bain, A. (1894). The senses and the intellect, 4th edition. London:
A second difficulty was that during the 1960s Longmans, Green.
there were many problems with incentive motivation. Bolles, R. C. (1975). Learning theory. New York: Holt.
Guthrie, E. R. (1935). The psychology of learning. New York: Har-
It was based on the little response rG, which had all
pers.
the conceptual properties of a response (i.e., it was Hull, C. L. (1943). Principles of behavior. New York: Appleton.
elicited by stimuli, it was conditionable, and it could (1952). A behavior system. New Haven, CT: Yale University
be motivated). The problem was that it did not seem Press.
to be observable. The rG concept was needed to ac- Miller, N. E., and Dollard, J. (1941). Social learning and imitation.
count for Tolmans discovery that animals learn New Haven, CT: Yale University Press.
places (Hull held that it was elicited by spatial stimu- Morgan, C. L. (1894). An introduction to comparative psychology. Lon-
don: Scott.
li), and it was needed to explain a variety of other ef-
Mowrer, O. H. (1939). A stimulus-response analysis of anxiety and
fects; but it was beginning to look like a fiction, a fig- its role as a reinforcing agent. Psychological Review 46, 553
ment. The Hullians said that when the animal looks 564.
to the left, it encounters stimuli that elicit rG and so (1947). On the dual nature of learning: A reinterpretation
it moves in that direction. Tolman said that the ani- of conditioning and problem-solving. Harvard Education-
mal expects food to be off to the west, and since it is al Review 17, 102148.
hungry and values food, it moves in that direction. If Pavlov, I. P. (1927). Conditioned reflexes, trans. G. V. Anrep. London:
Oxford University Press.
you cannot measure rG, then you have no way to test
Sheffield, F. D., and Roby, T. B. (1950). Reward value of a non-
Hulls view against Tolmans view of the situation. nutritive sweet taste. Journal of Comparative and Physiological
Eventually psychologists figured out that Tolmans Psychology 43, 461481.
theory is untestable because one cannot measure ex- Thorndike, E. L. (1898). Animal intelligence: An experimental
pectancies or values, and that Hulls theory is untest- study of the associative processes in animals. Psychological Re-
able because one cannot measure rG. Learning theo- view Monograph Supplement 2 (8).
ries are basically untestable. The great promise of (1911). Animal intelligence. New York: Teachers College
Press.
Hulls theory was slipping away. The golden age was
ending. York: Century.
Tolman, E. C., Ritchie, B. F., and Kalish, D. (1946). Studies in spa-
tial learning: II. Place learning versus response learning. Jour-
B. F. Skinner nal of Experimental Psychology 36, 221229.
Watson, J. B. (1914). Behavior. An introduction to comparative psycholo-
Some found comfort in B. F. Skinners approach.
gy. New York: Holt.
It could have been an alternative learning theory but Watson, J. B., and Rayner, R. (1920). Conditioned emotional reac-
chose to present itself as theoretically neutral. Cer- tions. Journal of Experimental Psychology 3, 114.
tainly Skinnerians did not worry about theoretical
matters as such. And they were eager to leave learning Robert C. Bolles
LEARNING THEORY: Current Status 329
LEARNING THEORY: Figure 1

CURRENT STATUS
When most students of psychology hear the term
learning theory, they probably think about the history
of psychology. After all, what comes to mind when you
think about learning theory? Perhaps one thinks of
the names of many great psychologists, such as
Thorndike, Pavlov, Tolman, Hull, and Skinner. Per-
haps you recall Skinners pigeons pecking at lights in
an operant chamber to obtain morsels of food. Or
maybe you remember Pavlovs famous discovery that
the sound of a metronome can be conditioned to
make dogs salivate if it is presented together with
powdered meat. But is learning theory a relic of psy-
chologys past, as these vignettes might suggest? The
answer is no. In fact, learning theory is an active and
vibrant area of study in psychology, with a steady
stream of fresh theories that are attempting to tacklel-
earning about complex stimuli, the involvement of Learning curves computed with the Rescorla-Wagner model.
context in memory retrieval, and the role of time in The asymptote of each curve (the point on the curve where
learning is maximal) approaches , which represents the
learning. Moreover, neuroscientists are increasingly
magnitude of the US. The rate at which learning occurs is
relying on learning theory in the quest to dissect the
determined by , which represents CS salience. The values
brain systems involved in learning and memory. chosen for each variable are arbitrary. Note that the amount of
Learning theory is alive and well. learning on each trial (the change from one trial to the next)
decreases over the course of training. In other words, learning is
the greatest on the first trial.
Modeling Learning
One of the most influential modern learning the-
ories was crafted in the early 1970s by Robert A. Res- of learning that has already accrued to all of the CSs
corla and Allan R. Wagner (Rescorla and Wagner, presented on that trial (VT). A key feature of the
1972). The Rescorla-Wagner model accounts for sev- model is that the amount that is learned on any given
eral features of a form of associative learning called trial (CS, US, or CS-US presentation) is determined
Pavlovian conditioning. Pavlovian or classical condi- by how much the animal has already learned about all
tioning is a type of learning in which neutral or con- of the CSs present on that trial.
ditional stimuli (CS), such as tones or lights, can pre-
dict the occurrence of biologically significant or To illustrate some of the properties of the Rescor-
unconditional stimuli (US), such as food or illness. la-Wagner model, Figure 1 displays a family of learn-
After having been paired with a US, a CS elicits a ing curves computed with Equation 1. Note that the
learned or conditional response (CR) that often is highest point on each learning curve (the asymptote)
similar to the unlearned or unconditional response approaches the value of . The rate at which the as-
(UR) elicited by the US. ymptote is approached is determined by . This
means that the amount that is learned during classical
Rescorla and Wagner used a simple mathematical
formula to model Pavlovian conditioning: conditioning is set by the strength of the US and the
speed of that learning is set by the salience of the CS.
For example, Pavlovs dogs learned to salivate pro-
VCS = ( VT) (1)
fusely to a CS that was followed by a large quantity of
powdered meat (the US). Similarly, the speed at
The equation has three critical variables: , the which the dogs attained that profuse quantity of sali-
salience of the CS, , the magnitude of the US, and vation was rapid when Pavlov used a salient CS, such
VT, the total amount of learning acquired by all of as a loud metronome. The Rescorla-Wagner model
the CSs present on the trial (For simplicity, a fourth
predicts these features of Pavlovian conditioning.
variable, , describing the salience of the US has been
omitted). According to this equation, the amount that Despite its simplicity, the Rescorla-Wagner
is learned on a conditioning trial (VCS) is the product model can account for a wide variety of interesting
of the salience of the CS () and the difference be- learning phenomena including extinction, blocking,
tween the magnitude of the US () and the amount and overshadowing. For example, in blocking condi-
330 LEARNING THEORY: Current Status
tioning, one CS (CSA) prevents conditioning to anoth- pound trials? You might imagine that there is profuse
er CS (CSB) when the two are presented together dur- salivation on these compound trials because both CSA
ing compound conditioning (CSA/CSB-US). Simply and CSB predict food. But animals and people readily
put, learning does not occur to CSB because the US learn to withhold salivation on the compound trials
is already perfectly predicted by CSA-CSB doesnt tell because no food is delivered on these trials.
the animal anything it doesnt already know. In terms
of the Rescorla-Wagner model, the value of (-VT) The Rescorla-Wagner model has trouble with this
is close to zero on the compound trial, because VCSA phenomenon because the compound stimulus
is close to 1 (CSA is already conditioned). The beauty CSA/CSB is treated as the sum of two elements (CSA
of the Rescorla-Wagner model is that this simple and CSB), which strongly predict the US. Despite the
mathematical equation can predict many complex fact that the US does not occur on compound trials,
learning phenomena. VCSA/CSB is always higher for the compound stimulus
than the individual CSs. This incorrectly predicts that
the compound stimulus will elicit a greater CR than
Is the Whole More Than the Sum of Its the elements presented alone. However, Pearces
Parts? model can readily account for the fact that animals
and people respond more on to the CSA and CSB
The Rescorla-Wagner model is an elemental compared to the compound stimulus. Pearce assumes
model of conditioning because it makes predictions that that compound stimulus is distinct from the ele-
about learning based on the individual stimuli in a ments that compose itin effect, it is like a third CS.
given trial. However, stimuli often occur together in Using a computational formula much like the Rescor-
complex combinations during classical conditioning. la-Wagner model, Pearce shows that animals rapidly
Consider, for example, the places where classical con- learn to respond discriminatively on CSA, CSB, and
ditioning happens, such as a dentists office. Here you CSA/CSB trials. By considering stimuli as configura-
learn to fear (CR) the sound of a dental drill (CS) that tions that are more than the sum of their parts, Pearce
predicts tooth pain (UR). But you also learn that the has provided a new and powerful explanation for sev-
dentists office itself, which is composed of many eral forms of learning.
unique stimuli (reclining chair, observation light, gur-
gling fountain, and the odor of the dentists breath,
to name a few), predict tooth pain and elicit learned Retrieving the Meaning of an Event
fear. Rescorla and Wagner would argue that each of
Of course, stimuli, whether they are CSs or words,
these contextual stimuli is an element, and that learn-
for that matter, often have ambiguous meanings.
ing to these elements happens individually. Accord-
Consider, for example, the word vessel. This word has
ing to their model, VCS is calculated for each element
several different meanings; it may refer to a water-
separately in order to determine the amount of learn-
craft, a container, or a tube containing blood, for ex-
ing that happens to these stimuli.
ample. How do we interpret the meanings of such
An alternative has been proposed by John M. ambiguous stimuli? Recent studies indicate that the
Pearce at Cardiff University in the United Kingdom context within which the stimulus occurs is important
(Pearce, 1994). Pearce suggests that stimuli that co- for understanding its meaning. For instance, sen-
occur should be treated as unique combinations or tences such as The Titanic was a grand vessel or
configurations of stimuli that are distinct from their The aorta is the largest vessel in the body provide
elements. Hence, rather than break down the den- a context in which to understand the meaning of the
tists office into many individual elements (such as the word vessel.
gurgling sound and dentists bad breath), Pearce ar-
Pavlovian conditioning experiments in animals
gues that we form a configuration of stimuli, such as
indicate the importance of context in understanding
dentists office, that includes all of the individual
the meaning of ambiguous CSs. For example, if a CS
stimuli and becomes associated with the US.
is first paired with food for several trials, but then
The power of this configural approach is appar- presented by itself (without food), the ability of the CS
ent in a type of learning called negative patterning. to elicit a CR is weakened. This process is called ex-
In this form of learning, two different CSs (CSA and tinction. Although one might imagine (as the Rescor-
CSB) are paired with a US, such as food. After several la-Wagner model predicts) that presenting the CS by
pairings, these CSs will elicit responses associated itself results in unlearning of the CS-US memory,
with food delivery, such as salivation. During this considerable data suggest that extinction itself repre-
training, trials are interspersed in which CSA and CSB sents new learning. In this case, animals learn that the
are presented together (a so-called compound trial) CS predicts the absence of the US. So how does the
without food delivery. What happens in these com- animal know how to respond to this ambiguous CS?
LEARNING THEORY: Current Status 331
After all, the CS has predicted both the presence and vals associated with the delay of reinforcement and
absence of the US and various times. the ITI. These computations are then used to make
Elegant work by Mark E. Bouton at the University decisions about how to respond to a particular stimu-
of Vermont has solved this enigma (Bouton, 1993). lus. In general, conditioning improves when the ratio
Bouton has found that the context in which condi- between the ITI and the delay of reinforcement is
tioning and extinction occur is critical for determin- large. And, as mentioned before, increasing the dura-
ing how animals respond to the CS. If animals are tion of the delay and reinforcement gradient results
conditioned (i.e., receive CS-US pairings) in one con- in weaker conditioning unless the ITI increases as
text but are extinguished (i.e., receive the CS by itself) well. These predictions are made without any appeal
in a different context, then the context controls how to associations between the CS and US; in essence, the
the animals respond. If they are placed in the context animals are just keeping time.
where the CS was presented by itself, they retrieve a
memory that the CS predicts the absence of the US Despite the computational power of rate-
and show little conditional responding. In contrast, if estimation theory, the model faces with a fundamen-
the animals are placed in the conditioning context, tal challenge: Does the brain learn this way? The ana-
they retrieve a memory that the US will follow the CS tomical and physiological facts about brain function
and show high levels of learned behavior. Hence, seem to be more consistent with associative models of
Bouton argues that contextswhich include physical learning and memory. For example, sensory path-
environments, internal states (such as hunger), and ways in the brain transmit information about CSs,
even timeregulate the ability of an ambiguous CS USs, contexts, and other stimuli. There is consider-
to evoke a learned response. able convergence among these pathways in several
brain areas thought to be involved in learning and
memory, such as the hippocampus and amygdala.
Its All in the Timing
Neurons in the brain exhibit changes in synaptic
The theories of Rescorla, Wagner, Pearce, and function after learning, and these changes are consis-
Bouton all assume that the foundation of learning is tent with the associative nature of Pavlovian condi-
an association, a bond or link formed between two or tioning, for example (Maren, 1999).
more stimuli. Pavlovian conditioning, for example,
assumes that an association forms between the CS and One of the promises of associative learning theo-
US, and that this association is necessary for the CS ry is that it is stimulating and guiding neuroscientists
to elicit a learned response (CR). In a radical depar- in their quest for discovering the brain mechanisms
ture from this fundamental assumption of associative of learning and memory. The continued elaboration
learning theory, Randy Gallistel and John Gibbon and refinement of learning theories are an integral
have argued that the time at which events occur, not component of understanding both brain and behav-
associations between the events, is at the heart of ior in humans and animals.
learning and memory (Gallistel and Gibbon, 2000).
In any conditioning experiment, stimuli occur in
a temporal as well as a physical context. Years of ani- See also: LEARNING THEORY: A HISTORY
mal and human research have yielded important in-
formation about the influence of timing on Pavlovian
Bibliography
conditioning. For example, conditioning is often op-
Bouton, M. E. (1993). Context, time, and memory retrieval in the
timal when the CS is turned on slightly before the US. interference paradigms of Pavlovian conditioning. Psychologi-
Long intervals between the onset of the CS and US cal Bulletin 114, 8099.
tend to produce weak learning; this is commonly Gallistel, C. R., and Gibbon, J. (2000). Time, rate, and condition-
called the delay of reinforcement gradient. However, ing. Psychological Review 107, 219275.
conditioning is related to both the delay of reinforce- Maren, S. (1999). Long-term potentiation in the amygdala: A
ment and the amount of time that elapses between mechanism for emotional learning and memory. Trends in
CS-US trials (the intertrial interval or ITI). Learning
Pearce, J. M. (1994). Similarity and discrimination: A selective re-
rates are comparable across short and long delays of view and connectionist model. Psychological Review, 587607.
reinforcement long as the ITI is varied in accordance Rescorla, R. A., and Wagner, A. R. (1972). A theory of Pavlovian
with the CS-US delay. conditioning: Variations in the effectiveness of reinforcement
Gallistel and Gibbons temporal model, which eds., Classical Conditioning II: Current Research and Theory. New
they call rate estimation theory, explains this and York: Appleton-Century-Crofts.
other properties of conditioning. In their model, ani-
mals compute the ratios between the temporal inter- Steve Maren
332 LEFT HEMISPHERE
LEFT HEMISPHERE memory (mass action). Walter Hunter was quick to

point out that removing more cerebral cortex re-
See: KNOWLEDGE SYSTEMS AND MATERIAL- moved more sensory information (visual, auditory,
SPECIFIC MEMORY DEFICITS
kinesthetic), in effect reducing the number of avail-
able cues (e.g., animals that are blind do not learn
mazes well). This issue has never really been resolved,
at least for complex maze learning in the rat, al-
LOCALIZATION OF MEMORY TRACES though we now know that the hippocampus is impor-
tant for such memories.
The brain consists of a vast number of individual cells
called neurons. Individual neurons form highly com- Following Lashleys failure to localize memory
plex patterns of interconnections with many other traces, some scientists adopted the view that they were
neurons. Each of these connections is called a synapse distributed either widely throughout the brain or
and a collection of interconnected neurons is called widely within certain brain structures like the cerebral
a neural network. It is within these networks of neurons cortex. But as more was learned about the anatomical
and synapses that memories are formed and stored. and functional organization of the brain, it became
The term memory trace, also called the engram, broadly clear that the brain does not have a diffusely distribut-
refers to the change(s) in the brain that serves to store ed organization; instead, it has a highly structured or-
a memory. To fully understand the nature of a memo- ganization. Donald Hebb, in his important and influ-
ry trace, at least three different but interrelated prop- ential book The Organization of Behavior (1949),
erties must be elucidated. First, the precise region proposed a resolution of this dilemma by assuming
within the brain where the memory dependent that the organization of a memory trace can be com-
changes occur must be localized. This entails identify- plex and involve a number of brain areas but that the
ing the specific neural network (or neural circuit) that trace can involve specific connections in particular
subserves the formation, storage, and retrieval of the areas. This remains a common view. Hebb also pro-
particular memory and then localizing the site(s) of posed a possible mechanism of memory trace forma-
change(s) within that network that mediates storage tion that has come to be known as the Hebb synapse.
of the memory. Second, once the site of memory stor- In brief, he argued that at neurons where traces are
age has been identified, the biophysical properties of formed, there must be active input from a to-be-
the changes that occurred within the neural network learned source (e.g., conditioned stimulus in Pavlov-
as a result of memory formation must be identified. ian terms) at the same time the neuron is firing action
For instance, these changes might involve strength- potentials. The Hebb synapse has come to be viewed
ening synaptic connections between different neu- more generally as a strengthening or weakening of
rons, a process that might entail expression of differ- one input (synapses) to a neuron if this input is active
ent, memory related genes. Finally, in addition to concurrent with activation from another input to the
identifying the site of memory formation and the na- neuron.
ture of the changes that occur, the memory specific
pattern of neural activity within the network that oc- By the end of the twentieth century, the focus had
curs during recall of a memory must be delineated. shifted away from memory traces in complex tasks to
This entry will focus primarily on the first step in un- more specific and discrete learning and memory
derstanding the nature of memory traces: localizing tasks, and research emphasized identifying the entire
traces within the brain. circuitries essential (necessary and sufficient) for par-
ticular forms of memory. Only after this has been ac-
complished can the memory traces be localized and
History of Memory Localization analyzed. Well-established methods of lesions, electri-
In the early days of behaviorism, the observable, cal recording of neuronal activity, and electrical stim-
quantifiable study of behaviors, it was thought that ulation of brain tissue, together with anatomical trac-
each memory was represented as a change in the ing of pathways in the brain, have enabled much
brain at one particular place. Karl Lashley began the progress in the identification of essential memory cir-
search for the memory trace, stressing the now obvi- cuits in the brain, at least for simpler forms of learn-
ous point that in order to analyze the nature of mem- ing, although the experimental difficulties are formi-
ory traces, it is necessary to find them. In his classic dable. Relatively new methods such as functional
1929 monograph, Brain Mechanisms and Intelligence, imaging, probes for genetic expression, or localized
he concluded that memories, at least memories for infusion of highly specific receptor antagonists or ag-
complex mazes in rats, did not have any particular onists have also become widely used tools in the
locus in the cerebral cortex (equipotentiality); the search for essential memory circuits. Once the com-
more cortex removed, the more the impairment in plete circuit for a particular form of learning has been
LOCALIZATION OF MEMORY TRACES 333
identified, the next step of localizing the memory mammals, involves associating a neutral stimulus,
trace(s) within that circuit is orders of magnitude such as a brief tone, with another stimulus, for in-
more difficult. Indeed, there are no universally stance an air puff to the eye, that evokes a specific
agreed upon methods for doing so. This aspect of the movement such as an eye blink. After presenting the
search for memory traces has become the conceptual tone paired with the air puff, subjects are conditioned
center of the field. to blink their eye to the tone alone. Using the meth-
ods of stimulation, lesions, and recordings described
above, the neural circuit that mediates this form of
Localization of Different Types of motor learning was found to critically involve the cer-
Memories ebellum and its associated brain-stem structures.
There are many different types of learning and Once this essential circuit was identified, the memory
memory; for instance, learning to ride a bicycle dif- trace was localized within the circuit to a particular re-
fers from memorizing a list of facts. A distinction is gion of the interpositus nucleus in the cerebellum. In
often made between two general categories of memo- addition to the interpositus, there appear to be addi-
ry: declarative (learning what) and procedural tional storage sites in the cerebellar cortex, and these
(learning how). Many other terms have been sug- sites certainly are distributed, in the sense that many
gested for this dichotomy; extreme examples of the thousands of neurons are involved. Localization of
two types of memory are ones memory of ones own the memory trace for eyeblink conditioning is a criti-
recent experiences (declarative) and classical or Pav- cal first step toward elucidating the mechanisms of
lovian conditioning, where a specific conditioned re- plasticity and the network-level properties mediating
sponse like salivation or eye blink is learned to a par- expression of the stored memory.
ticular conditioned stimulus (procedural). Although Unlike learned motor behaviors, such as eyeblink
both types of memory formation involve many re- conditioning, that involve the cerebellum, fear condi-
gions of the brain, the brain structures and systems es- tioning, as in conditioned changes in heart rate and
sential for the two types of memories are quite differ- blood pressure following pairing of a tone or light
ent. Indeed, there are several different memory with a painful electric shock, critically involves the hy-
circuits and systems in the mammalian brain. Some pothalamus and amygdala but not the cerebellum. As
of these will be noted here; each is treated in a sepa- with eyeblink conditioning, much of the circuitry es-
rate article in this volume. sential for conditioned fear has been identified. For
In humans and other mammals, the hippocam- this particular form of learning and memory, the
pus appears to play a key role in recent experiential amygdala is critically involved. In particular, the criti-
memory (declarative). Extensive damage to the hip- cal region for memory formation for conditioned fear
pocampus can markedly impair recent memory in hu- appears to be localized to a region of the amygdala
mans and monkeys. Evidence suggests that the im- called the basolateral complex. However, it is uncer-
pairment is more in the establishing of memoriesa tain whether the amygdala is the site of long-term
process that appears to take weeks in monkeys and storage of this type of memory. The amygdala is also
may take years in humansthan in their retrieval. In critically involved in hormonal modulation of memo-
rodents, recent working memory and spatial mem- ry storage.
ory are impaired by hippocampal lesions. Very recent Because the memory traces for conditioned
or short-term memory in monkeys also involves the motor responses or conditioned fear are fairly local-
prefrontal areas of the cerebral cortex. The thalamus, ized, the memory traces for procedural learning tasks
the largest subdivision of the diencephalon, also plays in general may also be relatively localized. In contrast,
a role in recent memory in humans. However, long- memory traces for declarative memories may be
term permanent memories, representing our knowl- much more widely distributed. On the other hand,
edge and our life experiences, are not stored in the the fact that damage to speech areas in the human ce-
hippocampus, prefrontal cortex, or thalamusand rebral cortex appears to abolish memory for language
thus are not impaired by damage to these structures. suggests that this complex learning and memory pro-
The cerebral cortex is often suggested as the storage cess, perhaps the most complex yet evolved in nature,
site for these long-term memories, but definitive evi- may have a considerable degree of localization.
dence is lacking.
A somewhat different approach has been taken in
In contrast, the clearest evidence for a high de- the study of simplified neuronal circuits in certain
gree of localization of a memory trace exists for classi- invertebrate preparations where the number of neu-
cal conditioning of discrete behavioral responses rons is small, their sizes are large, and their intercon-
for instance, the conditioned eyeblink response. This nections are well specified. Here, simplified neural
type of learning, which occurs in humans and other circuits containing only a few identified neurons can
334 LOCALIZATION OF MEMORY TRACES
be isolated and particular training procedures, usual- synapses on neuron dendrites, and even in changes
ly classical conditioning, can result in the circuits in the number of dendritic branches in certain types
showing changes in activity that can be long-lasting of neurons. Possibly all these processes and many
and can closely resemble similar associative learning more are involved in memory formation. A great deal
in mammals. In these preparations it is possible to lo- of progress has been achieved in the identification of
calize the learning-induced changes in the activities essential memory circuits in the brain. The search for
of the neurons and analyze the underlying mecha- the memory trace has become one of the most active
nisms in some detail. These mechanisms can then and exciting fields in neuroscience and psychology.
provide models of putative mechanisms of memory
See also: AMNESIA, ORGANIC; CODING PROCESSES:
storage in the mammalian brain.
ORGANIZATION OF MEMORY; EMOTION, MOOD,
AND MEMORY; GENETIC SUBSTRATES OF
MEMORY: CEREBELLUM; GUIDE TO THE
Mechanisms of Memory Formation ANATOMY OF THE BRAIN: AMYGDALA; GUIDE TO
As more and more memory traces are localized in THE ANATOMY OF THE BRAIN: SYNAPSE; HEBB,
the brain, understanding the biophysical properties DONALD; HORMONES AND MEMORY;
of the changes that occur as a result of memory for- INVERTEBRATE LEARNING; KNOWLEDGE
mation becomes possible. Theories abound regard- SYSTEMS AND MATERIAL-SPECIFIC MEMORY
ing the nature of the mechanisms of memory trace DEFICITS; LASHLEY, KARL; LONG-TERM
DEPRESSION IN THE CEREBELLUM,
formation. One early notion held that each memory
HIPPOCAMPUS, AND NEOCORTEX; LONG-TERM
was stored in a particular protein molecule. This view
POTENTIATION; MEMORY CONSOLIDATION:
is no longer tenable, but proteins of course play key MOLECULAR AND CELLULAR PROCESSES;
roles in the structure and functioning of nerve cells. MEMORY CONSOLIDATION: PROLONGED
Another early view was that the brain grew new path- PROCESS OF REORGANIZATION;
ways; thus, in Pavlovian conditioning a new pathway MORPHOLOGICAL BASIS OF LEARNING AND
would grow to connect the conditioned stimulus re- MEMORY; NEURAL SUBSTRATES OF CLASSICAL
gion of the brain to the unconditioned stimulus or re- CONDITIONING; NEUROTRANSMITTER SYSTEMS
sponse region. This does not occur. Instead, evidence AND MEMORY; PREFRONTAL CORTEX AND
is uniformly consistent with the more modest view MEMORY IN PRIMATES; PROTEIN SYNTHESIS IN
that there are changes in the actions of the synapses LONG-TERM MEMORY IN VERTEBRATES; SECOND
MESSENGER SYSTEMS; SPATIAL LEARNING:
that are the sites of the interconnections and interac-
ANIMALS; SPATIAL MEMORY; WORKING
tions among the neurons of the brain. Changes in
MEMORY: ANIMALS
synaptic actions can occur in many ways: changes in
amount of neurotransmitter release, changes in re- Bibliography
ceptor molecules, and a variety of other biochemical Hebb, D. O. (1949). The organization of behavior. New York: Wiley.
processes, ranging from calcium entry into neurons Lashley, K. S. (1929). Brain mechanisms and intelligence. Chicago:
to second messenger systems (cyclic AMP, cyclic GNP, University of Chicago Press.
protein kinases, and so forth) to changes in gene ex- of Neuroscience 23, 155184.
pression. Maren, S. (2001). Neurobiology of Pavlovian fear conditioning.
Although the biophysical properties of synaptic Annual Review of Neuroscience 24, 897931.
Squire, L. R. (1986). Mechanisms of memory. Science 232, 1,612
plasticity have been extensively studied, definitive 1,619.
proof that these mechanisms actually mediate memo- Squire, L. R., Knowlton, B., and Musen, G. (1993). The structure
ry storage remains elusive. Perhaps the clearest evi- and organization of memory. Annual Review of Psychology 44,
dence that synaptic plasticity is a mechanism of mem- 453495.
Thompson, R. F. (1990). The neurobiology of learning and memo-
ory storage is in conditioned fear. Here, the evidence
ry. In K. L. Kelner and D. E. Koshland Jr., eds., Molecules to
strongly indicates that a strengthening (potentiation) models: Advances in neuroscience, 219234. Washington, DC:
of synaptic transmission in the amygdala is critically American Association for the Advancement of Science.
involved in memory formation. However, whether Thompson, R. F., and Kim, J. J. (1996). Memory systems in the
this form of plasticity mediates long-term storage of brain and localization of a memory. Proceedings of the National
Academy of Sciences of the United States of America 93, 13,438
the memory or whether it is simply an intermediate
13,444.
process in the long-term memory storage is unknown. Thompson, R. F., and Krupa, D. J. (1994). Organization of memo-
The strongest evidence we have for a biological sub- ry traces in the mammalian brain. Annual Review of Neuro-
strate of long-term memory storage concerns long- science 17, 519549.
lasting structural changes in the synaptic interconnec- Zola-Morgan, S., and Squire, L. R. (1993). Neuroanatomy of mem-
ory. Annual Review of Neuroscience 16, 547563.
tions among neurons. Enriched environments and
even particular learning experiences can result in Richard F. Thompson
changes in the numbers and distributions of spine Revised by David J. Krupa
LONG-TERM DEPRESSION 335
LONG-TERM DEPRESSION Brains of certain fish species contain a cerebel-

IN THE CEREBELLUM, lum-like structure that lacks climbing fibers. LTD oc-
curs in parallel fiber-evoked excitatory postsynaptic
HIPPOCAMPUS, AND NEOCORTEX
potentials (EPSPs) in Purkinje celllike neurons when
Long-term depression (LTD) is a type of synaptic EPSPs repeatedly precede postsynaptic spikes within
plasticity in which the efficacy of signal transmission 60 milliseconds (Bell, Han, Sugawara, and Grant,
across a synapse is persistently reduced after a certain 1997).
triggering activity. LTD in the cerebellum was pro-
In the Hippocampus
posed as a theoretical possibility around 1970 and was
detected a decade later (Ito, Sakurai, and Tongroach, In a pyramidal cell of the hippocampal CA1 re-
1982). So far, several subtypes of LTD varying in cel- gion, a low-frequency stimulation (LFS) of a bundle
lular and molecular mechanisms have been found in of presynaptic fibers (1 hertz for five to fifteen min-
the cerebellum, hippocampus, and neocortex. LTD utes) typically induces LTD, while a high-frequency
occurs not only in excitatory synapses, but also in in- stimulation (five stimuli at 100 hertz repeated at 200-
hibitory ones. LTD may weaken or functionally inter- millisecond intervals for two seconds) induces LTP.
rupt useless or erroneous synaptic connections be- Associative stimulation of presynaptic fibers with
tween neurons, providing an opposing mechanism postsynaptic membrane depolarization (5 hertz for
against long-term potentiation (LTP) in various sixteen seconds) also induces LTD or LTP depending
forms of learning and memory. on the time from pre- to postsynaptic activities (Nishi-
yama et al., 2000). While LTP is induced when pre-
synaptic stimulation falls within the window from -2
Induction and Observation of LTD to 12 milliseconds after postsynaptic spikes, LTD oc-
The various forms of LTD revealed to occur in curs when presynaptic stimulation either precedes the
the cerebellum, cerebellum-like structures in fish, postsynaptic spikes by 16 to 28 milliseconds or follows
hippocampus and neocortex have the following char- with a delay of 15 to 20 milliseconds (see Figure 2).
acteristic features. In this article, references for the Associative LTD occurs in both stimulated (homo-
cerebellum are largely omitted because they can be synaptic LTD) and unstimulated (heterosynaptic
found in recent review articles (Ito, 2001; Hansel, LTD) synapses, while LTP occurs only homosynapti-
Linden, and DAngelo, 2001). cally. As revealed in a triple chain of cultured hippo-
In the Cerebellum and Cerebellum-Like campal neurons, induction of associative LTD in a
Structures synapse is accompanied by back-propagated induc-
tion of LTD in a synapse on a presynaptic cell (Fitzsi-
In the cerebellar cortex each Purkinje cell re-
monds, Song, and Poo, 1997). Through these LTD/
ceives two distinct types of excitatory synapses, one
LTP inductions, temporal information coded in the
from parallel fibers and the other from a (normally
timing of individual spikes may be converted into spa-
single) climbing fiber. LTD occurs when these two
tially distributed patterns of persistent synaptic modi-
types of synapses are activated repeatedly in approxi-
fications in a neural network (Bi and Poo, 1999).
mate synchrony, leading to an enduring decrease in
synaptic strength of the parallel fibers. In in vitro cer- In the Neocortex
ebellar slices, LTD is typically induced by conjunctive LFS-induced LTD occurs in layer III neurons of
stimulation at 1 hertz for five minutes (300 stimuli), the visual cortex following stimulation of the white
which reduces the transmission efficacy by 30 to 40 matter or layer IV (Kirkwood and Bear, 1994). Asso-
percent (see Figure 1). To induce LTD, climbing fiber ciative LTP or LTD is induced when presynaptic stim-
stimulation is often replaced by current-induced ulation is paired with membrane depolarization. A
membrane depolarization that enhances entry of relatively large membrane depolarization induces
Ca2+ ions into Purkinje cell dendrites as in the case LTP, while a relatively small membrane depolariza-
of climbing fiber stimulation (see below). Conjunctive tion causes LTD (Artola, Brocher, and Singer, 1990).
LTD can be followed for one to three hours without This dichotomy depends on the amount of Ca2+ in-
sign of recovery. It is robust in the sense that its induc- flux evoked by the depolarization (see below). LTD
tion does not depend critically on the timing between has also been observed in the sensorimotor (Bind-
climbing fiber and parallel fiber stimulations. man, Murphy, and Pockett, 1988) and prefrontal cor-
Strong stimulation of parallel fibers alone causes tex (Hirsch and Crepel, 1990).
homosynaptic LTD. Conjunctive LTD is accompa-
nied by LTD in the neighboring synapses located
within a distance of 100 micrometers, which are not Signal Transduction Underlying LTD
involved in conjunctive stimulation. Other types of Conjunctive LTD is a purely postsynaptic event,
LTD also occur in the cerebellum. but in other LTD subtypes, the contribution of pre-
336 LONG-TERM DEPRESSION
Figure 1
LTD observed in Purkinje cells in cerebellar slices. Left, inset, EPSPs induced by stimulation of parallel fibers (PF-EPSP). Six traces of
EPSPs recorded before (control) and 10 to 50 minutes after conjunctive stimulation of parallel fibers and a climbing fiber are
superposed. Each trace indicates EPSPs averaged for 5 sweeps repeated every 5 seconds. The graph plots the rising slopes of the EPSPs
against time. Shaded column indicates conjunctive stimulation. Right, a and b, climbing fiber-evoked EPSP superposed with Ca2+
spikes (CF-EPSP), recorded at the moments indicated in the graph. a+b, superposition of a and b on an expanded time scale.
synaptic factors is not excluded. The following post- climbing fiber synapses contain AMPA receptors,
synaptic signal transduction processes have been ana- type-1 corticotropin-releasing factor receptors
lyzed in synaptically induced as well as in various (CRFR1s), and type-1 insulin-like growth factor re-
reduced forms of LTD induced with chemical or elec- ceptors (IGF-1Rs). Inhibition of any of these recep-
trical stimulation of neurons in place of synaptic acti- tors results in the blockage of LTD. The 2 receptor
vation (Ito, 2001). is an orphan receptor with an unknown function.
mGluR1s (see below), CRFR1s, and IGF-1Rs are asso-
Receptors ciated with G-proteins which are coupled with certain
LTD in excitatory synapses is, at least in part, due second messenger processes.
to a reduced number of functional AMPA receptor
molecules in the postsynaptic membrane. Other types Calcium Entry and Release from
of receptors play roles in eventually inducing this Intracellular Stores
change. Activation of NMDA receptors is required in Induction of conjunctive LTD requires enhance-
LTD induction in excitatory synapses, but not in in- ment of the intracellular Ca2+ concentration, for it is
hibitory synapses, of hippocampal neurons (Fitzsi- blocked by injection of a Ca2+ chelator, EGTA, into
monds, Song, and Poo, 1997). NMDA receptors are Purkinje cells. Climbing fiber impulses evoke Ca2+ in-
not functional in adult Purkinje cells. Parallel fiber flux into dendrites of Purkinje cells through voltage-
synapses on Purkinje cells contain, besides AMPA re- sensitive Ca2+ channels. As underlying homosynap-
ceptors, type-1 metabotropic glutamate receptors tic LTD, Ca2+ influx also occurs in association with
(mGluR1s) and 2 glutamate receptors, whereas parallel fiber-induced EPSPs, if the EPSPs are suffi-
LONG-TERM DEPRESSION 337
ciently large to activate voltage-sensitive Ca2+ chan- Figure 2

nels. Ca2+ ions are also released from intracellular
stores in endoplasmic reticula, when inositol trisphos-
phate (IP3) receptors or ryanodine receptors on the
reticula are activated. Inhibition of IP3 receptors, de-
pletion of Ca2+ stores, or genetically induced loss of
endoplasmic reticulum in dendritic spines results in
the blockage of conjunctive LTD. However, intracel-
lular Ca2+ release is not required for the reduced
form of LTD in isolated Purkinje cells.
In hippocampal and neocortical neurons, LTD
induction depends on the membrane potential, which
determines the entry of Ca2+ ions into neurons
through cation channels associated with a NMDA re-
ceptors. These channels are normally blocked by
Mg2+ ions that are removed at a depolarized mem-
brane potential level. Modest and strong activations
of NMDA receptors lead to LTD and LTP, respective-
ly. Reduction of postsynaptic Ca2+ entry by partial
blockage of NMDA receptors converts LTP to LTD
(Nishiyama et al., 2000). LTP is also converted to
LTD by injecting EGTA into neocortical neurons (Ki-
mura, Tsumoto, Nishigori, and Yoshimura, 1990). In-
duction of homosynaptic and heterosynaptic LTD in
the hippocampus requires functional ryanodine and
IP3 receptors, respectively. Functional blockade or
genetic deletion of type-1 IP3receptors leads to a con-
version of LTD to LTP and elimination of hetero-
synaptic LTD, while blockage of ryanodine receptors
eliminated only homosynaptic LTD (Nishiyama et al.,
2000).
Metabotropic Glutamate Receptor-Driven

Processes
Activation of mGluR1s in Purkinje cells results in
activation of phopholipase C (PLC), which generates
IP3 (see above) and diacylglycerol (DAG) from mem-
brane phospholipid. DAG activates PKC. mGluR1s
could also be coupled with phospholipase A2 (PLA2),
which produces arachidonic acid and oleic acid from Critical time windows for the induction of LTP/LTD in the
membrane phospholipid. Inhibition of mGluR1s, hippocampus by correlated pre- and postsynaptic activation. The
PKC, PLA2 or receptors of IP3 prevents LTD from in- pre-post time interval refers to the time between the onset of the
duction, indicating that these are involved in LTD in- EPSC and the peak of the postsynaptic action potential during
each correlated activation as indicated by arrows in the records
duction. Application of an mGluR5 agonist, DHPG,
on the top. In a and b, normalized mean EPSC amplitudes
effectively induces LTD in hippocampal CA1 neurons
measured at 30 to 40 minutes after correlated activation are
(Fitzjohn, Kingston, Lodge, and Collingridge, 1999), plotted against the pre-post time intervals. a) Summary of
which has been studied as a reduced form of hippo- changes in the homosynaptic pathway (n=50). b) Summary of
campal LTD. changes in the heterosynaptic (unstimulated) pathway (n=34).
Data are from a subset of experiments such as those shown in a,
Nitric Oxide and cGMP in which the control pathway was monitored throughout the
Cerebellar granule cells contain a neuronal type experiment.
of nitric oxide synthase (nNOS), and parallel fiber
terminals release NO upon stimulation. NO, so re-
leased, diffuses into Purkinje cells, and activates protein kinase (PKG). The role of NO in the induc-
guanylate cyclase to increase the level of cyclic GMP tion of conjunctive LTD is evident because it is
(cGMP). cGMP in turn activates cGMP-dependent blocked when an inhibitor of NO synthase, L-NMMA,
338 LONG-TERM DEPRESSION
or hemoglobin (HB) which absorbs NO, is applied to Phosphorylation and Inactivation of AMPA
cerebellar slices, or in those mice deficient of nNOS. Receptors
Furthermore, sodium nitroprusside that releases NO Among subunits constituting AMPA receptors,
or a membrane-soluble derivative of cGMP induces the GluR2 subunits are dominant in parallel fiber-
an LTD-like phenomenon when applied to cerebellar Purkinje cell synapses. Evidence suggests that at the
slices together with AMPA. However, there is no evi- end stage of cerebellar conjunctive LTD, GluR2 is
dence that a reduced form of LTD in isolated Purkin- phosphorylated at its serine-880 anchorage to GRIP,
je cells requires a NO-cGMP pathway. and freed AMPA receptors are removed from post-
synaptic membrane by internalization via endocyto-
Protein Kinases and Phosphatases
sis. Similar consideration applies to associative LTD
Induction of cerebellar conjunctive LTD requires in hippocampal neurons (Luscher, Nicoli, Malenka,
activities of PKC, PKG, protein tyrosin kinase (PTK) and Muller, 2000; Man et al., 2000).
and mitogen-activated protein kinase (MAPK) but not
PKA. PKC phosphorylates serine-880 of AMPA recep-
tors (see below), and PTK interacts with PKC. PKG Functional Roles of LTD
acts on a PKG-specific substrate, G-substance, which In the cerebellum-like structure of fish, LTD
is richly contained in Purkinje cells. The phosphoryl- plays a clear role of lessening the effects of sensory in-
ated G-substrate was found to be a potent inhibitor of puts coinciding with command signals in order to
protein phosphatases, presumably type 2A. Inhibitors allow only unpredicted sensory inputs to stand out,
of PP1/2A induce LTD when combined with stimula- thus forming an adaptive sensory processor. Involve-
tion of AMPA receptors. ment of LTD in certain forms of cerebellar learning
PKC and PTK, but not PKA, are involved in has been demonstrated using pharmacological or ge-
mGluR-induced LTD in neurons of hippocampal netic means that impair LTD (Ito, 2001). Adaptation
dentate gyrus (Camodeca, Breakwell, Rowan, and of reflexive eye movements was abolished by applying
Anwyl, 1999). A Ca2+/ Calmodulin-dependent pro- an NO scavenger or transfecting a pseudo inhibitor
tein phosphatase, calcineurin (or PP2B), mediates of PKC to the cerebellum. An NO scavenger or NOS
LTD induction in the hippocampus because a cal- inhibitor also blocked adaptation in smooth pursuit
cineurin inhibitor, FK506, blocks LFS induction of eye movement to repeated sudden increases in the ve-
LTD in the visual cortex (Torii, Kamishita, Otsu, and locity of the moving spotlight. Mice deficient in glial
Tsumoto, 1995). Mice deficient in the calcineurin cat- fibrillary acidic protein or phospholipase C 4 (Mi-
alytic unit B1 exhibited significantly diminished LFS- yata et al., 2001) lacked LTD and did not exhibit the
induced LTD in the hippocampus (Zeng et al., 2001). classical conditioning of eye blinking. A walking cat
This is in contrast to the observation in Purkinje cells or mouse normally adapted to a sudden change in the
in which inhibition of PP2A facilitates conjunctive running belt conditions, but not under the influences
LTD. Calcineurin is also involved in LTD induction of NO scavenger or NOS inhibitor or mGluR1 defi-
downregulating GABAA receptors in inhibitory syn- ciency, which block LTD.
apses of the hippocampus (Lu, Mansuy, Kandel, and Mice in which hippocampal LTD was diminished
Roder, 2000). due to conditional knockout of calcineurin were spe-
cifically impaired in working memory and episodic-
Protein Synthesis
like memory tasks, including the delayed matching-
Translational inhibitors depressed the late phase to-place task and the radial maze task (Zeng et al.,
of LTD in cultured Purkinje cells, but they abolished 2001).
the entire LTD including its early phase in cerebellar
slices. Studies using five-minute pulses of translation- Accumulating evidence indicates that LTD is a
al inhibitors revealed that a quickly turned over pro- major type of synaptic plasticity that plays roles in var-
tein synthesis was required for LTD induction only ious forms of learning and memory. It will be a future
during the fifteen-minute period from the onset of task to confirm whether LTD is converted to a perma-
conjunctive stimulation (Karacho et al., 2001). A five- nent memory, and if so, by what mechanism is it con-
minute pulse application of transcriptional inhibitors verted. Researchers should also determine the specif-
also effectively blocked LTD induction with a delay of ic functional roles of LTD in various forms of learning
thirty minutes. Protein synthesis also plays roles in and memory.
hippocampal neurons, in which a translational inhibi- See also: GLUTAMATE RECEPTORS AND THEIR
tor blocked the late phase of LFS-induced LTD CHARACTERIZATION; GUIDE TO THE ANATOMY
(Kauderer and Kandel, 2000) as well as the entire OF THE BRAIN; NEURAL COMPUTATION:
course of DHPG-induced LTD (Huber, Kayser, and CEREBELLUM; NEURAL SUBSTRATES OF
Bear, 2000). CLASSICAL CONDITIONING: DISCRETE
LONG-TERM POTENTIATION 339
BEHAVIORAL RESPONSES; SECOND MESSENGER Man, H. Y., Lin, J. W., Ju, W. H., Ahmadian, G., Liu, L., Becker,
SYSTEMS; VESTIBULO-OCULAR REFLEX (VOR) L. E., Sheng, M., and Wang, Y. T. (2000). Regulation of
PLASTICITY AMPA receptor-mediated synaptic transmission by clathrin-
dependent receptor internalization. Neuron 25, 649662.
Bibliography Miyata, M., Kim, H.-T., Hashimoto, K., Lee, T.-K., Cho, S.-Y.,
Artola, A., Brocher, S., and Singer, W. (1990). Different voltage- Jiang, H., Wu, Y., Jun, K., Kano, M., and Shin, H.-S. (2001).
dependent thresholds for inducing long-term depression and Deficient long-term synaptic depression in the rostral cerebel-
long-term potentiation in slices of rat visual cortex. Nature lum correlated with impaired motor learning in phospholi-
347, 6972. pase C b4 mutant mice. European Journal of Neuroscience 13,
Bell, C. C., Han, V. Z., Sugawara, Y., and Grant, K. (1997). Synaptic 111.
plasticity in a cerebellum-like structure depends on temporal Nishiyama, M., Hong, K., Mikoshiba, K., Poo, M. M., and Kato, K.
order. Nature 387, 278281. (2000). Calcium stores regulate the polarity and input speci-
Bi, G-Q., and Poo, M-m. (1999). Distributed synaptic modification ficity of synaptic modification. Nature 408, 584588.
in neural networks induced by patterned stimulation. Nature Torii, N., Kamishita, T., Otsu, Y., and Tsumoto, T. (1995). An in-
401, 792796. hibitor for calcineurin, FK506, blocks induction of long-term
Bindman, L. J., Murphy, K. P. S., and Pockett, S. (1988). Postsyn- depression in rat visual cortex. Neuroscience Letters 185, 14.
aptic control of the induction of long-term changes in efficacy Zeng, H., Chattarji, S., Barbarosie, M., Rondi-Reig, L., Philpot, B.
of transmission at neocortical synapses in slices of rat brain. D., Miyakawa, T., Bear, M. F., and Tonegawa, S. (2001). Fore-
Journal of Neurophysiology 60, 1,0531,065. brain-specific calcineurin knockout selectively impairs bidi-
Camodeca, N., Breakwell, N. A., Rowan, M. J., and Anwyl, R. rectional synaptic plasticity and working/episodic-like memo-
(1999). Induction of LTD by activation of group I mGluR in ry. Cell 107, 617629.
the dentate gyrus in vitro. Neuropharmacology 38, 1,5971,606.
Fitzjohn, S. M., Kingston, A. E., Lodge, D., and Collingridge, G.
L. (1999). DHPG-induced LTD in area CA1 in juvenile rat
Masao Ito
hippocampus; characterization and sensitivity to novel mGlu
receptor antagonists. Neuropharamacology 38, 1,5771,583.
Fitzsimonds, R. M., Song, H-j., and Poo, M-m. (1997). Propagation
of activity-dependent synaptic depression in simple neural
networks. Nature 388, 439448.
Hansel, C., Linden, D., and DAngelo, E. (2001). Beyond parallel LONG-TERM MEMORY
fiber LTD: The diversity of synaptic and non-synaptic plastici-
ty in the cerebellum. Nature Neuroscience 4, 467475. See: APLYSIA: MOLECULAR BASIS OF LONG-TERM
Hirsch, J. C., and Crepel, F. (1990). Use-dependent changes in SENSITIZATION; KNOWLEDGE SYSTEMS AND
synaptic efficacy in rat prefrontal neurons in vitro. Journal of MATERIAL-SPECIFIC MEMORY DEFICITS;
Physiology (London) 427, 3149. LONG-TERM DEPRESSION IN THE
Huber, K. M., Kayser, M. S., and Bear, M. F. (2000). Role for rapid CEREBELLUM, HIPPOCAMPUS, AND
dendritic protein synthesis in hippocampal mGluR- NEOCORTEX; LONG-TERM POTENTIATION;
dependent long-term depression. Science 288, 1,2541,256. PROTEIN SYNTHESIS IN LONG-TERM MEMORY
Ito, M. (2001). Long-term depression: Characterization, signal
IN VERTEBRATES
transduction and functional roles. Physiology Review 81, 1,143
1,195.
Ito, M., Sakurai, M., and Tongroach, P. (1982). Climbing fibre in-
duced depression of both mossy fibre responsiveness and glu-
tamate sensitivity of cerebellar Purkinje cells. Journal of Physi-
ology (London) 324, 113134.
Karachot, L., Shirai, Y., Vigot, R., Yamamori, T., and Ito, M. LONG-TERM POTENTIATION
(2001). Induction of long-term depression in cerebellar Pur-
kinje cells requires a quickly turned over protein. Journal of [Long-term potentiation (LTP) is defined as a persistent en-
Neurophysiology 86, 280289. hancement in the strength of a synaptic connection produced
Kauderer, B. S., and Kandel, E. R. (2000). Capture of a protein syn-
thesis-dependent component of long-term depression. Pro- as a result of delivering a brief high frequency burst of neu-
ceedings of the National Academy of Sciences of the United States of ral activity (i.e., tetanus) to a presynaptic neuron or path-
America 97, 13,34213,347. way. The enhanced synaptic efficacy can persist for hours,
Kimura, F., Tsumoto, T., Nishigori, A., and Yoshimura, Y. (1990). days, or even weeks, depending on the stimulus protocol.
Long-term depression but not potentiation is induced in
Ca2+-chelated visual cortex neurons. Neurological Report 1, LTP is believed to be a prime candidate for a synaptic memo-
6568. ry mechanism because of its persistence and the fact that it
Kirkwood, A., and Bear, M. F. (1994). Homosynaptic long-term de- is found in brain regions that have been implicated in memo-
pression in the visual cortex. Journal of Neuroscience 14, 3,404 ry (e.g., the amygdala, hippocampus, cerebral cortex, and
3,412.
Lu, Y. M., Mansuy, I. M., Kandel, E., and Roder, J. (2000). Cal- cerebellum). Five entries are devoted to this important mech-
cineurin-mediated LTD of GABAergic inhibition underlies anism. They include an OVERVIEW, discussions of the dif-
the increased excitability of CA1 neurons associated with ferent forms and mechanisms of LTP in different brain re-
LTP. Neuron 26, 197205. gions, and BEHAVIORAL ROLES of LTP. The reader should
Luscher, C., Nicoli, R. A., Malenka, R. C., and Muller, D. (2000).
Synaptic plasticity and dynamic modulation of the postsynap- also see the entry on the related phenomenon of LONG-TERM
tic membrane. Nature Neuroscience 3, 545550. DEPRESSION.]
340 LONG-TERM POTENTIATION: Overview: Cooperativity and Associativity
OVERVIEW: COOPERATIVITY AND existence of Hebb synapses. This requirement for

ASSOCIATIVITY postsynaptic depolarization of sufficient strength to
evoke postsynaptic spike activity explains cooperativi-
Long-term potentiation (LTP) is the collective name for ty. The requirement for coincident presynaptic activi-
synaptic plasticity processes in which brief (less than ty explains associativity.
one second) episodes of intense synaptic activity lead
to an enhancement of synaptic efficacy lasting hours In 1983, G. L. Collingridge and colleagues
to weeks, or longer. LTP in many regions has associa- showed that blocking another glutamate receptor, the
tive induction properties based on a requirement for NMDA receptor, prevented LTP induction. In the
coincident presynaptic and postsynaptic activity. This same year Dingledine described that NMDA channels
requirement is similar to that proposed on theoretical permeate calcium ions, and G. Lynch and colleagues
grounds by Donald Hebb in 1949 for a synaptic modi- found that a rise in postsynaptic calcium was neces-
fication involved in learning and memory, known as sary for LTP induction. Moreover, several groups
the Hebb synapse. Because of its longevity and asso- showed that the NMDA receptor was coupled to a
ciative induction properties, scientists regard LTP as voltage-sensitive channel; that is, it needed, in addi-
the prime neuronal model for learning and memory tion to glutamate, membrane depolarization to open.
(see Figure 1). These results, together with the strict dependence of
the induction on coincident presynaptic and postsyn-
Scientists first described LTP in the rabbit hippo- aptic activity, led H. Wigstrm and B. Gustafsson in
campal formation, and later in the neocortex and a 1985 to propose that LTP is initiated as a conse-
variety of other regions in the vertebrate (including quence of calcium influx through NMDA receptor
human) nervous system. It has mostly been found for channels co-localized with the AMPA receptors on the
excitatory synapses in principal cells. These synapses postsynaptic spine membrane. The NMDA receptor
are typically spine synapses; that is, located on small would act as a coincidence (and cooperativity) detec-
protuberances (spines) on the dendrite. Excitatory actor because of its need for both transmitter binding
tion in these synapses is mediated by glutamate acting and membrane depolarization for activation. The co-
on AMPA receptors located postsynaptically, and localization of the NMDA and AMPA receptor chan-
LTP is seen as an increase of this AMPA receptor- nels on the subsynaptic spine membrane, together
mediated transmission. with a restricted localization of the rise in postsynap-
tic calcium due to the spine location of the synapse,
would secure input specificity for LTP.
Associative LTP: A History
T. Lo mo reported in a short note in 1966 that a
few seconds of repetitive synaptic activation (10 to 15 LTP Induction: Modulation of Threshold
hertz) led to a prolonged potentiation (LTP) of excit- Although many studies use unnatural stimulus
atory action on granule cells in the dentate gyrus of conditions to induce LTP, such as a one-second syn-
the hippocampus. In 1973, T. V. P. Bliss and A. R. chronous activation of afferents at high frequency,
Gardner-Medwin reported that LTP could last for physiological stimulus patterns have also been shown
weeks and that its induction appeared to depend on to be effective in inducing it. Brief burst stimulation
the number of activated presynaptic fibers, or rather that simulates the 5 to 7 hertz hippocampal EEG wave
the number of activated synapses. This dependence activity (theta rhythm) effectively produces LTP. This
on the co-activation of many presynaptic fibers (coop- is because this stimulus pattern depresses inhibitory
erativity), and the restriction of LTP to the synapses circuits and thereby enables a brief burst to produce
of the activated fibers (input specificity), were demon- substantial postsynaptic depolarization. High fre-
strated by several laboratories in the late 1970s. LTP quency presynaptic activation is not necessary for
induction was also found to be associative: A brief acti- LTP induction. LTP can be induced in synapses acti-
vation of a weak synaptic input that did not induce vated by single stimuli at low frequency (such as 0.2
LTP by itself did so when occurring in close temporal hertz) provided that this activation is associated with
contiguity with brief activation of a separate strong sufficient postsynaptic depolarization. Experimental-
synaptic input to the same target neurons. Thus, LTP ly this depolarization can be provided by current-
can form an associative connection between a weak induced action potentials in the cell body that either
input and a strong one or, alternatively seen, an asso- passively spread or actively back-propagate into the
ciation between a weak input and the response elicit- dendritic tree where the synapses are located. Under
ed by the strong one. In the mid-1980s several groups physiological conditions when action potential activi-
showed that LTP induction requires coincident spike ty is generated by synaptic excitation, the necessary
activity of the presynaptic terminal and of the target depolarization will in addition be provided by the
postsynaptic neuron; that is, they demonstrated the passive spread from active synapses. The synaptic ex-
LONG-TERM POTENTIATION: Overview: Cooperativity and Associativity 341
citation may actually play a triple role for generating Figure 1

the necessary conditions for LTP induction. First, by
generating local depolarization in the synaptic re-
gion. Second, by generating action potentials in the
cell body region that may back-propagate into the
dendrite. Third, by its depolarization of the dendritic
membrane facilitate back-propagation into the active
dendritic region.
The threshold for inducing LTP is controlled by
a number of factors. These include metaplasticity,
neuromodulators, and pharmacological agents that
all, directly or indirectly, interfere with the opening
of NMDA receptor channels. Metaplasticity (plasticity
of synaptic plasticity), a concept introduced by W. C.
Abraham in 1995, is a change in induction threshold
of LTP induced by prior synaptic, or cellular, activity,
that is not necessarily expressed as change in efficiacy
of normal synaptic transmission. One possible mech-
anism would be a prolonged activity-dependent
change in NMDA receptor-mediated signaling, such
as a LTP of that signaling. In fact, such activity-
dependent changes have been reported. Neuro-
modulators, such as norepinephrine, acetylcholine,
serotonin, dopamine, glutamate (via its metabotropic
receptors), and neuropeptides regulating attention,
motivation, emotion, and wakefulness, will all modu-
late LTP induction. As these neuromodulators exert
their neuromodulatory role by their action on synap-
tic transmission and on cellular excitability, the
NMDA receptor dependent requirements for LTP in- Schematic representation of the model for associative LTP
duction will concurrently be affected. Similarly, phar- induction. A presynaptic terminal (left) releases the transmitter
macological agents that affect excitatory or inhibitory glutamate (dots) onto a spine of a pyramidal cell dendrite, on
synaptic transmission, or cellular excitability, will which AMPA and NMDA types of glutamate receptor channels
alter the threshold for LTP induction. For example, are co-localized. Whereas AMPA receptor channels are opened
drugs that enhance inhibitory transmission such as by the transmitter alone, NMDA receptor channels require in
ethanol and benzodiazepines will impair the genera- addition coincident membrane depolarization. This dual
tion of LTP. requirement for the opening of NMDA receptor channels
explains associative induction.
LTP Expression
Independently of the induction intensity (above posed to rely on synthesis of new proteins whereas
threshold) LTP is established with a similar time E-LTP relies on modification of pre-existing ones.
course and within less than a minute. Its duration is,
however, variable, depending on the intensity and/or What aspect of synaptic transmission is actually
the repetition of the synaptic activity. Thus, following modified in NMDA receptor-dependent LTP? This
weak activation it may decay within minutes whereas issue has been intensively debated since the mid-
an intense activation can lead to a LTP lasting weeks. 1980s. Despite numerous studies, the question of
The more prolonged LTP given by more intense syn- whether there is a more efficient release of glutamate,
aptic activation may not only be explained by a great- or an increase in AMPA receptor channel number or
er extent of NMDA receptor activation but also by the efficiency, is still unsettled. A complicating factor may
concurrent activation of voltage-gated calcium chan- be that NMDA receptor-dependent LTP is based on
nels and/or other receptor/channel systems. Based on different expression mechanisms dependent on ex-
the effect of protein synthesis inhibitors, LTP has perimental conditions, for example cultured versus
been divided in an early phase (E-LTP, less than a few intact tissue, or on other factors such as brain region
hours), and a late LTP (L-LTP). L-LTP is thus sup- and animal age.
342 LONG-TERM POTENTIATION: Amygdala
Nonassociative LTP See also: GLUTAMATE RECEPTORS AND THEIR

CHARACTERIZATION; HEBB, DONALD
Long-term synaptic plasticity exists not only in
the form of NMDA receptor-dependent LTP but also Bibliography
as nonassociative forms not relying on NMDA recep- Abraham, W. C., and Bear, M. F. (1996). Metaplasticity: The plas-
tor activation. The foremost example of such NMDA ticity of synaptic plasticity. Trends in Neurosciences 19 (4), 126
receptor-independent LTP, first described by E. W. 130.
Bailey, C. H., Giustetto, M., Huang, Y. Y., Hawkins, R. D., and
Harris and C. W. Cotman in 1986, is that in the syn-
Kandel, E. R. (2000). Is heterosynaptic modulation essential
apse that connects granule cells in the dentate gyrus for stabilizing Hebbian plasticity and memory? Nature Reviews
with CA3 pyramidal neurons in the hippocampus. In Neuroscience 1 (1), 1120.
contrast to NMDA receptor-dependent LTP its ex- Gustafsson, B., and Wigstrom, H. (1988). Physiological mecha-
pression appears undisputed whereas its induction nisms underlying long-term potentiation. Trends in Neuro-
sciences 11 (4), 156162.
mechanism is debated. A number of studies by R. A.
Hanse, E., and Gustafsson, B. (1994). Onset and stabilization of
Nicoll and colleagues in the 1990s have established NMDA receptor-dependent hippocampal long-term potenti-
that NMDA receptor-independent LTP is presynapti- ation. Neuroscience Research 20 (1), 1525.
cally expressed as an increased efficiacy of transmitter Johnston, D., Williams, S., Jaffe, D., and Gray, R. (1992). NMDA-
release. These studies have also indicated its induc- receptor-independent long-term potentiation. Annual Review
of Physiology 54, 489505.
tion to be noncooperative, only related to presynaptic
Linden, D. J. (1999). The return of the spike: Postsynaptic action
calcium accumulation. On the other hand, studies by potentials and the induction of LTP and LTD. Neuron 22,
D. Johnston and colleagues have indicated a coopera- 661666.
tive induction, related to postsynaptic calcium influx Malenka, R. C., and Nicoll, R. A. (1999). Long-term potentia-
via voltage-gated calcium channels and release from tiona decade of progress? Science 285, 1,8701,874.
McBain, C.J., Freund, T. F., and Mody, I. (1999). Glutamatergic
internal calcium stores. A similar NMDA receptor-
synapses onto hippocampal interneurons: Precision timing
independent LTP is also described for the synapse without lasting plasticity. Trends in Neurosciences 22 (5), 228
that connects granule cells and Purkinje cells in the 235.
cerebellum. Sanes, J. R., and Lichtman, J. W. (1999). Can molecules explain
long-term potentiation? Nature Neuroscience 2 (7), 597604.
Bengt Gustafsson
LTP on Interneurons
Eric Hanse
Glutamatergic afferent fibers do not only make
excitatory synapses with principal cells, but also with
interneurons. Scientists have debated whether these AMYGDALA
interneuronal synapses exhibit LTP. It would appear
that LTP does not exist in most of these synapses. Long-term synaptic potentiation (LTP) has emerged
However, LTP with associative properties has been as the likeliest synaptic substrate for rapid associative
found in some interneuron types. Interestingly, this learning (Brown, Chapman, Kariss, and Keenan,
associative induction was found not to rely on NMDA 1988; Kelso and Brown, 1986). LTP has been report-
receptor activation as in the principal cells, indicating ed in numerous brain structures, including the amyg-
the existence of other mechanisms for coincidence dala, where it may participate in emotional learning
detection. (Blair et al., 2001; Chapman, Kairiss, Keenan, and
Brown, 1990). This article examines an unexpected
property of LTP in synapses of the amygdala and the
Conclusion implications of this finding on attempts to link amyg-
LTP denotes forms of synaptic plasticity with as- dalar LTP to emotional learning.
sociative as well as nonassociative induction (i.e., they
depend on temporal contiguity between activity in Amygdalar Neuroanatomy and
different pathways and on activity in a single pathway, Neurophysiology
respectively). Non-associative LTP will allow for a
more efficient transmission in specific pathways that Synaptic transmission and its plasticity seem to
are intensely used, irrespective of activity in others. depend jointly on the identity of the presynaptic and
Associative LTP on the other hand strengthens syn- postsynaptic neurons. A consideration of LTP in the
apses in a manner that relies on contiguity of activity amygdala thus requires some knowledge of the anato-
in those neurons that are connected via the modifi- my and physiology of the neurons in this brain re-
able synapse. This Hebbian modification rule is a gion. The functional neuroanatomy is also important
powerful device used in neural network models of for interpreting electrophysiological recordings.
nonsupervised learning to produce, for example, self- The amygdaloid nuclear complex is a collection
organizing capabilities. of subnuclei that fall into two major classes based on
LONG-TERM POTENTIATION: Amygdala 343
cell morphology: those containing cortexlike neurons (Chapman, Kairiss, Keenan, and Brown, 1990; Kee-
and those containing striatallike neurons (Swanson nan, Chapman, Chang, and Brown, 1988). The gross
and Petrovich, 1998). The cortexlike group includes cytoarchitecture of the slice was vividly revealed
the lateral, basal, and basolateral nuclei, sometimes through the use of differential-interference contrast
collectively termed the basolateral amygdaloid com- (DIC) optics and infrared (IR) illumination, which
plex (BLA). The more medially located nuclei, in- Brown and cowokers had been developing for this
cluding the central and medial nuclei, are part of the purpose (Keenan, Chapman, Chang, and Brown,
ventromedial expanse of the striatum. 1988). At low magnification, IR-DIC optics sharply
At a circuit level, the functional neuroanatomy of resolved the key landmarks needed to position the sti-
the amygdala remains a mystery, but there have been mulating and recording electrodes to repeatable loca-
a few studies of the anatomy and physiology of indi- tions that can be reliably matched to corresponding
vidual neurons (Chapman, Kairiss, Keenan, and plates of a rat brain atlas. Critically important were
Brown, 1990; Faulkner and Brown, 1999; Washburn the borders of EC, the basolateral nucleus, the lateral
and Moises, 1992). The organization of cells in BLA, nucleus, and the central nucleus (Chapman, Kairiss,
although different from that of any other brain re- Keenan, and Brown, 1990).
gion, perhaps resembles most closely a combination EC was selected as the placement site for the stim-
of the adjacent perirhinal cortex and the CA3 region ulation electrode for four reasons. First, it was the
of the hippocampus. While the unusual cell-firing only visually discernible large population of known
types are similar to those in the perirhinal cortex afferent inputs to BLA. Second, in contrast to other
(Faulkner and Brown, 1999; McGann, Moyer, and possible stimulation sites, this allowed a more homo-
Brown, 2001), the massive system of excitatory recur- geneous population of synapses, thus partaking of
rents and consequent tendency toward epileptiform some of the virtues of studies of the Schaeffer collater-
activity when pharmacologically disinhibited is remi- al/commissural (Sch/com) inputs to pyramidal neu-
niscent of the CA3 region of the hippocampus (Brown rons in hippocampal region CA1. Third, because EC
and Zador, 1990). contains fast-conducting fibers, these were presumed
Combined studies of BLA by Brown and co- to produce the earliest detectable postsynaptic re-
workers (Chapman, Kairiss, Keenan, and Brown, sponses. Fourth, neurophysiological results suggest-
1990; Faulkner and Brown, 1999) and Washburn and ed that EC-elicited PSCs in BLA neurons contain
Moises (1992) suggest that there may be as many as monosynaptic responses from the stimulation site.
seven firing patterns, including regular spiking, sin- Regarding this last point, the critical question was
gle spiking, late spiking, burst spiking, slow charging, not whether EC fibers project to BLAthey are
irregular spiking, and fast spiking. Small, aspiny stel- known to do so. The question was whether the PSCs
lates tend to be fast-spiking cells, presumably inhibi- recorded in BLA neurons in response to EC stimula-
tory interneurons. Pyramids and large stellates can be tion reflected direct synaptic inputs from stimulated
regular spiking, single spiking, burst spiking, or late EC fibers. The obvious alternative was that PSCs re-
spiking. Little is known about other structure- corded in response to EC stimulation were produced
function relationships. by inputs from other BLA neurons whose firing was
Spontaneous postsynaptic potentials (PSPs) and caused by EC stimulation. The difference can be im-
postsynaptic conductances (PSCs) are commonly portant for designing and interpreting neurophy-
much larger than those of the CA1 region of the hip- siology experiments (Xiang and Brown, 1998).
pocampus but similar to those in CA3 pyramidal neu- To minimize activation of recurrent circuitry and
rons (Faulkner and Brown, 1999; Johnston and to avoid epileptiform activity (Johnston and Brown,
Brown, 1984; Xiang and Brown, 1998). Electrical 1983; Xiang and Brown, 1998), the slices were not
stimulation of the external capsule (EC) commonly disinhibited pharmacologically. The excitatory com-
elicits in BLA neurons an excitatory-inhibitory con- ponent of the synaptic conductance waveform was
ductance sequence (Chapman, Kairiss, Keenan, and first evaluated (Griffith, Brown, and Johnston, 1986)
Brown, 1990), similar to that which Johnston and during baseline testing (thirty to ninety responses at
Brown and coworkers reported in hippocampal re- 0.1 Hz). An attempt was then made to induce LTP by
gion CA3 (Barrionuevo, Kelso, Johnston, and Brown, delivering three to ten trains (300 msec each) of high-
1986; Griffith, Brown, and Johnston, 1986; Johnston frequency (100 Hz, tetanic) electrical currents (each
and Brown, 1983). 0.1 msec) through the stimulation electrode that was
positioned in EC. Following tetanic stimulation, test-
Discovery of LTP in the Amygdala ing was resumed at 0.1 Hz for as long as two hours.
Browns team looked for LTP elicited by EC stim- The tetanic stimulation produced an early synap-
ulation in horizontal brain slices containing BLA tic enhancement that relaxed back to a sustained level
344 LONG-TERM POTENTIATION: Amygdala
of synaptic potentiation. Because of posttetanic po- that NMDARs are only part of the story behind LTP
tentiation (PTP), the early enhancement lasted less induction in the hippocampus (Magee, Hoffman,
than fifteen minutes, as in the hippocampus (Barrio- Colbert, and Johnston, 1998). Recent work by John-
nuevo, Kelso, Johnston, and Brown, 1986; Kelso and ston and coworkers (Magee and Johnston, 1997)
Brown, 1986; Kelso, Ganong, and Brown, 1986). LTP demonstrated the critical importance of backpropa-
was measured conventionally as the stable enhance- gating Ca2+ spikes in controlling the induction of one
ment that clearly outlasted PTP. The tetanic stimula- form of LTP in the Sch/com synaptic input to CA1 py-
tion protocol induced LTP in 80 percent of the neu- ramidal neurons. Blocking L-type Ca2+ channels pre-
rons studied, which averaged a 46 percent increase. vented or greatly inhibited LTP induction in these py-
ramidal neurons (Magee and Johnston, 1997). In a
parallel finding, Teyler and coworkers (Cavus and
Kinds of LTP in the Amygdala Teyler, 1996) have shown that the Sch/com synapses
Two obvious questions arise from the countless on CA1 pyramidal neurons can undergo two forms of
studies of LTP in the hippocampus. The first question LTP: one that is dependent on NMDARs and another
is whether there is a kind of LTP in the amygdala that depends on L-type Ca2+ channels.
whose induction depends on the activation of N-
methyl-d-aspartate receptors (NMDARs). NMDAR- Teyler and coworkers noted that NMDAR-
dependent LTP in the Sch/com input to CA1 is one dependent LTP and Ca2+ channel-dependent LTP
of the most commonly studied forms of LTP. There engage distinct signal transduction cascades and thus
is not much neurophysiological evidence for a similar could have different functional significance for en-
kind of plasticity in the amygdala. coding (Cavus and Teyler, 1996; Morgan, Coussens,
and Teyler, 2001). Teyler hypothesized that
Chapman and Bellavance (1992) found that APV
NMDAR-dependent LTP might be more critically
(50 M), a competitive antagonist for glutamate at the
involved in short-term memory, whereas Ca2+-
NMDAR, did not block LTP induction produced by
channel-dependent LTP could be more important
tetanic stimulation of EC. These investigators report
in long-term memory. This suggestion was based
that LTP induction could be inhibited by APV when
partly on findings by Teyler and coworkers (Mor-
the concentration was increased to 100 M, but this
concentration also strongly attenuated EC-evoked gan, Coussens, and Teyler, 2001) that Ca2+-
PSPs. The recording methods used by Chapman and channel-dependent LTP has a slower onset, it is more
Bellavance presumably tended to sample from a pop- persistent, and it is less subject to elimination by long-
ulation of larger BLA neurons, but other researchers term depression (LTD)at least in brain slices.
have reported similar results in BLA using selection Some additional findings from the previously
criteria designed to sample small interneurons. mentioned study by LeDoux and coworkers
Mahanty and Sah (1998) examined LTP in a sub- (Weisskopf, Bauer, and LeDoux, 1999) are relevant to
set of cells that have fast-action potentials and that Teylers hypothesis. Recall that LeDoux and co-
lack spike-frequency adaptation. EC stimulation (two workers examined LTP in a pathway to the lateral nu-
trains of 100 stimuli presented at 30 Hz) resulted in cleus by stimulation of a region of the striatum that
LTP in seven of seven of these putative interneurons. is near the central nucleus of the amygdala. Although
Pharmacological and electrophysiological analysis they found no effect of bath application of 50 M APV
suggested that NMDARs contribute relatively little to on LTP induction, bath application of the L-type
the EC-evoked PSCs. Application of APV (50 M) had Ca2+ channel blocker, Nifedipine (30 M), prevented
no effect on LTP induction in these putative interneu- LTP.
rons.
Results from the handful of studies that have
LeDouxs team (Weisskopf, Bauer, and LeDoux, been done on the neuropharmacology of LTP in BLA
1999) reported similar results in a previously unex- neurons do not seem to fit the pattern seen in the
plored synaptic circuit. They positioned a stimulating most commonly studied form of LTP that can be in-
electrode in a region of the striatum immediately dor- duced in the Sch/com synaptic input to CA1 pyra-
sal to the central nucleus and medial to the lateral nu- mids. With the Sch/com inputs, 100 M APV effective-
cleus, where they recorded evoked PSPs. LTP induc- ly prevents LTP induction but has no detectable effect
tion was unaffected when 50 M APV was added to on evoked PSPs and no detectable effect on the ex-
the bathing medium. pression of previously induced LTP (see Brown et al.,
The second question to emerge from research on 1989). NMDARs play a key role in inducing this form
hippocampal LTP concerns the role of backpropagat- of LTP in Sch/com synapses (Brown, Chapman,
ing dendritic Ca2+ spikes (Zador, Koch, and Brown, Kairiss, and Keenan, 1988). However, under normal
1990). Research in Johnstons laboratory has shown experimental conditions, NMDARs are not required
LONG-TERM POTENTIATION: Amygdala 345
for evoking PSPs or for expressing this form of LTP calized in the same synapses. Since the discovery of
after it has been established. amygdalar LTP by Brown and coworkers (Chapman,
Kairiss, Keenan, and Brown, 1990) it has become
clear that this form of LTP can be induced in the pres-
Behavioral Implications of Amygdalar LTP ence of APV unless the concentration is high enough
Information about amygdalar synapses has obvi- to interfere with experimentally evoked PSPs. This
ous implications for studies on the role of the amyg- form of LTP is especially interesting in the context of
dala in emotional learning. In one set of conditioning Teylers hypothesis that APV-resistant LTP is prefer-
experiments, Davis and coworkers (Campeau, Mis- entially involved in long-term memory, an idea that
erendino, and Davis, 1992; Miserendino, Sananes, could be relevant to the persistence of emotional
Melia, and Davis, 1990) perfused the amygdala with memory. Since1990, when LTP was first observed in
APV and claimed to have shown that the drug affects amygdala brain slices, the reliability and resolution of
learning but not ordinary synaptic transmission nor in vitro technology has matured enough to hasten
the ability to recall previous learning. Direct studies progress in subsequent research (Faulkner and
of amygdalar synapses, however, predict that infusing Brown, 1999; Moyer and Brown, 2002).
BLA with APV at a concentration that blocks amygda-
la-dependent learning could also impair retrieval of See also: GUIDE TO THE ANATOMY OF THE BRAIN:
previously established memories by interfering with AMYGDALA; NEURAL SUBSTRATES OF CLASSICAL
synaptic signalling and therefore access to memory. CONDITIONING: FEAR-POTENTIATED STARTLE;
NEURAL SUBSTRATES OF EMOTIONAL MEMORY
Other laboratories (Fendt, 2001; Lee, Choi,
Brown, and Kim, 2001; Lindquist and Brown, 2002) Bibliography
convincingly demonstrated that APV infusion directly Barrionuevo, G., Kelso, S. R., Johnston, D., and Brown, T. H.
into the amygdala does impair its functioning in a (1986). Conductance mechanism responsible for long-term
more general manner than had been previously potentiation in monosynaptic and isolated excitatory synaptic
inputs to hippocampus. Journal of Neurophysiology 55, 540
claimed by Davis and coworkers. As expected, infus- 550.
ing the amygdala with APV during the time of condi- Blair, H. T., Schafe, G. E., Bauer, E. P., Rodrigues, S. M., and Le-
tioning impaired learning (Lee, Choi, Brown, and Doux, J. E. (2001). Synaptic plasticity in the lateral amygdala:
Kim, 2001), as revealed by subsequent testing (done A cellular hypothesis of fear conditioning. Learning and Memo-
in the absence of APV). Contrary to claims by Davis ry 8, 22942.
Brown, T. H., Chapman, P. F., Kairiss, E. W., and Keenan, C. L.
and coworkers (Campeau, Miserendino, and Davis, (1988). Long-term synaptic potentiation. Science 242, 724
1992; Miserendino, Sananes, Melia, and Davis, 1990), 728.
these laboratories also found a profound memory im- Brown, T. H., Ganong, A. H., Kairiss, E. W., Keenan, C. L., and
pairment when the amygdala was perfused with APV Kelso, S. R. (1989). Long-term potentiation in two synaptic
at the time of testing (but not during the time of con- systems of the hippocampal brain slice. In J. H. Byrne and W.
O. Berry, eds., Neural models of plasticity. San Diego: Academic
ditioning). Press.
Multiple behavioral measures have shown that in- Brown, T. H., and Zador, A. M. (1990). The hippocampus. In G.
fusing BLA with APV profoundly impairs both the ac- Shepherd, ed., Synaptic organization of the brain. New York: Ox-
ford University Press.
quisition and expression of several classical, BLA- Campeau, S., Miserendino, M. J., and Davis, M. (1992). Intra-
dependent, conditioned-fear responses (CRs). Per- amygdala infusion of the N-methyl-D-aspartate receptor an-
haps even more surprising and dramatic was the tagonist AP5 blocks acquisition but not expression to an audi-
finding that infusing BLA with APV also greatly atten- tory conditioned stimulus. Behavioral Neuroscience 106, 569
uated or even eliminated unconditioned responses 574.
Cavus, I., and Teyler, T. J. (1996). Two forms of long-term potenti-
(URs) that are normally elicited during conditioning ation in area CA1 activate different signal transduction path-
trials. UR reactivity during conditioning predicted ways. Journal of Neurophysiology 76, 3,0383,047.
CR production during subsequent testing (Lee, Choi, Chapman, P. F., and Bellavance, L. L. (1992). Induction of long-
Brown, and Kim, 2001). The collective research on term synaptic potentiation in the basolateral amygdala does
LTP in amygdalar neurons correctly anticipated that not depend on NMDA receptor activation. Synapse 11, 310
318.
the effect of APV on amygdala function is much less Chapman, P. F., Kairiss, E. W., Keenan, C. L., and Brown, T. H.
specific than Davis and coworkers contended (Cam- (1990). Long-term synaptic potentiation in the amygdala.
peau, Miserendino, and Davis, 1992; Miserendino, Synapse 6, 271278.
Sananes, Melia, and Davis, 1990). Faulkner, B., and Brown, T. H. (1999). Morphology and physiolo-
gy of neurons in the rat perirhinal-lateral amygdala area.
Journal of Comparative Neurology 411, 613642.
Conclusion Fendt, M. (2001). Injections of the NMDA receptor antagonist
aminophosphonopentanoic acid into the lateral nucleus of
There are multiple forms of LTP/LTD that can be the amygdala block the expression of fear-potentiated startle
spatially segregated in different neurons or even colo- and freezing. The Journal of Neuroscience 21, 4,1114,115.
346 LONG-TERM POTENTIATION: Behavioral Roles
Griffith, W. H., Brown, T. H., and Johnston, D. (1986). Voltage- Zador, A., Koch, C., and Brown, T. H. (1990). Biophysical model
clamp analysis of synaptic inhibition during long-term poten- of a Hebbian synapse. Proceedings of the National Academy of Sci-
tiation in hippocampus. Journal of Neurophysiology 55, 767 ences of the United States of America 87, 6,7186,722.
775.
Thomas H. Brown
Johnston, D., and Brown, T. H. (1983). Voltage-clamp analysis of
Derick H. Lindquist
mossy fiber synaptic input to hippocampal neurons. Journal
of Neurophysiology 50, 464486.
(1984). Biophysics and microphysiology of synaptic trans-
mission in hippocampus. In R. Dingledine, ed., Brain slices. BEHAVIORAL ROLES
New York: Plenum Press.
Keenan, C. L., Chapman, P. F., Chang, V., and Brown, T. H. Long-term potentiation (LTP) refers the enhanced
(1988). Videomicroscopy of acute brain slices from hippocam- ability of a neuron to excite a neuron with which it is
pus and amygdala. Brain Research Bulletin 21, 373383. connected as a result of previous successful activation
Kelso, S. R., and Brown, T. H. (1986). Differential conditioning of
in the same pathway. LTP is typically produced by
associative synaptic enhancement in hippocampal brain
slices. Science 232, 8587.
electrically stimulating a group of input neurons in a
Kelso, S. R., Ganong, A. H., and Brown, T. H. (1986). Hebbian syn- way that produces simultaneous and repetitive activa-
apses in hippocampus. Proceedings of the National Academy of tion of their target neurons. Following this high fre-
Sciences of the United States of America 83, 5,3265,330. quency repetitive activation, future stimulation of the
Lee, H. J., Choi, J.-S., Brown, T. H., and Kim, J. J. (2001). Amygda- input neurons produces larger and more rapid activa-
lar N-methyl-D-aspartate (NMDA) receptors are critical for tion of the target neuron. The phenomenon of LTP
the expression of multiple conditioned fear responses. Journal
has many of the features of memory, including a last-
of Neuroscience 21, 4,1164,124.
Lindquist, D. H., and Brown, T. H. (2001). Antagonizing NMDA
ing change in the responsiveness of neurons follow-
receptors in the basolateral complex of the amygdala pre- ing brief experience with specific inputs, suggesting
vents conditioned enhancement of the rat eyeblink reflex. In- that LTP may be the cellular basis of memory. At the
tegrative Physiological and Behavioral Science 36, 171. same time, one should not confuse LTP with memory.
Magee, J., Hoffman, D., Colbert, C., and Johnston, D. (1998). Elec- LTP is a laboratory phenomenon that involves mas-
trical and calcium signaling in dendrites of hippocampal py- sive levels of activation never observed in nature.
ramidal neurons. Annual Review of Physiology 60, 327346.
Thus, the best we can hope is that LTP and memory
Magee, J. C., and Johnston, D. (1997). A synaptically controlled as-
sociative signal for Hebbian plasticity in hippocampal neu-
share a common basis in cellular mechanisms.
rons. Science 275, 209213. This section will summarize some of the recent
Mahanty, N. K., and Sah, P. (1998). Calcium-permeable AMPA re- research on the possible linkage between LTP and
ceptors mediate long-term potentiation in interneurons in
memory, using two main approaches: demonstrations
the amygdala. Nature 394, 683687.
McGann, J. P., Moyer, J. R., Jr., and Brown, T. H. (2001). Predomi-
of changes in synaptic efficacy consequent to a learn-
nance of late-spiking neurons in layer VI of rat perirhinal cor- ing experience and attempts to prevent learning by
tex. Journal of Neuroscience 21, 4,9694,976. pharmacological or genetic manipulation of the mo-
Miserendino, M. J. D., Sananes, C. B., Melia, K. R., and Davis, M. lecular mechanisms of LTP induction.
(1990). Blocking of acquisition but not expression of condi-
tioned fear-potentiated startle by NMDA antagonists in the
amygdala. Nature 345, 716718. Do Conventional Learning Experiences
Morgan, S. L., Coussens, C. M., and Teyler, T. J. (2001). Depoten- Produce Changes in Synaptic Efficacy
tiation of vdccLTP requires NMDAR activation. Neurobiology
Similar to Those That Occur after LTP?
of Learning and Memory 76, 229238.
Moyer, J. R., Jr., and Brown, T. H. (2002). Patch-clamp techniques Tying changes in synaptic physiology to learning
applied to brain slices. In A. Walz, A. Boulton, and G. B. seems like a daunting task because of the expectation
Baker, eds., Advanced techniques for patch-clamp analysis. To- that the magnitude of such observable changes in
towa, NJ: Humana Press. gross field potentials would be vanishingly small fol-
Swanson, L. W., and Petrovich, G. D. (1998). What is the amygdala?
lowing any normal learning experience. In addition,
Trends in Neuroscience 21, 323331.
Washburn, M. S., and Moises, H. C. (1992). Electrophysiological it is likely that learning involves both positive and
and morphological properties of rat basolateral amygdaloid negative changes in synaptic efficacy, that is, both
neurones in vitro. Journal of Neuroscience 12, 4,0664,079. LTP and long-term depression (LTD). Thus learning
Weisskopf, M. G., Bauer, E. P., and LeDoux, J. E. (1999). L-type could result in changes in the distribution of potenti-
voltage-gated calcium channels mediate NMDA-independent ated and depressed synapses with little or no overall
associative long-term potentiation at the thalamic input syn- shift and consequently no change or even an overall
apses to the amgydala. Journal of Neuroscience 19, 10,512
reduction in the averaged evoked potentials com-
10,519.
Xiang, Z., and Brown, T. H. (1998). Complex synaptic current monly used to measure LTP.
waveforms evoked in hippocampal pyramidal neurons by ex- Despite these concerns, an early study reported
tracellular stimulation of dentate gyrus. Journal of Neurophy- enhanced excitability of the hippocampal perforant
siology 79, 2,4752,484.
pathway in rats who had been exposed for prolonged
LONG-TERM POTENTIATION: Behavioral Roles 347
periods to an enriched environment (group hous- Do Treatments That Block LTP Prevent
ing with toys) as compared to impoverished envi- Memory?
ronment (solitary housing without toys). Environ-
The major limitation of the foregoing approach
mental enrichment resulted in an increased slope of
is that the experiments only provide correlationsbetw-
the synaptic potential and larger population action een aspects of LTP and memory. The converse ap-
potentials, consistent with the pattern of increased proach is to establish causal links between the LTP
synaptic efficacy observed following. Perhaps the and memory by seeing whether memory is disrupted
strong and long duration of learning involved in envi- by blocking LTP with drugs or genetic manipulations.
ronmental enrichment overcame the needle-in-the- This approach appeared fruitful because of the as-
haystack problem by enhancing the excitability of sumption that the manipulations would target plastic-
many hippocampal neurons. ity, not normal information processing in the brain,
and that they would knock out a critical kind of plas-
More recently, Joseph LeDoux and his colleagues ticity. This assumption arose from the observation
demonstrated LTP-like changes in neural responses that drugs such as D-2-amino-5-phosphonovalerate
in the amygdala. They trained rats to fear tones by (AP5) selectively block the NMDA receptor and thus
presenting repeated pairings of auditory stimuli and prevent hippocampal LTP while sparing normal syn-
foot shocks. Subsequently, in trained rats, condi- aptic transmissionhence the expectation that, to
tioned tones produced evoked potentials of greater the extent that the role of the NMDA receptor is re-
slope and amplitude similar to the characteristics of stricted to plasticity, these drugs would indeed block
LTP observed in the auditory pathway to the amygda- new learning without affecting nonlearning perfor-
la. There was no observable enhancement when the mance or retention of learning normally accom-
same tones and foot shocks arrived separately, and plished prior to drug treatment.
are therefore not associated with one another. Fur- Some of the earliest and strongest evidence sup-
thermore, the synaptic enhancements observed in porting a connection between LTP and memory came
trained rats were enduring, lasting as long as the be- from studies on spatial learning by Richard Morris
havioral response. and his colleagues. These studies exploited a water-
maze task in which rats learn to find an escape hidden
John Donoghue and his colleagues extended this
in a pool. Initially Morris and his colleagues showed
approach to the motor cortex and another form of
that AP5 prevented new spatial learning in the water
learning. They trained rats to reach with a paw maze. Drug-treated rats swam normally but did es-
through a small hole in a chamber to retrieve food cape as rapidly as the normal rats did. To assess the
pellets. Following the training, Donoghue and his col- rats knowledge of the escape locus, the researchers
leagues removed the brain and measured the used probe tests in which they removed the escape
strength of connections between cells within the area site and measured swimming near the location of the
of the motor cortex that controls hand movements. former escape site. Untreated rats showed a distinct
They used an in vitro preparation and evoked synap- preference for swimming in the vicinity of the former
tic potentials (EPSPs) in a principal cell layer of the escape locus, but drug-treated rats showed little or no
motor cortex by stimulating horizontal fibers that such bias, indicating the absence of memory of the es-
connect neighboring cells to one another. They found cape location. Further experiments showed no effect
that for the same or lower intensity of input stimula- of AP5 on memory when training was accomplished
tion, the magnitude of the EPSPs on the side of the prior to drug treatment. This is the expected result,
brain that controlled the trained paw (i.e., in the con- because NMDA receptors are necessary only for the
tralateral hemisphere) were consistently larger than induction of LTP, not for its maintenance.
those on the side of the brain that controlled the un- In other research by Morris and his colleagues
trained paw. Furthermore, they found it difficult to have shown how NMDA-receptor-dependent LTP
induce LTP by electrical stimulation in the trained might play a continuing role in updating memory. To
hemisphere but not in the untrained hemisphere. this end they varied the water-maze task by changing
Thus, training produced an anatomically localized in- the location of the escape platform every day. The
crease in synaptic efficacy that occluded the capacity rats consistently found the platform very rapidly on
for LTP. These observations show that synaptic po- the second trial on a given day. The animals were
tentiation results from motor learning and that the then tested with different memory delays inserted be-
real plasticity phenomenon shares common resources tween the first and second trial on each day. On some
with the artificial one. This study provides strong evi- days, AP5 was infused into the hippocampus, and on
dence for common cellular mechanisms of LTP and other days a placebo was given. AP5 treatment result-
learning. ed in a deficit on the second trial. Moreover, this defi-
348 LONG-TERM POTENTIATION: Behavioral Roles
cit varied with the duration between the first and sec- spite these highly selective temporal and anatomical
ond trials: there was no impairment with a fifteen- restrictions, the mice with this mutation were severely
second intertrial interval, but significant deficits deficient in spatial learning and other types of memo-
ensued with a delay of at least twenty minutes. These ry dependent on hippocampal function. A comple-
data suggest that memory for specific episodes of spa- mentary recent study showed that a mutation that re-
tial learning remains dependent on NMDA receptors sults in overexpression of NMDA receptors can
and LTP, even after the animals have learned the en- enhance several kinds of memory dependent on the
vironment and the general rules of the spatial task. hippocampus. Molecular genetic manipulations in-
creasingly indicate that interference with other as-
Other studies suggest that the cascade of molecu-
pects of the LTP molecular cascade, specifically PKC
lar events occasioned by LTP may also mediate the
and MAPK, also impair memory. Thus it seems likely
cortical plasticity that underlies memory. Yadin
that the full set of cellular events that mediate LTP
Dudai and his colleagues have focused on taste learn-
play critical roles in memory.
ing mediated by the gustatory cortex of rats. When
rats are exposed to a novel taste and subsequently be-
come ill, they develop a conditioned aversion to that Conclusion
taste, and this learning is known to depend on the
LTP and memory are not the same thing, and
gustatory cortex. AP5 produced an impairment in
there is no universal acceptance of evidence for
taste-aversion learning, whereas the same injections
shared mechanisms between LTP and memory. And,
given prior to retention testing or into an adjacent
notwithstanding some contradictory evidence not
cortical area had no effect. It is likely, then, that modi-
covered in this article, there is, nevertheless, compel-
fications in cortical taste representations rely on
ling evidence that learning enhances synaptic poten-
NMDA-dependent LTP.
tials in circuits relevant to memory. There is corre-
Furthermore, the blockade of protein synthesis in spondingly strong evidence that blocking LTP with
the gustatory cortex by infusion of an inhibitor prior drugs or genetic manipulations can impair memory
to learning also prevents development of the condi- and destabilize relevant neural representations.
tioned-taste aversion. MAP kinase and a downstream
protein kinase were activated selectively in the gusta- See also: GENETIC SUBSTRATES OF MEMORY:
tory cortex within ten minutes of exposure to a novel HIPPOCAMPUS; GLUTAMATE RECEPTORS AND
taste; activation peaked at thirty minutes, whereas ex- THEIR CHARACTERIZATION; LONG-TERM
posure to a familiar taste had no effect. Conversely, POTENTIATION; NEURAL SUBSTRATES OF
an MAP kinase inhibitor retarded conditioned-taste EMOTIONAL MEMORY; TASTE AVERSION AND
PREFERENCE LEARNING IN ANIMALS
aversion. This combination of findings strongly im-
plicate NMDA-mediated plasticity and subsequent
specific protein synthesis as critical factors in the cor- Bibliography
Green, E. J., and Greenough, W. T. (1986). Altered synaptic trans-
tical modifications that mediate this type of learning.
mission in dentate gyrus of rats reared in complex environ-
Other research has used targeted genetic manip- ments. Evidence from hippocampal slices maintained in vitro.
Journal of Neurophysiology 55, 739750.
ulations to show that blocking the cascade of molecu-
Martin, S. J., Grimwood, P. D., and Morris, R. G. M. (2000). Synap-
lar triggers for LTP also results in severe memory im- tic plasticity and memory: An evaluation of the hypothesis.
pairment. In one such early study, mice with a Annual Review of Neuroscience 23, 649711.
mutation of one form of CaMKII had deficient LTP Rioult-Pedotti, M.-S., Friedman, D., Hess, G., and Donoghue, J. P.
and were selectively impaired in learning the Morris (1998). Strengthening of horizontal cortical connections fol-
lowing skill learning. Nature Neuroscience 1, 230234.
water maze. Manipulation of biochemical mecha-
Rogan, M. T., Staubli, U. V., and LeDoux, J. E. (1997). Fear condi-
nisms by interference with specific genes has allowed tioning induces associative long-term potentiation in the
a highly specific identification of some of the critical amygdala. Nature 390, 604607.
molecular events. One study by Alcino Silva and his Rosenblum, K., Berman, D.E., Hazvi, S., Lamprecht, R., and
colleagues showed that substituting a single amino Dudai, Y. (1997). NMDA receptor and the tyrosine phospho-
rylation of its 2B subunit in taste learning in the rat insular
acid in CaMKII that prevents its autophosphorylation
cortex. Journal of Neuroscience 17, 5,1295,135.
resulted in a severe learning and memory deficit. Silva, A. J., Smith, A. M., and Giese, K. P. (1997). Gene targeting
Other new genetic approaches are providing a great- and the biology of learning and memory. Annual Review of Ge-
er temporal- and region-specific blockade of gene netics 31, 527547.
activation. Susumu Tonegawa and his colleagues Steele, R. J., and Morris, R. G. M. (1999). Delay dependent impair-
ment in matching-to-place task with chronic and intrahippo-
created a genetic block that was limited to postde-
campal infusion of the NMDA-antagonist D-AP5. Hippocam-
velopment activation of the genes for the NMDA re- pus 9, 118136.
ceptor in the CA1 subfield of the hippocampus. This Stevens, C. F. (1998). A million dollar question: Does LTP = mem-
mutation selectively blocked LTP in that region. De- ory? Neuron 20, 12.
LONG-TERM POTENTIATION: Maintenance 349
Tang, Y-P., Shimizu, E., Dube, G. R., Rampson, C., Kerchner, G. had been induced. Since the induction of LTP in-
A., Zhuo, M., Liu, G., and Tsien, J. Z. (1999). Genetic en- volves the postsynaptic activation of NMDA recep-
hancement of learning and memory in mice. Nature 401, 63
69.
tors, a retrograde signal has been postulated that re-
Tsien, J. Z., Huerta, P. T., Tonegawa, S. (1996). The essential role lays the postsynaptic activation to the presynaptic
of hippocampal CA1 NMDA receptor-dependent synaptic terminal. Researchers initially proposed arachidonic
plasticity in spatial memory. Cell 87, 1,3271,338. acid to be such a retrograde signal, since activation of
Howard Eichenbaum NMDA receptors stimulates phospholipase A2
(PLA2), and PLA2 inhibitors prevent LTP formation
(Bliss and Lynch, 1988). Arachidonic acid could also
provide a link with the presynaptic machinery in-
MAINTENANCE volved in the regulation of transmitter release be-
cause it activates protein kinases that have been
One of the most striking properties of memory is its
shown to participate in phosphorylation reactions im-
extreme duration; people have memories dating as
portant for transmitter release. Researchers later pro-
far back as twenty years or more. Similarly, one of the
posed that nitric oxide as well as carbon monoxide
features of long-term potentiation or LTP (at least in
were retrograde signals, although the evidence for
certain brain structures) is its duration; it has been a
this conclusion remains highly controversial (Haw-
major challenge for neurobiology to provide a cellu-
kins, Son, and Arancio, 1998; Malenka and Nicoll,
lar mechanism for long-lasting changes in synaptic
1999). While this mechanism could conceivably pro-
transmission that outlast protein turnover. The dis-
vide a satisfactory explanation for a short-lasting en-
covery of the critical role of the NMDA receptors in
triggering LTP indicated the existence of distinct hancement of transmitter release, in its present ver-
stages in the establishment of LTP, which are general- sion it does not account for a long-lasting increase.
ly defined as induction and maintenance (Bliss and Quantal analysis of synaptic transmission (Mar-
Lynch, 1988). The former refers to events activated tin, 1966) is an approach that, in principle, could pro-
during the high-frequency stimulation, whereas the vide an unambiguous answer to the question of the
latter designates the processes that are responsible locus of the changes underlying LTP. It uses the in-
for the changes in synaptic transmission that underlie trinsic variability of transmitter release and statistical
LTP and their long-term maintenance. In general, methods to determine the parameters generally
three types of mechanisms have been proposed as thought to govern synaptic transmission; in effect, the
maintenance mechanisms: those producing increased probability of release, the number of release sites, and
transmitter release; those resulting in changes in the elementary size of a postsynaptic response elicited
spine electrical properties; and those implicated in by a quantum of neurotransmitter. Three groups
regulating receptor properties (Landfield and Dead- have reported results obtained with quantal analysis
wyler, 1988). Since the 1980s evidence against and for in field CA1 before and after LTP induction in hippo-
each of these mechanisms has been obtained, and the campal slices. Two groups observed an increase in
most salient features of the arguments are presented quantal content (Malinow and Tsien, 1990; Bekkers
in this entry. and Stevens, 1990), whereas one group observed an
increase in quantal size (Foster and McNaughton,
Presynaptic Mechanisms 1991). In addition to this ambiguity in the results,
quantal analysis requires a number of assumptions
An increase in transmitter release could account that might not be satisfied at hippocampal synapses.
for an increase in synaptic transmission; it is generally
In particular, several studies have indicated the exis-
admitted, for instance, that short-term potentiation is
tence of synapses with functional NMDA receptors
due to increased transmitter release (Zucker, 1989).
but few, if any, functional AMPA receptors (the so-
Biochemical and electrophysiological evidence has
called silent synapses). Transformation of silent syn-
been obtained in support of the hypothesis that in-
apses into active synapses by LTP-inducing stimula-
creased transmitter release accounts for LTP. Gluta-
tion could account for the inconsistencies in results
mate is likely to be the neurotransmitter used by syn-
from quantal analysis (Malenka and Nicoll, 1999).
apses exhibiting LTP in various hippocampal
pathways, and increased glutamate levels have been Finally, increased transmitter release could be ac-
found after LTP induction in perfusates obtained counted for by an increase in the number of synapses.
with push-pull cannulas implanted in the dentate Anatomical studies using quantitative electron mi-
gyrus (Dolphin, Errington, and Bliss, 1982). In- croscopy have provided evidence that the number of
creased glutamate release was also found after LTP certain categories of synapses, in particular sessile
induction in hippocampal slices and in synaptosomes synapses, is increased in field CA1 following LTP in-
prepared from hippocampus of rats in which LTP duction (Lee, Schottler, Oliver, and Lynch, 1980;
350 LONG-TERM POTENTIATION: Maintenance
Chang and Greenough, 1984). Although an increased in two closely similar versions generated by alterna-
number of synapses could explain the duration of tive splicing of the genes, and designated flip and
LTP and account for its maintenance, it is not yet flop. Researchers have proposed that AMPA recep-
clear whether the increase in sessile synapses repre- tors exist as oligomers composed of a number of pos-
sents the formation of new synapses or the transfor- sible arrangements of flip and flop elements, with the
mation of existing synapses by shape modifications flip elements providing larger currents than the flop
(Toni et al., 1999). (Sommer et al., 1990). Twenty-first-century research
has suggested that LTP could result from a change in
the composition of receptor subunits or in changes in
Changes in Spine Electrical Properties the rates of receptor insertion or internalization in
Since their discovery, there has been much specu- the synaptic membrane from a subsynaptic pool of re-
lation concerning the roles of dendritic spines in syn- ceptors (Malinow, Mainen, and Hayashi, 2000). An-
aptic transmission and the possibilities that synaptic other dimension related to the stabilization of the
plasticity could be due to alterations in their electrical changes underlying LTP maintenance has recently
properties (Rall, 1978; Coss and Perkel, 1985). Be- been included to the search for the mechanisms of
cause of their shapelarge heads relative to long, LTP. It is now apparent that LTP stabilization in-
narrow necksthey are considered to be high- volves adhesion molecules and that there is a time
resistance elements coupling voltage changes at the window following LTP induction during which pro-
synapses with voltage changes in dendritic shafts. cesses leading to LTP consolidation can be reversed,
Based upon calculation and computer simulation, re- thus producing depotentiation (Lynch, 1998). There
searchers have proposed that decreased neck resis- has been considerable discussion to determine wheth-
tance could account for increased synaptic responses er depotentiation represents the same process as
(Wilson, 1988), which could affect the AMPA recep- long-term depression of synaptic transmission. Al-
tors more specifically than the NMDA receptors though both processes share a number of features,
(Lynch and Baudry, 1991). Such a decrease in spine there is not yet a general consensus as to the mecha-
resistance could be due to an increase in neck dimen- nisms and the significance of these forms of synaptic
sion, and anatomical evidence indicates that LTP is plasticity.
accompanied by an increase in spines with wider and
shorter necks. The recent discoveries of active ele-
ments (channels or pumps) in dendritic spines pro- Conclusion
vide alternative means of modifying their electrical Although several mechanisms discussed in this
properties (Johnston, Magee, Colbert, and Cristie, entry could account for some characteristics of LTP,
1996), although there is no evidence that these ele- most of them run into difficulties when they have to
ments are modified following LTP induction. explain its stability. This feature is almost certain to
eliminate most mechanisms, however attractive they
might appear, that are based solely on conformation-
Changes in Receptor Properties al or posttranslational changes in proteins. In particu-
One of the most important characteristics of LTP lar, several speculative mechanisms of LTP mainte-
is that it is expressed by an increase in the component nance have been based on the idea of biochemical
of the synaptic response resulting from the activation switches, constituted of biochemical reactions involv-
of the AMPA receptors with little changes in the coming positive feedback. The most popular of these bio-
ponent generated by the activation of NMDA recep- chemical switches is derived from the properties of
tors. Thus an alteration of the properties of the AMPA autophosphorylation of calcium/calmodulin kinase
receptors has been proposed as a potential mainte- type II, which is a very prominent protein in postsyn-
nance mechanism. Several studies have shown that aptic densities and especially in hippocampus (Crick,
both the ligand binding and the ionic conductance 1984; Lisman, 1994). The mechanism involved in
properties of the AMPA receptors are affected by a va- LTP maintenance therefore will probably require
riety of manipulations. In particular, changes in the structural modifications that can confer lasting stabili-
lipid environment of the AMPA receptors modify its ty to biochemical modifications responsible for
affinity for agonists, an observation that could ac- changes in synaptic transmission. The coupling of
count for the involvement of PLA2 in LTP. Studies these processes with those responsible for the mainte-
conducted in the 1990s on the molecular biology of nance of the adhesive properties of synaptic contacts
the AMPA receptors have provided exciting results represents a key feature in the evolution of adaptive
concerning the nature and properties of these recep- properties of brain neuronal networks. Finally, the
tors. AMPA receptors belong to a family of receptors various mechanisms described above underline the
encoded by at least four related genes, each existing richness of mechanisms that have evolved to produce
LONG-TERM POTENTIATION: Signal Transduction Mechanisms and Early Events 351
synaptic plasticity and that are likely to contribute to Sommer, B., Keinanen, K., Verdoorn, T. A., Wisden, W.,
the multiple forms of potentiation that are being un- Burnashev, N., Herb, A., Kohler, M., Takagi, T., Sakmann,
B., and Seeburg, P. H. (1990). Flip and flop: A cell-specific
covered at a variety of central synapses. functional switch in glutamate-operated channels of the CNS.
Science 249, 1,5801,585.
See also: GLUTAMATE RECEPTORS AND THEIR Toni, N., Buchs, P.-A., Nikonenko, I., Bron, C. R., and Muller, D.
CHARACTERIZATION; GUIDE TO THE ANATOMY (1999). LTP promotes formation of multiple spine synapses
between an axon terminal and a dendrite. Nature 401, 421
OF THE BRAIN: NEURON; MORPHOLOGICAL
425.
BASIS OF LEARNING AND MEMORY: Wilson, C. J. (1988). Cellular mechanisms controlling the strength
VERTEBRATES of synapses. Journal of Electron Microscope Technology 10, 293
313.
Zucker, R. S. (1989). Short-term synaptic plasticity. Annual Review
Bibliography
Bekkers, J. M., and Stevens, C. F. (1990). Presynaptic mechanism
for long-term potentiation in the hippocampus. Nature 346, Michel Baudry
724729. Gary Lynch
Bliss, T. V. P., and Lynch, M. A. (1988). Long-term potentiation:
Mechanisms and properties. In P. W. Landfield and S. A. De-
adwyler, eds., Long-term potentiation: From biophysics to behavior.
New York: Liss.
Chang, F., and Greenough, W. T. (1984). Transient and enduring
SIGNAL TRANSDUCTION MECHANISMS
morphological correlates of synaptic activity and efficacy AND EARLY EVENTS
change in the rat hippocampal slice. Brain Research 309, 35
46. Scientists believe that long-lasting changes in synaptic
Coss, R. G., and Perkel, D. H. (1985). The function of dendritic function are essential for learning and memory in the
spines: A review of theoretical issues. Behavior, Neurology, and mammalian brain. A widely studied example of such
Biology 44, 151185. synaptic plasticity is long-term potentiation (LTP).
Crick, F. (1984). Memory and molecular turn-over. Nature 312, The remarkable feature of LTP is that a short burst
101.
Dolphin, A. C., Errington, M. L., and Bliss, T. V. P. (1982). Long- of synaptic activity can trigger persistent enhance-
term potentiation of the perforant path in vivo is associated ment of synaptic transmission lasting for at least sev-
with increased glutamate release. Nature 297, 496498. eral hours, and possibly weeks or longer. There is
Foster, T. C., and McNaughton, B. L. (1991). Long-term synaptic great interest in understanding the cellular and mo-
enhancement in CA1 is due to increased quantal size, not lecular mechanisms that underlie this form of synap-
quantal content. Hippocampus 1, 7991.
Hawkins, R. D., Son, H., and Arancio, O. (1998). Nitric oxide as tic plasticity. First found in the hippocampus, this
a retrograde messenger during long-term potentiation in hip- phenomenon is now known to exist in cerebral cortex
pocampus. Progress in Brain Research 118, 155172. and other areas of the mammalian central nervous
Johnston, D., Magee, J. C., Colbert, C. M., and Cristie, B. R. system (CNS). Indeed, damage to the hippocampus
(1996). Active properties of neuronal dendrites. Annual Re- can result in certain defects in memory acquisition
Landfield, P. W., and Deadwyler, S. A., eds. (1988). Long-term poten- (see Milner, Squire, and Kandel, 1998).
tiation: From biophysics to behavior. New York: Liss. Most studies on LTP focus on the synapse be-
Lee, K., Schottler, F., Oliver, M., and Lynch, G. (1980). Brief bursts tween Schaffer collaterals and hippocampal CA1 neu-
of high frequency stimulation produce two types of structural
changes in rat hippocampus. Journal of Neurophysiology 44, rons. In this system, a brief burst of afferent stimula-
247258. tion leads to induction of LTP in postsynaptic CA1
Lisman, J. (1994). The CaM kinase II hypothesis for the storage of cells through a combination of (1) membrane depo-
synaptic memory. Trends in Neurosciences 17, 406412. larization and (2) activation of glutamate receptors of
Lynch, G. (1998). Memory and the brain: Unexpected chemistries the NMDA subtype (e.g., Collingridge, Kehl, and
and a new pharmacology. Neurobiology of Learning and Memory
70, 80100. McLennan, 1983). Researchers generally agree that
Lynch, G., and Baudry, M. (1991). Re-evaluating the constraints on the depolarization relieves Mg2+ (magnesium ion)
hypotheses regarding LTP expression. Hippocampus 1, 914. block of NMDA receptor channels and allows a Ca2+
Malenka, R. C., and Nicoll, R. A. (1999). Long-term potentia- (calcium ion) influx into dendritic spines that some-
tiona decade of progress? Science 285, 1,8701,874. how triggers LTP (Nicoll, Kauer, and Malenka, 1988).
Malinow, R., Mainen, Z. F., and Hayashi, Y. (2000). LTP mecha-
nisms: From silence to four-lane traffic. Current Opinion in
Neurobiology 3, 352357. Back-Propagating Action Potentials
Malinow, R., and Tsien, R. W. (1990). Presynaptic enhancement
shown by whole-cell recordings of long-term potentiation in In hippocampal pyramidal neurons an important
hippocampal slices. Nature 346, 177180. component of the membrane depolarization that al-
Martin, A. R. (1966). Quantal nature of synaptic transmission. Phys- lows opening of NMDA receptors is back-propagating
iological Review 46, 5166.
Rall, W. (1978). Dendritic spines and synaptic potency. In R. Por- action potentials. While action potentials are of
ter, ed., Studies in Neurophysiology. Cambridge, UK: Cam- course triggered in the active zone of the cell body,
bridge University Press. hippocampal pyramidal neurons along with many
352 LONG-TERM POTENTIATION: Signal Transduction Mechanisms and Early Events
other types of CNS neurons can actively propagate pathway, the ras/ERK MAP kinase pathway, and calci-
action potentials into the dendritic regions. These um-responsive nitric oxide (NO) synthase (Adams
dendritic action potentials are just like action poten- and Sweatt, 2002; Lisman and Zhabotinsky, 2001; Lu,
tials propagated down axons in that they are carried Kandel, and Hawkins, 1999; Sweatt, 1999; Hrabetova
predominantly by voltage-dependent ion channels and Sacktor, 1996).
such as sodium channels. The penetration of the
Involvement of Ca2+ -dependent protein kinases
back-propagating action potential into the dendritic
has been extensively tested and a wide variety of evi-
region provides a wave of membrane depolarization
dence is compatible with the general hypothesis of a
that allows for the opening of the voltage-dependent
necessity for postsynaptic protein kinase activation as
NMDA receptor/ion channels. In fact, the timing of
being necessary for triggering LTP (see Figure 1). A
the arrival of a dendritic action potential with synap-
necessity for PKC and CaMKII has been tested by in-
tic glutamate input appears to play an important part
tracellular injection of peptides that are potent and
in precise, timing-dependent triggering of synaptic
selective inhibitors of either PKC or CaMKII (Mali-
plasticity in the hippocampus (Magee and Johnston,
now, Schulman, and Tsien, 1989), and by character-
1997). Moreover, modulatory neurotransmitter sys-
ization of kinase-deficient mice generated through
tems can regulate the likelihood of action potential
gene knockout technology (Abielovich et al., 1993;
back-propagation through controlling dendritic
Silva, Stevens, Tonegawa, and Wang, 1992). Pharma-
potassium channels, allowing for sophisticated
cologic inhibitors of MAP kinase activation have also
information processing through an interplay of ac-
been shown to block LTP induction (Adams and
tion potential propagation, glutamate release, and
Sweatt, 2002). Involvement of the PKA pathway has
neuromodulation (Johnston, Hoffman, Colbert, and
been probed using pharmacologic and transgenic an-
Magee, 1999).
imal approaches, as has the involvement of the NO
synthase/cGMP pathway (Lu, Kandel, and Hawkins,
Importance of a Rise in Postsynaptic [Ca2+]i 1999; and Sweatt, 1999).
NMDA receptor activation leads to a transient The results obtained with inhibition of PKC,
[Ca2+]i increase arising in the postsynaptic neuron, CaMKII, PKA, MAP Kinase, and the NO/cGMP path-
an effect that researchers have measured by the use way can be interpreted in terms of a network of pro-
of fluorescent Ca2+ indicator dyes in pyramidal cell tein kinases, with protein phosphorylation as a link
dendrites within hippocampal slices (Regehr and between the rise in [Ca2+]i and the eventual expres-
Tank, 1990). A variety of experiments have demon- sion of enhanced synaptic function. However, the to-
strated the importance of the rise in [Ca2+]i for LTP. pology of the network and the nature of the interac-
Ca2+ buffers such as EGTA or BAPTA have been in- tions remain undefined. Indeed, there is no evidence
troduced with the aim of suppressing the transient to date to exclude the idea that one or more of these
Ca2+ increase; such maneuvers are effective in pre- enzymes act in a merely permissive way. At one ex-
venting the induction of LTP (Lynch et al., 1983). treme, background activity of a particular kinase
Moreover, a rise in postsynaptic Ca2+. independent might be necessary only prior to induction, to set the
of glutamate receptors, has been imposed by pho- stage for some other signaling mechanism triggered
toactivation of a caged Ca2+ compound, nitr-5; this by Ca2+.
method for Ca2+ delivery causes a sustained synaptic
potentiation (Malenka, Kauer, Zucker, and Nicoll,
1988). Evidence for Presynaptic Expression of
LTP
The question of how the kinases act leads to con-
Involvement of Protein Kinases sideration of ongoing debate about the nature of the
The key question at this point is how a relatively maintained synaptic enhancement in LTP. There is
brief rise in [Ca2+]i can lead to a long-lasting en- a variety of evidence to support the view that both in-
hancement of synaptic function. One popular hy- creased presynaptic transmitter release and en-
pothesis is that Ca2+ acts by activating signal trans- hanced glutamate receptor function postsynaptically
duction pathways sensitive to the elevation of are involved. The results indicating enhancement
postsynaptic Ca2+. The pathways that have been im- presynaptically support the idea of a retrograde sig-
plicated are quite varied and include: the nal that travels from the postsynaptic cell back to the
Ca2+-phospholipid dependent protein kinase (pro- presynaptic terminal. This line of reasoning has led
tein kinase C, PKC), multifunctional Ca/calmodulin- to a search for specific compounds that might act as
dependent protein kinase (CaMKII), calcium/ the retrograde messenger, with arachidonic acid, ni-
calmodulin sensitive adenylyl cyclase and the PKA tric oxide, and superoxide being the most widely con-
LONG-TERM POTENTIATION: Signal Transduction Mechanisms and Early Events 353
Figure 1
A typical LTP experiment and the effect of a protein kinase inhibitor on LTP. This plot shows the Excitatory Postsynaptic Potential
(EPSP) magnitude, measured as the initial slope, over time. The arrow indicates a brief (2 seconds total) period of high-frequency
synaptic stimulation. A fairly typical effect of application of a protein kinase inhibitor is shownin this case the compound SL327,
which inhibits activation of the ERK MAP Kinase cascade. The drug was present throughout the recording period.
sidered possibilities (Sweatt, 1999). The precise Mechanisms of Postsynaptic Enhancement

mechanisms by which these retrograde signals might In terms of mechanisms for enhancing postsyn-
affect neurotransmitter release are unclear, although
aptic responsiveness, two types of mechanisms are
presynaptic activation of PKC and the cGMP-
being actively pursued experimentally. One type of
dependent protein kinase are appealing possibilities.
process is referred to as activation of silent synapses
Little is known for certain about possible pre- (Poncer and Malinow, 2001; Isaac, Nicoll, and
synaptic mechanisms that might be set in motion by Malenka, 1999). The concept is that some latent syn-
putative retrograde messengers. A widely considered apses have available only NMDA receptors and thus
mechanism for synaptic potentiation involves a per- are inactive in terms of normal baseline synaptic
sistent enhancement of a presynaptic protein kinase, transmission. Activation of NMDA receptors by the
such as PKC (Linden and Routtenberg, 1989). The mechanisms described above leads to calcium influx
maintained expression of LTP can be reversibly and calcium-mediated insertion of AMPA receptors
blocked by bath application of a relatively nonspecific locally. Thus previously silent synapses become ca-
kinase inhibitor, H-7 (Malinow, Madison, and Tsien, pable of baseline synaptic transmission and the net
1988), and biochemical measurements show a sus- result is an increased efficacy of transmission.
tained enhancement of PKC in hippocampal slices
(Klann, Chen, and Sweatt, 1993; Hrabetova and Sack- A second mechanism under investigation is ki-
tor, 1996). PKC is known to increase the efficiency of nase-dependent regulation of AMPA receptors
excitation-secretion coupling in many systems, in- through direct phosphorylation. It is known that
cluding chromaffin cells, motor nerve terminals, and phosphorylation of AMPA receptors by PKC and
cultured hippocampal neurons. Further exploration CaMKII leads to increased conductance of ions
of presynaptic mechanisms is needed to determine through the AMPA channel, and evidence is available
which steps leading to exocytosis are enhanced in that phosphorylation at these regulatory sites in-
LTP. creases with LTP-inducing stimulation (Lee et al.,
354 LONG-TERM POTENTIATION: Signal Transduction Mechanisms and Early Events
2000; Derkach, Barria, and Soderling, 1999; Barria nases in LTP, the role of morphological changes and
et al., 1997). Thus it is hypothesized that through in- local protein synthesis in early stages of LTP, and the
creased postsynaptic phosphorylation of AMPA re- seemingly perpetual question of pre- versus postsyn-
ceptors synaptic efficacy is increased by enhancement aptic loci of LTP expression. Scientists are hopeful
of AMPA receptor function. The silent synapse hy- that continued research and progress will shed light
pothesis and the enhanced AMPA receptor conduct- on these topics.
ance hypothesis are not mutually exclusive; indeed
both could be simultaneously occurring at different See also: GLUTAMATE RECEPTORS AND THEIR
synapses between the same cells, or at the same syn- CHARACTERIZATION; GUIDE TO THE ANATOMY
apses at different points in time. OF THE BRAIN: HIPPOCAMPUS AND
PARAHIPPOCAMPAL REGION; LONG-TERM
POTENTIATION: OVERVIEW: COOPERATIVITY
Potential Roles for Protein Phosphatase AND ASSOCIATIVITY; SECOND MESSENGER
Regulation SYSTEMS
Finally, while increased protein phosphorylation Bibliography

has been hypothesized to play a key role in triggering Abeliovich, A., Chen, C., Goda, Y., Silva, A. J., Stevens, C. F., and
LTP for quite some time, studies conducted since the Tonegawa, S. (1993). Modified hippocampal long-term po-
turn of the century highlight a similar importance of tentiation in PKC gamma-mutant mice. Cell 75 (7), 1,253
tightly regulating protein dephosphorylation as one 1,262.
Adams, J. P., and Sweatt, J. D. (2002). Molecular psychology: Roles
of the mechanisms controlling the induction of long-
for the ERK MAP kinase cascade in memory. Annual Review
term synaptic change and lasting memory (Winder of Pharmacology and Toxicology 42, 135163.
and Sweatt, 2001). In one series of studies, Isabel Barria, A., Muller, D., Derkach, V., Griffith, L. C., and Soderling,
Mansui and her colleagues (Mallaret et al., 2001) en- T. R. (1997). Regulatory phosphorylation of AMPA-type glu-
gineered a mouse line in which the activity of the cal- tamate receptors by CaM-KII during long-term potentiation.
Science 276 (5,321), 2,0422,045.
cium-dependent phosphatase calcineurin (protein
Collingridge, G. L., Kehl, S. J., and McLennan, H. (1983). Excit-
phosphatase 2B) could be regulated, in a reversible atory amino acids in synaptic transmission in the Schaffer col-
manner, by tetracycline administration to adult ani- lateral-commissural pathway of the rat hippocampus. Journal
mals. In engineering their mouse they capitalized on of Physiology 334, 3346.
the known mechanism of regulation of calcineurin; Derkach, V., Barria, A., and Soderling, T. R. (1999).
Ca2+/calmodulin-kinase II enhances channel conductance of
that is, its regulation by an intrinsic autoinhibitory do-
alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionate type
main. Mansuis group engineered a mouse in which glutamate receptors. Proceedings of the National Academy of Sci-
they could regulate the expression of the isolated au- ences of the United States of America 96 (6), 3,2693,274.
toinhibitory domain of the calcineurin A subunit. In Hrabetova, S., and Sacktor, T. C. (1996). Bidirectional regulation
their mice, treatment of animals with tetracycline led of protein kinase M zeta in the maintenance of long-term po-
to expression of the calcineurin autoinhibitory do- tentiation and long-term depression. Journal of Neuroscience 16
(17), 5,3245,333.
main protein in several brain regions, including the Isaac, J. T., Nicoll, R. A., and Malenka, R. C. (1999). Silent gluta-
hippocampus. They then investigated long-term po- matergic synapses in the mammalian brain. Canadian Journal
tentiation in hippocampal slices, specifically in area of Physiology and Pharmacology 77 (9), 735737.
CA1 where many of the earlier studies implicating ki- Johnston D., Hoffman, D. A., Colbert, C. M., and Magee J. C.
nases had been performed. Calcineurin inhibition en- (1999). Regulation of back-propagating action potentials in
hippocampal neurons. Current Opinion in Neurobiology 9 (3),
hanced the magnitude and duration of long-term po- 288292.
tentiation in this region when LTP was triggered by Klann, E., Chen, S.-J., and Sweatt, J. D. (1993). Mechanism of PKC
high-frequency stimulation. These studies and a vari- activation during the induction and maintenance of LTP
ety of others (Winder and Sweatt, 2001) suggest the probed using a novel peptide substrate. Proceedings of the Na-
tional Academy of Sciences of the United States of America 90,
possibility of an important role for calcium-
8,3378,341.
dependent phosphatases in regulating the magnitude Lee, H. K., Barbarosie, M., Kameyama, K., Bear, M. F., and Hu-
of LTP induction in the hippocampus. ganir, R. L. (2000). Regulation of distinct AMPA receptor
phosphorylation sites during bidirectional synaptic plasticity.
Since the early 1990s, much progress has been
Nature 405 (6,789), 955959.
made in scientific understanding of the mechanisms Linden, D. J., and Routtenberg, A. (1989). The role of protein ki-
involved in LTP induction. A role for protein phos- nase C in long-term potentiation: A testable model. Brain Re-
phorylation in LTP induction is clear. Dendritic ac- search Reviews 14, 279296.
tion potentials were inconceivable in the early 1990s Lisman, J. E., and Zhabotinsky, A. M. (2001). A model of synaptic
memory: A CaMKII/PP1 switch that potentiates transmission
but in the twenty-first century are known to be critical
by organizing an AMPA receptor anchoring assembly. Neuron
components of information processing in neurons, 31 (2), 191201.
through their contributions to LTP induction. Many Lu, Y. F., Kandel, E. R., and Hawkins, R. D. (1999). Nitric oxide
questions remain concerning the targets of protein ki- signaling contributes to late-phase LTP and CREB phospho-
LORENZ, KONRAD 355
rylation in the hippocampus. Journal of Neuroscience 19 (23),

10,25010,261.
Lynch, G., Larson, J., Kelso, S., Barrionuevo, G., and Schottler, F.
(1983). Intracellular injections of EGTA block induction of
hippocampal long-term potentiation. Nature 305, 719721.
Magee, J. C., and Johnston, D. (1997). A synaptically controlled,
associative signal for Hebbian plasticity in hippocampal neu-
rons. Science 275 (5,297), 209213.
Malenka, R. C., Kauer, J. A., Zucker, R. S., and Nicoll, R. A. (1988).
Postsynaptic calcium is sufficient for potentiation of hippo-
campal synaptic transmission. Science 242, 8184.
Malinow, R., Madison, D. V., and Tsien, R. W. (1988). Persistent
protein kinase activity underlying long-term potentiation. Na-
ture 335, 820824.
Malinow, R., Schulman, H., and Tsien, R. W. (1989). Inhibition of
postsynaptic PKC or CaMKII blocks induction but not expres-
sion of LTP. Science 245, 862866.
Malleret, G., Haditsch, U., Genoux, D., Jones, M. W., Bliss, T. V.,
Vanhoose, A. M., Weitlauf, C., Kandel, E. R., Winder, D. G.,
and Mansuy, I. M. (2001). Inducible and reversible enhance-
ment of learning, memory, and long-term potentiation by ge-
netic inhibition of calcineurin. Cell 104, 675686.
Milner, B., Squire, L. R., and Kandel, E. R. (1998). Cognitive
neuroscience and the study of memory. Neuron 20 (3), 445
468.
Nicoll, R. A., Kauer, J. A., and Malenka, R. C. (1988). The current
excitement in long-term potentiation. Neuron 1, 97103.
Poncer, J. C., and Malinow, R. (2001). Postsynaptic conversion of
silent synapses during LTP affects synaptic gain and transmis-
sion dynamics. Nature Neuroscience 4 (10), 989996.
Regehr, W. G., and Tank, D. W. (1990). Postsynaptic NMDA recep-
tor-mediated calcium accumulation in hippocampal CA1 py-
ramidal cell dendrites. Nature 345, 807810.
Silva, A. J., Stevens, C. F., Tonegawa, S., and Wang, Y. (1992). Defi- Konrad Lorenz (AP/WideWorld Photos)
cient hippocampal long-term potentiation in alpha-calcium-
calmodulin kinase II mutant mice. Science 257 (5,067), 201
206.
Sweatt, J. D. (1999). Toward a molecular explanation for long-term Anatomical Institute in Vienna; he later became pro-
potentiation. Learning and Memory 6, 399416. fessor of psychology at the University of Knigsberg.
Winder, D. G., and Sweatt, J. D. (2001). Roles of serine/threonine
phosphatases in hippocampal synaptic plasticity. Nature Re- Lorenzs flirtation with Nazi ideology after the
views Neuroscience 2, 461474. German annexation of Austria in 1938 tarnished his
Richard W.Tsien
scientific reputation, especially in the United States
Revised by J. David Sweatt
and Great Britain. In a 1940 paper he mixed Nazi jar-
gon into a discussion of one of his favorite themes, the
destructive effects of domestication on animals and
humans. Lorenzs ideas sat well with Nazi theoreti-
cians. He later repudiated this paper and acknowl-
LORENZ, KONRAD (19031989) edged that it had brought him undeniable discredit.
The Austrian zoologist Konrad Lorenz was one of the During World War II Lorenz served as an army
pioneers of modern ethology, the comparative study physician on the eastern front, where he was captured
of animal behavior. Also noted for his views on ag- by Soviet troops. It was as a prisoner of war in Arme-
gression, Lorenz was a corecipient of the Nobel Prize nia from 1944 to 1948 that Lorenz wrote his first
for physiology or medicine in 1973. book. The Russian manuscript (published posthu-
mously in 1996) anticipates many of his later ideas on
Konrad Zacharias Lorenz was born November 7, animal behavior, evolution, instinct, and comparative
1903, in Vienna, Austria. By his own account he was methodology.
fascinated from an early age with keeping and observ-
ing animals. He attended Columbia University in In 1958 Lorenz and Erich von Holst cofounded
New York for one term in 1922, received an M.D. the Max-Planck-Institut fr Verhaltensphysiologie at
from the University of Vienna in 1928, and was Seewiesen, Germany, where Lorenz remained as di-
awarded a Ph.D. in zoology by the same university in rector until his retirement in 1973. Following his re-
1933. His first professional appointment was at the tirement, the Austrian Academy of Sciences created
356 LORENZ, KONRAD
the Institute of Comparative Ethology, based at Lo- ically accumulated informationhence animals ca-
renzs family home in Altenberg, where he served as pacity for immediate, appropriate, unlearned
director of the department of animal sociology until responses to food, mates, offspring, or predators. Lo-
his death. In 1973 Lorenz, Niko Tinbergen, and Karl renz contrasted such genetically determined re-
von Frisch shared the Nobel Prize in physiology or sponses with learned behavior and regarded behavior
medicine for their pioneering work in ethology, the as a sequence of learned and innate components. This
scientific study of animal behavior. In addition to his view was criticized by Lehrman for ignoring impor-
many scientific papers and books, Lorenz reached a tant findings about the development of behavior. In
wide audience with two popular books, King Solomons his theory of instinct, Lorenz conflated a number of
Ring (1952), a reminiscence on the delights of animal very different ideas about innate behavior, genetics,
study, and the more controversial On Aggression and evolution. The debate between Lorenz and Lehr-
(1966), a discussion of the origins of animal and man led to a clarification and revision of both posi-
human aggression. tions.
Lorenz was among the first to insist on the biolog-
The modern view of animal behavior preserves
ical roots of animal behavior, asserting that behavior
some of Lorenzs ideas, especially his insistence that
could be described with the same precision and scien-
functional behavior can occur in the absence of prior
tific rigor as anatomy. Though this was probably his
relevant experience. It is not correct, however, to re-
major contribution to the study of animal behavior,
gard such behavior as any more genetic than any-
he always acknowledged his debt to two of his pre-
thing else that animals do, or to suppose that such be-
decessors, Oskar Heinroth and Charles O. Whitman.
havior is developmentally inflexible. Both genes and
Lorenz further argued that the comparative study of
the environment can account for differences in be-
different species can reveal the effects of evolutionary
havior, and these two influences interact in their ef-
adaptation on behavior. Behavior varies among spe-
fects on behavior. Less controversy surrounds anoth-
cies because it is shaped by natural selection and
er of Lorenzs theoretical contributions: his insistence
shows the influence of the animals evolutionary his-
that learning in not simply a matter of open-ended
tory.
behavioral plasticity but rather consists of evolved
Behavior, Lorenz asserted, can be described with rules for acquiring information and modifying behav-
the same precision and scientific rigor as anatomy. ior.
He further argued that the comparative study of vari-
ous species can reveal the effects of evolutionary ad- Lorenz performed little experimental work, pre-
aptation on behavior, which varies among species be- ferring to draw on his enormous store of personal ob-
cause natural selection channels it to distinctive servation of animals. Many of his pronouncements on
functions in the physical and social environments of human behavior relied heavily on opinion and intu-
different animals. ition. Nevertheless, his compelling insights and per-
Lorenzs influence on contemporary research in suasive exposition of his ideas brought a new vigor to
learning and memory comes from two sources: his the scientific study of animal behavior.
work on imprinting and his ideas about the interplay Konrad Lorenz died on February 27, 1989, in Al-
between learned and innate behavior. Imprinting is tenberg, Austria.
the formation of a powerful bond between young pre-
cocial birds and their parents. Research findings on
imprinting, song-learning by birds, and other special- See also: IMPRINTING; OPERANT BEHAVIOR
ized forms of learning undermined the strict operant
reinforcement theories of behaviorism, the school of
thought that once dominated American psychology. Bibliography
Building upon Lorenzs pioneering work, later re- Lorenz, K. (1940). Durch Domestikation verursachte Strungen ar-
searchers have continued to do important work on teigenen Verhalten. Zeitschrift fr angewandte Psychologie und
the nature of imprinting, its neural basis, and the ef- Characterkunde 59, 1B81.
fects of imprinting on later mate choice. (1952). King Solomons ring. London: Methuen.
(1965). Evolution and modification of behavior. Chicago: Uni-
The second source of Lorenzs influence on cur- versity of Chicago Press.
rent research in learning and memory was his debate (1966). On aggression. London: Methuen.
with the American psychologist Daniel S. Lehrman (1996). The natural science of the human species. An introduction
to comparative behavioral research. Cambridge, MA: MIT Press.
during the 1960s on the relative importance of nature Nisbett, A. (1976). Konrad Lorenz. New York: Harcourt Brace
and nurture in the development of behavior. Much of Jovanovich.
Lorenzs scientific work is an analysis of instinctive be-
havior, which he regarded as the expression of genet- David F. Sherry
LURIA, A. R. 357
LURIA, A. R. (19021977)
The Soviet psychologist Alexandr Romanovich Luria
is best known for his work in neuropsychology but
also contributed to developmental and crosscultural
psychology. The son of a prominent physician, Luria
graduated from the University of Kazan in 1921 and
joined the staff of the Moscow Institute of Psychology
in 1923. He tempered his early interest in psycho-
analysis with a growing awareness of the need for ob-
jective methods. His attempts to study the uncon-
scious by objective methods resulted in the book The
Nature of Human Conflicts (1932). In 1924 Luria met
Lev Vigotsky, and their subsequent collaboration
until Vigotskys premature death in 1934 had a life-
long influence on Lurias work. Together they devel-
oped the concept of historicocultural psychology, ad-
vancing the thesis that higher cognitive functions A. R. Luria (Dr. Lena Moskovich)
result from the internalization of external cultural de-
vices and codes.
life. Although Luria believed in the invariant localiza-
Vigotsky and Luria embarked on two lines of re- tion of basic cognitive dimensions, he thought that
search driven by the historicocultural theory: devel- cultural and developmental factors governed the di-
opmental studies of language acquisition and the reg- mensional composition of the functional systems
ulatory role of language in behavior, and cross- corresponding to complex processes.
cultural studies of modes of inference among the
tribes of Central Asia. Both lines of research ran afoul Given Lurias cultural-developmental interests, it
of the official Marxist dogma of Soviet authorities, is not surprising that the neuropsychology of lan-
who pressured Luria was to terminate his develop- guage was foremost on his agenda in major works
mental and cross-cultural work. such as Traumatic Aphasia (1970), The Role of Speech in
the Regulation of Normal and Abnormal Behavior (1961),
Lurias scientific interests were gradually evolving and The Higher Cortical Functions in Man (1966).
toward more biological aspects of psychology and Lurias distinct taxonomy of aphasias reflects his in-
neuropsychology. In the early 1930s Luria turned his terest in refined linguistic operational analysis.
attention to the nature-nurture debate and em- Lurias interest in the frontal lobes was a continuation
barked on a series of studies of cognitive processes in of his earlier concern with self-regulation and con-
identical and fraternal twins. In the late 1930s, Luria, sciousness. His emphasis on the hierarchic nature of
already a professor of psychology, earned a medical cognitive control undoubtedly reflected his interac-
degree, perhaps because of an authentic intellectual tions with Nicholas Bernstein, a Soviet physiologist
evolution or perhaps because psychology had grown and mathematician who foreshadowed some of the
too ideologically parlous an endeavor amid the Sta- basic concepts of cybernetics and who, like Vigotsky,
linist strictures of Soviet academia. Lurias interest in suffered official ostracism for deviation from the Pav-
neuropsychology intensified duringWorld War II, lovian doctrine.
when he helped to develop remedial procedures for
soldiers with head injuries. In the later part of his career, Luria became deep-
ly interested in memory. The monograph Neuropsy-
Luria was more interested in developing a com- chology of Memory (1976) describes his attempts to root
prehensive theory of brain-behavioral relations rath- amnesic syndromes in subcortical neuroanatomy. He
er than in describing clinical syndromes or develop- distinguished the midline amnesic syndromes (cu-
ing diagnostic and remedial techniques. His early riously, without further distinguishing between the
eclectic interests shaped his brand of neuropsy- diencephalic and mesiotemporal variants) and pitu-
chology. Luria entered the field at the time of the rag- itary, mesiofrontal, massive prefrontal, and cortical
ing debate between narrow localizationism and memory deficits. He was particularly interested in the
equipotentialism. He went beyond this simplistic relationship between consolidation and executive
dilemma and formulated his concept of dynamic functions.
functional systems, which captures the relationship
between the localizable dimensions of cognition and Although Luria is widely known in the West for
the complex traits, skills, and behaviors of everyday various adaptations of his diagnostic approaches, he
358 LURIA, A. R.
never bothered to compile them into a battery. In few. The centennial anniversary of his birth in 2002
fact, he abhorred the notion of a battery and always has occasioned a number of scientific symposia world-
advocated a flexible, hypothesis-testing approach. He wide.
also disdained the notion of instrument standardiza-
tion. Although Lurias bias against quantification is See also: AMNESIA, ORGANIC; MEMORY
often cast in East-West terms, it is probably a genera- CONSOLIDATION: MOLECULAR AND CELLULAR
PROCESSES; MEMORY CONSOLIDATION:
tional phenomenon. After all, Luria the clinician de-
PROLONGED PROCESS OF REORGANIZATION;
scended from the the great turn-of-the-century Euro-
MNEMONISTS
pean neurologist-phenomenologists. What makes his
diagnostic approach singularly attractive is that every Bibliography
procedure targets a cognitive dimension accounted Luria, A. R. (1932). The nature of human conflicts. New York: Live-
for in his brain-behavioral theory. right.
(1961). The role of speech in the regulation of normal and abnor-
During the later part of his career, Luria was pro- mal behavior, ed. J. Tizard, London: Pergamon.
fessor of psychology at Moscow State University, (1966; reprint 1980). The higher cortical functions in man,
where he held the chair of neuropsychology, and di- trans. B. Haigh. New York: Basic Books.
rector of the Neuropsychology Laboratory at the (1970). Traumatic aphasia, trans. D. Bowden. The Hague:
Mouton.
Bourdenko Institute of Neurosurgery in Moscow. He
(1976). Neuropsychology of memory, trans. B. Haigh. Washing-
received numerous awards and was elected to various ton, DC: Winston.
academies and learned societies in the Soviet Union Goldberg, E. (1990). Contemporary neuropsychology and the legacy of
and in the West. Luria. Hillsdale: Erlbaum.
(2001). The executive brain: Frontal lobes and the civilized mind.
The enduring influence of Lurias seminal work New York: Oxford University Press.
has been the subject of several books: Contemporary Sacks, O. W. (1996). Scotoma: Forgetting and neglect in science.
Neuropsychology and the Legacy of Luria (1990), Hidden In R. B. Silvers, ed., Hidden histories of science. New York: New
Histories of Science (1996), and The Executive Brain: York Review of Books.
Frontal Lobes and the Civilized Mind (2001), to name a Elkhonon Goldberg
M
MASSED TRAINING mathematical analysis. However, the early efforts
were confined to the routine exercise of finding equa-
See: DISTRIBUTED PRACTICE EFFECTS tions that could provide compact descriptions of the
average course of improvement with practice in sim-
ple laboratory tasks (Gulliksen, 1934). Unfortunately,
these efforts yielded no harvest of new insights into
MATHEMATICAL LEARNING the nature of learning and did not enter into the de-
THEORY sign of research.
Theories of learning were enormously visible and in-
fluential in psychology from 1930 to 1950, but Clark L. Hull and Habit Strength
dropped precipitously from view during the next two The agenda for a new approach that might more
decades while the information-processing approach nearly justify the label mathematical learning theory was
to cognition based on a computer metaphor gained set by Clark L. Hull, who distilled much of the learn-
ascendancy. The groundwork for a resurgence of ing theory extant at the end of the 1930s into a system
learning theory in the context of cognitive psychology of axioms from which theorems might be derived to
was laid during the earlier period by the development predict aspects of conditioning and simple learning
of a subspecialty that may be termed mathematical in a wide variety of experimental situations (Hull,
learning theory. 1943). One of the central ideas was a hypothetical
In psychology, just as in physical and biological measure of the degree of learning of any stimulus-
sciences, mathematical theories should be expected response relationship. This measure, termed habit
to aid in the analysis of complex systems whose behav- strength (H), was assumed to grow during learning of
ior depends on many interacting factors and to bring the association between any stimulus and response ac-
out causal relationships that would not be apparent cording to the quantitative law
to unaided empirical observation. Mathematical rea-
soning was part and parcel of theory construction H = M (1 - ecN) (1)
from the beginnings of experimental psychology in
some research areas, notably the measurement of sen- where N denotes number of learning experiences, M
sation, but came on the scene much later in the study is the maximum value of H, and c is a constant repre-
of learning, and then got off to an inauspicious start. senting speed of learning. Thus, H should increase
By the 1930s, laboratory investigation of conditioning from an initial value of zero to its maximum as a
and learning at both the animal and the human level smooth diminishing returns function of trials.
had accumulated ample quantitative data to invite However, knowing the value of H does not enable
359
360 MATHEMATICAL LEARNING THEORY
prediction of what the learner will do at any point in search over the next two decades, it became apparent
practice for two reasons. First, a test for learning may that learning depends also on the degree to which an
be given in a situation different from that in which experience provides the learner with new informa-
practice occurred (as when a student studies at home tion. If, for example, an animal has already learned
but is tested in a classroom); according to Hulls prin- that a particular stimulus, Sl, predicts the occurrence
ciple of stimulus generalization, the effective habit of shock, then on a trial when S1 and another stimu-
strength on the test is assumed to be some reduced lus, S2, precede shock in combination, the animal
value H that differs from H as a function of the differ- may learn nothing about S2 (Kamin, 1969). This ob-
ence between the two situations. Further, motivation servation and many related ones can be explained by
must be taken into account. If the student in the ex- a new theory, which can be viewed as a refinement
ample studies in a relaxed mood but arrives at the test and extension of one component of Hulls axiom
situation in a state of anxiety, performance may be af- about habit growth. In the new version, the change in
fected and fail to mirror the actual level of learning. habit strength on any learning trial is described by the
In Hulls theory, it is assumed that the strength of the function
tendency to perform a learned act, termed excitatory
potential (E), is determined by habit strength and level H = c (M - HT), (4)
of motivation or emotion (denoted D) acting jointly
where H denotes strength of a particular habit under-
according to the multiplicative relation
going acquisitionfor example, the relation between
S2 and shock in the illustrationand HT denotes the
E = E D. (2) sum of strengths of all habits active in the situation
for example, the habits relating both Sl and S2 to
These conceptions were implemented in vigorous re- shock in the illustration (Rescorla and Wagner, 1972).
search programs mounted by Kenneth W. Spence, Studying the new learning function, one can see that
one analyzing the relation between anxiety and learn- if habit strength for S1 is large enough so that HT is
ing in human beings and the other producing the first equal to M, then there will be no increment in habit
quantitative accounts of the ability of some animals to strength for S2. The prediction that learning about S2
exhibit transposition (i.e., to generalize from learning is blocked by the previous learning about S1 has been
experiences in terms of relationships such as brighter borne out by many experiments. This success, and
than or larger than, as distinguished from absolute others like it, in dealing with novel and sometimes
values of stimulus magnitude). surprising findings has put the body of theory built by
Rescorla and Wagner and others on the groundwork
of Hulls system into a commanding position in pres-
The Law of Habit Growth
ent-day research on animal learning.
Hulls theory as a whole did not prove viable. It
depended on too many assumptions and was too
loose-jointed in structure to be readily testable, and Probability Matching
its deterministic axioms lacked the flexibility needed At one time it appeared that the basic ingredients
to interpret varieties of learning more complex than of Hulls theory could be adapted in a straightforward
simple conditioning. However, some components and simple way to apply to human learning. In the
have survived to reappear as constituents of new theo- early 1950s, a physicist converted to mathematical
ries. A notable instance is the law of habit growth. An psychology, in collaboration with an eminent statisti-
implication of Equation 1 is that the change in habit cian, arrived at the idea that the hypothetical notion
strength, H, on any trial can be expressed as of habit strength could be dispensed with so that
learning functions like Equation 3 would be ex-
H= c (M - H), (3) pressed in terms of changes in response probabilities
(Bush and Mosteller, 1955). If, for example, a rat was
in which form it is apparent that learning is fastest learning to go to the correct (rewarded) side of a sim-
when the difference between current habit strength ple T maze, they assumed that when a choice of one
and its maximum is large, and slow when the differ- side led to reward, the probability, p, of that choice
ence is small. According to Hulls assumptions, and would increase in accord with the function
indeed nearly all learning theories of his time, the
p = c (1 - p), (5)
function should be applicable on any trial when the
response undergoing learning occurs in temporal and when it led to nonreward, p would decrease in ac-
contiguity to reinforcement (that is, a positive incen- cord with
tive such as food for a hungry animal or a to-be-feared
event such as an electric shock). In the course of re- p = -cp (6)
MATHEMATICAL LEARNING THEORY 361
where c is a constant reflecting speed of learning. By Figure 1

a simple derivation it can be shown that if reward on
one side of the maze has some fixed probability (0, 1,
or some intermediate value), Equations 5 and 6, taken
together, imply a learning function of the same form
as Hulls function for growth of habit strength. Also,
the prediction can be derived that the final level of
the learning function will be such that, on the aver-
age, the probability of choosing a given side is equal
to the probability of reward on that side, as illustrated
in Figure 1. This prediction of probability matching has
been confirmed (under some especially simplified ex-
perimental conditions) in a number of studies. In an
elegant series of elaborations and applications of this
simple model, Bush and Mosteller (1955) showed that
it could be carried over to human learning in situa-
tions requiring repeated choices between alternatives
associated with different magnitudes and probabili-
ties of reward, as in some forms of gambling.
However, only a few such applications appeared,
and in time it became apparent that, although Bush Theoretical predictions of percentages of responses to a given
and Mostellers methods proved widely useful, nei- side of a choice point on each trial of an illustrative experiment
ther their approach nor others building on Hulls sys- for groups of learners who had different probabilities of reward
tem could be extended to account for many forms of (.85, .70, or .30) on that side. Details of the model are given in
human learning. the text.
Statistical Learning Theory theory, the associations simply related stimuli and re-
The limitations of models based on conceptions sponses. If, say, a rat was learning the correct turn in
of habit growth or similarly gradual changes in re- a T maze, then on any trial associations would be
sponse probabilities were anticipated by the great pi- formed between the response made, together with the
oneer of modern neuroscience, D. O. Hebb. In a sem- ensuing reward or nonreward, and whatever aspects
inal monograph, Hebb (1949) noted that, whereas of the choice situation the animal happened to attend
slow and gradual learning characterizes animals low to (the notion of sampling).
on the phylogenetic scale and immature organisms
An interesting mathematical property of this
higher on the scale, the learning of the higher forms,
model is that for simple learning situations, if the
most conspicuously human beings, as adults is nor-
total population of stimulus aspects available for sam-
mally much faster, and often takes the form of abrupt
pling is large, the basic learning functions take a form
reorganizations of the products of previous learning
identical to those of Bush and Mostellers model
(as when a person trying to master a technique of
(Equations 5 and 6), with the learning-rate parameter
computer programming suddenly grasps a key rela-
c interpreted as the proportion of stimulus aspects
tionship).
sampled on any trial. When the population is not
A theoretical approach that would prove more large, however, the model is better treated as a type
adaptable to the treatment of human learning ap- of probabilistic process (technically, a Markov chain)
peared in the 1950s, contemporaneously with the in which learning is conceived in terms of discrete
work of Bush and Mosteller, in the development of transitions between states. A state is defined by the
statistical learning theory, also known as stimulus way the total set of memory elements is classified with
sampling theory (Estes, 1950). This approach is based respect to alternative responses. The model is thus
on the ideas that even apparently simple forms of able to handle the fact that learning is under some
learning entail the formation of associations among conditions slow and gradual but under other condi-
representations in memory of many aspects of the sit- tions fast and discontinuous. In a special case of the
uation confronting the learner, and that more com- model, it is assumed that the learner perceives the en-
plex forms are largely a matter of classifying or reclas- tire stimulus pattern present on a trial as a unit, so
sifying already existing associations (now more often that the system has only two states, one in which the
termed memory elements). In the earliest versions of the unit is associated with the correct response and one
362 MCGEOCH, JOHN A.
in which it is not. Surprisingly at the time, this ultra- new work applying and extending what was originally
simplified model was found to account in detail for conceived as a model of animal conditioning to a vari-
the data of experiments on the learning of elementa- ety of work on human learning and categorization.
ry associations (as between faces and names) and con- Thus, by the recurring process of rediscovery and re-
cept identification, even for human learners (Bower, interpretation, mathematical learning theory has be-
1961). come active in a new guise as a component of the new
The stimulus-sampling model automatically ac- discipline of cognitive science.
counts for the phenomenon of generalization, which See also: DISCRIMINATION AND GENERALIZATION;
required a special postulate in Hulls system. If learn- HEBB, DONALD; HULL, CLARK L.; LEARNING
ing has occurred in one situation, the probability that THEORY: A HISTORY; LEARNING THEORY:
the learned response will transfer to a new situation CURRENT STATUS; SPENCE, KENNETH
is equal to the proportion of aspects of the second sit-
uation that were sampled in the first. The sampling Bibliography
model has the serious limitation that it cannot explain Bower, G. H. (1961). Application of a model to paired-associate
learning. Psychometrika 26, 255280.
how perfect discriminations can be learned between Bush, R. R., and Mosteller, F. (1955). Stochastic models for learning.
situations that have aspects in common. However, this New York: Wiley.
deficiency was remedied in a new version formulated Estes, W. K. (1950). Toward a statistical theory of learning. Psycho-
by Douglas L. Medin and his associates especially for logical Review 57, 94107.
the interpretation of discrimination and categoriza- Gluck, M. A., and Bower, G. H. (1988). From conditioning to cate-
gory learning: An adaptive network model. Journal of Experi-
tion (Medin and Schaffer, 1978). With a new rule for
mental Psychology: General 117, 225244.
computing similarity between stimuli or situations, Gulliksen, H. A. (1934). A rational equation of the learning curve
this formulation, generally known as the exemplar- based on Thorndikes law of effect. Journal of General Psycholo-
memory model, has been shown to account for many gy 11, 395434.
properties of human category learning in simulated Hebb, D. O. (1949). Organization of behavior: A neurophysiological the-
ory. New York: Wiley.
medical diagnosis and the like.
Hull, C. L. (1943). Principles of behavior. New York: Appleton.
Kamin, L. J. (1969). Predictability, surprise, attention, and condi-
tioning. In B. A. Campbell and R. M. Church, eds., Punishment
Connectionism and aversive behavior. New York: Appleton-Century-Crofts.
McClelland, J. L., and Rumelhart, D. E. (1986). Parallel distributed
An interesting aspect of science is the way prog-
processing: Explorations in the microstructure of cognition, Vol. 2.
ress often occurs by seeing older theories in a new Cambridge, MA: MIT Press.
light. A striking instance is the observation that Res- Medin, D. L., and Schaffer, M. M. (1978). Context theory of classi-
corla and Wagners extension of Hulls learning theo- fication learning. Psychological Review 85, 207238.
ry can be viewed as a special case of a type of adaptive Rescorla, R. A., and Wagner, A. R. (1972). A theory of Pavlovian
conditioning: Variations in the effectiveness of reinforcement
network model that has been imported from engi-
and non-reinforcement. In A. H. Black and W. F. Prokasy,
neering into the new branch of cognitive science eds., Classical conditioning, Vol. 2: Current research and theory.
known as connectionism (Rumelhart and McClelland, New York: Appleton-Century-Crofts.
1986). A connectionist network is composed of layers Rumelhart, D. E., and McClelland, J. L. (1986). Parallel distributed
of elementary units (nodes) interconnected by path- processing, Vol. 1. Cambridge, MA: MIT Press.
ways. When the network receives an input at the bot- W. K. Estes
tom layer, as from the stimulus display in a learning
situation, activation is transmitted through the net-
work to output nodes at the top layer that lead to re-
sponse mechanisms. During learning, memories are
MCGEOCH, JOHN A. (18971942)
stored not in individual units but as distributions of
strengths of connections in groups of units. These John A. McGeoch (18971942) was the single most
strengths are modified during learning by a process seminal figure in defining the research area known as
that is driven by an overall tendency toward error cor- verbal learning. Trained in the Chicago functionalist
rection. Connectionist models have been developed tradition by Harvey A. Carr, McGeoch received the
and applied for the most part in relation to complex Ph.D. in 1926 with a dissertation titled A Study in the
cognitive activities such as speech perception and lan- Psychology of Testimony. His professional career
guage acquisition (McClelland and Rumelhart, 1986), was notable for its meteoric advancement, its geo-
but it has also been discovered that the descendant of graphic mobility, and its massive research energy.
Hulls learning theory embodied in the Rescorla- After rising through the academic ranks at Washing-
Wagner model can be viewed as a special case of a ton University in St. Louis, he became professor at the
connectionist network (Gluck and Bower, 1988). The University of Arkansas in 1928, at the age of 31. Two
consequence of this observation has been a wave of years later he went as chairman of the Psychology De-
MCGEOCH, JOHN A. 363
partment to the University of Missouri, a position he

subsequently filled at Wesleyan University and at the
University of Iowa (Pratt, 1943; Wolfle, 1943).
McGeochs major work was The Psychology of
Human Learning: An Introduction (1942). It was not in-
tended as such: He had been at work, during the
1930s, with the collaboration of Carr, on a projected
two-volume manual covering all published work on
human learning. The first draft of the manual was
about 80 percent complete in 1936, and had received
chapter-by-chapter feedback from Carr (see Carrs in-
troduction to McGeoch, 1942). After several years
delay, caused by ill health and frequent moves, Mc-
Geoch was persuaded by a publisher, on assuming the
Iowa chairmanship, to write a digest of the manu-
script, intended for use as a textbook. McGeochs im-
pact on the field has come through the comprehen-
siveness and authority of this book and its revision a
decade later by his student Arthur L. Irion (McGeoch
and Irion, 1952).
McGeochs reputation rests additionally on a the- John A. McGeoch (Psychology Archives, University of Akron)
oretical assertion and a related empirical corpus. The
theoretical assertion is in a paper (McGeoch, 1932)
that argues that the process of forgetting, which was In the context of McGeochs total career, this rep-
by then his special commitment, could not in princi- utation is largely erroneous: His dissertation was a se-
ple be explained by the mere passage of time, as was ries of field studies on memory as testimony in which
asserted by the decay theory of forgetting. Rather, he McGeoch tested children (ages 9 to 14) in intact class-
said, activities that occurred during that time must be rooms within the East St. Louis school system (Mc-
responsible. It follows that to penetrate the mecha- Geoch, 1928a, 1928b, 1928c). Among many other
nism of forgetting, one should study retroactive inhi- comparisons, he was interested in the relation be-
bition. McGeochs most visible research activities re- tween accuracy and intelligence, which he operation-
spected that logical agenda. alized by contrasting normal and institutionalized
children and by the administration to each group of
His scientific style remained close to the data and the Army Alpha test (McGeoch, 1928b). An earlier
was conservative with respect to theoretical leaps. Mc- publication (McGeoch, 1925a) had compared (1) the
Geoch believed that the psychology of human learn- staged event (eyewitness reality) with (2) objects or (3)
ing and memory was in a basically pretheoretical words as stimuli. Thus, McGeoch may be among a
phase, in which the field was responsible for deter- handful of pioneers in developing the study of eyewit-
mining, first of all, what the data were before propos- ness testimony in American psychology (all but one of
ing elaborate explanations of them. Postman (1972) the references he cited in this work were from Ger-
has described this and other characteristics of the man laboratories). Nor was this work on practical as-
functionalist attitude McGeoch received from such pects of memory isolated: Others of his early publica-
predecessors as Carr and J. R. Angell and passed tions concerned memory for poetry (Whitely and
along to such students as A. W. Melton (an under- McGeoch, 1928), emotional measurement (McGeoch
graduate at Washington University) and B. J. Under- and Bunch, 1930), and the relation between suggest-
wood (a graduate student at Iowa). The austerity of ibility and juvenile delinquency (McGeoch, 1925b).
his book, the programmatic drive of his experiments The catholicism of his original research interests is
on retroactive inhibition, and his atheoretical bias es- solidly within the functionalist tradition. One specula-
tablished McGeoch as an archetype for the contrived tion is that McGeoch later turned to the precision and
laboratory approach to human learning and memory.
control of experimental studies because of his com-
In the more recent days of cognitive psychology, nei-
mitment to finding answers for everyday problems in
ther McGeochs reputation nor the reputation of ver-
human life, not in order to escape from them.
bal learning has fared well. The latter tradition is held
up as an example of the sterility into which orthodox
science can pass. See also: INTERFERENCE AND FORGETTING
364 MEASUREMENT OF MEMORY
Bibliography Wechsler Adult Intelligence Scale. Span can also be

McGeoch, J. A. (1925a). The fidelity of report of normal and sub- measured using items other than digits: letters or
normal children. American Journal of Psychology 36, 434445. words, for example. However, it should be noted that
(1925b). The relationship between suggestibility and intelli-
gence in delinquents. Psychological Clinic 6, 133134.
the span measure obtained may vary with the kinds
(1928a). The influence of sex and age upon the ability to of items employed. Thus word span is typically some-
report. American Journal of Psychology 40, 458466. what shorter than digit span, and word span itself will
(1928b). Intelligence and the ability to report. American vary depending on various features of the words, such
as their length and familiarity.
(1928c). The relation between different measures of the
ability to report. American Journal of Psychology 40, 592596.
(1932). Forgetting and the law of disuse. Psychological Re- Free Recall
view 39, 352370. In free recall, a list of items is presented and the
(1942). The psychology of human learning. New York: Long-
task is to recall them in any order. The measure of
mans, Green.
McGeoch, J. A., and Bunch, M. E. (1930). Scores in the Pressey memory is the number (or proportion) of these items
X-O tests of emotion are influenced by courses in psychology. recalled. If people are allowed to begin recalling im-
Journal of Applied Psychology 14, 150159. mediately after the presentation of the last item, this
McGeoch, J. A., and Irion, A. L. (1952). The psychology of human
simple measure is subject to a strong recency effect:
learning. New York: Longmans, Green.
Postman, L. (1972). The experimental analysis of verbal learning The last few items in the list will be recalled very well
and memory: Evolution and innovation. In C. P. Duncan, L. (and usually recalled first). There is also a weaker pri-
Seechrest, and A. W. Melton, eds., Human memory: Festschrift macy effect: The first few items in the list will be better
for Benton J. Underwood, pp. 123. New York: Appleton- recalled than those in the middle positions. In some
Century-Crofts.
Pratt, C. C. (1943). John A. McGeoch: 18971942. American Journal applications these serial position effects are eliminat-
of Psychology 56, 134136. ed by inserting buffer items at the beginning and
Whitely, P. L., and McGeoch, J. (1928). The curve of retention for end of the list that are not counted as part of the recall
poetry. Journal of Educational Psychology 19, 471479. measure.
Wolfle, D. (1943). McGeochs psychology of human learning: A
special review. Psychological Bulletin 40, 350353.
Cued Recall
Robert G. Crowder
In cued recall, the task is to recall each item in re-
sponse to a cue provided by the tester. This cue may
have been presented along with the item at the time
of study (an intra-list cue) or be an item not studied
MEASUREMENT OF MEMORY before (an extra-list cue). For example, a participant
There are many techniques for measuring memory. may study word pairs such as dog-tail and when tested
Some methods are highly specialized, either for the be given the intralist cue dog and asked to recall the
purpose of addressing a particular research question word with which it was paired. Alternatively, he or she
or for diagnosing a specific memory dysfunction, but may study a list of single words (including tail but not
most measurement procedures are variants of the dog) and when tested be given the extra-list cue dog as
basic techniques described in this entry. a potential aid to recall. A special case of extra-list
cuing is category cuing. If the list contained items
Measures of Recall from several different categories (fruits and animals),
cued recall could take the form of providing the par-
Immediate Serial Recall: Memory Span ticipant with the category labels as cues.
In the most commonly used version of this task,
a randomly ordered sequence of digits (e.g., 4-7-8-2- A potential difficulty in measuring cued recall is
5-9) is read once to the participant, who is required guessing. This problem arises when the items to be re-
to repeat them in the same order. The resulting mea- called have strong prior associations with the cue. For
sure is known as the digit span, defined as the number example, suppose the word orange was studied and at
of digits that can be repeated in the correct order recall the extra-list cue fruit is provided. The partici-
without error. An immediate difficulty is that an indi- pant may have forgotten the item orange but offer it
viduals performance may fluctuate slightly from one merely because it is a plausible guess to the cue fruit.
occasion to the next, six digits being correctly repeat- The usual solution to this difficulty is to obtain a base-
ed on one occasion, for example, and seven on the line measure of how often the item is wrongly recalled
next. The formal measurement of span is therefore in response to the cue when it was not on the study
usually taken as the number of digits that can be cor- list. In this example, the control measure for the word
rectly recalled on 50 percent of occasions. The digit orange would be its recall rate to the cue fruit when or-
span test is found in such standard test batteries as the ange was not part of the study list.
MEASUREMENT OF MEMORY 365
Measures of Recognition Memory Clearly what is needed is some way of adjusting

the hit rate to take into account criterion differences
Forced-Choice Recognition reflected in the false alarm rate. One method is to
In this procedure memory is measured by pre- take as a score the difference between hits and false
senting each of the previously studied items (the alarms, a procedure that has little theoretical justifica-
old items) with one or more new items or lures tion but offers a simple, and in many circumstances
and instructing the participant to choose which of adequate, measure. A more sophisticated method is
these items is old. The measure is then the number to make use of a model known as the theory of signal
or proportion of items correctly identified as old. detection, a decision model taken from psychophys-
There are two difficulties with this measure. Guessing ics. The model yields a measure, termed d (d-prime),
poses an obvious problem because in the case of a that is independent of the participants criterion and
two-alternative forced choice, someone who remem- can be interpreted as a measure of the ability to dis-
bered nothing at all could guess correctly half the criminate difference in subjective familiarity between
time. Increasing the number of lures reduces the ex- the old and the new items.
pected rate of correct guessing but does not eliminate
the problem entirely. Researchers have proposed var- Other Measures of Memory
ious methods of correcting for guessing, but they
are often not used, partly because the problem of Measures of Prospective Memory
guessing is not as serious as it may seem. Normally Prospective memory is remembering to perform
one is interested not in the absolute number of items an action at a future point in time: to follow an in-
recognized but in the comparison of recognition rates struction to buy milk on the way home from work, for
across different occasions or conditions. If the influ- example. In order to measure prospective memory it
ence of guessing is the same across these conditions, is necessary to designate the point at which the future
then the differences in the recognition rates between act of remembering is to be performed. In this regard
conditions will provide an adequate comparative it is possible to distinguish between time-based and
measure. event-based tasks. In time-based prospective remem-
bering, an action must be performed at a specified
A second potential source of difficulty is the na-
time in the future, such as remembering to turn off
ture of the lures. A recognition test can be made more
the oven in ten minutes. In event-based prospective
or less difficult by altering the degree of similarity be-
remembering, an action must be performed contin-
tween the correct (old) items and the lures. Thus er- gent on the occurrence of some other event, such as
rors are more likely if the lure is a synonym of the old remembering to follow an instruction to lock the door
item than if it is an unrelated word chosen at random. on the way out.
Single-Item (Yes/No) Recognition Tests Measures of Implicit Memory
In what may seem the simplest form of a recogni- Implicit measures are obtained through observ-
tion test, participants are shown each test item in turn ing performance on a task that indirectly reveals the
and asked to respond yes if they have seen it before influence of past experience. A common example of
(an old item) and no if they have not (a new item). an implicit memory test is word-fragment comple-
The test list contains a mixture of old and new items. tion. Suppose a participant has seen a list that in-
A possible measure of memory would be the propor- cludes the word assassin. Some time later the partici-
tion of items correctly identified as old, a measure re- pant is given a fragment completion test consisting of
ferred to as the hit rate. However, this measure has some but not all of the letters of the word (e.g., aa
a serious shortcoming: It will be influenced by the in) and asked to find the word that could be formed
participants criterion for saying yes. Adopting a by filling in the blanks. No reference is made to the
lax criterion (that is, saying old even if the item is prior list. The essential aspect of an implicit memory
only faintly familiar) can yield a high hit rate, usually test is that participants receive no instructions to re-
at the expense of mistakenly saying old to a large member items from the prior list, nor are they in-
number of new items. Such errors are termed false formed of the lists possible relevance. They may
positives or false alarms. Thus, from the participants therefore be quite unaware that their performance is
point of view, responding in a yes/no recognition test being influenced by a past experience. The measure
becomes a trade-off between hits and false alarms. of implicit memory is the improvement in perfor-
The difficulty from a measurement point of view is mance on the word-fragment completion task. Im-
that different participants may adopt different provement is measured relative to word-fragment
criteria, so that comparing hit rates alone can be very performance obtained under control conditions in
misleading. which the solution word (assassin) was not on the prior
366 MEMORY CONSOLIDATION: Molecular and Cellular Processes
study list. Among the other tasks that can be used to nisms required for triggering the intraneuronal syn-
measure implicit memory is stem completion. In this aptic processes underlying the initial stages of
task participants are given the initial letters of a word memory, little is known about the mechanisms that
and asked to add letters to complete a word. The consolidate memories. Among the processes most in-
major difficulty with implicit measures is to ensure tensively studied, mechanisms that regulate tran-
that the participant is not making use of explicit scription seem to have a clear role in memory consoli-
memory strategies. One way to achieve this is to predation. Also, many studies have demonstrated that
vent participants from becoming aware of the rela- memory consolidation involves multiple brain sys-
tionship between the initial study and subsequent test tems. For example, while the hippocampus has a criti-
phases of the procedure. Another is to instruct partic- cal role in the initial stages of memory consolidation,
ipants that they should avoid conscious strategies by remote memories seem to be dependent on cortical
responding with the first thing that comes to mind. storage sites.
The Relationship among the Different Memory and Protein Synthesis

Measures Evidence from a variety of systems and organisms
The various measurement procedures described demonstrates that protein synthesisduring or
in this entry are not merely alternative ways of esti- shortly after trainingis essential for the formation
mating a single true quantity that could be thought of long-term memory (LTM) (Davis and Squire,
of as the amount remembered. In this regard the 1984). For example, the protein synthesis inhibitor
measurement of memory is very different from the anisomycin, given systemically before or immediately
measurement of physical quantities such as length or after training, blocks LTM (typically tested twenty-
weight. Various techniques for measuring the dis- four hours after training) but not short-term memory
tance between two points should yield the same result, (STM; 30120 minutes after training) tested in a wide
and any variation should be regarded as reflecting er- spectrum of behavioral tasks. Anisomycin can also
rors of measurement. The situation is quite different block the later phases of long-term potentiation
in the case of memory measurement. For example, in (LTP) without affecting its early stages (Frey et al.,
evaluating memory for an event, a recognition test 1993). LTP is an experimental model of the synaptic
and a recall test should not be thought of as alterna- plasticity mechanisms underlying learning and mem-
tive methods of measuring a single ideal quantity (the ory.
amount remembered) but as different measures of Unfortunately, most of the protein-synthesis in-
memory performance on two distinct, although possi- hibitor drugs have a variety of side effects that compli-
bly related, tasks. Much research suggests that under cate the interpretation of the findings. Protein-
many circumstances different measures of memory synthesis inhibitors can make animals sick, and some
are quite dissociated or uncorrelated, reflecting the of these inhibitors have nonspecific effects on the le-
fact that different measurement procedures tap dif- vels of housekeeping proteins necessary for cell
ferent aspects of the memory system. health. Thus, the dramatic effects of protein-synthesis
inhibitors on memory triggered a search for the tran-
See also: IMPLICIT MEMORY; MEMORY SPAN
scription factors that direct the gene expression re-
Bibliography quired for memory. Transcription factors bind to spe-
Tulving, E., and Craik, F. I. M., eds. (2000). The Oxford handbook cific sequences in the promoters of genes and serve
of memory. New York: Oxford University Press. as attachment sites for the machinery that transcribes
Robert S. Lockhart genes. Alterations of these transcription factors have
far more specific biological effects than the transcrip-
tion and translation inhibitors used to study memory.
MEMORY CONSOLIDATION: CREB and Memory

MOLECULAR AND CELLULAR Several lines of evidence from studies with multi-
PROCESSES ple organisms and brain systems demonstrate that the
Memory is a complex biological process involving cAMP Responsive Element Binding Protein (CREB)
multiple brain systems, each with a specialized func- is one of the transcription factors regulating the syn-
tion, and many molecular and cellular mechanisms thesis of proteins necessary for the formation of LTM.
that process and consolidate information in the brain. Learning activates several signaling pathways, in-
Although studies in recent years have made consider- cluding cAMP, CaMKIV and MAPK signaling cas-
able inroads into the molecular and cellular mecha- cades. Activation of these signaling cascades seems to
MEMORY CONSOLIDATION: Molecular and Cellular Processes 367
lead to the phosphorylation and activation of CREB. Downstream Targets of CREB and Their
Phosphorylated CREB bound to cAMP Responsive Role on Memory
Elements (CRE sites) in the promoter of specific
CREB controls the expression of other proteins
genes, such as 14-3-3 eta, can then bind to a complex
that seem to regulate the complex molecular and cel-
of proteins that transcribes those genes. Insights into lular events underlying long-term memory. For ex-
the biochemistry of CREB directed the design of example, multiple pulses of serotonin activate CREB in
periments that tested the impact on memory of both Aplysia and induce a set of genes, many of which may
increases and decreases in CREB function. regulate synaptic plasticity and memory. Researchers
have identified several synaptic plasticity candidate
In a wide range of species, including Aplysia, Dro-
genes, including the CCAAT enhancer-binding pro-
sophila, song birds, mice, and rats, CREB-dependent
tein, ApC/EBP. ApC/EBP contains CRE sites in the
transcription has been demonstrated to be crucial for promoter region, is rapidly induced by cAMP, and ex-
the formation of LTM. Furthermore, the levels of ac- hibits properties consistent with an immediate early
tive CREB may also be an important determinant of gene. Low levels of ApC/EBP are present in unsti-
the amount and schedule of training required for mulated sensory neurons, but there are higher levels
LTM. Higher levels of CREB facilitate LTM, while during the initial phase of long-term facilitation
LTM formation in animals with lower CREB levels re- (LTF), a synaptic model of LTM in Aplysia. Decreas-
quires more training with longer intervals between ing the expression of ApC/EBP selectively blocks the
trials. The involvement of CREB in memory forma- formation of LTF but not short term facilitation or
tion does not seem to be restricted to certain forms STF (Alberini et al., 1994). Remarkably, inhibitory
of memory. Tests as diverse as olfactory conditioning avoidance training in rats induces two homologs of
in flies, fear conditioning, spatial memory, condi- ApC/EBP (C/EBP and C/EBP ), and down-
tioned taste aversion, social recognition, and social regulation of C/EBP with antisense oligonucleotides
transmission of food preferences in rodents demon- leads to deficits in long- but not in short-term memo-
strate the involvement of CREB in memory forma- ry in rats tested in this fear conditioning task (Tau-
tion. Similarly, CREB has a role in the stability of syn- benfeld et al., 2001).
aptic changes in a variety of species, including Aplysia, C/EBP may regulate another transcription factor,
rats, and mice. Furthermore, CREB also has a role in ApAF (Aplysia Activating Factor) (Bartsch et al., 2000),
other forms of plasticity, including topographical cor- which shares homology with the mammalian PAR
tical reorganization and circadian rhythms (Silva et family of transcription factors. Injection of recombi-
al., 1998)). nant ApAF into a cocultured Aplysia neuronal prepa-
ration converts the STF normally produced by one
There is compelling evidence that cAMP- pulse of serotonin into LTF. Additionally, injection of
dependent signaling can activate CREB (Silva et al., antibodies against ApAF or injection of a dominant
1998). Deleting both calcium/calmodulin induced ad- negative form of ApAF that contains only the bZIP
enyl cyclases (1 and 8), enzymes that generates cAMP, domain blocks the LTF produced by five pulses of se-
block the maintenance but not the induction of LTP rotonin.
in the CA1 region. Similarly, these genetic manipula- Zif268 is another transcription factor whose ex-
tions also cause deficits in hippocampal long-term pression seems to be regulated by CREB. The pro-
memory. Pharmacological activation of adenyl cy- moter region of Zif268 contains two CRE sites and six
clases turns on CREB and rescues the LTP and mem- SRE sites. The levels of Zif268 are upregulated short-
ory deficits of this mutant (Wong et al., 1999). Anoth- ly following LTP-inducing stimulation. Furthermore,
er kinase that can phosphorylate and activate CREB mutant mice lacking Zif268 show specific deficits in
is calcium/calmodulin kinase IV (CaMKIV). Experi- the late phases of LTP. Similarly, LTM, but not STM
ments with a dominant-negative form of CaMKIV tested with a variety of tasks, is abnormal in these mu-
(dnCaMKIV) expressed in the postnatal forebrain tants (Jones et al., 2001). Like the CREB-/- mutant
showed that this kinase has a role in the later stages mice, the Zif268 mutants show deficits in spatial
of LTP. Similarly, behavioral tests indicated that this learning in the water maze following massed training,
kinase modulates long-term memory formation: the and these deficits are rescued by spaced training.
dnCaMKIV mice showed specific long-term memory
deficits in both spatial learning and contextual fear
conditioning tasks. CREB activation was impaired Hippocampal/Neocortical Interactions
by the dnCaMKIV transgene, confirming that this ki- During Memory Consolidation
nase activates CREB during learning (Kang et al., Memory consolidation involves the molecular
2001). cascades described above, but it also requires multiple
368 MEMORY CONSOLIDATION: Molecular and Cellular Processes
brain regions. A number of studies have indicated and experiences. Instead, there is a overwhelming
that whereas memory is initially stored in the hippo- amount of data that show that memory is a construc-
campus, eventual storage takes place in cortical net- tive and dynamic process, designed to fine tune be-
works. Studies examining patients with hippocampal havioral responses and not to faithfully record experi-
damage have found a loss of memory called tempo- ences. The very usefulness of memory has to do with
rally-graded retrograde amnesia (Squire, 1992): Pa- this ability to extract and generalize from experience
tients with hippocampal damage suffer severe amne- adaptive responses and rules that improve survival
sia for memories that are a few years old at the time and fitness. It is not useful for the mouse to remember
of injury (recent memory) but not for memories that that a particular cat is dangerous. It is far more useful
are many years old (remote memory). Similar find- for it to learn that cats in general are dangerous.
ings have been made in rodents. For example, con- Moreover, a number of studies show that the very pro-
textual fear conditioning is severely disrupted by hip- cess of recall can bring memories to a labile state in
pocampal lesions made one day, but not several which they can be edited or reconsolidated (Sara,
weeks, after training. These studies suggest that the 2000). Some of the very molecular mechanisms that
hippocampus processes new memories but that they lay down memories in the first place may be called
are eventually stored elsewhere, presumably the cor- upon during recall to edit these memory traces. Thus,
tex. blocking protein synthesis during recall interferes
with memory just as dramatically as blocking it during
A study examining patients with cortical brain training (Nader et al., 2000). Similarly, blocking
damage showed that they have recent but not remote CREB function during recall interferes with memory
memory, a result that is consistent with the theory just as blocking it during training (Kida et al., 2002),
that memory is only temporarily stored in the hippo- an eloquent demonstration of the constructive es-
campus and that it is eventually consolidated in the sence of recall and of the fragile nature of memory.
cortex. Interestingly, imaging brain function with the
deoxyglucose technique showed that metabolic activi-
ty was high in the hippocampus when maze learning See also: APLYSIA: MOLECULAR BASIS OF LONG-TERM
was recent, but relatively high in the cortex when SENSITIZATION; PROTEIN SYNTHESIS IN LONG-
TERM MEMORY IN VERTEBRATES; SECOND
maze learning was remote (Bontempi et al., 1999).
MESSENGER SYSTEMS
Performance on the maze correlated only with hippo-
campal metabolism for recent memory and only with
cortical metabolism for remote memory. Bibliography
Alberini, C. M., Ghirardi, M., Metz, R., and Kandel, E. R. (1994).
A study using molecular genetic techniques came C/EBP is an immediate-early gene required for the consolida-
to a strikingly similar conclusion (Frankland et al., tion of long-term facilitation in Aplysia. Cell 76, 1,0991,114.
2001). In the mutants studied, LTP is normal in the Bartsch, D., Ghirardi, M., Casadio, A., Giustetto, M., Karl, K. A.,
Zhu, H., and Kandel, E. R. (2000). Enhancement of memory-
hippocampus but deficient in the cortical regions related long-term facilitation by ApAF, a novel transcription
thought to store memory. Remarkably, these mice factor that acts downstream from both CREB1 and CREB2.
showed normal memory for up to three days after Cell 103, 595608.
training but dramatic forgetting over longer reten- Bontempi, B., Laurent-Demir, C., Destrade, C., and Jaffard, R.
tion intervals (ten to fifty days). These data suggest (1999). Time-dependent reorganization of brain circuitry un-
derlying longterm memory storage. Nature 400, 671675.
that the abnormal LTP in the cortex interfered with Davis, H. P., and Squire, L. R. (1984). Protein synthesis and memo-
the storage of memory there and that the normal LTP ry. Psychology Bulletin 96, 518559.
in the hippocampus may explain why memory was Frankland, P., OBrien, C., Ohno, M., Kirkwood, A., and Silva, A.
normal for up to three days after training. It is strik- (2001). CaMKII-dependent plasticity in the neocortex is re-
ing that a wide range of studies, ranging from genetic quired for remote memory. Nature 411, 309313.
Frey, U., Huang, Y.-Y., and Kandel, E. R. (1993). Effects of cAMP
and imaging studies in mice to neuroanatomical le- simulate a late stage of LTP in hippocampal CA1 neurons.
sions in rats, monkeys, and humans all demonstrate Science 260, 1,6611,664.
this striking interaction between the hippocampus Jones, M. W., Errington, M. L., French, P. J., Fine, A., Bliss, T. V.,
and the neocortex in memory consolidation. Garel, S., Charnay, P., Bozon, B., Laroche, S., and Davis, S.
(2001). A requirement for the immediate early gene Zif268 in
the expression of late LTP and long-term memories. Nature
Editing Memory During Recall Kang, H., Sun, L. D., Atkins, C. M., Soderling, T. R., Wilson, M.
Thus, memory goes through several stages where A., and Tonegawa, S. (2001). An important role of neural ac-
tivity-dependent CaMKIV signaling in the consolidation of
information is consolidated with different molecular
long-term memory. Cell 106, 771783.
mechanisms and in different brain structures. Howev- Kida, S., Josselyn, S., Ortiz, S., Kogan, J., Chevere I., Masushige,
er, it would be misleading to think of animal memory S., and Silva, A. (2002). CREB required for the stability of new
as computer memory, a permanent etching of events and reactivated fear memories. Nature Neuroscience 5, 34855.
MEMORY CONSOLIDATION: Prolonged Process of Reorganization 369
Nader, K., Schafe, G. E., and LeDoux, J. E. (2000). The labile na- that the former refers to molecular and cellular events
ture of consolidation theory. Nature Reviews Neuroscience 1, within neurons and the latter to what retrograde am-
216219.
Sara, S. J. (2000). Retrieval and reconsolidation: Toward a neuro-
nesia has suggested about the brain systems involved
biology of remembering. Learning and Memory 7, 7384. in elaborating long-term memory. When one speaks
Silva, A. J., Kogan, J. H., Frankland, P. W., and Kida, S. (1998). about prolonged processes of reorganization (or pro-
CREB and memory. Annual Review of Neuroscience 21, 127 longed consolidation), one is usually referring to this
148. latter usage.
Squire, L. R. (1992). Memory and the hippocampus: A synthesis
from findings with rat, monkeys, and humans. Psychology Re- There are also other ways in which memory can
view 99, 195231. undergo gradual change and reorganization across
Taubenfeld, S. M., Milekic, M. H., Monti, B., and Alberini, C. M. time that may be related to consolidation. For exam-
(2001). The consolidation of new but not reactivated memory
requires hippocampal C/EBPbeta. Nature Neuroscience 4, 813
ple, forgetting itself operates over a long period, with
818. the result that details are lost and one is left with ker-
Wong, S. T., Athos, J., Figueroa, X. A., Pineda, V. V., Schaefer, M. nels of the past, the central meanings. One can forget
L., Chavkin, C. C., Muglia, L. J., and Storm, D. R. (1999). Cal- the particulars while abstracting and retaining the
cium-stimulated adenylyl cyclase activity is critical for hippo- main points. Thus, forgetting is not a passive process
campus-dependent long-term memory and late phase LTP.
Neuron 23, 787798.
of progressive loss and dissolution but a dynamic pro-
cess during which representations of the past within
Alcino Silva neocortex are continually reconstructed and reorga-
nized. Events such as rehearsal and episodes of new
learning also influence the structure of already estab-
lished representations. These continuing influences
MEMORY CONSOLIDATION: on the structure of long-term memory are beyond the
PROLONGED PROCESS OF scope of this article.
REORGANIZATION
The origin of the concept of memory consolidation is A Brain System for Memory in the
generally credited to Georg Elias Mller and his stu- Temporal Lobe
dent Alfons Pilzecker. Their 300-page monograph, In 1957 Scoville and Milner described the pro-
published in 1900, proposed that memory is not found effects on memory following bilateral removal
formed instantaneously at the time of learning but of the medial temporal lobe in a patient who became
takes time to be fixed (or consolidated). The studies known as H.M. Comprehensive study of this patient
involved lists of nonsense syllables and focused espe- established the fundamental principle that the ability
cially on retroactive inhibition, the finding that when to acquire new memories is a distinctly cerebral func-
two lists are learned in succession, learning the section, separable from other perceptual and intellectual
ond list interferes with subsequent recall of the first functions. This discovery inspired efforts to develop
list. On the basis of this finding, they suggested that an animal model of human memory impairment in
the processes needed to form memory continue for a the monkey and led eventually to the identification of
period of time after learning, during which time they the hippocampus and adjacent, anatomically related
are vulnerable to interference. While this origin of the structures in the medial temporal lobe as components
consolidation concept is widely known, it is not gener- of a memory system important for the formation of
ally known that the interval across which the putative long-term memory (Squire and Zola-Morgan, 1991).
consolidation process operated in these early experi- Disrupting the function of this system impairs the
ments was less than ten minutes. ability to form declarative memories (the ability to ac-
The consolidation concept subsequently came quire new facts and events) while leaving intact a col-
into widespread usage and found application in a lection of nondeclarative memory abilities, including
number of contexts. The term is often used to refer habits and skills, simple forms of conditioning, and
to the cascade of molecular events, including protein other means by which experience can change how
synthesis, that unfolds during the hours after learning one interacts with the world. Disruption of medial
and ultimately results in morphological growth and temporal lobe function also causes retrograde amne-
change at synapses. The term is also used in the con- siathat is, the loss of memories that were acquired
text of observations that damage to the hippocampus before the onset of amnesia. The finding that retro-
or related structures can produce temporally graded grade amnesia is typically temporally graded (most
retrograde amnesia: impairment of memories that severe for recent events and less severe for remote
were acquired weeks, months, and even years earlier, events) suggested that medial temporal lobe struc-
despite sparing of more remote memories. There tures are essential at the time of learning as well as
need be no confusion among these usages if one notes after learning during a prolonged and gradual period
370 MEMORY CONSOLIDATION: Prolonged Process of Reorganization
of reorganization and consolidation. During consoli- Uncertainties about Temporally Graded

dation, long-term memory is gradually stabilized Retrograde Amnesia
within the neocortex.
Some uncertainties remain about temporally
graded retrograde amnesia and its interpretation.
Temporally Graded Retrograde Amnesia First, there is not yet agreement as to whether spatial
memories undergo the same kind of transition as
The earliest evidence for temporally graded ret- nonspatial memories dofrom a form that depends
rograde amnesia came from clinical observations of on the hippocampus and related structures to a form
human patients recorded more than 100 years ago. that is independent of these structures. On the one
Quantitative studies of retrograde amnesia began in hand, the available data from humans suggest that
the 1970s. For example, on a test of former one- spatial and nonspatial memory have the same status.
season television programs, designed to permit the For example, a profoundly amnesic patient with large
equivalent sampling of past time periods, psychiatric medial temporal lobe lesions was able to recall the
patients prescribed bilateral electroconvulsive thera- spatial layout of the neighborhood where he had lived
py (ECT) for depressive illness exhibited temporally- as a child, and his memory in this respect was as good
graded retrograde amnesia covering about three as the memory of healthy individuals who had lived
years. Evidence relating temporally graded retro- as children in the same neighborhood (Teng and
grade amnesia to the medial temporal lobe has come Squire, 1999). On the other hand, the findings from
from patients for whom detailed neuropsychological experimental animals are less clear. Some studies of
information is available together with postmortem in- rodents have found retrograde amnesia for spatial
formation about which structures were damaged. memory to be temporally graded after hippocampal
Such case material has shown that retrograde amne- lesions (Ramos, 1998; Squire, Clark, and Knowlton,
sia is brief when damage is limited to the CA1 field 2001), but other studies have found memory to be af-
of the hippocampus (Zola-Morgan, Squire, and Ama- fected similarly across past time periods (Sutherland
ral, 1986) and more extensive, covering fifteen years et al., 2001).
or more, when the damage is more extensive in this
A second uncertainty about temporally graded
region (the CA fields, dentate gyrus, subiculum, and
retrograde amnesia is whether episodic memory (that
some cell loss in the entorhinal cortex) (Rempel-
is, the recollection of autobiographical events) be-
Clower, Zola, Squire, and Amaral, 1996). Despite re-
comes independent of the medial temporal lobe in
ports of temporally graded retrograde amnesia in
the same way that memory for general facts becomes
well-described patients, it is nevertheless the case that
independent. Some patients with medial temporal
such studies depend on retrospective methods that
lobe lesions appear able to recollect specific events
make it difficult to document precisely the time
from their remote past as well as healthy individuals
course of the phenomenon. Accordingly, when pro-
(Reed and Squire, 1997), whereas at least one patient
spective studies of retrograde amnesia in experimen-
with damage thought to be limited to the medial tem-
tal animals began in the 1990s, the nature of retro-
poral lobe is described as quite incapable of remote
grade amnesia, as well as the concept of
episodic recollections (Cipolotti et al., 2001). A third
consolidation, were placed on firmer ground.
source of uncertainty is that it is in principle possible
Of more than ten studies that have been carried to construct alternatives to a consolidation account of
out in mice, rats, rabbits, and monkeys, most have retrograde amnesia (Nadel and Moscovitch, 1997),
found temporally graded retrograde amnesia follow- although these ideas have difficulty accounting for
ing damage to the hippocampal region (Rempel- the available facts (Knowlton and Fanselow, 1998).
Clower, Zola, Squire, and Amaral, 1996; Squire,
Clark, and Knowlton, 2001). The retrograde amnesia
typically covers about one month. In addition, related Looking for Consolidation with Brain-
work suggests that retrograde amnesia can become Imaging Techniques
more extensive as damage extends beyond the hippo- One difficulty with the concept of gradual memo-
campus into adjacent medial temporal lobe struc- ry consolidation is that this idea has depended largely
tures. This pattern of findings implies that the hippo- on only one kind of evidence, namely, the evidence
campus itself is important for memory for a relatively of temporally graded retrograde amnesia. Some ef-
short period of time after learning and that the adja- forts have been made to find direct evidence for
cent perirhinal and parahippocampal cortices remain memory consolidation in neuroimaging studies (func-
important for a longer time. Ultimately, memory de- tional magnetic resonance imaging [fMRI]) that as-
pends on widespread areas of the neocortex and is in- sess medial temporal lobe activity while individuals
dependent of the medial temporal lobe. recollect recent and more remote memories. These
MEMORY CONSOLIDATION: Prolonged Process of Reorganization 371
studies have yielded mixed results (Ryan et al., 2001; Figure 1

Niki and Luo, 2002; Haist, Bowden Gore, and Mao,
2001; Maguire, Henson, Mummery, and Frith, 2001).
A difficulty in interpreting findings obtained with this
technique is that the neuroimaging method provides
information only about regions where activity corre-
lates with a particular cognitive operation and does
not provide information about which regions are es-
sential. Perhaps the medial temporal lobe is very
often engaged during memory retrieval but is essen-
tial only when recent memories are being retrieved.
Or perhaps activation of the medial temporal lobe
during recollection of recent and remote events is
driven to a large extent by the encoding into long-
term memory of the events of the test session, and by
the re-encoding of the recollections themselves.
It is notable that work in experimental animals
has found activity in the hippocampal region to differ
depending on the length of the retention interval.
Regional brain activity was mapped in mice tested five
days or twenty-five days after learning. Increasing the
retention interval from five to twenty-five days result-
ed in decreased hippocampal activity during reten-
tion testing (Bontempi, Laurent-Demir, and Jaffar,
1999). Additional studies in humans and experimen-
tal animals using neuroimaging techniques should be
of value. A neural network model of memory consolidation. (top)
Schematic diagram of the neural network model. Areas cortex 1
and cortex 2 represent association neocortex, and area MTL
Gradual Memory Consolidation represents the medial temporal lobe memory system. Each unit
in each area is reciprocally connected to all units in all other
Another reason that discussion continues about areas. A key feature of the model is that changes in the
the concept of memory consolidation is that the idea connections to and from the MTL area are fast and short-
itself is rather vague and depends on mechanisms not lasting, whereas changes in the connections between the
yet identified. The central idea is that medial tempo- neocortical areas are slow and long-lasting. Consolidation occurs
ral lobe structures direct the gradual establishment of when random activity in the MTL co-activates the stored
memory representations in the neocortex, and this patterns in cortex 1 and cortex 2, which results in strengthening
interaction between the medial temporal lobe and the of the cortico-cortical connections. (bottom) Performance of this
neocortex embodies the consolidation process. Some model in a retrograde amnesia experiment. The network learned
researchers have proposed that the hippocampus can two different patterns concurrently, and was tested on how well
it was able to reproduce the patterns after varying amounts of
serve as a temporary memory store because hippo-
time, given presentation of a part of each pattern. The
campal synapses change quickly. The neocortex can
Normal curve (thick line) shows the performance the intact
store information only gradually as an accumulation network. The Lesioned curve shows the performance of the
of small synaptic changes (McClelland, McNaughton, network when the MTL area was inactivated immediately prior
and OReilly, 1995). Consolidation occurs when the to testing.
hippocampus (and the system to which it belongs) re-
peatedly activates representations in the neocortex,
with the result that the neocortex can eventually sup- Rapid modification of neocortical representations
port memory independently of the hippocampal sys- would lead to instability, for example, if one rapidly
tem. introduced into a network the fact that penguins are
Computational considerations have suggested birds, but cannot fly, when birds are already repre-
that consolidation is important precisely because it sented in the network as animals that can fly. In such
enables the neocortex to slowly incorporate into its a case, McClelland and colleagues showed that the
representations the regularities of the environment, performance of the network would be markedly dis-
such as facts about the world (McClelland, McNaugh- rupted because the network begins to apply the char-
ton, and OReilly, 1995; OReilly and Rudy, 2001). acteristics of the penguin to other birds.
372 MEMORY CONSOLIDATION: Prolonged Process of Reorganization
One proposal is that the hippocampal system Knowlton, B. J., and Fanselow, M. S. (1998). The hippocampus,
stores only a compressed form of the memory but consolidation and on-line memory. Current Opinion in Neuro-
biology 8, 293296.
stores sufficient information to activate relevant sites
Maguire, E. A., Henson, R. N. A., Mummery, C. J., and Frith, C.
in the neocortex. Consolidation occurs when the neo- D. (2001). Activity in prefrontal cortex, not hippocampus, var-
cortex is repeatedly activated by the medial temporal ies parametrically with the increasing remoteness of memo-
lobe, either during rehearsal, reminiscence, or per- ries. NeuroReport 12, 441444.
haps as a result of spontaneous reactivation during McClelland, J. L., McNaughton, B. L., and OReilly, R. C. (1995).
sleep. When activation occurs, gradual and long- Why there are complementary learning systems in the hippo-
campus and neocortex: Insights from the successes and fail-
lasting changes occur in the connections between the
ures of connectionist models of learning and memory. Psycho-
cortical sites. Eventually, these connections become so logical Review, 419437.
strong that the medial temporal lobe is not needed to Mller, G. E., and Pilzecker, A. (1900). Experimentelle beitrage
recreate the original representation (see Figure 1). zur lehre vom Gedachtniss. Zeitschrift fur Psychologie, Ergan-
zungsband 1,1300.
Nadel, L., and Moscovitch, M. (1997). Memory consolidation, ret-
Conclusion rograde amnesia and the hippocampal complex. Current
The idea that memory can consolidate or reorga- Opinion in Neurobiology 7, 217227.
Niki, K., and Luo, J. (2002). An fMRI study on the time-limited
nize after learning in a lengthy process rests largely
role of the medial temporal lobe in long-term topographical
on the phenomenon of temporally graded retrograde autobiographical memory. Journal of Cognitive Neuroscience 14,
amnesianamely, the finding in studies of both hu- 500507.
mans and experimental animals that damage to the OReilly, R. C., and Rudy, J. W. (2001). Conjunctive representa-
medial temporal lobe can impair memories that were tions in learning and memory: Principles of cortical and hip-
acquired recently while sparing more remote memo- pocampal function. Psychology Review 108, 311345.
Ramos, J. M. J. (1998). Retrograde amnesia for spatial informa-
ries. These results have suggested that medial tempo-
tion: A dissociation between intra and extramaze cues follow-
ral lobe structures are required for a limited period ing hippocampus lesions in rats. European Journal of Neuro-
after learning, during which time they direct a pro- science 10, 3,2953,301.
cess of consolidation in the neocortex by gradually Reed, J. M., and Squire, L. R. (1997). Impaired recognition memo-
binding together the multiple, anatomically separate ry in patients with lesions limited to the hippocampal forma-
sites that together represent memory for a whole tion. Behavioral Neuroscience 111, 667675.
Rempel-Clower, N., Zola, S. M., Squire L. R., and Amaral, D. G.
event.
(1996). Three cases of enduring memory impairment follow-
Efforts to observe consolidation directly in ing bilateral damage limited to the hippocampal formation.
neuroimaging studies have yielded mixed re- Journal of Neuroscience 16, 5,2335,255.
sults.Uncertainties also remain concerning whether Ryan, L., Nadel L., Keil K., Putnam K., Schnyer D., Trouard T.,
and Moscovitch, M. (2001). Hippocampal complex and re-
spatial memory operates by the same rules as nonspa-
trieval of recent and very remote autobiographical memories:
tial memory and whether memory for specific events Evidence from functional magnetic resonance imaging in
that are unique to time and place consolidate in the neurologically intact people. Hippocampus 11, 707714.
same way as facts, which can be repeated in multiple Scoville, W. B., and Milner, B. (1957). Loss of recent memory after
contexts. Computational principles suggest that con- bilateral hippocampal lesions. Journal of Neurology, Neurosur-
solidation occurs because new information must be gery, and Psychiatry 20, 1121.
Squire, L., and Kandel, E. R. (1999). Memory: From mind to molecules.
incorporated gradually into preexisting representa-
New York: Scientific American Library.
tions in the neocortex. Simple network models have Squire, L. R., and Alvarez, P. (1995). Retrograde amnesia and
been constructed that capture these ideas. memory consolidation: A neurobiological perspective. Current
Opinion in Neurobiology 5, 169177.
See also: GUIDE TO THE ANATOMY OF THE BRAIN: Squire, L. R., Clark, R. E., and Knowlton, B. J. (2001). Retrograde
CEREBRAL CORTEX; GUIDE TO THE ANATOMY amnesia. Hippocampus 11, 5055.
OF THE BRAIN: HIPPOCAMPUS AND Squire, L. R., and Zola-Morgan, S. (1991). The medial temporal
PARAHIPPOCAMPAL REGION; MLLER, GEORG lobe memory system. Science 253, 1,3801,386.
ELIAS Sutherland, R. J., Weisend, M. P., Mumby, D., Astur, R. S., Hanlon,
F. M., Koerner, A., Thomas, M. J., Wu, Y., Moses, S. N., Cole,
Bibliography C., Hamilton, D. A., and Hoesing, J. M. (2001). Retrograde
Bontempi, B., Laurent-Demir, C., and Jaffar, R. (1999). Time- amnesia after hippocampal damage: Recent versus remote
dependent reorganization of brain circuitry underlying long- memories in two tasks. Hippocampus 11, 2742.
term memory storage. Nature 400, 671675. Teng, E., and Squire, L. R. (1999). Memory for places learned long
Cipolotti, L., Shallice T., Chan, D., Fox, N., Scahill, R., Harrison, ago is intact after hippocampal damage. Nature 400, 675677.
G., Stevens, J., and Rudge, P. (2001). Long-term retrograde Zola-Morgan, S., Squire, L. R., and Amaral, D. G. (1986). Human
amnesia: The crucial role of the hippocampus. Neuropsy- amnesia and the medial temporal region: Enduring memory
chologia 39, 151172. impairment following a bilateral lesion limited to field CA1
Haist, F., Bowden Gore, J., and Mao, H. (2001). Consolidation of of the hippocampus. Journal of Neuroscience 6, 2,9502,967.
human memory over decades revealed by functional magnetic
resonance imaging. Nature Neuroscience 11, 1,1391,145. Larry R. Squire
MEMORY SEARCH 373
MEMORY DRUGS Figure 1
See: COGNITIVE ENHANCERS; DRUGS AND

MEMORY; PHARMACOLOGICAL TREATMENT
OF MEMORY DEFICITS
MEMORY SEARCH
Encoding refers to the content or form in which infor-
mation is stored in memory; forgetting is loss of the
stored information with the passage of time or with
exposure to interfering materials; and retrieval refers
to accessing information from memory. Any observa-
tion of memory reflects all three components, but
measurements of recognition time, reflecting memo-
ry search, emphasize retrieval. Recognition time is
the time required to respond whether a visually pres-
ented test item was part of a previously studied list.
For an example of a recognition test following a
short-term memory list, see Figure 1. Recognition
time and accuracy for individual items in both short-
term and long-term memory are consistent with re-
trieval mechanisms involving parallel, or direct ac-
cess, operations rather than unguided search through
many memories. Some aspects of the increase in rec-
ognition time with the length of small short-term
memory lists suggest a serial, or sequential, search of
the list. However, when more detailed analysis of re-
sponse times and accuracies is carried out, serial (a) A sample test sequence in an STM item-recognition
experiment (see, e.g., Stemberg, S. [1966]. High speed scanning
search is ruled out as the sole mechanism of retrieval
in human memory. Science 153, 652654). In this sample, the
of the prior experience of an item. Instead, response
items are digits presented visually over time (t) and the list
times inversely related to recency of items in particu- length (4) is within the short-term range. Subjects press one of
lar list positions. Recency determines the efficiency of two keys to indicate whether or not the test item appeared on
a simple, direct retrieval process. In contrast with the the current list. Time and accuracy of key presses are measured;
direct-access retrieval of item information, informa- (b) Typical results of the item-recognition experiments,
tion about order generally reflects a slower, serial graphing the average (correct) response time (RT) as a function
process of recovery from memory in which list items of the length of the list for list members and nonmembers. RT
are searched in order. increases approximately linearly with list length, with roughly
equal slopes for list members and nonmembers (RT = bR + sL,
R=Yes for list members or No for non-list members). The slope,
Retrieval in Short-Term and Long-Term s, has been used as a measure of retrieval efficiency.
Memory
Primary, or short-term, memory (STM) for a few
recently active concepts is distinct from secondary, or searches through irrelevant memories. In contrast,
long-term, memory (LTM), which is the repository recall or production (of a previously experienced but
for all the varied information we retain over extended unpresented item) may involve a series of retrieval
periods of time (James, 1890, 1950). Short-term operations in which memory is successively sampled,
memory has been described by some theorists as an using general list cues and previously retrieved items
active subset of long-term memory (Cowan, 1995). as retrieval cues (Raaijmakers and Shiffirin, 1981).
Retrieval from LTM during recognition (of a pres-
ented item) has many properties of direct access or In the 1960s, strong claims were made that access
content addressability (Gillund and Shiffrin, 1984; to STM was fundamentally different from the direct-
Murdock, 1971). Direct-access retrieval occurs when access retrieval from LTM. STM retrieval was assert-
a cue or set of cues makes contact with memory in a ed to involve a series of comparisons to all currently
unitary process without recourse to a sequence of active concepts (Sternberg, 1966). Although this view
374 MEMORY SEARCH
fails to give a full account of memory search under all exhaustive serial search in the retrieval of informa-
circumstances, it is a useful starting point. tion from STM.
However, linear increases in mean RT with list
Recognition Time and STM length are also consistent with a parallel retrieval
mechanism in which all comparisons take place at the
Anecdotal information about human memory
same time, but with an efficiency that depends on the
largely reflects memory failures. Scientists use both
number of concurrent comparisons (see Figure 2b).
memory failures (errors) and measures of retrieval
This is called mechanism mimicry. Mimicry of serial
time to infer properties of memory. The ease of mem-
mechanisms by parallel mechanisms occurs when
ory retrieval is often measured in terms of the average
only simple measures such as the average RT are
time to recognize (respond to) a test item (response
available (see Townsend and Ashby, 1983, for a math-
time, RT). In S. Sternbergs (1966) classic short-term
ematical treatment of this mimicry). Hence, the regu-
memory experiments, subjects decided whether a dis-
larities noted by Sternberg (1975; see Figure 1a) may
played digit (or, in other studies, letter, word, syllable,
result from exhaustive and serial comparisons, or
or outline shape) had appeared on a short study list
from a set of parallel comparisons whose efficiency is
of one to six items. One key-press indicated a yes re-
affected by the number of comparisons, or by a direct-
sponse, another no (see Figure 1a). RT increased ap-
access process (see Figure 2c) whose efficiency is af-
proximately linearly for each additional list item (see
fected by other factors that covary with list length.
Figure 1b): mean RT = bR + sL, (R = Y or N), where
More detailed analyses discriminate the serial and
L is the list length, s is the additional average RT per
parallel mechanisms. In either case, time to retrieve
list item, and bR is a base time. This characteristic lin-
information from STM depends on the number of
ear (or approximately linear) relationship holds for
items currently being remembered.
lists of up to six to nine items (Burrows and Okada,
1975), at which point the increment for additional
items decreases sharply, possibly reflecting a transi- Direct-Access in Retrieval of Item
tion from list lengths within the capacity of STM to
Information
those exceeding that capacity.
A serial exhaustive search mechanism has these
properties: (1) linear increases in RT with list length
Underlying Retrieval Mechanisms where (2) slopes for positive and negative tests are
The linear relation of mean RT to list length, equal; (3) RT should not depend on the position of
where both positive (yes) and negative (no) test items a positive test item on the list; (4) the fastest (mini-
frequently yield approximately the same slope, was mum) RTs should increase with the length of the list;
interpreted by Sternberg as evidence for a serial and and (5) the RT variance (a measure of variation from
exhaustive set of comparisons between the test item test to test) should increase linearly with list length for
and all items in the list representation in short term both positive and negative recognition tests. Over
memory (see Figure 2a). The slope s of the RT func- many variants of STM item recognition experiments,
tion (the added time for each additional list member) properties (1) and (2) hold approximately, although
was identified as the time to compare the test item RT increases with list length may be more logarithmic
with each item on the list, one at a time, in series. The than linear (Briggs, 1974). Properties (3) and (4),
intercept bR then reflects all the processing mecha- which require more detailed breakdowns of data, fail
nisms (encoding of the test stimulus, organization systematically. Retrieval of item information from
and execution of the response) that do not depend on STM, when considered in more detail, is inconsistent
list length. If the list items were searched in series, with serial exhaustive processing.
equality of positive and negative slopes implies that RT, the time to recognize an item from STM, de-
the test item is compared exhaustively against all list pends strongly on its recency (property 3 fails). Test
items. items experienced very recently yield fast RTs, with
A search, or sequence of comparisons, that termi- RT increases for each less recent item. There is also
nated upon finding a match would finish, on average, a small advantage for the first list item. Recency in the
about halfway through the list when the item was posi- study list is the controlling factor whenever rehearsal
tive, but would go through the list when the test item is minimal or constrained to match study order. This
was negative. The slope of the negative tests would is easily seen when the data are graphed appropriate-
then be twice that of the positives; the 2:1 slope ratio ly (i.e., Monsell, 1978; see Figure 3a). Longer lists
is a property of some searches of items in visible dis- yield slower mean RTs because they include items of
plays but not of items in memory. Thus, parallel and less recency. Averaging over list positions yields ap-
linear list length functions are often thought to reflect proximately linear (or logarithmic) increases in mean
MEMORY SEARCH 375
Figure 2
Schematics illustrating serial and parallel retrieval mechanisms for comparing a test item against the memory representation of the
list. All these schematics distinguish nonretrieval processes (stimulus encoding, decision, response selection, and execution) from the
retrieval processes or memory comparisons. Parallel retrieval mechanisms can mimic serial mechanisms in predicting average
response time (RT; see Townsend, J. T., and Ashby, F. G. [1983]. The stochastic modeling of elementary psychological processes. New York:
Cambridge University Press). a. Sternberg ([1966]. High speed scanning in human memory. Science 153, 652654) proposed that
recognition following short memory lists of length L consisted of a serial and exhaustive sequence of comparisons of the test item
against a representation of each list item in memory. The test item is compared with all list items regardless of whether a match is
found, and a new comparison does not begin until the preceding one has been completed; b. Recognition as a set of parallel,
concurrent comparisons of the test item against representations of L list items. Efficiency depends on the number of concurrent
comparisons; c. Recognition as direct access to a relevant list memory. Efficiency depends on familiarity (see McElree, B., and Dosher,
B. [1989]. Serial position and set size in short-term memory: The time course of recognition. Journal of Experimental Psychology: General
118, 346373).
RT with list length. Failure of the prediction that the probability, in situations where list items are the same
minimum RT should depend fairly strongly on list over long sets of trials.
length (property 4) is implied by the recency data (see The list position effects and aspects of the RT dis-
Figure 3a): RT for the most recent items depends tributions in item recognition tasks contradict prop-
only weakly on the length of the list. Distributions of erties 3, 4, and 5 of the serial exhaustive scan. The ap-
RTs from shortest to longest show only tiny shifts of proximate equality of positive and negative slopes
the minimum with list length; average RT increases (property 2) contradicts terminating (nonexhaustive)
with list length largely reflect shifts in the longer RTs scan. For item memory (as distinct from order memo-
(see Figure 3b) (Hockley, 1984). Predictions of linear ry see subsequent explanation), the data are consis-
increases in variability (property 5) also fail. Other tent with a direct-access retrieval mechanism in which
findings that contradict the exhaustive serial search decreased availability of less recent items determines
mechanism are decreases in RT when a stimulus is re- average RT for particular list positions and hence for
peated and decreased RT for stimuli with high test different list lengths (McElree and Dosher, 1989).
376 MEMORY SEARCH
Figure 3 that the rate of retrieval of items from lists of different

length was fastest for the single most recent itema
case of an immediate match between the last item
studied and the test itembut is otherwise unaffected
by either list length or list position (see Figure 4). Re-
trieval from STM was parallel or direct access, yet the
ultimate success of retrieval was limited by familiarity
in memory. Recent items have been least affected by
forgetting due to the passage of time or intervening
items between study and test. The strength of items
when measured by errors and the accessibility of
items when measured by RT are both directly related
to the recency of study, with a small additional advan-
tage for the primacy or first item on the list.
LTM Retrieval
Recognition of items presented in longer lists
that exceed estimates of STM capacity shows many of
the same properties as recognition of items from
short, recent lists. Recognition from longer lists leads
to more errors than for STM lists, where the error
rates may be less than 5 percent. However, as in STM,
list position is an important factor in LTM, affecting
both RT and accuracy. As in shorter lists, items near
the end of the list are recognized more quickly and
accurately. When longer lists are used, study is usually
followed by many test trials, and location in the test
(a) Recency and primacy in STM recognition. The response time protocol also has a powerful effect. Earlier tests yield
(RT) is shown for each list position for each of several list faster RT and accuracy. Later in the test sequence, ad-
lengths (list position is labeled in terms of recency, with -1 ditional time and materials are interpolated between
denoting the last item studied, -2 the next-to-last item studied, encoding and retrieval; this is another manipulation
and so on). Recognition RT is fastest for items appearing closest
of recency. As with the STM data, study and test posi-
to the end of short lists (right of graph), increasing as items
tion effects on average RT primarily reflect shifts in
become less recent, except for the first (primacy) item in each
list (data from Monsell, S. [1978]. Recency, immediate
the long tail of the RT distributions. Full time course
recognition memory, and reaction time. Cognitive Psychology 10, of recognition is fastest for the single most recent
465501). Recency effects on RT of both list members and item, and otherwise equivalent but limited by famil-
nonmembers are not predicted by serial exhaustive mechanisms; iarity (Wickelgren, Corbett, and Dosher, 1980). These
(b) Distributions of RT (the number of responses in each RT findings rule out recency-dependent serial compari-
band) for lists of length 3 to 6 (data from Hockley, W. E. [1984]. sons that terminate on a match. The details of these
Analysis of response time distribution in the study of cognitive data, when examined carefully, are accounted for by
processes. Journal of Experimental Psychology: Learning, Memory and a parallel, direct-access retrieval process with shifts in
Cognition 10, 598615). List length primarily affects the longest estimated familiarity (Ratcliff, 1978).
RTs. Serial mechanisms predict shifts in the entire distribution
with list length. Indeed, direct access appears to be a general
property of item retrieval. The same time-course pat-
terns are found in both supra- and sub-span lists
Time Course of STM Retrieval for Item (Wickelgren, Corbett, and Dosher, 1983; McElree,
2001), indicating that direct access is a property of
Information
both short- and long-term representations. Direct-
A direct-access retrieval mechanism was directly access retrieval is also evident in the recovery of infor-
confirmed by more detailed RT methods, which allow mation from more complex representations and
inferences about the full time course of retrieval. based on altered cues. For example, direct-access re-
These methods interrupt retrieval at various times trieval characterizes the recognition of an item that is
after onset of the test display and observe the rate of part of a hierarchically coded group (McElree, 1998),
increase in correct responding with additional re- as well as recognition that is based on component
trieval time. B. McElree and B. Dosher (1989) showed properties (e.g., phonological and semantic proper-
MEMORY SEARCH 377
ties) of the memory representation. The latter finding Figure 4

in particular suggests that direct access arises from a
content-addressable retrieval operation that may op-
erate either on an exact representation or via reinte-
gration of the studied item from related cues during
retrieval.
Search Processes in Recovery of Order

Information
Although item information appears to be recov-
ered through a direct-access content-addressable re-
trieval process, this is not so for certain other forms
of information. Studies indicate that the retrieval of
relational information, including temporal order in-
formation in STM and positional information in LTM
often require a slow serial search. For example, if in-
formation about the order of items in STM must be
retrieved, both accuracy and retrieval speed depend
strongly upon item recency (McElree and Dosher,
1993). Temporal order information in STM was ex-
amined with a judgment of recency task, in which sub-
jects were presented two test probes from a short list
of five or six items and asked to select the item that
occurred more recently. The accuracy and speed of
the order judgment, as measured either by mean RT
or by time course analysis, was directly related to the
recency of the most recent item. A related phenome-
non occurs in n-back tasks, in which a positive re-
sponse is restricted to a particular ordinal position
(e.g., a 1-back match, a 2-back match) in a long ongo-
ing sequence instead of an overt judgment of order
(McElree, 2001). In both kinds of relational or order
tasks, retrieval speed is directly controlled by the re-
cency of the most recent relevant item, strongly impli- Full-time course of STM retrieval for list positions is consistent
with parallel, direct-access mechanisms. Time course is
cating a memory search process that begins with the
measured by examining accuracy after allowed recognition times
most recent item. One possibility is that ordered rep-
between 0.1 second and 2 seconds for list lengths of 3 to 6; d is
resentations in memory are serially scanned from the a bias-free measure of accuracy. Retrieval is fastest for the single
most recent, moving backwards in time (for specific most recent item in each list (labeled -1). Maximum accuracy
models see McElree and Dosher, 1993; McElree, late in retrieval varies with recency and primacy. Related effects
2001). Alternatively, order information may be recon- on response times are shown in Figure 3a and Figure 3b. Data
structed at retrieval by a serial chain process (Mur- from McElree, B., and Dosher, B. (1989). Serial position and set
dock, 1982), in which the last item on the list is used size in short-term memory: The time course of recognition.
as a cue to recover the next item on the list from Journal of Experimental Psychology: General 118, 346373.
memory, and so on. In all cases in which the retrieval
of relational or order information have been evaluat-
ed, some form of (terminating) successive or serial which ordered recall is perfect 50 percent of the time
process operates. This stands in contrast to the direct- (usually an interpolated value) is called the memory
access recovery of item information from either STM span. Span and the RT and accuracy for recognizing
or LTM. a single list item both measure aspects of STM func-
tion. Historically, span has been taken as a measure
of capacity (Miller, 1956; Baddeley, 1986), while rec-
Relation of STM Item and Order Retrieval ognition RT was taken as a measure of retrieval effi-
to STM Recall ciency. Although the two measures do not strongly
Another classic measure of STM is the ordered correlate with each other across individuals, they vary
recall of all items on a short list. The list length at together across different to-be-remembered materi-
378 MEMORY SEARCH
Figure 5
STM recognition and STM recall are strongly coupled over study material. The additional recognition response time (milliseconds) for
each additional list member (s in Figure 1b) is correlated with the inverse of the longest list recalled in perfect order half the time, or
memory span (items). Response time per item is a measure of STM retrieval efficiency and span is a measure of STM capacity.
Familiar materials such as digits or letters yield larger measured capacity and more efficient retrieval than unfamiliar materials such as
nonsense syllables or random shapes. Data from Cavanaugh, J. P. (1972). Relation between the immediate memory span and the
memory search rate. Psychological Review 79, 525530; and Brown, H. L., and Kirsner, K. (1980). A within-subjects analysis of the
relationship between memory span and processing rate in short-term memory. Cognitive Psychology 12, 177187.
als. J. P. Cavanaugh (1972) compared, via a survey of ent retrieval demands of the two tasks. Item recogni-
the research, the memory spans and the RT list- tion reflects the strength of direct-access content ad-
length slopes of digits, letters, words, shapes, and dressable information in memory, while span, a
nonsense materials. Materials yielding higher spans measure of ordered recall, additionally requires the
(longer list lengths supporting 50% recall) exhibit rel- sequential access of ordered representations, either
atively shallower slopes (less increase in average RT by sequential access of an intrinsically ordered repre-
with increasing list length; see Figure 5). The primary sentation, or by chained retrieval from a content-
factor producing differences in materials in both addressable store. In support of this view that STM
measures of STM may be overall familiarity (Puckett recall is composed of a sequence of retrieval or scan-
and Kausler, 1984). ning operations, the time to complete ordered recall
The relationship between item recognition and of a span length list may be as long as 5 to 8 s (Dosher
memory span in ordered recall may reflect the differ- and Ma, 1998), generally compatible both with the
MEMORY SPAN 379
time course of individual retrievals (see Figure 4) and McElree, B. (1998). Attended and non-attended states in working
with the time course of recovery of item information. memory. Journal of Memory and Language 38, 225252.
(2001). Working memory and focal attention. Journal of Ex-
perimental Psychology: Learning, Memory, and Cognition 27, 817
835.
Relation to Other Abilities McElree, B., and Dosher, B. (1989). Serial position and set size in
The capacity of STM has been viewed as an ele- short-term memory: The time course of recognition. Journal
mentary information-handling process, related to ef- of Experimental Psychology: General 118, 346373.
(1993). Serial retrieval processing in the recovery of order
ficiency in a variety of mental tasks (Baddeley, 1986). information. Journal of Experimental Psychology: General 122,
The speed of access to information in STM, defined 291315.
as the slope of the dependence of RT on list length Miller, G. E. (1956). The magic number seven, plus or minus two:
(s in the linear equation above), has been tested as a Some limits of our capacity for processing information. Psy-
correlate of the quality of performance on general chological Review 63, 8197.
Monsell, S. (1978). Recency, immediate recognition memory, and
cognitive indices such as aptitude scores. Correlations reaction times. Cognitive Psychology 10, 465501.
of capacity with psychometric measures are usually Murdock, B. B., Jr. (1971). A parallel processing model of scan-
higher than those of retrieval time with psychometric ning. Perception and Psychophysics 10, 289291.
measures (Sternberg, 1975). However, various special (1982). A theory for the storage and retrieval of item and
populations, such as the young and the elderly, have associative information. Psychological Review 89, 609626.
Palmer, J., MacLeod, C. M., Hunt, E., and Davidson, J. E. (1985).
been shown to have characteristic increases in STM Information processing correlates of reading. Journal of Mem-
recognition times, either in base times or in slopes, ory and Language 24, 5988.
compared with the performance of young adults Puckett, J. M, and Kausler, D. H. (1984). Individual differences
(Sternberg, 1975). STM function is one of the impor- and models of memory span: A role for memory search rate?
tant information-processing correlates with verbal in- Journal of Experimental Psychology: Learning, Memory, and Cogni-
tion 10, 7282.
telligence (Palmer, MacLeod, Hunt, and Davidson, Raaijmakers, J. G., and Shiffrin, R. M. (1981). Search of associative
1985). The concepts of STM and its close relative, memory. Psychological Review 88, 93134.
working memory, have been central in recent years in Ratcliff, R. (1978). A theory of memory retrieval. Psychological Re-
the development of theories and predictive indices of view 85, 59108.
general cognitive functions (Engle, Tuholski, and Sternberg, S. (1966). High speed scanning in human memory. Sci-
ence 153, 652654.
Kane, 1999). (1975). Memory-scanning: New findings and current con-
troversies. Quarterly Journal of Experimental Psychology 27, 132.
See also: MEMORY SPAN; RETRIEVAL PROCESSES IN Townsend, J. T., and Ashby, F. G. (1983). The stochastic modeling of
MEMORY; WORKING MEMORY: HUMANS elementary psychological processes. New York: Cambridge Uni-
versity Press.
Wickelgren, W. A., Corbett, A. T, and Dosher, B. A. (1980). Prim-
Bibliography
ing and retrieval from short-term memory: A speed-accuracy
Baddeley, A. D. (1986). Working memory. New York: Oxford Univer- tradeoff analysis. Journal of Verbal Learning and Verbal Behavior
sity Press. 19, 387404.
Briggs, G. E. (1974). On the predictor variable for choice reaction
time. Memory & Cognition 2, 575580. Barbara Anne Dosher
Burrows, D., and Okada, R. (1975). Memory retrieval from long Brian McElree
and short lists. Science 188, 1,0311,033.
Cavanaugh, J. P. (1972). Relation between the immediate memory
span and the memory search rate. Psychological Review 79,
525530.
Cowan, N. (1995). Attention and memory: An integrated framework.
New York: Oxford University Press. MEMORY SPAN
Dosher, B., and Ma, J. J. (1998). Output loss or rehearsal loop? The term memory span refers to the maximum length
Output time versus pronunciation time limits in immediate
recall for forgetting matched material. Journal of Experimental of a sequence of items that can be reproduced from
Psychology: Learning, Memory, and Cognition 24, 316335. memory following a single presentation. Scientists
Engle, R. W., Tuholski, S., and Kane, M. (1999). Individual differ- have been interested in memory span since the publi-
ences in working memory capacity and what they tell us about cation of the first important study of memory, nine-
controlled attention, general fluid intelligence and functions teenth-century German experimental psychologist
of the prefrontal cortex. In A. Miyake and P. Shah, eds., Mod-
els of working memory: Mechanisms of active maintenance and exec- Hermann Ebbinghauss monograph in 1885. Using
utive control. Cambridge, MA: Cambridge University Press. himself as his only subject, Ebbinghaus determined
Gillund, G., and Shiffrin, R. M. (1984). A retrieval model for both the number of presentations necessary for an error-
recognition and recall. Psychological Review 91, 167. free reproduction of a sequence of items; he found
Hockley, W. E. (1984). Analysis of response time distribution in the that this number decreased dramatically as the length
study of cognitive processes. Journal of Experimental Psychology:
Learning, Memory, and Cognition 10, 598615. of the sequence decreased until the sequence includ-
James, W. (1890; reprint 1950). The principles of psychology. New ed only seven items, at which point only a single pre-
York: Dover. sentation was needed. Ebbinghaus showed no partic-
380 MEMORY SPAN
ular interest in this finding, but others did. Within two ing procedure. More precise estimates, as may be
years, memory span was shown to increase systemati- needed for certain research purposes, may be ob-
cally during childhood and to be appreciably shorter tained with the staircase, or up-and-down, method.
for the mentally impaired. Within a decade, memory This involves the presentation of a series of lists, the
span was firmly established in what was then an length of any given list being one more than that of
emerging field of mental abilities testing, where it has the immediately preceding list if the latter was cor-
remained ever since. rectly reproduced and one less if it was not. Memory
span is given by averaging the list lengths. In deter-
Most often the procedure for testing memory mining this average the first few lists should be disre-
span calls for the recall of the items in the order in garded because they will reflect the arbitrary length
which they were presented. Sometimes the order in of the first list. Also, the length of what would have
which the items have to be recalled is entirely uncon- been the next list in the series, as given by the length
strained; sometimes the items have to be recalled in and outcome for the final presented list, should be in-
their reverse order of presentation. But only rarely cluded.
does the procedure call for more difficult transforma-
tions, such as recall of the items in alphabetical or nu- Ebbinghauss finding of seven items as the maxi-
merical order. Such transformations would draw on mum length for a reproducible list provides a first ap-
what is often referred to as working memory and, in- proximation of memory span. Closer approximations
terestingly, would probably provide a more valid will vary with the age of the rememberer, for span in-
measure of mental ability. Incidentally, the measure creases through childhood and declines in old age.
of choice for this ability has been a variant of memory Moreover, span varies among individuals of the same
span known as working memory span, which is the age.
number of items that can be retained while perform-
ing some other cognitive task.
Factors in Memory Span
In addition to varying among individuals, memo-
Measuring Memory Span ry span varies for a given individual according to a
Order of reproduction aside, there has never considerable number of factors. For example, span
been a fixed procedure by which memory span is can be increased by presenting the items at an irregu-
measured. Most of the methods reviewed long ago by lar rate, so that they appear temporally grouped.
J. Paul Guilford and Karl Dallenbach (1925) are still Also, span for verbal items tends to be slightly greater
extant. Sometimes the sequence of items is deliber- with auditory presentation than with visual presenta-
ately set somewhat too long for perfect reproduction, tion. Of particular interest is the effect of the nature
and memory span is defined as the number of items of the list item. The most common kind of item, espe-
from this sequence that are recalled. This method is cially in mental abilities testing, is the digit. Digit span
quick to administer but too problematic to be consid- is roughly one item greater than letter span, which in
ered anything more than rough and ready. One prob- turn is roughly one item greater than word span. Also
lem is that the usual requirement that the items be re- used as list items have been nonsense syllables (which
called in their exact order of presentation has to be is what Ebbinghaus used), geometric designs, and pic-
modified to allow for imperfect recall. Another is that, tures of objects. One finding that has emerged from
whatever the criterion regarding recall order, the a comparison of memory span for different kinds of
number of items recalled is likely to vary according to items is an impressive linear relation between memo-
the number presented, even in the supraspan range. ry span and a hypothetical search rate, operational-
For such reasons, memory span is usually determined ized in terms of the slope of the roughly linear func-
by presenting lists of several different lengths and as- tion that relates the time to decide whether a test, or
certaining the maximum length for which recall is probe, item was included in a just-presented short se-
perfect. In the procedure typical of most mental abili- quence to the length of the sequence. Specifically, a
ties tests, list length is at first so short that perfect re- relatively small increment in decision time is incurred
production is virtually certain, and then is gradually by increasing the length of a sequence of items of the
increased until errors are made. kind that yields a relatively large memory span. Al-
though the significance of this finding remains uncer-
Like any other psychological measure, memory tain, it clearly raises the question of how memory span
span is not entirely reliable, and is therefore a statisti- should be conceptualized.
cal conceptthe sequence length for which there is
an even chance of perfect reproduction. For this rea- One way of conceptualizing memory span is as a
son, most tests include two or three lists of each measure of the capacity of what the American psy-
length, a specific stopping rule, and a specific averag- chologist and philosopher William James called pri-
MENTAL RETARDATION (INTELLECTUAL DISABILITIES) 381
mary memory. In other words, memory span can be tween 1.5 and 2.0 seconds. And, indeed, items that
considered as the number of items that can be held yield long spans, such as digits, tend to be those that
in conscious mind at any given instance. Aside from are articulated relatively quickly. There are, however,
its intuitive appeal, this interpretation is supported by exceptions to this rule, leaving memory span still
evidence that memory span varies according to cer- without a satisfactory interpretation.
tain characteristics of the list items that are salient in
conscious experience during performance of the task. Bibliography
Brener, R. (1940). An experimental investigation of memory span.
In particular, the representation of the list items in
mind usually takes the form of inner speech, and Brooks, J. O., III, and Watkins, M. J. (1990). Further evidence of
memory span has been found to depend on factors the intricacy of memory span. Journal of Experimental Psycholo-
that can reasonably be regarded as relevant to the gy: Learning, Memory, and Cognition 16, 1,1341,141.
spoken form of the items. Thus, span is shorter when Cavanaugh, J. P. (1972). Relation between immediate memory
span and the memory search rate. Psychological Review 79,
the list items are phonemically similar to one another 525530.
than when they are phonemically dissimilar (e.g., Cowan, N. (2001). The magical number 4 in short-term memory:
shorter for GBVDPTZ than for GMRKSQY), when the A reconsideration of mental storage capacity. Behavioral and
list items are phonemically lengthy than when they Brain Sciences 24, 87114.
are phonemically short (e.g., shorter for lists of poly- Dempster, F. N. (1981). Memory span: Sources of individual and
developmental differences. Psychological Bulletin 89, 63100.
syllabic words than for lists of monosyllabic words), Guilford, J. P., and Dallenbach, K. M. (1925). The determination
and when subjects engage in irrelevant vocalization of memory span by the method of constant stimuli. American
during list presentationan activity that is likely to Journal of Psychology 36, 621628.
suppress covert naming of the items. Other support Miyake, A., ed. (2001). Individual differences in working memory.
Journal of Experimental Psychology: General 130, 163237.
for the idea of memory span as the product of prima-
Schiano, D. J., and Watkins, M. J. (1981). Speech-like coding of
ry memory derives from evidence that memory span pictures in short-term memory. Memory & Cognition 9, 110
for one kind of item predicts fairly well memory span 114.
for another kind of item but is a poor predictor of Schweikert, R., and Boruff, B. (1986). Short-term memory capaci-
performance on supraspan tasks that clearly cannot ty: Magic number or magic spell? Journal of Experimental Psy-
chology: Learning, Memory, and Cognition 12, 419425.
be performed on the basis of primary memory.
Underwood, B. J., Boruch, R. F., and Malmi, R. A. (1978). Compo-
Such supportive evidence notwithstanding, the sition of episodic memory. Journal of Experimental Psychology:
General 107, 393419.
idea that memory span measures primary memory
does little to shape twenty-first-century theorizing. Michael J. Watkins
One reason for this is that virtually all of todays mem-
ory theorists adopt an information-processing per-
spective, and they give little or no consideration to
the conscious realization of the mechanisms they hy- MENTAL RETARDATION
pothesize. A more particular reason is that primary (INTELLECTUAL DISABILITIES)
memory has been recast as a short-term store, the ca-
pacity of which has been estimated at three or four In any society, certain individuals fail to make normal
items. This estimate, which was based on the number progress in intellectual, social, and linguistic growth
of items recalled from toward the end of a supraspan and development, exhibiting marked difficulties in
sequence, is appreciably smaller than memory span. learning. The need to distinguish these individuals
One way to account for the discrepancy is to make the led the scientific community to develop intelligence
tests, as well as to create an operational definition of
plausible assumption that the capacity of primary
mental retardation. Mental retardation (also called in-
memory varies with the nature of the task and that,
tellectual disabilities) is defined as significantly subaver-
unlike supraspan recall tasks, the memory span task
age general intellectual functioning existing concur-
stretches primary memory to its maximum. Another
rently with deficits in adaptive behavior, and
possibility, supported by recent evidence, is that
manifested during the developmental period. Clini-
memory span represents primary memory supple-
cians typically interpret the condition to include those
mented by secondary memorythat is, by items that
individuals who obtain an IQ (intelligence quotient)
have dropped out of conscious mind and have to be
score two or more standard deviations below the
recollected.
mean (i.e., IQ less than 70) on an intelligence mea-
A conception of memory span more in keeping sure such as the Stanford-Binet or the Wechsler Adult
with the zeitgeist, though not inherently incompatible Intelligence Scale. Clinicians usually include two
with the primary memory interpretation, is that it other criteriaone including deficits in everyday
represents the amount of information that can be ar- functioning (so-called adaptive behavior) and onset
ticulated in a certain time, variously estimated as be- during the childhood years (i.e., before age eigh-
382 MENTAL RETARDATION (INTELLECTUAL DISABILITIES)
teen)in the three-criteria definitions of mental re- are retarded remains unclear, but some (unspecified)
tardation. interaction of biological, familial, psychological, so-
cial, and environmental risks seems likely. And, if (as
Within the population with mental retardation it-
in nonretarded persons) approximately half of the
self, researchers and clinicians have further classified
variation in IQ scores is due to the workings of many
individuals due to the individuals degree of intellec-
genes, such individuals may show their lower IQ
tual impairment. Thus, persons with mental retarda-
scores from having received a poor polygenetic
tion have often been described as having mild (IQ 55
draw. With the success of behavior geneticists, at
to 69), moderate (40 to 54), severe (25 to 39), or pro-
least one of these polygenes for intelligence has been
found (below 25) degrees of impairment. Especially
identified (Chorney et al., 1998). Future work will
for persons with severe or profound mental retarda-
tease apart the effects of genetic and various environ-
tion, most individuals also show correspondingly
mental factorsworking separately or in tandem
lower levels of functioning on a wide variety of cogni-
over timeon ones overall IQ.
tive-linguistic and everyday tasks. For individuals with
mild and moderate mental retardation, however, life
success (e.g., performing well at work) seems much Organic Mental Retardation
more determined by ones personality, perseverance,
socialization skills, and lack of maladaptive behaviors. Organic mental retardation includes all persons
whose intellectual deficits can be attributed to one or
IQ-related classifications have limited utility for more specific organic insults. Such insults can occur
understanding how different intellectual abilities may before birth (as in any of the 750-plus genetic disor-
or may not be connected. In particular, an overreli- ders associated with mental retardation); at or around
ance on an omnibus IQ score fails to acknowledge ei- the time of birth (due to anoxia, prematurity); or in
ther the composite nature of this score or perspectives the years after birth (due to meningitis, head trauma).
on the multidimensional nature of intelligence, such
as those espoused by Howard Gardner (1983) and Since the 1990s, scientists have made much prog-
Jerry Fodor (1983). In fact, a single IQ score (e.g., 50) ress in mental retardation by focusing on well-defined
may be associated with markedly different patterns of organic syndromes, particularly those related to such
development, with one child failing to make progress genetic disorders as Down syndrome, Prader-Willi
on language and performing near age level on visuo- syndrome, and Williams syndrome (Dykens and Ho-
spatial tasks and another showing the exact opposite dapp, 2001). Such research ties these forms of mental
pattern. retardation to particular patterns of cognitive, lin-
guistic, adaptive, and maladaptive deficits. In exam-
Another way to classify persons with mental retar- ining functioning in persons with specific genetic eti-
dation, then, relies on the cause or etiology of ones ologies of mental retardation, researchers also gain
mental retardation. From the late 1960s until the insights into the structure of human intelligence. Sev-
early twenty-first century, several researchers have eral of the major genetic forms of mental retardation
advocated what has been called the two-group ap- are discussed in this entry, together with presenting
proach to mental retardation (Hodapp, 1997). Ad- symptoms, characteristic cognitive deficits, explora-
herents of this approach argue that the retarded pop- tions of their neurobiological underpinnings, and at-
ulation can be divided into those whose mental tempts at treatment.
retardation is due to the same polygenetic and/or en-
vironmental factors causing IQ differences within typ- Down Syndrome
ically developing individuals (cultural-familial mental Down syndrome (DS), trisomy 21, is the best-
retardation) versus those whose mental retardation is known form of mental retardation and the most prev-
caused by a specific organic insult (organic mental re- alent form known to be associated with a chromosom-
tardation). This entry explores the cultural-familial al abnormality, occurring once in every 800 to 1,000
and organic groups in turn. live births. The syndrome was identified and de-
scribed in 1866 by Langdon Down, and the extra
twenty-first chromosome was discovered in 1959.
Cultural-Familial Mental Retardation Since that time, Down syndrome has been the focus
The first group is comprised of those individuals of intensive genetic and behavioral research. Chil-
whose mental retardation has no obvious organic eti- dren with DS usually have characteristic physical fea-
ology. Scientists do not know the cause of mental dis- tures including the epicanthic fold (leading to its orig-
ability in approximately 50 percent of persons with inal label, mongolism), a protruding tongue, short
mental retardation. One hypothesis is that this group stature, and hypotonia (weak muscle tone). DS often
simply forms the lower portion of the normal, Gaussi- occurs together with such medical conditions as heart
an distribution of intelligence. Why such individuals defects, leukemia, and gut atresia. Although in the
MENTAL RETARDATION (INTELLECTUAL DISABILITIES) 383
past individuals with DS had a short life span and Intellectually, most children with Prader-Willi
were often institutionalized, medical treatments have syndrome show relative weaknesses on tasks involving
improved their life span and they are now typically consecutive, step-by-step order in problem solving, or
raised at home. Persons with DS have received consid- sequential processing. In contrast, these children per-
erable attention for educational successes that have form well on tasks requiring integration and synthesis
far outstripped earlier predictions. DS may be detect- of stimuli as a unified whole, or simultaneous process-
ed during pregnancy through chorionic villus sam- ing. In addition to these more general cognitive pro-
pling, or amniocentesis; this procedure is usually rec- files, research demonstrates that many individuals
ommended in women above age thirty-five, who bear with Prader-Willi syndrome exhibit particularly high-
a substantially higher risk of carrying a Down syn- level abilities in jigsaw puzzles. Such skills, which on
drome baby (Pueschel, 2001). average are way above typical children of comparable
chronological ages, are especially shown in those hav-
Three behavioral characteristics appear in many ing the deletionas opposed to the maternal diso-
(possibly most) individuals with Down syndrome myform of this disorder.
(Rondal, Perera, and Nadel, 1999). The first involves
a specific set of cognitive-linguistic strengths and Williams Syndrome
weaknesses. In various studies, persons with Down Williams syndrome is caused by a micro-deletion
syndrome appear particularly impaired in language. on chromosome 7. Children with this syndrome gen-
Such impairments, which are more pronounced than erally show a characteristic, elfin-like facial appear-
overall levels of mental age (MA) per se, occur in lin- ance, along with heart and other health problems
guistic grammar, in expressive (as opposed to recep- (Dykens, Hodapp, and Finucane, 2000). As many as
tive) language, and in articulation. Conversely, per- 95 percent of these children suffer from hyperacusis,
sons with Down syndrome often show relatively or a hypersensitivity to sound. Along with an almost
higher performance on tasks of visual short-term overly social, outgoing style of personality, children
memory. with Williams syndrome also show a wide variety of
anxieties and fears. Although not every child with
A second behavioral issue involves the rate of de-
Williams syndrome shows any or all of these fears, the
velopment, with children with Down syndrome devel-
vast majority do appear to be overly fearful compared
oping at slower rates as they get older. Such slowings
to most children with mental retardation.
of development may relate to age-related changes or
to difficulties these children have in achieving certain Apart from these medical and psychiatric issues,
cognitive tasks (e.g., language). A third, possibly relat- scientific attention has strongly focused on the inter-
ed change concerns Alzheimers disease. It is known estingpossibly uniquecognitive-linguistic profile
that neuropathological signs of Alzheimers disease shown by most of these children. Children with Wil-
appear to be universal in individuals with DS by age liams syndrome show relative strengths in language;
thirty-five or forty. Geneticists continue to explore for many years researchers thought that these chil-
the connection of DS with Alzheimers, and continue dren might even perform at chronological age levels
to learn more about pathological segments of chro- on a variety of linguistic tasks. Although late-
mosome 21 involved in overlapping conditions. twentieth-century research has found that age-
appropriate performance in language occurs in only
Prader-Willi Syndrome small percentages of children with Williams syn-
Prader-Willi syndrome is caused by missing ge- drome (Mervis, Morris, Bertrand, and Robinson,
netic material from the chromosome 15 derived from 1999), these children nevertheless show relative
the fathereither a deletion on the paternally de- strengths in language, as well as in auditory process-
rived 15 or two chromosome 15s from the mother (so- ing and in some areas of music. Conversely, many
called maternal disomy). Most individuals with children with Williams syndrome perform especially
Prader-Willi syndrome are short in stature (about 5 poorly on a variety of visuo-spatial tasks.
feet tall in adulthood) and show extreme hyperphagia Taken together, these patterns of strengths,
(overeating). Indeed, hyperphagia and resultant obe- weaknesses, and maladaptive behaviors have led to
sity have long been considered the hallmarks of calls for interventions that might be tailored to differ-
Prader-Willi syndrome, and most cases of early death ent etiological groups. If, for example, persons with
in the syndrome relate to obesity and its related heart Williams syndrome are social and show relative
and circulatory problems (Dykens and Cassidy, 1999). strengths in language, they might benefit greatly
In addition to hyperphagia, many individuals show from various group and talk therapies (Dykens and
obsessions and compulsions that are similar in level Hodapp, 1997). In Prader-Willi syndrome, strict su-
to those with clinically diagnosed obsessive- pervision around food and eating has historically
compulsive disorder. proven effective, but various drug regimens (mainly
384 MENTAL RETARDATION (INTELLECTUAL DISABILITIES)
involving growth hormone) and clinical interventions notions to be incorrect. Specifically, only about 5 per-
(for obsessions and compulsions) are increasingly cent of children with Williams syndrome show age-
being tried. Educationally, too, calls have recently equivalent levels of language, and language, while a
been made for etiology-based interventions (Hodapp relative strength, appears connected to at least cer-
and Fidler, 1999). In England, for example, S. Buck- tain other cognitive abilities. Thus, in Williams syn-
ley (1999) has long advocated reading instruction for drome (as with typically developing children), abili-
children with Down syndrome. The idea is that if for ties in auditory short-term memory are strongly
most of these children visual skills are relatively related with abilities in linguistic grammar. Such
strong and language skills relatively weak, then read- studies, which continue in many of these syndromes,
ing may help them as an entryway into language. Al- promise to tell researchers both about strengths and
though few such suggestions have so far been tested, weaknesses and inter-domain connections within
future work promises to specify etiology-based inter- each specific syndrome, as well how various aspects of
vention proposals and to evaluate if indeed such strat- cognition and language might be organized in nonre-
egies are more effective than contemporary, more ge- tarded persons.
neric intervention approaches.
Bibliography
Buckley, S. (1999). Promoting the cognitive development of chil-
Developmental Accounts of Mental dren with Down syndrome: The practical implications of re-
Retardation cent psychological research. In J. A. Rondal, J. Perera, and L.
Nadel, eds., Downs syndrome: A review of current knowledge.
Despite the extreme variability of cause in retar-
London: Whurr Publishers Ltd.
dation, various researchers have for many years ex- Burack, J. A., Hodapp, R. M., and Zigler, E., eds. (1998). Handbook
amined development in children with mental retar- of mental retardation and development. Cambridge, UK: Cam-
dation (see articles in Burack, Hodapp, and Zigler, bridge University Press.
1998). Such studies have generally focused on two re- Chorney, M. J., Chorney, K., Seese, N., Owen, M. J., Daniels, J.,
McGuffin, P., Thompson, L. A., Detterman, D. K., Benbow,
lated topics. In the first, studies examine the sequences
C., Lubinski, D., Eley, T., and Plomin, R. (1998). A quantita-
or orderings in development. With only a few excep- tive trait locus associated with cognitive ability in children.
tions, most children develop in the universal order- Psychological Science 9, 159166.
ings of development in most cognitive and linguistic Dykens, E. M., and Cassidy, S. B. (1999). Prader-Willi syndrome.
domains. In S. Goldstein and C. R. Reynolds, eds., Handbook of neurode-
velopmental and genetic disorders in children. New York: Guilford
A second developmental concern relates to the Press.
structure of development, or the ways in which devel- Dykens, E. M., and Hodapp, R. M. (1997). Treatment issues in ge-
netic mental retardation syndromes. Professional Psychology:
opments in various domains go together. Although Research and Practice 28 (3), 263270.
individuals with cultural-familial mental retardation (2001). Research in mental retardation: Toward an etiolog-
likely show equal or near-equal levels of abilities ic approach. Journal of Child Psychology and Psychiatry and Allied
across a variety of cognitive-linguistic domains, such Disciplines 42, 4971.
does not appear to be the case in several genetic men- Dykens, E. M., Hodapp, R. M., and Finucane, B. (2000). Genetics
and mental retardation syndromes: A new look at behavior and inter-
tal retardation disorders. Thus, children with Down vention. Baltimore, MD: Paul H. Brookes Publishers.
syndrome show worse-than-MA-level performance in Fodor, J. (1983). Modularity of mind: An essay on faculty psychology.
linguistic grammar, those with Prader-Willi syndrome Cambridge, MA: MIT Press.
show relative strengths in simultaneous processing Gardner, H. (1983). Frames of mind. New York: Basic Books.
(and weaknesses in sequential, step-by-step process- Hodapp, R. M. (1997). Developmental approaches to children with
disabilities: New perspectives, populations, prospects. In S. S.
ing), and those with Williams syndrome display rela- Luthar, J. A. Burack, D. Cicchetti, and J. R. Weisz, eds., Devel-
tively strong language and weak visuo-spatial skills. opmental psychopathology: Perspectives on adjustment, risk and dis-
Furthermore, in many cases, this pattern of strengths order. Cambridge, UK: Cambridge University Press.
and weaknesses becomes more pronounced with age, Hodapp, R. M., and Fidler, D. J. (1999). Special education and ge-
such that children showing slight patterns of netics: Connections for the 21st century. Journal of Special Edu-
cation 22, 130137.
strength-over-weakness at early ages show more ex- Mervis, C. B., Morris, C. A., Bertrand, J. M., and Robinson, B. F.
treme patterns as they get older. (1999). Williams syndrome: Findings from an integrated pro-
gram of research. In H. Tager-Flusberg, ed., Neurodevelopmen-
Until the mid-1990s, some researchers felt that tal disorders: Contributions to a framework from the cognitive sci-
such patterns provided evidence for modular skills ences. Cambridge, MA: MIT Press.
in several genetic mental retardation disorders. In Pueschel, S. M., ed. (2001). A parents guide to Down syndrome, rev.
Williams syndrome, for example, some researchers edition. Baltimore, MD: Paul H. Brookes.
held that language occurred at normal, age- Rondal, J., Perera, J., and Nadel, L., eds. (1999). Down syndrome:
A review of current knowledge. London: Whurr.
equivalent levels and was separate from other areas of
cognition. Later studies have generally proven such Robert M. Hodapp
METACOGNITION ABOUT MEMORY 385
METACOGNITION ABOUT MEMORY Judgments of Learning

Metacognition about memory, sometimes called The next kind of monitoring occurs during and/
metamemory, refers to the self-monitoring and self- or at the end of learning. The persons judgment of
control of ones own memory in the acquisition and learning is the evaluation of how well he or she has
retrieval of information. It is a relatively new topic, learned a given item; it is a prediction of the likeli-
having been investigated by psychologists for approx- hood that the item will be recalled correctly on a fu-
imately forty years. Before then, researchers viewed ture test. The psychologists Arbuckle and Cuddy
learners as passive, as blank slates onto which new (1969) first showed that the predictive accuracy of
ideas were etched through repetition. By contrast, peoples judgment of learning is above chance but far
subsequent researchers viewed the learner as an ac- from perfect, like ease-of-learning judgments. Re-
tive controller of his or her learning, whether acquir- search by Leonesio and Nelson (1990) has shown,
ing new or retrieving old information. Moreover, re- hwoever, that judgments of learning are more accu-
searchers now know that people can monitor their rate than ease-of-learning judgments for predicting
progress during both learning and retrieval. eventual recall; this is probably because the judg-
ments of learningbut not ease-of-learning judg-
For example, imagine a student who is studying mentscan be based in part on what the learner no-
for an examination that will occur tomorrow in tices about how well he or she is mastering the items
French class, say on French-English vocabulary such during learning.
as chateau/castle and rouge/red. Let us keep that Research by Nelson and Dunlosky (1991) showed
student in mind as we consider the monitoring and a case in which judgments of learning are especially
the control of the students learning of the new vocab- accurate: when they are made after a filled delay of
ulary and his or her attempts to retrieve the answers at least thirty seconds following the offset of study of
during the test the next day. an item. This phenomenon, in which the accuracy of
judgment of learning in predicting future recall is
A theoretical framework that integrates all of nearly perfect, is known as the delayed JOL effect.
these processes into an overall system can be found
in Nelson and Narens (1990). Feeling-of-Knowing Judgments
These judgments are peoples prediction of
Different kinds of monitoring processes can be whether they will eventually remember an answer that
distinguished in terms of when they occur in the they do not now recall. This was the first kind of
learning/retrieval sequence and whether they pertain metamemory judgment to be examined in the labora-
to the persons future performance (in which case the tory. The pioneering researcher, Hart (1965), found
focus is said to be on prospective monitoring) or to that these feeling-of-knowing judgments were some-
the persons past performance (in which case the what accurate in predicting subsequent memory per-
focus is said to be on retrospective monitoring). formance. The subsequent likelihood of correctly rec-
ognizing a nonrecalled answer was higher for
nonrecalled items that people said were stored in
Prospective Monitoring their memory than for those reported as not stored
in memory. However, as in the case of other
Ease-of-Learning Judgment metamemory judgments, the accuracy of this judg-
The first metamemory judgment made by some- ment was far from perfect; people often did not rec-
one who is getting ready to learn new information oc- ognize answers they had claimed they would recog-
curs before learning begins. This ease-of-learning nize, and they sometimes recognized those they
judgment is the persons evaluation of how easy or thought they would not. The accuracy of predicting
difficult the items will be to learn. For instance, the other kinds of memory performance (e.g., ease of rel-
person might believe that the overall set of items will earning) on nonrecalled items was subsequently re-
take a given amount of time to learn and that chateau/ viewed and investigated by Nelson, Gerler, and
castle will be more difficult to learn that rouge/red. Narens (1984), who also offered several theoretical
Underwood (1966) first showed that people are mod- explanations for how people might make their feel-
erately accuratenot perfectly accurate but well ing-of-knowing judgments.
above chanceat predicting which items will be easi-
est or hardest to learn. Peoples predictions of how
easy it will be to learn each item, made in advance of Retrospective Confidence Judgments
learning, are moderately correlated (i.e., covaried) In contrast to attempts by people to predict their
with their subsequent recall after a constant amount future memory performance, retrospective confi-
of study time on every item. dence judgments occur after the venturing of the an-
386 METACOGNITION ABOUT MEMORY
swerwhether correct or incorrectand pertain to Even adults do not always allocate more study
degrees of confidence in the correctness of the an- time to more difficult items. Son and Metcalfe (2000)
swer. For instance, if our hypothetical student were have shown that when there is insufficient time to
asked to recall the English equivalent of chateau, the study all items, subjects may allocate the most study
person might say castle (correct answer) or might time to easier items. For instance, if our hypothetical
say red (incorrect answer); then he or she would student judged his or her learning to be greater for
make a confidence judgment about the likelihood rouge/red than for chateau/castle, then, if he or she had
that the recalled answer was correct. Fischhoff, Slovic, unlimited study time, that knowledge should result in
and Lichtenstein (1977) found that retrospective con- extra study time for it than for rouge/red. If study time
fidence judgments were substantially accurate but were limited, the student might allocate more time to
often marred by overconfidence. For instance, for chateau/castle than to rouge/red, perhaps reasoning
those items to which people had given a confidence that this strategy would enhance performance,
judgment of 80 percent likely to be correct, the ac- though Son and Metcalfes (2000) research shows no
tual percentage of correct recognition was much evidence that this strategy aids test performance.
lower than 80 percent.
Strategy Employed During Self-Paced Study.
This overconfidence occurs primarily for confi- Not only can people control how much study time they
dence judgments made for individual items. Confi- allocate to various items (as discussed in the previous
dence judgments can also be made for a list of items paragraph), but they can also control which strategy they
(e.g., How many of the fifty items on the test did you employ during that study time. For many kinds of learn-
recall correctly?), which is labeled an aggregate con- ing, there are strategies that are more effective than rote
fidence judgment. Griffin and Tversky (1992) report- repetition. For instance, a mnemonic strategy for learn-
ed that aggregate confidence judgments are usually ing foreign-language vocabulary was investigated by Pre-
less overconfident (and sometimes even underconfi- ssley, Levin, and Ghatala (1984). After people had
dent), a phenomenon known as the aggregation ef- learned some foreign-language vocabulary by rote and
fect. other foreign-language vocabulary by the mnemonic
strategy, they were given a choice of using whichever
strategy they wanted for a final trial of learning some ad-
Control ditional foreign-language vocabulary. Only 12 percent
Not only can people monitor their progress dur- of adults chose the mnemonic strategy if they had not re-
ing learning and retrieval, but they can also control ceived any test trial during the earlier learning phase,
many aspects of their learning and retrieval. First, we whereas 87 percent chose the mnemonic strategy if they
will consider aspects they can control during self- had received test trials during the earlier learning phase.
paced learning, and then we will consider aspects they Apparently, people should have test trials to help them
can control during retrieval. realize the effectiveness of different strategies for learn-
ing. Moreover, getting children to adopt the mnemonic
Control During Self-Paced Learning strategy spontaneously required test trials during the
Allocation of Self-Paced Study Time During earlier learning phase. The children also needed to have
Learning. Our hypothetical student who is learning feedback after those test trials to tell them how well they
foreign-language vocabulary can choose to allocate large had done on the rote-learned items versus the mnemon-
or small amounts of study time to each item and subse- ic-learned items.
quently can allocate extra study time to some items.
Moreover, as shown by Bisanz, Vesonder, and Voss Control During Retrieval
(1978), the allocation of study time may be made in con- Control of Initiating Ones Retrieval. Immedi-
junction with the judgments of learning described ately after a questions is posed, before someone memory
above. In an investigation of learners of various ages, Bi- for the answer, he or she makes a metamemory decision
sanz and colleagues discovered that learners in the early about whether the answer is likely to be found in memo-
years of primary school might make accurate judgments ry. For instance, if you were asked the telephone number
of learning but would not use those judgments when al- of the president of the United States, you probably
locating additional study time across the items, whereas would decide immediately that the answer is not in your
slightly older children would use those judgments when memory. Notice that you do not need to search through
allocating additional study time. In particular, the older all the telephone numbers that you know, nor do you
children allocated extra study time to items that they be- need to search through all the information you have
lieved they had learned but not to those that they be- stored in your memory about the president. You proba-
lieved they had not learned. By contrast, the younger bly realize that the president does have a telephone
children were not systematic in allocating extra study number, but you know you dont know it, and therefore
time primarily to unlearned items. you dont initiate attempts to retrieve that answer. Con-
METAPLASTICITY 387
sider how different that situation is from one in which Hart, J. T. (1965). Memory and the feeling-of-knowing experience.
you are asked the telephone number of one of your Journal of Educational Psychology 56, 208216.
Leonesio, R. J., and Nelson T. O. (1990). Do different measures of
friends. metamemory tap the same underlying aspects of memory?
This initial feeling-of-knowing judgment that Journal of Experimental Psychology: Learning, Memory, and Cogni-
tion 16, 464470.
precedes an attempt to retrieve an answer was investi-
Nelson, T. O., and Dunlosky, J. (1991). When peoples judgments
gated by Reder (1987). She found that people were of learning (JOLs) are extremely accurate at predicting subse-
faster at making a feeling-of-knowing decision about quent recall: The Delayed-JOL Effect. Psychological Science
whether they knew the answer to a general- 2, 267270.
information question (e.g., What is the capital of Fin- Nelson, T. O., Gerler, D., and Narens, L. (1984). Accuracy of feel-
ing-of-knowing judgment for predicting perceptual identifi-
land?) than they were at answering the question
cation and relearning. Journal of Experimental Psychology: Gen-
(e.g., saying Helsinki). This finding demonstrates eral 113, 282300.
that the metamemory decision is made before, not Nelson, T. O., and Narens, L. (1984). Metamemory: A theoretical
after, the retrieval of the answer. If and only if people framework and some new findings. In G. H. Bower, ed., The
feel that they know the answer will they initiate at- psychology of learning and motivation. San Diego, CA: Academic
Press.
tempts to retrieve the answer from memory. When
Pressley, M., Levin, J. R., and Ghatala, E. (1984). Memory strategy
they feel that they do not know the answer, they dont monitoring in adults and children. Journal of Verbal Learning
even attempt to search memory (as in the example of and Verbal Behavior 23, 270288.
the presidents telephone number). Reder, L. M. (1987). Strategy selection in question answering. Cog-
nitive Psychology 19, 90138.
Control of the Termination of Extended At- Son, L. K., and Metcalfe, J. (2000). Metacognitive and control strat-
tempts at Retrieval. People may initially feel strongly egies in study-time allocation. Journal of Experimental Psycholo-
enough that they know an answer to begin searching gy: Learning: Memory and Cognition 26, 204221.
Underwood, B. J. (1966). Individual and group predictions of item
memory for it, but after fruitless extended attempts at
difficulty for free learning. Journal of Experimental Psychology
retrieval, they will terminate the search. Nelson, Gerler, 71, 673679.
and Narens (1984) found that people would search lon-
ger for a sought-but-not-retrieved answer if they felt Thomas O. Nelson
strongly that they knew the answer. Put differently, the Revised by Thomas O. Nelson and Petra Scheck
amount of time that elapses before someone gives up
searching memory for a nonretrieved answer tends to in-
crease with the degree of the feeling that the person
knows the answer. METAPLASTICITY
For example, our hypothetical student men- Most neural connections exhibit synaptic plasticity,
tioned above might spend a long time during the ex- increases or decreases in synaptic efficacy. Several dis-
amination attempting to retrieve the English equiva- tinct forms of synaptic plasticity exist, differing in
lent of chateau (which the person had studied the both their induction requirements and time course of
night before) but little or no time attempting to re- expression. Synaptic plasticity allows for dynamic
trieve the English equivalent of cheval (which the per- modification of neural circuitry that can act on time
son did not study previously). The metamemory deci- scales ranging from milliseconds to potentially life-
sion to continue or terminate attempts at retrieving times, and has been implicated in a wide range of
an answer from memory may also be affected by other neural and behavioral phenomena including learning
factors, such as the total amount of time available dur- and memory. A body of literature has developed
ing an examination. demonstrating that the ability to induce synaptic plas-
ticity is itself modifiable; that is, plasticity is plastic.
This is referred to as metaplasticity, a higher-order
Bibliography
form of synaptic plasticity. Metaplasticitys critical
Arbuckle, T. Y., and Cuddy, L. L. (1969). Discrimination of item
strength at time of presentation. Journal of Experimental Psy- feature is that, once instantiated, it modifies the abili-
chology 81, 126131. ty of subsequent activity to alter synaptic efficacy, fun-
Bisanz, G. L., Vesonder, G. T., and Voss, J. F. (1978). Knowledge damentally altering the rules that govern when and
of ones own responding and the relation of such knowledge how changes are expressed.
to learning. Journal of Experimental Child Psychology 25, 116
128. Metaplasticity has been observed in a variety of
Fischhoff, B., Slovic, P., and Lichtenstein, S. (1977). Knowing with brain systems and across different species, suggesting
certainty: The appropriateness of extreme confidence. Jour- that it may be a ubiquitous feature of synaptic opera-
nal of Experimental Psychology: Human Perception and Perfor- tion. Of particular interest, metaplasticity is observed
mance 3, 552564.
Griffin, D., and Tversky, A. (1992). The weighing of evidence and in brain structures that have been implicated in learn-
the determinants of confidence. Cognitive Psychology 24, 411 ing and memory (e.g., hippocampus, amygdala, cor-
435. tex), suggesting that it may play a key role in the regu-
388 METAPLASTICITY
lation of information processing. A state of An important theoretical feature of m is that it

metaplasticity can be created by the same factors that represents a sliding modification threshold, the level
can induce plasticity itself: the intrinsic activity of a of which depends on the recent average activity of the
neuron (homosynaptic factors) as well as through the synapse. This has been confirmed experimentally in
actions of hormones and neuromodulators (hetero- the hippocampus, where prior activation of a synapse
synaptic factors). To distinguish between these two was found to modify activity thresholds for LTP and
causal factors, a distinction can be made between ac- LTD (i.e., modify m). For example, high frequency
tivity-dependent metaplasticity, which refers to states priming stimuli increases the threshold activity level
created homosynaptically; and modulatory metaplas- required for LTP (modifies m due to an increase in
ticity, which refers to states created heterosynaptical- average activity), making it less likely that LTP will be
ly. induced and in parallel increasing the ability to in-
duce LTD. Conversely, low frequency priming stimuli
modifies m to decrease the threshold activity require-
Activity-Dependent Metaplasticity ment for LTP; thus reducing the probability of LTD.
The intrinsic activity of a neuron has long been A behavioral means of altering m was demonstrated
known to result in synaptic plasticity. The duration of in the developing visual cortex by Bear and Ritten-
the induced change, as well as the type of plasticity in- house. In these experiments, early visual deprivation
volved, depends upon the frequency, duration, and (which would decrease the average activity of cortical
pattern of activity. For example, high frequency acti- neurons) facilitated the induction of LTP compared
vation can produce long-term potentiation (LTP), an to light-reared animals of similar age. These effects
enhancement of synaptic strength that can last for were reversed by as little as two days of visual experi-
hours. LTP has emerged as a leading cellular mecha- ence, demonstrating the importance of neural activity
nism underlying learning, and is perhaps best charac- in adjusting the modification threshold.
terized in the hippocampus. It is now understood that The underlying cellular mechanisms regulating
the ability to induce LTP is not a static feature of hip- plasticity are only partially understood. One possibili-
pocampal neurons, but instead is strongly influenced ty is that Ca++ entry during NMDAR activation may
by the recent activation history of the neuron. For ex- activate a molecular cascade leading to an alteration
ample, a low frequency priming burst, which is insuf- of the NMDAR itself. Because the NMDAR is a critical
ficient for producing LTP, was found to increase the point of calcium entry into the postsynaptic neuron,
subsequent activity requirements for inducing LTP. these changes would significantly alter the properties
This effect was transient, synapse-specific, and de- of activity-dependent synaptic plasticity. In addition
pended (as does LTP) on postsynaptic N-methyl-D- to signaling through the NMDAR, activation of meta-
aspartate receptor (NMDAR) activation. botropic glutamate receptors (mGluRs) has also been
shown to induce metaplasticity. For example, activa-
It is now understood that the induction of plastic-
tion of group one mGluRs has been shown to facilitate
ity at hippocampal synapses is more complex than
LTP. The effects of group two mGluRs seem more
first appreciated. Depending upon the activity of the
complex, with reports of group two mGluR activation
neuron, the expression of plasticity can be bidirec-
facilitating the induction of LTD in the dentate gyrus
tional, either exhibiting a long-term enhancement
in vivo, but inhibiting LTD and facilitating LTP in
(LTP), or a long-term decrement (long-term depres-
area CA1 of the hippocampus in vitro. Thus, it ap-
sion; LTD). For example, 900 pulses at 1-3 Hz pro-
pears that there are multiple regulatory mechanisms
duces LTD, 900 pulses at 10 Hz results in no lasting
underlying activity-dependent metaplasticity, sug-
change, and 900 pulses at 50 Hz produced LTP.
gesting that there may be multiple modes of control
These effects are blocked by the application of an
over long-term synaptic plasticity.
NMDA receptor antagonist, suggesting a relatively
straightforward relationship between the activity of a Two possible functions for activity-dependent
presynaptic neuron, postsynaptic Ca++ entry, and the metaplasticity have been proposed. First, it may pro-
direction of change. These data suggest the existence vide a mechanism to promote stability within a neural
of a crossover point, where activity below this point network by maintaining synaptic parameters within
produces LTD, activity at this point (e.g., 10 Hz) pro- some optimal operating range. For example, in-
duces no change, and activity above this point pro- creased levels of activity (which lead to enhanced syn-
duces LTP. This crossover point is represented as the aptic efficacy) would concurrently raise the threshold
modification threshold theta (m) proposed as a theo- for subsequent enhancement, thus applying a brake
retical neural plasticity algorithm in artificial neural on facilitation. This would prevent saturation of the
networks in what is now commonly known as the BCM neural network at some ceiling level. Further, the
rule. sliding threshold would lead to the promotion of syn-
METAPLASTICITY 389
aptic decrement, perhaps providing a means for re- quences: Following tail shock, the same tactile experi-
setting synaptic weights within a network. Conversely, ence that normally produced reflex inhibition instead
low activity at a synapse would slide m in the opposite resulted in either no change or a trend towards reflex
direction, promoting synaptic enhancement and pre- enhancement. Thus modifying the capacity for STP
venting saturation at some floor level. Second, meta- fundamentally altered a basic regulatory response
plasticity may provide a functional means for inte- normally exhibited by the animal.
grating neural activity over extended periods of time.
This function may have an especially important role
in regulating synaptic connectivity in development, Stress and Metaplasticity
allowing experience occurring over extended periods A rapidly growing body of literature demon-
of time to exert top-down control over the refinement strates that behavioral stress can regulate LTP and
of synaptic connectivity. serve as an example of modulatory metaplasticity in
the mammalian brain. Studies have demonstrated
Modulatory Metaplasticity that acute and severe behavioral stress (e.g., by re-
straining and shocking the animal) produces a
Modulatory metaplasticity is the regulation of marked impairment of LTP. Interestingly, the same
synaptic plasticity through the actions of extrinsic fac- stress that blocked the induction of LTP facilitates the
tors such as neuromodulators and hormones. It is im- induction of LTD, suggesting that stress may in fact
portant to distinguish modulatory metaplasticity be capable of modifying m. Consistent with this,
from the myriad number of other effects that neuro- more mildly stressful situations can in fact produce a
modulators can have on the process of synaptic com- facilitation of LTP, suggesting that the level of stress
munication. Modulatory metaplasticity refers to the is analogous to the level of average activity in deter-
regulation of the induction of synaptic plasticity. It
mining the direction of m modification. Prime candi-
differs from a more generalized modulatory scaling
dates for mediating these effects of stress on synaptic
of synaptic efficacy, which would affect all aspects of
plasticity are the corticosteroids released by the adre-
synaptic transmission and plasticity in equal fashion.
nal cortex in times of stress. The hippocampus con-
A second characteristic of this form of metaplasticity
tains an abundance of corticosteroid receptors which
is that it does not require the activity of the synapse
are of two types: (1) a high-affinity mineralocorticoid re-
whose plasticity is being regulated. This feature dis-
ceptor (MR); and (2) a lower-affinity glucocorticoid recep-
tinguishes modulatory metaplasticity from the activi-
tor (GR). Differential binding to these receptor types
ty-dependent forms described above.
may underlie the differential effects of mild and se-
A primary example of modulatory metaplasticity vere stress. MR agonists facilitate LTP, whereas GR
has been described in the synaptic regulation of the agonists impair LTP and facilitate the induction of
defensive siphon withdrawal reflex (SWR) in the marine LTD. Therefore low levels of corticosteroids would
invertebrate Aplysia. The SWR is subject to multiple primarily act through the high-affinity MR receptor
forms of regulation including a dynamic form based and facilitate LTP, with higher levels leading to great-
upon the recent tactile experience of the animal. For er activation of the GR receptor leading to suppres-
example, a local water disturbance (turbulence) can sion of LTP.
produce a transient inhibition of the SWR that is a
consequence of an elevation of reflex threshold, es- The cellular mechanism underlying this form of
sentially increasing the signal requirement for reflex modulatory metaplasticity may be similar to that of
initiation in the face of greater environmental noise. activity-dependent metaplasticity, in that both appear
This form of behavioral regulation can be mediated to involve Ca++ influx into the postsynaptic cell. Cor-
by an activity-dependent form of synaptic enhance- ticosteroids have been shown to induce a rise in Ca++
ment called short-term plasticity (STP) that is ex- influx, the levels of which may depend upon levels of
pressed by identified inhibitory interneurons in re- corticosteroids. In a similar fashion as proposed for
sponse to tactile stimulation. This transient elevation activity-dependent metaplasticity, low levels of Ca++
of inhibitory synaptic strength reduces the ability to entry produced by mild stress may lead to a functional
activate excitatory interneurons in the SWR circuit; shift in m leading to a promotion of LTP, whereas
thus imposing a requirement for greater sensory more severe stress results in greater levels of Ca++
input in order to activate these neurons (and subse- entry that shifts m in the opposite direction, impair-
quently the reflex). That specific forms of STP ex- ing LTP and promoting LTD. Thus an extrinsic mod-
pressed by these inhibitory neurons could be signifi- ulator may regulate plasticity in fundamentally the
cantly suppressed by tail shock and the subsequent same way as was observed with intrinsic activity. It is
release of serotonin was an important finding. This important to note that other factors besides the corti-
regulation of STP had direct behavioral conse- costeroids have been implicated in stress-induced
390 MIGRATION, NAVIGATION, AND HOMING
regulation of LTP, including the cytokine inter- postnatal development. Proceedings of the National Academy of
leukin-1beta and endogenous opioids. Sciences of the United States of America 96, 12,87612,880.
Thomas J. Carew
Thomas M. Fischer
Conclusion
It is now understood that the ability to induce
many different forms of synaptic plasticity is not a
static feature of a neuron, but is regulated both by the MIGRATION, NAVIGATION,
recent activity history of the synapse, as well as by the AND HOMING
presence of a number of neuromodulators. While the
detailed means by which regulation is achieved can be The prevailing view among behavioral biologists and
diverse, most seem to share a common feature in that ethologists of the 1950s was that the remarkable abili-
they act through the same signaling pathways that in- ty of migratory animals, especially birds, to return to
duce plasticity, most notably through Ca++ signaling the same breeding and wintering area year after year
in the neuron. The apparent complexity of regulation was based on innate mechanisms of orientation and
poses a fundamental challenge in trying to relate syn- navigation. Later this emphasis on hereditary control
aptic events to behavioral output, because the direc- yielded to more dynamic conceptualizations of spa-
tion of plasticity and even the capacity for induction tial-behavior mechanisms. The revised theory assigns
may vary based upon a number of intervening factors. a larger role to environmental influences and learn-
Clearly, metaplasticity endows the nervous system ingexperience-dependent change in behaviorin
with additional degrees of freedom in the dynamic animal orientation and navigation.
regulation of neural circuits and systems, adding yet
another dimension to the already remarkable com- Orientation
plexity of the brain.
Orientation refers to a heading or directed move-
ment that bears a specific spatial relationship to some
See also: STRESS AND MEMORY environmental or proprioceptive reference. It is typi-
cally discussed metaphorically in terms of compass di-
rections when the sun, stars, or the earths magnetic
Bibliography field are used as orientation stimuli. Birds that mi-
Abraham, W. C., and Bear, M. F. (1996). Metaplasticity: The plas- grate at night use all three of these environmental
ticity of synaptic plasticity. Trends in Neuroscience 19, 126130. stimuli to orient their seasonal movements. Nocturnal
Bear, M. F., and Rittenhouse, C. D. (1999). Molecular basis for in-
duction of ocular dominance plasticity. Journal of Neurobiology
migrants show an impressive ability to vary their re-
41, 8391. sponse in accordance with changes of information
Bienenstock, E. L., Cooper, L. N., and Munro, P. W. (1982). Theo- from one orientation stimulus to another. For exam-
ry for the development of neuron selectivity: Orientation ple, several species of birds have been shown to
specificity and binocular interaction in visual cortex. Journal change their response based on experience with dif-
ferent ambient magnetic fields. Wiltschko and
de Kloet, E. R., Oitzl, M. S., and Joels, M. (1999). Stress and cogni-
tion: Are corticosteroids good or bad guys? Trends in Neuro-
Wiltschko found that shifting the orientation of an
science 22, 422426. ambient magnetic field resulted in birds correspond-
Dudek, S. M., and Bear, M. F. (1992). Homosynaptic long-term de- ingly shifting their migratory orientation. The birds
pression in area CA1 of hippocampus and effects of N- likewise shifted their orientation with respect to the
methyl-D-aspartate receptor blockade. Proceedings of the Na- stars, which were not subjected to experimental shift-
tional Academy of Sciences of the United States of America 89,
ing. This behavior suggests that birds could learn a
4,3634,367.
Fischer, T. M., Blazis, D. E., Priver, N. A., and Carew, T. J. (1997). new orientation response to the stars by using the
Metaplasticity at identified inhibitory synapses in Aplysia. Na- magnetic field as a calibrating reference. Bingman
ture 389, 860865. and Wiltschko described similar results in birds of an-
Huang, Y. Y., Colino, A., Selig, D. K., and Malenka, R. C. (1992). other species that learned a new orientation response
The influence of prior synaptic activity on the induction of to the setting sun based on their magnetic field expe-
long-term potentiation. Science 255, 730733.
Kim, J. J., and Yoon, K. S. (1998). Stress: Metaplastic effects in the
rience.
hippocampus. Trends in Neuroscience 21, 50559. But the earths magnetic field is not always the
Kirkwood, A., Rioult, M. C., and Bear, M. F. (1996). Experience- primary calibrating orientation stimulus. Able and
dependent modification of synaptic plasticity in visual cortex.
Able found that in one species orientation to the
Nature 381, 526528.
Quinlan, E. M., Olstein, D. H., and Bear, M. F. (1999). Bidirection- earths magnetic field varied with information
al, experience-dependent regulation of N-methyl-D-aspartate gleaned from celestial cues. These findings demon-
receptor subunit composition in the rat visual cortex during strate that the migratory orientation of birds is modi-
MIGRATION, NAVIGATION, AND HOMING 391
fiable by experience and hence, to that extent, Navigation

learned. Homing is the ability of an animal to return to
The results described above were taken from ex- some goal location or home. Navigation is the range
periments with birds that had already experienced at of spatial behavior mechanisms that facilitate hom-
least one migration. Work with birds prior to their ing. Traditionally, true navigation has encompassed
first migration has revealed that experience during both the ability to specify ones location in space rela-
their first summer may have an even larger effect on tive to an undetectable goal location and the subse-
subsequent migratory behavior. Evidence suggests quent ability to determine a goal-oriented directional
that migrant birds are born with an inherited disposi- bearing.
tion to orient in a particular direction with respect to The spatial orientation evinced by migratory
the axis of celestial rotation and the earths magnetic birds does not necessarily involve true navigation, as
field. Emlen has shown that to learn a migratory- was shown in Perdecks experiments. Perdeck cap-
orientation response to specific star patterns, birds tured and marked thousands of starlings (Sturnus vul-
rely on the rotation of the night sky about its axis as garis) that were migrating through the Netherlands in
a directional reference. Once directional information autumn. The birds were then transported to Switzer-
from celestial rotation is transferred to the star pat- land and released. An examination of the locations
terns during the birds first summer, they can use the where the birds were subsequently recaptured re-
patterns as an independent source of directional in- vealed an important difference between adults that
formation. had already experienced one migration cycle and
Surprisingly, the inherited migratory-orientation young birds that were migrating for the first time.
response to the earths magnetic field can vary with Adult starlings were recovered primarily northwest of
a birds first summer experience. Bingman (1983) the release point. The northwesterly orientation cor-
responded to the direction needed to reach the birds
found that varying magnetic-field experiences during
normal wintering homes near the northern coast of
the summer resulted in birds learning different au-
France. Therefore, the adults displayed true naviga-
tumn migratory-orientation responses to the earths
tional behaviorthey succeeded in determining a
magnetic field. The advantage of being able to
course from an unfamiliar location that brought them
change the innate orientation preference to the
close to their traditional wintering quarters. Young
earths magnetic field is that birds raised in areas of
birds, in contrast, were recovered primarily southwest
different magnetic field declination (the angular dif-
of the release point. The southwesterly orientation
ference between magnetic north and geographic
corresponded to the direction required to reach the
north) are still able to maintain both seasonally and
birds winter homes from the area from which they
geographically appropriate migratory orientation.
were captured but not from the release location. Al-
Able and Able found that celestial rotation is the ref-
though displaying good orientation, the young birds
erence used by young birds to override their innate
failed to orient in a manner consistent with goal-
orientation response to the earths magnetic field and directed, true navigation.
learn a new response better suited to reaching their
winter homes. Like experienced birds, young birds In addition to emphasizing the difference be-
who have yet to engage in their first migration mani- tween orientation and navigation, Perdecks results
fest learned changes in orientation behavior based on demonstrate the importance of experiential learning
interpreting the directional relationships among a va- in at least some aspects of avian navigation toward a
riety of environmental stimuli. goal location. Young birds that had never been to the
wintering area could not navigate a course to it.
Birds are unique neither in their highly directed
Young birds on their first migration appear to em-
movements nor in their ability to learn new orienta-
ploy vector navigation, an innate disposition to fly in
tion responses. From its birth, the beachhopper
a certain direction for a fixed period of time in order
(Talitrus saltator), a small crustacean that inhabits the
to arrive in the general vicinity of their populations
shoreline of the Mediterranean, displays an orienta- wintering range. However, it is only after experienc-
tion response to the sun that enables it to quickly ing their winter home that they develop the ability to
move perpendicular to the shoreline axis in order to navigate to it from unfamiliar locations. It appears
avoid danger. Ugolini and Macchi have shown that that a similar learning process supports the ability of
exposing young animals to a different environment birds to navigate to and recognize the same breeding
can modify this innate orientation response. This is site year after year.
another example of altered orientation based on
learning the spatial relationship among salient envi- Salmon evince a similar aptitude for learning in
ronmental stimuli, in this case sun and shoreline. returning to breed in their natal stream, as has been
392 MIGRATION, NAVIGATION, AND HOMING
well documented by Cooper et al., who placed young navigational systems seem to be parts of a single
coho salmon (Oncorhynchus kisutch) in a tank of water learning mechanism, Ioal et al. (for the navigational
containing a specific odorant. They marked the fish map) and Gagliardo et al. (for landmark navigation)
and later released them into Lake Michigan. They have shown that brain lesions to the hippocampus can
then placed the odorant at the mouth of one stream impair navigational map and landmark navigation
that feeds into the lake. Sometime later, as the fish under some training conditions but not others. The
began to enter streams for breeding, the fish that had implications of the hippocampus research is that
been exposed to the odorant in the tank were much there is more than one kind of navigational map and
more likely to enter the specially odorized stream. more than one type of landmark navigation differing
The fish apparently learned the characteristic odor of at least in the neural mechanisms that support their
their stream during early development and used that acquisition.
information to guide their return home. Like orientation, learning in navigation is not
Perhaps the best-studied navigational system in unique to vertebrates. Gould has shown that honey
animals is that of the homing pigeon. A number of bees (Apis mellifera) can learn a familiar landmark map
distinct spatial-behavior mechanisms, nearly all of in a manner similar to that of homing pigeons and
them influenced by environmental experience, gov- that they are able to use this map from locations
ern pigeon homing. To orient in space, homing pi- where they have never been before as long as they can
geons prefer to rely on the sun, but they can also have maintain sensory contact with familiar landmarks.
recourse to the earths magnetic field. Wiltschko and What is remarkable about naturally occurring
Wiltschko have shown that sun orientation is not in- spatial learning is that it often occurs in the absence
nate but depends on a young birds learning the suns of any tangible external reward. For example, what is
path across the sky. Although they further reported rewarding about associating atmospheric odors with
that pigeons could learn to use the sun for orientation wind direction to a young pigeon enclosed in an out-
only during intervals of direct exposure to the sun, door aviary? What is rewarding about learning the
Budzynski et al. have shown that when young pi- suns path? It is difficult to explain such phenomena
geons, like bees and other animal groups, are allowed without assuming that animals are biologically predis-
to experience the sun for only one part of the day, posed to learn spatial relationships among environ-
that experience is sufficient for them to use the sun mental stimuli in a rapid and efficient way without de-
for orientation at any time of day. pendence on environmental reinforcement. Natural
Pigeons rely on at least two navigational mecha- selection has seemingly endowed animals with ner-
nisms: a navigational map that they can use from dis- vous systems that predispose intrinsic reinforcement
tant locations where they have never been before, and of exploration. A spontaneous proclivity to explore,
landmark navigation, which they can use when they therefore, is a crucial element in most learned spatial
are in sensory contact with familiar environmental behavior.
stimuli. In some young pigeons the ability to learn a See also: SPATIAL LEARNING: ANIMALS
navigational map depends on the opportunity to as-
sociate atmospheric odors with wind directions. Bibliography
Using fans to alter the relationship between wind di- Able, K., and Able, M. (1990). Calibration of the magnetic compass
rection and atmospheric odors experienced by differ- of a migratory bird by celestial rotation. Nature 347, 378380.
(1995). Interaction in the flexible orientation system of a
ent groups of young pigeons, Ioal et al. raised pi-
migratory bird. Nature 364, 230232.
geons that learned navigational maps that varied with Baker, R. (1984). Bird navigation: The solution of a mystery. London:
the birds experience of odors and wind direction. Houder and Stoughton.
Berthold, P. (2001). Bird migration: A general survey. Oxford: Ox-
Landmark navigation is similarly dependent on ford University Press.
experience. Wallraff has shown that young birds con- Bingman, V. (1983). Magnetic field orientation of migratory Sa-
fined to an outdoor aviary are able to learn a naviga- vannah sparrows with different first summer experience. Be-
tional map, as evidenced by their ability to orient to- haviour 87, 4353.
Bingman, V., and Wiltschko, W. (1988). Orientation of dunnocks
ward home when released from a distant, unfamiliar
(Prunella modularis) at sunset. Ethology 7, 19.
location. Such birds are nonetheless impaired in re- Budzynski, C., Dyer, F., and Bingman, V. (2000). Partial experi-
turning home because of an inability to navigate in ence with the suns arc is sufficient for all day sun compass ori-
the vicinity of their home aviary. The lack of opportu- entation in homing pigeons, Columba livia. Journal of Experi-
nity to fly outside their aviary rendered them unable mental Biology 203, 2,3412,348.
Cooper, J., Scholz, A., Horrall, R., Hasler, A., and Madison, D.
to learn to navigate by familiar landmarks, which is
(1976). Experimental confirmation of the olfactory hypothe-
important for navigation near home. Gagliardo et al. sis with artificially imprinted homing coho salmon (On-
have shown that landmark navigation is at least par- corhynchus kisutch). Journal of the Fisheries Resources Board of
tially based on visual information. Although the two Canada 33, 703710.
MNEMONIC DEVICES 393
Emlen, S. (1970). Celestial rotation: Its importance in the develop- tions (loci), such as those in a public building, were
ment of migratory orientation. Science 170, 1,1981,201. memorized. Second, some object was thought of to
Gagliardo, A., Ioal, P., and Bingman, V. (1999). Homing in pi-
geons: The role of the hippocampal formation in the repre-
represent each important part of the oration, such as
sentation of landmarks used for navigation. Journal of Neuro- a spear to represent the tenth topic, war. Third, the
science 19, 311315. image created for each topic was combined with the
Gagliardo, A., Odetti, F., and Ioal, P. (2001). Relevance of visual image of its corresponding location. The spear might
cues for orientation at familiar sites by homing pigeons: An be imaged as penetrating the tenth locus, a door.
experiment in a circular arena. Proceedings of the Royal Society
of London. 268, 2,0652,070.
While making his speech, the orator thought of each
Gould, J. (1986). The local map of honeybees. Do insects have cog- location in turn and used the image seen in his minds
nitive maps? Science 232, 861863. eye as the prompt for the next part of his address.
Healy, S. (1998). Spatial representation in animals. Oxford: Oxford After a few days, the images from the speech would
University Press. fade from memory, but the more highly learned loci
Ioal, P., Gagliardo, A., and Bingman, V.P. (2000). Hippocampal
participation in navigational map learning in young homing
could be used to memorize a new speech.
pigeons is dependent on training experience. European Jour- During the Middle Ages and the Renaissance
nal of Neuroscience 12, 742750. mnemonic devices were used not so much by orators
Ioal, P., Papi, F., Fiaschi, V., and Baldaccini, N. (1978). Pigeon
navigation: Effects upon homing behavior by reversing wind
for memorizing speeches as by scholars to classify and
direction at the loft. Journal of Comparative Physiology 128, 285 memorize all knowledge. This fascinating aspect of
295. mnemonic devices is surveyed by Frances Yates
Perdeck, A. (1958). Two types of orientation in migrating starlings, (1966).
Sturnus vulgaris L., and chaffinches, Fringilla coelebs L., as re-
vealed by displacement experiments. Ardea 55, 194202.
Ugolini, A., and Macchi, T. (1988). Learned component in the Principles of Mnemonic Learning
solar orientation of Talitrus saltator Montagu (Amphipoda:
Talitridae). Journal of Experimental Marine Biology and Ecology Most mnemonic procedures utilize the three
12, 7987. memory processes of symbolizing, organizing, and as-
Wallraff, H. (1966). ber die Heimfindeleistungen von Brieftau- sociating. Symbolizing is finding a memorable, pref-
ben nach Haltung in verschiedenartig abgeschirmten Vo- erably imageable, representation for what is to be
lieren. Zeitschrift fr vergleichende Physiologie 52, 215259.
Wiltschko, R., and Wiltschko, W. (1980). The development of sun
learned. In the aforementioned example, a spear rep-
compass orientation in young homing pigeons. Behavioral resents war. Organizing involves activating a knowl-
Ecology and Sociobiology 9, 135141. edge structure in memory, such as a set of loci, to
Wiltschko, W., and Wiltschko, R. (1990). Magnetic orientation and which new information can be associated. The new in-
celestial cues in migratory orientation. Experimentia 46, 343 formation must then be associated to components of
351.
the knowledge structure. These knowledge-structure
Verner P. Bingman components, labeled mental cues, must have certain
properties for mnemonic learning to be effective. De-
signing a mnemonic device must take into account
how easily mental cues can be reconstructed, as well
MISINFORMATION as how easily they can be associated with new informa-
tion and how easily they can be discriminated from
See: RECONSTRUCTIVE MEMORY
other mental cues (Bellezza, 1981). Visual imagery is
often used by the learner to create the association be-
tween the mental cue and the symbol to be remem-
bered.
MNEMONIC DEVICES
Mnemonic devices are methods for memorizing. The Types of Mnemonic Devices
ancient Greek poet Simonides of Ceos is the legend-
ary discoverer of mnemonic devices. Pleased by Si- Higbee (1988) discusses the wide variety of mne-
monidess praise, the twin gods Castor and Pollux monic procedures and some of the research done on
called him from a banquet just before the hall col- them. This entry attempts only a brief overview. Ex-
lapsed. The other guests were mangled beyond rec- amples of some of the mnemonic devices mentioned
ognition, but Simonides remembered the places they appear in Table 1.
had been sitting and so was able to identify the dead. The process of organizing is particularly impor-
Such was the discovery of the method of loci (or loca- tant when using a mnemonic technique such as the
tions). It became so much a part of the study of rheto- method of loci (see Figure 1), the story mnemonic, or
ric that the most venerable of the Roman orators used the link mnemonic. When using the story mnemonic,
the method of loci for memorizing their speeches. a list of words is memorized by creating a story from
Their procedure was as follows: First, a series of loca- them. The words become organized in memory by the
394 MNEMONIC DEVICES
Table 1
theme and context of the story. When the list words The keyword for hegemony might be gem. Next, a vi-
later have to be recalled, the story can be reviewed by sual image or sentence is formed that associates the
the learner and the list words recognized. key word with the meaning of the word to be learned.
The link mnemonic is somewhat different from In this instance, the sentence might be The nation with
the method of loci or the story mnemonic: The suc- the most gems rules the others. When coming across the
cessive words forming a pair in a list are associated word hegemony again, the learner must first think of
using a visual image. Each image can be distinct and the word gem, which then acts as a prompt for the
separate from the other images. All the words end up mnemonic sentence containing the meaning of the
joined together in memory by visual images like links word.
in a chain. A useful but difficult variation of the keyword
In other mnemonic devices the memory process mnemonic is the face-name mnemonic used for asso-
of symbolizing is paramount. When memorizing ciating names and faces. First, the persons name is
numbers, a system that dates back to the seventeenth transformed into a meaningful and concrete key-
century, the digit-consonant mnemonic, can be used. word, such as Cushing into cushion. The keyword then
In this system numbers are changed to words because has to be associated to a salient bodily feature. In this
words are more easily memorized than numbers. The example one might think of Mr. Cushing as having
words can then be recalled and changed back into the a cushion on his head because of his thick hair. When
numbers they represent. meeting Mr. Cushing again, one must recognize his
hair as his critical physical feature. His hair should act
One of the most useful mnemonics has been the
as a cue for the word cushion, which in turn will act as
keyword mnemonic, in which symbolizing processes
a prompt for the name Cushing.
play an important role. The keyword mnemonic is
particularly useful for learning the meanings of The mnemonic devices just reviewed are used to
words. For example, when learning the word hegemo- remember facts. Other less widely known mnemonic
ny, meaning authority of one nation over others, the techniquescalled process mnemonicsare de-
word must first be correctly pronounced (see Figure signed to help remember rules, principles, and proce-
2). It is pronounced something like he-GEM-oh-knee. dures. Process mnemonics have been used in Japan
Next, a keyword must be chosen that is familiar, to help teach mathematical rules and computational
meaningful, and sounds like the word to be learned. skills and chemical formulas. Process mnemonics in-
MNEMONISTS 395
corporate symbolizing, such as representing mathe- Figure 1

matical rules (e.g., to divide fractions, multiply by a
reciprocal) with more memorable and imageable
phrases (flip the fool into the pool, which refers to
a jogger who is fooling around on a diving board by
standing on his head). Association is used as well. In
this example, the division sign between fractions (two
joggers) is associated with a diving board. Higbee
(1987) provides a more complete description of pro-
cess mnemonics, which can be relatively complex.
Practical Applications
The more complicated mnemonic devices, such
as the digit-consonant mnemonic, the keyword mne-
monic, the face-name mnemonic, and process mne-
monics require study and practice in order to be effec-
tive. But little research has been performed to
determine if the investment in time and effort to be-
come proficient in these mnemonics results in im-
proved performance in the classroom, in the work-
place, in situations in which many peoples names recited the first 31,811 digits of pi from memory and
have to be remembered, and so on. However, interest that in 1987 Hideaki Tomoyori broke that record by
in this topic seems to be growing. The keyword mne- reciting the first 40,000 digits. People performing
monic has been shown to be useful to students learn- such feats of memory are called mnemonists or me-
ing classroom-type materials. This is true for both av- morists. Although feats like these are rare, since the
erage students and students with learning disabilities 1890s there have been several scientific accounts of
(Mastropieri et al., 1987). Using the pool process people with prodigious memories. Starting with the
mnemonic described above, one researcher found pioneering work of Alfred Binet, the scientific litera-
that after three one-hour class sessions, third graders ture describes over a dozen people showing excep-
performed calculations with fractions as well as sixth tional memory for verbal materials. Brown and Def-
graders and better than fourth and fifth graders (Hig- fenbacher (1975) give a comprehensive review of
bee, 1987). these studies. Studies of exceptional memory perfor-
mance contribute to our understanding of memory by
See also: MNEMONISTS
describing the processes memorists use and by com-
Bibliography paring them with processes used by people with ordi-
Bellezza, F. S. (1981). Mnemonic devices: Classification, character- nary memories.
istics, and criteria. Review of Educational Research 51, 247275.
Higbee, K. L. (1987). Process mnemonics: Principles, prospects,
and problems. In M. A. McDaniel and M. Pressley, eds., Imag- Representative Case Studies
ery and related mnemonic processes. New York: Springer-Verlag.
(1988). Your memory: How it works and how to improve it, 2nd The studies on memorists have shown that they
edition. New York: Prentice-Hall. use a variety of techniques to remember material.
Mastropieri, M. A., Scruggs, T. E., and Levin, J. R. (1987). Mne- Four memorists are presented here to demonstrate
monic instruction in special education. In M. A. McDaniel and
that variety: Shereshevskii, Alexander Craig Aitken,
M. Pressley, eds., Imagery and related and mnemonic processes.
New York: Springer-Verlag. VP, and Rajan.
Yates, F. A. (1966). The art of memory. London: Routledge and
Kegan Paul. Shereshevskii
A. R. Luria (1968) has made Shereshevskii (S) the
Francis S. Bellezza
Revised by Mark A. McDaniel
most famous mnemonist. (Luria referred to him only
as S, but his real name later became known.) S was al-
most thirty when Luria began his studies, and the re-
search continued for almost thirty years. Somewhat
surprisingly, S was unaware that his memory was un-
MNEMONISTS usual until Luria began his investigations.
The Guinness Book of World Records reports that in S used three basic processes, usually in combina-
1981 Rajan Srinavasen Mahadevan (known as Rajan) tion, for remembering verbal material. The first was
396 MNEMONISTS
Figure 2 until . . . I cant make anything out (Luria, 1968, p.

39).
Professor Aitken
Many psychologists think that Professor Aitken,
who lived from 1895 to 1967, was the best all-around
mnemonist. In a summary of the work on Aitken,
Hunter (1977) points out that he was a brilliant math-
ematician, an excellent mental calculator, and an ac-
complished violinist with an extraordinary memory.
His primary method for learning was to search out
meaningful relationships within the material and with
previously learned information. Hunter provides a
quote from Aitken that best captures his approach:
to generate rich visual images to represent informa- Musical memory can . . . be developed to a
tion. When he became a stage performer, he trained more remarkable degree than any other, for
himself to convert senseless words into meaningful we have a metre and a rhythm, a tune, or
images so that he could remember nonsense words or more than one, the harmony, the instrumen-
words from unfamiliar languages. The second process tal color, a particular emotion or sequence of
was to use familiar locations, such as stops on an oft- emotion, a meaning, . . . in the executant an
traveled street, to place the images mentally for later auditory, a rhythmic and a muscular and
retrieval. This procedure is the method of locations functional memory; and secondarily in my
(or loci) developed by the ancient Greek poet Simoni- case, a visual image of the page . . . perhaps
des of Ceos in about 500 B.C. The method of locations also a human interest in the composer, with
has been discussed by authors as diverse as Aristotle whom one may identify oneself . . . and an es-
and Thomas Aquinas. S apparently developed the thetic interest in the form of the piece. They
technique independently. The third process was to are so many, and they are so cumulative, that
create a story with appropriate images to retrieve the the development of musical memory, and ap-
information. preciation, has a multitude of supports.
(1977, p. 157)
With these techniques, S was able to remember
any information presented. Luria was unable to find Although Aitkens memory was prodigious, it was
any limit to the amount of material S could recall in not infallible. For example, in 1936 he correctly re-
this fashion. More surprisingly, there appeared to be called sixteen three-digit numbers after four presen-
no limit to the duration of Ss memory. Luria reports tations. Two days later, he recalled all but one of the
a request for recall of a fifty-word list given without numbers and, after an additional presentation, he re-
warning sixteen years after presentation of the list. called them all. In 1960, without further study, he re-
That request, like all the others Luria reports, result- called twelve of the numbers but also produced eight
ed in successful retrieval of the list. incorrect numbers.
VP
S had strong synesthesia, which appears to be
unique among the memorists who have been investi- VP (identified only by these initials in the pub-
gated. Synesthesia is said to occur when information lished report) is an excellent chess player whose
coming into one sensory system (e.g., audition) pro- memory has been investigated by Hunt and Love
duces an effect in another sensory system (e.g., vi- (1972). VP has an exceptional memory but, like Ait-
sion). S once said to the Russian psychologist Vy- kens, it is not infallible. For example, Hunt and Love
gotsky, What a crumbly, yellow voice you have reproduced VPs recall of an Indian story after inter-
(Luria, 1968, p. 24). On another occasion, Luria was vals of 1 hour and 6 weeks. Although there were small
concerned that S might not remember his way in an changes in both recalls, his overall accuracy was re-
unfamiliar location. S replied that he couldnt possi- markable.
bly forget because heres this fence. It has such a salty VP learned material by relating it to prior infor-
taste and feels so rough; furthermore, it has such a mation. For example, he knew several languages and
sharp, piercing sound. . . . (Luria, 1968, p. 38). Syn- could associate any three-letter string with a word. He
esthesia interfered with the images S produced and learned number matrices by rows and sometimes re-
presented an enduring problem for him. For exam- coded the row as a date. It is also clear that he spent
ple, S once noted that Other times smoke or fog ap- a great deal of time practicing memorizing so that he
pears . . . and the more people talk, the harder it gets, became very adept at recoding information.
MODALITY EFFECTS 397
Rajan Aitken uses that technique and also searches for rela-
Rajan has an exceptional memory for digits but tionships within the material to be learned. S uses im-
not for other material. A group of researchers from agery and the classic method of locations as his pri-
Kansas State University (Thompson, Cowan, and mary means for learning material. Because Ericsson
Frieman, 1993) performed extensive tests on his and his colleagues clearly refer to relating the materi-
memory. Their studies showed that Rajan learned al to preexisting verbal knowledge, Ss procedures do
sets of digits more rapidly than VP or S. He used a not conform to their theory. Rajan also does not fit
procedure pairing locations and digits to learn the the theory at all. His procedure, pairing locations and
material. He also encoded the digits in chunks (such digits, cannot be construed as encoding by relation to
as a row in a matrix). Thus, he learned that the fifth preexisting knowledge.
digit in the fourth row was 3 rather than using preex-
isting knowledge to encode the information. He ex- Conclusion
plicitly attached cues to the chunks for retrieval. For
The four memorists use quite different tech-
example, he learned the first column in a matrix as
niques to remember information, some of which call
a cue for retrieving each row of the matrix.
into question a portion of the theory of skilled memo-
Once the material was learned, Rajans procedure ry. Their memories are unusually good, but the pro-
allowed for extremely effective retrieval of informa- cesses they use to remember can all be used by people
tion. Working in the first 10,000 decimal digits of pi, with ordinary memories. In short, their memories are
he could retrieve a digit at a specified location (e.g., unusual in the amount they can remember but not in
digit 4,765) in an average time of twelve seconds. He the processes they use to remember.
had the digits of pi chunked in groups of 10 digits.
When he was given the first five digits of a ten-digit See also: MNEMONIC DEVICES
group in the first 10,000 digits of pi, he could give the
next five digits in an average time of seven seconds. Bibliography
Brown, E. (1988). Superior memory performance and mnemonic
encoding. In L. K. Obler and D. Fein, eds., The exceptional
brain. New York: Guilford Press.
The Memorists and a Theory of Skilled Brown, E., and Deffenbacher, K. (1975). Forgotten mnemonists.
Memory Journal of the History of the Behavioral Sciences 11, 342349.
Ericsson, K. A., and Chase, W. G. (1982). Exceptional memory.
In several papers, Ericsson and his colleagues American Scientist 70, 607615.
(e.g., Ericsson and Chase, 1982) suggest three general Hunt, E., and Love, T. (1972). How good can memory be? In A.
principles for skilled memory and illustrate these W. Melton and E. Martin, eds., Coding processes in human memo-
principles with people skilled at some aspect of mem- ry. Washington, DC: John Wiley.
ory. The three principles they propose are meaning- Hunter, I. M. L. (1977). An exceptional memory. British Journal of
ful encoding (the use of preexisting knowledge to Luria, A. R. (1968). The mind of a mnemonist. New York: Basic Books.
store the presented information in memory), retrieval Obler, L. K., and Fein, D. (1988). The exceptional brain. New York:
structure (explicitly attaching cues to the encoded Guilford Press.
material to allow efficient retrieval), and speedup (a Thompson, C. P., Cowan, T. M., and Frieman, J. (1993). Memory
reduction in study time with further practice). They search by a memorist. Hillsdale, NJ: Erlbaum.
claim that ordinary subjects, as well as skilled memo- Charles P. Thompson
rists, show these principles.
Consistent with this theory, all four mnemonists
described here attach retrieval cues when learning
material to ensure accurate and fast retrieval. Fur- MODALITY EFFECTS
ther, three of them show a reduction in study time In the classic modality effect, immediate recall of the
with practice. There is no clear evidence available on last few items from a verbal sequence is influenced by
this point for Professor Aitken. However, it seems the presentation modality: Recall is more likely if the
likely that he would show a similar effect. sequence is spoken aloud than if it is read silently.
The data from these memorists suggest that the Most people are familiar with the experience of brief-
skilled-memory theory founders on the third princi- ly retaining speech as if in a mental tape recorder and
ple. All four memorists use procedures for encoding occasionally using this echoic memory to do a dou-
the material that are available to and used by people ble take. An example is when one is asked the time
with ordinary memories. But, contrary to the theory, while reading. The sounds linger in memory and one
not all of them encode the material by relating it to can recover them to get the meaning even if one was
preexisting knowledge. VP fits the theory because he not paying attention at the exact moment when the
encodes material by relating it to prior information. question was asked. The auditory modality advantage
398 MODALITY EFFECTS
Figure 1 The Classic Effect

The classic modality effect occurs in comparing
the immediate recall of sequences of verbal items that
are presented in written or in spoken form. It does
not matter much whether it is the experimenter or
the subject who reads the spoken sequences aloud.
Recall performance is usually plotted as a function of
each items serial position in the sequence. The mo-
dality effect occurs both when recall itself is serial
that is, when the items have to be recalled in their
order of occurrenceand when recall is free, in the
sense that subjects are free to choose any order of re-
call. Digit sequences are frequently used for serial re-
call; sometimes letters or syllables are used, some-
times unrelated words. In free recall, the sequences
are almost always unrelated words. Two of the earliest
demonstrations of this modality effect are repro-
duced in Figure 1 (serial recall) and Figure 2 (free re-
call).
The modality effect is intimately bound up with

two other effects, the recency effect and the suffix ef-
fect. The recency effect is the finding that the last few
(or most recent) items from a sequence are more like-
has been widely attributed to an echoic memory sys- ly to be recalled than the preceding items. As Figures
tem that stores raw acoustic information for at least 1 and 2 illustrate, in serial recall there is little recency
several seconds no matter how ones attention is di- effect for the silent sequences, but in free recall both
rected during stimulus presentation. input modes show a large recency effect and the mo-
dality effect appears as an enhanced recency effect.
Evidence shows that modality effects can occur The suffix effect is the finding that in serial recall, a
for sounds that were presented too long ago for echo- single spoken item at the end of the sequence, such
ic memory still to be used, and in situations in which as zero after the last to-be-recalled digit, essentially
there are no sounds. There is an advantage, for exam- wipes out the modality effect. With the suffix, there is
ple, in recalling lip-read or signed words over silently as little recency effect for spoken as for silent se-
read words. There are even certain circumstances that quences.
favor visual as opposed to auditory presentations. As
a result, psychologists often use the term modality effect The original explanation of the modality effect
to refer to any differences in memory performance was the Precategorical Acoustic Storage (PAS) model
that are associated with differences in stimulus mo- described by Robert G. Crowder and John Morton in
dality. Such modality effects are more pervasive, and 1969. PAS is an auditory sensory memory store, the
of more fundamental importance, than researchers echoic counterpart of the visual sensory store (iconic
had previously thought. They show how various memory). Its function is to retain speech input at pre-
codes, or derived types of information, are used in categorical levelthat is, prior to analysis of mean-
memory. For example, when one hears a spoken ingfor further processing. It was supposed to last
word, one can reflect upon how it sounds (an acoustic about two seconds unless it was overwritten or erased
code), how the word is formed from consonants and by subsequent speech input. Thus, the modality effect
vowels (a phonological code), how one would
presumably occurs because echoic memory of the last
pronounce it (an articulatory code), how the word
few items, unlike iconic memory, can persist long
would look if printed (an orthographic code), and
enough to contribute to immediate recall. This theory
what the word means (a semantic code). Such codes
was useful in understanding modality, recency, and
are preserved and processed to varying degrees
suffix effects as well as speech perception. The limita-
and are used together as cues assisting the later recall
of the word. The presentation modality influences tion in the PAS account is its inability to explain other
how well or how easily various codes can be modality effects that have been discovered since
formed. 1969.
MODALITY EFFECTS 399
Subsequent Developments
Three topics of special importance have led to
modifications in scientific understanding since
Crowder and Mortons PAS theory. They are modality
effects in the absence of sound, long-term modality
effects, and the role of stimulus timing in causing mo-
dality effects.
Modality Effects in the Absence of Sound

Modality and suffix effects in serial recall occur
when the items are lip-read or mouthed silently in-
stead of being spoken aloud, as was shown for exam-
ple by Kathryn T. Spoehr and William J. Corin in
1978. These findings led to a revision of the PAS
model and to several other theoretical developments.
The revised PAS model, described by Crowder in
1983, assumes that information in PAS can also be ac-
tivated by the facial-gestural features involved in lip-
reading and mouthing, on the grounds that these fea-
tures are involved in speech perception. Evidence of
modality and suffix effects with American Sign Lan-
guage in the congenitally deaf led Michael A. Shand
and Edward S. Klima to propose a primary linguistic
coding hypothesis, which assumes that such effects
occur whenever the stimulus modality is compatible
with the short-term memory code involved in a per-
sons primary mode of communication. The revised There are also modality effects in long-term as
hypotheses have merit but it is not clear if they can ex- well as short-term memory. For example, a spoken
plain all modality effects. For instance, they cannot
modality effect occurs in a free recall task, in which
explain modality and suffix effects that have been dis-
each word in the sequence is preceded and followed
covered with tactile stimuli and with musical notes or
by lengthy periods of spoken distractor activity, as
rhythms.
John M. Gardiner and Vernon H. Gregg showed in
Long-Term Modality Effects 1979. Such interference is assumed to eliminate any
There are various indications that auditory- contribution from sensory or short-term memory be-
modality-specific information persists longer than the cause it wipes out modality and recency effects if it oc-
two seconds that was originally assumed by the PAS curs only after the last word in the sequence.
theory. One is that modality effects have been shown
A few modality effects have been discovered in
to result largely from the ability of auditory informa-
other long-term memory tasks, including recognition
tion to survive in the presence of what is called output
memory, where Martin A. Conway and Susan E. Ga-
interference, or interference from the early part of
thercole found an effect across all serial positions.
the participants own response. Several investigators
Some of these effects are a reversal of the usual audi-
(including Fergus I. M. Craik in 1969, C. Philip Bea-
man and John Morton in 2000, and Nelson Cowan tory superiority, and they seem to interact with other
and others in 2002) have found that when partici- factors in ways that are as yet poorly understood. In
pants recall items from the end of the list first, lists in long-term recall with written responses, either visual
both modalities are recalled equally well. When par- presentation or spoken presentation along with in-
ticipants recall items from the beginning of the list structions to imagine how the word is printed help
first, however, items from the end of the list are re- participants to avoid false memories. In such
called much better for spoken lists than for printed studies, when numerous words from a semantic cate-
lists. Part of these effects may occur because the writ- gory are presented (e.g., sleeve, button, collar, and so
ten or typewritten responses that were used overwrite on) there is a tendency to recall a central word that
orthographic codes but Fergus Craik showed that actually was not presented (e.g., shirt) and the avail-
much of the auditory advantage appears to persist ability of an orthographic code in memory appears to
even with spoken responses. Auditory codes are thus prevent that sort of mistake, as Ronald T. Kellogg
more likely to survive output interference. showed in 2001.
400 MORPHOLOGICAL BASIS OF LEARNING AND MEMORY
Modality and Stimulus Timing al, aspects of memory function than originally envis-
The long-term modality effect led Arthur M. aged in echoic memory theory.
Glenberg and Naomi G. Swanson in 1986 to propose See also: MEMORY SPAN
a detailed model that assumes that temporal informa-
tion is more finely represented in the auditory mode Bibliography
and that the modality effect reflects greater temporal Beaman, C. P., and Morton, J. (2000). The separate but related ori-
distinctiveness. Because it is possible that temporal gins of the recency effect and the modality effect in free recall.
Cognition 77, B59B65.
distinctiveness underlies the recency effect in both
Conway, M. A., and Gathercole, S. E. (1987). Modality and long-
short-term and long-term memory, it can potentially term memory. Journal of Memory and Language 26, 341361.
provide a quite general account of modality and re- Cowan, N., Saults, J. S., Elliott, E. M., and Moreno, M. (2002). De-
cency effects. A number of current studies have inves- confounding serial recall. Journal of Memory and Language,
tigated temporal and ordinal factors. Though re- 46,153177.
Crowder, R. G. (1983). The purity of auditory memory. Philosophi-
search indicates that serial-order information is cal Transactions of the Royal Society of London B302, 251265.
better retained in the auditory mode, the evidence Crowder, R. G., and Morton, J. (1969). Precategorical acoustic stor-
makes it less clear if the same is true for temporal in- age (PAS). Perception and Psychophysics 5, 365373.
formation as such, as Ian Neath and Robert Crowder Frankish, C. (1989). Perceptual organization and precategorical
acoustic storage. Journal of Experimental Psychology: Learning,
showed in 1990.
Memory, and Cognition 15, 469479.
Gardiner, J. M. (1983). On recency and echoic memory. Philosophi-
One indication that the temporal-information ac-
cal Transactions of the Royal Society of London B302, 267282.
count has merit comes from studies of the grouping Gardiner, J. M., and Gregg, V. H. (1979). When auditory memory
of items. In 1989 Clive Frankish showed that the ben- is not overwritten. Journal of Verbal Learning and Verbal Behav-
efit that comes from items in a list being separated by ior 18, 705719.
gaps into small groups is much greater for spoken Glenberg, A. M., and Swanson, N. G. (1986). A temporal distinc-
tiveness theory of recency and modality effects. Journal of Ex-
lists than for printed lists. Nelson Cowan and others
perimental Psychology: Learning, Memory, and Cognition 12, 315.
(2002) set up a situation in which lists of nine random Kellogg, R. T. (2001). Presentation modality and mode of recall in
digits were presented at a steady pace on some trials verbal false memory. Journal of Experimental Psychology: Learn-
but, on other trials, were grouped into three sets of ing, Memory, and Cognition 27, 913919.
three with one-second gaps between groups. The re- Morton, J., Crowder, R. G., and Prussin, H. A. (1971). Experiments
with the stimulus suffix effect. Journal of Experimental Psycholo-
sponse boxes to be filled in by the participant always gy 91, 169190.
appeared as three groups of three across the screen. Nairne, J. S. (1990). A feature model of immediate memory. Memo-
This encouragement of grouping had a small benefit ry & Cognition 18, 251269.
for the recall of printed lists, equivalently no matter Neath, I., and Crowder, R. G. (1990). Schedules of presentation
and temporal distinctiveness in human memory. Journal of Ex-
whether the actual list was grouped or not. However,
perimental Psychology: Learning, Memory, and Cognition 16, 316
for spoken lists, there was a very large benefit of actu- 327.
al stimulus groupingmuch larger than for visual Penney, C. G. (1989). Modality effects and the structure of short-
listsand no noticeable mental grouping for steadily term memory. Memory & Cognition 17, 398422.
paced lists. Thus, the mental representation of the list Shand, M. A., and Klima, E. S. (1981). Nonauditory suffix effects
in congenitally deaf signers of American Sign Language. Jour-
was much more dependent on the actual stimulus nal of Experimental Psychology: Human Learning and Memory 7,
timing for spoken lists than for printed lists. 464474.
Spoehr, K. T., and Corin, W. J. (1978). The stimulus suffix effect
Modality effects have a broader empirical base as a memory coding phenomenon. Memory & Cognition 6,
than psychologists once realized, and late-twentieth- 583589.
century discoveries have led to fresh theoretical ap- John M. Gardiner
proaches, among which temporal distinctiveness the- Nelson Cowan
ory perhaps is the most influential. In addition to the
alternative hypotheses already mentioned, there are
other suggestions that there has been less time to
evaluate. James S. Nairne suggested (in 1990) con- MORPHOLOGICAL BASIS OF
ceptualizing the memory trace in terms of modality- LEARNING AND MEMORY
independent and modality-dependent features,
which may or may not reflect sensory aspects of the [An increasing body of evidence indicates that learning and
stimulus. Catherine G. Penney (in 1989) developed a memory involve changes in the strength of the connections
separate streams hypothesis, which assumes that between neurons. However, changes in synaptic strength
modality separation of verbal items is inherent to the can be realized in many ways. One exciting possibility is that
structure of short-term memory. All these ideas assign changes in strength are produced by growth of new synaptic
modality effects to more fundamental, less peripher- contacts. This possibility was suggested more than a century
MORPHOLOGICAL BASIS OF LEARNING AND MEMORY: Invertebrates 401
ago, but has received strong experimental support since the learning: the siphon-gill withdrawal reflex and the
1980s. These results are reviewed in the following two entail-siphon withdrawal reflex.
tries, which focus on studies in INVERTEBRATE model sys-
tems and VERTEBRATE model systems respectively.] Both withdrawal reflexes can be enhanced by sen-
sitization, a form of nonassociative learning (Castel-
lucci, Pinsker, Kupfermann, and Kandel, 1970; Wal-
ters, Byrne, Carew, and Kandel, 1983). Sensitization
INVERTEBRATES is the enhancement of a behavioral response resulting
from the application of a novel or noxious stimulus
There is much evidence that learning and memory to the animal. Sensitization in a variety of animals
result from cellular changes occurring within individ- seems to be a short-lived phenomenon, recovering
ual neurons (Byrne, 1987), but the cellular mecha- over the course of minutes to an hour. The long-term
nisms underlying these changes have resisted defini- form of sensitization lasts from one day to roughly
tive elucidation. The mechanisms that produce a three weeks. Recent evidence suggests that the situa-
change in neuronal structure have received scant at- tion may be more complex. For example, there is a
tention even though this class of mechanisms was first distinct form of sensitization with intermediate dura-
suggested more than a century ago (Ramn y Cajal, tion (Sutton and Carew, 2000). Investigators have
1988). Researchers have observed morphological cor- identified several elements of the circuits mediating
relates of learning in several vertebrate and inverte- these reflexes. In both circuits, the sensory neurons
brate model systems (Bailey and Kandel, 1993), but are sites of short- and long-term plasticity. Neurite
no one has proved that modifications of neuronal outgrowth followed long-term but not short-term sen-
structure affect behavior. Nevertheless, the variety of sitization.
model systems with morphological correlates sug-
gests that this is one of the mechanisms that underlies Following sensitization of the siphon-gill with-
learning. drawal reflex, training increased the complexity of
the axonal arbor of sensory neurons, as measured by
An investigation of the morphological correlates total neurite length (Bailey and Chen, 1988a). In ad-
of learning and memory should incorporate several dition, the number of varicosities per neuron in-
features. First, there should be a readily observable creased. These observations were consistent with en-
behavior that can be modified by training. Second, hanced convergence onto follower motor neurons.
the circuit mediating that behavior should be well Tail sensory neurons yielded similar results (Wain-
known. Third, the morphology of the neurons in that wright, Zhang, Byrne, and Cleary, 2002). The effects
circuit must be accessible to analysis. Although no of training, however, were limited to a region of the
real-world model system meets all of these criteria, a arborization in the pedal ganglion, which is where fol-
number of investigators have turned to invertebrates lower motor neurons are located. The total arboriza-
for model systems with few compromises. tion length in this region was increased, as was the
The structural modifications vary among differ- number of branch points and varicosities.
ent systems. This is not surprising because morpho-
The honeybee is an interesting case because it has
logical features may contribute to neuronal function
a powerful facility for learning the features and loca-
in several ways. This article will focus on two broad
tion of food. One of the most profound learning ex-
classes of morphological changes: the large-scale re-
periences for a foraging bee occurs on its first flight
modeling of neuronal arborizations arising from out-
from the hive. Presumably, the animal learns naviga-
growth that seem to result in the formation of new
tional cues that the target reinforces, but the precise
synaptic connections and more localized structural
coding of the information remains unclear.
changes that presumably affect preexisting connec-
tions. Given the complexity of learning, it likely draws
on large portions of the brain. Some research has
pointed to a correlation between learning and an in-
Neurite Outgrowth crease in volume of two brain regions: the mushroom
Cajals original hypothesis suggested that learn- body and antennal lobe glomeruli (Farris, Robinson,
ing is mediated by the outgrowth of new axons and and Fahrbach, 2001; Sigg, Thompson, and Mercer,
the formation of new synaptic connections, a process 1997). In the mushroom body, the growth appears to
that recapitulates brain development (Ramn y Cajal, be due at least in part to an increase in sprouting from
1988). This hypothesis has been tested in detail in the Kenyon cell dendrites (Farris, Robinson, and Fahr-
marine mollusk Aplysia. Two simple defensive reflexes bach, 2001). A form of olfactory learning in the fruit
in this animal have proved useful for studies of the fly Drosophila yielded similar results (Balling et al.,
cellular and molecular mechanisms of nonassociative 1987).
402 MORPHOLOGICAL BASIS OF LEARNING AND MEMORY: Invertebrates
A distinctive approach to morphological corre- does the size of the axonal arborization. This out-
lates has been taken with Drosophila. Rather than growth results in an increase in the number of active
looking directly for correlates following a training ses- zones, in the length of the active zone, and the num-
sion, investigators generate mutants that are deficient ber of vesicles in proximity to the release site (i.e., the
in learning and observe effects of the mutations on vesicle complement; Bailey and Chen, 1983). These
neuronal structure. In some cases these mutations same features are smaller in animals subjected to
have been found to affect the morphology of neurons long-term habituation. For example, sensory neurons
in the central nervous system. Two commonly used from these animals had on average 35 percent fewer
mutations are dnc and rut. Both mutations impair a varicosities compared to controls. Quantitative analy-
form of olfactory conditioning in which flies associate sis of the time course over which these anatomical
a particular odor with an electrical shock. Their criti- changes occur during long-term sensitization sug-
cal effect is on the cAMP second-messenger pathway: gests that alterations in the number of sensory neuron
dnc increases cAMP levels, and rut reduces them. Al- varicosities and active zones persist in parallel with
though they have opposite biochemical effects, both the behavioral retention of the memory, whereas ac-
mutations result in larger numbers of spines and vari- tive zone length and vesicle complement do not (Bai-
cosities in a central process of the sensory neuron in- ley and Chen, 1989). Thus, there may be an overlap-
nervating the antero-notopleural thoracic bristle ping cascade of mechanisms that sustain
(Corfas and Dudai, 1991). More frequently, the ef- modifications in synaptic function.
fects of the mutations have been observed at the neu-
romuscular junctions of animals at the larval stage of In contrast with the extensive structural changes
development. In this preparation, the effects of the following long-term training, the morphological cor-
two mutations are different: dnc mutations increase relates of short-term memory in Aplysia (lasting min-
motor neuron branching (Schuster et al., 1996), utes to hours rather than days to weeks) are far less
whereas rut mutations decrease branching (Cheung, pronounced and are primarily restricted to shifts in
Shayan, Boulianne, and Atwood, 1999). the proximity of synaptic vesicles adjacent to sensory
The mollusk Hermissenda has been the subject of neuron active zones, a phenomenon that may reflect
studies of associative conditioning between visual and altered levels of transmitter mobilization (Bailey and
rotational stimuli. A crucial site of plasticity is the Chen, 1988b). These studies in Aplysia suggest a clear
photoreceptor. Following training, the arborization difference in the time course of structural events that
of the medial type B photoreceptor was reduced com- underlie memories of differing durations. The tran-
pared to controls (Alkon et al., 1990). This result is sient duration of short-term memories probably in-
consistent with other biophysical and biochemical volves the covalent modification of preexisting pro-
changes resulting in decreased synaptic strength to teins and is accompanied by modest structural
follower neurons. remodeling in the vicinity of the active zone, such as
In these model systems changes in the complexity the translocation of synaptic vesicles to the release
of the arbor parallels the changes in synaptic efficacy. site.
Neurite outgrowth presumably strengthens the be- Another model system in which synaptic structure
havioral response by forming new synapses to follow- seems correlated with function is the crayfish neuro-
er neurons. Neurite retraction presumably weakens muscular junction. This system does not permit an
the behavioral response by reducing the number of examination of the effects of a specific learning para-
synapses. Because these changes appear to be associ- digm because the cellular plasticity underlying behav-
ated only with long-lasting forms of learning, it seems ioral modifications does not occur at the neuromuscu-
that there is a high threshold for induction. Such lar junction. Instead, attention is focused on the
changes would not seem to admit of easy reversal, but effects of high-frequency stimulation, which produces
no testing has confirmed this hypothesis. both short- and long-term enhancements in synaptic
strength.
Modification of Synaptic Ultrastructure
In the crayfish preparation, a single axon makes
Synaptic strength can be modulated by changes numerous synaptic contacts with the muscle fiber.
in neuronal structure that are more subtle than those These synapses have different functional properties
outlined in the previous section. Usually such changes (Wojtowicz et al., 1994); some are affected by long-
involve the structural modification of preexisting syn- term facilitation, and some are not. Moreover, many
apses. This possibility has been examined in great de- synapses that release transmitter do not contribute to
tail in Aplysia. the EPSP, suggesting that changes in EPSP amplitude
Following long-term sensitization training, the can be effected by modifying only a fraction of syn-
number of sensory neuron varicosities increases as apses.
MORPHOLOGICAL BASIS OF LEARNING AND MEMORY: Invertebrates 403
Early ultrastructural studies supported the idea portant determinants of synaptic strength between
that high-frequency stimulation produced an increase the two different types of motor neurons.
in the percentage of synapses exhibiting presynaptic
active zones. These observations suggest that alter- In an elegant series of experiments at the devel-
ations in neuronal activity can induce rapid structural oping neuromuscular junction of larval Drosophila,
transformations at the synapse, affecting primarily moderate underexpression of a cell adhesion mole-
the active-zone. However, subsequent studies suggest- cule resulted in axonal outgrowth and varicosity for-
ed that the effects of high-frequency stimulation were mation (Schuster et al., 1996). More severe underex-
not persistent. Stimulation at 20 Hz for ten minutes pression resulted in axonal retraction and decreased
produced an increase in the number of synapses per varicosity formation. These morphological changes
unit area of terminal membrane, a decrease in the did not by themselves produce changes in synaptic
synaptic area, and an increase in the number of dense strength, however. Unknown intrinsic mechanisms
bodies per unit of terminal membrane area or syn- maintained synaptic strength at an appropriate level
apse area. Forty-five minutes later, however, the syn- by adjusting the number and size of active zones
apse was still facilitated, but these structural features (Stewart, Schuster, Goodman, and Atwood, 1996).
were no longer different from control. Only one Enhancing synaptic strength required an activation of
structural change was observed following long-term the cAMP second-messenger pathway and phospho-
facilitation: In synapses that had more than one dense rylation of the nuclear regulatory protein CREB.
body, there was an increase in the number of synapses Thus, in addition to the difficulties of demonstrating
per unit area of terminal. the structural mechanisms of learning, there are the
further challenges of demonstrating that those
In Drosophila, mutants that interfere with learning changes contribute to the modulated function of the
also produce changes in the structure of the neuro- neural circuit underlying the behavior.
muscular junction (Renger et al., 2000). For example,
dnc mutations, which increase cAMP levels, produced In recent studies in Aplysia discussed earlier
relatively modest structural changes: the ratio of (Wainwright, Zhang, Byrne, and Cleary, 2002), the
docked to undocked vesicles increased, and pre- structural changes induced by four days of long-term
synaptic and postsynaptic specializations became less sensitization training did not occur after only one day
densely stained. Structural changes following the rut of long-term sensitization training. Moreover, the
mutation, which decreases cAMP levels, were more neurite outgrowth that accompanied four days of
profound. The number of synapses per unit area of training occurred on both sides of the animal, even
terminal decreased, but the size of each synapse in- though the behavioral modification was limited to the
creased. The ratio of docked to undocked vesicles de- trained side. These dissociations do not rule out the
creased, and staining intensity of presynaptic and role of structural changes in learning but caution
postsynaptic specializations was unaffected. If the against the overinterpretion of results in the absence
structural changes at these peripheral synapses reflect of appropriate controls.
similar changes within the central nervous system,
It is difficult indeed to nail down the possible
then they could contribute to the learning deficits.
causal relationship between structural changes and
behavioral changes. Structural changes require a
Causal Relationship Between Structure and large number of cellular processes, such as protein
Function synthesis, cytoskeletal reorganization, organelle
translocation, and membrane-membrane interac-
Notwithstanding the structural changes that seem tions. Therefore, in most model systems there is no
to correlate with learning, the relationship between single drug or manipulation that can block selected
structure and function is not necessarily a simple one. structural changes without affecting other normal cel-
For example, at the crayfish neuromuscular junction, lular processes. Researchers are always seeking sys-
muscle fibers are innervated by both tonic and phasic tems and control experiments that might permit a
motor neurons. Tonic motor neurons are typically more precise determination of the possible causal
highly active, and their terminals typically evince role of structural changes.
large synaptic varicosities with large numbers of vesi-
cles. Phasic motor neurons with low levels of impulse
activity have thinner terminals with fewer vesicles. Bibliography
Aside from the greater number of tonic terminals, Alkon, D. L., Ikeno, H., Dworkin, J., McPhie, D. L., Olds, J. L.,
Lederhendler, I., Matzel, L., Schreurs, B. G., Kuzirian, A.,
there are no marked differences in synaptic structure
Collin, I., and Yamoah, E. (1990). Contraction of neuronal
in the two populations of neurons (Msghina, Govind, branching volume: An anatomical correlate of Pavlovian con-
and Atwood, 1998). Therefore other mechanisms, ditioning. Proceedings of the National Academy of Sciences of the
presumably biophysical or biochemical, are more im- United States of America 87, 1,6111,614.
404 MORPHOLOGICAL BASIS OF LEARNING AND MEMORY: Vertebrates
Bailey, C. H., and Chen, M. (1983). Morphological basis of long- Wainwright, M. L., Zhang, H., Byrne, J. H. and Cleary, L. J. (2002).
term habituation and sensitization in Aplysia. Science 220, Localized neuronal outgrowth induced by long-term sensiti-
9,193. zation training in Aplysia. Journal of Neuroscience 20, 4,132
(1988a). Long-term memory in Aplysia modulates the total 4,141.
number of varicosities of single identified sensory neurons. Walters, E. T., Byrne, J. H., Carew, T. J., and Kandel, E. R. (1983).
Proceedings of the National Academy of Sciences of the United States Mechanoafferent neurons innervating tail of Aplysia. II. Mod-
of America 85, 2,3732,377. ulation by sensitizing stimuli. Journal of Neurophysiology 50,
(1988b). Morphological basis of short-term habituation in 1,5431,559.
Aplysia. Journal of Neuroscience 8, 2,4522,459. Wojtowicz, J. M., Marin, L., and Atwood, H. L. (1994). Activity-
(1989). Time course of structural changes at identified sen- induced changes in synaptic release sites at the crayfish neu-
sory neuron synapses during long-term sensitization in Aply- romuscular junction. Journal of Neuroscience 14, 3,6883,703.
sia. Journal of Neuroscience 9, 1,7741,780. Craig H. Bailey
Bailey, C. H., and Kandel, E. R. (1993). Structural changes accom-
Revised by Len Cleary
panying memory storage. Annual Review of Physiology 55, 397
426.
Balling, A., Technau, G. M., and Heisenberg, M. (1987). Are the
structural changes in adult drosophila mushroom bodies
memory traces? Studies on biochemical mutants. Journal of
VERTEBRATES
Neurogenetics 4, 6573. The central issue in morphology of learning and
Byrne, J. H. (1987). Cellular analysis of associative learning. Physio-
memory is how such constructs are stored in the ner-
Castellucci, V. F., Pinsker, H., Kupfermann, I., and Kandel, E. R. vous system. The basic shape of neurons and synapses
(1970). Neuronal mechanisms of habituation and dishabitua- is illustrated in Figure 1. In the late nineteenth centu-
tion of the gill withdrawal reflex in Aplysia. Science 167, 1,745 ry, Santiago Ramn y Cajal (Spanish histologist,
1,748. 18521934) suggested that learning might involve
Cheung, U. S., Shayan, A. J., Boulianne, G. L., Atwood, H. L. changes in the synaptic connections through which
(1999). Drosophila larval neuromuscular junctions responses
neurons communicate. Such synaptic change could
to reduction of cAMP in the nervous system. Journal of Neuro-
biology 40, 113.
take at least four possible forms. First, forming new
Corfas, G., and Dudai, Y. (1991). Morphology of a sensory neuron synapses or removing existing synapses could alter
in Drosophila is abnormal in memory mutants and changes the pattern of functional connections. Second, selec-
during aging. Proceedings of the National Academy of Sciences of tively strengthening or weakening some synapses
the United States of America 88, 7,2527,256. could alter the pattern of functional connections.
Farris, S. M., Robinson, G. E., and Fahrbach, S. E. (2001). Experi-
Very strong evidence exists for both of these possibili-
ence- and age-related outgrowth of intrinsic neurons in the
mushroom bodies of the adult worker honeybee. Journal of
ties during learning, and in models of learning such
Neuroscience 21, 6,3956,404. as long-term potentiation (LTP). Third, by forming new
Msghina, M., Govind, G. K., and Atwood, H. L. (1998). Synaptic neurons (neurogenesis), neuronal circuits could be
structure and transmitter release in crustacean phasic and modified; there is increasing evidence that neuro-
tonic motor neurons. Journal of Neuroscience 18, 1,3741,382. genesis occurs in the mature brain. Fourth, changes
Ramn y Cajal, S. (1988). Anatomicophysiological considerations
could occur in both the nonsynaptic regions of neu-
on the cerebrum. In J. DeFelipe and E. G. Gros, eds., Cajal
on the cerebral cortex. New York: Oxford University Press. rons and in nonneural elements of the brain such as
Renger, J. J., Ueda, A., Atwood, H. L., Govind, C. K., and Wu, C.-F. glial cells and the vascular system subserving the
(1996). Genetic dissection of structural and functional compo- brain, for which evidence exists as well.
nents of synaptic plasticity. II. Fasciclin II controls presynap-
tic structural plasticity. Neuron 17, 655667.
(2000). Role of cAMP cascade in synaptic stability and plas- Changes in Synapse Number
ticity: Ultrastructural and physiological analyses of individual
synaptic boutons in drosophila memory mutants. Journal of
Important roots of memory research lie in studies
Neuroscience 20, 3,9803,992. of the effects of experience upon brain development.
Schuster, C. M., Davis, G. W., Fetter, R. D., and Goodman, C. S. For example, visual experience is necessary to devel-
(1996). Genetic dissection of structural and functional compo- op normal visual ability in mammals. Searching for a
nents of synaptic plasticity. I. Fasciclin II controls synaptic sta- basis for this in brain anatomy, Cragg (1975) and oth-
bilization and growth. Neuron 17, 641654.
ers, noted that animals deprived of visual experience
Sigg, D., Thompson, C. M., Mercer, A. R. (1997). Activity-
dependent changes to the brain and behavior of the honey
had fewer synaptic connections per nerve cell in the
bee, Apis mellifera (L.). Journal of Neuroscience 17, 7,1487,156. visual cortex. These studies profoundly influenced
Stewart, B. A., Schuster, C. M., Goodman, C. S. and Atwood, H. L. thinking about the processes by which the brain stores
(1996). Homeostasis of synaptic transmission in Drosophila information, because they showed that 1. brain struc-
with genetically altered nerve terminal morphology. Journal ture is malleable; 2. synaptic organization can be or-
chestrated into different configurations by behavioral
Sutton, M. A. and Carew, T. J. (2000). Parallel molecular pathways
mediate expression of distinct forms of intermediate-term fa- experience; 3. both forming new connections and
cilitation at tail sensory-motor synapses in Aplysia. Neuron 26, pruning existing connections are involved in altering
219231. brain organization; and 4. differential experience can
MORPHOLOGICAL BASIS OF LEARNING AND MEMORY: Vertebrates 405
modify the structure of synapses, suggesting that syn- findings, reporting that visual cortical neurons of rats
aptic efficacy (or strength) also can change. These ef- reared in enriched environments had larger dendritic
fects of experience on synaptic connections during fields than did those of cage-housed controls. Den-
development led to proposals that such changes drites of neurons receive the bulk of the synaptic
might underlie adult learning as well. input (see Figure 1), so the implication was that new
A separate developmental approach that was very synapses formed. Similar findings were subsequently
fruitful in understanding brain substrates of learning reported in other areas of the cerebral cortex and in
and memory involved enriching the lives of young an- brain regions such as the hippocampus, superior col-
imals with additional stimulation. Donald Hebb (psy- liculus, and cerebellum. The enriched environment
chobiologist, 19041985) proposed ways in which changed brain anatomy in adult rats, which is impor-
synaptic change could be incorporated meaningfully tant evidence for involvement in learning and memo-
into functional circuitry. With his students, he showed ry. Turner and Greenough (1985) found that rats
that enriching the rearing environment of rats with reared in enriched environments had more synapses
cagemates and toys improved the animals ability to per neuron in the visual cortex, compared with rats
solve complex problems. Hebb concluded that behav- reared alone or in pairs in standard laboratory cages.
ior, and by implication brain organization, was per- Moreover, similar changes occur in the striatum as
manently altered by this early experience. Subse- well (Comery, Shah, and Greenough, 1995), suggest-
quently, Rosenzweig, Bennett, and Diamond (1972) ing that the experience-dependent changes in neuro-
found that regions of the cerebral cortex were thicker nal morphology influence multiple levels or systems
and heavier in rats reared in enriched environments, in the brain. The general conclusion from these en-
compared with rats reared in solitary or group cages. riched environment studies is that when animals are
Volkmar and Greenough (1972) followed up these placed in an environment in which they store infor-
mation that affects later behavior, they form new syn- number changes in invertebrate plasticity paradigms
apses. (The term enriched is used in comparison with also exists.
the typical laboratory environment and does not Neurogenesis presents yet another mechanism
imply superiority to the natural environment.) whereby brain organization could be modified in re-
Follow-up studies have explored the effects of sponse to experience. For example, increased num-
specific learning tasks upon these same measures. bers of new neurons have been reported in the den-
tate gyrus of the hippocampus of animals exposed to
There is compelling evidence that many forms of
a complex environment (Kempermann, Kuhn, and
learning change both the amount of dendrite per
Gage, 1998) or permitted access to an exercise wheel
neuron and the number of synapses per neuron. For
(Van Praag, Christie, Sejnowski, and Gage, 1999).
example, dendritic branching is increased in visual
More specific to learning and memory, Gould et al.
cortical neurons following twenty-five days of expo-
(1999a) reported that the number of new neurons in
sure to a series of maze problems (Greenough,
rat dentate gyrus dramatically increased following as-
Juraska, and Volkmar, 1979). Subsequent work used
sociative conditioning. While most reports of neuro-
split-brain rats, severing the nerve fibers that allow
genesis have been limited to rodents and to relatively
the right and left hemispheres to communicate, and
unique areas in the brain such as the hippocampus or
opaque contact lenses that restricted visual input
olfactory bulb, neurogenesis has been reported in the
from training to one eye. Neurons on the trained side primate cortex (Gould, Reeves, Graziano, and Gross,
of the brain selectively exhibited dendritic field size 1999b) and disputed (Kornack and Rakic, 2001).
increases; thus these changes were not of the general While the issues of the extent, location, and function-
sort that might be due to stress or arousal associated al relevance of neurogenesis has yet to be resolved,
with the task, which should affect both sides of the the mere possibility of such a mechanism in the ma-
brain equally. These studies, and others, indicated ture central nervous system has spurred a great deal
that the altered dendritic fields were associated with of excitement and hope for many areas of neural re-
neural input and output related to the training. search.
Synaptogenesis is also implicated in associative
learning. Tsukahara (1981) investigated associative Role of Synapse Formation in Plastic
limb flexion conditioning, using stimulation to the ce- Neural Change
rebral peduncle as the conditioned stimulus and elec-
An issue that affects all of these studies is whether
tric shock to the forelimb as the unconditioned stimu-
anatomical changes that are seen following training
lus. In this paradigm, red nucleus lesions abolish the
result merely from increased neural activity involved
conditioned response, implicating the involvement of
in performing the learned task. Muscles grow larger
this structure. Electrophysiological studies following
in response to exercise; perhaps neurons do too, such
conditioning indicated enhanced input to the red nu-
that these structural changes have nothing to do with
cleus from the cerebral cortex. Subsequently, Tsuka-
learning or memory per se. (This issue is, of course,
haras coworkers (Murakami, Higashi, Katsumaru,
not unique to morphological studies; proposed mo-
and Oda, 1987) reported morphological evidence for
lecular and other aspects of the cellular mechanisms
formation of new corticorubral synapses in condi-
of memory may similarly be artifacts of activity.)
tioned animals.
There have been direct tests of the effects of neu-
Similar anatomical effects of training have been ral activity versus learning on synapse change. For ex-
observed in other behavioral tasks. Stewart (1991) example, Black et al. (1990) compared a cerebellar cor-
amined day-old chicks that learned to avoid pecking tical region in rats that had learned a complex series
a bad tasting food particle and found an elevated of motor tasks to that of rats that performed one of
number of synapses in a forebrain region previously two forms of physical exertion involving little learn-
shown to be involved in the learning. In another ining: running on a treadmill or in an activity wheel.
volved brain region there were also increases in the Exercise alone had no effect on synapse number
number of spines (see Figure l), the dendritic compo- whereas rats that had learned increased the number
nent of one type of synapse. Likewise, the density of of synapses per neuron in the cerebellar cortex. Simi-
spines was reported to be elevated in hippocampal lar effects have been observed in areas such as the
area CA1 in adult rats following spatial learning motor cortex (Kleim et al., 1996). In contrast, blood
(Moser, Trommald, and Andersen, 1994). Compara- vessel density was elevated in affected regions in rats
ble changes have been observed in numerous other that had exercised, whereas the motor learning rats
paradigms (e.g., bird song learning and imprinting in had the same blood vessel density as control animals.
birds). Finally, while this discussion is confined to ver- These results indicate that activity and learning have
tebrates, excellent evidence for comparable synaptic very different effects on brain tissue.
MORPHOLOGICAL BASIS OF LEARNING AND MEMORY: Vertebrates 407
Additional support for the role of synapse forma- changes in synapse strength. Synaptic vesicle num-
tion in plastic neural change has come from studies bers have been reported to decrease with visual depri-
of long-term potentiation. LTP involves an increase vation (e.g., Tieman, 1985) and there are reports of
in the response of postsynaptic neurons following both decreased vesicle density and altered vesicle lo-
high frequency bursts of presynaptic firing. LTP in- cation within the presynaptic terminal following LTP
duction has been shown to increase spine density and induction (e.g., Applegate, Kerr, and Landfield,
modify dendritic morphology in the motor cortex 1987; Fifkova and Van Harreveld, 1977). In contrast,
(Ivanco, Racine, and Kolb, 2000) and in hippocampal vesicle numbers have been shown to increase in rats
subfield CA1 (Lee, Oliver, Schottler, and Lynch, reared in enriched environments (Nakamura, Ko-
1981). Chang and Greenough (1984) included a type bayashi, Ohashi, and Ando, 1999; Sirevaag and Gree-
of high frequency stimulation that did not induce nough, 1991).
LTP, and subsequently did not alter synapse forma- At least three other synapse features appear to be
tion, suggesting that LTP-induced synaptogenesis sensitive to experience. First, small discontinuities in
was not caused by high frequency stimulation alone. the postsynaptic density, termed perforations, have
In contrast to CA1, induction of LTP in the hippo- been found to increase in number following complex
campal dentate gyrus does not cause synapse forma- environment exposure and to decrease in affected
tion but changes synapse structure (Geinisman, de synapses subsequent to sensory deprivation (Gree-
Toledo-Morrell, and Morrell, 1991). Recent advances nough, West, and DeVoogd, 1978). Moreover, Vren-
in microscopy have permitted near real time evalua- sen and Cardozo (1981) found that the number of
tion of changes in morphology of living neurons. perforated synapses increased in the visual cortex fol-
Using these tools, Engert and Bonhoeffer (1999) re- lowing visual discrimination learning. The function of
ported that postsynaptic spines are formed de novo these perforations is unknown. Second, the incidence
in response to stimulation. Taken together, these of multiple synaptic boutons (presynaptic elements that
findings reinforce the view that numerous different synapse with multiple postsynaptic components) is el-
cellular changes may be involved in learning, and evated in numerous brain areas following exposure to
other forms of neural plasticity as well. a complex environment and in animals that have
been trained in an associative learning paradigm
Changes in Synapse Structure: Indications (Geinisman et al., 2001; Jones et al., 1997). Third, the
of Synapse Efficacy Change cellular organelles that synthesize protein, polyriboso-
mal aggregates, are frequently found in the heads and
Several structural features of synapses have been necks of spines during periods of synapse formation
found to be altered by behavioral experience. One of (Steward and Falk, 1985). They are also found more
the most obvious features is the size of synapses. Larg- frequently in spines of animals in complex environ-
er synapses may release more neurotransmitter or ments, possibly reflecting greater rates of synapse for-
have more receptors, such that a size change could in- mation. Protein synthesis at the synapse has been pro-
dicate a strength change. Early findings indicated posed as a memory mechanism (Steward and
smaller synapses in the visual cortex of animals visual- Schuman, 2001).
ly deprived during development, and Tieman (1985)
reported smaller geniculocortical synapses in
monocularly deprived cats. Conversely, synapse size Changes in Nonneural Elements
increased following imprinting in day-old chicks, and The enriched environment work indicated from
similar size changes were found after avoidance learn- its earliest days that morphological changes were not
ing. Larger synapses were also observed in layer IV of restricted to neurons. Glial cells, supportive elements
the visual cortex of rats reared in enriched environ- that maintain ionic, metabolic, and neurotransmitter
ments, compared with individually caged controls. homeostasis, respond similarly to environmental
Likewise, Van Harreveld and Fifkova (1975) de- complexity. Sirevaag and Greenough (1991) reported
scribed changes in the size of synaptic spine heads that astrocytes grow larger and extend additional
and necks (see Figure 1) following LTP induction in processes into the tissue during the first phase of their
the dentate gyrus. Koch and Zador (1993) have sug- response to the animals housing in an enriched envi-
gested that larger spine components, and the associ- ronment. In a second phase, astrocytes divide, in-
ated spine neck restriction, may permit activation of creasing their numbers, and shrink, on average, to-
focalized intracellular cascades and reduce electrical ward their preexposure size. These stages are
resistance, thereby facilitating the passage of synaptic qualitatively comparable with those of gliosis, the glial
current into the dendrite (Harris and Kater, 1994). reaction to injury, yet protracted. Moreover, Ander-
Synaptic vesicles are believed to contain neurotran- son et al. (1994) have shown an increase in the volume
smitters, and changes in their numbers could indicate of glia per Purkinje cell in the cerebellum of animals
that learned a complex motor skill, but not those who Engert, F., and Bonhoeffer, T. (1999). Dendritic spine changes as-
simply exercised. Likewise, blood vessel density in- sociated with hippocampal long-term synaptic plasticity. Na-
ture 399, 6670.
creased in rats placed in an enriched environment at
Fifkova, E., and Van Harreveld, A. (1977). Long-lasting morpho-
the age of weaning. In animals that are older at the logical changes in dendritic spines of dentate granular cells
time they are first exposed to enrichment, this blood following stimulation of the entorhinal area. Journal of Neuro-
vessel response diminishes with increasing age. cytology 6, 211230.
Geinisman, Y., Berry, R. W., Disterhoft, J. F., Power, J. M., and Van
der Zee, E. A. (2001). Associative learning elicits the forma-
Conclusion tion of multiple-synapse boutons. Journal of Neuroscience 21,
5,5685,573.
Morphological research has provided strong evi- Geinisman, Y., de Toledo-Morrell, L., and Morrell, F. (1991). In-
dence for both forms of synaptic change that have duction of long-term potentiation is associated with an in-
been proposed to underlie learning and memory. crease in the number of axospinous synapses with segmented
Formation, and occasionally loss, of synapses occurs postsynaptic densities. Brain Research 566, 7788.
both during periods of development when the brain Gould, E., Beylin, A., Tanapat, P., Reeves, A., and Shors, T. J.
(1999a). Learning enhances adult neurogenesis in the hippo-
is storing information and during exposure to specific campal formation. Nature Neuroscience 2, 260265.
learning tasks. Various control procedures have Gould, E., Reeves, A. J., Graziano, M. S., and Gross, C. G. (1999b).
largely ruled out the possibility that these synaptic Neurogenesis in the neocortex of adult primates. Science 286,
changes are artifactual results arising from factors 548552.
other than learning. Changes in the structure of syn- Greenough, W. T., Juraska, J. M., and Volkmar, F. R. (1979). Maze
apses, such as in the size or shape of synaptic compo- training effects on dendritic branching in occipital cortex of
adult rats. Behavioral and Neural Biology 26, 287297.
nents, also occur during learning and in other situa- Greenough, W. T., West, R. W., and DeVoogd, T. J. (1978). Sub-
tions in which functional brain organization is synaptic plate perforations: Changes with age and experience
altered, such as LTP. Many of these structural in the rat. Science 202, 1,0961,098.
changes have been associated with synapse strength Harris, K. M., and Kater, S. B. (1994). Dendritic spines: Cellular
differences in other research. Developing research specializations imparting both stability and flexibility to syn-
implicates addition of new neurons via neurogenesis aptic function. Annual Review of Neuroscience 17, 341371.
Ivanco, T. L., Racine, R. J., and Kolb, B. (2000). Morphology of
and changes in nonneuronal elements of the brain. layer III pyramidal neurons is altered following induction of
Thus the weight of the evidence indicates that both LTP in sensorimotor cortex of the freely moving rat. Synapse
synapse formation/removal and synapse strength 37, 1622.
changes are involved in learning and memory, but Jones, T. A., Klintsova, A. Y., Kilman, V. L., Sirevaag, A. M., and
additional mechanisms are also likely. Greenough, W. T. (1997). Induction of multiple synapses by
experience in the visual cortex of adult rats. Neurobiology of
See also: MORPHOLOGICAL BASIS OF LEARNING AND Learning and Memory 68, 1320.
MEMORY: INVERTEBRATES; PROTEIN SYNTHESIS Kempermann, G., Kuhn, H. G., and Gage, F. H. (1998). Experi-
IN LONG-TERM MEMORY IN VERTEBRATES ence-induced neurogenesis in the senescent dentate gyrus.
Journal of Neuroscience 18, 3,2063,212.
Bibilography Kleim, J. A., Lussnig, E., Schwarz, E. R., Comery, T. A., and Gree-
nough, W. T. (1996). Synaptogenesis and FOS expression in
Andersen, B. J., Li, X., Alcantara, A. A., Isaacs, K. R., Black, J. E.,
the motor cortex of the adult rat after motor skill learning.
and Greenough, W. T. (1994). Glial hypertrophy is associated
with synaptogenesis following motor-skill learning, but not
with angiogenesis following exercise. Glia 11,7380. Koch, C., and Zador, A. (1993). The function of dendritic spines:
Applegate, M. D., Kerr, D. S., and Landfield, P. W. (1987). Redis- Devices subserving biochemical rather than electrical com-
tribution of synaptic vesicles during long-term potentiation in partmentalization. Journal of Neuroscience 13, 413422.
the hippocampus. Brain Research 401, 401406. Kornack, D. R., and Rakic, P. (2001). Cell proliferation without
Black, J. E., Isaacs, K. R., Anderson, B. J., Alcantara, A. A., and neurogenesis in adult primate neocortex. Science 294, 2,127
Greenough, W. T. (1990). Learning causes synaptogenesis, 2,130.
whereas motor activity causes angiogenesis, in cerebellar cor- Lee, K. S., Oliver, M., Schottler, F., and Lynch, G. (1981). Electron
tex of adult rats. Proceedings of the National Academy of Sciences microscopic studies of brain slices: The effects of high fre-
of the United States of America 87, 5,5685,572. quency stimulation on dendritic ultrastructure. In G. A.
Chang, F.-L. F., and Greenough, W. T. (1984). Transient and en- Kerkut, and H. V. Wheal, eds. Electrophysiology of isolated mam-
during morphological correlates of synaptic activity and effi- malian CNS preparations. New York: Academic Press.
cacy change in the rat hippocampal slice. Brain Research 309, Moser, M.-B., Trommald, M., and Andersen, P. (1994). An in-
3546. crease in dendritic spine density on hippocampal CA1 pyra-
Comery, T. A., Shah, R., and Greenough, W. T. (1995). Differential midal cells following spatial learning in adult rats suggests the
rearing alters spine density on medium-sized spiny neurons formation of new synapses. Proceedings of the National Academy
in the rat corpus striatum: Evidence for the association of of Sciences of the United States of America 91, 12,67312,675.
morphological plasticity with early response gene expression. Murakami, F., Higashi, S., Katsumaru, H., and Oda, Y. (1987). For-
Neurobiology of Learning and Memory 63, 217219. mation of new corticorubral synapses as a mechanism for clas-
Cragg, B. G. (1975). The development of synapses in kitten visual sical conditioning in the cat. Brain Research 437, 379382.
cortex during visual deprivation. Experimental Neurology 46, Nakamura H., Kobayashi, S., Ohashi, Y., and Ando, S. (1999). Age-
445451. changes of brain synapses and synaptic plasticity in response
MOTOR SKILL LEARNING 409
to an enriched environment. Journal of Neuroscience Research architecture of the brain (Houk, 2001). Definitions of
56, 307315. motor skills typically pertain to the movements of the
Rosenzweig, M. R., Bennett, E. L., and Diamond, M. C. (1972). limbs and torso as opposed to those of perceptions
Chemical and anatomical plasticity of brain: Replications and
extensions. In J. Gaito, ed. Macromolecules and behavior, 2nd and the formulation of ideas, but the conceptual
edition. New York: Appleton-Century-Crofts. boundaries blur in the face of the planning that pre-
Sirevaag, A. M., and Greenough, W. T. (1991). Plasticity of GFAP- cedes elaborate motor acts.
immunoreactive astrocyte size and number in visual cortex of
rats reared in complex environments. Brain Research 540, How do people learn and remember how to
273278. dance, type, hop, play the piano, and tie their shoe-
Steward, O., and Falk, P. M. (1985). Polyribosomes under develop- laces? Bartlett (1932) said of the skilled tennis player,
ing spine synapses: Growth specializations of dendrites at sites When I make the stroke I do not, as a matter of fact,
of synaptogenesis. Journal of Neuroscience Research 13, 7588.
produce something absolutely new, and I never mere-
Steward O., and Schuman E. M. (2001). Protein synthesis at synap-
tic sites on dendrites. Annual Review of Neuroscience 24, 299 ly repeat something old. A central issue in the learn-
325. ing of motor skills is how the movement form is ac-
Stewart, M. G. (1991). Changes in dendritic and synaptic structure quired through practice and retained over time. A
in chick forebrain consequent on passive avoidance learning. related issue is the role that variations of movement
In R. J. Andrew, ed. Neural and behavioral plasticity. London:
form play in realizing the goal of the act. These two
Oxford University Press.
Tieman, S. B. (1985). The anatomy of geniculocortical connections issues, movement invariance and motor equivalence,
in monocularly deprived cats. Cellular and Molecular Neuro- have been the focus of the theorizing about the acqui-
biology 5, 3545. sition and retention of motor skills.
Tsukahara, N. (1981). Sprouting and the neuronal basis of learn-
ing. Trends in Neurosciences 4, 234237. Motor skills involve two distinctive operations:
Turner, A. M., and Greenough, W. T. (1985). Differential rearing One is to select, recall, and initiate the movement seg-
effects on rat visual cortex synapses. I. Synaptic and neuronal ment required at each stage in a task; the other is to
density and synapses per neuron. Brain Research 329, 195 guide the trajectories of the movement segments so
203.
that they achieve the subgoals required to complete
Van Harreveld, A., and Fifkova, E. (1975). Swelling of dendritic
spines in the fascia dentata after stimulation of the perforant the task. Much of the work on skill learning and re-
fibers as a mechanism of post-tetanic potentiation. Experimen- tention has emphasized the second operation, the re-
tal Neurology 49, 736749. finement of trajectories based on the experience
Van Praag, H., Christie, B. R., Sejnowski, T. J., and Gage, F. H. gained through the sensory consequences of the
(1999). Running enhances neurogenesis, learning, and long-
movement (Adams, 1971).
term potentiation in mice. Proceedings of the National Academy
of Sciences of the United States of America 96, 13,42713,431. The emerging conceptions of cognitive psycholo-
Volkmar, F. R., and Greenough, W. T. (1972). Rearing complexity gy challenged the one-to-one memory accounts of
affects branching of dendrites in the visual cortex of the rat.
movement representation. An outgrowth of this trend
Science 176, 1,4451,447.
Vrensen, G., and Cardozo, J. N. (1981). Changes in size and shape was Schmidts (1975) schema theory of motor learn-
of synaptic connections after visual training: An ultrastructur- ing, which promoted a one-to-many representational
al approach of synaptic plasticity. Brain Research 218, 7997. construct for both recall and recognition processes of
William T. Greenough movement control. The representation for each
Revised by William T. Greenough and James D. Churchill memory state consisted of the relations between task,
organism, and environmental variables rather than
the absolute levels of the variables themselves. The
schema was a generic rule for a given class of move-
ments that allowed the generalization of movement
MOTOR SKILL LEARNING outcome to a variety of task and environmental cir-
A variety of motor skills occur in various forms of cumstances. Schema theory proposed that the more
movement: work, play, sport, communication, dance, variable the practice within the potential class of
and so on. Psychophysical studies of the learning and movements (e.g., variations in the length, velocity,
retention of motor skills date from the 1890s, with and/or angle of a forehand drive in tennis), the more
neurophysiological studies coming later. Attempts to general the schema rule would become for that activi-
combine cognitive and neural approaches flourished ty. Neurobiologically based theories of movement
in the twentieth century (Bernstein, 1967) and persist control seem to be able to account for most of this
unabated, capitalizing on advances in technology. framework (Bertier et al., 1993).
The theoretical and operational emphases of this The schema theory seemed to provide a solution
field parallel those in other subdomains of learning, to two enduring problems in motor-skill acquisition
in part because motor, perceptual, and cognitive skills and retention: novelty and the limited storage capaci-
are not mutually exclusive and in part because of ana- ty. The novelty problem addresses the question of
tomical advances that show the underlying modular how the performer accommodates to novel tasks and
410 MOTOR SKILL LEARNING
environmental circumstances. The limited storage ca- are shedding light on the changes in the motor cortex
pacity of the CNS arises as a consequence of the many that occur with practice (Classen et al., 1998). Func-
one-to-one representations that would be generated tional imaging of the brain is helping to define the
from an individuals lifetime movement experience, networks that participate in motor programming (van
especially in the absence of schema theory. The Mier et al., 1998).
Schmidt schema theory could not resolve the novelty
problem because it did not explain the initial estab- Conclusion
lishment of the movement class; it accounted only for
changes in the scaling of force, velocity, or position Traditional theories of motor learning and reten-
of a given action pattern (such as a tennis forehand tion fail to capture many of the qualities of the pro-
drive) rather than the generation of pattern of the gression from novice to expert in skill acquisition.
forehand drive movementperhaps because it did The development of a skill is a continuous explorato-
not interface with the emerging knowledge about the ry activity, not a replica of a static representation of
basal ganglia (Houk and Wise, 1995). action. Neurobiological cognitive models, based on
new techniques for studying brain activation, com-
During the 1980s one-to-one and one-to-many bined with the ecological approach to perception and
prescriptive accounts of motor-skill learning were action, are beginning to capture some of the impor-
challenged by the tenets of the ecological approach tant qualities, both invariant and changing, of the dy-
to perception and action (Kugler and Turvey, 1987). namics of motorskill acquisition and retention.
The ecological approach seeks the solution to motor
learning through the mapping of perception and ac- Bibliography
tion with minimal appeal to the representational pro- Adams, J. A. (1971). A closed-loop theory of motor learning. Jour-
cesses typically posited by cognitive psychologists. A nal of Motor Behavior 3, 111150.
Bartlett, F. C. (1932). Remembering: A study of experimental and social
central concern has been the appropriateness of cog-
psychology.Cambridge: Cambridge University Press.
nitive strategies proposed to map the emergent Barto, A. G., Fagg, A. H., Sitkoff, N., and Houk, J. C. (1999). A cer-
movement dynamics into a rule-based symbolic rep- ebellar model of timing and prediction in the control of
resentation. reaching. Neural Computation 11, 565594.
Bernstein, N. A. (1967). The coordination and regulation of movements.
The emergent structure and variability of the London: Pergamon.
movement sequence was subsequently analyzed in Berthier, N. E., Singh, S. P., Barto, A. G., and Houk, J. C. (1993).
terms of physical-systems solutions to the mapping of Distributed representation of limb motor programs in arrays
of adjustable pattern generators. Journal of Cognitive Neuro-
the gradient and equilibrium regions of the perceptu-
science 5, 5678.
al and motor processes (Kugler and Turvey, 1987; Sc- Classen, J., Liepert, J., Wise, S. P., Hallett, M., and Cohen, L. G.
honer and Kelso, 1988). A physical approach to the (1998). Rapid plasticity of human cortical movement repre-
study of the learning and retention of motor skills sentation induced by practice. Journal of Neurophysiology 79,
must contend with the question of how information 1,1171,123.
Fowler, C. A., and Turvey, M. T. (1978). Skill acquisition: An event
and dynamics relate to the intention of the perform-
approach with special reference to searching for the optimum
er. Learning can be viewed as an exploratory activity, of a function of several variables. In G. E. Stelmach, ed., Infor-
with the performer searching for stable regions of the mation processing and motor control. New York: Academic Press.
perceptual and motor dynamics that realize the goal Houk, J. (2001). Neurophysiology of frontal-subcortical loops. In
of the act (Newell et al., 1989). D. G. Lichter and J. L. Cummings, eds., Frontal-subcortical cir-
cuits in psychiatry and neurology. New York: Guilford Publica-
The 1990s saw a rising interest in understanding tions.
how the networks of the brain might learn to generate Houk, J. C., and Wise, S. P. (1995). Distributed modular architec-
tures linking basal ganglia, cerebellum, and cerebral cortex:
motor command signals capable of controlling skilled
Their role in planning and controlling action. Cerebral Cortex
movements. Neural-network models of the cerebel- 5, 95110.
lum based on the anatomy and physiology of these Hua, S. E., and Houk, J. C. (1997). Cerebellar guidance of premo-
circuits helped to relate skill acquisition and perfor- tor network development and sensorimotor learning. Learn-
mance to many fundamental features of motor per- ing and Memory 4, 6376.
Kugler, P. N., and Turvey, M. T. (1987). Information, natural law,
formance, such as the one-to-many representational
and the self-assembly of rhythmic movement. Hillsdale, NJ: Erl-
construct for motor programs (Berthier et al., 1993) baum.
and the predictive capacity required to prevent insta- Newell, K. M., Kugler, P. N., van Emmerik, R. E. A., and McDon-
bility caused by closed-loop control (Barto et al., ald, P. V. (1989). Search strategies and the acquisition of coor-
1999). Models of the interaction between the cerebel- dination. In S. A. Wallace, ed., Perspectives on the coordination
of movement. Amsterdam: North-Holland.
lum and the cerebral cortex suggested the manner in
Schmidt, R.A. (1975). A schema theory of discrete motor skill
which memories might be translocated to improve learning. Psychological Review 82, 225260.
the speed and automaticity of practiced skills. New Schoner, G., and Kelso, J. A. S. (1988). A dynamic theory of behav-
techniques such as transcranial magnetic stimulation ioral change. Journal of Theoretical Biology 135, 501524.
MLLER, GEORG ELIAS 411
van Mier, H., Tempel, L. W., Perlmutter, J. S., Raichle, M. E., and
Petersen, S. E. (1998). Changes in brain activity during motor
learning measured with PET: Effects of hand of performance
and practice. Journal of Neurophysiology 80, 2,1772,199.
Karl M. Newell
Revised by James C. Houk
MLLER, GEORG ELIAS (18501934)

During his tenure at the University of Gttingen from
1881 to 1921, the German psychologist Georg Elias
Mller helped to spearhead major advances in theory
and research into perception, learning, and memory.
Early Life and Career

Georg Elias Mller was born on July 20, 1850,
into a clerical family in Grimma, Saxony, Germany.
As a schoolboy he showed a precocious interest in nat-
ural science, philosophy, poetry, and history. In 1868
he began to study philosophy and history at the Uni-
versity of Leipzig and moved to the University of Ber-
lin in 1869. In Berlin he became acquainted with Ru-
dolph Hermann Lotzes writings, which shifted his Georg Elias Mller (Georg-Elias-Mller Institute of Psychology)
focus from history to science as the subject closest to
his principal interest, philosophy. After interrupting
his studies to fight in the Franco-Prussian war of Main Work
18701871, he moved to Gttingen in 1872 to study
with Lotze. The following year he submitted his doc- Mllers main contributions to experimental psy-
toral dissertation on sensory attention (Mller, 1873) chology fall into three areas: psychophysics, visual
and he completed his postdoctoral dissertation on perception, and memory.
psychophysics (Mller, 1878) in 1876. Five years later Psychophysics
he succeeded Lotze as chair of philosophy in Gtting- Working from the philosophical background of
en, a position he held for forty years, until his manda- the mind-body problem, which energized his re-
tory retirement in 1921. search all his life, Mller first directed his attention
By then Mllers institute had become one of the to psychophysics, the scientific discipline that spear-
most important centers for experimental psychology headed the establishment of psychology as an objec-
in all of Europe, attracting students from all over the tive science. The Weber-Fechner law, which describes
world. He also opened his laboratory to women at a the relationship between physical and perceived stim-
time when research opportunities for women were ulus intensity, had appeared in Gustav Fechners
nearly nonexistent. Among his students were Alfred groundbreaking work Elements of Psychophysics in
Binet, E. R. Jaensch, Adolf Jost, David Katz, Oswald 1860. While the idea of a lawful relationship between
Kroh, Oswald Klpe, Lillian J. Martin, Eleanor Mc- physical objects and perceived stimuli had been wide-
Gamble, Alfons Pilzecker, Gza Rvsz, Edgar Rubin, ly accepted, the experimental methods for obtaining
and Friedrich Schumann. Mller, who was often de- measurements and the possibility of a mathematically
scribed as a somewhat gruff character, was feared by precise treatment of subjective perceptions were the
his students as a relentless critic and a stickler for sci- subject of extended scientific debate. Mllers post-
entific rigor, but he was also cherished as a skillful lis- doctoral dissertation, On the Foundations of Psycho-
tener and as an unfaltering supporter of their work. physics (1878), and his paper on the method of right
His meticulous, diligent, and innovative experimen- and wrong cases (later known as the method of cons-
tal approach earned his laboratory the reputation as tant stimuli) established Mller as an independent,
the premier center for experimental psychology in all critical thinker. His contributions were acknowledged
of Europe. Mllers pioneering vision also led to the in Fechners Revision of the Main Points of Psycho-
formation of the Society for Experimental Psycholo- physics. His methods still are widely used in psycho-
gy, which he headed from 1904 to 1927. physical experiments.
412 MLLER, GEORG ELIAS
Visual Perception complete mastery, or Erlernungsmethode). They then

Visual perception, especially color perception, explored the conditions under which previously
occupied Mller in the late 1890s and led to his au- learned material facilitated the memorization of reor-
thoring four publications on this topic between 1896 ganized syllable lists (the method of savings, or Er-
and 1897. In these papers he argued that the three sparnismethode). Using these methods, they deter-
photochemical substances proposed by Ewald Hering mined that reading the pairs of syllables in trochaic
as the basis of color vision were reversible by a chemi- style resulted in stronger associations between the syl-
cal rather than a metabolic process. Moreover, he lables within the pair than between the adjacent sylla-
proposed that equal excitation by red/green, blue/ bles of two pairs. They also explored a number of con-
yellow, and black/white resulted in the perception of ditions (e.g., retention delay) that either facilitated or
cortical gray rather than the visual silence as- weakened acquisition of the study material.
sumed by Hering. Mller returned to color percep- Mller and Pilzecker significantly improved on
tion after his retirement and remained with it until his these early experiments by replacing the method of
final years, even at the expense of completing his au- complete mastery with the method of right associates
tobiography. Instead, he summarized his observa- (Treffermethode), shifting the focus from acquisition to
tions and theories in two books: On Sensations of Color: recall of learned material. In these experiments the
Psychophysical Investigations (1930) and Brief Contribu- number of learning trials for each list of syllable pairs
tions to the Psychophysics of Color Sensations (1934). was fixed. Then the subjects were cued with the first
Memory syllable of each pair and asked to recall the second syl-
lable of each pair. This method, developed by Adolph
Mllers work on memory is probably his best Jost in Mllers laboratory, provided three ways to as-
known. He inherited the topic from Hermann Eb-
sess memory performance: correct answers, incorrect
binghaus, whose monograph On Memory (1885) first
answers, and failures to recall. Moreover, the data
demonstrated that even cognitive phenomena such as
could be supplemented by the subjects response
verbal memory could be successfully studied and
times as measured by a novel apparatus. The shutter
quantified in the laboratory. When Ebbinghaus
in front of the memory drum was now operated by an
showed no further interest in memory research, Ml-
electromagnet. Opening the shutter exposed the cue
ler seized the opportunity. During the next thirty
syllable for each trial and triggered a chronometer.
years Mller and his students made a number of semi-
The subjects response was registered by a lip key or
nal contributions. The most notable were Experi-
vibration-sensitive device that would stop the chro-
mental Contributions to the Investigation of Memo-
nometer, close the shutter, and advance the memory
ry (1894) and Experimental Contributions to the
drum to the next cue syllable. Armed with this meth-
Theory of Learning and Memory (1900). Both arti-
odology, Mller and Pilzecker began a series of ex-
cles describe sophisticated experimental methods for
periments that quickly exceeded the scope of Eb-
studying verbal memory, which were developed by
binghauss original studies. After studying the effects
Mller and his students Friedrich Schumann, Adolf
of list repetition and retention interval on cued recall,
Jost, and Alfons Pilzecker and became standard in ex-
they explored interference between associations.
perimental psychology for many decades.
They first presented a pair of syllables (e.g., ser-lad)
Mllers first contribution was to recognize the and in a later list used the same cue syllable (ser) in
importance of constructing lists of syllables that were another association (e.g., ser-kum). Recall cued by am-
equally difficulty to read and learn. To this end he biguous syllables was consistently lower than recall
tightened Ebbinghauss rules for generating and se- cued by unambiguous syllables.
lecting nonsense syllables. Mllers second contribu-
These findings became fundamental to learning
tion was to control the presentation of study material.
and memory research. Yet Mllers monograph with
Lists of syllables were mounted on a rotating cylinder,
Pilzecker is probably best known for its novel propos-
called a memory drum, and syllables were presented
al that learning does not induce instantaneous, per-
one at a time through a shutter behind which the
manent memories but that memory requires an inter-
memory drum was mounted. The subjects were in-
val of consolidation. He and Pilzecker arrived at that
structed to respond by reading each syllable aloud
conclusion after a number of experiments on retroac-
with emphasis on every odd-numbered syllable, thus
tive inhibition, a painstaking analysis of the types in-
creating paired associates (pairs of emphasized and
correct answers volunteers were prone to make, and
nonemphasized syllables, or trochees).
a thoughtful consideration of a phenomenon they
Mller and Schumann first adopted Ebbinghauss termed perseveration (i.e., the tendency of study
method of determining how many trials were needed material to inger in a subjects mind after learning).
to memorize a list of twelve syllables (the method of They wrote (Mller and Pilzecker, 1900, 196197):
MULTIPLE-MEMORY SYSTEMS 413
After all this, there is no alternative but to as- (1882). Revision der Hauptpuncte der Psychophysik [Revision of
sume that, after a list of syllables has been the main points of psychophysics]. Leipzig: Breitkopf and
Hrtel.
read, certain physiological processes that Haupt, E. J. (1998). Origins of American psychology in the work
serve to strengthen the associations induced of G. E. Mller, classical psychophysics and serial learning. In
during reading of that list continue with de- R. W. Rieber and K. Salzinger, eds., Psychology Theoretical-
creasing intensity for a period of time. These Historical perspectives, 2nd edition. Washington, DC: American
processes and their facilitating effects on Sociological Association.
Jost, A. (1897). Die Associationsfestigkeit in ihrer Abhngigkeit von
these associations are being weakened to a der Verteilung der Wiederholungen [The strength of associa-
greater or lesser extent if the experimental tions in relation to the distribution of repetitions]. Zeitschrift
subject experiences further mental exhaus- fr Psychologie 14, 436472.
tion immediately after reading a list. . . . It Lechner, H. A., Squire, L. R., and Byrne, J. H. (1999). 100 years
seems justified to suppose that the physiolog- of consolidationremembering Mller and Pilzecker. Learn-
ing and Memory 6, 7787.
ical processes mentioned her are the same McDougall, W. (1901). Experimentelle Beitraege zur Lehre vom
that underlie perseverative tendencies. . . . Gedaechntis, von G. E. Mueller und A. Pilzecker [Experimen-
Mental exertion in an experimental subject tal contributions to the theory of memory by G. E. Mueller
after reading a list of syllables has, first, the and A. Pilzecker].Mind 10, 388394.
direct effect of weakening the perseverative Mller, G. E. (1873). Zur Theorie der sinnlichen Aufmerksamkeit [On
the theory of sensory attention]. Ph.D. diss. Leipzig: Edelman.
tendencies of these syllables and, second, be- (1878). Zur Grundlegung der Psychophysik [On the Founda-
cause the effect of these perseverative ten- tions of Psychophysics], Vol. 4: Kritische Beitrge. Bibliothek fr
dencies is to consolidate syllable associations, Wissenschaft und Literatur. 23 vols. Philosophische Abtheilung.
the additional effect of impairing these asso- Berlin: Greiben.
ciations (retroactive interference). (1879). ber die Maasbestimmung des Ortsinnes der Haut
mittels der Methode der richtigen und falschen Flle [On the
Considering their observations on perseveration, quantification of the location sense of the skin by way of the
Mller and Pilzecker proposed that consolidation oc- method of right and wrong cases]. Pflgers Archiv fr die ge-
samte Physiologie 19, 191235.
curred within about ten minutes. Yet their experi-
(1930). ber die Farbenempfindungen, Psychophysische
ments did not systematically explore this time course. Untersuchungen. Band 1 und 2 [On Sensations of Color, Psy-
McDougall and Burnham quickly recognized the idea chophysical Investigations]. Zeitschrift fr Psychologie, Ergn-
of memory consolidation because it provided a way to zungsband 17 and 18. Leipzig: Barth.
understand temporally graded retrograde amnesia (1934). Kleine Beitrge zur Psychophysik der Farbempfindungen
[Brief Contributions to the Psychophysics of Color Sensa-
that results from traumatic head injury. Later, Carl
tions]. Leipzig: Barth.
Duncans experiments on ECS-induced retrograde Mller, G. E. and Pilzecker, A. (1900). Experimentelle Beitrge zur
amnesia in rats launched decades of research in many Lehre vom Gedchtnis [Experimental contributions to the
laboratories to study consolidation and its time theory of memory]. Zeitschrift fr Psychologie, Ergnzungsband
course. 1, 1300.
Mller, G. E., and Schuman, F. (1894). Experimentelle Beitrge
With the passing decades, awareness of Mllers zur Lehre vom Gedchtnis [Experimental contributions to the
importance to the field of experimental psychology, investigation of memory]. Zeitschrift fr Psychologie 6, 81190,
especially memory research, began to fade. But his 257.
groundbreaking experiments on memory still inspire Sprung, L., and Sprung, H. (2000). Georg Elias Mller and the be-
ginnings of modern psychology. In G. A. Kimble and M. Wer-
productive research in cellular neurobiology and sys-
theimer, eds., Portraits of pioneers in psychology, Vol. 4. Mahwah,
tems neuroscience. NJ: Erlbaum.
An indefatigable scholar, Mller worked long Hilde A. E. Lechner
after his retirement in 1921, producing important
publications until his death, on December 23, 1934.
Bibliography
Boring, E. G. (1950). History of experimental psychology. New York:
Appleton-Century-Crofts. MULTIPLE-MEMORY SYSTEMS
Burnham, W. H. (1903). Retroactive amnesia, illustrative cases and In 1950 Karl Lashley published his influential manu-
a tentative explanation. American Journal of Psychology 14, 382
396. script In Search of the Engram, in which he concluded
Duncan, C. P. (1949). The retroactive effect of electroshock on that memory was widely distributed in the mamma-
learning. Journal of Comparative Physiological Psychology 42, 32 lian brain and that there is no apparent localization
44. of mnemonic traces within specific brain structures.
Ebbinghaus, H. (1885). ber das Gedchtnis, Untersuchungen zur ex- Five decades worth of research since then suggests
perimentellen Psychologie [On memory, Investigations in experi-
mental psychology]. Leipzig: Duncker and Humbolt. that his conclusion may have been partially incorrect.
Fechner, G. (1860). Elemente der Psychophysik [Elements of psycho- Whereas it is clear that distributed brain structures do
physics], 2 vols. Leipzig: Breitkopf and Hrtel. indeed participate in mnemonic functions, it is also
414 MULTIPLE-MEMORY SYSTEMS
the case that there is some degree of neuroanatomical human amnesic syndrome, and the mid-1960s saw
localization of learning and memory. There is exten- several studies examining the effects of hippocampal
sive evidence that memory is organized in multiple system damage in various learning tasks. However,
systems that differ in terms of the type of memory these early studies often failed to reveal memory defi-
they mediate. The multiple-memory-systems hypoth- cits following lesions of the hippocampus, leading
esis is supported by findings of neuroscientific re- some investigators to conclude that this brain struc-
search in several mammalian species, including rats, ture is not particularly important for mnemonic pro-
monkeys, and humans (Hirsh, 1974; OKeefe and cesses. One explanation for the failure of some ani-
Nadel, 1978; Olton, Becker and Handelmann, 1979; mal studies to reveal memory deficits following
Packard, Hirsh, and White, 1989; Kesner, Bolland, hippocampal damage is that this structure mediates
and Dakis, 1993; Mishkin and Petri, 1984; Zola- the acquisition of a specific type of memory. This lat-
Morgan, Squire, and Mishkin, 1982; Cohen and ter view implies that evidence for a mnemonic role for
Squire, 1980; Warrington and Weiskrantz, 1982; the hippocampus would emerge only with the use of
Knowlton, Mangels, and Squire, 1996). In addition to appropriate learning tasks, and this hypothesis be-
providing neuroanatomical dissociations of the role came one of the primary tenets of several dual-
of various brain structures in different memory tasks, memory theories abroad in the mid-1970s.
an important goal of multiple-memory-systems re-
search is elucidation of the psychological operating In an extensive early review of the effects of hip-
principles that distinguish different types of memory. pocampal system damage on performance of various
learning and memory tasks in rats, R. Hirsh (1974)
published perhaps the earliest example of a neuro-
Early Evidence of Multiple-Memory biological-based dual-memory theory. Hirsh pro-
Systems posed that one way to describe the psychological op-
W. B. Scoville and B. Milner (1957) provided erating principles that distinguished putative
early indirect evidence suggesting the existence of hippocampus-dependent and nonhippocampus-
multiple-memory systems in the human brain. In an dependent memory systems involved consideration
attempt to alleviate seizure activity in epilepsy and to of the historic debate between cognitive (e.g. Tolman,
develop possible alternatives to the practice of per- 1932) and stimulus-response (e.g. Hull, 1943) learn-
forming frontal lobotomies in treating psychosis, they ing theorists. In view of the extraordinary emphasis
excised large regions of the medial temporal lobe of placed by early learning theorists on the question of
the brain in a number of patients. The discovery that what animals learn, it is not surprising that the debate
a severe anterograde human amnesic syndrome re- between cognitive and S-R learning theorists was ex-
sulted from removal of temporal lobe neural tissue tremely influential. Hirsh suggested, The behavior
sparked immense interest in the role of this brain re- of hippocampally ablated animals is held to be every-
gion in learning and memory. A process of elimina- thing for which early S-R theorists could have wished
tion, in which mnemonic deficits were contrasted fol- (Hirsh, 1974, p. 439). That same year, J. OKeefe and
lowing varying degrees of damage to different L. Nadel introduced a prcis of their dual-memory
temporal-lobe structures, led to the hypothesis that theory, later elaborated on in their influential book
damage to the hippocampus was primarily responsi- The Hippocampus as a Cognitive Map (1978). Their re-
ble for the memory deficits observed in these patients. markable discovery that hippocampal neurons fired
However, these so-called temporal lobe amnesics in relationship to a rats spatial location led OKeefe
performed normally on some perceptual and motor- and Nadel to suggest that the hippocampus is the
skill learning tasks (e.g. Milner, 1962; Corkin, 1968). neuroanatomical substrate of E. C. Tolmans (1932)
These early discoveries of a pattern of spared and im- cognitive spatial map. Expanding on these ideas, M.
paired learning abilities following temporal lobe Mishkin and colleagues (1984) subsequently offered
damage was consistent with the hypothesis that mem- a distinction between a hippocampal-based cognitive
ory is organized in multiple brain systems. However, memory system and a nonhippocampal based S-R
the development of dual-memory theories, which habit memory system in the monkey brain.
might distinguish independent memory systems, did
not begin in earnest for several more years. Several other dual-memory theories have
emerged based on research in rats, monkeys, and hu-
mans. One prominent theoretical distinction between
Lower Animal Research and the different types of memory in humans is based on a
Development of Dual-Memory Theories pattern of spared and impaired learning in temporal
The ability to produce localized brain damage in lobe amnesia; it draws a distinction between hippo-
animals allowed experimental psychologists to pur- campus-dependent declarative memory and nonhip-
sue the development of an animal model of the pocampus-dependent procedural memory (Cohen
MULTIPLE-MEMORY SYSTEMS 415
and Squire, 1980). Declarative memory involves action methodology to examine the hypothesis that the
quisition and expression of specific facts and events, hippocampal system and caudate nucleus are parts of
whereas procedural memory involves acquisition of independent memory systems: 1. Performance in the
instrumental S-R habits and various forms of classical- win-shift task requires rats to remember those arms
ly conditioned behaviors. A theoretical distinction be- that have been previously visited, while performance
tween explicit and implicit memory processes in hu- in the win-stay task simply requires rats to learn to ap-
mans (Graf and Schacter, 1985) has also received proach lit maze arms. Therefore, the win-shift task is
extensive investigation and support. Explicit memory often considered a prototypical test of spatial working
is held to involve the conscious recollection of facts memory and appears to involve the use of a cognitive
and events, while implicit memory involves uncon- mapping strategy. In contrast, the win-stay task is es-
scious acquisition and expression of learned behav- sentially a simultaneous visual discrimination and
ior. E. B Tulvings distinction between episodic and may involve acquisition of an S (light)-R (approach)
semantic memory represents an additional example habit; 2. The two tasks share the same motivational
of a multiple-memory-systems approach to memory (appetitive), sensory (primarily visual), and motoric
organization in humans (Tulving, 1987). (maze running) characteristics; therefore, any differ-
ential effects of damage to separate brain areas on
performance in the two tasks can be more readily as-
Neuroanatomical Bases of Multiple-
cribed to mnemonic processes.
Memory Systems
When rats were tested in the acquisition of these
Each dual-memory theory was developed from a
two tasks following lesions of the hippocampal system
consideration of the pattern of spared and impaired
and caudate nucleus, a double dissociation resulted:
learning that follows damage to one particular brain
region: the hippocampus. Thus, the existence of mul- Hippocampal-system lesions selectively impaired ac-
tiple-memory systems has often been suggested based quisition of the win-shift task, while caudate-nucleus
on the observations of single dissociationsthat is, lesions selectively impaired acquisition of the win-stay
damage to the hippocampus impairs performance of task (Packard, Hirsh, and White, 1989). A double dis-
task X but not task Y. But, postulating functional in- sociation between the mnemonic functions of the hip-
dependence of memory systems based on a single dis- pocampus and caudate nucleus in these two radial
sociation is difficult because of the potential existence maze tasks has also been demonstrated using post-
of hierarchal relationships among the mnemonic training intracerebral infusions of memory-
role(s) of particular brain structures and task (Weisk- enhancing drugs (Packard and White, 1991). Evi-
rantz, 1989). A more compelling argument for the ex- dence that the caudate nucleus is part of a memory
istence of multiple-memory systems in the brain re- system that mediates at least one form of nonhippo-
quires the demonstration of a double dissociation, (a campal-dependent memory also comes from research
term coined by Teuber, 1955), in which damage to in nonhuman primates and humans (Knowlton, Man-
brain structure X impairs performance on task A but gels, and Squire, 1996; Fernandez-Ruiz et al., 2001).
not task B, and damage to brain structure Y impairs Besides the several studies that have dissociated
performance on task B but not task A. the mnemonic functions of the mammalian hippo-
An early study in rats used two eight-arm radial- campus and caudate nucleus, other studies suggest
maze tasks to demonstrate a double dissociation be- that the cerebellum and amygdala mediate additional
tween the mnemonic functions of the hippocampal forms of learning and memory. R. Thompson and
system and caudate nucleus (Packard, Hirsh, and colleagues have described a role for the cerebellum in
White, 1989). The radial maze consisted of a center classically conditioned eyeblink behavior, providing
platform with eight alley-shaped arms radiating away a detailed analysis of the neural circuitry mediating
from the center and containing food cups at the distal the flow of information underlying the processing of
end-points. In a standard win-shift version of the ra- unconditioned and conditioned stimuli in this form
dial-maze task, rats obtained food rewards by visiting of Pavlovian conditioning (McCormick et al., 1981;
each arm of the radial maze once in a daily training Kim and Thompson, 1997). Other behavioral neuro-
session, and reentries into previously visited maze science research indicates that the mammalian amyg-
arms within a session were scored as errors. In a win- dala mediates stimulus-affect memory, as evidence
stay version of the radial-maze task, rats obtained by a selective role for this brain region in fear condi-
food rewards by visiting four randomly selected and tioning (LeDoux, 1992; Davis, 1992), and stimulus-
illuminated maze arms twice within a daily training reward learning (e.g. Cador, Everitt, and Robbins,
session, and the spatial location of these maze arms 1989; McDonald and White, 1993). Activation of ef-
varied by day. Two important features of these radial- ferent amygdala pathways by emotionally arousing
maze tasks make them ideal for employing dissocia- events also modulates the distinct cognitive and S-R
416 MULTIPLE PERSONALITY
habit memory process subserved by the hippocampus Lashley, K. S. (1950). In search of the engram. Symposium of Society
and caudate nucleus (Packard, Cahill, and McGaugh, for Experimental Biology 4, 454482.
LeDoux, J. E. (1992). Emotion and the amygdala. In J. P. Aggleton,
1994). ed., The amygdala: Neurobiological aspects of emotion, memory, and
Behavioral neuroscience research conducted mental dysfunction. New York: Wiley-Liss.
largely over the last two decades has dramatically in- McCormick, D. A., Lavond, D. G., Clark, G., Kettner, R. E., Rising,
C. E., and Thompson, R. F. (1981). The engram found? Role
creased our understanding of neuroanatomical sys- of the cerebellum in classical conditioning of nictitating mem-
tems that mediate different types of memory in mam- brane and eyelid responses. Bulletin of the Psychonomic Society
malian species ranging from the rat to the human. 18, 103105.
These findings suggest that elucidation of the neuro- McDonald, R. J., and White, N. M. (1993). A triple dissociation of
biological bases of memory organization will remain memory systems: Hippocampus, amygdala, and dorsal stria-
tum. Behavioral Neuroscience 107, 322.
a complex task involving a consideration of patterns Milner, B. (1962). Les troubles de la memorie accompagnat des
of spared and impaired memory function following lesion hippocampiquesblilaterales. In P. Passant, ed., Physi-
various brain manipulations in lower animals, in ologie de lHippocampae. Paris: Centre de la Recherche Scienti-
neuropsychological studies of brain damaged hu- fique.
mans, and in neuroimaging studies of memory pro- Mishkin, M., and Petri, H. L. (1984). Memories and habits: Some
implications for the analysis of learning and retention. In L.
cesses in the intact human brain. R. Squire and N. Butters, eds., Neuropsychology of Memory. New
York: Guilford.
See also: AMNESIA, ORGANIC; DECLARATIVE OKeefe, J. and Nadel, L. (1978). The hippocampus as a cognitive map.
MEMORY; EPISODIC MEMORY; IMPLICIT Oxford, UK: Oxford University Press.
MEMORY; LOCALIZATION OF MEMORY TRACES; Olton, D. S., Becker, J. T., and Handelmann, G. E. (1979). Hippo-
MEMORY CONSOLIDATION: PROLONGED campus, space, and memory. Behavioral and Brain Sciences 2,
PROCESS OF REORGANIZATION; NEURAL 313365.
SUBSTRATES OF CLASSICAL CONDITIONING: Packard, M. G., Cahill, L., and McGaugh, J. L. (1994). Amygdala
DISCRETE BEHAVIORAL RESPONSES; NEURAL modulation of hippocampal-dependent and caudate nucleus-
SUBSTRATES OF EMOTIONAL MEMORY dependent memory processes. Proceedings of the National Acad-
emy of Sciences of the United States of America 91, 8,4778,481.
Packard, M. G., Hirsh, R, and White, N. M. (1989). Differential ef-
Bibliography
fects of fornix andcaudate nucleus lesions on two radial maze
Cador, M., Robbins, T. W., and Everitt, B. J. (1989). Involvement tasks: Evidence for multiple memory systems. Journal of
of the amygdala in stimulus-reward associations: interaction Neuroscience 9, 1,4651,472.
with the ventral striatum. Neuroscience 30, 7786. Packard, M. G., and White, N. M. (1991). Dissociation of hippo-
Cohen, N. J., and Squire, L. R. (1980). Preserved learning and re- campus and caudate nucleus memory systems by posttraining
tention of pattern analyzing skill in amnesics: Dissociation of intracerebral injection of dopamine nists. Behavioral Neuro-
knowing how and knowing that. Science 210, 207210. science 105, 295306.
Corkin, S. (1968). Acquisition of motor skill after bilateral medial Scoville, W. B., and Milner, B. (1957). Loss of recent memory after
temporal lobe excision. Neuropsychologia 6, 255265. bilateral hippocampal lesions. Journal of Neurology, Neurosur-
Davis, M. (1992). The role of the amygdala in conditioned fear. In gery, and Psychiatry 20, 1121.
J. P. Aggleton ed., The amygdala: Neurobiological aspects of emo- Sutherland, R. J., and Rudy, J. W. (1989). Configural association
tion, memory, and mental dysfunction. New York: Wiley-Liss. theory: The role of the hippocampal formation in learning,
Fernandez-Ruiz, J., Wang, J., Aigner, T. G., and Mishkin, M. memory, and amnesia. Psychobiology 17, 129144.
(2001). Visual habit formation in monkeys with neurotoxic le- Teuber, H. L. (1955). Physiological Psychology. Annual Review of
sions of the ventrocaudal neostriatum. Proceedings of the Na- Psychology 6, 267296.
tional Academy of Sciences of the United States of America 98, Tolman, E. C. (1932). Purposive behavior in animals and men. New
4,1964,201. York: Appleton-Century-Crofts.
Graf, P., and Schacter, D. L. (1985). Implicit and explicit memory Tulving, E. (1987). Multiple memory systems and consciousness.
for new associations and amnesic patients. Journal of Experi- Human Neurobiology 6, 6780.
mental Psychology: Learning, Memory, and Cognition 11, 501 Warrington, E. K, and Weiskrantz, L. (1982). Amnesia: A discon-
518. nection syndrome? Neuropsychologia 20, 233247.
Hirsh, R. (1974). The hippocampus and contextual retrieval of in- Weiskrantz, L. (1989). Remembering dissociations. In H. L.
formation from memory: A theory. Behavioral Biology 12, 421 Roediger, and E. Tulving, eds., Varieties of memory and con-
444. sciousness: Essays in honour of Endel Tulving. Hillsdale, NJ: Erl-
Hull, C. L. (1943). Principles of Behavior. New York: Appleton- baum.
Century-Crofts. Zola-Morgan, S., Squire, L., and Mishkin, M. (1982). The
Kesner, R. P., Bolland, B. L., and Dakis, M. (1993). Memory for neuroanatomy of amnesia: Amygdala-hippocampus versus
spatial locations, motor responses, and objects: Triple dissoci- temporal stem. Science 218, 1,3371,339.
ation among the hippocampus, caudate nucleus, and ex-
trastriate visual cortex. Experimental Brain Research 93, 462 Mark G. Packard
470.
Kim, J. J., and Thompson, R. F. (1997). Cerebellar circuits and syn-
aptic mechanisms involved in classical eyeblink conditioning.
Trends in Neuroscience 20, 177181.
Knowlton, B. J., Mangels, J. A., and Squire, L. R. (1996). A MULTIPLE PERSONALITY
neostriatal habit learning system in humans. Science 273,
1,3991,402. See: AMNESIA, FUNCTIONAL
N
NATURAL SETTINGS, MEMORY IN view, in which the witness is asked to recall events in
different orders and from different perspectives.
The study of memory as it is used in natural settings
is now an accepted part of the scientific study of mem- F. C. Bartletts 1932 book Remembering included
ory, engaging the attention of many researchers. This the first systematic research on memory for stories, a
is a relatively new development; the classical study of field that was then largely neglected until the 1970s.
memory took little interest in naturalistic studies. In Since that time, however, Bartletts concept of sche-
the century between 1880 and 1980, an emphasis on ma has been widely adopted not only in the study of
story recall but wherever memory is facilitated by pre-
experimental control led most experimenters to use
existing mental structures.
specially prepared tasks and materials in their work,
even ifas often happenedthose tasks seemed A related subject is the study of memory as it ap-
meaningless to their subjects. The study of memory pears in oral poetry and oral traditionfor example,
in natural settings was largely ignored. in the preliterate heroic epics of ancient Greece. For
decades this field was the province of humanistic
scholars and classicists, but David Rubins 1995 analy-
Early Research sis based on the principles of cognitive psychology has
been very successful.
Two important exceptions to this trend are worth
mentioning: eyewitness testimony and memory for
stories. In the early 1900s, William Stern established Naturalistic versus Laboratory Methods
a journal (Beitrage zur Psychologie der Aussage) devoted A 1978 conference at Cardiff, Wales, brought to-
entirely to the study of testimony. In his experiments, gether many researchers with an interest in practical
Stern staged unexpected or dramatic events in the aspects of memory. In the opening address Ulric
presence of groups of people who were then interro- Neisser attacked the traditional study of memory as
gated as if they had witnessed a real crime. Such testi- largely unproductive and called for a new emphasis
mony is surprisingly unreliable. Witnesses often give on more naturalistic studies. Since that time many
highly confabulated accounts of the event itself and such studies have been conducted, and several related
are rarely able to describe the criminals. Confi- conferences have been held. This work is not without
dence is no guarantee of accuracy in such reports. its critics. Some proponents of more traditional mem-
Modern research confirms that eyewitnesses may be ory research believe that naturalistic studies have lit-
entirely wrong even when they are quite sure they are tle value and that scientific progress can result only
right. One documented way to improve eyewitness from controlled laboratory experiments. In an article
testimony is Edward Geiselmans cognitive inter- titled The Bankruptcy of Everyday Memory Mah-
417
418 NATURAL SETTINGS, MEMORY IN
zarin Banaji and Robert Crowder asserted that a de- riences were remembered best of all. There are prob-
cade of naturalistic research had produced no impor- ably individual differences in these types of bias.
tant findings. A symposium incorporating nine
In a different research tradition, Michael Ross
replies to Banaji and Crowder appeared in the Janu-
has proposed a specific theory of the bias that appears
ary 1991 American Psychologist. Since that time the con-
when people try to remember their own past charac-
troversy seems to have abated, with research on mem-
teristics. (How severe were my headaches last month?
ory in natural settings now recognized as a significant
How fast did I read before I took this study-skills
component of the study of memory as a whole.
course?) Such estimates are not so much recalled as
deduced from the traits current level (my headache
Long-Term Memories today, my reading speed now) together with an im-
plicit theory of its stability or tendency to change.
Because laboratory experiments are necessarily When that implicit theory suggests that a trait is stable
limited in time, studies of very long-term memory al- (my headaches are chronic, I always read this fast), its
most always take advantage of natural settings. The past value is remembered simply by reference to its
most systematic studies of this kind have been con- present level. The situation is different when the im-
ducted by Harry Bahrick and his associates, who have plicit theory suggests that a change should have oc-
found that some forms of memory are very persistent. curred. Thus people who have just been through a
The ability to recognize yearbook photographs of headache treatment program (or taken a study-skills
ones high school classmates, for example, remains course) will tend to remember their earlier pain le-
strong even thirty-five years after graduation. In con- vels as worse (or their earlier reading speed as slower)
trast, recall of classmates names declines steadily than they really were. After all, the treatment (or the
over the same period. Recall for the material actually course) must have done some good! This is only one
learned in school (e.g., a foreign language) tends to of several possible sources of error in recalling medi-
drop off for the first five years, but after that there is cal information. Information that confirms patients
little more forgetting over the next twenty-five. Inter- previous beliefs seems better remembered than infor-
estingly, even very brief reminders may bring much mation that gainsays those beliefs.
of the apparently forgotten material to mind again.
Memories Triggered by Cues

Recall of Life Experiences
Another way to study memory for life events
A number of psychologists have studied their own pioneered by Francis Galton more than a century
memories by making brief notes of one or more expe- agois to provide random words as cues and ask sub-
riences each day for a prolonged period. Then, after jects to report whatever personal experiences (if any)
a delay of months or years, they test their own recall the cue brings to mind. When the number of memo-
of the recorded events. These diary studies have ries retrieved is plotted on a log-log scale as a function
produced several clear results. For one thing, forget- of the time that has passed since the original event,
ting is not all-or-none: an event that seemed at first an approximately linear forgetting function appears.
to be completely forgotten may be retrieved if more There are, however, two significant departures from
cues are provided. Another fairly obvious finding is linearity in forgetting. One is that middle-aged and
that experiences rated as important are recalled elderly individuals recall disproportionately many
somewhat better than unimportant ones. But impor- events from their years of adolescence and early
tance turns out to be very difficult to rate: the true sig- adulthood (roughly from ages ten to twenty-five).
nificance of an experience may depend on later de- This pattern is called the reminiscence bump. The
velopments and hence may not be apparent when it bump appears not only in memories of life experi-
occurs. ences but also in the recall of music, books, and news
events: apparently it is the most memorable period of
Some scholars have suggested that memory is
life in every sense.
egotistically biased, that we recall our own actions as
more praiseworthy (and perhaps our own experi- The other departure from linearity appears at the
ences as more pleasant) than they really were. Such beginning of the forgetting function: few if any events
distortions certainly do happen and have been veri- are recalled from the first few years of life. The near
fied in research settings. Nevertheless, the diary absence of memories from early childhood, which
studies confirm this tendency only in part. On the av- Sigmund Freud called infantile amnesia, has been
erage, experiences rated as pleasant are indeed better the object of considerable interest in its own right.
remembered than unpleasant ones. But at least one Current interpretations of this phenomenon have
study found that personally relevant unpleasant expe- been much influenced by modern studies of young
NEOCORTICAL PLASTICITY 419
childrens memory. Although children as young as Ross, M. (1989). Relation of implicit theories to the construction
two years have some ability to recall past experiences, of personal histories. Psychological Review 96 341357.
Rubin, D. C., Rahhal, T. A., and Poon, L. W. (1998). Things
memories at that age have little internal organization learned early in adulthood are remembered best. Memory &
and require much cueing by adults. Perhaps it is be- Cognition 26, 319.
cause young children lack the narrative skills that are
Ulric Neisser
necessary to link individual memories to one another
and to a develop life story that adults can recall so lit-
tle from that period of life.
NEOCORTICAL PLASTICITY
Flashbulb Memories
[A fundamental question in philosophy, psychology, and
Some experiences are remembered so vividly and
neuroscience concerns the development of our perceptions of
confidently that they seem unforgettable, almost as
the world. Are our perceptions completely determined by the
though the brain took a snapshot of the scene at that
intrinsic brain circuitry and its development, independent of
moment; such recollections are often called flash-
experience (nativism), or are our perceptions and the under-
bulb memories. Most flashbulb memories are of
lying neural circuitry determined by experience (empiri-
unique and personal experiences, but some are public
cism)? A truly major contribution of neuroscience in the
(in a sense): they capture the moment when one first
twentieth century was the resolution of this issue, based
learned of some surprising and significant public
largely on study of the development of the mammalian visual
event. Psychologists have studied the flashbulb
system. As might be expected the truth lies somewhere in be-
memories created by hearing the news of the assassi-
tween.
nations of President Lincoln, President Kennedy, and
Martin Luther King Jr. (among others) as well as of This extraordinary story is told in the third entry in this
other dramatic events such as the explosion of the series, on the DEVELOPMENT OF THE VISUAL SYSTEM. In
space shuttle Challenger, the verdict in the O. J. Simp- brief, the basic wiring diagram of the mammalian visual sys-
son trial, and so on. They have found that despite the tem is specified by genetic and developmental factors. How-
strong subjective conviction that typically accompa- ever, normal visual stimulation (i.e., experience) is neces-
nies such memories, they are not neessarily accurate. sary for the system to develop normally. Thus, occluding one
eye in developing mammals (most work has been done on
False memories can also occur in other contexts,
cats and monkeys) at a particular critical period markedly
often as a result of suggestion. Such suggestion can
take many forms, but one particularly important con- alters the normal wiring diagram in the primary visual sys-
text is psychotherapy. It is now well established that tem, resulting in markedly impaired vision. Patterned visual
false memories of childhood sexual abuse can be pro- stimulation must be temporally synchronized (simultaneous)
duced by suggestion during therapy and that this can between the two eyes and must also be spatially synchronous
have serious consequences. To be sure, not all recol- for normal binocular vision to develop. This synchronized
lections of childhood abuse are false; many are all too pattern leads to changes in the strength of the connections
valid. Unfortunately, valid and invalid memories are between active neurons, a phenomenon called synaptic plas-
not distinguishable by any simple internal character- ticity.
istic. The only way to establish the accuracy of a given In addition to the extraordinary degree of plasticity seen
recollection is with external evidence, which is not al- in the development of the visual system, the ADULT VISUAL
ways available. CORTEX is also capable of a remarkable degree of adapta-
tion and reorganization. Thus, infusing neuromodulatory
See also: AMNESIA, INFANTILE; EXPERTS MEMORIES; and growth factors in the visual cortex of adult mammals
FALSE MEMORIES; ORAL TRADITIONS can reinstate the plasticity in ocular dominance characteris-
Bibliography tics. Retinal lesions result in a wide range of biochemical
Bahrick, H. P. (1984). Semantic memory content in permastore: changes in the deafferented visual cortex. Perhaps more im-
Fifty years of memory for Spanish learned in school. Journal portant, the adult visual cortex exhibits marked synaptic
of Experimental Psychology: General 117, 129. plasticity: processes of LONG-TERM POTENTIATION
Banaji, M. R., and Crowder, R. G. (1989). The bankruptcy of every- (LTP) and LONG-TERM DEPRESSION (LTD). The Mc-
day memory. American Psychologist 44, 1,1851,193. Colloch effect is an example of perceptual plasticity that per-
Bartlett, F. C. (1932). Remembering. Cambridge, UK: Cambridge
University Press. sists for up to twenty-four hours; after exposure to colored
Geiselman, R. E., Fisher, R. P., MacKinnon, D. P., and Holland, H. gratings, black and white gratings appear to be limited in
L. (1985). Eyewitness memory enhancement in the police in- complementary colors.
terview: Cognitive retrieval mnemonics versus hypnosis. Jour-
nal of Applied Psychology 70, 401412. Another system that exhibits dramatic neocortical plas-
Gruneberg, M. M., Morris, P. E., and Syjes, R. N., eds. (1978). Prac- ticity is the SOMATOSENSORY CORTEX, which consists of a
tical Aspects of Memory. Chichester, UK: Wiley. number of processing areas with representations or maps
420 NEOCORTICAL PLASTICITY: Adult Visual CortexAdaptation and Reorganization
of the body surface. Experience can alter the organization of Dark rearing tends to maintain the visual cortex
these maps dramatically. Thus, amputation of digits or limbs in an immature, plastic state. This extended period
in monkeys and humans results in marked changes in soma- of plasticity may be due to the prolonged expression
tosensory cortex such that the maps of the adjacent intact of Cpg15, an activity-regulated gene that encodes a
body parts expand. Similarly, particular types of training membrane-bound ligand that coordinates the growth
(e.g., in tactile discriminations or motor skills) can result in of dendrites and axonal arbors and the maturation of
expansion of the relevant regions of somatosensory cortex in their synapses. Normally, cpg15 levels decline during
both monkeys and humans. These experience-dependent al- the critical period but dark-rearing results in the fail-
terations in somatosensory cortex appear to involve alter- ure to down-regulate this gene and thus the cortex re-
ations in both the cortex itself and the relays of sensory infor- mains in a state of enhanced ability to form new syn-
mation in the brain stem and thalamus. aptic connections.
The MOTOR CORTEX also exhibits marked plasticity. Finally, the study of visual cortical development
Training in particular motor skills can alter to some degree and plasticity has turned to mouse models, where de-
the organization of motor representation in the motor area spite the lack of ocular dominance segregation of
of the neocortex. However extensive repetition of a particu- thalamocortical afferents, knockouts of specific genes
lar movement does not necessarily result in expansion of rep- allows for the unprecedented ability to examine the
resentation of the movements in the motor cortex. There is role of receptors, ligands, and enzymes in the devel-
evidence that processes of LTP and LTD are involved in opment and maintenance of visual cortical structure,
motor learning. The cerebellum is also much involved in the function, and plasticity.
learning of motor skills.]
LTP and LTD in Visual Cortex
Adult visual cortex shows evidence for synaptic
ADULT VISUAL CORTEXADAPTATION modifications in both in vitro and in vivo prepara-
AND REORGANIZATION tions. Tetanic stimulation of the LGN in adult rats
The adult visual cortex is capable of a remarkable de- leads to the induction of long-term potentiation
gree of plasticity in the face of alterations of its nor- (LTP) that is dependent on NMDA-receptor activa-
mal pattern of input or in response to sensory adaptation. The result is enhanced field potentials in layer
tion or perceptual practice. This adult plasticity 4 and deep layer 3 and increased visual evoked poten-
complements the use-dependent plasticity that plays tials to flash and grating stimulation. This LTP indi-
an important role in the establishment of the normal cation is paralleled by an increase in zif-268 expres-
organization during development. sion in layers 2 and 3. In contrast, long-term
depression, which has been hypothesized to underlie
many plastic changes in development, has been in-
Monocular Deprivation Plasticity in Adults consistently demonstrated in adult visual cortex. Con-
The shift in ocular dominance of cortical neurons sistent low-frequency stimulation (1 hertz) in cortical
as a consequence of monocular deprivation is one of slices can induce LTD, but more physiologically
the hallmarks of cortical plasticity during develop- based stimulation, based on a Poisson distribution of
ment. The potential for ocular dominance plasticity interpulse intervals, fails to produce LTD.
is greatest during a critical period of development,
during which time thalamocortical axons segregate
into ocular dominance columns. However, residual
Plasticity Following Retinal and Cortical
plasticity remains during later development even Lesions
after this structural remodeling has occurred. Howev- Plastic changes in the topographic organization
er, later deprivation fails to induce plastic changes in and receptive field properties have been reported in
ocular dominance to cortical neurons. Several lines of adult visual cortex following retinal and cortical le-
evidence suggest that ocular dominance plasticity can sions. Following binocular retinal lesions or monocu-
be restored to visual cortex through neuromodulators lar retinal lesions paired with enucleation of the other
and growth factors. For example, modulation of eye, researchers have observed a profound topo-
cAMP activity through NA-beta adrenoreceptor acti- graphic reorganization of visual cortex. Immediately
vation, cholera A subunit, or forskolin produces a oc- after the retinal lesion there is an unmasking of ectop-
ular dominance plasticity in adult cats that results in tic receptive fields in the deafferented cortical zone.
either a shift of ocular dominance or a reduction in These large, unmasked receptive fields can be dis-
binocularity. Cortical infusion of nerve growth factor, placed up to fifteen degrees, reflecting the unmask-
NGF, in adult cats, results in a paradoxical shift of oc- ing of intrinsic horizontal connections that extend up
ular dominance toward the deprived eye. to 5 millimeters. Over the following weeks to months,
NEOCORTICAL PLASTICITY: Adult Visual CortexAdaptation and Reorganization 421
the cortical scotoma becomes nearly completely driving extensive computer-based visual field training. In
en by receptive fields arising from non-lesioned por- addition, there is functional MRI evidence for topo-
tions of the retina. This reorganization is described by graphic reorganization extrastriate cortical areas fol-
a progressive shift in the topographic map into the lowing restricted lesions of primary visual cortex in
deafferented zone. Neurons within the reorganized adult humans.
zone demonstrate nearly normal orientation, direc-
tion, and spatial frequency tuning. However, the con-
trast thresholds of cells are markedly elevated. This Adaptation and Perceptual Learning
second phase of cortical reorganization appears due The perceptual phenomenon of adaptation after-
in part to the growth of new cortical connections into effects suggests the possibility of use-dependent and
the deafferented zone. long-lasting (minutes to hours) modifications of re-
sponsiveness of feature-specific neurons in the visual
Retinal lesions in adults lead to a wide range of
cortex. Prolonged exposure to patterned stimuli
biochemical changes in the deafferented cortex. The
leads to a feature-specific elevation of perceptual
growth factors BDNF, NT-3, and NGF are elevated
thresholds that lasts several minutes. Such adaptation
from three days to two years post-lesion. Neu-
has been shown for the movement and the orienta-
rotrophin receptors p75, TrkB, and TrkC increase.
tion of contours, for the spatial frequency of gratings,
The expression of CREB, CaMKII, and MAP2 are
and for color contrast. A related and very long-lasting
also increased. In contrast, zif268 is decreased at
(up to twenty-four hours) adaptation phenomenon is
twenty hours and recovered by thirteen months, while
the McCulloch effect. After exposure to colored grat-
synaptophysin and GAP-43 increased at six months
ings of different orientation, black and white gratings
post-lesion. Neurotransmitters also show changes in
appear to a subject to be tinted in complementary col-
the deafferented zone. GABA and glutamic acid de-
ors that remain associated with the respective orienta-
carboxylase levels decrease within the deafferented
tions of the gratings. All these adaptation aftereffects
zone while glutamate levels increase initially along
show interocular transfer, indicating that adaptation
the edge of the deafferented zone and this peak pro-
has occurred at the cortical level. Recordings from
gressively shifts into the deafferented zone.
single cells support the idea that adaptation occurs in
Similar to the plasticity evoked by retinal lesions visual cortex. Optical and single cell recordings in cat
in adults, a temporary artificial scotoma produces visual cortex demonstrate that adaptation to a given
rapid increases in receptive field sizes of neurons orientation produces a repulsive shift in the orienta-
along the border of the scotoma. Receptive fields in- tion tuning of single cells. This phenomenon is most
crease nearly twofold in width during the scotoma pronounced at orientation singularities in visual cor-
and rapidly return to normal following removal of the tex where the representations of the full range of ori-
scotoma. Similar results have been observed in areas entations are brought into close physical proximity.
V2 and V3 of macaque monkeys and may be related Perceptual learning refers to the increase in per-
to perceptual filling-in and color-constancy phenom- formance in discrimination tasks that results from
ena. practice. In general, this enhanced performance
There is evidence for functional reorganization shows interocular transfer and is specific for the topo-
after cortical lesions. In patients suffering from visual graphic location and specific stimulus features of the
field defects due to cortical lesions, intensive visual practiced task. These results have been interpreted as
training can lead to shrinkage of the scotoma. This enhanced performance by task-specific cortical mod-
phenomenon suggests that cortical tissue adjacent to ules that receive retintopic input and learn to solve a
the zone of destruction can recover functions in a use- task after a short training phase.
dependent manner. Such an interpretation is sup- In conclusion, behavioral, electrophysiological,
ported by studies of cortical lesions in adult cats. Re- and morphological evidence confirms the persistence
stricted lesions of visual cortex lead to plastic changes of use-dependent plasticity in the striate cortex of
in visual field topography. Immediately after a corti- adult mammals and humans. Furthermore, twenty-
cal lesion, there is little if any change in visual topog- first century research has begun to uncover the bio-
raphy, but there is some change in spontaneous activ- chemical mechanisms that support this cortical plas-
ity and excitability of cells surrounding the lesion. ticity. This evidence supports the notion that adap-
Within two to three months, cells within this border tivity is a constituent property of cortical networks.
region develop significantly larger receptive fields
that result in a partial filling in of the cortical scoto- See also: GLUTAMATE RECEPTORS AND THEIR
ma. This increase in RF size is limited to three to four CHARACTERIZATION; GUIDE TO THE ANATOMY
degrees, which is approximately the same dimension OF THE BRAIN: CEREBRAL CORTEX; LONG-TERM
reported for shifts in human cortical scotomas follow- DEPRESSION IN THE CEREBELLUM,
422 NEOCORTICAL PLASTICITY: Auditory Cortex
HIPPOCAMPUS, AND NEOCORTEX; LONG-TERM or US) to the face. Evoked potentials elicited by the
POTENTIATION; SECOND MESSENGER SYSTEMS CS in the ACx became larger during conditioning.
This was soon followed by a two-tone discrimination
Bibliography
study showing that the increased magnitude of re-
Chino, Y., Smith, E. L., Whang, B., Matsuura, K., Mori, T., and
Kaas, J. H. (2001). Recovery of binocular responses by cortical
sponse was caused by the specific association of a CS
neurons after early monocular lesions. Nature Neuroscience 4, and US versus a second, unpaired tone. Many more
689690. studies in various animals and training situations also
Crist, R. E., Li, W., and Gilbert, C. D. (2001). Learning to see: Ex- demonstrated that the responses of the auditory cor-
perience and attention in primary visual cortex. Nature Neuro- tex to sounds were affected by the learned psychologi-
science 4, 519525.
Galuske, R. A., Kim, D-S., Castren, E., and Singer, W. (2000). Dif-
cal importance of acoustic stimuli.
ferential effects of neurotrophins on ocular dominance plas- But these results had little effect on the neuro-
ticity in developing and adult cat visual cortex. European Jour-
nal of Neuroscience 12, 3,3153,330.
biology of learning/memory or auditory neurophy-
Gilbert, C. D. (1998). Adult cortical dynamics. Physiological Reviews siology. Auditory physiologists took scant notice of
78, 467485. learning studies that used only one or two tones be-
Heynen, A. J., and Bear, M. F. (2001). Long-term potentiation of cause they shed little light on how the ACx processed
thalamocortical transmission in adult visual cortex in vivo. the many acoustic frequencies to which it is sensitive.
Imamura, K., Kasamatsu, T., Shirokawa, T., and Ohashi, T. (1999).
Learning/memory workers persisted in the shared as-
Restoration of ocular dominance plasticity mediated by aden- sumption that the auditory cortex is a sound analyz-
osine 3,5-monophosphate in adult visual cortex. Proceedings er but not a learning machine. But this assump-
of the Royal Society of London 266, 1,5071,516. tion is not corroborated by the brain.
Kirkwood, A., Lee, H-K., and Bear, M. (1995). Co-regulation of
long-term potentiation and experience-dependent plasticity
in visual cortex by age and experience. Nature 375, 328331.
Nedivi, E., and Lee, W-C. A. (2002). Extended plasticity of visual
Receptive Field Plasticity in Learning and
cortex in dark-reared animals may result from prolonged ex- Memory
pression of cpg15-like genes. Journal of Neuroscience 22, 1,807
1,815.
A solution to the impasse involved combining
Obata, S., Obata, J., Das, A., and Gilbert, C. D. (1999). Molecular methods from auditory neurophysiology with those
correlates of topographic reorganization in primary visual from learning and memory. Researchers determined
cortex following retinal lesions. Cerebral Cortex 9, 238248. the effects of classical conditioning on the receptive
Daniel J. Felleman fields of neurons in the ACx. The receptive field (RF)
of a cell is the part of the stimulus environment to
which it is sensitive. Thus, the RF for frequency is de-
AUDITORY CORTEX scribed by a tuning curve that plots the number of
cellular discharges as a function of acoustic frequency.
Learning is acquiring information and memory is The best frequency (BF) of a cell is the frequency
storing it. A common belief is that sensory systems are that elicits the most discharges. Demonstrating that
only sensory analyzers that provide instantaneous learning involves a systematic change in the process-
knowledge of the world, but are not sites of informa- ing of acoustic-frequency information involves show-
tion storage. However, during the last decade of the ing an alteration in receptive fields when a subject
twentieth century, neurophysiological recordings in learns that a particular frequency is behaviorally im-
the auditory cortex revealed that learning and memo- portant.
ry involve specific and enduring changes in the way
that the brain analyzes and stores sound. This article This approach was first applied to the ACx in
covers the topic of neural plasticity in the auditory 1990 at the University of California at Irvine. RFs
cortex (ACx) during learning and memory in adult were obtained before and immediately after a single
animals and humans. As used here, the term neural conditioning session in which a tone was paired with
plasticity refers to systematic, long-term (minutes to a mild shock. Behavioral conditioned responses de-
lifetimes) changes in the responses of neurons to the veloped. More importantly, the RFs developed sys-
same physical stimulus (e.g, a tone) arising out of an tematic plasticity. Responses to the CS frequency in-
experience. creased, whereas responses to the best frequency and
many other frequencies decreased. These opposing
changes were often sufficient to shift frequency tun-
Background ing toward or to the frequency of the CS so that it be-
In 1955 Galambos and his colleagues performed came the new BF (see Figure 1). Tuning shifts were
a seminal experiment in which cats were classically caused only by association between a tonal CS and the
conditioned by pairing an auditory conditioned stim- shock US. Also, this plasticity was not caused by arous-
ulus (CS) with a puff of air (unconditioned stimulus, al to the CS frequency because RF shifts are still
NEOCORTICAL PLASTICITY: Auditory Cortex 423
Figure 1
Learning shifts receptive field tuning to the frequency of the conditioned stimulus in tone-shock conditioning (30
trials). The best frequency before learning was 0.75 kHz. The subject was trained with a CS of 2.5 kHz. A.
Pretraining and posttraining receptive fields. Note that tuning shifted so that 2.5 kHz became the best frequency.
The inset shows the opposite changes of response in poststimulus time histograms for the pretraining BF and the
CS frequency during the pre- and post-periods. B. Difference in receptive fields due to conditioning (Posttraining
minus Pretraining RFs). Note that the maximum increase of response was at the CS frequency and the maximum
decrease in response was at the pretraining best frequency.
424 NEOCORTICAL PLASTICITY: Auditory Cortex
Figure 2
CS-specific induction of behavioral memory (indexed by respiration changes) following pairing tone (6 kHz) with NB stimulation in
another context. A) Examples of individual respiration records (with value of respiration change index, RCI) to three frequencies (2, 6,
and 12 kHz) for one animal each from the paired and unpaired groups. The largest response was at 6 kHz for the paired animal (RCI
= 0.50). Horizontal bar indicates tone duration. B) Frequency generalization functions. Left panel, group mean ( SE) change in
respiration to all tones for both groups. The maximal response was at the CS frequency of 6 kHz for the paired group, but not for the
unpaired group. Right panel, the group difference function (Paired minus Unpaired) shows a high degree of specificity of respiratory
responses to 6 kHz.
observable after training when subjects are placed RF plasticity possesses all of the major character-
under general anesthesia and thus not aroused by istics of memory. In addition to being associative, it
stimuli. is highly specific to the CS frequency, develops very
NEOCORTICAL PLASTICITY: Development of the Visual System 425
rapidly (within five trials), exhibits consolidation (i.e., deoxyglucose: II. Auditory cortex plasticity. Behavioral Brain
becomes stronger without additional training), and is Research 20, 281293.
retained indefinitely (tested to eight weeks). More- Gluck, H., and Rowland, V. (1957). Defensive conditioning of elec-
trographic arousal with delayed and differentiated auditory
over, RF plasticity exhibits generality across different stimuli. Electroencephography and Clinical Neurophysiology 11,
types of training (e.g., one- and two-tone discrimina- 485491.
tion training, in both in classical and instrumental McLin, D. E. III, Miasnikov, A. A., and Weinberger, N. M. (2002).
conditioning), and motivation (appetitive as well as Induction of behavioral associative memory by stimulation of
aversive). RF plasticity also develops in humans and the nucleus basalis. Proceedings of the National Academy of Sci-
ences of the United States of America.
produces an expanded representation of the CS fre- Morris, J. S., Friston, K. J., and Dolan, R. J. (1998). Experience-
quency in the ACx. dependent modulation of tonotopic neural responses in
RF plasticity may provide a memory code for ac- human auditory cortex. Proceedings of the Royal Society of Lon-
don 265, 649657.
quired stimulus importance: the greater the behav- Recanzone, G. H., Schreiner, C. E., and Merzenich, M. (1993).
ioral significance, the greater the number of cells that Plasticity in the frequency representation of primary auditory
become tuned to that stimulus. Such a memory code cortex following discrimination training in adult owl mon-
could explain aspects of selective attentionfor ex- keys. Journal of Neuroscience 13, 87103.
ample, why one is more likely to hear ones own name Weinberger, N. M. (2002). Experience-dependent response plas-
ticity in the auditory cortex: Issues, characteristics, mecha-
in a noisy room than a random name. While back- nisms and functions. In T. Parks, ed., Handbook of Auditory Re-
ground noise may control many neurons, the attune- search. New York: Springer-Verlag.
ment of still more more networks of neurons to our Weinberger, N. M., and Diamond, D. M. (1987). Physiological plas-
names increases the probability that the name will en- ticity of single neurons in auditory cortex: Rapid induction by
gage some of these cells. This process could explain learning. Progress in Neurobiology 29, 155.
why the loss of memory in aging and brain degenera- Norman M. Weinberger
tion take less of a toll on ones most important memo-
ries. If the memory is represented by more cells, then
important information has a safety factor in num-
bers, which blunt the impact of cell loss. DEVELOPMENT OF THE VISUAL SYSTEM
The neocortex, or cerebral cortex, is a thick layer of
Mechanisms and Functions cells (neurons) covering the outer surface of the fore-
Subcortical neuromodulatory transmitter systems brain. In humans and some other mammals it is wrin-
probably play a role in ACx plasticity in learning. The kled, with many folds separated by deep fissures. In
preponderance of evidence implicates the nucleus humans 75 percent or more of the neurons in the
basalis cholinergic system (NB/ACh) in enabling the forebrain are in the cerebral cortex, a much higher
auditory cortex to store specific information. For ex- proportion than in for other animals. Partly for this
ample, RF plasticity induced in a variety of tasks, mo- reason, and also because of clinical observations on
tivations, and species also can be induced by substitut- people with brain damage, most researchers believe
ing NB stimulation for appetitive or aversive that the cortex plays a significant role in uniquely
reinforcement. That this plasticity can be blocked by human behaviors, such as highly complex learning
the administration of atropine directly to the ACx and information processing.
shows that the induced RF plasticity requires the en- Our knowledge of how the cortex works has
gagement of cholinergic muscarinic receptors in the grown enormously since the 1950s and 1960s, mostly
ACx. Moreover, pairing a tone with stimulation of the through experiments in which the activity of cortical
NB/ACh is sufficient to induce actual specific behav- neurons is studied one neuron at a time. This proce-
ioral memory in the rat (see Figure 2). Although dure may seem paradoxical, since there is great re-
much more research is necessary, researchers have es- dundancy in the brain and the activity of any particu-
tablished the outlines of a reductionistic account of lar neuron cannot be crucially important. But
learning as the storage of specific information in the through painstaking studies of hundreds of thou-
auditory cortex. sands of neurons in different cortical regions, a pic-
ture has emerged that allows us to understand many
Bibliography
basic operating principles of the cortex.
Bakin, J. S., and Weinberger, N. M. (1990). Classical conditioning
induces CS-specific receptive field plasticity in the auditory An important function of the cortex is percep-
cortex of the guinea pig. Brain Research 536, 271286. tion. Just about all the information we receive from
Galambos, R., Sheatz, G., and Vernier, V.G. (1955). Electrophy- the world around us, through our senses, arrives
siological correlates of a conditioned response in cats. Science
123, 376377.
eventually in the cortex. The visual cortex is one of
Gonzalez-Lima, F., and Scheich, H. (1986). Neural substrates for the best-understood cortical regions. It has a compli-
tone-conditioned bradycardia demonstration with 2- cated, elegant structural and functional organization
426 NEOCORTICAL PLASTICITY: Development of the Visual System
that appears to underlie many processes of visual per- weeks, controlling or manipulating their visual expe-
ception and recognition. rience to depart from the norm in various ways. Ensu-
Researchers have devoted extensive experimen- ing microelectrode studies determine whether and
tation to answering the intriguing question of how the how the abnormal experience has altered the activity
brain develops these sophisticated mechanisms for of visual cortical cells.
analyzing and interpreting the visual world. It is now During the first weeks or months of postnatal life,
clear that visual experience early in life is extremely called the critical or sensitive period, the developing
important: The system is highly malleable or plastic mammals visual system undergoes a period of special
because its development depends on the type and vulnerability or susceptibility to the effects of many al-
amount of visual information the young, developing tered rearing conditions. As the critical period
organism receives; this property may prefigure the evolves into adulthood, the same manipulations of
kind of plasticity of the adult brain that enables hu- the animals visual experience generally do not affect
mans and animals to learn new behaviors. the physiological organization of the visual cortex in
terms of neurons response characteristics (although
some kinds of plasticity persist).
Basic Properties of Visual Cortical Neurons
For instance, early studies examined the effects of
Information from the two eyes converges on indi- completely depriving kittens of experience with visual
vidual neurons in the visual cortex; most of these neu- form or pattern by closing both eyelids during the
rons can respond to visual stimuli presented in either first few postnatal weeks. Initially it appeared that this
eye and so are termed binocular neurons. Many cells, manipulation does not produce major changes: Visu-
though, respond more strongly to stimulation of one al cortical cells binocularity and the distribution of
eye than of the other; the term ocular dominance refers ocular dominance are unaffected, and these kittens
to the relative influence of the two eyes on a cells re- cells also show orientation selectivity resembling that
sponse. Visual cortical neurons have another interest- of normally reared kittens, although the precision of
ing property: They are feature detectors. Almost orientation detection is slightly reduced. Subsequent-
every neuron responds best to a specific stimulus: ly, more detailed experiments showed that the system
usually a line, bar, or straight edge having a particular of disparity-detecting neurons does not develop nor-
orientation (e.g., horizontal, vertical, diagonal) and mally in binocularly deprived kittens.
precise location in the visual field. A third principal
characteristic of the responses of neurons in the visual By contrast, kittens raised with just one eyelid
cortex is the relationship between a binocular cells closed (monocular deprivation, or MD) show dramat-
preferred or optimal stimulus in each eye. Many cells ic changes in their visual cortical organization. Nearly
show interocular matching; that is, the stimulus orien- all the cells are responsive only to stimulation of the
tation that produces the best response is the same in eye that was open, and almost none to stimulation of
both eyes and has the same location within the visual the formerly closed eye. As might be expected, behav-
field. Many other cells have slightly different optimal ioral tests of visual acuity reveal deficits in the de-
stimulus orientations or locations in the two eyes. prived eye.
These small differences are called interocular dispari- This cortical ocular dominance shift is not due
ties, and neurons that exhibit this property are termed simply to the presence or absence of patterned visual
disparity detectors. input but to the absence of simultaneous stimulation
of both eyes, as shown by experiments in which kit-
tens were raised with one eye closed for several days
Plasticity in Visual Development or a few weeks, after which that eye was opened and
The question of how this cortical system for the the opposite eye was closed for a comparable period
analysis of visual space develops early in life has been (reverse suture), and also by experiments involving
the subject of hundreds of experiments since the alternating daily MD. Kittens raised using these
1960s. An important overall conclusion is that many methods experience the same amount of patterned
response characteristics of visual cortical neurons de- visual input through both eyes, but at any given mo-
pend on early visual experience for their normal de- ment only one eye is receiving stimulation. When cor-
velopment; these include binocularity and ocular tical ocular dominance and binocularity are studied
dominance, orientation selectivity, and interocular in these kittens, almost all neurons are visually re-
matching of cells optimal stimulus requirements, es- sponsive, but each responds only to stimulation of
pecially location and orientation (i.e., disparity detec- one eye; there are almost none of the binocular cells
tion). Typically, experiments of this sort involve rais- that make up 80 to 90 percent of the visual cortical
ing young animals (usually kittens or monkeys, whose neurons in normally reared kittens. Furthermore, if
visual systems resemble those of humans) for days or the relative amount (time) of stimulation given the
NEOCORTICAL PLASTICITY: Development of the Visual System 427
two eyes is made unequal in these experiments, there each eye), there is a disruption of binocularity: Most
is a corresponding change in the proportion of cells cells respond only to stimulation in one eye but not
activated by stimulation of each eye. in both eyes. In this respect the effects of large in-
terocular rotations are like those of strabismus or al-
Not only must patterned visual stimulation be
ternating monocular deprivation.
temporally synchronized (simultaneous) between the
two eyes in order for cortical binocularity to develop From an evolutionary standpoint, cortical plastic-
normally, it must also be spatially synchronous; that ity in the development of interocular relationships
is, each part of the pattern must stimulate the same has clear adaptive significance. The developing ani-
point on both retinas (corresponding points). Some mal undergoes relatively rapid changes in height and
humans exhibit an oculomotor disorder called strain the lateral separation of the two eyes. There is thus
bismus, in which the two eyes are misaligned. People a continually changing relation between the interocu-
with this disorder often have poor visual acuity in one lar image disparity of objects in the environment and
eye and almost always have deficient binocular depth the distance of those objects from the observer. The
perception. An animal model of strabismus, in which existence of neuronal disparity-detecting mecha-
some of the muscles that control the position of the nisms able to adjust to these changes during early life
eyes are severed, has proved useful in studying the provides an advantage in capturing prey, eluding
cortical effects of this disorder. Kittens raised with ex- predators, and so forth. Detailed reviews of the many
perimental strabismus show a marked loss of binocu- studies on this issue have been written by Frgnac and
lar neurons; in fact, the physiological organization of Imbert (1984) and by Shinkman, Isley, and Rogers
the visual cortex resembles that of kittens reared with (1985).
alternating monocular deprivation.
The system of orientation-detecting neurons in
the visual cortex is also susceptible to the effects of
Some Theoretical Considerations
early visual experience. Kittens raised viewing con- A major problem is identifying underlying mech-
tours or edges confined to a single orientation (e.g., anisms responsible for plasticity in the developing vi-
vertical stripes) have a preponderance of visual corti- sual system. One appealing idea involves binocular
cal cells whose preferred receptive fields are at or competition, in which fibers from the thalamus, some
near the experienced orientation; this finding is in carrying information from one eye and some from
marked contrast with the cortical organization of nor- the other, compete to form synapses on binocular cor-
mally reared kittens, in which all possible orientations tical cells. Although the underlying competitive
are represented about equally among cells receptive mechanism it is not yet clear, recent work points to a
fields. major role for nerve growth factor (NGF): rats with
intraventricular injections of NGF fail to show the ex-
In addition to the dramatic alterations in binocu- pected ocular dominance shift when subjected to MD
larity and orientation selectivity consequent upon ex- during the critical period (Pizzorusso and Maffei,
perimental manipulations of early visual experience, 1996). In any case, when researchers place one eye at
there is also a degree of plasticity in the development a disadvantage during MD, fibers representing the
of visual cortical cells specialized for disparity detec- other eye are more successful at making cortical con-
tion, especially interocular orientation disparity. For nections.
instance, there are changes in the visual cortices of
kittens raised wearing prism goggles that introduce At the same time, it is clear that neural activity
rotations of the images seen by the left and right eyes. originating in the deprived eye continues to play a
These rotations are around the visual axis (line of role in animals subjected to MD. Some researchers
sight), and are opposite in the two eyes, producing a have raised kittens with MD combined with damage
controlled amount of interocular orientation dispari- to a small area of the retina in the open eye. Later,
ty; that is, an edge or contour in the field of view does these kittens show the usual shift in ocular dominance
not give rise to parallel images on the two retinas, as toward the experienced eye, except that cortical cells
is normally the case, but instead is displaced clockwise that would otherwise have been responsive to stimula-
in one eye and counterclockwise in the other. If these tion of the damaged retinal area are instead respon-
rotations are small (e.g., eight degrees in each eye), sive to stimulation of the deprived eye. The important
there is a corresponding shift in the average disparity role of intracortical inhibitory mechanisms is consis-
of cortical neurons preferred receptive-field orienta- tent with this finding. For example, if a drug that
tions between the two eyes: Most cells show an in- blocks the action of inhibitory neurotransmitters is
terocular orientation disparity that matches the expe- administered to a previously monocularly deprived
rienced image rotation. On the other hand, if the kitten while recordings from cells in the visual cortex
rotations are large (e.g., sixteen degrees or more in are in progress, responsiveness to stimulation of the
428 NEOCORTICAL PLASTICITY: Development of the Visual System
deprived eye increases immediately; this effect con- ensues from the combination of a conditioned stimu-
tinues until the drug wears off and then disappears. lus and an unconditioned stimulus in a learning ex-
What, exactly, is the role of visual experience in periment. Indeed, drugs that block the NMDA recep-
neocortical development? Does it simply maintain the tor interfere with or even prevent the normal
feature-detecting capabilities and the interocular reacquisition of learned responses in experimental ani-
lationships of visual cortical neurons, or does it sharp- mals. As with NA and ACh depletion, this effect has
en and even alter these properties? Some researchers been obtained using several different kinds of condi-
have related this question to the age-old philosophi- tioning procedures.
cal issue of nature versus nurture; however, it is now In kittens, the pharmacological blocking of
clear that both genetic influences (nature) and the in- NMDA receptors also blocks the neural plasticity
fluences of the individuals unique visual environ- shown in the loss of cortical binocular cells following
ment (nurture) are crucial. The real question con- monocular deprivation. NMDA receptors are com-
cerns the relative degree of these influences on the posed of subunits that experimentation can selective-
development of the organization of our visual system ly alter; recent studies have begun to delimit specific
and of our perceptual capacities. Most researchers be- subunits responsible for various plastic changes in
lieve that early visual experience can, within limits, cortical neuronal function, both in vivo and also in
modify the formation of connections in the central vitro (Bear and Rittenhouse, 1999; Philpot et al.,
nervous system, thereby exerting substantial control 2001). Furthermore, plasticity that arises from the ac-
over the final outcome. This conclusion is borne out tions of neurotransmitters and from the activation of
both by studies using experimental animals and by NMDA receptors has been demonstrated at the cellu-
clinical observations on humans who have experi- lar level through iontophoresis, recording from a sin-
enced visual disorders in early childhood. gle neuron while using minute electrical currents to
eject small quantities of neurotransmitter substance
or of NMDA from the electrode into the immediate
Relation of Developmental Plasticity to vicinity of the neuron under study. The responses of
Learning and Memory in Adults many visual cortical cells to a visual stimulus pres-
ented to the nondominant eye, or at a nonoptimal
Candidate Pharmacological and
orientation, show a substantial temporary increase in
Neurochemical Mechanisms
strength when these stimuli are repeatedly paired
We are now beginning to understand some mech- with the iontophoretic application of NA and ACh, or
anisms that may underlie both neural plasticity in of NMDA and glutamate (an excitatory transmitter
early development and plasticity as manifested in that acts upon NMDA receptors). This effect is evi-
adult learning and memory. These may include some dent in kittens but not in adult cats. Thus these neuro-
dynamic aspects of the synaptic relations within neu- transmitter systems and the NMDA receptors have
ronal networks of the cerebral cortex and some been clearly implicated in neuronal plasticity early in
neurochemical, especially neurotransmitter, changes life and also in adult learning and memory.
that accompany (and may ultimately be responsible
for) some of the phenomena of neuronal plasticity de- Application of Transgenic Models
scribed above. For instance, the neurotransmitters Some researchers have examined the visual corti-
norepinephrine (NE) and acetylcholine (ACh) play a cal plasticity in genetically altered subjects. Gordon
critical role in the formation and maintenance of (1997) has provided an excellent review of results ob-
adult memories, and visual cortical plasticity is retained using transgenic mice. For instance, ocular
duced or abolished when levels of these neurotran- dominance shifts consequent upon MD occur in some
smitters are substantially depleted experimentally. mice bred to lack the gene encoding a particular form
Neurotransmitters exert many of their effects in of calcium/calmodulin-dependent protein kinase II.
the brain by acting on their receptors located on post- Thus calcium, with its well-known role in presynaptic
synaptic cells. There are numerous classes of recep- events, may help to signal activity-dependent plastici-
tors in the central nervous system; one notable type ty using this particular protein. This novel line of re-
is N-methyl-D-aspartate (NMDA) receptor, whose ac- search might yield major advances in our understand-
tion is voltage-dependent; its properties come into ing of molecular mechanisms of both adult learning
play only with some depolarization of the postsynap- and memory and developmental plasticity.
tic cell. It may therefore be a kind of gate, permitting
additional excitation only atop some excitatory effects Behavioral State
already present in the postsynaptic neuron. Activa- Various aspects of animal behavior that contrib-
tion of the NMDA receptor may thus be a neuro- ute to learning and memory appear to enable or en-
chemical analogue of the behavioral excitation that hance cortical plasticity. For instance, the well-
NEOCORTICAL PLASTICITY: Motor Cortex 429
studied effects of sleep upon memory consolidation al plasticity within the motor cortex and that this plas-
apply in much the same fashion to MD-induced plasticity represents the neural encoding of motor skill.
ticity in visual cortex (Frank, Issa, and Stryker, 2001).
Kittens that underwent anesthesia-induced sleep fol-
lowing MD exhibited more pronounced shifts in ocu- Motor Map Plasticity
lar dominance than kittens treated comparably but More than a century ago Fritsch and Hitzig
kept awake, either in a lighted environment or in (1870) found that electrical stimulation applied to re-
darkness. Perhaps other behavioral-state variables gions of frontal cortex could evoke movement and
can similarly affect cortical plasticity. Hitzig. The subsequent refinements of intracortical
microstimulation (Asanuma and Ward, 1971) have
Conclusion identified two basic principles of organization within
the motor cortex. The first is that movements corre-
There has been a dramatic increase in experi- sponding to body parts over which there is a high de-
mental attention to the relation between brain mech- gree of control, such as the hand, are multiply repre-
anisms of developmental plasticity and of learning sented over a larger area of cortex than those of less
and memory. The search for general mechanisms of dexterous body parts such as the foot. The second is
neural plasticity is likely to remain a central concern that, although the general location of the map within
of neuroscience, which will benefit especially from the cortex is constant, the particular organization of
studies of visual cortical neuronal networks. movement representations is dynamic and can
See also: GLUTAMATE RECEPTORS AND THEIR change in response to a variety of manipulations
CHARACTERIZATION (Sanes and Donoghue, 2000).
Bibliography These two principles of cortical organization also

Bear, M. F., and Rittenhouse, C. D. (1999). Molecular basis for in- appear to govern the encoding of skilled movement
duction of ocular dominance plasticity. Journal of Neurobiology within the motor cortex. Motor-skill training leads to
41, 8391. a reorganization of movement representations that
Frank, M. G., Issa, N. P., and Stryker, M. P. (2001). Sleep enhances expands the representations corresponding to
plasticity in the developing visual cortex. Neuron 30, 275287.
trained movements. For example, squirrel monkeys
Frgnac, Y., and Imbert, M. (1984). Development of neuronal se-
lectivity in primary visual cortex of cat. Physiological Reviews trained to retrieve a food pellet from successively
64, 325434. smaller food wells exhibit an expansion of wrist and
Gordon, J. A. (1997). Cellular mechanisms of visual cortical plastic- digit representations into elbow and shoulder repre-
ity: A game of cat and mouse. Learning and Memory 4, 245 sentations (Nudo, Milliken, Jenkins, and Merzenich,
261.
1996). Rats trained to reach outside their cages to re-
Philpot, B. D., Weisberg, M. P., Ramos, M. S., Sawtell, N. B., Tang,
Y.-P., Tsien, J. Z., and Bear, M. F. (2001). Effect of transgenic trieve a food pellet show a comparable expansion of
overexpression of NR2B on NMDA receptor function and wrist and digit representations (Kleim, Barbay, and
synaptic plasticity in visual cortex. Neuropharmacology 41, 762 Nudo, 1998). There is also evidence for a similar re-
770. organization following motor-skill learning in human
Pizzorusso, T., and Maffei, L. (1996). Plasticity in the developing
subjects (Pascuel-Leoni et al., 1995).
visual system. Current Opinion in Neurology 9, 122125.
Shinkman, P. G., Isley, M. R., and Rogers, D. C. (1985). Develop- Although these experiments show the concur-
ment of interocular relationships in visual cortex. In R. N.
Aslin, ed., Advances in neural and behavioral development, Vol. 1. rence of skill learning and map reorganization, they
Norwood, NJ: Ablex. do not reveal whether these changes result from
learning or drive the development of skill. A recent
Paul G. Shinkman
experiment examining the motor maps of rats at dif-
ferent times during skilled forelimb training suggests
that reorganization occurs after the acquisition of
motor skill. Significant expansion of wrist and digit
MOTOR CORTEX
movements requires ten days of skilled reach, train-
Nearly all adult behavior expresses an acquired ing whereas significant improvements in reaching ac-
motor skill. Because mammals have highly evolved curacy occurs after three days (Kleim et al., 2000).
motor regions of frontal cortex, including the prima- This temporal discrepancy seems to show that func-
ry motor cortex, they can acquire skilled motor be- tional reorganization requires sufficient performance
haviors involving a wide range of complex and adapt- of the skilled movements once they have been ac-
able movements. The remarkable adaptability of quired. Rats trained for three days followed by ten
mammalian motor behavior suggests a high degree of days without training do not exhibit any significant
functional flexibility in the motor cortex. There is change in the distribution of movement representa-
now a large body of evidence demonstrating function- tions (Hogg et al., 2001). However, simple movement
430 NEOCORTICAL PLASTICITY: Motor Cortex
repetition alone is not sufficient to drive changes in shown that the pattern of intracortical connections
the motor maps. In one experiment, squirrel mon- correlates with the functional reorganization that fol-
keys were trained daily to retrieve pellets from a large lows peripheral nerve lesions. Further, motor-skill
food well that does not require the development of training paradigms that induce motor map changes
skilled wrist movements but does involve extensive also cause changes in the morphology of cortical neu-
wrist and digit use; there was no significant map reor- rons. For example, cortical pyramidal neurons of rats
ganization despite thousands of movement repeti- trained on a skilled reaching task have a significantly
tions (Plautz, Milliken, and Nudo, 2000). Similarly, increased dendritic arbor within the trained versus
rats housed with running wheels for a month show no untrained motor cortex (Withers and Greenough,
significant change in the distribution of forelimb 1989; Greenough, Larson, and Withers, 1985). Fur-
movement representations (Kleim et al., 2002). Thus, ther, rats trained daily to traverse a complex set of ob-
motor map reorganization is dependent upon the ac- stacles have significantly more synapses per neuron
quisition and subsequent performance of novel within the motor cortex than rats forced to run a flat,
skilled movements but does not occur in response to obstacle-free runway (Kleim, Lussnig, Schwarz, and
repetition of existing movements. Greenough, 1996). Finally, Kleim et al. (2002b) have
shown that increases in synapse number within the rat
motor cortex following skilled reach training is con-
Neural Substrates of Motor Map Plasticity fined to those areas of cortex that also underwent re-
In the microcircuitry of the motor cortex there organization of movement representations.
are extensive recurrent axon collaterals that span sev- All of these data suggest that the development of
eral millimeters across the cortex (Huntley and Jones, skilled movement is encoded within the motor cortex
1991; Keller, 1993). Intracortical stimulation evokes through changes in the strength of intracortical affer-
movement by activating these horizontal afferents ents that may involve increases in synapse numbers.
(Jankowska, Padel, and Tanaka, 1975). Hence These adaptations of cortical circuitry then show up
changes in intracortical connectivity might mediate as changes in the distribution of cortical-movement
changes in the organization of movement representa- representations. The coordinated anatomical and
tions. Several experiments have supported this hy- physiological plasticity thus represents a neural
pothesis by demonstrating activity-dependent mechanism by which motor memories (engrams) are
changes in synaptic efficacy within the horizontal con- represented within the brain.
nections of the motor cortex via differential patterns
of electrical stimulation. Long-term potentiation See also: MOTOR SKILL LEARNING
(LTP) of synaptic responses occurs in response to
high frequency stimulation (Hess and Donoghue,
Bibliography
1994), whereas lower frequency stimulation leads to
Asanuma, H., and Ward, J. E. (1971). Patterns of contraction of dis-
long-term depression (LTD) of these same synapses tal forelimb muscles produced by intracortical stimulation in
(Hess and Donoghue, 1996). Although these changes cats. Brain Research 27, 97109.
are induced through artificial patterns of cortical Fritsch, G., and Hitzig, E. (1870). On the electrical excitability of
stimulation, they demonstrate that the horizontal the cerebrum, trans. G. von Bonin. In Some Papers on the Cere-
connections within the motor cortex are modifiable in bral Cortex. Springfield, MA: Thomas, 1960.
Greenough, W. T., Larson, J. R., and Withers, G. S. (1985). Effects
response to differential motor experience. of unilateral and bilateral training in a reaching task on den-
dritic branching of neurons in the rat sensory-motor forelimb
More direct evidence for the role of horizontal af-
cortex. Behavioral and Neural Biology 44, 301314.
ferent plasticity in motor learning has come from ex- Hess, G., and Donoghue, J. P. (1994). Long-term potentiation of
periments showing that synaptic potentials following horizontalconnections provides a mechanism to reorganize
stimulation of intracortical afferents were significant- cortical motor maps. The Journal of Neurophysiology 71, 2,543
ly greater in the trained than in the untrained hemi- 2,547.
(1996). Long-term potentiation and long term depression
sphere (Rioult-Pedotti, Friedman, Hess, and Donog-
of horizontal connections in motor cortex. Acta Neurobiologiae
hue, 1998). Further, relative to the untrained Experimentalis 56, 397405.
hemisphere, LTP was reduced and LTD was en- Hogg, T., Cooper, S., Vozar, D., VandenBerg, P. M., and Kleim,
hanced in the trained hemisphere (Rioult-Pedotti, J. A. (2001). Expansion of distal forelimb representations
Friedman, Hess, and Donoghue, 2000). These results within rat motor cortex is dependent upon performance and
not acquisition of skilled forelimb movements. Society for
demonstrate that motor learning increases intracorti-
Neuroscience Abstracts 27, 775.
cal synaptic efficacy while maintaining the range with- Huntley, G. W. (1997). Correlation between patterns of horizontal
in which synapses can be modified. connectivityand the extent of shortterm representational
plasticity in motor cortex. Cerebral Cortex 7, 143156.
Changes in the efficacy of intracortical afferents Huntley, G. W., and Jones, E. G. (1991). Relationship of intrinsic
may have an anatomical basis. Huntley (1997) has connections to forelimb movement representations in mon-
NEOCORTICAL PLASTICITY: Somatosensory Cortex 431
key motor cortex: A correlative anatomical and physiological fairly consistent organizations from individual to in-
study. Journal of Neurophysiology 66, 390413. dividual within a species and conform to a general
Jankowska, E., Padel, Y., and Tanaka, R. (1975). The mode of acti- plan across species. Yet the organizations of the maps
vation of pyramidal tract cells by intracortical stimuli. Journal
of Physiology 249, 617636. of the body surface can be altered in developing or
Keller, A. (1993). Intrinsic synaptic organization of the motor cor- adult mammals by removing or changing the signifi-
tex. Cerebral Cortex 3, 430441. cance of some of the inputs, or by damaging parts of
Kleim, J. A., Barbay, S., Cooper, N., Hogg, T., Reidel, C., Remple, maps. These changes in map organizations result
M., and Nudo, R. J. (2002b). Motor-learning dependent syn-
from neurons losing their original receptive fields
aptogenesis is localized to functionally reorganized motor
cortex. Neurobiology of Learning and Memory 77, 6377. and acquiring new ones on different body parts.
Kleim, J. A., Barbay, S., and Nudo, R. J. (1998). Functional reorga- Thus, the organization of somatosensory cortex can
nization of the rat motor cortex following motor skill learn- change. Best documented in primates and rats, such
ing. The Journal of Neurophysiology 80, 3,3213,325. changes have also been observed in humans.
Kleim, J. A., Cooper, N. R., and VandenBerg, P. (2002a). Exercise
inducesangiogenesis but does not alter movement representa- Plasticity in the somatosensory cortex may be im-
tions within rat motor cortex. Brain Research, 934, 16. portant in development by allowing the sensory sys-
Kleim, J. A., Hogg, T., Whishaw, I. Q., Reidel, C., Cooper, N., and tems of individuals to adjust to bodily growth through
VandenBerg,P. (2000). Time course and persistence of motor
learning-dependent changes in the functional organization of
the use of information from the environment. Plastic-
the motor cortex. Society of Neuroscience Abstracts, 27, 775. ity in adults may also be important in reassigning
Kleim, J. A., Lussnig, E., Schwarz, E. R., and Greenough, W. T. neurons to new roles after damage to the system in
(1996). Synaptogenesis and fos expression in the motor cor- order to promote recovery. Following amputations,
tex of the adult rat following motor skill learning. The Journal researchers have noted changes in patients soma-
Nudo, R. J., Milliken, G. W., Jenkins, W. M., and Merzenich, M.
tosensory cortex that often correlate with phantom
M. (1996). Use-dependent alterations of movement represen- pains. After a stroke, recovery of function is also relat-
tations in primary motor cortex of adult squirrel monkeys. ed to changes in cortical organization.
Pascual-Leone, A., Nguyet, D., Cohen, L. G., Brasil-Neto, J. P., The plasticity of the somatosensory cortex is
Cammarota, A., and Hallett, M. (1995). Modulation of muscle probably most important for allowing adjustments in
responses evoked by transcranial magnetic stimulation during the neural network that may be critical in learning
the acquisition of new fine motor skills. Journal of Neurophy- sensorimotor skills. Monkeys trained in various types
siology 74, 1,0371,045.
Plautz, E. J., Milliken, G. W., and Nudo, R. J. (2000). Effects of re-
of tactile discrimination tasks exhibit changes in the
petitive motortraining on movement representations in adult cortical organization of the trained skin surfaces that
squirrel monkey: Role of use versus learning. Neurobiology of reflect the temporal and spatial structure of the
Learning and Memory 74, 2755. trained stimuli. In humans, training for skilled motor
Rioult-Pedotti, M. S., Friedman, D., and Donoghue, J. P. (2000). behaviors evinces changes in SI. For example, trained
Learning-induced LTP in neocortex. Science 290, 533536.
Rioult-Pedotti, M. S., Friedman, D., Hess, G., and Donoghue, J. P.
violinists have larger SI representations of the hand
(1998). Strengthening of horizontal cortical connections fol- used to finger the instrument than that used to bow
lowing skill learning. Nature Neuroscience 1, 230234. it, and both representations are larger than in un-
Sanes, J. N., and Donoghue, J. P. (2000). Plasticity and primary trained subjects.
motor cortex. Annual Review of Neuroscience 23, 393415.
Withers, G. S., and Greenough, W. T. (1989). Reach training selec- Researchers have used two dominant experimen-
tively alters dendritic branching in subpopulations of layer II/ tal methods to examine mechanisms underlying so-
III pyramids in rat motor-somatosensory forelimb cortex. matosensory plasticity. In one, they induce plasticity
Neuropsychologia 27, 6169.
by cutting the sensory nerves to part of the skin or by
Jeffrey A. Kleim dorsal rhizotomy. For example, over several weeks,
neurons in the somatosensory cortex formerly activat-
ed by inputs from the denervated hand of a monkey
acquire new receptive fields on adjoining parts of the
SOMATOSENSORY CORTEX
hand that have intact sensory afferents. Some reacti-
The somatosensory cortex of mammals has a number vation may occur sooner, within seconds or hours of
of subdivisions or processing areas. Although the the nerve section. In the other method, they induce
number varies by species, each cerebral hemisphere plasticity by manipulating the patterns of use of the
in all mammals contains at least two systematic repre- skin surface, thus obviating the need for a destructive
sentations of the tactile receptors of the contralateral lesion. For example, trimming all but two adjacent
body surface, the primary somatic area SI, and the whiskers in a rat strengthens those whiskers abilities
secondary somatic area SII. Neurons throughout to excite SI, both in their own somatotopic territory
these representations can be activated by stimuli on and in that of the other untrimmed whisker. Overall,
restricted portions of the body surface, the receptive there are likely to be several different cellular mecha-
fields of the neurons. These two representations have nisms of cortical plasticity with different intervals.
432 NEURAL COMPUTATION: Approaches to Learning
It is not clear how the cortex reorganizes, but Plasticity of sensory representation in cortex has
there are several obvious possibilities. In both adult been most extensively studied for the somatosensory
and developing brains, plasticity, particularly its ini- cortex, but reorganizations occur after partial remov-
tial stages, seems to result from alterations in the ef- als of inputs for visual, auditory and motor areas as
fectiveness of anatomical connections. Indeed, well. Thus, plasticity is a basic feature of important
changes in synaptic efficacies of thalamocortical and systems throughout the brain.
intracortical projections have been detected after
Bibliography
whisker pairing in intact animals and in brain slices.
Buonomano, D. V., and Merzenich, M. M. (1998). Cortical plastici-
Such changes are likely based on Hebbian processes ty: From synapses to maps. Annual Review of Neuroscience 21,
such as long-term potentiation (LTP) and depression 149186.
(LTD), both of which can be elicited in SI. There is Ebner, F. F., Rema, V., Sachdev, R., and Symons, F. J. (1997). Ac-
more evidence for these types of mechanisms: Block- tivity-dependent plasticity in adult somatic sensory cortex.
Seminars in Neuroscience 9, 4758.
ade of the N-methyl-D-aspartate (NMDA) subtype of Jones, E. G. (2000). Cortical and subcortical contributions to activi-
glutamate receptor, which is necessary for the induc- ty-dependent plasticity in primate somatosensory cortex. An-
tion of cortical LTP, can also block plasticity of SI. nual Review of Neuroscience 23, 137.
Kaas, J. H. (1991). Plasticity of sensory and motor maps in adult
In addition to Hebbian processes, some immedi- mammals. Annual Review of Neuroscience 14, 137167.
ate recovery of cortical responsiveness might be due Jon H. Kaas
to a reduction in the lateral spread of inhibition of ex- Revised by Peter W. Hickmott
citatory pathways, a result of removal of a source of
activation for the inhibitory neurons. Other changes
might be due to a reduction in the production of in-
hibitory or neuromodulatory (such as acetylcholine,
norepinephrine, or serotonin) neurotransmitters in NEURAL COMPUTATION
the deprived zones of cortex. Over longer times, new
connections are likely to be formed, such as those ob- [Learning and memory emanate from the engagement of
served in the hindbrain nuclei and cortex. Increases complex biochemical and molecular networks within neu-
in the number, size, or strength of excitatory synaptic rons and complex interactions of neurons within neural net-
contacts probably play a major role, and this could in- works. Neural computation seeks to understand the ways in
volve the formation of new synapses and the elabora- which these various elements interact to process information
tion of axon arbors, dendrites, and dendritic spines. and generate behavior in general; and, with respect to learn-
ing and memory, the ways in which these processes give rise
The changes in the cortex reflect, in part, adjust- to the encoding, storage, and retrieval of memories. Five en-
ments made in the relays of sensory information in tries on neural computation follow. The first, APPROACHES
the brain stem and thalamus before they reach the TO LEARNING, provides an overview of the field. This entry
cortex. There is evidence for changes in the receptive is followed by specific entries on the neural computations that
fields of spinal cord neurons after section or anesthet- occur in each of four selected brain regions that are involved
ic block of peripheral nerves. Other changes, particu- in learning and memory: CEREBELLUM, HIPPOCAMPUS,
larly sprouting of new connections after early lesions NEOCORTEX, and OLFACTORY CORTEX.]
and transneuronal atrophy, have occurred in the
hindbrain and thalamus and undoubtedly contribute
to cortical reorganization. Probably all the mecha-
nisms cited above underlie the plasticity in these sub- APPROACHES TO LEARNING
cortical somatosensory areas. There is evidence for Nervous systems are capable of solving extraordinari-
purely cortical changes, however, including observa- ly sophisticated computational problems. The visual
tions of receptive field changes in cortex that are not or tactile recognition of an object in a cluttered scene
observed in thalamus, reorganization of SII after paris childs play, but well beyond the capability of the
tial lesions of SI (which provides the only input to SII), fastest digital computers. Most animals can navigate
and changes in isolated slices of cortex after reorgani- over rough surfaces with great agility, but present-day
zation. Interestingly, rats implanted with multielect- robots are limited in their movements to a very nar-
rode arrays demonstrate rapid changes at all levels of row range of terrains. We can learn to use language
the somatosensory pathway in response to anesthetic and to read and write well beyond anything so far ac-
block of peripheral nerves; the dynamics of these complished by artificial intelligence. We take all of
changes suggest that the changes are relayed from the these abilities for granted because we are so good at
cortex to thalamus and hindbrain, not just in the as- them; trying to duplicate them with machines has
cending direction. made their great difficulty more apparent.
NEURAL COMPUTATION: Approaches to Learning 433
Neural computation is the systematic study of the influences on learning may be more diffuse and wide-
computational principles underlying the function of spread.
neural systems, from the level of molecular mecha- Neurons have limited dynamic range. Unlike dig-
nisms to the organization of brain systems (see Figure ital systems, which are capable of accurately repre-
1). This computational approach to neuroscience is senting very large and very small numbers, the range
still in its infancy (Sejnowski, Koch, and Churchland, of membrane potentials and firing rates found in neu-
1988). There has been a recent emphasis on studying rons is limited. Also, the variability in the properties
neural networks, small groups of highly connected of neurons within the same population is significant,
neurons; however, as shown in Figure 1, neural net- and the properties of the same neuron can vary with
works are only one level of investigation in the ner- time. The same is true for analog VLSI (very-large-
vous system, and neural computation depends on scale integration) circuits that are designed to mimic
computational principles at each of these levels. A few the processing that occurs in neurons (Mead, 1989).
general principles have emerged from the study of This variability and limited dynamical range have
abstract models of neural systems that are likely to be consequences for the way that information is coded
important for the biological study of learning and and the way that neural circuits are designed. One
memory. way to preserve information is to process relative le-
Some Principles of Neural Computation vels, or differences, rather than absolute levels. Thus,
visual neurons are more sensitive to contrast (spatial
In the von Neumann architecture commonly
differences) and changes (temporal differences) than
used in digital computers, the memory and the pro-
to absolute intensity levels. Another mechanism for
cessor are physically separated. This separation gives
preserving information is dynamically altering base-
rise to a bottleneck in the flow of information between
line activity in neurons. Adaptive biochemical mecha-
the two. In neural systems, memory and processing
nisms inside cells, such as light adaptation in photore-
are intertwined; the same circuits that process sensory
ceptors, allow neurons to remain in their most
and motor information are involved in learning and
sensitive range. Adaptive mechanisms have been
the storage of new information. A unified processor-
found for calibrating sensorimotor coordination,
memory system allows many circuits to work together
such as slow adaptation of the vestibulo-ocular reflex
in parallel and, as a consequence, the solutions to
(VOR) to changes in the magnification of the lens
many commonly occurring problems can be comput-
(Lisberger, 1988).
ed in only a few serial steps. The representation of
sensations and memories in such an architecture is Taxonomy of Learning Systems
more difficult for us to imagine and to use than one Adaptation to ongoing sensory stimulation does
in which the functions are segregated (Churchland not require an additional source of information out-
and Sejnowski, 1992). The brain, however, did not side the processing stream; this type of learning is
evolve to make it easier for us to analyze. called unsupervised. In contrast, the type of adaptation
Locality is an important constraint that arises that occurs in response to sensorimotor mismatch
when hardware for artificial neural networks is de- does require an outside error signal; this is called su-
signed (Mead, 1989). Wires are expensive on comput- pervised learning. In the case of VOR learning, the
er chips, just as they are in the brain, so only limited error signal is the slip of the image on the retina, and
connectivity is possible between processing elements. the gain of the reflex is changed to reduce the slip.
The organization of sensory processing into a hierar- The amount of supervision can vary from a crude
chy of maps and the laminar organization of cortical good/bad reinforcement signal to very detailed feed-
structures is wire-efficient. This also places con- back of information about complex sensory signals
straints on the organization of learning systems, from the environment that might be termed a teach-
which share the same circuitry. In particular, the deci- er. Supervised learning is sometimes called error-
sion to store a piece of information at a particular lo- correction learning.
cation in the brain is a local one that depends on elec- It is not necessary for the error signal to come
trical and chemical signals that are spatially and from outside the organism; important information
temporally restricted. The Hebbian mechanism for about the proper operation of a circuit can be provid-
synaptic plasticity that has been found in the hippo- ed by another internal circuit, or even by internal con-
campus and neocortex obeys this principle of locality: sistency within the same circuit. For example, a senso-
The presynaptic release of neurotransmitter and the ry area that was trying to predict future inputs could
postsynaptic depolarization needed to trigger the compare its prediction against the next input to im-
long-term potentiation at these synapses are spatially prove its performance. Such an unsupervised system
contiguous, and there is a brief temporal window dur- with an internal measure of error is termed monitored
ing which both signals must be present. Modulatory (Churchland and Sejnowski, 1992). As shown in Fig-
434 NEURAL COMPUTATION: Approaches to Learning
ure 2, all possible combinations of supervised and un- Selected Examples

supervised, monitored and unmonitored, learning An interesting example of a monitored system is
systems are possible. song learning in the white-crowned sparrow. In this
NEURAL COMPUTATION: Approaches to Learning 435
species of songbird, the male hears the local dialect

after hatching but does not produce the song until the
next spring. At first the song is imperfect, but with
each repetition the details improve until it is a good
reproduction of the original song heard the previous
year, even though there is no external teacher dur-
ing the refinement. The internally stored template is
compared with the imperfect song production; the
error between them drives learning mechanisms to
improve the song. This learning is monitored because
the error is derived from an internal template of the
desired sound. We may use a similar strategy when
learning to produce new sounds in a foreign lan-
guage.
Transformations between two populations of
neurons can be learned with Hebbian mechanisms at
the synapses between the input fibers and the output
neurons. The pattern of activity on the input fibers is
matched with the desired pattern on the output neu-
rons. In some models of the cerebellum, for example,
associative motor learning is mediated by climbing fi-
bers, which provide a teaching signal to the output
Purkinje cells. By including feedback projections of
the output neurons back onto themselves, a partial
features that separate classes (Churchland and Sej-
input cue can regenerate the entire stored pattern. In
nowski, 1992).
this mode, the system is unsupervised because the de-
sired output pattern of activity during learning is the As computers grow more powerful, it will become
same as the input pattern. Such content-addressable possible to simulate more complex models of neural
recurrent networks have been suggested as models systems; however, even these simulations will fall
for the piriform cortex and area CA3 of the hippo- short of the richness of real neural systems and the
campus. Some properties such as memory capacity of complex environments that confront biological crea-
associative networks of simplified processing units tures. Hardware emulations that interact with the real
have been well studied; the analysis of networks based world in real time would greatly improve our ability
on model neurons with more complex properties is to test hypotheses about the organization of neural
just beginning. systems (Mead, 1989). Ultimately we will need to
study complete model systems in order to understand
Learning mechanisms have also been used to the multiple levels of adaptation and learning that
model the development of cortical systems. One of provide flexibility and stability in a changing world.
the best-explored areas of unsupervised learning in
artificial networks is competitive learning, in which
See also: BIRDSONG LEARNING; HEBB, DONALD;
incoming sensory information is used to organize the
NEOCORTICAL PLASTICITY
internal connections of a sensory map. For example,
the formation of ocular dominance columns in visual
cortex of cats and monkeys during development de- Bibliography
pends on competitive synaptic mechanisms. The de- Churchland, P. S., and Sejnowski, T. J. (1992). The computational
brain. Cambridge, MA: MIT Press.
velopment of ocular dominance columns can be mim- Kohonen, T. (1984). Self-organization and associative memory. Berlin:
icked in a computer model that uses Hebbian Springer-Verlag.
learning in the spatially restricted terminal arbors of Lisberger, S. G. (1988). The neural basis for motor learning in the
axons projecting to the cortex from the lateral genic- vestibulo-ocular reflex in monkeys. Science 242, 728.
ulate nucleus (Miller et al., 1989). Similar mecha- Mead, C. (1989). Analog VLSI and neural systems. Reading, MA: Ad-
dison-Wesley.
nisms can also be used in neural systems to learn Miller, K. D., Keller, J. B., and Stryker, M. P. (1989). Ocular domi-
more complex features that distinguish among differ- nance column development: Analysis and simulation. Science
ent types of sensory inputs (Kohonen, 1984). It is also 245, 605615.
likely that other learning mechanisms are used to dis- Sejnowski, T. J., Koch, C., and Churchland, P. S. (1988). Computa-
tional neuroscience. Science 241, 1,2991,306.
cover invariants of sensory patterns, which are often
as important in pattern recognition as the distinctive Terrence J. Sejnowski
436 NEURAL COMPUTATION: Cerebellum
CEREBELLUM Each Purkinje cell also normally receives a single

climbing fiber exclusively from the inferior olive.
The cerebellum is located at the rear of the brain Climbing fibers project to the cerebellar cortex, form-
above the brain stem. It is connected to the spinal ing strong synaptic connections with Purkinje cells,
cord, brain stem, and cerebral cortex. The function and send collaterals to the cerebellar nuclei. Climb-
of the cerebellum has been the topic of classic studies ing-fiber input occurs at a relatively low frequency
of the effects of lesions of the cerebellumsuch as (around 1 hertz) and over a narrow dynamic range
discoordination and dysmetriain animals and but powerfully excites Purkinje cells, causing them to
human patients. These symptoms suggest that the generate bursts of three to four high-frequency spikes
cerebellum helps us to move precisely, smoothly, (complex spikes).
quickly, and even without visual feedback. Because
learning is essential to the acquisition of such skills, A theory propounded in 1970 indicated that con-
junctive activation of a mossy-fiber/parallel-fiber
one can infer that the cerebellum learns by itself, al-
pathway and a climbing fiber onto a Purkinje cell ei-
though it is possible it connects to a learning mecha-
ther strengthens or weakens the transmission from
nism elsewhere in the brain.
the parallel fiber to the Purkinje-cell synapse. Climb-
A detailed analysis of neuronal circuits in and ing-fiber signals will thus modify the signal flow
around the cerebellum (see Figure 1) has yielded net- through the cerebellar cortex. This operation is anal-
work models, modular models, and control-system ogous to that in which all connections from the sec-
models of the cerebellum to represent its computa- ond to the third layer are strengthened or weakened
tional capabilities. The cerebellar cortex is modeled by an outside teacher who recognizes correct or incor-
by the three-layered simple perceptron, which, in rect performance of the simple perceptron. Indeed,
combination with nuclear structures, forms a func- long-term depression (weakening) of synaptic
tional module of the cerebellum, a cerebellar corti- strength (LTD) was discovered a decade later. The oc-
conuclear microcomplex. The microcomplex pro- currence of LTD suggests that the cerebellar cortex
vides an adaptive control mechanism to a spinal-cord/ learns by means of weakening connections responsi-
brain-stem function. In cortical sensorimotor ble for erroneous performance.
functions, it provides an internal model that helps In the simple-perceptron model, the cerebellum
feedforward control. In mental thought processes in processes spatial information; in the adaptive-filter
the association cortex as well, it seemingly plays a sim- model, it processes temporal patterns. In the adap-
ilar role. tive-filter model, the inhibitory connection between
Golgi cells and granule cells constructs an integrator
with a long time constant, which activates granule
Neuronal Network in the Cerebellar Cortex cells with varied latencies. A mossy-fiber signal would
The cerebellar cortex is made up of five principal then be converted to time-scattered parallel-fiber sig-
types of neurons (Purkinje, basket, stellate, Golgi, and nals from which Purkinje cells select an appropriate
granule cells) that are arranged in a very uniform pat- temporal pattern learned through climbing-fiber sig-
tern. The cerebellar cortex receives mossy fibers from nals.
various precerebellar nuclei, which also supply excit-
atory synapses to cerebellar nuclei by way of their col- Functional Module of the Cerebellum:
laterals. Input from the mossy fibers is relayed to the
Microcomplex
granule cells, which send out bifurcating axons, called
parallel fibers, that run in beams. The axons of the The cerebellar cortex is organized into seven lon-
granule cells supply excitatory synapses to Purkinje gitudinal (A, B, C1, C2, C3, D1, and D2) zones. Each
cells, which provide the sole output of the cerebellar zone sends Purkinje cell axons to a certain cerebellar
cortex. Each Purkinje cell receives as many as 175,000 or vestibular nucleus. Thus, zone A projects to fastigi-
inputs from the granule cells. Thus, the connections al and vestibular nuclei, zone B to vestibular nuclei,
from mossy fibers to granule cells to Purkinje cells are zones C1 and C3 to the rostral part of the interpositus
the major signal-flow pathway, allowing the distribu- nucleus and zone C2 to the caudal part of it, and
tion of input information received by mossy fibers to zones D1 and D2 to the medial and lateral parts of the
numerous granule cells from which Purkinje cells se- lateral (dentate) nucleus.
lect information to generate their output signals. Each longitudinal zone of the cerebellar cortex is
Normally, Purkinje cells fire simple spikes at approxi- composed of a number of microzones. Each micro-
mately 50 hertz. The discharge frequency of simple zone projects to a small group of vestibular or cerebel-
spikes are modulated in response to mossy-fiber lar nuclear neurons and receives climbing fibers from
input. a small group of inferior olive neurons, which project
NEURAL COMPUTATION: Cerebellum 437
collaterals to a small group of nuclear neurons pro- Figure 1

jected by the microzone. A microcomplex is an inter-
connected set of a microzone, a small group of nucle-
ar neurons, and a small group of inferior olive
neurons. The microzones defined in the paravermis
and flocculus may measure about 10 mm2. In rats,
one microzone of this size contains about 10,000 Pur-
kinje cells and 2,740,000 granule cells. The human
cerebellum is about 50,000 mm2 wide, so that it may
contain as many as 5,000 microzones as its functional
unit.
The microcomplex would function as a module of
the cerebellum in the following manner. First, input
signals from a precerebellar nucleus (except those
from the inferior olive) drive the nuclear neurons,
which generate output signals of the microcomplex
under inhibitory influences of Purkinje cells. Second,
the same input signals pass via mossy fibers to a
microzone, where they are relayed by granule cells
and in turn excite Purkinje cells and other cortical
neurons, eventually evoking simple spikes in Purkinje
cells. Simple spike discharges of Purkinje cells driven
by moss-fiber signals produce a unique functional
state of a microzone arising from concerted activities
of excitatory and inhibitory synapses. Third, climbing
fibers convey error signals pertaining to the opera-
tion of the neural system that includes the micrcom-
plex, as generated by various neuronal mechanisms
in diverse preolivary structures. Fourth, climbing-
trol systems in the spinal cord and brain stem, which
fiber error signals induce LTD in the conjointly acti-
are converted to adaptive control systems by an at-
vated parallel fiber-Purkinje cell synapses (learning
tached microcomplex. Typical examples are the vesti-
rule) and thereby modify the operation of the micro-
buloocular reflex and eyeblink conditioning. Com-
complex until the error signals are minimized.
pound-movement systems such as those for
Climbing-fiber signals evoke complex spikes in Pur-
locomotion or saccades are control systems equipped
kinje cells and thereby induce conducting impulses in
with a function generator. Innate behavior such as
Purkinje cell axons, which eventually evoke IPSPs in
food intake, drinking, and reproductive activity is a
the nuclear neurons. However, the effects of the
form of control involving a motor program. The
IPSPs on nuclear neurons are counteracted by the
microcomplex introduces adaptability into these con-
EPSPs evoked via collaterals of olivocerebellar fibers.
trol systems.
The signal content of complex spikes has been ana-
lyzed in various motor behaviors and is related partly When a sensorimotor cortex develops, the micro-
to consequence errors and partly to internally com- complex forms a cerebrocerebellar communication
puted errors. loop, which may provide an internal model of the
The postulated operation of the microcomplex, skeletomuscular system to be controlled by the motor
including inhibitory neurons in the cerebellar cortex, cortex. Command signals generated by the motor
has been computer-simulated. However, since several cortex may perform precise control using an internal
forms of synaptic plasticity other than conjunctive feedback through this internal model of the cerebel-
LTD were observed in the cerebellum, further studies lum even without referring to the external feedback
are needed to reproduce the performance of a micro- (see Figure 2A). In the Smith predictor model, the in-
complex that incorporate these forms of synaptic ternal model not only represents dynamics of the con-
plasticity. trol object but also incorporates the delay time to be
spent in the external feedback, so that it reproduces
exactly the same effect as the external feedback. The
Roles in Neural Control microcomplex may also be attached parallel to the
The microcomplex gives an extracerebellar sys- motor cortex and provides an inverse model of the
tem an adaptive mechanism. Reflexes are classic con- skeletomuscular system, which also allows the control
438 NEURAL COMPUTATION: Cerebellum
Figure 2 level of neuronal circuit, movement and thought

could be controlled by the same mechanism. The the-
sis of the involvement of the cerebellum in mental ac-
tivities was based on anatomical connections between
the association cortex and cerebellar hemisphere.
Brain-imaging studies have afforded evidence of an
internal model.
See also: GUIDE TO THE ANATOMY OF THE BRAIN;
LONG-TERM DEPRESSION IN THE CEREBELLUM,
HIPPOCAMPUS, AND NEOCORTEX; NEURAL
SUBSTRATES OF CLASSICAL CONDITIONING:
DISCRETE BEHAVIORAL RESPONSES; VESTIBULO-
OCULAR REFLEX (VOR) PLASTICITY
Bibliography
Albus, A. (1971). A theory of cerebellar function. Mathematical Bio-
sciences 10, 2561.
Blakemore, S. J., Frith, C. D., and Wolpert, D. M. (2001). The cere-
bellum is involved in predicting the sensory consequences of
action. NeuroReport 12, 1,8791,884.
Dow, R. S., and Moruzzi, G. (1958). The physiology and pathology of
the cerebellum. Minneapolis: University of Minnesota Press.
Fujita, M. (1982). Adaptive filter model of the cerebellum. Biologi-
cal Cybernetics 45, 195206.
Groenewegen, H. J., and Voogd, J. (1977). The parasagittal zona-
tion within the olivocerebellar projection. I. Climbing fiber
distribution in the vermis of cat cerebellum. Journal of Compar-
ative Neurology 174, 417488.
Groenewegen, H. J., Voogd, J., Freedman, S. L. (1979). The para-
sagittal zonation within the olivocerebellar projection. II.
Climbing fiber distribution in the intermediate and hemi-
spheric parts of cat cerebellum. Journal of Comparative Neurolo-
gy 183, 551601.
Hansel, C., Linden, D., and DAngelo, E. (2001). Beyond parallel
fiber LTD: The diversity of synaptic and nonsynaptic plastici-
ty in the cerebellum. Nature Neuroscience 4, 467475.
Imamizu, H., Miyauchi, S., Tamada, T., Sasaki, Y., Takino, R.,
A. Seemingly feedforward control system with an internal Futz, B. Yoshioka, T., and Kawato, M. (2000). Human cere-
feedback through a forward model. AM (actual movement) bellar activity reflecting an acquired internal model of a new
becomes equivalent to IM (instructed movement) if G (signal tool. Nature 403, 192195.
transfer characteristics of the forward model) = G (that of the Ito, M. (1984). The cerebellum and neural control. New York: Raven
actual hand/arm system), even when the external feedback is Press.
interrupted. g is the signal transfer characteristics of the cerebral (1993). Movement and thought: Identical control mecha-
nisms by the cerebellum. Trends in Neuroscience 16, 448450.
cortex. Originally proposed by Ito, M. (1984). The cerebellum and
(2001). Long-term depression: Characterization, signal
neural control. New York: Raven Press. B. Two-degrees-of-
transduction and functional roles. Physiological Reviews 81,
freedom control system. Originally proposed by Kawato, M., 1,1431,195.
Furukawa, K., and Suzuki, R. (1987). A hierarchical neuronal Kawato, M., Furukawa, K., and Suzuki, R. (1987). A hierarchical
network model for control and learning of voluntary movement. neuronal network model for control and learning of voluntary
Biological Cybernetics 57, 169185. f(G) represents the signal movement. Biological Cybernetics 57, 169185.
transfer characteristics of the inverse model. If f(G) = 1/G, Leiner, H. C., Leiner, A. L., and. Dow, R. S. (1986). Does the cere-
AM=IM even when the external feedback is interrupted. CM, bellum contribute to mental skill? Behavioral Neuroscience 100,
command signals. 443453.
Marr, D. (1969). A theory of cerebellar cortex. Journal of Physiology
202, 437470.
Miall, R. C., Weir, D. J., Wolpert, D. M., and Stein, J. F. (1993). Is
to be performed without an external feedback (see the cerebellum a Smith Predictor? Journal of Motor Behavior
Figure 2B). 25, 203216.
Oscarsson, O. (1976). Spatial distribution of climbing and mossy
The control with a forward or inverse model fibre inputs into the cerebellar cortex. In O. Creutzfeldt, ed.,
shown in Figure 2 can be generalized to the thought Afferent and intrinsic organization of laminated structures in
the brain. Experimental Brain Research Supp. 1. Berlin: Spring-
in which we move images, concepts, or ideas repre- er-Verlag.
sented in the temporoparietal cortex by command Rosenblatt, F. (1962). Principles of neurodynamics: Perceptron and the
signals generated by the prefrontal cortex. At the theory of brain mechanisms. Washington, DC: Spartan Books.
NEURAL COMPUTATION: Hippocampus 439
Schweighofer, N., Arbib, M., and Dominey, P. F. (1996). A model several clear principles as to how actual neural circuits
of the cerebellum in adaptive control of saccadic gain. I. The might accomplish this.
model and its biological substrate. Biological Cybernetics 75,
1928. The essence of these principles is illustrated in
Schweighofer, N., Arbib, M. and Kawato, M. (1998). Role of the Figure 1, which may be thought of as an incomplete
cerebellum in reaching movements in humans. I. Distributed
inverse dynamics control. European Journal of Neuroscience 10,
and crude model for hippocampal regions CA3 and
8694. fascia dentata, and their neocortical inputs. The
axons from the granule cells of the fascia dentata
Nelson Donegan
make sparse but strong contacts with the CA3 pyrami-
Revised by Masao Ito
dal cells. The axons of the pyramidal cells feed back
into the pyramidal layer, making contacts that are ini-
tially ineffective but can be made effective by imple-
HIPPOCAMPUS menting Hebbian synaptic enhancement. That is,
whenever a pyramidal cell is strongly activated by its
Both clinical neuropsychological studies and animal
granule cell input, those synapses that it receives from
experiments involving damage to the hippocampal
other pyramidal cells that have been activated by the
formation indicate that this structure plays a funda-
same input become strengthened (in this illustration,
mental role in at least the initial establishment of
the strength goes from 0 to 1). By strengthening the
long-term associative memory; however, the exact
connections among neurons that have been active to-
role of the hippocampus in associative memory, and
gether during an event, it becomes possible later to
the physiological and computational mechanisms by
recall that event in its entirety, given only a fragment
which this role is accomplished, remain subjects of in-
of it. One of Marrs important contributions was the
tense study and debate. Although by no means prov-
en, evidence favors the hypothesis that the hippocam- idea that a small population of inhibitory interneu-
pus acts as a simple interim repository for memories rons plays the crucial role of assessing the total num-
of certain kinds of events, and that other (neocortical) ber of active inputs and adjusting the threshold of the
circuitry draws on this repository during a process memory (pyramidal) cells such that only a fixed pro-
known as memory consolidation (e.g., Zola-Morgan portion of them are allowed to fire. This adjustment
and Squire, 1990). The following is a brief overview is accomplished essentially by dividing the total exci-
of some current ideas about how the unique circuitry tation of a given cell by the total number of active in-
of the hippocampal formation might enable rapid as- puts. In this way, when an incomplete event is pres-
sociative memory, and why such an interim repository ented, the effective threshold of the pyramidal cells
might be necessary. is lowered (because there is less inhibition). If the re-
duced input is part of a stored event, all of the corre-
Specific events are generally represented in the sponding synapses will have been strengthened and
nervous system as distributed patterns of activity within the correct cells will fire. Incorrect cells will not fire
rather large populations of cells, and rarely by the ac- because they will, on average, not have as many en-
tivity of single cells. The activity pattern may be hanced inputs in that particular event. Thus, provid-
thought of as a vector, that is, a list of ones and zeros, ed it is unique, even a small fragment of a stored
indicating which neurons are firing and which are si- event will cause activation of the full event. The read-
lent, or as a list of positive real numbers indicating the er unfamiliar with these principles might benefit by
firing rates of the neurons over some short interval. working through the example in Figure 1. For more
Associative recall, by its most general definition, is the detailed discussion, see McNaughton and Nadel
ability of the brain to reconstruct the vector corre-
(1989) or McNaughton and Barnes (1990).
sponding to a stored event when presented with a vec-
tor that is missing some of the original information, There are, of course, constraints on the amount
has been corrupted somehow by noise, or merely (and kinds) of information that can be stored by such
bears some significant resemblance to the original. a memory, as can readily be understood by imagining
This ability is often called either pattern completion or the consequences of attempting to store so much in-
autoassociative recall. Using the pre- and postsynaptic formation that all of the synapses have been en-
conjunction principle for modifying the connection hanced. At this point no information is recoverable.
strengths between neurons originally elaborated by One solution is to encode an event with the minimum
Donald Hebb (1949), work in the 1960s and early number of fibers necessary to capture its vital features
1970s by Kohonen (1972), Mart (1971), Steinbuch (sometimes called redundancy removal). The hippo-
(1961), Willshaw and colleagues (1969), and others campus constitutes the highest level of association
laid the theoretical foundations for how a simple pat- cortex in the nervous system, receiving its input from
tern-completion network might operate. In particu- polymodal association cortex. Much of the processing
lar, the work of Marr was seminal, because it outlined that occurs in these areas can be thought of as redun-
440 NEURAL COMPUTATION: Hippocampus
Figure 1
dancy removal or feature extraction. Moreover, hippo- will not work. An alternative solution is to use an extra
campal cells are often silent for prolonged periods, layer of very many cells between the input and the
suggesting that rather few of them are active at any memory cells. By spreading the connections from the
one time. Another solution is to make the patterns to input to this layer more or less randomly, yet adjust-
be stored as different from each other as possible, a ing the thresholds so that about the same total number
process known as orthogonalization, because orthogo- of intermediate cells as input cells are active, the de-
nal vectors are uncorrelated. gree of overlap between input events will be reduced.
This is merely a consequence of the facts that, in the
Although feature extraction is itself an orthogon- larger population, the probability that a given cell is
alization process, sometimes it is necessary to store as active in any one event is reduced, and the probability
separate, one-time events that may differ only in of its being active in any two events is the square of
some arbitrary but important way. Because feature the single-event probability. Marr called this solution
detector formation requires knowledge of the long- codon formation. It is likely that the granule cells of the
term statistical regularities in the input, this solution fascia dentata perform something like this function.
NEURAL COMPUTATION: Neocortex 441
Although greatly oversimplified, models such as subset of pyramidal cells and their recurrent collater-
this go a long way toward accounting for the known als (2). Each pyramidal cell then adds up how many
anatomical and physiological organization of the hip- of its currently active recurrent collateral synapses
pocampal formation. Many theoretical and experi- have previously been enhanced, and divides this by
mental neuroscientists believe that something like the total number of active inputs. If this is equal to or
this process goes on in the hippocampus during the greater than threshold (i.e., 1), the unit fires. If not
original encoding of long-term associative memory. too many patterns are stored, the result will be the
One of the major outstanding questions in the field, output of the complete original pattern (3). The in-
however, is why the hippocampus, and presumably hibitory interneuron sums the total input and sets the
the information stored there, is necessary only for a divisor for the pyramidal cells accordingly. There is
limited period after the initial registration. Estimates some evidence that something of this sort actually
of exactly how long vary from hours to years in differ- happens in the brain.
ent species, and for different kinds of memory. At
present only educated guesses can be offered in an- See also: PARALLEL DISTRIBUTED PROCESSING
MODELS OF MEMORY
swer to this question. The first is that it is probably
both unnecessary and biologically expensive simply Bibliography
to store all experience; however, often it cannot be Hebb, D. O. (1949). The organization of behavior. New York: Wiley.
predicted at the time of an event whether the infor- Kohonen, T. (1972). Correlation matrix memories. IEEE Transac-
mation is sufficiently important or reliable to be tions on Computers C-21, 353359.
stored permanently. Second, if a set of events does Marr, D. (1971). Simple memory: A theory for archicorrex. Philo-
sophical Transactions of the Royal Society of London B262, 2381.
contain some statistical regularity, it may be possible McNaughton, B. L., and Barnes, C. A. (1990). From cooperative
to generate a new feature detector for that regulari- synaptic enhancement to associative memory: Bridging the
ty, and hence to achieve redundancy reduction in abyss. Seminars in the Neurosciences 2, 403416.
permanent memory. In order to assess this, however, McNaughton, B. L., and Nadel, L. (1989). Hebb-Marr networks
some representation of the raw events must be stored and the neurobiological representation of action in space. In
M. A. Gluck and D. E. Rumelhart, eds., Neuroscience and con-
for comparison. J. L. McClelland (personal communi-
nectionist theory. Hillsdale, NJ: Erlbaum.
cation) has proposed a somewhat related hypothesis, Steinbuch, K. (1961). Die Lernmatrix. Kybernetik 1, 3645.
expressed in terms of the parallel distributed process- Willshaw, D. J., Buneman, O. P., and Longuet-Higgins, H. C.
ing or connectionist models of cognitive psychology. (1969). Nonholographic associative memory. Nature 222,
Briefly, the argument is that learning appropriate 960962.
Zola-Morgan, S. M., and Squire, L. R. (1990). The primate hippo-
and efficient internal representations for a particular campal formation: Evidence for a time-limited role in memo-
problem or set of events generally requires multiple ry formation. Science 250, 288289.
exposures to the events (at least with currently discov-
Bruce L. McNaughton
ered algorithms). During these exposures a kind of
global error term for the performance of the memory
can be computed and used to correct (in small steps)
the set of synaptic connection strengths (weights). The
NEOCORTEX
argument, then, is that the hippocampus makes use One of the central goals of neuroscience research is
of many cells to store a set of memories that later can to understand the nature of the computations carried
be played back to the neocortex for the purpose of out in the neocortex. The neocortex is a thin sheet of
computing an appropriate set of connection weights roughly 1010 neurons and fibers that forms the exter-
to store the information with the minimal number of nal surface of much of the brain. The neocortex arose
units. Ideas such as these, it is hoped, will eventually recently in evolution, around the time that mammals
provide a firm computational explanation for the branched off from reptiles. Yet over the past 60 mil-
process of memory consolidation and the crucial role lion years, the neocortex has shown the most dramat-
of the hippocampal formation. ic expansion and elaboration of any brain system. In
A simple (and incomplete) model for the fascia fact, differences in the neocortex best distinguish the
dentata and CA3 subfields of the hippocampus illus- human brain from those of other mammals. The neo-
trates how this system may implement autoassociation, cortex is involved in higher perceptual and cognitive
using simple Hebbian synaptic enhancement. In- processing and contains areas specialized for decision
puts from fascia dentata essentially impose output making, executive function, sensory perception (vi-
patterns in CA3. The synapses of the activated recur- sion, hearing, touch, and smell), generating action,
rent collaterals that terminate on active pyramidal language, mathematics, and complex memory tasks.
cells are enhanced (in this simple illustration, the The functional properties of neocortex depend
weights go from 0 to 1). Later, input of some frag- upon its structure. There are four major principles of
ment of a stored pattern (1) activates a corresponding neocortical organization: First, the cortex is orga-
442 NEURAL COMPUTATION: Neocortex
Figure 1 Seeing is a complex act that requires a parallel

analysis of shape, motion, depth, color, texture, and
other attributes, all leading to the discrimination and
recognition of an object. For example, a major cue to
depth comes from stereopsis, a process of pattern
matching in which the slightly disparate views of the
world seen by the two eyes are compared; the distance
of various objects is computed from the shift in view.
Perception of depth thus results from a cortical com-
putation.
Computations in another set of cortical areas re-
sult in perception of color. Contrary to the implica-
tions of simple optics, our perceptions of color are not
solely determined by the wavelengths of light reflect-
ed from an object. As Edwin Land demonstrated,
color perception is determined by the relative
Illustrations of the aperture problem. (A) Cortical neurons receive
amount of light of different wavelengths reflected by
inputs from spatially restricted regions of the world, depicted in an object as opposed to that reflected by its surround-
the figure as the shaded circular region. If a long line is viewed as ing environment. Neurons in cortical area V4 have
it moves through this region, only the component of motion been found to respond to the color of an object as we
perpendicular to the line can be detected, since motion along the perceive it; those in the lower-level area V1 respond
lines axis is undetectable. (B) This leads to an indeterminacy in to the wavelength of the light, not the color (Zeki,
detecting the true direction of motion of the line, since all motions 1980); thus, the color computation probably depends
sharing the same perpendicular component will be treated as on computations carried out in area V4.
equal.
The analysis of motion is similarly carried out in
several stages and provides the best-investigated ex-
nized vertically into six layers, each layer receiving ample of a cortical computation. In some species,
from and sending connections to various brain sites. such as the rabbit and the housefly, neurons in the
Second, there is a horizontal organization of the cor- retina are capable of detecting motion. In most mam-
tex into columns, such that neurons in a vertical col- mals, however, sophisticated motion detection first
umn (of approximately 0.5 millimeter diameter) ex- arises in the cortex. Each cortical neuron receives vi-
sual inputs from only a small region of spacethat is,
hibit similar physiological properties. Third, most
it sees only a limited visual field. As Marr (1982)
cortical areas contain one or more topographic maps.
first pointed out, this leads to the so-called aperture
(In a topographic map, adjacent areas of the periph-
problem (see Figure 1). If you look at a moving line
ery are represented in adjacent areas of cortex. For
through a small aperture, you can determine only the
example, in visual cortex, area Vl contains a map of
component of motion perpendicular to the line be-
the visual field; in somatosensory cortex, area SI con-
cause it is impossible to tell whether the line is moving
tains four separate maps of the body surface.) Finally,
along its own axis. Each neuron faces essentially the
cortical areas are functionally segregatedcells in same problem. Neurons in V1 are orientationally and
each area are specialized for particular functional directionally selective, which means that they respond
tasks. (For example, cells in visual area MT are spe- only to lines and edges whose orientation falls within
cialized for motion detection, among other things, a certain narrow range (e.g., within ten degrees of
whereas those in visual area V4 are specialized for vertical) and that are moving within a narrow range
color and pattern analysis, among other things). The of directions (usually perpendicular to the preferred
maps in different cortical areas are interconnected in orientation). The only way around the aperture prob-
a vast and intricate scheme. lem is to combine information from cells with differ-
In the somatosensory cortex (which mediates ent directional selectivities. This neural synthesis
touch, temperature, and pain), maps are dynamically probably occurs in area MT (a small region of higher-
organized such that they change continuously in re- level visual neocortex that receives input from area
sponse to skin stimulation (Buonomano and Mer- V1), as was shown in an ingenious experiment on
monkey cortex by Movshon and colleagues (1985).
zenich, 1998; Kaas, 2000). However, the visual neo-
cortex has been the most intensively studied region The key to Movshons experiment is the visual
and provides the clearest example of the kinds of stimulus presented to the monkey. The stimulus, as
computations the cortex must carry out. shown in Figure 2, consisted of two gratings (arrays
NEURAL COMPUTATION: Neocortex 443
of parallel, evenly spaced, lines), oriented at different Figure 2

angles, and each moving in a direction perpendicular
to the orientation of their lines. If one draws two such
gratings on transparent plastic sheets, superposes
them, and moves them with respect to each other,
then, instead of two sets of moving lines, one sees a
checkerboard pattern moving in a third direction.
This third direction is not the vector sum of the two
component directions; rather, it is a more complicat-
ed function that reflects the intersection of the lines
of uncertainty as specified by the constraints of the
aperture problem (Movshon, Adelson, Grizzi, and
Newsome, 1985).
Researchers presented this stimulus to alert mon-
keys and observed the responses of neurons in areas
V1 and MT. In area MT, they found neurons that re-
sponded to the direction of motion of the checker-
board. In V1, they found no such neurons; rather,
neurons responded to the direction of motion of ei-
ther one grating or the other. Thus, to speak some- The Movshon experiment (Movshon, J. A., Adelson, E. H.,
what broadly, V1 sees only the component gratings, Grizzi, M. S., and Newsome, W. T. [1985]. The analysis of
moving visual patterns. In C. Chagas, R. Gattass, and C. Gross,
whereas MT sees the checkerboard. This is a situa-
eds., Pattern recognition mechanisms. Experimental Brain Research
tion analogous to color vision: A higher-level cortical
Supp. 11, 117151). A. Two gratings are shown, oriented at
area computes an object-related attribute based on in- different angles. Each grating is moved in a direction
formation from lower cortical areas. perpendicular to the orientation of its lines, as indicated by the
Taking this one step further, Bill Newsome and arrow. B. The two gratings are superposed and moved, as in
colleagues, in a remarkable series of experiments, (A). The resulting perception is of a checkerboard pattern
demonstrated that the direction of motion of a stimu- moved in a third, unrelated direction (see text for details). This
experiment illustrates how the cortex computes the direction of
lus, as viewed and reported by an awake, trained mon-
motion of an object from sets of indeterminate information (due
key, correlates precisely with the activation of individ-
to the aperture problem).
ual neurons in area MT. By determining the
selectivities of MT cells to various directions of mo-
tion, the experimenters could read off the neural re-
hippocampus can produce and maintain specific pat-
sponse to any test stimulus. When the monkey occa-
terns of firing in response to a stimulus. As Earl Miller
sionally made an error in judging the direction or
velocity of the stimulus, it was correlated with a failure and colleagues have shown, if a monkey is shown a
of the appropriate cells to fire. The investigators picture such as a photo of another monkey, specific
could even change the motion direction that the mon- recognition cells in high-level visual cortical areas
key perceived by stimulating particular cells to fire via (such as the inferotemporal IT cortex) will fire and
implanted microelectrodes (Britten et al., 1992). A continue to fire even after removal of the stimulus.
number of computational models have proposed Neurons in areas of frontal cortex have an even
mechanisms by which area MT may carry out these more remarkable property. They can maintain a pat-
computations (Schrater, Knill, and Simoncelli, 2000). tern of firing (in which certain neurons in the region
Each area of the cortex is dedicated to particular continually fire while others are silent) even in the
computations; there are, however, similarities to the presence of distracting stimuli. The ability of a net-
neuronal and architectural structure across the cor- work to maintain a fixed pattern of activity in the ab-
tex. This suggests that a basic set of canonical compu- sence of direct stimulus input is attractor behavior.
tations may be carried out by all cortical regions and Attractor states are stable and self-perpetuating be-
that differences in cortical computations may be cause the inputs each cell receives via connections
largely due to the type of inputs received. It is not from other active neurons in the network are suffi-
clear what the canonical cortical computations are, cient to maintain its firing. Based on theoretical con-
but researchers have suggested several candidate siderations, John Hopfield (1982) proposed that at-
mechanisms. tractor states might correspond to stored memories.
Neurons in many regions of the neocortex and in Each attractor state (pattern of activity) has a basin
phylogenetically older cortical regions such as the of attraction, a space of related activity patterns that
444 NEURAL COMPUTATION: Neocortex
transform into the attractor state, much like particles A cortical hypercolumn integrates bottom-up, top-
gravitationally attracted into a black hole. This con- down, and horizontal (contextual) inputs. Taking our
fers on the network the property of auto-association earlier example of color vision, the perceived color
or error-correction; if a noisy, corrupted or partial depends upon the incident wavelengths within the re-
memory is used as input, the complete stored pattern ceptive field (bottom-up), the contextual inputs from
can be retrieved. neighboring regions of the visual field (horizontal),
and top-down knowledge. Integration and inference
The ability to stably maintain a piece of informa-
tion in active or working memory is essential to any are the central principles of cortical function, and
computational process. If we attempt to add two plus mechanisms for these processes have been envisioned
three, we have to store the first number while we fetch in several recent theories. Rajesh Rao and Dana Bal-
the second and perform the operation. Any type of lard (1999) have proposed that feedback loops be-
matching process in recognition (e.g., Wait a min- tween higher and lower areas carry out a kind of pre-
ute; I know that face!) requires holding a stimulus in dictive coding. Higher areas predict the activity in
working memory. Equally important is the ability to the lower areas and send this prediction back via de-
release the network from the memory or to use one scending connections, and feedforward connections
memory as a prompt to the next associated memory to the higher areas convey the residual errors in the
in a sequence. In the frontal cortex, where neurons predictiona kind of neural Kalman filter. Several
have the unique ability to hold on to the attractor information-theory-based learning algorithms take a
state even in the face of distracting stimuli, the neuro- similar approach, in which the key operation is maxi-
modulator dopamine may control attractor stability mal compression of the represented information with
(Durstewitz, Seamans, and Sejnowski, 2000). When the minimal loss of information about the inputs. Shi-
dopamine levels are high, the attractors are robustly mon Ullman has shown that a cortically inspired net-
stable against distraction; when the dopamine level work based on maximizing mutual information can
falls, distracting stimuli knock the system into a new achieve remarkable image-recognition performance,
pattern of activity corresponding to the novel stimu- including the ability to categorize novel objects into
lus. Some investigators have suggested that disorders previously learned categories.
of the dopamine system, such those typical of schizo- Cortical processing is extremely rapid. M. Fabre-
phrenia, may underlie the symptoms of perseveration Thorpe and colleagues (2001) have shown that well-
(inability to get a thought out of ones head) and delu- known images can be recognized when presented at
sions/hallucinations (possibly the inappropriate link- interstimulus intervals of fifty milliseconds. Given
ing of distracting stimuli or memories). typical cortical firing rates, this result suggests that
A cortical circuit can carry out different computa- recognition and other higher processes require only
tions at different times. The system of neuromodula- one or two spikes from each cellthere is no time for
tors, including acetylcholine, serotonin, noradrena- any iterations of an algorithm. This harks back to a
line, dopamine, and other agents are a control system rule suggested by David Marr that a good neural algo-
regulating the kinds of computations and the flow of rithm operates in the time required for information
information between brain regions. For example, J. to reach the relevant circuit but does not require any
Lisman and N. A. Otmakhova (2001) have proposed additional processing time. These timing constraints
that in the hippocampus bursts of dopamine may call for representations with fast dynamics. One likely
transiently switch a circuit into a learning mode (as candidate is the space-rate code (Maass, 2001) in
opposed to an on-line information processing mode). which the degree of activation is represented by the
Similarly, Michael Hasselmo (1999) has demonstrat- fraction of neurons in a population that fire in a short
ed that acetylcholine may switch between learning (e.g., 5 ms) time window. Since, in each subsequent
and recall, and may selectively suppress intrinsic as time window, the fraction firing can substantially
opposed to extrinsic inputs. The release of acetylcho- change, information can be rapidly communicated.
line and noradrenaline correlate closely with visual How might recognition make use of spike-timing
attention, and both acetylcholine and dopamine are
information? John Hopfield has demonstrated one
associated with inducing rapid changes in cortical
possible mechanism using the example of speech rec-
connectivity. Disorders of the neuromodulatory sys-
ognition. Hopfields network (2001) makes use of an
tem, coupled with structural damage to cells and syn-
array of feature detectors, each tuned to an onset or
apses, are the hallmark of most of the neurodegen-
offset of sounds in a particular frequency range. Any
erative diseases with cognitive deficits (e.g.,
speech stimulus (e.g., a speaker saying the word one)
Alzheimers disease).
produces a spatiotemporal pattern of activation of the
Perhaps the greatest mystery in understanding feature detectors. In Hopfields network, the feature
cortical processing is the mechanism of integration. detectors activate cells in a second network whose fir-
NEURAL COMPUTATION: Olfactory Cortex as a Model for Telencephalic Processing 445
ing fate adaptsthat is, their firing rate slows with Maass, W. (2001). Computation with spiking neurons. In M. A.
time after the stimulus. The rate of adaptation can be Arbib, ed., The handbook of brain theory and neural networks.
varied and is learnedit is set so that at some time Marr, D. (1982). Vision: A computational investigation into the human
point, all cells responding to the stimulus will have representation and processing of visual information. San Francisco:
adapted to fire at the same rate. This common rate W. H. Freeman.
is transientcells will continue to adapt, and their fir- Miller, E. K. (2000). The prefrontal cortex and cognitive control.
ing rates will diverge; except for a brief time window, Nature Reviews Neuroscience 1, 5965.
Movshon, J. A., Adelson, E. H., Grizzi, M. S., and Newsome, W. T.
all cells responding to the stimulus share a similar fir- (1985). The analysis of moving visual patterns. In C. Chagas,
ing rate. Hopfield proposes that this situation, where R. Gattass, and C. Gross, eds., Pattern recognition mechanisms.
a number of cells in a network fire at approximately Experimental Brain Research Supp. 11, 117151.
the same rate, is statistically significant. It could be Rao, R. P., and Ballard, D. H. (1999). Predictive coding in the visu-
one of the fundamental kinds of computations car- al cortex: A functional interpretation of some extra-classical
receptive-field effects. Nature Neuroscience. 2, 7987.
ried out by cortical networks. Schrater, P. R., Knill, D. C., and Simoncelli, E. P. (2000). Mecha-
Understanding cortical computation remains a nisms of visual motion detection. Nature Neuroscience. 3, 64
challenge. But advances in neuroscience, particularly 68.
Thorpe, S., Fize, D., and Marlot, C. (1996). Speed of processing in
the emergence of optical imaging techniques, cou- the human visual system. Nature 381, 520522.
pled with the development of sophisticated informa- VanRullen, R., and Thorpe, S. J. (2001). The time course of visual
tion-theory models, offer the promise of new insights. processing: From early perception to decision-making. Jour-
Such advances promise a new era in artificial intelli- nal of Cognitive Neuroscience 13, 454461.
gence and the creation of information technologies Zeki, S. M. (1980). The representation of colours in the cerebral
cortex. Nature 284, 412418.
powered by biology-based algorithms.
Leif H. Finkel
See also: GUIDE TO THE ANATOMY OF THE BRAIN:

CEREBRAL CORTEX
OLFACTORY CORTEX AS A MODEL FOR
TELENCEPHALIC PROCESSING
Bibliography
Britten, K. H., Shadlen, M. N., Newsome, W. T., and Movshon, J. Changes to myriad synapses throughout the brain
A. (1992). The analysis of visual motion: A comparison of neu- must be coordinated every time a memory is estab-
ronal and psychophysical performance. Journal of Neuroscience lished, and these synapses must be appropriately re-
12, 4,7454,765.
Buonomano, D. V., and Merzenich, M. M. (1998). Cortical plastici- activated every time the memory is retrieved. Once
ty: From synapses to maps. Annual Review of Neuroscience 21, stored, memories can be recognized (as a re-
149186. experienced input) or recalled (via different input,
Durstewitz, D., Seamans, J. K., and Sejnowski, T. J. (2000). Neuro- such as a name evoking the memory of a face or a
computational models of working memory. Nature Neuro- scene evoking memories of an experience) by many
science 3, 1,1841,191.
Fabre-Thorpe, M., Delorme, A., Marlot, C., and Thorpe, S. (2001). routes. We remember what tables are as well as we re-
A limit to the speed of processing in ultra-rapid visual catego- member a specific table, and we recognize objects de-
rization of novel natural scenes. Journal of Cognitive Neuro- spite seeing them from quite different angles, under
science 13, 171180. different lighting, in different settings. Computation-
Hasselmo, M. E. (1999). Neuromodulation: Acetylcholine and al simulations of synaptic modifications (e.g., long-
memory consolidation. Trends in Cognitive Science 3, 351359.
Hopfield, J. J. (1982). Neural networks and physical systems with term potentiation) in distinct brain-circuit architec-
emergent collective computational abilities. Proceedings of the tures illustrate how these minute changes can give rise
National Academy of Sciences of the United States of America 79, to coherent properties of memory; how analyses of
2,5542,558. different brain areas yield derivations of disparate
Hopfield, J. J., and Brody, C. D. (2001). What is a moment? Tran- memory functions; and how interactions among con-
sient synchrony as a collective mechanism for spatiotemporal
integration. Proceedings of the National Academy of Sciences of the nected regions give rise to still new operating princi-
United States of America 98, 1,2821,287. ples beyond those of their constituents.
Kaas, J. H. (2000). The reorganization of somatosensory and The principal anatomical designs in mammalian
motor cortex after peripheral nerve or spinal cord injury in
primates. Progress in Brain Research 128, 173179. brain are cortical: planar arrays of neurons, arranged
Land, E. H. (1983). Recent advances in retinex theory and some with their cell bodies in sheets and their apical den-
implications for cortical computations. Proceedings of the Na- drites standing in parallel. This laminar pattern con-
tional Academy of Sciences of the United States of America 80, trasts with that of most reptilian brain structures, in
5,1635,169. which neurons are grouped in globular clusters (nu-
Lisman, J. E., and Otmakhova, N. A. (2001). Storage, recall, and
novelty detection of sequences by the hippocampus: Elaborat-
clei); an exception is the cortically organized reptil-
ing on the Socratic model to account for normal and aberrant ian pallium. The phylogenetic origins of the mamma-
effects of dopamine. Hippocampus 11, 551568. lian neocortex (perhaps including transformed
446 NEURAL COMPUTATION: Olfactory Cortex as a Model for Telencephalic Processing
Figure 1 riform cortex, for its pearlike shape; or primary olfac-

tory cortex, for its placement as the first cortical struc-
ture to receive olfactory input relayed from the
periphery.) Abstract models have been constructed
based on four fundamentals of the olfactory system:
its anatomical structure, its physiological operation
during behavior, the characteristics of synaptic
change caused by LTP, and the nature of the inputs
that arrive naturally at the system during olfactory-
related behaviors.
Figure 2 schematically illustrates the anatomical
structure of a typical mammals olfactory system
(adapted from Shepard, 1991). In the figure the ani-
mals nose is to the left, with the axons from the nose
Characteristics of a cortical layer. Axons (horizontal lines) course
comprising the first cranial nerve (Nerve I) making
through the apical dendrites of a layer of neurons making
synaptic contact (in the regions termed glomeruli)
sparse, random synaptic contacts corresponding to entries in a
matrix.
with the primary excitatory (mitral) cells of the olfac-
tory bulb. Mitral cells are inhibited by granule cells
via specialized synapses (see Haberly and Shepard,
this volume), and mitral cell axons (comprising the
nonpallial precursors as well as pallium) are the sub-
lateral olfactory tract) project to cortex, where they
ject of continuing controversy. The difference is one
form synaptic contacts with the apical dendrites of the
of function, not just form. With cells arrayed in a
primary cortical excitatory layer II and III cells.
plane, the axons providing input to the structure pass
Those cells in turn project both forward, to provide
through the dendritic field, making random and
sparse synaptic contacts. This anatomical arrange- the input to downstream brain structures (such as en-
ment creates a biological version of a three- torhinal cortex), and backward, to provide feedback
dimensional array or matrix in which the rows corre- to the bulb both directly and by way of the anterior
spond to the input axons, the columns are the den- olfactory cortex (often termed the anterior olfactory
drites, and each matrix entry is a synaptic connection nucleus, despite its laminar rather than nuclear struc-
between an axon and dendrite (see Figure 1). The ture).
neocortex undergoes vast expansion with mamma-
lian evolution, and as the cortex comes to dominate
Simple Emergent Computations from
the brain, cortical computation comes to dominate
behavior.
Feedforward Operation of the Bulb-Cortex
System
When an animal is actively engaged in olfactory
The Olfactory Bulb and Paleocortex learning behavior, the entire bulb-cortex system, its
The olfactory paleocortex, one of the oldest relics primary target output structures (entorhinal cortex
in mammals of the reptilian pallium, is an apt starting and hippocampus), and even the overt behavioral
point for evaluation of cortical computation. One rea- sniffing activity of the animal operate in synchrony,
son is its relative simplicity (for instance, it has three at a rate of about four to eight cycles per second (Ma-
primary layers instead of the six layers of the neo- crides, 1975; Macrides et al., 1982; Vanderwolf, 1992;
cortex). Another is its relative proximity to its input Wiebe and Staubli, 2001). As the animal repeatedly
environment. In other sensory systems, inputs typi- samples or sniffs the olfactory environment, neurons
cally proceed from a peripheral organ (e.g., cochlea) through the entire assembly line of olfactory-
to one or more lower-brain structures (e.g., cochlear hippocampal structures send spikes down their
nucleus, colliculus), then to a noncortical (nuclear) axons, in bursts occurring approximately every fourth
structure in the thalamus (e.g., medial geniculate nu- to eighth of a second. Computer simulations of the re-
cleus), and only then on to the primary cortex for the sulting feedforward neuronal activity in the cortex
appropriate sense (e.g., auditory cortex). By compari- have shown that LTP-like synaptic-change incre-
son, olfactory receptors (activated by chemical odor- ments cause specific cortical target cells that initially
ants drawn in through the nose) project to the olfacto- responded to a particular odor to become increasing-
ry bulb and thence straight to olfactory cortex. (The ly responsive not only to that odor but also to a range
structure is variously termed olfactory paleocortex, of similar odors. Figure 3 uses broad simplifying as-
for its phylogenetic age; piriform, pyriform, or prepy- sumptions to illustrate this straightforward principle.
Figure 2
Schematic diagram of mammalian olfactory system anatomy. Input from receptor cells in the nose arrive via axons comprising the first
cranial nerve, making synaptic contact with the dendrites of mitral cells in the olfactory bulb. Mitral cell axons in turn make synaptic
contact with the apical dendrites (projecting downward, toward the cortical surface) of primary cells in the olfactory cortex. Cortical cell
axons project forward to become input to successive anatomical structures (entorhinal cortex, hippocampus) as well as projecting
backward to become feedback input to the inhibitory cells of the olfactory bulb.
(Models of the olfactory bulb [Anton et al., 1991; flower scent (and different odors elicit only their
1992] not discussed here are assumed). cluster responsese.g., meat scent, smoke scent).
This cluster responsecan give rise to useful general-
In the left-hand panel of the diagram, input ization properties, informing the organism of the
axons b, c, and d are active (arrows), and are assumed category of an otherwise unfamiliar odor, but, some-
to be sufficient to elicit firing responses from three what counterintuitively, it prevents the system from
target cells (darkened). Synapses whose inputs and making fine distinctions among members of a cluster.
targets are coactive (highlighted) will potentiate. These results are almost generic, as many computa-
After potentiation, the enhanced synapses (enlarged, tional frameworks with very different characteristics,
right panel) confer more voltage change than they including competitive networks (e.g., von der Mals-
did in their unpotentiated state, so fewer active inputs burg, 1973; Grossberg, 1976; Rumelhart and Zipser,
should suffice to elicit a target neuron response. Thus 1985; Coultrip et al., 1992); backpropagation (Ru-
any of the depicted input patterns P, Q, and R may melhart et al., 1986); and dynamical or excitatory
now suffice to activate the same three target cells, feedback networks (e.g., Hopfield, 1982) can exhibit
whereas none of these inputs would have activated similar properties.
these neurons prior to synaptic potentiation.
After potentiation episodes, inputs with highly Complex Computations from Combined
overlapping activation patterns tend to educe identi-
cal neuronal response patterns in the cortex. The re-
Feedforward and Feedback Olfactory
sult is the mathematical operation of clustering, in Operation
which sufficiently similar inputs are placed into a sin- Absent from the foregoing analysis is the exten-
gle category or cluster. The odor of a rose, a violet, sive inhibitory feedback projection from cortical neu-
or a lily might, after long-term potentiation, elicit rons to granule cells in the bulb. This pathway selec-
only an undifferentiated response corresponding to tively inhibits those bulb inputs that generate cluster
Figure 3
Simple effects of synaptic potentiation on cell response to feedforward inputs. (Left) Before potentiation, if three active synapses suffice
to elicit a response from target cells, then the three darkened cells will respond to input S (the combined activation of axons b, c, and
d), and their active synapses (highlighted) will potentiate. (Right) After potentiation, strengthened synapses (enlarged) contribute more
voltage change to a cell whenever activated, so that the same three neurons may now fire in response to reduced inputs P, Q, and R,
which would have been insufficient to elicit responses from these neurons before potentiation.
responses in the cortex, thereby unmasking the re- stance in which a novel and efficient algorithm for a
mainder of the bulbs activity. That remainder be- well-studied computational problem emerged from
comes the subsequent input to the cortex on the next simulation and analysis of a specific cortical network.
activity cycle, whereupon the same cortical operations The method was readily generalized to modalities
are performed. The result is that the second cortical other than olfaction. For instance, input patterns cor-
response (one fourth to one eighth of a second later) responding to speech sounds yielded naturally occur-
will consist of a quite distinct set of neurons from the ring clusters and subclusters on successive samples
initial response, since most of the input components (see Figure 4). Elaboration of the algorithm gave rise
giving rise to that first response are now inhibited by to families of computational signal-processing meth-
the feedback from cortex to bulb. Analysis of the sec- ods whose performance on complex signal classifica-
ond (and ensuing) responses has shown successive tion tasks has consistently equaled or outperformed
subclustering of an input: the first cycle of response those of competing methods (interested readers are
identifies the odors membership in a particular clus- referred to: Kowtha et al., 1994; Coultrip and Grang-
ter (e.g., floral), the next response (a fraction of a sec- er, 1994; Granger et al., 1997).
ond later) identifies its membership in a particular
subcluster (rose), then in a sub-subcluster (particular
variety of rose), and so on. Roughly five levels of Biological Findings and Psychological
subclustering occur in the simulation before the in- Implications
hibitory feedback to the bulb runs its course. That is,
If the olfactory system operates in this way, it
the system uses an unexpected type of temporal cod-
should show striking behavioral and electrophy-
ing, using specific target neurons selectively activated
siological results. Behavioral experiments showed
at a series of different time points to discriminate
that rats recognized novel similar odors as members
among inputs.
of a category yet also distinguished and recognized
This iterative subclustering activity turned out to individual category members, providing evidence
be mathematically expressible as a novel algorithm that animals build unsupervised similarity-based per-
for the well-studied statistical task of hierarchical clus- ceptual clusters (Granger et al., 1991). Individual ol-
tering. All such algorithms have differential costs or factory cortical neurons, measured chronically in be-
complexity in terms of the time (number of mathe- having animals, responded selectively when tested on
matical steps) and space (amount of storage) required very different odors. Moreover, responses were tran-
for each operation. Surprisingly, the derived olfacto- sient, corresponding to the interval of a specific sniff
ry algorithm exhibited computational costs that com- cycle but not to multiple cycles, a result a result that
pared favorably with those in the (extensive) litera- also corresponds with the computer simulations (Mc-
ture on such methods (Ambros-Ingerson et al., Collum et al., 1991). Findings arrived at under differ-
Kilborn et al., 1996). These studies represent an in- ent experimental conditions have yielded various hy-
Figure 4
Hierarchy created by computer simulation of successive feedforward and feedback activity in an olfactory bulb-cortex-like structure,
operating on spoken sounds rather than on olfactory input. Each sound is a letter of the alphabet. After simulated long-term
potentiation, the initial simulated cortical response (1) does not differentiate among any letters, all of which are similar enough to
each other (and different enough from other auditory inputs, from traffic noises to bird whistles) to belong to a single cluster. The
next cortical response (2) differentiates A, J, K sounds from B, C, T sounds, and others. Successive responses (35) make
iteratively finer distinctions. Eventually each letter belongs to its own sub-cluster.
potheses of olfactory function (e.g., Schoenbaum and hierarchically configured and sequentially traversed
Eichenbaum, 1995; Haberly, 2001). Further studies during recognition.
of unit neuron recordings in behaving animals will be Modeling and analysis of other brain areas, in-
needed to settle competing interpretations of the ob- cluding constituents of the hippocampal formation,
served data. auditory neocortex, the striatal complex, and thala-
The computational and neurobiological findings mo-cortical loops, has yielded a range of additional,
yield hypotheses about psychological function. Oper- starkly different emergent fundamental computa-
ations emerging from cortical circuits presumably tions for each structure, as well as novel complex op-
erations from combinations of these (e.g., Lynch and
constitute elemental psychological acts and contrib-
Granger, 1992; Gluck and Granger, 1993; Granger et
ute, through combination, to more complex mental
al., 1994; 1997; Myers et al., 1995; Kilborn et al.,
processes in ways not yet understood. The evocation
1996; Aleksandrovsky et al., 1996; 1997). As in the
of successively finer-grained information about a
case of the hierarchical clustering algorithm identi-
stimulus via sequential cortical responses suggests a
fied in the olfactory system, new functions derived
fundamental operation of repetitive perceptual sam-
from other brain regions exhibited computational
pling. Visual, auditory, and somatosensory cortices
characteristics comparable to algorithms of known
have anatomical architectures analogous to the olfac-
power, often equaling or surpassing the best algo-
tory bulb-cortex template, including excitatory feed-
rithms in cost and efficacy. Moreover, as in the case
forward and inhibitory feedback interactions with
of the olfactory system, the results suggested new in-
thalamic nuclei (see Herkenham, 1986; Jones, 1998, terpretations of both simple and complex psychologi-
for reviews). Perhaps the second glance of a scene cal operations, intimating the development of more
educes qualitatively different information from the advanced hypotheses of human brain function.
first glance (even when such glances are covert cy-
cles operating within these cortical structures, rather
See also: NEURAL COMPUTATION: APPROACHES TO
than behavioral eye movements). Humans exhibit
LEARNING; NEURAL COMPUTATION:
synchronized rhythmic firing during learning and CEREBELLUM; NEURAL COMPUTATION:
during complex sensory processing (Caplan et al., HIPPOCAMPUS; NEURAL COMPUTATION:
2001; Sobotka and Ringo, 1997). And human subjects NEOCORTEX
in perceptual and conceptual studies robustly recog-
nize objects first at categorical levels and subsequently Bibliography
at successively subordinate levels (Mervis and Rosch, Aleksandrovsky, B., Brucher, F., Lynch, G., and Granger, R.
1981; Schlaghecken, 1998; Kuhl et al., 2001), sug- (1997). Neural network model of striatal complex. In Biologi-
gesting the presence of structured memories that are cal and Artificial Computation, from Neuroscience to Technology.
IWANN97 International Conference on Artificial and Natural Kilborn, K., Granger, R., and Lynch, G. (1996). Effects of LTP on
Neural Networks. Berlin: Springer-Verlag, pp.104115. response selectivity of simulated cortical neurons. Journal of
Aleksandrovsky, B., Whitson, J., Garzotto, A, Lynch, G., and Cognitive Neuroscience 8, 338353.
Granger, R. (1996). An algorithm derived from thalamocorti- Kowtha, V., Satyanarayana, P., Granger, R., and Stenger, D.
cal circuitry stores and retrieves temporal sequences. Proceed- (1994). Learning and classification in a noisy environment by
ings of the International Conference on Pattern Recognition, IEEE a simulated cortical network. Proceedings of the Third Annual
Computer Society Press 4, 550554. Computer and Neural Systems Conference Boston: Kluwer.
Ambros-Ingerson, J., Granger, R., and Lynch, G. (1990). Simula- Kuhl, P., Tsao, F., Zhang, Y., DeBoer, B. (2001). Language, cul-
tion of paleocortex performs hierarchical clustering. Science ture, mind, brain, Progress at the margins between disci-
247, 1,3441,348. plines. Annals of the New York Academy of Science 935, 136174.
Anton, P., Granger, R., and Lynch, G. (1993). Simulated dendritic Lynch, G., and Granger, R. (1992). Variations in synaptic plasticity
spines influence reciprocal synaptic strengths and lateral and types of memory in corticohippocampal networks. Jour-
inhibition in the olfactory bulb. Brain Research 628, nal of Cognitive Neuroscience 4, 189199.
157165. Macrides, F. (1975). Temporal relations between hippocampal
Anton, P., Lynch, G., and Granger, R. (1991). Computation of fre- slow waves and exploratory sniffing in hamsters. Behavioral Bi-
quency-to-spatial transform by olfactory bulb glomeruli. Bio- ology 14, 295308.
logical Cybernetics 65, 407414. Macrides, F., Eichenbaum, H. B., and Forbes, W. B. (1982). Tem-
Bliss, T. V. P., and Lo mo, T. (1973). Long-lasting potentiation of poral relationship between sniffing and the limbic (theta)
synaptic transmission in the dentate area of the anesthetized rhythm during odor discrimination reversal learning. Journal
rabbit following stimulation of the perforant path. Journal of of Neuroscience 2, 1,7051,717.
Physiology 232, 331356. McCollum, J., Larson, J., Otto, T., Schottler, F. Granger, R., and
Caplan, J., Madsen, J., Raghavachari, S., and Kahana, M. (2001). Lynch, G. (1991). Short-latency single-unit processing in
Distinct patterns of brain oscillations underlie two basic pa- olfactory cortex. Journal of Cognitive Neuroscience 3, 293
rameters of human maze learning. Journal of Neurophysiology 299.
86, 368380. Mervis, C., and Rosch, E. (1981). Categorization of natural objects.
Coultrip, R., and Granger, R. (1994). LTP learning rules in sparse Annual Review of Psychology 32, 89115.
Myers, C., Gluck, M., and Granger, R. (1995). Dissociation of hip-
networks approximate Bayes classifiers via Parzens method.
pocampal and entorhinal function in associative learning, A
Neural Networks 7, 463476.
computational approach. Psychobiology 23, 116138.
Coultrip, R., Granger, R., and Lynch, G. (1992). A cortical model
Puelles, L. (2001). Thoughts on the development, structure and
of winner-take-all competition via lateral inhibition. Neural
evolution of the mammalian and avian telencephalic pallium.
Networks 5, 4754.
Philosophical Transactions of the Royal Society of London 356,
Gluck, M., and Granger, R. (1993). Computational models of the
(Biol.) 1,5831,598.
neural bases of learning and memory. Annual Review of Neuro-
Rosch, E., and Lloyd, B. B. (1978). Cognition and categorization.
science 16, 667706.
Granger, R., Staubli, U., Powers, H., Otto, T., Ambros-Ingerson,
Rumelhart, D., Hinton, G., and Williams, R. (1986). Learning rep-
J., and Lynch, G. (1991). Behavioral tests of a prediction
resentations by back-propagating errors. Nature 323, 533
from a cortical network simulation. Psychological Science 2,
536.
116118.
Rumelhart, D., and Zipser, D. (1985). Feature discovery by compet-
Granger, R., Whitson, J., Larson, J., and Lynch, G. (1994). Non-
itive learning. Cognitive Science 9, 75112
Hebbian properties of LTP enable high-capacity encoding of
Schlaghecken, F. (1998). On processing beasts and birds, an
temporal sequences. Proceedings of the National Academy of Sci- event-related potential study on the representation of taxo-
ences of the United States of America 91, 10,10410,108. nomic structure. Brain and Language 64, 5382.
Granger, R., Wiebe, S., Taketani, M., Ambros-Ingerson, J., and Schoenbaum, G., and Eichenbaum, H. (1995). Information coding
Lynch, G. (1997). Distinct memory circuits comprising the in the rodent prefrontal cortex. I. Single-neuron activity in or-
hippocampal region. Hippocampus 6, 567578. bitofrontal cortex compared with that in pyriform cortex.
Grossberg, S. (1976). Adaptive pattern classification and universal Journal of Neurophysiology 74, 733750.
recoding. I. Parallel development and coding of neural fea- Shepherd, G. (1991). Computational structure of the olfactory
ture detectors. Biological Cybernetics 23, 121134. system. In J. L. Davis and H. Eichenbaum, eds., Olfaction,
Haberly, L. (2001). Parallel-distributed processing in olfactory cor- a model system for computational neuroscience. Cambridge: MIT
tex, New insights from morphological and physiological anal- Press.
ysis of neuronal circuitry. Chemical Senses 26, 551576. Sobotka, S., and Ringo, J. (1997). Saccadic eye movements, even
Herkenham, M. (1986). New perspectives on the organization and in darkness, generate eventrelated potentials recorded in
evolution of nonspecific thalamocortical projections. In E. medial septum and medial temporal cortex. Brain Research
Jones and A. Peters, eds., Cerebral cortex, Vol. 5. New York: 756, 168173.
Plenum Press. Vanderwolf, C. (1992). Hippocampal activity, olfaction, and sniff-
Hopfield, J. (1982). Neural networks and physical systems with ing: An olfactory input to the dentate gyrus. Brain Research
emergent collective computational abilities. Proceedings of the 593, 197208.
National Academy of Sciences of the United States of America 79, von der Malsburg, C. (1973). Self-organization of orientation sensi-
2,5542,558. tive cells in the striate cortex. Kybernetik 14, 85100.
Jones, E. G.(1998). A new view of specific and nonspecific thalamo- Wiebe, S., and Staubli, U. (2001). Recognition memory correlates
cortical connections. Advances in Neurology 77, 4971. of hippocampal theta cells. Journal of Neuroscience 21, 3,955
Karten, H. J. (1997). Evolutionary developmental biology meets 3,957.
the brain, the origins of mammalian neocortex. Proceedings of
the National Academy of Sciences of the United States of America 94,
2,8002,804. Richard H. Granger, Jr.
NEURAL SUBSTRATES OF AVOIDANCE LEARNING 451
NEURAL SUBSTRATES OF sions of the cingulate cortical areas that receive input
AVOIDANCE LEARNING from these thalamic nuclei, render rats, cats and rab-
bits incapable of active avoidance learning. Lesions in
People and animals learn to avoid pain provided that only one of the nuclei, or in the cingulate cortical pro-
warning stimuli are available to signal pain-inducing jection field of a single nucleus, yield partial learning
events. Such learning is generally of two types, active deficits. In addition, lesions of the amygdala block
and inhibitory. Active avoidance refers to movements learning. In these studies laboratory tasks involve ani-
learned in response to warning stimuli for the pur- mals learning to jump over a barrier or learning to
pose of avoiding pain. Inhibitory avoidance refers to in- step in an activity wheel to avoid a mild electric shock
action, learned because action in the presence of the signaled by tone or light warning stimuli. The deficit
warning stimuli has previously led to pain. in animals with these lesions is a true learning deficit,
not an inability to move or to perceive the warning
Limbic and Motor Systems stimuli.
Research implicates the brains limbic and motor The specific involvement of the limbic areas in
systems in the mediation of avoidance learning. The avoidance learning receives further support from
limbic system is a vast network of interconnected re- data indicating that cerebellar lesions, which block
gions including the amygdala, hippocampus, limbic classical conditioning of eyeblink responses, do not
thalamus and the cingulate area of the cerebral cor- affect avoidance learning in rabbits.
tex. Relevant parts of the motor system include the Studies of neuronal activity during active avoid-
striatum and the nucleus accumbens. Many laborato- ance learning by rabbits in the activity wheel task have
ry studies of the neural substrates of avoidance learn- shown that amygdalar, limbic thalamic, and cingulate
ing involve locomotion (or its inhibition). Therefore, cortical neurons learn to produce impulses in re-
areas of the brain concerned with the initiation and sponse to the warning tone that signals shock. Im-
maintenance of locomotion are also involved. pulse rate in trained rabbits increased just after the
warning stimulus and reached its maximum rate just
Theoretical Overview: WHAT and WHEN before the rabbits stepped, suggesting that stepping
was triggered by the neuronal activity. That this activ-
Available data indicate that in avoidance learning ity is truly related to learning and not to generalized
the motor system acts as a WHAT system, and the lim- arousal is indicated by the fact that the activity is selec-
bic system acts as a WHEN system. The WHAT system tive, for example, greater in response to the warning
determines what is to be done, that is, the particular tone than to a second tone which is presented as often
behavior to be performed. Its functions include learn- as the warning stimulus but is not predictive of shock.
ing and remembering the response to be performed;
making ready or priming the response when the avoid- Cingulate cortical neurons send axons to striatal
ance situation is encountered; and executing the re- motor areas such as the caudate nucleus. Active avoid-
sponse. The WHAT system is relatively poor when it ance learning is impaired in animals with lesions in
comes to remembering important signals in the envi- the caudate nucleus. Thus, the flow of nerve impulses
ronment, including the warning stimuli that call for from the cingulate cortex to the caudate nucleus and
avoidance behavior; the WHEN system is specialized possibly to other motor areas is likely responsible for
to handle these functions. The WHEN system learns initiating active avoidance responses. Theoretically,
about and remembers warning stimuli, and issues the role of the caudate nucleus represents the func-
command volleys of neuronal activity that tell the tion of the WHAT system, and the information flow
WHAT system precisely when to execute the avoid- from the limbic thalamus and the cingulate cortex to
ance response (see Figure 1). the caudate nucleus represents the command volley
issued by the WHEN system.
Data supporting these ideas come from studies of
the effects of experimentally induced brain damage During the acquisition of the stepping avoidance
(lesions) and from studies of the activity of brain neu- response by rabbits, two forms of neuronal learning
rons during avoidance learning. Gabriel (1993) re- have been noted in the limbic thalamus and in the
viewed much of the research data summarized below; cingulate cortex. One form is discriminative or selective
citations of studies not included in the review are neuronal activity (SA), as defined above, and the sec-
noted in the text that follows. ond form, excitatory modulation (EM), is a dramatic in-
crease of impulse firing rate in trained rabbits in re-
sponse to both warning and nonwarning tones. Two
Active Avoidance Learning facts indicate that EM is, like SA, clearly a learning-
Experimental lesions in the medial dorsal (MD) related change and not merely a reflection of general
and anterior (AN) nuclei of the limbic thalamus, or le- arousal: 1. EM does not occur when rabbits experi-
452 NEURAL SUBSTRATES OF AVOIDANCE LEARNING
Figure 1
Schematic diagram of the information flows among brain structures involved in avoidance learning. The plus and minus signs
represent, respectively, transmissions relevant to active and inhibitory avoidance behavior.
ence repeated sessions with shock and nonwarning ceptors increases in the AN during learning (Vogt et
tones, that is, tones that do not predict the occurrence al., 1991), and scopolamine hydrobromide, which
of shock; and 2. EM occurs at different rates in differ- blocks these receptors, abolishes both the EM in the
ent nuclei of the limbic thalamus. (General arousal or AN, as well as performance of the avoidance behavior.
excitement would be expected to increase neuronal These results suggest that EM is due to stimulation by
activity simultaneously in many brain areas.) Al- acetylcholine of the increased numbers of receptors
though they are combined in the limbic thalamus, EM in the AN. Acetylcholine is released by the terminals
and SA have different brain origins. of brain stem tegmental neurons that project to the
AN. Cholinergic stimulation may increase the excita-
Origins of EM tion of AN neurons by enhancing the release of the
Experimental lesions in the large fiber tract (the mammillothalamic tract neurotransmitter or by in-
mammillothalamic tract) that runs from the hypothala- creasing the excitability of AN neurons in response to
mus to the AN block the development of EM in the that neurotransmitter. To summarize, EM originates
AN; these lesions diminish performance efficiency of in the limbic thalamus and has the function of ampli-
the avoidance response. Binding of the high-affinity fying the selective discharges as they are relayed to
ligand oxotremorine to muscarinic acetylcholine re- the cingulate cortex.
NEURAL SUBSTRATES OF AVOIDANCE LEARNING 453
Selective Neuronal Activity in Two alley for food or water reinforcement), or assessing
Functional Circuits choice behavior such as the relative amount of time
Immediately after the onset of training, SA in the spent in neutral areas compared with time spent in
anterior cingulate cortex develops very rapidly, areas previously established as dangerous. The tim-
whereas neurons of the posterior cingulate cortex and ing of avoidance responses is not greatly critical for
in the AN develop selective activity gradually. Sepa- learning of these tasks because subjects are not re-
rate lesions in these areas induced before training im- quired, as they are in active avoidance tasks, to pro-
pair respectively, performance in the early and late duce discrete behaviors at particular moments. In-
stages of learning. These two circuits represent re- stead, behavior must be suppressed in response to
spectively the operations of an immediate or primary configurations of static, continuously present envi-
working memory circuit and a secondary or interme- ronmental stimuli.
diate-term memory circuit. Development of SA in the The hippocampus, a limbic area of the cerebral
secondary circuit is a result of short-term consolida- cortex, is important to complex memory functions.
tion processes. SA also develops very rapidly in the Hippocampal lesions are detrimental to inhibitory
parts of the ascending sensory (auditory) pathway for avoidance learning. Hippocampal involvement may
hearing, including the medial geniculate nucleus reflect the fact that the warning stimuli in inhibitory
(MGN). Because MGN neurons project to the amyg- avoidance tasks are often experimental environments
dala, it was believed that SA in the MGN engenders or places, rather than discrete stimuli such as tones.
the SA in the amygdala. However, SA in the MGN was Successful performance in such tasks depends on the
blocked during training of subjects with temporary cognitive mapping functions of the hippocampus, in-
amygdalar lesions, indicating that the amygdala en- cluding remembering whether particular environ-
genders SA in the MGN. ments are dangerous or safe (see Nadel, OKeefe, and
Black, 1975).
Origins of Selective Activity
As in active avoidance learning, outputs of the
Additional studies have shown that the amygdala
amygdala are be involved in initiating learning-
modulates the MGN by way of an amygdalar pathway
related plasticity in other areas in response to aversive
to the auditory cortex and from there to the MGN.
stimulation received during inhibitory avoidance
Auditory cortical lesions abolish the rapid SA in all
training. Research of James L. McGaugh (2000) indi-
areas, including the MGN, basolateral nucleus of the
cates that amygdalar outputs induce the hippocam-
amygdala and the anterior cingulate cortex (Duvel et
pus to store or consolidate memory underlying inhib-
al., 2001). Subjects with these lesions make fewer
itory avoidance learning. Hippocampal cognitive
learned responses than controls during the early ses-
mapping operations could give rise to the suppres-
sions of discriminative avoidance learning, and they
sion of movement toward dangerous environments as
fail to discriminate between the CS+ and the CD-
a result of information flow over a massive WHEN sys-
during the early sessions. However, the slower SA of
tem pathway from the hippocampal formation to the
the secondary circuit remains intact and these sub-
nucleus accumbens. The latter area is a WHAT system
jects eventually learn to normative levels of perfor-
component that has been implicated in the suppres-
mance. Large permanent lesions of the amygdala
sion of locomotion.
block the development of gradual SA in the secondary
memory circuit and neurons of the lateral anterior Finally, substantial evidence indicates that learn-
nucleus of the amygdala exhibited slow SA. These re- ing of immobility in response to discrete cues using
sults indicate that 1. the amygdala initiates SA in the Pavlovian training procedures appears to be mediat-
MGN and possibly in the auditory cortex; 2. the audi- ed by plasticity that develops within the amygdala. In
tory cortex is the origin of SA in the remaining areas this instance, the amygdala is viewed as a site of mem-
of the primary memory circuit; and 3. the lateral ante- ory storage rather than a plasticity-initiating agent.
rior area of the amygdala may initiate gradual SA of
the secondary memory circuit. See also: ACTIVE AND PASSIVE AVOIDANCE
LEARNING: BEHAVIORAL PHENOMENA; GUIDE
TO THE ANATOMY OF THE BRAIN: AMYGDALA;
Inhibitory Avoidance Learning NEURAL SUBSTRATES OF CLASSICAL
CONDITIONING: FEAR CONDITIONING,
Experimental procedures used to study inhibito- FREEZING; PASSIVE (INHIBITORY) AVOIDANCE,
ry avoidance learning include the delivery of shock FEAR LEARNING
after performance of an unlearned response by rats
(moving from a lighted area to an innately preferred Bibliography
dark area), the delivery of shock after performance of Duvel, A. D., Smith, D. M., Talk, A., and Gabriel, M. (2001). Medial
a previously learned response (running in a maze geniculate, amygdalar and cingulate cortical training-induced
454 NEURAL SUBSTRATES OF CLASSICAL CONDITIONING: Cardiovascular Responses
neuronal activity during discriminative avoidance learning in freezing to context. Another productive method to study
rabbits with auditory cortical lesions. Journal of Neuroscience learned fear is FEAR-POTENTIATED STARTLE. Here a light
27, 3,2713,281. CS is paired with shock to establish conditioned fear to the
Gabriel, M. (1993). Discriminative avoidance learning: A model
system. In M. Gabriel and B. Vogt, eds. Neurobiology of cingu- light. This CS is then given together with a loud acoustic
late cortex and limbic thalamus, 478523. Toronto: Birkhauser. stimulus that elicits behavioral startle response. Presentation
McGaugh, J. L. (2000). Memorya century of consolidation. Sci- of the light CS enhances the startle response to the acoustic
ence 287, 248251. stimulus. Much of the circuitry for the startle response and
Nadel, L., OKeefe, J., and Black, A. (1975). Slam on the brakes: fear potentiation of the response has been identified. As in
A critique of Altman. Brunner and Bayers response-
inhibition model of hippocampal function. Behavioral Biology freezing, the amygdala plays a critical role.
14, 151162. The article on NEURAL SUBSTRATES OF AVOIDANCE
Vogt, B. A., Gabriel, M., Vogt, L. J., Poremba, A., Jensen, E. L., Ku-
LEARNING is included in this book in part for contrast. It
bota, Y., and Kang, E. (1991). Muscarinic receptor binding in-
creases in anterior thalamus and cingulate cortex during dis- describes an example of instrumental learning, where the
criminative avoidance learning. Journal of Neuroscience 11, animal can influence the outcome, unlike classical condi-
1,5081,514. tioning. The focus in this entry is on active avoidance, where
Michael Gabriel the animal can make a response (e.g., locomotion) when a
CS occurs to avoid a shock US. But until the animal first
moves to avoid the shock, the training is Pavlovian. Critical
neural structures for this form of learning include the amyg-
dala, certain nuclei of the thalamus, and the cingulated area
NEURAL SUBSTRATES OF CLASSICAL of the cerebral cortex.]
CONDITIONING
[Classical or Pavlovian conditioning, first described by Ivan
Pavlov (see PAVLOV, IVAN ), is a procedure where a neutral CARDIOVASCULAR RESPONSES
stimulus such as a light or sound (conditioned stimulus, CS)
is presented together with an unconditioned stimulus (US) Learning is an enduring change in behavior arising
that elicits a behavioral response (UR). As a result of pair- from experience-induced structural changes in the
ing, the CS comes to elicit a conditioned response (CR). In brain. For example, during Pavlovian (classical) aver-
Pavlovs original experiments with dogs, a bell (CS) was sive learning, an organism learns that an auditory or
paired with meat powder in the mouth (US). The salivation visual stimulus (the conditioned stimulus, CS) that re-
UR elicited by the meat powder came to be elicited by the bell, peatedly precedes an electric shock (the uncondi-
the CR. The CS must precede the US for learning to occur; tioned stimulus, UCS) predicts the occurrence of the
in the delay procedure the CS and US co-terminate; in the shock. This learned association is reflected in a vari-
trace procedure the CS offset occurs prior to the onset of the ety of adaptive cardiovascular responses to the CS
US. Although contiguity of the CS and US is necessary for and is a specific consequence of the associative rela-
learning, the contingency between them, the probability that tionship between the CS and UCS. The search for the
the CS will predict the occurrence of the US, is critically im- structural changes of Pavlovian learning is governed
portant. by the assumption that an association between the CS
and the UCS hinges on a convergence of information
The second entry in this section, on conditioning of DIS- about both at a common brain structure or structures.
CRETE BEHAVIORAL RESPONSES, uses eyeblink condition- It is this convergence that results in structural
ing as the prototypic example (tone CS, corneal airpuff US). changes.
The cerebellum and its associated brainstem circuitry is the
necessary and sufficient circuit for this form of learning; Researchers have used learned cardiovascular re-
however, the hippocampus also becomes important in the sponses, particularly changes in heart rate (HR) in re-
trace procedure. Elsewhere in this book, NEURAL SUB- sponse to the CS during Pavlovian learning, as mod-
STRATES OF EMOTIONAL MEMORY provides an over- els for assessing learning and for identifying the brain
view of fear learning, where a neutral CS is paired with an regions in which learning-related structural changes
aversive, emotionally arousing US such as shock. The amyg- occur. HR responses are especially useful because re-
dala plays a key role in all aspects of fear conditioning. A searchers have identified the location of the motor
major component of fear learning is classical conditioning neurons that produce these responses and can thus
of CARDIOVASCULAR RESPONSES. Here, the amygdala, identify the central brain structures that activate these
prefrontal cortex, and cerebellum are all involved. The most neurons. This, then, sets the stage for determining
widely used behavioral index of FEAR CONDITIONING is the specific location(s) and nature of the structural
freezing. Here it appears that critical components of the fear changes.
memory are stored in a region of the amygdala. However the The conditioned cardiovascular changes that
hippocampus also becomes critically important in learned occur during Pavlovian conditioning represent only
NEURAL SUBSTRATES OF CLASSICAL CONDITIONING: Cardiovascular Responses 455
one of a whole complex of nonspecific visceral re- tures other than auditory cortex that receive MGm
sponses that occur during classical conditioning. Ac- projections may be essential components of the cir-
quisition of these learned autonomic adjustments oc- cuit.
curs rapidly, often within just two to three CS/UCS
presentations (Powell, McLaughlin, and Chachick,
2000). These changes may also accompany later- Central Structures Involved in Learned
occurring specific learned somatomotor responses, Bradycardia
such as the conditioned eye blink or the leg-flexion There is substantial evidence that three separate
response. Learned autonomic changes are nonspecif- but partially overlapping higher-level CNS structures
ic responses because they occur regardless of the na- mediate Pavlovian heart-rate conditioning; these in-
ture of the conditioning contingencies. A CS that sig- clude several subnuclei of the amygdala, the medial
nals an electric shock UCS delivered to either the prefrontal cortex, and the cerebellar vermus. Interac-
orbital region or the footpad of animals, for example, tions between these three structures are illustrated in
results in a host of nonspecific CRs to the signals that Figure 1.
are quite similar regardless of the application of the
The work of David Cohen and associates (1984)
UCS. However, the learned somatomotor response is
on the pigeon, Bruce Kapp and colleagues on the rab-
specific to the UCS, consisting of an eyeblink CR in
bit (Kapp et al., 1991), and LeDoux and colleagues on
the former case and leg-flexion CR in the latter.
the rat (LeDoux, 2000) have demonstrated a thalam-
Hence such responses are usually referred to as spe-
ic-amygdaloid circuit that is necessary for the elabora-
cific conditioned responses.
tion of conditioned cardiovascular changes. These
There are many parametric differences between studies have shown that one recipient of projections
the acquisition of specific and nonspecific responses from the MGm is the lateral nucleus of the amygdala
during classical conditioning. A wealth of research in- (LeDoux, Farb, and Ruggiero, 1990), which projects
dicates that the the brain mechanisms responsible for to the amygdala central nucleus (Ace) and the baso-
the early-occuring heart rate and other nonspecific lateral (BL) nucleus. Cells in the somatic thalamus
changes are quite different from those that mediate also project to the lateral nuclei. Since the UCS is
the later acquisition of somatomotor changes. We will electric shock, such inputs to the amygdala may well
concentrate on the former mechanisms, with specific reflect UCS information; there is some evidence that
reference to cardiovascular changes, although there the critical association between the CS and UCS takes
is substantial evidence suggesting that there are inter- place in this nucleus and is relayed to the ACe and BL
actions between the structures that mediate specific nucleus.
and nonspecific conditioning (Powell, McLaughlin,
and Chachick, 2000). The contribution of the ACe to learned bradycar-
dia in the rabbit has been extensively investigated
(Kapp et al., 1991). The ACe projects directly to the
Pathways Transmitting CS Information region of cardiodecelerative motor neurons within
the medulla, and electrical stimulation of the ACe
Investigations of HR conditioning generally have
elicits vagal bradycardia and excitation of these neu-
used auditory CSs that precede an electric-shock
rons. Hence, the ACe is located between the CS path-
UCS. Research by Joe LeDoux, using an auditory CS
way and the motor neurons that produce the re-
in the rat during Pavlovian aversive learning, has
sponse. Interference with its normal functioning
demonstrated that destruction of the inferior col-
markedly attenuates the learned response. Further,
liculus (a structure relaying auditory information
the activity of its neurons in response to the CS
from the most peripheral structures of the auditory
changes over repeated CS-UCS pairing; and, for
system) blocks the expression of several nonspecific
some neurons, the greater the neurons excitatory re-
learned responses (LeDoux, Sakaguchi, and Reis,
sponse to the CS, the greater the conditioned brady-
1984). The inferior colliculus projects to the medial
cardiac response. These combined observations sug-
geniculate nucleus, and lesions of its magnocellular
gest the ACe excites motor neurons leading to
component (MGm) produce deficits in the acquisition
response expression.
of the learned bradycardiac responses (i.e., HR de-
creases) in the rabbit (McCabe, McEchron, Green, While the ACe may be a critical structure within
and Schneiderman, 1993). Destruction of the audito- the circuit, as noted above, other structures, including
ry cortex, which receives projections from the MGm, the cerebellar vermis and medial prefrontal cortex,
does not interfere with the initial acqusition of this also appear to be important (Supple and Kapp, 1989;
learned response (Teich et al., 1988), suggesting that Powell, McLaughlin, and Chachick, 2000). Lesions of
the auditory cortex is not necessary for learning the either produce a marked attenuation of the response,
relationship between the CS and UCS. Thus, struc- and the activity of neurons to the CS in each region
456 NEURAL SUBSTRATES OF CLASSICAL CONDITIONING: Cardiovascular Responses
Figure 1
A schematic diagram of the components of the putative neural circuit mediating learned bradycardia in the rabbit. The left side shows
auditory structures that carry information about the tone CS. Components are believed to include primary auditory relay nuclei from
the ear to the medial geniculate nucleus. CS information diverges from the primary auditory pathway at the level of the medial
geniculate and is carried to the amygdala. From the amygdala, the central nucleus exerts both direct and indirect influences on motor
neurons in the medulla that decrease heart rate when activated. The cerebellar vermis and medial prefrontal cortex, both important in
learned bradycardia, may influence the central nucleus through indirect routes. The prefrontal cortex also projects directly to the
medulla. Its control of learned HR adjustments may be influenced by neurons in the hippocampus. Evidence also suggest that
information about the UCS may access the medial geniculate, the lateral nucleus of the amygdala, prefrontal cortex, cerebellar vermis,
and perhaps other structures of the circuit. However, the pathways that transmit UCS information have not been extensively
investigated. Nevertheless, there are multiple sites of CS and UCS convergence within the circuit, suggesting multiple potential sites of
structural change.
changes as a function of repeated CS-UCS pairings. ents from the prefrontal cortex project directly to the
As with the ACe, in each region increased neuronal cardiac nuclei that control HR changes in the medulla
responses to the CS predicted greater CR response (Buchanan et al., 1994). There is also considerable ev-
magnitude. Both regions are anatomically associated idence that the hippocampus may be intimately in-
with the amygdala: the medial prefrontal cortex via volved in the stimulus processing associated with clas-
a direct pathway to the BL and ACe, and the vermis sical conditioning of HR responses. There is evidence
via an indirect one. It has also been shown that effer- that the ventral hippocampus, possibly via its effer-
NEURAL SUBSTRATES OF CLASSICAL CONDITIONING: Cardiovascular Responses 457
ents to the medial prefrontal cortex, plays a role in both the CS and UCS, and, like more central struc-
conditioning of nonspecific responses (Powell, tures, neuronal activity changes as a function of re-
McLaughlin, and Chachick, 2000). Information pro- peated CS-UCS pairings (Supple and Kapp, 1989).
cessed in the hippocampus may thus be responsible Likewise, neurons in the amydala, cerebellar vermis,
for the participation of the medial prefrontal cortex and medial prefrontal cortex are responsive to the
in HR conditioning. UCS, although the pathways by which the UCS is
transmitted to these structures are in need of further
analysis. The evidence therefore suggests that struc-
Functional Considerations tural changes may occur at multiple sites along this
Do these three different areas mediate nonspecif- critical circuit.
ic responses under different parametric and behav-
ioral circumstances? LeDoux has pointed out that See also: NEURAL SUBSTRATES OF EMOTIONAL
automatic subcortical processing of fearful stimuli by MEMORY
the amygdala offers the advantage of allowing the or-
Bibliography
ganism to a make a quick response to dangerous envi-
Buchanan, S. L., Thompson, R. H., Maxwell, B. L., and Powell, D.
ronmental stimuli (LeDoux, 2000). This subcortical A. (1994). Efferent connection of the medial prefrontal cortex
circuit is thus important in the immediate response to in the rabbit. Experimental Brain Research 100, 469483.
environmental signs of danger. On the other hand, a Cohen, D. H. (1984). Identification of vertebrate neurons modified
host of studies in human subjects have revealed that during learning: An analysis of sensory pathways. In D. L.
an emotional component associated with prefrontal Alkon and J. Farley, eds., Primary neural substrates of learning
and behavioral change. Cambridge, UK: Cambridge University
processing guides and affects complicated decision- Press.
making. The work of Damasio (1999), for example, Damasio, A. (1999). The feeling of what happens: Body and emotion in
indicates that the prefrontal cortex is important as an the making of consciousness. New York: Harvest/Harcourt.
emotional component of memory selection, which Ghelarducci, B., and Sebastiani, L. (1996). Contribution of the cer-
guides response selection based upon its conse- ebellar vermis to cardiovascular control. Journal of the Auto-
nomic Nervous System 56, 149156.
quences. The role of the prefrontal cortex in condi- Kapp, B. S., Markgraf, C. G., Wilson, A., Pascoe, J. P., and Supple,
tioned HR responding may, therefore, be related to W. F. (1991). Contributions of the amygdala and anatomical-
its integration with more complicated somatomotor ly-related structures to the acquisition and expression of aver-
response selection. Prefrontal projections bypass hy- sively conditioned responses. In L. Dachowski and C. F.
pothalamic and brain-stem mechanisms, thus project- Flaherty, eds., Current topics in animal learning: Brain, emotion
and cognition. Hillsdale, NJ: Erlbaum.
ing directly to autonomic regulatory mechanisms in LeDoux, J. E. (2000). Emotion circuits in the brain. Annual Review
the medulla and spinal cord. Such processing by the of Neuroscience 23, 155184.
prefrontal cortex may be related to integration of au- LeDoux, J. E., Farb, C., and Ruggiero, D. A. (1990). Topographic
tonomic activity that is part of the emotional process- organization of neurons in the acoustic thalamus that project
ing involved in normal adaptive behavior. to the amygdala. Journal of Neuroscience 10, 4354.
LeDoux, J. E., Sakaguchi, A., and Reis, D. J. (1984). Subcortical ef-
The role of the vermis in learned HR changes is ferent projections of the medial geniculate nucleus mediate
unclear, although cerebellar control of HR changes in emotional responses conditioned to acoustic stimuli. Journal
a number of situations is well established (Ghelarduc- of Neuroscience 4, 683698.
McCabe, P. M., McEchron, M. D., Green, E. J., and Schneiderman,
ci and Sebastiani, 1996). For example, pronounced N. (1993). Electrolytic and ibotenic acid lesions of the medial
bradycardia occurs during the diving reflex. More- subnucleus of the medial geniculate prevent the acquisition of
over, such learned HR changes may be related to the classically conditioned heart rate to a single acoustic stimulus
integration of autonomic activity with somatomotor in rabbits. Brain Research 619, 291298.
learned responses, which are controlled by the cere- Powell, D. A., McLaughlin, J., and Chachich, M. (2000). Classical
conditioning of autonomic and somatomotor responses and
bellum. These conclusions are, however, highly spec- their central nervous system substrates. In J. E. Steinmetz and
ulative at the present time. D. S. Woodruff-Pak, eds., Eyeblink classical conditioning, Vol. 2:
Animal models. Boston: Kluwer.
Supple, W. F., Jr., and Kapp, B. S. (1989). Response characteristics
Conclusion of neurons in the medial component of the medial geniculate
nucleus during Pavlovian differential fear conditioning in
So where is the occurrence, within the compo- rabbits. Behavioral Neuroscience 103, 1,2761,286.
nents of this circuit, of the convergence of CS and Teich, A. H., McCabe, P. M., Gentile, C. G., Jarrell, T. W., Winters,
UCS information that is necessary for the structural R. W., Liskowsky, D. R., and Schneiderman, N. (1988). Role
changes responsible for HR learning? The research of auditory cortex in the acquisition of differential heart rate
conditioning. Physiology and Behavior 44, 405412.
suggests that multiple components may represent
sites of structural change that depend on continual Bruce S. Kapp
cognitive and somatomotor behaviors. On the senso- William F. Supple
ry side, neurons within the MGm are responsive to Revised by Donald A. Powell
458 NEURAL SUBSTRATES OF CLASSICAL CONDITIONING: Discrete Behavioral
DISCRETE BEHAVIORAL RESPONSES tially causes no movement, while the US elicits move-
ment of the nictitating membrane (the NM, a third
Much of the research on how the brain codes and in- eyelid found in some species) and closure of the outer
fluences behavior has included studies of the neural eyelids. After 100 to 150 of these pairings, an NM or
bases of learning and memory. Physiological psychol-
eyelid movement is elicited by the CS even in trials
ogists, neurophysiologists, neuroanatomists, and
when the air puff or shock US is not presented. Be-
neuropsychologists have made major interdisciplin-
cause many of the parametric features of this behav-
ary contributions in this field, using a variety of ex-
ioral paradigm have been well documented by Gor-
perimental techniques.
mezano and associates, it has been adopted by a
One such productive research strategy in study- number of laboratories for use as a model system for
ing learning-related phenomena in the nervous sys- the study of the neural substrates of learning (see
tem is the use of simple mammalian models. The Steinmetz, 2000 and Steinmetz et al., 2001, for re-
model-systems approach assumes that the most views).
promising path to an understanding of complex
human learning and memory phenomena is the study Data from a variety of animal preparations and
of neural processes associated with simple learning from the human amnesia literature suggested that the
and memory tasks in nonhumans. The governing as- hippocampus, a limbic system structure, was involved
sumption in such work is that the manner in which the in coding learning and memory. Because of these ob-
brain codes learning and memory events is less com- servations, Thompson and associates attempted to as-
plex in simple learning tasks than in more elaborate sess hippocampal involvement in classical NM condi-
ones. This reduction in task complexity has made it tioning (e.g., Berger and Thompson, 1978).
possible to begin an analysis of the brain pathways Recordings from the hippocampus revealed neurons
and structures governing learning and memory pro- that altered their firing patterns during paired but
cesses. not unpaired presentations of the CS and US. Even
before CRs were observed, neurons in the hippocam-
Much useful data about brain substrates has pus began discharging in a pattern that preceded and
emerged from classical conditioning of discrete be-
modeled the amplitude and time course of the
havioral responses. This approach involves present-
learned behavioral response (i.e., the unit activity
ing a conditioned stimulus (CS) just before a second,
looked as though it could be producing the behavior-
unconditioned stimulus (US). Before training, the CS
al response). But lesions of the hippocampus failed to
typically elicits no overt response while the US reli-
abolish learning or retention of the simple motor re-
ably elicits a discrete, reflexive response called the un-
sponse, even when the lesions included all of the neo-
conditioned response (UR). After a number of CS-US
cortex as well as the hippocampus. These data indi-
pairings, however, the animal begins to execute the
cated that the hippocampus was probably involved in
discrete response after presentation of the CS but
coding the classically conditioned NM response but
before presentation of the US. This learned, antici-
that it was not essential for producing CRs. More re-
patory response is called the conditioned response
cent data suggest that in addition to possibly modu-
(CR). Many investigations of the neural substrates
lating the conditioning process during simple delay
of learning and memory have used this simple
conditioning, an intact hippocampus may be neces-
learning paradigm. The classical eyeblink-
sary for more complex classical conditioning prepara-
conditioning paradigm has significantly advanced
tions like trace conditioning and discrimination-
the understanding of how the brain encodes learning
reversal conditioning. Furthermore, studies in hu-
processes. Two eyeblink-conditioning preparations
mans suggest that awareness of the training
are presented here: nictitating membrane/eyeblink
contingencies may be an important determinant of
conditioning in the rabbit and eyeblink conditioning
the involvement of the hippocampus in conditioning
in the freely moving rat.
(e.g., Manns, Clark, and Squire, 2000).
The interpositus nucleus of the cerebellum
Nictitating Membrane/Eyelid Conditioning appears to be essential for classical conditioning
in the Rabbit (see Steinmetz, 2000 and Thompson, 1986, for
In the early 1960s, Gormezano developed a clas- reviews). The cerebellum is a structure that plays
sical conditioning preparation in rabbits that has a major role in motor control. Recordings from
proved valuable for the study of the neural bases of the interpositus nucleus as well as portions of
conditioning (see Gormezano, Kehoe, and Marshall, cerebellar cortex revealed populations of neurons
1983 and Woodruff-Pak and Steinmetz, 2000). Dur- that formed amplitude-time course models of
ing simple classical delay conditioning, a tone or light the CR during paired CS-US presentations.
CS is paired with a shock or air puff US. The CS ini- Furthermore, electrolytic or chemical lesions of the
NEURAL SUBSTRATES OF CLASSICAL CONDITIONING: Discrete Behavioral 459
interpositus nucleus (as small as one cubic millimeter) commutator that allows the rat free movement within
permanently abolished CRs in trained rabbits. Cere- the conditioning chamber.
bellar lesions before training prevented the forma-
Using rats instead of rabbits in eyeblink condi-
tion of CRs. The interpositus is known to output to
tioning studies has some advantages: rats are less
the red nucleus, which in turn sends projections to
costly to purchase and maintain; more is known about
brain-stem nuclei that control the musculature in-
the neuroanatomy of the rat; rats have a wider reper-
volved in generating NM movements and eyelid clo-
toire of other behaviors that can be studied at the
sure.
same time as eyeblink conditioning; and the shorter
Brain stimulation, recording, and lesion methods lifespan of rats is better suited to developmental and
have helped to delineate possible pathways involved aging studies.
in projecting the CS and US to the cerebellum. It ap- Studies concerning the basic circuitry underlying
pears that an acoustic CS may be projected to a num- eyeblink conditioning in the rat indicated that the
ber of primary brain-stem auditory nuclei that, in basic neural substrates of this form of conditioning in
turn, relay parallel projections to lateral regions of rats are nearly identical to those of the rabbit. The
the pontine nuclei. Cells in the lateral pontine nucle- cerebellum plays a critical role in encoding the learn-
ar regions may then project axons to the cerebellum. ing and memory of the response. Other brain areas,
The air puff US appears to be projected from the cor- like the hippocampus and neostriatum, are also in-
nea of the eye to the trigeminal nucleus in the brain volved in the conditioning process, and these regions
stem to the inferior olivary complex (also in the brain seem to play a modulatory role in the learning and
stem). Cells in the inferior olive then appear to send memory of this simple behavior.
axons to regions of the cerebellum. Data from tempo-
rary lesions (using the GABA agonist, muscimol, or The rat preparation has proved quite valuable for
brain cooling methods) and recording and stimu- two lines of research. First, Stanton and colleagues
lation studies have provided evidence that have very successfully used the rat eyeblink condition-
neuronal plasticity is established in cerebellar ing preparation to study neural and behavioral corre-
regions that receive convergent CS and US input lates of development of this simple form of learning.
and then relayed to brain-stem nuclei responsible For example, their elegant studies have shown that
for generating the motor response. Sites in the the development of conditioned responses parallels
interpositus nucleus and discrete regions of the closely the development of the cerebellum and relat-
cerebellar cortex receive this convergent input; ed brain circuits. Second, the rat eyeblink condition-
plasticity mechanisms at these sites are under inves- ing preparation has been used successfully to study
tigation. Figure 1 shows a schematic diagram of the behavioral and neurological effects of early alco-
neural circuitry that seems to play a role in classical hol exposure, a model of the human condition known
eyelid conditioning. as fetal alcohol syndrome. In this model, neonatal rats
are given binge levels of alcohol over a few days. Once
The rabbit classical NM conditioning paradigm the rats reach adulthood, researchers use eyeblink
has produced a wealth of data concerning the neural conditioning and neural recording and neuroanato-
substrates of a simple form of motor learning. The my methods to study the long-term effects of neonatal
careful control over stimulus presentation and re- alcohol exposure on behavioral and neural function.
sponse elicitation afforded by this learning prepara- Such studies indicate that neonatal alcohol exposure
tion has allowed the analysis of critical stimulus path- results in a permanent loss of neurons in the cerebel-
ways and potential regions of stimulus convergence, lar cortex and the deep cerebellar nuclei and that that
thus advancing the study of the cellular bases of this loss of these neurons, in turn, affects physiological
form of learning. and behavioral response during eyeblink condition-
ing, which requires the cerebellum.
Eyeblink Conditioning in the Freely
Moving Rat Conclusion
Rats have enjoyed a booming popularity as sub- The two different classical eyeblink conditioning
jects in more recent classical eyeblink conditioning discussed above have provided basic data on how the
experiments. Several laboratories have begun using brain codes the learning of simple behavioral re-
variations of a preparation described by Ronald Skel- sponses. Future work in this area will likely continue
ton (1988) to conduct classical eyeblink conditioning in two directions: further delineation of essential cel-
experiments in freely moving rats. This preparation lular processes that actually code the conditioning
uses tones or lights as CSs and periorbital stimulation process (e.g., possible mechanisms in the cerebellum
as a US. Connections to the rat are made through a and brain stem that account for the conditioning);
460 NEURAL SUBSTRATES OF CLASSICAL CONDITIONING: Discrete Behavioral
Figure 1
A schematic diagram depicting brain regions and circuitry proposed to be involved in classical eyeblink conditioning in rabbits and
rats. The (-) depicts synapses where inhibition occurs. All other synapses are thought to be excitatory. Note the convergence of CS- and
US-related information in the cerebellum.
and studies aimed at delineating the interactions be- and memory, especially by providing key data about
tween higher (e.g., cerebral cortex) and lower (e.g., how the brain codes simple learning tasks like classi-
brain stem and cerebellum) brain areas during classical conditioning of discrete responses.
cal conditioning. Genetic knockout and mutant prep-
arations, reversible brain inactivation methods, basic See also: ACTIVE AND PASSIVE AVOIDANCE
LEARNING: BEHAVIORAL PHENOMENA;
molecular biology techniques, used in conjunction
CLASSICAL CONDITIONING: BEHAVIORAL
with eyeblink classical conditioning should advance PHENOMENA; GUIDE TO THE ANATOMY OF THE
our understanding of the neurobiology of learning BRAIN: CEREBELLUM; NEURAL SUBSTRATES OF
NEURAL SUBSTRATES OF CLASSICAL CONDITIONING: Fear Conditioning, Freezing 461
CLASSICAL CONDITIONING: CARDIOVASCULAR except for breathing. Freezing is not passive, howev-
RESPONSES; NEURAL SUBSTRATES OF CLASSICAL er; it is a coordinated, protective defense against dan-
CONDITIONING: FEAR CONDITIONING, ger.
FREEZING; NEURAL SUBSTRATES OF CLASSICAL
CONDITIONING: FEAR-POTENTIATED STARTLE Why should a rat freeze and not run away when
threatened? The utility of freezing can be understood
Bibliography by observing the response of a cat to a ball suspended
Berger, T. W., and Thompson, R. F. (1978). Neuronal plasticity in on a piece of string. The cat will likely ignore the ball
the limbic system during classical conditioning of the rabbit if it is stationary. In contrast, if you wiggle the string,
nictitating membrane response. I. The hippocampus. Brain
Research 145, 323346.
the cat becomes vigilant and vigorously pounces or
Gormezano, I., Kehoe, E. J., and Marshall, B. S. (1983). Twenty paws the ball. This pouncing is predatory hunting be-
years of classical conditioning research with the rabbit. In J. havior. The moving ball elicits this behavior for two
M. Sprague and A. N. Epstein, eds., Progress in physiological reasons: the cats visual system is very good at detect-
psychology, Vol. 10, pp. 197275. New York: Academic Press. ing movement, and movement triggers predatory at-
Green, J. T., Rogers, R. F., Goodlett, C. R. and Steinmetz, J. E.
(2000). Impairment in eyeblink classical conditioning in adult
tack. Freezing evolved to counter the predators sen-
rats exposed to ethanol as neonates. Alcoholism: Clinical and sitivity to movement. Prey species freeze because
Experimental Research 24, 438447. many predators see stationary objects poorly, and
Manns, J. R., Clark, R. E., and Squire, L. R. (2000). Parallel acquisi- they are less likely to detect immobile prey. Hence, in
tion of awareness and trace eyeblink classical conditioning. the vicinity of a predator, a still rat is a safer rat.
Skelton, R. W. (1988). Bilateral cerebellar lesions disrupt condi- Animals typically spend more time freezing as the
tioned eyelid responses in unrestrained rats. Behavioral Neuro- perceived danger increases. Therefore, freezing is a
science 102, 586590.
useful index of fear. For example, a rat will spend
Stanton, M.A., and Freeman, Jr., J. H. (2000). Developmental
studies of eyeblink conditioning in the rat. In D. S. Woodruff- more time freezing in a chamber in which it received
Pak and J. E. Steinmetz, eds., Eyeblink classical conditioning, five aversive electric shocks than in a chamber in
Vol. 2: Animal models. Boston: Kluwer. which it received one.
Steinmetz, J. E. (2000). Brain substrates of classical eyeblink condi-
tioning: A highly localized but also distributed system. Be- Freezing is best measured with time sampling. In
havioural Brain Research 110, 1324. this procedure, at regular intervals of a few seconds,
Steinmetz, J. E., Gluck, M. A., and Solomon, P. R., eds. (2001). a trained observer makes a judgment as to whether
Model systems and the neurobiology of associative learning: A fest-
the animal is freezing or not at a given instant. The
schrift in honor of Richard F. Thompson. Mahwah, NJ: Erlbaum.
Thompson, R. F. (1986). The neurobiology of learning and memo- percentage of instantaneous samples scored as freez-
ry. Science 233, 941947. ing provides a probability estimate of the behavior
Woodruff-Pak, D. S., and Steinmetz, J. E., eds. (2000). Eyeblink clas- that is appropriate for statistical analysis. This sam-
sical conditioning, Vol. 2: Animal models. Boston: Kluwer. pling procedure is accurate, reliable, and reproduc-
Joseph E. Steinmetz ible.
Pavlovian Fear Conditioning

FEAR CONDITIONING, FREEZING Fear is rapidly learned and measured in the labo-
When threatened, some animals simply freeze. This ratory with a procedure called Pavlovian fear condi-
defensive behavior has been studied in greatest detail tioning. This method has become a standard means
in rats. Freezing behavior has been one index of fear of exploring the behavioral processes and neural
in Pavlovian conditioning experiments investigating mechanisms of learning. In a typical Pavlovian fear-
fear learning and memory. This article describes the conditioning procedure, a rat is placed in a chamber
adaptive value of freezing, its use in Pavlovian fear where it is presented with a tone that is followed by
conditioning, and its neural substrates. a brief but mildly aversive foot shock. Later, during
a test session, the rat is reexposed to either the cham-
ber or the tone. During this reexposure the rat will
Freezing Behavior display fear. With this preparation, the tone and the
Animals have developed numerous behaviors chamber serve as conditional stimuli (CS). They were
that help them avoid threats and danger. For exam- originally neutral signals that did not predict danger,
ple, when a deer mouse encounters a predator such but after they were paired with an unconditional stim-
as a snake, it may attempt to run away. Or it might try ulus (US)the foot shockthe animal responds fear-
to fend off the predator by biting or fighting with it. fully to the CS. Such a response to the CS is called a
When threatened by predators or dominant conspe- conditional response (CR); freezing is an example of
cifics, rats, mice, squirrels, and other prey species also a CR. The shock-paired stimuli trigger CRs, which
freeze, often in a crouch, utterly devoid of movement are measures of learning in Pavlovian experiments.
462 NEURAL SUBSTRATES OF CLASSICAL CONDITIONING: Fear Conditioning, Freezing
Brain Circuit of Fear Conditioning and CEA, which in turn projects to a variety of structure
Freezing that include the periaqueductal gray (PAG), the retic-
ular formation, and the lateral hypothalamus. Dam-
Scientists use several experimental techniques to
age to the CEA disrupts the expression of a wide
characterize brain circuits. Lesion studies seek to
range of defensive behaviors, including freezing.
damage a brain structure with surgical techniques. In
this procedure a scientist may inject a neurotoxin into The periaqueductal gray, or PAG, is highly inter-
a specific target region to kill the cells there. After the connected with the CEA. This region coordinates de-
animal recovers from surgery, it is tested to deter- fensive behaviors. Chemical or electrical stimulation
mine what effect the lesion has on its behavior. Infu- of the dorsal-lateral PAG (dlPAG) triggers bursts of
sion studies seek to temporarily alter function in a tar- forward movements resembling flight. Damaging this
get brain structure through the injection of chemicals region disrupts flightlike behavior. Consequently, the
through surgically implanted injectors mounted to dlPAG seems to coordinate defensive reactions like
the animals skull. Drugs can be infused before, dur- flight. In contrast, chemical or electrical stimulation
ing, or after a behavioral treatment. This technique of the ventral PAG (vPAG) triggers freezing, and
has the advantage of being temporary and reversible. damage to this structure disrupts it. Interconnections
Electrophysiological studies seek to measure the elec- between these areas may allow the animal to rapidly
trical properties of targeted neurons associated with shift between freezing and flight.
specific behaviors such as freezing. The hippocampus is also plays a role in the ex-
Rat, humans, and other mammals share funda- pression of some types of fear behavior. When a rat
mentally similar brain circuits that underlie fear. In- is trained with a tone-shock pairing inside a chamber,
deed, behavioral neuroscientists have described a set the animals will later display fear of both the tone and
of interconnected brain regions that constitute a fear chamberthe context. However, if the animals dor-
circuit that is similar in humans, rats, mice, rabbits, sal hippocampus is damaged after training, it will
and monkeys. In all of these species a structure called freeze in response to the tone but not the context.
the amygdala is a prominent component of the fear Presumably the hippocampus plays a role in recogni-
circuit tion of the chamber. An important aspect of context
fear is that the effect of the hippocampal damage
The amygdala is composed of a set of intercon- grades with time. That is, if the hippocampus is dam-
nected clusters of neurons called nuclei that lie in the aged one day after the training session, the amount
interior portion of the temporal lobe. The Pavlovian of contextual fear disruption is large; if the damage
fear-conditioning paradigm has shown that three nu- occurs two months after the training session, the
clei within the amygdala make major contributions to amount of contextual fear disruption is small. This
fear behavior: the lateral (LA), basal (BA), and central property of contextual fear is similar to the retro-
nuclei (CEA). The LA and BA nuclei make up the grade amnesia seen in human patients who have hip-
frontotemporal amygdala (FTA). This part of the pocampal damage: these patients cannot recall events
amygdala communicates most closely with the frontal immediately before their brain damage, but they can
and temporal lobes of the brain, and it is important recall events from years before.
for fear learning.
Moreover, critical components of the memory es- Conclusion
tablished during fear conditioning are located in the
FTA. First, the FTA is connected to auditory, visual, Freezing, a defensive response common in ro-
olfactory, and brain regions that govern pain sensa- dents and other prey animals, has been employed as
tion. Thus, sensory information of the CS and pain in- an index of fear in Pavlovian conditioning experi-
formation of the US converge in the FTA. Second, ments that investigate fear learning and memory.
Pavlovian fear conditioning enhances the response of The amygdala plays a role in fear responses in all
cells in the FTA that respond to tone CSs. Third, mammalian species; contextual fear shares properties
damage to the FTA produces a pronounced and often with human declarative memory. Thus, freezing be-
total loss of many Pavlovian fear responses such as havior and Pavlovian fear conditioning provide a use-
freezing. Fourth, chemical inactivation of this struc- ful animal model of learning and memory.
ture disrupts fear learning. Thus, the FTA is critical
for learning fear, and it may be the locus of the estab- See also: GUIDE TO THE ANATOMY OF THE BRAIN:
lishment of fear memory. AMYGDALA; GUIDE TO THE ANATOMY OF THE
BRAIN: HIPPOCAMPUS AND PARAHIPPOCAMPAL
The CEA may be the output of the amygdala. It REGION; NEURAL SUBSTRATES OF AVOIDANCE
is closely tied with the striatum and specializes in LEARNING; NEURAL SUBSTRATES OF CLASSICAL
modulating motor outflow. The FTA projects to the CONDITIONING: CARDIOVASCULAR RESPONSES;
NEURAL SUBSTRATES OF CLASSICAL CONDITIONING: Fear-Potentiated Startle 463
NEURAL SUBSTRATES OF CLASSICAL Neural Systems Involved in Fear-

CONDITIONING: FEAR-POTENTIATED STARTLE; Potentiated Startle
A major advantage of the fear-potentiated startle
paradigm is that fear is measured by a change in a
Bibliography simple reflex. Hence, with potentiated startle, fear is
Edmunds, M. (1974). Defence in animals: A survey of antipredator de-
expressed through some neural pathway(s) activated
fences. Burnt Mill, England: Longman.
Fanselow, M. S. (1994). Neural organization of the defensive be- by the conditioned stimulus that connects to the star-
havior system responsible for fear. Psychonomic Bulletin & Re- tle pathway. Figure 1 shows a schematic summary dia-
view 1, 429438. gram of the neural pathways we believe are required
(2000). Contextual fear, gestalt memories, and the hippo- for fear-potentiated startle, given a visual conditioned
campus. Behavioural Brain Research 110, 7381.
stimulus and foot shock as the unconditioned stimu-
Fanselow, M. S., and LeDoux, J. E. (1999). Why we think plasticity
underlying Pavlovian fear conditioning occurs in the baso- lus.
lateral amygdala. Neuron 23, 229232.
Fendt, M., and Fanselow, M. S. (1999). The neuroanatomical and
neurochemical basis of conditioned fear. Neuroscience and Bio- The Acoustic Startle Pathway
behavioral Reviews 23, 743760. In the rat, the latency of acoustic startle is six mil-
Squire, L. R., Clark, R. E., and Knowlton, B. J. (2001). Retrograde
amnesia. Hippocampus 11, 5055.
liseconds, recorded electromyographically in the
Swanson, L. W., and Petrovich, G. D. (1998). What is the amygdala? foreleg, and eight milliseconds in the hind leg. This
Trends in Neurosciences 21, 323331. very short latency indicates that only a few synapses
Bill P. Godsil
can be involved in mediating acoustic startle. Using
Michael S. Fanselow
a variety of techniques, we showed that acoustic startle
was mediated by a pathway that includes auditory
neurons embedded in the auditory nerve called coch-
lear root neurons, an area just dorsal to the superior
FEAR-POTENTIATED STARTLE olives in the nucleus reticularis pontis caudalis, and
motoneurons in the spinal cord. Bilateral chemical le-
When a stimulus such as a light, which engenders sions of these cell groups eliminate startle, whereas le-
little behavioral effect before pairing, is paired with sions in a variety of other auditory or motor areas do
an aversive stimulus such as a foot shock, the light not. Startlelike responses can be elicited electrically
(conditioned stimulus) can elicit seemingly fearful re- from each of these nuclei, with progressively shorter
sponses in animals: autonomic changes, freezing, and latencies as the electrode is moved down the pathway.
an increase in the amplitude of the startle reflex elicit-
ed by an auditory stimulus in the presence of the
light. The last is called the fear-potentiated startle effect Where Does Fear Activate the Startle
and can occur with an auditory, visual, tactile, or ol- Pathway?
factory conditioned stimulus under conditions where By eliciting startlelike responses electrically from
startle is elicited by either a loud sound or an air puff. various points along the startle pathway in the pres-
Fear-potentiated startle is a valid measure of clas- ence and absence of a light previously paired with a
sical conditioning because it occurs only following shock, we concluded that fear ultimately alters trans-
paired rather than unpaired or random presenta- mission at the nucleus reticularis pontis caudalis. In-
tions of the conditioned stimulus. Potentiated startle jection of retrograde tracers into this part of the star-
shows considerable temporal specificity because its tle pathway indicated that it receives direct
magnitude in testing is greatest at the interval after projections from the central nucleus of the amygdala,
light onset that matches the light-shock interval in an area of the brain long implicated in fear, as well
training. This paradigm offers a number of advan- as from an area in the mesecenphalic reticular forma-
tages as an alternative to most animal tests of fear or tion and deep layers of the superior colliculus.
anxiety because it involves no operant and is reflected
by an enhancement rather than a suppression of con-
tinuing behavior. Drugs like clonidine, morphine,
Lesions of the Amygdala Block Fear-
diazepam, and buspirone, which differ in their mech- Potentiated Startle
anism of action yet all reduce fear or anxiety in hu- Chemical lesions of either the basolateral or cen-
mans, decrease potentiated startle in rats. Conversely, tral nucleus of the amygdala following fear condition-
drugs like yohimbine, piperoxane, and B-carbolines, ing completely eliminate potentiated startle. In con-
which induce anxiety in normal people and exagger- trast, lesions of a variety of other brain areas,
ate it in anxious people, increase the magnitude of including the frontal cortex, insular cortex, visual cor-
fear-potentiated startle in rats. tex, hippocampus, septal nuclei, superior colliculus,
464 NEURAL SUBSTRATES OF CLASSICAL CONDITIONING: Fear-Potentiated Startle
red nucleus, and cerebellum, do not. Low-level elec- and the posterior intralaminar nuclei of the thalamus,
trical stimulation of the amygdala markedly increases but not lesions of either structure alone, blocked the
acoustic startle amplitude at stimulus currents and acquisition of fear-potentiated startle. However, post-
durations that do not produce any other signs of be- training combined lesions of these areas together did
havioral activation. not prevent expression of conditioned fear. These re-
sults suggest that parallel cortical and subcortical
pathways are involved in relaying shock information
Role of Different Amygdala Efferent
during fear conditioning.
Projections in Fear-Potentiated Startle
The pathway between the central nucleus of the
amygdala and the part of the nucleus reticularis Role of the Anterior Perirhinal and Insular
pontis caudalis that is critical for startle involves the Cortex in Fear-Potentiated Startle
caudal division of the ventral amygdalofugal pathway,
The perirhinal cortex, which receives either visu-
which also sends collaterals to many brain-stem target
al or auditory conditioned stimulus information,
areas involved in the somatic and autonomic symp-
projects directly to the lateral and basolateral amyg-
toms of fear and anxiety. Lesions along this pathway
completely block potentiated startle. In contrast, le- dala. Posttraining lesions of the anterior perirhinal
sions of other major projections from the central nu- cortex completely blocked the expression of fear-
cleus of the amygdala do not. In addition to this direct potentiated startle when a visual conditioned stimulus
pathway, an indirect pathway between the central nu- is used, provided the lesion destroyed both the dy-
cleus of the amygdala and the deep superior col- granular and agranular portions of the perirhinal
liculus/mesecephalic reticular formation is required cortex. Posttraining lesions of the perirhinal area (in-
for fear-potentiated startle because inactivation of cluding secondary auditory cortices) blocked fear-
this region completely blocked expression but not ac- potentiated startle to both an auditory and visual con-
quisition of fear-potentiated startle using a visual conditioned stimulus. However, reliable potentiated star-
ditioned stimulus. tle was observed after retraining in animals sustaining
main geniculate body lesions (which would destroy
cortical connections between the thalamus and per-
Convergence of Light and Shock Input at irhinal cortex) or following pretraining lesions of the
the Amygdala perirhinal area. These data suggest that cortical areas
Figure 1 shows that visual information reaches normally are used for the expression of fear condi-
the amygdala by way of two parallel pathways. One in- tioning but that subcortical areas can take over if the
volves direct retinal inputs to the lateral posterior nu- cortex is damaged. Finally, as mentioned before,
cleus of the thalamus, which projects directly to the shock information seems to require the insular cor-
basolateral amygdala and perirhinal cortex. The tex, which in turn projects directly to the lateral nu-
other involves retinal inputs to the dorsal lateral ge- cleus of the amygdala.
niculate nucleus, which projects indirectly to the per-
irhinal cortex via the visual cortex. Lesions of both of
these visual thalamic nuclei together, but not either Glutamate Receptors in the Amygdala Are
one alone, blocked fear-potentiated startle when a vi- Critical for Fear-Potentiated Startle
sual, but not an olfactory, conditioned stimulus was
Because conditioned and unconditioned stimulus
used. When an auditory conditioned stimulus is used,
information converge at the lateral and basolateral
this involves parallel inputs from the auditory thala-
mus to the perirhinal cortex either directly or indi- amygdala nuclei, this could be the site of plasticity for
rectly via the auditory cortex. Pretraining or post- fear-potentiated startle. In fact, local infusion of
training lesions of the entire auditory thalamus NMDA antagonists into the amygdala blocked the ac-
completely blocked fear-potentiated startle to an au- quisition but not the expression of fear-potentiated
ditory but not to a visual conditioned stimulus. Post- startle when a visual, auditory, or olfactory condi-
training lesions restricted to the main body of the me- tioned stimulus was used. This blockade of fear acqui-
dial geniculate, which projects to the perirhinal sition probably was not the result of preventing shock
cortex via auditory cortex, also specifically blocked information from getting to the amygdala because
fear-potentiated startle to the auditory CS. Pain infor- this treatment also blocked the acquisition of second
mation reaches the amygdala via parallel pathways order fear-potentiated startle, which does not involve
that include the caudal granular/dysgranular insular shock during second order training. Following condi-
cortex and the posterior intralaminar nuclei of the tioning, local infusion of non-NMDA antagonists
thalamus. Pretraining lesions of both insular cortex block the expression of fear-potentiated startle.
NEURAL SUBSTRATES OF CLASSICAL CONDITIONING: Fear-Potentiated Startle 465
Figure 1
Schematic diagram of the neural pathways believed to be involved in the fear-potentiated startle effect using a visual conditioned
stimulus and a footshock unconditioned stimulus.
Fear-Potentiated Startle in Humans anxiety. Conditioned fear appears to result when a

formerly neutral stimulus comes to activate the amyg-
Fear-potentiated startle also can be measured in
dala after being paired with an aversive stimulus. Acti-
humans. In one test people are told that when a cer-
vation of the central nucleus of the amygdala in-
tain colored light comes on, they might get a shock
creases startle via direct and indirect connections
on the wrist, whereas when a different-colored light
between the amygdala and a specific point in the
comes on, they will not get a shock. Startle is elicited
acoustic startle pathway. More generally, the central
with bursts of noise through earphones, and the eye-
nucleus of the amygdala and its efferent projections
blink component of startle is measured electromyo-
to the brain stem may constitute a central fear system
graphically from the obicularis oculi muscles. Startle
that produces a constellation of fearlike behaviors in
amplitude was consistently higher in the presence of
animals and people. Finally, the acquisition of condi-
the light that signals shock. The size of this increase
tioned fear may involve an NMDA-dependent pro-
depended on when the subject expected the shock
cess at the level of the amygdala.
based on verbal instructions; the increase was also evi-
dent using conditioning procedures. Brain imaging See also: NEURAL SUBSTRATES OF CLASSICAL
studies show that the amygdala is activated during CONDITIONING: CARDIOVASCULAR RESPONSES;
this verbally mediated fear-potentiated startle test NEURAL SUBSTRATES OF CLASSICAL
and that patients with lesions of the amygdala failed CONDITIONING: DISCRETE BEHAVIORAL
to display fear-potentiated startle. People also show RESPONSES; NEURAL SUBSTRATES OF CLASSICAL
an increase in startle when they see scary pictures, CONDITIONING: FEAR CONDITIONING,
such as a gun in the face, a dog about to bite, or muti- FREEZING; NEURAL SUBSTRATES OF EMOTIONAL
MEMORY; PASSIVE (INHIBITORY) AVOIDANCE,
lated bodies. The size of the increase in startle was di-
FEAR LEARNING
rectly related to the subjects degree of negative va-
lence and arousal. Bibliography
Berg, W. K., and Davis, M. (1985). Associative learning modifies
startle reflexes at the lateral lemniscus. Behavioral Neuroscience
Conclusion 99, 191199.
Brown, J. S., Kalish, H. I., and Farber, I. E. (1951). Conditional fear
The fear-potentiated startle paradigm has been as revealed by magnitude of startle response to an auditory
useful for elucidating neural substrates of fear and stimulus. Journal of Experimental Psychology 41, 317328.
466 NEURAL SUBSTRATES OF EMOTIONAL MEMORY
Campeau, S., and Davis, M. (1995a). Involvement of subcortical Rosen, J. B., Hitchcock, J. M., Miserendino, M. J. D., Falls, W. A.,
and cortical afferents to the lateral nucleus of the amygdala Campeau, S., and Davis, M. (1992). Lesions of the perirhinal
in fear conditioning measured with fear-potentiated startle in cortex but not of the frontal, medial prefrontal, visual, or in-
rats trained concurrently with auditory and visual conditioned sular cortex block fear-potentiated startle using a visual con-
stimuli. Journal of Neuroscience 15, 2,3122,327. ditioned stimulus. Journal of Neuroscience 12, 4,6244,633.
(1995b). Involvement of the central nucleus and basolateral Sananes, C. B., and Davis, M. (1992). N-Methyl-D-Aspartate le-
complex of the amygdala in fear conditioning measured with sions of the lateral and basolateral nuclei of the amygdala
fearpotentiated startle in rats trained concurrently with audi- block fear-potentiated startle and shock sensitization of star-
tory and visual conditioned stimuli. Journal of Neuroscience 15, tle. Behavioral Neuroscience 106, 7280.
2,3012,311. Shi, C. J., and Davis, M. (1999). Pain pathways involved in fear con-
Davis, M. (2000). The role of the amygdala in conditioned and un- ditioning measured with fear-potentiated startle: Lesion
conditioned fear and anxiety. In J. P. Aggleton, ed., The amyg- studies. Journal of Neuroscience 19, 420430.
dala, Vol. 2. Oxford, UK: Oxford University Press. (2001). Visual pathways involved in fear conditioning mea-
Davis, M., and Astrachan, D. I. (1978). Conditioned fear and startle sured with fear-potentiated startle: Behavior and anatomic
studies. Journal of Neuroscience 21, 9,8449,855.
magnitude: Effects of different footshock or backshock inten-
Walker, D. L., and Davis, M. (2002). The role of glutamate recep-
sities used in training. Journal of Experimental Psychology: Ani-
tors within the amygdala in fear learning, fear-potentiated
mal Behavior Processes 4, 95103.
startle, and extinction. Pharmacology, Biochemistry, and Behavior
Davis, M., Falls, W. A., Campeau, S., and Kim, M. (1993). Fear-
71, 379392.
potentiated startle: A neural and pharmacological analysis.
Behavior Brain Research 58, 175198. Michael Davis
Davis, M., Gendelman, D. S., Tischler, M. D., and Gendelman, P.
M. (1982). A primary acoustic startle circuit: Lesion and stim-
ulation studies. Journal of Neuroscience 6, 791805.
Davis, M., Schlesinger, L. S., and Sorenson, C. A. (1989). Temporal
specificity of fear-conditioning: Effects of different condi-
tioned stimulus-unconditioned stimulus intervals on the fear-
NEURAL SUBSTRATES OF
potentiated startle effect. Journal of Experimental Psychology: EMOTIONAL MEMORY
Animal Behavior Processes 15, 295310.
Falls, W. A., and Davis, M. (1994). Fear-potentiated startle using
Emotions are an integral part of our psychological
three conditioned stimulus modalities. Animal Learning and life. For years, neuroscientists have largely ignored
Behavior 22, 379383. emotion, in part because it was believed that emotions
Funayama, E. S., Grillon, C., Davis, M., and Phelps, E. A. (2001). were difficult to objectively define and measure, and
A double dissociation in the affective modulation of startle in therefore were outside the realm of legitimate scien-
humans: Effects of unilateral temporal lobectomy. Journal of
tific investigation. In recent years, however, great
Cognitive Neuroscience 13, 721729.
Gewirtz, J., and Davis, M. (1997). Second order fear conditioning
strides have been made in our understanding the
prevented by blocking NMDA receptors in the amygdala. Na- brain pathways and structures underlying one espe-
ture 388, 471474. cially important emotion: fear. In particular, much
Grillon, C., Amelia, R., Merikangas, K., Woods, S. W., and Davis, has been learned about how the brain learns to fear
M. (1993). Measuring the time course of anticipatory anxiety an object or situation, how learned fears can guide the
using the fear-potentiated startle reflex. Psychophysiology 30,
acquisition of behaviors that are instrumental in
340346.
Grillon, C., and Davis, M. (1997). Fear-potentiated startle condi- avoiding danger, and how fear can augment the
tioning in humans: Effects of explicit and contextual cue con- strength of memory formation of significant life
ditioning following paired versus unpaired training. Psycho- events.
physiology 34, 451458.
Lang, P. J., Bradley, M. M., and Cuthbert, B. N. (1990). Emotion,
attention, and the startle reflex. Psychology Review 97, 377 Classical Fear Conditioning
395.
Lee, Y., Lopez, D. E., Meloni, E. G., and Davis, M. (1996). A prima-
The fear learning system of the brain has been
ry acoustic startle circuit: Obligatory role of cochlear root neu- most extensively studied in the laboratory rat using a
rons and the nucleus reticularis pontis caudalis. Journal of simple, robust form of associative learning known as
Neuroscience 16, 3,7753,789. classical or Pavlovian fear conditioning (LeDoux,
Meloni, E. G., and Davis, M. (1999). Muscimol in the deep layers 2000; 2002). In this behavioral paradigm, an animal
of the superior colliculus/mesencephalic reticular formation
learns to respond defensively to an initially neutral
blocks expression but not acquisition of fearpotentiated star-
tle in rats. Behavioral Neuroscience 113, 1,1521,160.
stimulus (the conditioned stimulus; CS) after it has been
Paschall, G. Y., and Davis, M. (2002). Olfactory mediated fear po- associated or paired with a noxious stimulus (the un-
tentiated startle. Behavioral Neuroscience 116, 412. conditioned stimulus; US). In rats, presentation of the
Phelps, E. A., OConnor, K. J., Gatenby, J. C., Gore, J. C., Grillon, CS after conditioning elicits defensive behavior
C., and Davis, M. (2001). Activation of the left amygdala to a (freezing) and supporting autonomic and endocrine
cognitive representation of fear. Natural Neuroscience 4, 437
adjustments. This is an implicit form of memory,
441.
Rosen, J. B., and Davis, M. (1988). Enhancement of acoustic startle which means that the brain system involved functions
by electrical stimulation of the amygdala. Behavioral Neuro- independently of conscious mediation (LeDoux,
science 102, 195202. 2000, 2002). Nevertheless, conscious (or explicit)
NEURAL SUBSTRATES OF EMOTIONAL MEMORY 467
memories of the emotional experience can and often contiguity, a hallmark of associative learning. Finally,
are formed in parallel. amygdala LTP and fear conditioning have been
shown to be subserved by similar biochemical or mo-
In auditory fear conditioning, the most thor-
lecular mechanisms (Schafe, Nader, Blair, and Le-
oughly studied form of implicit fear learning, a tone
Doux, 2001).
(CS) is paired with brief electrical shock to the feet
(US). This form of fear conditioning involves trans- The LA is also critically involved in the expres-
mission of auditory and somatic sensory information sion of conditioned fear by way of its projections to
to the lateral nucleus of the amygdala (LA), an area the nearby central nucleus of the amygdala ([CE];
that is critical for fear conditioning. During fear con- Amorapanth, LeDoux, and Nader, 2000; Nader, Ma-
ditioning, individual neurons in the LA are well jidishad, Amorapanth, and LeDoux, 2001; Pitknen,
suited to integrate information about tone and foot Savander, and LeDoux, 1997). The CE is known to
shock. Cells in the LA, for example, receive direct project to areas of the forebrain, hypothalamus, and
projections from areas of the auditory thalamus and brain stem that control behavioral, endocrine, and
cortex. Inputs from each of these auditory areas con- autonomic conditioned responses (CRs) associated
verge onto single neurons in the LA, and these same with fear learning. Projections from the CE to the
cells are also responsive to the foot shock US. Further, midbrain periacqueductal gray, for example, have
auditory fear conditioning is disrupted either by per- been shown to be particularly important for mediat-
manent lesions or reversible functional inactivation ing behavioral and endocrine responses such as freez-
targeted to the LA and adjacent areas (LeDoux, 2000; ing and hypoalgesia (De Oca, DeCola, Maren, and
2002; Maren, 2001). Fanselow, 1998; Helmstetter and Tershner, 1994),
The LA is not only the principle site of sensory whereas projections to the lateral hypothalamus have
input in the amygdala; it also appears to be an essen- been implicated in the control of conditioned cardio-
tial locus of plasticity during fear conditioning. For vascular responses (Iwata, LeDoux, and Reis, 1986).
example, individual cells in the LA alter their re- Importantly, while lesions of these individual areas
sponse properties when CS and US are paired during can selectively impair expression of individual CRs,
fear conditioning; specifically, LA neurons that are damage to the CE interferes with the expression of all
initially weakly responsive to auditory input respond fear CRs (LeDoux, 2000, 2002). Thus, the CE is typi-
vigorously to the same input after fear conditioning cally thought of as the principal output nucleus of the
(Quirk, Armony, Repa, Li, and LeDoux, 1997; Repa fear system that acts to orchestrate the collection of
et al., 2001). Thus, a change occurs in the function of hardwired, and typically species-specific, responses
LA cells as the result of training, a finding that has that underlie defensive behavior.
contributed to the view that neural plasticity in the LA
encodes key aspects of fear learning and memory
storage (Blair et al., 2001; Fanselow and LeDoux,
Instrumental Fear Learning
1999; Maren, 1999; Quirk et al., 1997). In addition to its role in Pavlovian fear condition-
ing, the amygdala contributes to other fear-related as-
What mechanism may mediate the change that pects of behavior. For example, Pavlovian fear condi-
occurs in the LA as a result of conditioning? One of tioning is useful for learning to detect a dangerous
the leading candidates is long-term potentiation (LTP), object or situation, but the animal must also be able
an activity-dependent form of plasticity initially dis- to use this information to guide ongoing behavior
covered in the hippocampus (Bliss and Lomo , 1973). that is instrumental in avoiding that danger. In some
LTP is an attractive candidate for a cellular mecha- situations, the animal must learn to make a response
nism of learning and memory during fear condition- (i.e., move away, press a bar, or turn a wheel) that will
ing for several reasons. For one, LTP has been ob- allow it to avoid presentation of a shock or danger sig-
served in each of the major sensory input pathways to nal, a form of learning known as active avoidance. In
the LA, including the thalamic and cortical auditory other situations, the animal must learn not to re-
pathways (Maren, 1999, 2001). Second, fear condi- spond, also known as passive avoidance. Both of these
tioning itself has been shown to lead to electrophy- are examples of instrumental conditioning, and the
siological changes in the LA similar to those observed amygdala plays a vital role in each.
following artificial LTP induction, and these changes
persist over days (Maren, 1999). Third, associative The role of different amygdala nuclei, including
LTP in the LA has been shown to be sensitive to the the LA, CE, and the basal nucleus, in learning tasks
same contingencies as fear conditioning (Bauer, Le- that involve both classical and instrumental fear
Doux, and Nader, 2001). That is, LTP in the LA ap- learning components has recently been examined
pears to depend on the contingency between pre- and (Amorapanth, LeDoux, and Nader, 2000; Killcross,
postsynaptic activity rather than simply on temporal Robbins, and Everitt, 1997). In one such study, for ex-
468 NEURAL SUBSTRATES OF EMOTIONAL MEMORY
ample, the animal was first trained to associate a tone tional memory processes. The amygdala, and particu-
with foot shock (the Pavlovian component). Next, the larly the LA, participates both in reflexive or reactive
animal learned to move from one side of a two- forms of fear learning, where the animal learns to fear
compartment box to the other to avoid presentation one stimulus that has been associated with another.
of the tone (the instrumental component). In summa- However, the amygdala also appears to participate in
ry, while lesions of the LA appear to impair both the instrumental or active forms of fear learning, where
classical and instrumental aspects of fear learning, le- the animal must learn to cope behaviorally in the
sions of the CE impair only the Pavlovian component presence of fearful stimuli. In addition, the amygdala
(i.e., the tone-shock association). Conversely, lesions modulates the formation of memories in other sys-
of the basal nucleus impair only the instrumental tems of the brain, such as systems involved in explicit
component (learning to move to the second compart- or conscious memory.
ment). Thus, different outputs of the LA appear to
See also: GUIDE TO THE ANATOMY OF THE BRAIN;
mediate Pavlovian and instrumental behaviors elicit-
LONG-TERM POTENTIATION: AMYGDALA;
ed by a fear-arousing stimulus (Amorapanth, Le- NEURAL SUBSTRATES OF CLASSICAL
Doux, and Nader, 2000; Nader and LeDoux, 1998). CONDITIONING; PASSIVE (INHIBITORY)
This is not to say, however, that the basal nucleus is AVOIDANCE, FEAR LEARNING
a site of motor control or a locus of memory storage
for instrumental learning. Rather, it likely guides Bibliography
fear-related behavior and reinforcement via its pro- Amorapanth, P., LeDoux, J. E., and Nader, K. (2000). Different lat-
eral amygdala outputs mediate reactions and actions elicited
jections to nearby striatal regions that are known to
by a fear-arousing stimulus. Nature Neuroscience 3 (1), 7479.
be necessary for reinforcement learning (Everitt, Bauer, E. P., LeDoux, J. E., and Nader, K. (2001). Fear condition-
Cador, and Robbins, 1989; Everitt et al., 1999; Rob- ing and LTP in the lateral amygdala are sensitive to the same
bins, Cador, Taylor, and Everitt, 1989). stimulus contingencies. Nature Neuroscience 4, 687688.
Blair, H. T., Schafe, G. E., Bauer, E. P., Rodrigues, S. M., and Le-
In spite of the fact that fear conditioning itself is Doux, J. E. (2001). Synaptic plasticity in the lateral amygdala:
an implicit form of learning and memory, during A cellular hypothesis of fear conditioning. Learning and Memo-
most emotional experiences explicit or conscious ry 8, 229242.
Bliss, T. V. P., and Lomo , T. (1973). Long-lasting potentiation of
memories are also formed (LeDoux, 2000; 2002). synaptic transmission in the dentate area of the anesthetized
These occur through the operation of the medial rabbit following stimulation of the perforant path. Journal of
temporal lobe memory system involving the hippo- Psychology 232, 331356.
campus and related cortical areas (Eichenbaum, De Oca, B. M., DeCola, J. P., Maren, S., and Fanselow, M. S.
2000; Milner, Squire, and Kandel, 1998). The role of (1998). Distinct regions of the periaqueductal gray are in-
volved in the acquisition and expression of defensive re-
the hippocampus in the explicit memory of an emo- sponses. Journal of Neuroscience 18 (9), 3,4263,432.
tional experience is similar to its role in other kinds Eichenbaum, H. (2000). A cortical-hippocampal system for declar-
of experiences, with one important exception. During ative memory. Nature Reviews Neuroscience 1, 4150.
fearful emotional experiences, the amygdala activates Everitt, B. J., Cador, M., and Robbins, T. W. (1989). Interactions
between the amygdala and ventral striatum in stimulus-
neuromodulatory systems in the brain and hormonal
reward associations: Studies using a second-order schedule of
systems in the brain and body. Chemicals released by sexual reinforcement. Neuroscience 30, 63-75.
these systems modulate the function of forebrain Everitt, B. J., Parkinson, J. A., Olmstead, M. C., Arroyo, M., Roble-
areas such as the hippocampus and serve to enhance do, P., and Robbins, T. W. (1999). Associative processes in ad-
the storage of the memory in these areas (McGaugh, diction and reward: The role of amygdala-ventral striatal sub-
systems. Annual of the New York Academy of Science 877, 412
2000). The primary support for this model comes
438.
from studies of inhibitory avoidance learning, where Fanselow, M. S., and LeDoux, J. E. (1999). Why we think plasticity
the animal must learn not to enter a chamber in which underlying Pavlovian fear conditioning occurs in the baso-
it has previously received shock. In this paradigm, lateral amygdala. Neuron 23 (2), 229232.
various pharmacological manipulations of the amyg- Helmstetter, F. J., and Tershner, S. A. (1994). Lesions of the per-
iaqueductal gray and rostral ventromedial medulla disrupt
dala that affect neurotransmitter or neurohormonal antinociceptive but not cardiovascular aversive conditional re-
systems modulate the strength of the memory. For sponses. Journal of Neuroscience 14, 7,0997,108.
example, immediate posttraining blockade of adren- Iwata, J., LeDoux, J. E., and Reis, D. J., (1986). Destruction of in-
ergic or glucocorticoid receptors in the amygdala im- trinsic neurons in the lateral hypothalamus disrupts the classi-
pairs memory retention of inhibitory avoidance, while cal conditioning of autonomic but not behavioral emotional
responses in the rat. Brain Research 368 (1), 161166.
facilitation of these systems in the amygdala enhances Killcross, S., Robbins, T. W., and Everitt, B. J. (1997). Different
acquisition and memory storage (McGaugh, 2000; types of fear-conditioned behavior mediated by separate nu-
McGaugh et al., 1993). clei within amygdala. Nature 388 (6,640), 377380.
Thus, a picture has begun to emerge of the role of Neuroscience 23, 155184.
of the amygdala and its different subnuclei in emo- (2002). Synaptic self. New York: Viking.
NEUROGENESIS 469
Maren, S. (1999). Long-term potentiation in the amygdala: A What Is Adult Neurogenesis?

mechanism for emotional learning and memory. Trends in
Neuroscience 22 (12), 561567. Neurogenesis is the development of neurons
(2000). Neurobiology of Pavlovian fear conditioning. Annu- from neural stem or progenitor cells (see Figure 1).
al Review of Neuroscience 24, 897931. This developmental process begins with the division
McGaugh, J. L. (2000). Memorya century of consolidation. Sci- of the stem or progenitor cell and ends with a mature
ence 287 (5,451), 248251. and functioning new neuron. Neurogenesis of the
McGaugh, J. L., Introini-Collison, I. B., Cahill, L. F., Castellano,
vast majority of neurons occurs during intrauterine
C., Dalmaz, C., Parent, M. B., and Williams, C. L. (1993).
Neuromodulatory systems and memory storage: Role of the development and ceases postnatally. In the hippo-
amygdala. Behavioural Brain Research 58, 8190. campus and the olfactory system, however, neuro-
Milner, B., Squire, L. R., and Kandel, E. R. (1998). Cognitive genesis continues throughout life. Compared to the
neuroscience and the study of memory. Neuron 20, 445468. billions of neurons in the brain, the number of new
Nader, K., Majidishad, P., Amorapanth, P., and LeDoux, J. E. neurons produced in the postnatal brain is very low
(2001). Damage to the lateral and central, but not other, and even decreases with age. In the hippocampus new
amygdaloid nuclei prevents the acquisition of auditory fear
neurons are generated only in the granule cell layer
conditioning. Learning and Memory 8, 156163.
Pitknen, A., Savander, V., and LeDoux, J. E. (1997). Organization of the dentate gyrus, which increases in size with age.
of intra-amygdaloid cicuitries in the rat: An emerging frame- Under special circumstances single new neurons
work for understanding functions of the amygdala. Trends in might be generated in brain regions outside the hip-
Neuroscience 10 (11), 517523. pocampus or the olfactory system, most notably in the
Quirk, G. J., Armony, J. L., Repa, J. C., Li, X.-F., and LeDoux, J. neocortex. It would be interesting to learn whether
E. (1997). Emotional memory: A search for sites of plasticity. cortical neurogenesis under normal conditions would
Cold Spring Harbor Symposia on Biology 61, 247257.
Repa, J. C., Muller, J., Apergis, J., Desrochers, T. M., Zhou, Y., and
could contribute to learning and memory, because
LeDoux, J. E. (2001). Two different lateral amygdala cell pop- memory is stored in the neocortex. Elizabeth Gould
ulations contribute to the initiation and storage of memory. from Princeton University has reported adult neuro-
Nature Neuroscience 4, 724731. genesis in the neocortex of primates, but Pasko Rakic
Robbins, T. W., Cador, M., Taylor, J. R., and Everitt, B. J. (1989). and David Kornack from Yale University published a
Limbic-striatal interactions in reward-related processes. study in which they found no evidence of adult corti-
Neuroscience Biobehavioral Reviews 13, 155162.
cal neurogenesis. Although the function of neuro-
Schafe, G. E., Nader, K., Blair, H. T., and LeDoux, J. E. (2001).
Memory consolidation of Pavlovian fear conditioning: A cel- genesis in the olfactory system is not known, it ap-
lular and molecular perspective. Trends in Neuroscience 24, pears that adult neurogenesis pertaining to learning
540546. processes occurs only in the hippocampal dentate
gyrus.
Joseph E. LeDoux
Revised by Glenn E. Schafe and Joseph E. LeDoux Adult hippocampal neurogenesis in rats was first
described by Josef Altman in 1965, but at that time
the new neurons were considered to be a curiosity or
an atavism. The methods then available did not allow
a quantitative approach and a detailed phenotypic
analysis. Consequently, adult hippocampal neuro-
NEUROGENESIS
genesis was rediscovered several times, most nota-
Contrary to a still widely held belief, the adult brain bly by Michael Kaplan in the late 1970s and in the
routinely generates new neurons. Strikingly, one of early 1990s by Elizabeth Gould, then with Bruce Mc-
the two brain regions to which this adult neurogenesis Ewen at Rockefeller University in New York City.
is apparently restricted is the hippocampus. The hip- Georg Kuhn at the Salk Institute in La Jolla, Califor-
pocampus plays a key role in learning and memory, nia, was the first to use confocal microscopy in con-
so the role that new neurons can contribute to its nection with immunofluorescent labeling techniques
function is an inviting area for research. Most current to investigate adult hippocampal neurogenesis. As a
theories on how the hippocampus processes informa- result, new cells could be unambiguously identified as
tion for storage consider the brain to be a network neurons, and quantitative analyses became more fea-
that is static at the level of neuronal numbers and sible. In 1998 Peter Eriksson from Gothenborg, Swe-
plastic only at the level of neurites and synapses. den, used this technique to demonstrate that adult
However, mounting evidence indicates not only that hippocampal neurogenesis occurs in humans.
the hippocampus produces new neurons but also that In a parallel development, Fernando Nottebohm
adult neurogenesis is tightly linked to hippocampal and his coworkers from Rockefeller University stud-
function. New neurons might contribute not only to ied adult neurogenesis in birds. They were the first to
hippocampal function, but they might also be indis- add a strong functional context to this research. They
pensable to it. found that song learning in adult canaries correlates
470 NEUROGENESIS
Figure 1
New neurons are strategically added to the narrowest stretch in the hippocampal three-synapse circuit. The hippocampus is
the gateway to memory and its main function is to consolidate memory. Information is passed from cortical regions, including
the entorhinal cortex, to the dentate gyrus (1). The axons of the granule cell neurons of the dentate gyrus project via the
mossy fiber tract to area CA3 (2). The pyramidal neurons of CA3 connect to area CA1, from where the association cortex is
reached. Long-term storage occurs in neocortical association areas. Information is processed in the hippocampus to allow
long-term storage. It is not known in any detail how the hippocampus processes information; but new neurons might allow
the hippocampus to modify the network at the point of greatest convergence and in the long run to adjust to altered
functional needs. According to this theory, new neurons are not involved in the long-term storage of information. The
bottom part of the figure illustrates the development of a proliferating neuronal stem or progenitor cell in the subgranular
zone into a fully differentiated new granule cell.
NEUROGENESIS 471
strongly with adult neurogenesis in the brain center Physical activity enhances adult neurogenesis. In
responsible for this behavior. They later showed that, addition to providing a hint of potential links be-
in some food-storing bird species, neurogenesis is en- tween physical and mental activity and health, this
hanced in those seasons in which the birds have to re- finding suggests that active participation, in contrast
member the location of their food caches. to passive exposure to sensory stimuli, might affect
adult neurogenesis. This finding dovetails with psy-
chological learning theories that also emphasize ac-
How Is Adult Hippocampal Neurogenesis tive exploration and pursuit. Together these findings
Regulated? support the hypothesis that, in response to functional
The development of a functional interpretation demand, new neurons can be recruited into hippo-
of adult hippocampal neurogenesis in mammals campal neuronal circuitry. The cellular machinery
seemed more complex. Goulds early work empha- needed to achieve this goal goes far beyond a nonspe-
sized a suppression of adult neurogenesis arising cific mitogenic effect on the progenitor cells in the
from chronic, severe stress and mediators of a stress dentate gyrus.
reaction such as NMDA receptor activation and glu-
cocorticoids; on this account adult hippocampal
neurogenesis was due to stress-related damage to the
Hippocampal Function and Adult
hippocampus. Did this explanation imply that under Neurogenesis
normal conditions adult hippocampal neurogenesis The question of how the new neurons contribute
contributes to hippocampal function? The late 1990s to hippocampal function and thereby to learning and
brought a wealth of studies showing the effects of nu- memory is linked to the more fundamental issue of
merous manipulations on adult hippocampal neuro- the overall function of the hippocampus. Although
genesis. While no conclusive theory of the regulation the hippocampus is possibly the best-investigated
of adult hippocampal neurogenesis emerged from structure in the brain, with a rich store of data avail-
this research, it became apparent that adult neuro- able from researchers in anatomy, molecular biology,
genesis reacts to a wide range of stimuli with great biochemistry, electrophysiology, and behavioral psy-
sensitivity but low specificity. chology, no grand unifying theory of hippocampal
function has yet emerged. The hippocampus is classi-
One solution to this puzzle lies in the definition
cally characterized as the gateway to memory
of neurogenesis. The generation of a new neuron from
through which all information must pass to be re-
a neuronal stem or progenitor cell entails a series of
membered. Careful analysis of famous patient H.M.
developmental steps. What is sometimes seems to be
and animal research have indicated that bilateral hip-
an effect on neurogenesis is often an impact on cell
pocampal damage results in the inability to acquire
proliferation in which there is no certainty of a detect-
new memory. Recall of older information, however,
able change in the number of new mature neurons at
is spared. There is nonetheless a brief period of retro-
a later time. Regulation of adult neurogenesis, howev-
grade amnesia prior to the damage, indicating that a
er, occurs on several levels of neuronal development,
certain amount of time is needed for storage of infor-
including cell proliferation. Progenitor-cell prolifera-
mation in cortical regions. During this time the hip-
tion might represent a rather nonspecific part in this
pocampus processes information for storage.
regulation and can be influenced by many different
factors. Not all kinds of information require hippocampal
treatment. While declarative memory is considered
Learning stimuli and experience, however, ap-
hippocampus-dependent, procedural information,
parently do not influence progenitor-cell prolifera-
such as brushing your teeth (which is much more diffi-
tion in the dentate gyrus but exert a survival-
cult to declare in words than to do), can be learned
promoting effect on the progeny of the dividing cells.
without hippocampal involvement. In the context of
Animals that live in a so-called enriched environ-
experimental research with animals, declarative
ment (e.g., a cage providing much more sensory,
memory is difficult to assess. Spatial navigation, how-
physical, and social stimulation than the regular
ever, is hippocampus-dependent and can be easily in-
housing in laboratories) have a net increase in neuro-
vestigated in animals. Certain types of conditioning
genesis. This increase did not show a mandatory cor-
tasks are hippocampus-dependent, too, and there-
relation with changes in cell proliferation, but was
fore tests of spatial memory and conditioning tasks
due to a survival-promoting effect on the progeny of
dominate animal experiments on hippocampal func-
the dividing cells. In old mice, no effect on the total
tion.
number of surviving cells was detectable, but a pheno-
typic shift among these cells occurred toward more In 2000 Tracey Shors, in collaboration with Eliza-
new neurons and at the expense of new glial cells. beth Gould, showed that the elimination of cell prolif-
472 NEUROGENESIS
eration in the dentate gyrus by means of a cytostatic in single cells but is rather distributed over the synap-
drug is followed by reduced performance on a hippo- tic weights in a network of neurons. Synaptic plastici-
campus-dependent conditioning task, whereas per- ty, which is essentially changing synaptic weights, is at
formance on the hippocampus-independent form of the heart of the cellular mechanisms underlying
the test was spared. While this result conforms well to memory. Moreover, synapses come and go. Altered
theories linking adult hippocampal neurogenesis to connections cannot explain the recruitment of new
hippocampal function, the study does not explain the neurons. But neurons are more than the sum of their
function of the new neurons. This problem is similar connections. Neurons process information.
to one faced in many knockout studies. Elimination
of a particular gene often provides strong evidence
that this gene is involved in a biological process, even A Theory of New Neurons in Learning and
that it is necessary for this process, but it rarely ex- Memory
plains the specific nature of this involvement. One If the hippocampus is the gateway to memory,
reason why the Shors study does not yet provide the then adding new neurons in the hippocampus might
full answer is that hippocampal function certainly amount to adding new gatekeepers to this portal (see
goes beyond mediating eyeblink conditioning. In this Figure 1). The essentially three-synapse circuit within
sense the study even raises a new, puzzling question: the hippocampal formation includes a bottleneck
Why would the hippocampus require new neurons for within the connection between the dentate gyrus and
a task as simple and redundant as eyeblink condition- area CA3. The axons of granule cell neurons in the
ing? dentate gyrus form the mossy fiber tract, and adult
Regarding spatial memory as an indicator of hip- neurogenesis adds new granule cells that rapidly seek
pocampal function, our studies have shown that an in- connection to CA3.
crease in hippocampal neurogenesis correlates with Henriette van Praag and Alejandro Schinder at
an improvement on the Morris water maze task, a the Salk Institute in La Jolla studied the electrophy-
widely used test of hippocampal function in relation siological properties of newly generated granule cells
to spatial memory. However, this correlation never and found that the new neurons are indeed function-
seemed to be linear. Levels of adult neurogenesis do al. Information has to pass the narrows of the mossy
not seem to be indicative of memory as storage capac- fiber tract, and there appears to be a good reason for
ity. Rather, adult hippocampal neurogenesis seems this forced concentration. A small network is fast and
linked to the learning process itself; differences be- flexible. But the small network here has to cope with
tween animals living in an enriched environment a huge range of input information. It has to adjust to
emerged during the acquisition phase of the test. This novelty and complexity. If optimization here means
makes sense if one considers the hippocampus as a having the smallest possible network that can handle
gateway to memory, a processing unit that consoli- the situation, and if single new neurons can be strate-
dates information for storage but not as the brain gically introduced to the network, it becomes clearer
structure that provides long-term memory. But how how even a low rate of adult neurogenesis can be ben-
can new neurons contribute function to this process- eficial for the functional system. This view is also com-
ing unit? patible with the fact that neurogenesis decreases in
older age. If adult neurogenesis were to add units of
Joe Z. Tsiens group from Princeton University
memory to the hippocampus, old animals, who learn
found that presenilin-1 knockout mice had reduced
quite well, would have to rely on new neurons as much
experience-induced neurogenesis without any signs
as younger animals. If the modification of the net-
of impaired memory formation. The group reasoned
work were cumulative, however, older animals would
that new neurons might be necessary to forget old
require fewer new neurons because, given their great-
information, in the sense that the hippocampus
er experience, optimization of their network is al-
would periodically need to clear the structure from
ready further advanced and requires fewer strategic
outdated hippocampal memory traces after cortical
additions.
memory consolidation. This is a new twist on the
idea that new neurons increase (or modify) the pro- The new gatekeeper theory might explain how
cessors working memory (RAM in a computer analo- the hippocampus could benefit from adult neuro-
gy). New neurons would allow the hippocampus to genesis. It does not explain what the new neurons ac-
clear the structure for the processing of new informa- tually do. So far no conceptual links have been estab-
tion. Our own theory, explained below, is similar in lished between adult neurogenesis and existing
emphasizing the modification of the processor net- theories about the many specialized hippocampal
work. In our view, however, cells need not be replaced neurons, most notably place cells, with their response
to enable forgetting because information is not stored to a specific, recognized spatial pattern.
NEUROIMAGING 473
There is strong evidence that new neurons play maging methods, positron emission tomography
a role in hippocampal function and thus in learning (PET) and functional magnetic resonance imaging
and memory. Their contribution most likely lies in a (fMRI), measure signals related to blood flow or oxy-
refinement of the processing unit, not in an extension genation. Both methods are based on the correlation
of the storage capacity. But the precise manner in of blood flow and neuronal activity: as is the case with
which this refinement takes place (and how a growing muscles, blood flows more briskly in those parts of the
hippocampal network actually works) remains a ques- brain that are the most active neuronally. PET can
tion for current and future research. capture a single time-lapse image of forty seconds of
See also: GUIDE TO THE ANATOMY OF THE BRAIN: blood flow. Methodological necessity dictated that
HIPPOCAMPUS AND PARAHIPPOCAMPAL REGION PET studies of memory be done with blocked designs,
which involve a comparison of the blood flowand
Bibliography hence neuronal activityresulting from several trials
Altman, J., and Das, G. D. (1965). Autoradiographic and histologic of a task performed in two separate blocks. fMRI can
evidence of postnatal neurogenesis in rats. Journal of Compara-
tive Neurology 124, 319335.
take images much more rapidly, usually producing a
Feng, R., Rampon, C., Tang, Y. P., Shrom, D., Jin, J., Kyin, M., picture of the relative oxygenation level in blood ves-
Sopher, B., Martin, G. M., Kim, S. H., Langdon, R. B., Si- sels every two to three seconds. Although early fMRI
sodia, S. S., and Tsien, J. Z. (2001). Deficient neurogenesis in studies still use blocked designs similar to those used
forebrain-specific presenilin-1 knockout mice is associated
with reduced clearance of hippocampal memory traces. Neu-
in PET, the finer temporal capability of fMRI has also
ron 32, 911926. allowed event-related studies, in which trials of differ-
Gould, E., Tanapat, P., Hastings, N. B. and Shors, T. J. (1999). ent types can be intermixed, much as in a standard
Neurogenesis in adulthood: A possible role in learning. cognitive psychological experiment. Researchers in-
Trends in Cognitive Science 3, 186192.
Greenough, W. T., Cohen, N. J., and Juraska, J. M. (1999). New
terested in the cognitive neuroscience of memory
neurons in old brains: Learning to survive? Nature Neuro- have developed ingenious applications of event-
science 2, 203205. related designs for this technology. Although many
Kempermann, G., and Gage, F. H. (1999). New nerve cells for the imaging studies have confirmed the importance of
adult brain. Scientific American 280, 4853.
Kempermann, G., Kuhn, H. G., and Gage, F. H. (1997). More hip-
medial temporal cortex to memory, the studies have
pocampal neurons in adult mice living in an enriched envi- also highlighted important contributions of other
ronment. Nature 386, 493495. cortical areas, particularly the frontal cortex.
Rakic, P. (2002). Neurogenesis in adult primate neocortex: An
evaluation of the evidence. Nature Review Neuroscience 3, 65
71.
Shors, T. J., Miesegaes, G., Beylin, A., Zhao, M., Rydel, T., and Encoding
Gould, E. (2001). Neurogenesis in the adult is involved in the
A pair of papers, appearing in the same issue of
formation of trace memories. Nature 410, 372376.
van Praag, H., Kempermann, G., and Gage, F. H. (2000). Neural the journal Science, highlighted the utility of event-
consequences of environmental enrichment. Nature Review related designs. In these studies participants were
Neuroscience 1, 191198. presented with either words (Wagner et al., 1998) or
van Praag, H., Schinder, A., Christie, B., Toni, N., Palmer, T. D.,
indoor/outdoor scenes (Brewer et al., 1998) and then
and Gage, F. H. (2002). Functional neurogenesis in the adult
hippocampus. Nature 415, 1,0301,034. were asked to make judgments about the items while
undergoing imaging testing. These tasks were de-
Gerd Kempermann
signed to be good incidental encoding tasks. Follow-
Fred H. Gage
ing the imaging sessions, the researchers adminis-
tered recognition memory tasks for the studied items,
noting which items the subjects remembered and
which they forgot. They then sorted the images relat-
NEUROIMAGING ed to the encoding of the remembered or forgotten
Much of the cognitive neuroscience of memory has items. In both frontal and medial temporal regions,
taken advantage of lesion-behavior studies, which as- in the left hemisphere for the words and the right
sess behavioral deficits that ensue from damage to hemisphere for the scenes, activity was greater at en-
targeted parts of the brain. For example, lesions to coding for the remembered than for the forgotten
the medial temporal lobe, buried deep in the brain, items. In other words, the level of activity at the time
often produce profound anterograde amnesia. of encoding in these frontal and medial temporal re-
In the mid-1980s functional neuroimaging began gions predicted, on average, whether an item would
to play an increasingly important role in enhancing be remembered or forgotten. Several further studies
our understanding of the organization of memory in have replicated this subsequent memory effect in a
the human brain. Two of the main functional neuroi- number of encoding conditions.
474 NEURON
Retrieval how well visual experience will be remembered. Science 281,

1,1851,187.
Studies of memory retrieval have made several Donaldson, D. I., Petersen, S. E., Ollinger, J. M., and Buckner, R.
further associations between memory and cortical L. (2001). Dissociating state and item components of recogni-
areas outside the medial temporal lobe. Starting with tion memory using fMRI. Neuroimage 13, 129142.
early PET studies of memory and continuing through Karni, A., Meyer, G., Jezzard, P., Adams, M. M., Turner, R., and
Ungerleider, L. G. (1995). Functional MRI evidence for adult
many fMRI studies, right frontal activation was com- motor cortex plasticity during motor skill learning. Nature
mon when people retrieved information from episod- 377, 155158.
ic memory (Squire, 1992; Squire et al., 1992; Tulving Nyberg, L., Habib, R., McIntosh, A. R., and Tulving, E. (2000). Re-
et al., 1994). Considerable debate has focused on activation of encoding-related brain activity during memory
whether this right-frontal activation relates to the ef- retrieval. Proceedings of the National Academy of Sciences of the
United States of America 97, 11,12011,124.
fortful state accompanying retrieval, to the successful Schacter, D. L., and Buckner, R. L. (1998). Priming and the brain.
retrieval of items or to postretrieval processes. Re- Neuron 20, 185195.
searchers are striving to develop new, more complex Squire, L. R. (1992). Memory and the hippocampus: A synthesis
designs that aim to separate sustained, modelike pro- from findings with rats, monkeys, and humans. Psychological
cesses from the transient processing of items (Donald- Review 99, 195231.
Squire, L. R., Ojemann, J. G., Miezin, F. M., Petersen, S. E., Vid-
son, Petersen, Ollinger, and Buckner, 2001). een, T. O., and Raichle, M. E. (1992). Activation of the hippo-
Other, left-lateralized cortical regions in parietal campus in normal humans: A functional anatomical study of
and frontal cortex have been more directly associated memory. Proceedings of the National Academy of Sciences of the
with retrieval success. Event-related studies, mainly of United States of America 89, 1,8371,841.
Tulving, E., Kapur, S., Markowitsch, H. J., Craik, F. I. M., and
recognition memory tests, show greater activation in Habib, R. (1994). Neuroanatomical correlates of retrieval in
these regions for studied items than for previously episodic memory: Auditory sentence recognition. Psychology
unstudied, new words. Researchers have found these 91, 2,0122,015.
old-new effects in several kinds of experiments with Wagner, A. D., Schacter, D. L., Rotte, M., Koustaal, W., Maril, A.,
different types of material, but always in similar re- Dale, A. M., Rosen, B. R., and Buckner, R. L. (1998). Building
memories: Remembering and forgetting of verbal experi-
gions (see Figure 1). ences as predicted by brain activity. Science 281, 1,1881,191.
Finally, several studies (e.g., Nyberg, Habib,
Steven E. Petersen
McIntosh, and Tulving, 2000) have suggested that
recollection of, say, visual memories entails the reacti-
vation of visual perceptual-processing mechanisms.
In these studies, as people recall previously studied vi-
NEURON
sual information, extra-primary visual regions are ac-
tivated, even in the absence of a visual cue. It seems See: GUIDE TO THE ANATOMY OF THE BRAIN
that some top-down process is reactivating perceptual
mechanisms to retrieve specific perceptual content.
Other Studies
NEUROTRANSMITTER SYSTEMS
AND MEMORY
Other studies have focused on myriad other as-
pects of learning and memory, such as procedural Ever since the discovery of the chemical nature of syn-
learning (Karni et al., 1995), priming (Schacter and aptic transmission, the role of neurotransmitters in
Buckner, 1998), and so on. Each of these lines of the formation and retrieval of memories has been the
study seems to elucidate further neural mechanisms subject of intense scientific investigation. As the num-
related to distinct memory processes. Such studies ber of both neurotransmitters and forms of memories
have yielded insights into cognitive architectures as has been steadily increasing over the years, the task
well. For example, the earlier-noted right-prefrontal of uncovering general principles describing the in-
activation in many episodic retrieval tasks suggests volvement of neurotransmitter systems in memory
the presence of a specific retrieval mode. Advancing has become extremely difficult. Furthermore, the lack
neuroimaging technology promises to shed still more of understanding of the molecular and cellular mech-
light on the neural mechanisms that underlie human anisms of learning and memory has limited the ex-
memory and cognition. perimental approaches to two general strategies: (1)
an interventional strategy using pharmacological
See also: EPISODIC MEMORY; PROCEDURAL tools or lesion/stimulation of specific neurotransmit-
LEARNING
ter systems; and (2) a correlational strategy using
Bibliography naturally occurring conditions (neurological dis-
Brewer, J. B., Zhao, Z., Desmond, J. E., Glover, G. H., and Gabrieli, eases, aging) affecting specific neurotransmitter sys-
J. D. (1998). Making memories: Brain activity that predicts tems, or genetically engineered mutant mice. Based
NEUROTRANSMITTER SYSTEMS AND MEMORY 475
Figure 1
The left panel shows the location of a parietal region in a horizontal slice from the brain. The right panel shows the difference in
activation level between hits (correct identification of items that had been previously studied in a recognition test) and correct
rejections (items correctly identified as being unstudied). This region shows a retrieval success effect in that the activity is much larger
for hits and correct rejections.
on these studies, a number of neurotransmitters and brief bursts of high-frequency electrical stimulation of
neuronal pathways using these neurotransmitters monosynaptic pathways using glutamate as a neuro-
have consistently demonstrated an important role in transmitter (Malenka and Nicoll, 1999). The relation-
learning and memory (Chapoutier, 1989; Decker and ships among LTP, learning and memory, and gluta-
McGaugh, 1991). matergic systems have been further established by the
use of specific antagonists of different subtypes of glu-
tamate receptors (Morris et al., 1990), which clearly
Glutamate established that LTP induction is due to NMDA re-
In early work using an avian retina preparation, ceptor activation while LTP maintenance involves
proline was found to inhibit glutamate release; as pro- modification of AMPA receptors. Manipulations af-
line had also been shown to inhibit learning and fecting either type of receptors were found to have
memory in chicks, researchers suggested that gluta- significant effect on learning and memory processes.
mate played a significant role in learning and memo- For instance, positive modulators of AMPA receptors,
ry. Building on these results, Cherkin and his col- the ampakines, have been shown to be cognitive en-
leagues (1976) published a series of articles hancers (Ingvar et al., 1997), indicating that up-
demonstrating that various glutamate antagonists in- regulation of AMPA receptor-mediated synaptic re-
jected intracerebrally could retard memory consoli- sponses facilitate LTP formation and AMPA LTP-
dation in neonatal chicks, using a behavioral para- mediated learning and memory processes. Likewise,
digm involving flavor aversion learning. In the 1990s, overexpression of one subunit of NMDA receptors in
the role of glutamate in learning and memory re- mice, the NR2B subunit, resulted in increased LTP
ceived considerable support from two independent and learning and memory (Tsien, 2000). Second, var-
lines of research. First, the most studied cellular ious human conditions associated with major distur-
model of learning and memory is the long term po- bances of memory exhibit marked degeneration of
tentiation (LTP) of synaptic transmission elicited by glutamatergic neurons. Thus, severe cases of amnesia
476 NEUROTRANSMITTER SYSTEMS AND MEMORY
are the result of loss of glutamatergic neurons in the tylcholine-induced synchronization of neuronal firing
hippocampal formation (Squire, 1986), and one of in the theta range (5 to 7 hertz) produces optimal
the hallmarks of Alzheimers disease is a loss of gluta- conditions for summation of excitatory depolariza-
matergic neurons in entorhinal cortex and hippo- tion, thereby facilitating activation of NMDA recep-
campus (Hyman et al., 1987). All these studies point tors, LTP induction, and ultimately memory forma-
out to the critical role of glutamatergic pathways of tion.
the hippocampus in the formation of long-term
memories.
Catecholamines, Serotonin, and Histamine
Although noradrenergic systems have often been
GABA (-Aminobutyric Acid) implicated in learning and memory (Gold and Zor-
Historically, GABAergic neurons have been ne- netser, 1983), studies using pharmacological block-
glected as possible participants in memory processes. ade of norepinephrine receptors or destruction of
The discovery of the mode of action of benzodiaze- noradrenergic neurons indicate that these neurons
pines and the known effects of these compounds on do not directly participate in the learning of a variety
memory renewed the interest in the possible role of of tasks (Pontecorvo et al., 1988). However, the situa-
GABA in memory. Thus benzodiazepines potentiate tion is probably more complicated, for noradrener-
the effect of GABA at the GABAA receptors and gen- gic, serotonergic, and histaminergic systems appear
erally produce an impairment of learning, whereas li- to participate in the modulation of some memory
gands acting at the benzodiazepine sites but produc- processes (DHooge and De Deyn, 2001). Depending
ing an opposite effect (and therefore called reverse on which subtype of receptors is activated under par-
agonists) enhance learning in mice, chickens, and hu- ticular conditions, facilitation or inhibition or memo-
mans (Lister, 1985). Local administration of GABA ry formation has been observed. In particular, it is
agonists or antagonists in discrete brain regions have now clear that spatial learning, which is often used to
been effective in producing impairment or enhance- study mechanisms of learning and memory, depends
ment, respectively, in various learning tasks (Mc- upon the coordinated action of several brain regions
Gaugh, 1989). In this case as well, the links between and neurotransmitter systems. In addition, research-
the GABAergic systems and LTP have been proposed ers have repeatedly argued that these modulatory sys-
to account for the role of GABA in learning and mem- tems are activated under numerous behavioral condi-
ory, as inhibition of GABA receptors generally facili- tions (arousal and stress), which in turn affect
tate LTP induction, while potentiation of GABA re- information processing and storage in sensory and as-
ceptors prevent LTP induction. sociational pathways.
Acetylcholine Other Systems

Acetylcholine has long been involved in learning Although not generally considered as neurotran-
and memory (Deutsch, 1983). Pharmacological smitters, a growing number of peptides have been
studies using both acetylcholinesterase inhibitors found to be colocalized with traditional neurotran-
(producing increased levels of acetylcholine at syn- smitters and have been assigned the role of neuro-
apses) and blockers of acetylcholine receptors have modulators. Their role in synaptic function is still a
been performed in numerous learning tasks and ani- matter of debate, but it is impossible not to mention
mal species. The general conclusion is that impair- the evidence implicating a number of neuropeptides
ment of cholinergic transmission produces cognitive in memory processes. In particular, ACTH and frag-
impairment. This conclusion is also reinforced by ments of ACTH have repeatedly been shown to im-
studies in humans with pathological alterations in prove memory consolidation. Vasopressin has also
cholinergic function. Thus, Alzheimers disease is as- been suggested as participating in memory forma-
sociated with loss of cholinergic neurons in the nucle- tion, possibly through an interaction with nora-
us basalis of Meynert (Coyle et al., 1983). In Parkin- drenergic systems (DHooge and De Deyn, 2001).
sons disease, memory impairment is correlated with Similarly, opioid peptides have been clearly implicat-
a decrease in cholinergic function in the frontal cor- ed in memory processes, including the suggestion
tex (Chapoutier, 1989). A possible unifying mecha- (Gallagher, 1985) that opioid peptide-containing
nism for a role of cholinergic neurons in memory pro- neurons are part of a forgetting mechanism. Much at-
cesses has been ascribed to the participation of these tention has been devoted to the role of the steroid
neurons in the generation of rhythmical brain activity hormone estrogen in memory formation, as estrogen
( and rhythms) directly involved in information replacement therapy was found to alleviate cognitive
storage and in LTP (Winson, 1990). In particular, ac- deficits in postmenopausal women. The recent dis-
NEUROTRANSMITTER SYSTEMS AND MEMORY 477
covery that estrogen potentiates AMPA and NMDA Bibliography

receptor function and facilitates LTP induction pro- Chapoutier, G. (1989). The search for a biochemistry of memory.
vides a bridge between the cellular effects of estrogen Archives of Gerontology and Geriatrics Supp. 1, 719.
and its behavioral effects. This finding also raises in- Cherkin, A., Eckardt, M. J., and Gerbrandt, L. D. (1976). Memory:
Proline induces retrograde amnesia in chicks. Science 193,
teresting questions related to the evolution of memo-
242244.
ry systems and of their relationships to reproductive Coyle, J. T., Price, D. L., and DeLong, M. R. (1983). Alzheimers
and sexual behaviors. disease: A disorder of cortical cholinergic innervation. Science
219, 1,1841,190.
Decker, M. W., and McGaugh, J. L. (1991). The role of interactions
Conclusion between the cholinergic systems and other neuromodulatory
As it becomes more and more widely accepted systems in learning and memory. Synapse 7, 151168.
that learning and memory reflect the existence of a Deutsch, J. A. (1983). The cholinergic synapse and the site of mem-
ory. In J. A. Deutsch, ed., The physiological basis of memory. New
variety of synaptic plasticity mechanisms occurring at York: Academic Press.
numerous stages of information processing and in DHooge, R., and De Deyn, P. P. (2001). Applications of the Morris
different neuronal networks participating in informa- water maze in the study of learning and memory. Brain Re-
tion storage, it is not surprising that several neuro- search Reviews 36, 6090.
transmitter systems participate, directly or indirectly, Gold, P. E., and Zornetser, S. F. (1983). The mnemon and its
in the processes of information storage and retrieval. juices. Behavioral and Neural Biology 38, 151189.
Hyman, B. T., Van, H. G. W., and Damasio, A. R. (1987). Alz-
The initial notion that one neurotransmitter was in-
heimers disease: Glutamate depletion in the hippocampal
volved in one form of behavior has long been dis- perforant pathway zone. Annals of Neurology 22 (1), 3740.
missed in favor of the idea that distributed and paral- Ingvar, M., et al. (1997). Enhancement by an ampakine of memory
lel neuronal networks participate in multiple sensory, encoding in humans. Experimental Neurology 146, 553559.
motor, and cognitive operations. While the roles of Lister, R. G. (1985). The amnesic action of benzodiazepines in
glutamate and glutamatergic pathways as well as the man. Neuroscience and Biobehavioral Review 9, 8794.
Malenka, R. C., and Nicoll, R. A. (1999). Long-term potentia-
functional significance of modifications of synaptic ef-
tiona decade of progress? Science 285, 1,8701,874.
ficacy at glutamatergic synapses in this process are McGaugh, J. L. (1989). Involvement of hormonal and neuro-
starting to be well established, leading to the develop- modulatory systems in the regulation of memory storage. An-
ment of the first memory pill, the task of correlat- nual Review of Neuroscience 12, 255287.
ing the roles of other neurotransmitters and other Morris, R. G. M., Davis, S., and Butcher, S. P. (1990). Hippocampal
neuronal pathways to specific forms of learning and synaptic plasticity and NMDA receptors: A role in informa-
tion storage. Philosophical Transactions of the Royal Society of
memory still awaits the development of more specific
London B329, 187204.
pharmacological tools and behavioral tests. In partic- Pontecorvo, M. J., Clissold, D. B., and Conti, L. H. (1988). Age-
ular, it is of crucial importance to define learning related cognitive impairments as assessed with an automated
characteristics that are both species and time depen- repeated measures memory task: Implications for the possible
dent as the demands on the learning and memory role of acetylcholine and norepinephrine in memory dysfunc-
machinery are exquisitely sensitive to species require- tion. Neurobiology of Aging 9, 617625.
Squire, L. R. (1986). Mechanisms of memory. Science 232, 1,612
ments and time windows adapted to the particular
1,619.
species. It is clear, however, that this strategy will re- Tsien, J. Z. (2000). Building a brainier mouse. Scientific American
sult in the development of more refined pharmaco- 282, 6268.
logical treatments with potentials for alleviating more Winson, J. (1990). The meaning of dreams. Scientific American 263,
specific learning and memory disorders. 8696.
See also: COGNITIVE ENHANCERS; GLUTAMATE Michel Baudry

RECEPTORS AND THEIR CHARACTERIZATION Joel L. Davis
O
OBJECT CONCEPT, their own faces in order to see, as in games of peek-a-
DEVELOPMENT OF boo. From eight to twelve months (Stage 4), infants
search for a completely hidden object and generalize
The object concept is the knowledge that objects are their search to different objects and different covers
permanent, independent entities that exist in space or barriers. They do not, however, generalize to dif-
and time even when one cannot perceive or act on ferent locations. Infants who find a hidden toy in one
them. Humans would be almost unable to function location continue to search for it there even after see-
without this knowledge. Although children clearly ac- ing the toy hidden in a new location. Between twelve
quire the object concept early in development, re- and eighteen months (Stage 5), infants incorporate
searchers disagree about exactly when and how they location information into their object knowledge.
acquire it. They track the hidden object to its most recent hiding
location, provided they see the toy hidden there (visi-
ble displacement). However, from eighteen to twenty-
Piagets Theory four months (Stage 6), infants find the hidden object
The Swiss researcher Jean Piaget proposed the in a new location even without seeing it hidden there
earliest comprehensive account of object concept de- (invisible displacement). According to Piaget, this be-
velopment in the 1930s. Piaget believed that children havior demonstrates that infants fully acquire the ob-
gradually construct the concept over the first two ject concept between eighteen and twenty-four
years of life in a predictable and universal series of six months.
stages. From birth to three or four months (Stages 1
and 2), infants do not truly perceive objects; they
merely recognize stimulation associated with their Contradictory Evidence
own subjective experience, such as the reaction of Although the behaviors in Piagets manual search
pleasure connected with the sight of a caregiver or an tasks are highly replicable, many researchers disagree
attractive toy. By two months, infants turn to look at with his interpretations. When tasks are simplified,
an object that makes a sound, demonstrating an inte- infants appear sensitive to hidden objects much earli-
gration of vision and hearing that gives objects greater than Piaget proposed. Most research focuses on
er solidity. Between four and eight months (Stage 3), Stages 3 and 4, from four to twelve months. For exam-
infants noticeably progress toward acquiring the ob- ple, in the violation-of-expectation method, infants
ject concept. For example, infants visually or manual- simply watch events involving hidden objects while
ly follow the path of an object that they drop and re- their looking times are measured. In one series of
turn to an object after dropping it out of sight. They studies described by Renee Baillargeon in 1993, in-
also retrieve partially hidden objects and uncover fants between four and seven months watched a barri-
479
480 OBJECT CONCEPT, DEVELOPMENT OF
er hide an object by rotating in front of it like a draw- searchers such as Richard Bogartz and others, inter-
bridge. In one event, the barrier bumped into the pretations that transcend perceptual explanations by
hidden object and then rotated back to its starting po- appealing to mental representation may be too elabo-
sition (a possible event). In another event, the barrier rate. Although this work has clearly demonstrated the
rotated through the space occupied by the hidden ob- importance of perceptual factors, such challenges
ject (an impossible event) before returning to its start- cannot account for all violation-of-expectation re-
ing position. Infants looked significantly longer at the sults. Thus, cognitive factors also appear to affect in-
impossible event, suggesting they remembered the fants looking times, suggesting that infants may in-
hidden object, expected the barrier to stop when it deed have object knowledge early in development.
reached the object, and were surprised by the viola-
tion of this expectation. In many studies, described by To explain the difference between looking and
Baillargeon in 1993 and 1998 and by others, infants reaching performance, other accounts propose that
between four and eight months looked longer at im- the two tasks tap either different paths of knowledge
possible than possible events across a range of objects, or representations of different strengths. For exam-
barriers, and displays. Such evidence suggests that in- ple, according to Gavin Bremner, violation-of-
fants in Piagets Stage 3 not only mentally represent expectation studies tap perceptual capacities that be-
the existence of hidden objects but also other proper- come gradually incorporated with development into
ties such as the location, height, and number of hid- infants capacities for action in search tasks. Likewise,
den objects, and even whether objects behavior fol- Munakata and colleagues in 1997 proposed that in-
lows principles of gravity and inertia. These results fants object representations gradually become
suggest that infants understand the object concept stronger with development. Infants may succeed in
earlier than Piaget proposed. looking tasks with a weak representation of the hid-
den object but fail on reaching tasks because active
search requires a stronger object representation.
Why Do Looking and Reaching Differ? Although researchers have fervently debated in-
Why do infants between four and eight months terpretation of violation-of-expectation studies, an-
perform better on looking than reaching tasks? Re- other line of research supports the position that in-
searchers have proposed quite different explanations, fants have early sensitivity to hidden objects. In this
resulting in considerable controversy. According to approach, researchers present infants with a manual
one account endorsed in 1993 by Baillargeon and by search task but hide the object in the dark instead of
others, manual search tasks like those Piaget used with a cover or barrier. The object becomes hidden
may be inadequate for measuring infants object when the lights go off, but infants can retrieve it with
knowledge; such tasks require infants to move a cover a simple direct reach instead of a means-end action.
or barrier in order to retrieve an object. Infants may Using the reaching-in-the-dark method, researchers
know the hidden object exists but are unable to dem- such as Rachel Clifton and colleagues found that in-
onstrate this knowledge by retrieving the object be- fants between six and seven months reached for ob-
cause of a secondary deficit in means-end ability. That jects in the dark. In combination with violation-of-
is, infants may have difficulty planning a sequence of expectation results, these findings support the view
reaching for the cover and moving it out of the way that infants have object knowledge before they dem-
(the means) in order to retrieve the object (the end). onstrate it in means-end tasks.
In contrast, infants can demonstrate their knowledge
in violation-of-expectation studies because they do
not require means-end skill. Other work has chal- Object Location
lenged this account. For instance, Yuko Munakata In addition to research on infants object knowl-
and colleagues in 1997 demonstrated that when edge between four and eight months (Stage 3), in-
means-end demands were equated for retrieving visi- fants behavior with hidden objects from eight to
ble and hidden objects, infants succeeded more with twelve months (Stage 4) has also generated many
visible objects. Thus, a means-end deficit cannot by it- studies. As described earlier, infants younger than
self account for infants search problems. Whether the twelve months have difficulty tracking a hidden object
object is hidden or visible has a significant effect on that changes location. For instance, infants who find
infants success. a hidden toy in location A continue to search for it
According to other researchers, infants longer there, even after seeing the toy hidden in location B.
looking at impossible than possible events may not re- Although this A-not-B error is easy to replicate in
flect genuine object knowledge. Instead, infants look- reaching versions of the task, researchers such as Aye-
ing times may reflect simpler perceptual preferences sha Ahmed and others reported that infants erred less
for a familiar event or a novel event. According to re- often in looking versions of the task. Like violation-of-
OBJECT CONCEPT, DEVELOPMENT OF 481
expectation results with four- to eight-month-olds, fants encode object motion and location information
this finding suggests that infants know more than they earlier in development than they encode object prop-
can demonstrate in manual search tasks. erties or features, only later incorporating both kinds
Why then do infants make the A-not-B error in of information in their object representations. Like-
their manual search? Researchers disagree. Piaget wise, the 1997 account of Munakata and colleagues
originally proposed that infants cannot separate the proposes that knowledge is graded, with object repre-
objects existence from its location. In contrast, Esther sentations becoming stronger with development.
Thelen and colleagues suggested the error results not
from inadequate object knowledge but from the nor- Conclusion
mal development of the dynamic processes of look-
ing, remembering, planning for action, and reaching. Piaget originally proposed that infants gradually
This dynamic systems account proposes that these develop the object concept from birth to age two.
processes are continuously meshed or embodied such However, evidence collected from simplified tasks
that knowledge cannot be separated from perception, that measure looking suggests that infants have some
memory, and reaching. Memory and reaching expe- object knowledge several months earlier than Piaget
rience also play crucial roles in a different explana- believed. Despite challenges to interpretations of
tion of the A-not-B error that Adele Diamond reiter- looking studies and divergent explanations for why
ated in 1998. According to Diamond, the error results infants perform better in looking than reaching tasks,
from a combination of limitations in both memory most researchers agree that looking studies demon-
and inhibition. Infants memory for the toy in loca- strate early sensitivity to hidden objects. Results from
tion B fades quickly, and they have trouble inhibiting reaching-in-the-dark studies support this conclusion.
the previously reinforced reach to location A. This ac- Findings from looking versions of the A-not-B task
count shares some features with an alternative ac- likewise suggest that infants know more about objects
count of the A-not-B error that Munakata described than they demonstrate in their manual search,
in 1998. According to this account, competition be- though researchers disagree about why infants err in
tween latent memory traces for location A and active the reaching version of the task. The goal for future
memory traces for location B explains the error. In- research is to thoroughly test the predictions of these
fants repeated experience in finding the toy in loca- various accounts, with the hope of establishing a com-
tion A results in a latent bias toward A. Infants also prehensive framework for understanding infants ac-
form active memory traces for location B, maintain- quisition of the object concept.
ing an accessible representation of the hidden toy at
B. However, the strong latent memory traces for A Bibliography
override the weaker active memory for B, resulting in Ahmed, A., and Ruffman, T. (1998). Why do infants make A not
B errors in a search task, yet show memory for the location of
the error. hidden objects in a nonsearch task? Developmental Psychology
34, 441453.
Baillargeon, R. (1993). The object concept revisited: New direc-
Alternatives to Piagets Theory tions in the investigation of infants physical knowledge. In C.
Granrud, ed., Visual perception and cognition in infancy: Carne-
Beyond these accounts, which address specific
gie-Mellon symposia on cognition. Hillsdale, NJ: Erlbaum.
phenomena of object concept development (e.g., the (1998). Infants understanding of the physical world. In M.
A-not-B error or the difference between looking and Sabourin, F. Craik, and M. Robert eds., Advances in psychologi-
reaching performance), researchers have proposed cal science, Vol. 2: Biological and cognitive aspects. East Sussex,
more comprehensive theories of the development of UK: Psychology Press.
Bogartz, R. S., Shinskey, J. L., and Schilling, T. (2000). Object per-
the object concept as a whole. Some accounts, like
manence in five-and-a-half month old infants? Infancy 1, 403
that posed by Elizabeth Spelke, propose that some 428.
degree of core object knowledge is present from Bremner, J. G. (1998). From perception to action: The early devel-
birth. Other theories, like that described by Baillar- opment of knowledge. In F. Simion and G. Butterworth, eds.,
geon in 1998, suggest that from birth there is a spe- The development of sensory, motor and cognitive capacities in early
infancy: From perception to cognition. East Sussex, UK: Psycholo-
cialized learning mechanism that guides infants ac-
gy Press.
quisition of object knowledge. According to this Clifton, R. K., Rochat, P., Litovsky, R. Y., and Perris, E. E. (1991).
account, infants initially form a broad category for Object representation guides infants reaching in the dark.
events involving hidden objects and gradually ex- Journal of Experimental Psychology 17, 323329.
pand their understanding of such events by identify- Diamond, A. (1998). Understanding the A-not-B error: Working
memory vs. reinforced response, or active versus latent trace.
ing the important factors that affect it. In similar ac-
Developmental Science 1, 185189.
counts, Fei Xu and others suggest that infants Munakata, Y. (1998). Infant perseveration and implications for ob-
construct object representations from knowledge ject permanence theories: A PDP model of the AB task. Devel-
pathways that are initially separate. In particular, in- opmental Science 1, 161184.
482 OBSERVATIONAL LEARNING
Munakata, Y., McClelland, J. L., Johnson, M. H., and Siegler, R. social and symbolic environment, some pertaining to
(1997). Rethinking infant knowledge: Toward an adaptive cognitive skills, preconceptions, and value prefer-
process account of successes and failures in object perma-
nence tasks. Psychological Review 104, 686713.
ences of the observers, and others to the conspicuous-
Piaget, J. (1954). The construction of reality in the child. New York: ness, attractiveness, and functional value of the mod-
Basic. eled activities themselves. Still other factors concern
Spelke, E. (1995). Initial knowledge: Six suggestions. In J. Mehler the structural arrangements of human interactions
and S. Franck, eds., Cognition on cognition. Cognition special se- and social networks, which largely determine the
ries. Cambridge, MA: MIT Press.
Thelen, E., Schoner, G., Scheier, C., and Smith, L. B. (2001). The
types of models to which people have ready access.
dynamics of embodiment: A field theory of infant persevera- People cannot be much influenced by observed
tive reaching. Behavioral and Brain Sciences 24, 186.
events if they do not remember them. A second sub-
Xu, F. (1999). Object individuation and object identity in infancy:
The role of spatiotemporal information, object property in- function governing observational learning concerns
formation, and language. Acta Psychologica 102, 113136. centers on cognitive representational processes. Re-
tention involves an active process of transforming and
Jeanne L. Shinskey
restructuring information about modeled events into
rules and conceptions for generating new patterns of
behavior. Preconceptions and emotional states can
bias how observed information is transformed into
OBSERVATIONAL LEARNING memory codes. Similarly, recall involves reconstruc-
tion of past experiences rather than simple retrieval
Psychological theories have traditionally emphasized of registered past events.
learning from direct experience. If knowledge and
skills could be acquired only by trial and error, In the third subfunction in observational learn-
human development would be greatly retarded, not ingthe production processsymbolic conceptions
to mention exceedingly tedious and hazardous. are transformed into appropriate courses of action.
Moreover, limited time, resources, and mobility im- Conceptions are rarely translated into proficient ac-
pose severe limits on the places and activities that tion from the outset. Skills are usually perfected
people can directly explore to gain new knowledge through a conception-matching process. Conceptions
and competencies. Nevertheless, humans have guide the construction and execution of behavior pat-
evolved an advanced cognitive capacity for observa- terns. The behavior is modified as necessary to
tional learning that enables them to abbreviate achieve close correspondence between conception
knowledge acquisition by learning from the examples and action. The richer the repertoire of subskills that
provided by others. Indeed, nearly all behavioral, people possess, the easier it is to integrate them to
cognitive and emotional learning that results from di- produce the new forms of behavior.
rect experience can be duplicated by observation of The fourth subfunction in observational learning
others behavior. concerns motivational processes. People do not per-
A special power of social modeling is that it can form everything they learn. Performance of observa-
transmit new ways of thinking and behaving to count- tionally learned behavior is influenced by three major
less people in widely dispersed locales through sym- types of incentive motivators: direct, vicarious, and
bolic modes of communication. By drawing on this self-produced. People are more likely to perform ob-
symbolic modeling, observers can transcend the servationally learned behavior if it is rewarding rather
bounds of their immediate environment. With the ad- than punishing. The observed costs and benefits ex-
vent of enormous advances in the technology of com- perienced by others influence the performance of
munication, observational learning from the symbolic modeled patterns as much as directly experienced
environment is playing an increasingly powerful role consequences do. People are motivated by the suc-
in peoples everyday lives. cesses of others who are similar to them but are dis-
couraged their failures. Personal standards of con-
duct provide a further source of incentive motivation.
Subfunctions Governing Observational People pursue activities that they find satisfying and
Learning that confer a sense of self-worth; they reject those of
Observational learning is governed by four sub- which they personally disapprove.
functions that are summarized in Figure 1. Attention-
al processes determine what people selectively ob-
serve in the profusion of modeling influences and Abstract Observational Learning
what information they extract from modeled events. Observational learning is not merely a process of
A number of factors influence what people choose to behavioral mimicry. Highly functional patterns of be-
explore and how they perceive what is modeled in the havior, which constitute the proven skills and estab-
OBSERVATIONAL LEARNING 483
Figure 1
The four major subfunctions governing observational learning and the influential factors operating within each subfunction.
lished customs of a culture, may be adopted in essen- behavior through observational learning. In addition
tially the same form as they are exemplified. There to creating new competencies, modeling influences
is little leeway for improvisation on how to drive auto- have strong motivational effects. Seeing others gain
mobiles or solve arithmetic problems. However, in desired outcomes by their actions can foster the in-
many activities, subskills require improvisation in the centives of positive expectations; observed adversities
mastery of varied challenges. Modeling influences can serve as the disincentives of positive expectations.
can convey rules for generative and innovative behav- The motivational impact of modeled transactions de-
ior as well. For example, an individual may see others pends on observers outcome expectations, judg-
confront moral conflicts involving different matters ments of their ability to perform the modeled behav-
but apply the same moral standard to them. In ab- ior, their perceptions of the likely success or failure
stract observational learning, observers extract the of the modeled actions, and their expectation of simi-
rules or standards embodied in the specific judg- lar results if they undertook comparable activities.
ments or actions exhibited by others. Once they learn
a rule, they can use it to generate new instances of be- People are easily aroused by the emotional ex-
havior that go beyond what they have seen or heard. pressions of others. Therefore, observers can acquire
Thus, on the basis of modeled information, people lasting attitudes, values, and emotional dispositions
acquire standards for categorizing and judging toward persons, places, or things that they associate
events, linguistic rules of communication, thinking with modeled emotional experiences. They learn to
skills on how to gain and use knowledge, and personal fear the things that frightened others, to dislike what
standards for regulating motivation and conduct. Evi- repulsed them, and to like what gratified them. Fears
dence that generative rules of thought and behavior and intractable phobias can be weakened or eliminat-
can be created through abstract modeling attests to ed by modeling coping strategies for exercising con-
the broad scope of observational learning. trol over the things that are feared. Values can simi-
larly be developed and altered vicariously by repeated
exposure to the likes and dislikes of others.
Multiple Effects of Observational Learning During their daily lives, people have direct con-
So far the discussion has centered on the acquisi- tact with only a small sector of the physical and social
tion of knowledge, cognitive skills, and new styles of environment. They inhabit a limited locale, work in
484 OLDS, JAMES
the same setting, travel the same routes, visit the same Rogers, E. M., and Kincaid, D. L. (1981). Communication networks:
places, perform the same routines day in and day out, Toward a new paradigm for research. New York: Free Press.
Rosenthal, T. L., and Zimmerman, B. J. (1978). Social learning and
and interact with the same circle of friends and asso- cognition. New York: Academic Press.
ciates. Consequently, their conceptions of social reali-
ty with which they have little or no contact are greatly Albert Bandura
influenced by vicarious experiencesby what they see
and hearin the absence of direct experiential cor-
rectives. The more peoples conceptions of reality de-
pend upon the medias symbolic environment, the OLDS, JAMES (19221973)
greater is its social impact. Many of the shared mis- James Olds became a prominent figure in physiologi-
conceptions about occupational pursuits, ethnic cal psychology when he discovered, in 1953, that rats
groups, minorities, the elderly, social and gender could be trained to perform a variety of experimental
roles, and other aspects of life are at least partly en- tasks, some at very high rates, in order to obtain a
gendered through symbolic modeling of stereotypes. pleasurable electrical stimulus applied to a discrete
central nervous system site. This procedure, often
Social Diffusion Through Symbolic called self-stimulation or intracranial self-stimulation,
Modeling is intimately related to the brain pathways that medi-
ate positive reinforcement. At the time of his sudden
Much of the preceding discussion has concerned death, Olds was Bing Professor of Behavioral Biology
observational learning at the individual level. Video at the California Institute of Technology in Pasadena
systems feeding off telecommunications satellites and a member of the National Academy of Sciences.
have become the dominant vehicle for disseminating The events leading up to the serendipitous discovery
symbolic environments to vast populations. Social of pleasure centers in the brain were described both
practices are being widely diffused within societies, by Olds (1973b, 1975a, 1977) and others (Miller,
and ideas, values, and styles of conduct are being 1980; Milner, 1989).
modeled worldwide. Observational learning is thus
coming to play an increasingly influential role in so-
ciopolitical and transcultural change. In this broader Early Life
function of social diffusion, modeling through the Olds was born in Northbrook, Illinois, where his
mass media instructs people in new ideas and social father, Leland, was industrial editor for the Federated
practices. Positive and negative incentives determine Press. The family moved in 1931 to Nyack, New York,
which of the modeled innovations will be adopted. where Oldss father headed the New York State Power
People are linked together by networks of social rela- Commission. In 1939 the family moved to Washing-
tionships. Social networks provide diffusion paths for ton, DC, where the elder Olds served as commission-
the spread of new ideas and behavior. er and then chairman of the Federal Power Commis-
Observational learning takes many forms and sion until 1949. Olds began his college career as an
produces diverse outcomes. The modeling influences undergraduate at the University of Wisconsin in Mad-
on which it draws serve as instructors, motivators, in- ison. He transferred to St. Johns College in Annapo-
hibitors, disinhibitors, social facilitators, emotion ar- lis, Maryland, and then worked briefly for the Inter-
ousers and modifiers, constructors of social realities, national News Service before being drafted into the
and transnational acculturators. army during World War II. Trained in Arabic, he
spent most of his war service in Cairo, Egypt, with the
See also: LANGUAGE LEARNING: HUMANS Persian Gulf Command. After returning to the Unit-
ed States at the end of the war, Olds transferred to
Bibliography
Amherst College, of which his grandfather had been
Bandura, A. (1986). Social foundations of thought and action: A social
cognitive theory. Englewood Cliffs, NJ: Prentice-Hall. president, to finish his undergraduate studies.
(1997).Self-efficacy: The exercise of control. New York: Free- Olds received his doctoral training in the depart-
man. ment of social relations at Harvard University. A rela-
(2002). Environmental sustainability through sociocogni-
tive approaches to deceleration of population growth. In P. tively new department at that time, it included social,
Schmuch and W. Schultz, eds., The psychology of sustainable de- experimental, and clinical psychology and sociology.
velopment. Dordrecht, the Netherlands: Kluwer. Oldss mentor was Richard Solomon, who gave him
Braithwaite, J. (1994). A sociology of modelling and the politics of thorough training in experimental psychology and
empowerment. British Journal of Sociology 45, 445479. exposed him to the current literature on physiologi-
Gerbner, G., Gross, L., Morgan, M., and Signorielli, N. (1994).
Growing up with television: The cultivation perspective. In J.
cal psychology. Olds had a part-time job as editor of
Bryant and D. Zillman, eds., Media effects: Advances in theory a book by Talcott Parsons, chairman of the depart-
and research. Hillsdale, NJ: Erlbaum. ment, to supplement his graduate fellowship. Oldss
OLDS, JAMES 485
contributions were so substantial that Parsons made

him a coauthor. This was an unusual intellectual ap-
prenticeship for someone who later made major con-
tributions to the understanding the neural substrates
of reward learning. His contact with Olds during sub-
sequent years provided the biological perspective for
Parsonss sociological theorizing.
Oldss thesis dealt with motivation and was influ-
enced by Donald Hebbs landmark book, The Organi-
zation of Behavior (1949). He planned to train under
Hebb as a first step toward developing a neural real-
ization of a model of Edward Tolmans sign-gestalt
theory that ideas determine behavior (Olds, 1954).
The basic conviction underlying Oldss career plans
at this stage was that behavior had to be explained in
terms of underlying brain activity. Olds felt that the
two principal problems of physiological psychology
were motivation and learning, that these two prob-
lems were closely intertwined, and that their solution
depended on a detailed knowledge of the central ner-
vous system. These basic ideas motivated Oldss en-
tire professional career.
Professional Beginnings
After completing his doctoral thesis, Olds spent
a year as a lecturer at Harvard and then moved on to
McGills psychology department to spend two years as
a postdoctoral trainee under Hebb. During this peri-
od, Hebbs laboratory interacted extensively with the
Wilder Penfield and Herbert Jasper groups at the
Montreal Neurological Institute. When Olds arrived
at McGill to begin his training, he was placed under is interesting to note Milners description of the ener-
the guidance of Peter Milner, who was then finishing gy and organizational ability that Olds showed in
his doctoral thesis. Milner taught him how to implant fleshing out their original observation (Milner, 1989).
stimulating electrodes in the brain, and Olds pre- Olds went on to perform many systematic studies
pared a rat with stimulating electrodes in what he as- using a standard and sensitive technique to deter-
sumed was the reticular formation, an area of consid- mine which portions of the brain supported self-
erable interest at that time. The hypothesis to be stimulation. He also developed new technical ap-
tested was that the animal would avoid stimulation in proaches and took advantage of new developments in
the reticular formation (assumed to be the source of the brain/behavioral sciences. This approach also
neural activity to be reduced by behavior). But when characterized the contributions he made to the study
the rat received stimulation through the implanted of learning and memory.
electrode, quite the opposite effect occurred. The rat
actually was attracted to those locations that were as-
sociated with the stimulation. This observation in the Mapping the Brain for Learning
very first rat he studied turned Oldss initial hypothe- Oldss laboratory at Caltech in the early 1970s
sis on its head and also suggested that the brain con-
was the scene of a unique series of studies designed
tains reward or pleasure centers that mammals seek
to map the brain for learning (Disterhoft and Buch-
to activate during goal-seeking behaviors.
wald, 1980; Olds, 1973a). These studies used a tech-
Some unsuccessful attempts to replicate the origi- nique he had perfected for recording small groups of
nal electrode placement led to the determination that neurons in the freely moving rat (Olds et al., 1972;
the hypothalamus and septal regions supported self- Olds, 1975b). The basic objective was to determine
stimulation. Olds and Milner described their land- how neurons in various brain regions in animals
mark discovery in a paper that appeared in 1954. It learning the same associative response compared
486 OLDS, JAMES
over time. The regions that changed earliest, by defi- used it in their analysis of hippocampal-system activi-
nition, were especially notable in the formation and ty during learning of the nictitating membrane re-
readout of the learned response. sponse in the rabbit. Woody (1974) had begun studies
using temporal analysis of the motor side of the eye-
Studies of how the auditory system might change
blink pathways in overtrained cats at about the time
its processing of the tone-conditioned stimulus dem-
the Olds groups mapping studies began. More recent
onstrated that neurons from the inferior colliculus up
approaches have combined in vitro analyses of hippo-
to the auditory cortex showed alterations in firing
campal brain slices with in vivo recording to ensure
rate during differential auditory conditioning. The
precise study of localized cellular and subcellular al-
research focused intensively on the posterior nucleus
terations in the hippocampus during eyeblink condi-
of the thalamus, a region that receives multisensory
tioning (Disterhoft and McEchron, 2000).
afferent drive and showed large firing-rate changes
very early in the learning process. The involvement Considerable progress has been made in the
of the hippocampus, a region that Olds sometimes re- analysis of the mechanisms of fear conditioning over
ferred to as the Rosetta Stone of the brain, also re- the past two decades. The most successful studies of
ceived considerable attention. Oldss previous map- this phenomenon have used and extended the ap-
ping of the brain for self-stimulation had proach pioneered by Olds and his colleagues in their
demonstrated that widely spread brain regions sup- mapping studies (Davis and Whalen, 2001; Schafe et
ported this phenomenon. Studies of neurons in many al., 2001). These studies of emotional learning have
other regions of the brain, including the nonspecific resonated among nonneuroscientists interested in
thalamus, the basal ganglia, the reticular formation, brain mechanisms of learning and behavior (Dama-
and the hypothalamus showed altered firing rates sio, 1994; LeDoux, 1996), imparting renewed impe-
during an important behavioral event such as audito- tus to Oldss position that learning and motivation are
ry learning. pivotal problems in physiological psychology (now
subsumed under the more encompassing term neuro-
The series of studies done by Oldss group yield-
science).
ed some additional general contributions to the study
of learning and memory. First, they delineated the
advantage of using an apparently simple task as a Setting the Stage for Future Research
model system to study mechanisms of learning in
Olds spent considerable time and effort setting
the mammalian brain. Second, they clearly demon-
the stage for subsequent laboratory work; his col-
strated that many neurons in many brain regions do
leagues were intimately involved in the daily sequence
change during acquisition of even a simple learning
of events that led to the important observations that
task. These data firmly supported the idea that learn-
ensued. Oldss approach to postdoctoral training was
ing is truly a distributed process in the brain. Third,
modeled after that of Donald Hebblearning by
they demonstrated the advantages of formulating
doing. His experience in Hebbs laboratory, where he
temporal maps of alterations in the brain during
discovered self-stimulation as a postdoctoral fellow,
learning as a way to determine where important alter-
clearly impressed upon him the value of serendipity,
ations were occurring (Olds et al., 1972). The tempo-
personal effort, and scientific tinkering. The series of
ral maps were drawn in the two dimensions: from
brain-mapping studies used that general approach
conditioned-stimulus onset through conditioned-
and remain landmarks in the study of associative
response performance and from the first training trial
learning in mammalian brain.
through acquisition and overlearning of the condi-
tioned response. This approach allowed a clear rank- Olds was passionately dedicated to studying the
ing of the relative importance of the many alterations functioning of the brain. His one frustration was that
in single-neuron activity observed during learning in he felt he had not devoted enough time to studying
widely scattered brain regions. Finally, the studies the brain. He felt that the more facts he had at his dis-
clearly emphasized the importance and strength of an posal, the more likely it was that he could assemble
approach that dissected individual subsystems for de- them into compelling insights. He spent a good deal
tailed analysis during learning of a common task. of time thinking, talking, and writing about how the
Subsystem analysis allows a more precise delineation brain works (Olds, 1975a, 1977, 1980).
of the total system interactions.
Olds was also fascinated with computers and elec-
The approach Olds designed and that his group tronics. Many of his ideas about brain function, such
used to such obvious advantage was adapted by many as his speculations about memory-storage function in
laboratories and is still being profitably used in the the hippocampus, used computers and their memo-
continuing search for mechanisms of learning and ries as analogies (Olds, 1969). He expended much ef-
memory. For example, Berger and Thompson (1978) fort on developing and testing software and hardware
OLDS, JAMES 487
for the abundant computer equipment in his labora- sions that ranged far from the data at hand and dis-
tory. His studies of associative learning used a com- cussions of appropriate strategies to use in current or
bined hardware-software system simultaneously to future experiments. Descriptions of Oldss scientific
study a large number of brain regions in animals activities by Neal Miller (1980) and Peter Milner
learning the same relatively simple associative task. (1989) confirm Olds lifelong eagerness to share ideas
with colleagues and students and to approach brain
The technical problem of developing better ways
function with novel perspectives.
to study single neurons in conscious animals was one
to which Olds devoted onsiderable energy. One of the Olds was a gracious, urbane man with a good
reasons he came to Caltech was to take advantage of sense of humor. He was also a family man. His wife,
the possibilities for developing electronic gadgetry Marianne, worked closely with him in the lab. He also
for his experiments. He collaborated with an electri- lavished considerable attention on his son, James,
cal engineer from the Jet Propulsion Laboratory in who was in high school during the time Olds was on
the design and building of what must have been one the Caltech faculty.
of the earliest telemetry systems for multiple single-
neuron recording. The idea was to transmit signals See also: GUIDE TO THE ANATOMY OF THE BRAIN;
from ten microwire electrodes simultaneously without LOCALIZATION OF MEMORY TRACES
incurring the danger of cable artifacts. The rat looked
a little ungainly with the miniature transmitter on its Bibliography
head (considerably less miniature in 1971 than it Berger, T. W., and Thompson, R. F. (1978). Neuronal plasticity in
would be today), but the system worked pretty well. the limbic system during classical conditioning of the rabbit
Olds was always trying to come up with a better opera- nictitating membrane response. I. The hippocampus. Brain
tional amplifier than the ones he was using, although Research 145, 323346.
Davis, M., and Whalen, P. J. (2001). The amygdala: Vigilance and
the old standbys often worked better. He also was in-
emotion. Molecular Psychiatry 6, 1334.
volved in troubleshooting things like electronic wave- Damasio, A. R. (1996). Descartes error: Emotion, reason and the human
form identifiers; he wanted his to work more simply brain. New York: G. P. Putnam.
and efficiently. He would hook up a rat, sit down in Disterhoft, J. F., and Buchwald, J. S. (1980). Mapping learning in
front of an oscilloscope, and run a new waveform the brain. In A. Routtenberg, ed., Biology of reinforcement. New
York: Academic Press.
identifier through its paces by himself. He knew from
Disterhoft, J. F., and McEchron, M. D. (2000). Cellular alterations
experience that often the product did not work as in hippocampus during acquisition and consolidation of hip-
precisely as the engineerintended. pocampusdependent trace eyeblink conditioning. In D.
Woodruff-Pak and J. E. Steinmetz, eds., Eyeblink classical condi-
The portion of the laboratory used for the map- tioning, Vol. 2: Animal models. Norwell, MA: Kluwer.
ping of learning was set up with four training stations. Hebb, D. O. (1949). The organization of behavior. New York: John
Olds used a recording station of his own and carried Wiley.
on a series of experiments separate from those of the LeDoux, J. (1998). The emotional brain: The mysterious underpinnings
of emotional life. New York: Simon and Schuster.
postdoctoral fellows and graduate students. He trav-
Miller, N. E. (1980). Introduction: Brain stimulation reward and
eled extensively, but when he was in town, he came in theories of reinforcement. In A. Routtenberg, ed., Biology of
every morning to check the rat-in-training in his sta- reinforcement. New York: Academic Press.
tion and to check the setting of the waveform discrim- Milner, P. M. (1989). The discovery of self-stimulation and other
inators on the unit channels. Olds was very demand- stories. Neuroscience and Biobehavioral Reviews 13, 6167.
Olds, J. (1954). A neural model for sign-gestalt theory. Psychological
ing about the quality of the data he and his group
Review 61, 5972.
gathered. He was a firm believer that high-quality (1969). The central nervous system and the reinforcement
findings came from high-quality data. His system had of behavior. American Psychologist 24, 114132.
numerous checks for electronic noise and other arti- (1973a). Brain mechanisms of reinforcement learning. In
facts. He also took an intense interest in the experi- D. E. Berlyne and N. B. Madsen, eds., Pleasure, reward, prefer-
ence. New York: Academic Press.
ments as they were being run. Those with freely mov-
(1973b). Commentary on Olds and Milners Positive rein-
ing rats ran from evening until early morning, the forcement produced by electrical stimulation of septal area
peak of the rats diurnal cycle. A collaborator coming and other regions of rat brain. In E. S. Valenstein, ed., Brain
in during the evening to check experimental progress stimulation and motivation: Research and commentary. Glenview,
was likely to discover that Olds had been there shortly IL: Scott, Foresman.
(1975a). Mapping the mind onto the brain. In F. G. Wor-
before. Olds was almost always in the laboratory early
den, J. P. Swazey, and G. Adelman, eds., The neurosciences:
on Sunday to make notes on the printout or adjust- Paths of discovery. Cambridge, MA: MIT Press.
ments to the computer. (1975b). Unit recordings during Pavlovian conditioning. In
N. A. Buchwald and M. A. B. Brazier, eds., Brain mechanisms
Olds and his collaborators met in his office every in mental retardation. New York: Academic Press.
afternoon to discuss the data and their interpretation. (1977). Drives and reinforcements: Behavioral studies of hypotha-
These meetings often included theoretical discus- lamic functions. New York: Raven Press.
488 OLFACTORY CORTEX
(1980). Thoughts on cerebral functions: The cortex as an wrote his doctoral thesis on the behavioral effects of
action system. In A. Routtenberg, ed., Biology of Reinforcement. penicillin injected into the hippocampus. He joined
New York: Academic Press.
Olds, J., Disterhoft, J. F., Segal, M., Kornblith, C. L., and Hirsh,
the faculty of Johns Hopkins University at the age of
R. (1972). Learning centers of rat brain mapped by measur- twenty-six in 1969, the same year he was awarded his
ing latencies of conditioned unit responses. Journal of Neuro- Ph.D. Within ten years he was a full professor. He
physiology 35, 202219. served as department chairman from 1982 to 1987
Olds, J., and Milner, P. M. (1954). Positive reinforcement pro- and remained a member of the Johns Hopkins faculty
duced by electrical stimulation of septal area and other re-
gions of rat brain. Journal of Comparative and Physiological Psy-
until he was struck down by pancreatic cancer at the
chology 47, 419427. age of fifty-one.
Schafe, G. E., Nader, K., Blair, H. T., and LeDoux, J. E. (2001).
Memory consolidation of Pavlovian fear conditioning: A cel-
lular and molecular perspective. Trends in Neuroscience 24, Scientific Context
540546.
Woody, C. D. (1974). Aspects of the electrophysiology of cortical
Oltons central scientific interest was the brain
processes related to the development and performance of mechanisms of memory. His career spanned the end
learned motor responses. The Physiologist 17, 4969. of the behaviorist era, the emergence of modern
John F. Disterhoft
neuroscience and cognitive science, and the wide-
spread acceptance of animal cognition as a legitimate
discipline. Some of the key influences on Oltons
thinking were the ideas and findings of Scoville and
Milner (1957) on case H.M., Tolman (1948) on cogni-
OLFACTORY CORTEX tive maps, Mishkin and Delacour (1975) on primate
See: GUIDE TO THE ANATOMY OF THE BRAIN; models of amnesia, Tulving (1972) on episodic mem-
NEURAL COMPUTATION ory, Platt (1964) on strong inference, and Garner
(1956) on converging operations. A thorough empiri-
cist, Olton made his most important scientific contri-
butions through his rigorous analytic studies of the
OLTON, DAVID (19431994) contribution of the hippocampal system to memory
and cognition in rats. He developed novel techniques
David Olton was a psychologist who studied the and pioneered powerful combinations of converging
neuroscience of memory and animal cognition. He methods. He was a master at translating abstract theo-
discovered that rats not only learned and remem- retical concepts into powerful experiments, and he
bered places, but like humans, could keep lists of was particularly ingenious when translating human
places in memory for hours. His work on memory in neuropsychological findings into animal models.
animals, aimed toward modeling human memory and
amnesia, led to the discovery that remembering
places required the same brain structures in rats and The Radial Maze and Working Memory
humans: the hippocampal system. These basic discov- Theory
eries launched a research program into how the hip- Oltons most prominent methodological contri-
pocampal system normally supported memory, how bution was the radial maze, and his best-known con-
it was impaired by physical damage, aging, or by dis- ceptual contribution was the idea that the hippocam-
eases such as Alzheimers dementia, and how im- pus supported working memory. The method and the
paired memory might be restored. theory are linked historically and together constitute
David Olton was born in New Jersey and reared Oltons most influential work.
in Richmond, Virginia. He attended Haverford Col- Olton and Samuelson introduced the radial maze
lege, where he enjoyed the questions posed in philos- in 1976. The paper described the maze, an elevated
ophy classes, but he was more impressed with the central platform with eight arms extending symmetri-
surer methods of the natural sciences. After taking a cally like spokes from a wheel. In the original task,
course in psychophysics, he decided that the combi- the end of each arm had a food cup that was filled at
nation of psychological questions and empirical the start of a daily trial, and rats learned to forage op-
methods was the right fit. He earned a B.A. in psy- timallyto enter each arm once to obtain food with-
chology in 1964 and was awarded the Pennsylvania out reentering depleted arms. Probe experiments es-
Psychological Association prize for the best under- tablished that rats remembered the spatial location of
graduate research project. the arms they had visited to get food and excluded al-
Oltons interest in behavioral neuroscience grew ternative explanations such as scent marking, intra-
with his dissertation work at the University of Michi- maze cues, and response chaining. This demonstra-
gan, where he studied with Robert L. Isaacson. He tion of spatial memory in rats was a new and
OLTON, DAVID 489
important contribution to the emerging field of ani-

mal cognition. The result implied that rats represent-
ed a list of spatial locationskept several places in
mindand, through a flexible memory process that
was sensitive to interference, remembered which of
those locations they visited on a given day. This find-
ing was of such fundamental importance that the ra-
dial maze became a symbol of animal cognition.
After establishing these new facts on animal cog-
nition, Olton pursued the neural bases of radial maze
performance. Olton first demonstrated that perfor-
mance depended upon the major extrinsic anatomi-
cal connections of the hippocampus (Olton, Walker,
and Gage, 1978). Lesions of the entorhinal cortex,
the fimbria-fornix, the septum, or the postcommissu-
ral fornix produced chance performance on the radi-
al maze. Unilateral lesions of the fornix or entorhinal
cortex did not impair performance, whereas crossed
lesions that disconnected the hippocampus from cor-
tical and subcortical throughput did. Lesions restrict-
ed to hippocampal neurons produced similar impair-
ments to the disconnections (Handelmann and
Olton, 1981; Jarrard, 1986). Together, the results
were clear: lesions of the hippocampus, its extrinsic
connections, or its intrinsic circuitry impaired perfor-
mance in the radial maze. The task proved to be one
of the most sensitive and selective measures of hippo-
campal function ever devised and is still in use. The (Olton, 1978; Olton and Papas, 1978). Each arm in
basic cognitive requirements of the taskspatial dis- the baited set contained food at the start of a trial; the
crimination, recent memory, flexible memory expres- arms in the unbaited set never had food. For the bait-
sion, and resistance to interferenceinfluenced ed set, the contingencies were the same as in the stan-
Oltons thinking about hippocampal function for his dard radial maze task: during a given trial, the rat had
entire career. to choose each arm once and to avoid reentering that
Oltons initial analysis of the radial maze empha- arm. In contrast, the rat always had to avoid entering
sized both spatial representation and flexible memo- the arms in the unbaited set to perform efficiently.
ry processing. His later work addressed how the hip- The same spatial discrimination ability was required
pocampus contributes to each of these two important to distinguish arms in both sets, but flexible memory
task demands. He initially emphasized the impor- expression was required to remember which of the
tance of spatial representation and referred to con- baited arms had been entered in a given trial. Normal
verging evidence from lesion, stimulation, and re- rats entered each baited arm once during each trial
cording studies. Thus, he showed that hippocampal and avoided unbaited arms altogether. In contrast,
seizures impaired radial maze performance (Olton rats trained in the task and then given lesions of the
and Wolf, 1982), and that neurons had place fields in hippocampal system repeatedly reentered baited
the radial maze that, in principle, could represent the arms within a trial but avoided entering unbaited
different arms as well as a rats entrance into and exit arms. Thus, hippocampal lesions impaired flexible
from those arms (Olton, Branch, and Best, 1978). By memory expression, but not spatial discrimination.
the late 1970s, however, Olton became convinced that Around 1977, Olton adapted Honigs (1978)
the hippocampus was crucial for recent memory for terms working memory to emphasize task components
a wide range of stimuli aside from spatial ones. The that required flexible memory for items within a trial
evidence that swayed his view, together with the clari-
and reference memory to describe task components that
ty of Oltons arguments, had a powerful and enduring
were unchanged across trials. Olton and Papas (1978)
influence on the neuroscience of memory and hippo-
assembled these findings and concepts to claim that
campal function.
the hippocampus was required for working but not
Olton assigned the arms of a seventeen-arm radi- reference memory, even when the items to be remem-
al maze into either a baited or an unbaited set bered were spatial locations. After he investigated a
490 OLTON, DAVID
nonspatial working memory task, Olton became more The theory predicted that hippocampal lesions would
firmly convinced that memory processes better de- impair working-memory tasks and spare reference-
fined the unique contribution of the hippocampal sys- memory tasks. The clarity and simplicity of the ideas
tem than spatial representation. Rats were trained to were compelling, and they inspired an international
remember which of four visually and tactually distinct research effort that led to important advances in the
arms they had visited in a trial. The location of each neuroscience of memory, as described in detail below.
of the four arms was changed after each choice, pre- Both of the predictions succeeded often, but not al-
venting rats from using a spatial representation of ways. For example, some spatial-reference memory
their locations. This nonspatial working memory task tasks did require the hippocampus, most notably in
was severely impaired by lesions of the fimbria-fornix the water maze (Morris et al., 1982). The cumulative
(Olton and Feustle, 1981). data refuted the strongest predictions of working-
The pattern of intact and impaired performance memory theory, and Olton conceded that this version
in the radial maze tasks culminated in Oltons work- of the theory was insufficient.
ing memory theory of hippocampal function, which Although the distinction between working and
informed his most-cited work, Hippocampus, Space, reference memory did not account for the full range
and Memory (Olton, Becker, and Handelmann, 1979). of effects of hippocampal system lesions, working
The theory claimed that the hippocampus is required memory procedures often required the septo-
for behaviors that demand working memory, inde- hippocampal system and differentially activated hip-
pendent of whether the material to be remembered pocampal neurons. In collaboration with Warren
was spatial. Working and reference memory were de- Meck, Gary Wenk, and Russ Church, Olton showed
fined operationally, and testing procedures were ex- that hippocampal lesions impaired working memory
plicitly distinguished from memory processes. for the duration of recently presented stimuli (Meck
Oltons working memory theory also addressed one
et al., 1984 ; Olton et al., 1988). In a double dissocia-
of the two major deficits in amnesia associated with
tion, prefrontal cortical circuits were shown to be cru-
hippocampal damage: the inability of amnesic pa-
cial for attending to the duration of two simultaneous-
tients to remember recent events even (and perhaps
ly presented stimuli (Olton, Wenk, Church, and
especially) when those events are comprised of famil-
Meck, 1988). Single neurons in the hippocampus
iar items.
were more commonly and more strongly activated
during performance of a nonspatial working memory
Working Memory: Cognitive Basis and task than during either a spatial or a cued reference-
Operational Definitions memory task that required the same perceptual dis-
Olton intended working-memory tasks to empha- criminations and behaviors (Wible et al., 1986). Mem-
size an event-based, trial-unique memory process ory, rather than other task variables, strongly influ-
(Olton et al., 1979). From the outset, however, the enced hippocampal activity. The tasks required
choice of the term working memory posed a problem discrimination between cue boxes, and lesions of the
for students. Cognitive neuroscientists had replaced hippocampal system impaired only the working-
short-term memory with working memory to describe a memory procedure (Raffaele and Olton, 1988). Sei-
memory buffer that served as a representational zure stimulation of the hippocampus completely reset
workspace for manipulating items kept in mind (e.g., working memory (Olton and Wolf, 1981; Knowlton et
Baddeley, 1974). Although he did occasionally deal., 1985) but had no effect on the learning of a spatial
scribe it as a memory buffer, Oltons working memory reference-memory discrimination (Knowlton et al.,
was not defined by duration, computational work- 1989). Spatial working-memory performance in a T-
space, or consciousness. Rather, Olton used the term maze provided an especially sensitive, quantitative
to operationalize a memory process described in assay of septo-hippocampal function. Microinjections
ethology that was most similar to Tulvings (1972) de- of GABA agonists or acetylcholine antagonists dis-
scription of episodic memory (Olton et al., 1979). rupted working memory performance, reduced ace-
From Oltons view, working memory entailed memo- tylcholine release in the hippocampus, and sup-
ry for events, items that occurred in a specific tempo- pressed hippocampal theta in tightly correlated
ral and individual context. The distinct and varying patterns (Givens and Olton, 1990).
significance of individual items, encoded as events As the limitations of the original working-
within a temporal context, provided representations memory theory became clear, Olton sought to correct
that guide responses more flexibly than their unvary- the principles, logic, and analytic approaches that
ing stimulus content. misdirected the theory. Early in the 1990s he decided
Oltons working-memory theory made strong that his thinking had been too strongly influenced by
predictions through clear operational definitions. categorical interpretations of dissociation experi-
OLTON, DAVID 491
ments (the classic Olton model was a 2x2 table de- perienced stimuli or behaviors. The potential com-
scribing a double dissociation), and he began to con- plexity of the neural systems underlying operationally
sider quantitative, parametric experimental designs. defined working memory tasks was known from cog-
His reading of the data had already convinced him nitive neuropsychology, where different working
that the spatial theory of hippocampal function was memory systems were dissociatedcompare H.M.s
incomplete, that working-memory demand was a cru- short-term memory for verbal and nonverbal items
cial variable in the extent to which tasks were im- (Sidman, Stoddard, and Mohr, 1968), and from re-
paired by hippocampal lesions, and that the categori- search on nonhuman primates (Fuster, 1995). Refer-
cal, operational definition of working memory ence-memory tasks, defined operationally as trial-
limited its usefulness. He also began to reexamine the independent memory, can in principle depend upon
third cognitive requirement for performing the radial memory for repeated episodes as well as those for the
maze task to guide his thinking: resistance to interfer- rules and procedures that were originally described to
ence (Olton and Shapiro, 1992; Shapiro and Olton, be the information-processing core of reference
1994). memory tasks (Olton et al., 1979). Subsequent re-
Oltons new approach was fruitful. In a nonspa- search has revealed some of the complex circuitry un-
tial, continuous, conditional-discrimination task, rats derlying working and reference memory.
were trained to press one bar if two consecutive stimu- Research in Norman Whites lab at McGill Uni-
li were the same and another bar if they were different versity used the radial maze to show that various op-
(Wan, Pang, and Olton, 1994; see also Wible et al., erationally defined reference-memory tasks require
1992). By changing the frequency of stimulus repeti- distinct neural components that support independent
tion and delay interval, the experiment varied both memory strategies (e.g., McDonald and White, 1993).
proactive interference and working-memory de- The hippocampus, the amygdala, and the caudate
mand. Even after months of training, high proactive nucleus each support a different memory strategy for
interference and long delays revealed significant non- discriminating arms in the radial maze (Packard,
spatial working-memory impairments in rats with le- Hirsh, and White, 1989; McDonald and White, 1983).
sions of the hippocampus or fornix (Wan et al., 1994). In one such task, the hippocampal system was shown
His new approach to memory research had begun to be necessary for discriminating between adjacent
well when he became gravely ill. but not separated arms of the maze. To learn this spa-
tial reference memory discrimination, intact rats had
to visit the two arms in succession during each daily
Continuing Influence: Theory and trial, suggesting that learning the spatial discrimina-
Application tion depended upon comparisons supported by re-
cent working memory (White and Oellet, 1997).
Working-Memory Theory
Oltons clear predictions and conceptual analyses Working-memory mechanisms are also variable
posed a challenge that persists today: Can any current in rats, humans, and nonhuman primates. In rats the
variety of brain substrates underlying working-
theory of hippocampal function accurately describe a
memory tasks was shown by Kesner et al. (1993), who
priori, in operational terms, the full range of tasks
reported a triple dissociation among working-
that will be impaired or spared by damage to the hip-
memory tasks. Lesions of the caudate, visual cortex,
pocampal system? The varied-response strategies
and hippocampus selectively impaired working mem-
provided by multiple memory systems suggest that
ory for responses, visual objects, and locations, re-
any particular task can be guided by more than one
spectively (Kesner et al., 1993). Oltons latest ideas on
system. Even an operationally defined spatial working
hippocampal function, working memory, and inter-
memory task in the radial maze can be solved, in prin-
ference also continue to influence memory research
ciple, by response chaining and thus dispense with
(e.g. Long and Kesner, 1998; Gilbert et al., 1998;
the hippocampus. So, operational distinctions aside,
Hampton et al., 1998; Fortin et al., 2002).
what remains of Oltons analysis of hippocampal
function? Spatial versus Nonspatial Memory
The conceptual core of working-memory theory Oltons observations that the hippocampus is not
emphasized important aspects of hippocampal func- required exclusively for spatial memory have been
tion. Hindsight makes clear that the original opera- verified repeatedly. Many nonspatial memory tasks
tional definition Olton proposed for working memo- that require flexible responses to events have been
ry did not distinguish between tasks that required shown to require the hippocampus in both rats and
memory for recent events (e.g., episodic memory) nonhuman primates. Lesions of the hippocampus im-
from those that could be solved by other mechanisms pair social-recognition memory (Kogan et al., 2000),
for maintaining short-term memory for recently ex- social transmission of food preference (Winocur,
492 OLTON, DAVID
1990; Bunsey and Eichenbaum, 1995), transitive in- The Man

ference (Bunsey and Eichenbaum, 1996; Dusek and David Oltons contribution to science included
Eichenbaum, 1997), DRL performance (Sinden et al., far more than the sum of his research and ideas. Al-
1986), negative patterning (Sutherland and Rudy, ways serious about science, he was informal in his
1989), trace eyelid conditioning (Solomon et al., manner and had an ironic sense of humor about him-
1986), and memory for olfactory sequences (Fortin, self. These qualities made him a sought collaborator,
Agster, and Eichenbaum, 2002). Not all spatial dis- a valued and respected colleague, and an outstanding
crimination tasks require the hippocampus. Thus, mentor. Among his graduate, undergraduate, and
rats with fornix or hippocampal lesions not only learn postdoctoral students were Fred H. Rusty Gage of
to discriminate arms in a radial maze, but they can be the Salk Institute, now president of the Society for
trained to find a hidden platform in the Morris water Neuroscience; John Morrison, director of the Center
maze using a reference-memory procedure, whereas for Neurobiology at Mount Sinai School of Medicine;
they are unable to learn flexible responses such as a and Gary Wenk, professor of psychology at the Uni-
spatial-reversal learning in the same situation (Eich- versity of Arizona.
enbaum, Stewart, and Morris, 1990; Whishaw et al.,
1995). Olton argued that by helping to encode the See also: SPATIAL MEMORY; WORKING MEMORY:
temporal context of events, the hippocampus contrib- ANIMALS
utes to behavioral flexibility. Thus, the same stimulus
can be approached in one instance and avoided in a Bibliography
second instance because the unique sequence and the Arendash, G. W., King, D. L., Gordon, M. N., Morgan, D., Hatch-
outcome of the behavioral interaction with that stimu- er, J. M., Hope, C. E., and Diamond, D. M. (2001). Progres-
lus are remembered. A widespread view is that the sive, age-related behavioral impairments in transgenic mice
hippocampus is indeed crucial for remembering carrying both mutant amyloid precursor protein and pre-
items in their temporal context and that this memory senilin-1 transgenes. Brain Research 891, 4253.
Baddeley, A. D., and Hitch, G. (1974). Working memory. In G. A.
provides the necessary representation for flexible Bower, ed., Recent advances in learning and motivation. New
memory expression. Empirical studies (Fortin et al., York: Academic Press.
2002) and theoretical reviews (Manns and Squire, Bunsey, M., and Eichenbaum, H. (1996). Conservation of hippo-
2001; Eichenbaum et al., 1999) emphasize this aspect campal memory function in rats and humans. Nature 379
255257.
of hippocampal function, which was the cognitive
Diamond, D. M., Fleshner, M., Ingersoll, N., and Rose, G. M.
core of Oltons working memory theory. (1996). Psychological stress impairs spatial working memory:
Relevance to electrophysiological studies of hippocampal
Applications function. Behavioral Neuroscience 110, 661672.
Working memory procedures in general, and the Eichenbaum, H., Stewart, C., and Morris, R. G. M. (1990). Hippo-
radial maze in particular, remain in use in important campal representation in place learning. Journal of Neuro-
science 10, 3,5313,542.
studies to assess hippocampal function and memory Garner, R. W., Hake, H. W., and Eriksen, C. W. (1956). Operation-
in many species, including mice, nonhuman pri- ism and the concept of perception. Psychological Review 63 (3),
mates, and humans (Glassman et al., 1994). To cite 149159.
only a few examples: The tight relationship between Givens, B. S., and Olton, D. S. (1990). Cholinergic and GABAergic
modulation of medial septal area: Effect on working memory.
place field activity and spatial behavior in rats was first
Behavioral Neuroscience 104, 849855.
shown in a spatial-working memory task in a four-arm Glassman, R. B., Garvey, K. J., Elkins, K. M., Kasal, K. L., and
radial maze (OKeefe and Speakman, 1987). Psycho- Couillard, N. L. (1994). Spatial working memory score of hu-
logical stress and concomitant elevated corticoste- mans in a large radial maze, similar to published score of rats,
roids were shown to selectively impair working mem- implies capacity close to the magical number 7 2. Brain Re-
search Bulletin 34, 151159.
ory in a radial maze (Diamond et al., 1996). Age-
Hampton, R. R., Shettleworth, S. J., and Westwood, R. P. (1998).
related memory deficits in nonhuman primates were Proactive interference, recency, and associative strength:
examined using a formal variant of the radial-maze Comparisons of black-capped chickadees and dark-eyed jun-
task (Rapp et al., 1997). A working-memory proce- cos. Animal Learning and Behavior 26, 475485.
dure in the water maze helped to clarify the impor- Handelmann, G. E., and Olton, D. S. (1981). Spatial memory fol-
lowing damage to hippocampal CA3 pyramidal cells with
tance of hippocampal NMDA receptors in memory kainic acid: Impairment and recovery with preoperative train-
(Steele and Morris, 1999). Finally, recent efforts to ing. Brain Research 217, 4158.
model Alzheimers disease in transgenic mice have Honig, W. K. (1978). Animal memory and animal learning. In H.
used a working-memory task in a radial water maze L. Roitblat, T. G. Bever, and H. S. Terrace, eds., Animal Cogni-
to track the progression of age-related memory defi- tion. Hillsdale, NJ: Erlbaum.
Jarrard, L. E. (1986). Selective hippocampal lesions and behavior:
cits (e.g. Arandash et al., 2001). Oltons hypotheses, Implications for current research and theorizing. In R. L.
analyses, and approaches continue to influence mem- Isaacson and K. H. Pribram, eds., The hippocampus, Vol. 4.
ory science. New York: Plenum.
OPERANT BEHAVIOR 493
Kesner, R. P., Bolland, B. L., and Dakis, M. (1993). Memory for Tulving, E. (1972). Episodic and semantic memory. In E. Tulving
spatial locations, motor responses, and objects: Triple dissoci- and W. Donaldson, eds., Organization of memory. New York:
ation among the hippocampus, caudate nucleus, and ex- Academic Press.
trastriate visual cortex. Experimental Brain Research 93, 462 Wan, R. Q., Pang, K., and Olton, D. S. (1994). Hippocampal and
470. amygdaloid involvement in nonspatial and spatial working
Kogan, J. H., Frankland, P. W., and Silva, A. J. (2000). Long- term memory in rats: Effects of delay and interference. Behavioral
memory underlying hippocampus-dependent social recogni- Neuroscience 108, 866882.
tion in mice. Hippocampus 10, 4756. White, N. M., and Oellet, M.C. (1997). Roles of movement and
McDonald, R. J., and White, N. M. (1993). A triple dissociation of temporal factors in spatial learning. Hippocampus 7, 501510.
memory systems: Hippocampus, amygdala, and dorsal stria- Wible, C. G., Findling, R. L., Shapiro, M., Lang, E. J., Crane, S.,
tum. Behavioral Neuroscience 107, 322. and Olton, D. S. (1986). Mnemonic correlates of unit activity
Mishkin, M., and Delacour, J. (1975). An analysis of short-term vi- in the hippocampus. Brain Research 399, 97110.
sual memory in the monkey. Journal of Experimental Psychology Winocur, G. (1990). Anterograde and retrograde amnesia in rats
[Animal Behavior] 1, 326334. with dorsal hippocampal or dorsomedial thalamic lesions. Be-
Morris, R. G. M., Garrud, P., Rawlins, J. N. P., and OKeefe, J. havioural Brain Research 38, 145154.
(1982). Place navigation impaired in rats with hippocampal
lesions. Nature 297, 681683. Matthew L. Shapiro
OKeefe, J., and Speakman, A. (1987). Single unit activity in the rat
hippocampus during a spatial memory task. Experimental
Brain Research 68, 127.
Olton, D. S., Becker, J. T., and Handelmann, G. H. (1979). Hippo-
campus, space and memory. Behavioral and Brain Sciences 2,
313365.
OPERANT BEHAVIOR
Olton, D. S., Branch, M., and Best, P. J. (1978). Spatial correlates In the 1930s B. F. Skinner developed a new method-
of hippocampal unit activity. Experimental Neurology 58, 387
ology for the study of animal learning and behavior.
409.
Olton, D. S., and Feustle, W. A. (1981). Hippocampal function re- He called it operant behavior, to reflect the fact that the
quired for nonspatial working memory. Experimental Brain Re- animal operated on the environment to produce a
search 41, 380389. reward, or reinforcer. The Behavior of Organisms, pub-
Olton, D. S., and Papas, B. C. (1979). Spatial memory and hippo- lished in 1938, was the principal document in which
campal function. Neuropsychologia 17, 669682. he presented his findings and his conceptual ap-
Olton, D. S., and Samuelson, R. J. (1976). Rememberance of places
passed: Spatial memory in rats. Journal of Experimental Psychol-
proach to the study of animal learning and behavior.
ogy: Animal Behavior Processes 2, 97116. In the method that Skinner developed, the ani-
Olton, D. S., and Shapiro, M. L. (1992). Mnemonic dissociations:
The power of parameters. Journal of Cognitive Neuroscience 4,
mal (most often a rat, pigeon, or monkey) emits par-
200207. ticular behaviors, called instrumental responses (or be-
Olton, D. S., Walker, J. A., and Gage, F. H. (1978). Hippocampal haviors), to gain a reinforcer. Most often, these
connections and spatial discrimination. Brain Research 139, responses involve an operandum (formerly called
295308. manipulandum) that is suited to the subjects motor
Olton, D. S., Wenk, G. L., Church, R. M., and Meck, W. H. (1988).
abilities. Rats, monkeys, and other mammals press a
Attention and the frontal cortex as examined by simultaneous
temporal processing. Neuropsychologia 26, 307318. horizontal bar (or lever) in the experimental chamber
Olton, D. S., and Wolf, W. A. (1982). Hippocampal seizures pro- (often called a Skinner box), while pigeons peck at a
duce retrograde amnesia without a temporal gradient when vertical disk (or key); fish can be taught to swim
they reset working memory. Behavioral and Neural Biology 33, through a ring. Normally, the reinforcer immediately
437452. follows the response.
Platt, J. R. (1964). Strong inference. Science 146, 347353.
Raffaele, K. C., and Olton, D. S. (1988). Hippocampal and amyg- Animals learn to emit particular instrumental re-
daloid involvement in working memory for nonspatial stimu- sponses because the reinforcers shape behavior. Be-
li. Behavioral Neuroscience 102, 349355.
Rapp, P. R., Kansky, M. T., and Roberts, J. A. (1997). Impaired
haviors that are followed by a reinforcer increase in
spatial information processing in aged monkeys with pre- frequency, and behaviors that are not followed by a
served recognition memory. NeuroReport 8, 1,9231,928. reinforcer decrease in frequency. For example, to
Scoville, W. B., and Milner, B. (1957). Loss of recent memory after train a rat to press a lever, the experimenter may first
bilateral hippocampal lesions. Journal of Neurology Neurosur- reinforce the animal every time it approaches the
gery and Psychiatry 20, 1121.
lever. When the rat is reliably approaching the lever,
Sidman, M., Stoddard, L. T., and Mohr, J. P. (1968). Some addi-
tional quantiative observations of immediate memory in a pa- reinforcers are provided only if it actually touches the
tient with bilateral hippocampal lesions. Neuropsychologia 6, lever. Finally, only pressing the lever is reinforced.
245254. This shaping of behavior by progressively narrowing
Steele, R. J., and Morris, R. G. M. (1999). Delay-dependent impair- the range of behaviors that are reinforced (the operant
ment of a matchingtoplace task with chronic and intrahip-
class) is known as the method of successive approxima-
pocampal infusion of the NMDA-antagonist D-AP5. Hippo-
campus 9, 118136.
tion. If reinforcement for a behavior is discontinued,
Tolman, E. C. (1948). Cognitive maps in rats and men. Psychological the behavior will decrease in frequency and may stop
Review 56, 144155. completely. This process is known as extinction.
494 OPERANT BEHAVIOR
Figure 1
Sample cumulative records of lever pressing on various simple reinforcement schedules. Horizontal displacements in the records
indicate passage of time. Vertical displacements indicate cumulative responses. Hatch marks indicate times when the reinforcer is
delivered. CRF is continuous reinforcement; FR, fixed ratio; VR, variable ratio; FI, fixed interval; and VI, variable interval.
(Hypothetical data)
In discrete-trial procedures, the trial ends with a for a human subject. Skinner designed schedules of re-
single response, and the probability, latency, or force inforcement that provided reward only intermittently,
of that response is recorded as the measure of behav- in contrast with continuous reinforcement, where each
ior. Skinner developed another method of studying response is reinforced. The subject may be reinforced
behavior that he called free-operant procedures. Here, only after emitting a number of responses, on a ratio
the subject has access to the operandum for extended schedule, or for a response after a period of time has
periodssometimes an extended trial, on other occa- elapsed, on an interval schedule. The required ratio
sions an entire experimental sessionand can re- may be constant on all occasions; this is a fixed-ratio
spond repeatedly during that period. Therefore, the schedule. Or it may vary from trial to trial; this is a
rate of responding becomes the primary measure of be- variable-ratio schedule.
havior. Skinner developed an ingenious method for
displaying the rate with a cumulative record (see Figure Likewise, in an interval schedule the interval may
1). Each response displaces a pen upward by a small be fixed or variable. Skinner found that each of these
amount on a moving strip of paper. This makes the schedules produced distinctive cumulative records.
rate of responding immediately visible as the measure For example, in fixed-ratio schedules, animals fre-
of behavior. The higher the rate of responding, the quently do not respond immediately after a reinforc-
steeper the slope of the cumulative record. However, er; this is called a post-reinforcement pause. Then they
in most current experimental applications, counters emit responses in a high-rate burst to obtain the re-
and computers are used to record and analyze reinforcer. In fixed-interval schedules, the subject typi-
sponse output. These measures allow for more quan- cally does not respond immediately after the reinforc-
titative analyses of behavior. er, and the rate of responding steadily accelerates as
the end of the interval approaches. Variable-interval
and variable-ratio schedules usually generate steady
Schedules of Reinforcement rates of responding. Ratio schedules generally pro-
The designated instrumental response is followed duce high rates of responding because the rate of re-
on at least some occasions by a reinforcer such as a inforcement depends entirely on the rate of respond-
food pellet or liquid refreshment for the rat or mon- ing. However, ratio schedules requiring a large
key, grain for the bird, or money, tokens, or points number of responses for each reinforcer may induce
OPERANT BEHAVIOR 495
ratio strain in the form of extended periods of no re- attainment when the stimuli belong to different classes.
sponding. Pigeons, for example, readily learn to discriminate
between pictures containing images of one or more
These simple schedules of reinforcement can be
people and pictures without a person.
combined into more complex schedules. One sched-
ule may produce yet another schedule before a rein- Stimulus control is also studied using discretetrial
forcer is given, a chain schedule, or two schedules may choice procedures. A stimulus is presented as a sam-
regularly alternate on one operandum, a multiple ple, and then the animal must choose which of two re-
schedule. In these schedules, distinctive stimuli signal sponse alternatives is correct for that particular stimu-
which particular schedule is currently in effect. In a lus. Correct choices are reinforced. Such methods are
mixed schedule, the component schedules alternate, analogous to signal detection experiments with human
but they are not signaled by an external cue. subjects and have provided precise measurements of
animal perception. If a delay intervenes between the
In concurrent schedules, two (or more) schedules
sample stimulus and the choice, the short-term mem-
are simultaneously in effect and the subject can
ory or working memory of animals can be studied.
choose between them. These schedules can be ar-
Generally, the accuracy of choice decreases markedly
ranged on separate operanda or on one operandum.
with delays of even a few seconds.
In the latter procedure the subject can choose be-
tween schedules by performing a switching response
to a different operandum. It has been found that ani- Control with Aversive Stimuli
mals distribute the time spent responding to each
Positive reinforcers are normally appetitive stimu-
schedule in proportion to the rate of reinforcement
li. Aversive stimuli, such as electric shock or loud noise,
obtained from each. This relation is known as the
are also effective in the control of behavior. If aversive
matching law. Type of schedule, magnitude of the re-
stimuli are consequences for responding, they are
inforcers, and type of reinforcement are also impor-
punishers, and they reduce the rate of responding,
tant determinants of choice. For example, studies of
which is otherwise maintained by positive reinforce-
self-control have shown that animals are impulsive;
ment. Animals are very sensitive to both the strength
they choose small, immediate reinforcers over de-
and the frequency of the punishers. Aversive stimuli
layed, but much larger, reinforcers.
are also used in the study of escape and avoidance. The
latter is most often studied in a free-operant situation.
Stimulus Control The subject, most often a rat, is subjected to brief, in-
termittent shocks. By emitting a required response,
Discriminative stimuli can signal the effective such as bar pressing or crossing a hurdle, the subject
schedule of reinforcement. For rats, these can be dif- can postpone or cancel the shock. This procedure
ferent tones or the presence or absence of a house generates consistent rates of avoidance behavior in
light in the chamber. For pigeons, different colors rats, monkeys, and other organisms, especially when
or patterns may be projected onto the response key. each response guarantees a shock-free interval.
Monkeys are often presented with complex visual pat-
terns. The discriminative stimuli come to control the
rates of responding. For example, a pigeon will re- Summary
spond at the same rate to a key lit red or green if both
Operant methodology has shown that animal be-
colors signal a variable-interval (VI) schedule. Howev-
havior is an orderly function of its antecedents (dis-
er, if the VI schedule during the green-light compo-
criminative stimuli) and its consequences (reinforce-
nent is removed, then the rate of responding to this
ment and punishment). It has also enabled
negative stimulus rapidly decreases. The response
experimenters to explore various areas of animal per-
rate to the red light, the positive stimulus, will actually
ception, cognition, and choice. Furthermore, the
increase over its previous level, a phenomenon called
principles of operant behavior have application to
behavioral contrast. New stimuli from the same stimulus
humans. Operant techniques have been employed in
dimension can be presented in a generalization test. For
personal instruction and in the treatment of dysfunc-
example, if the discriminative stimuli used in training
tional human behavior.
are two tones, then a rat may be tested with a range
of tonal frequencies. Gradients of generalization (or dis- See also: CONDITIONING, CLASSICAL AND
crimination) are readily obtained; that is, the amount INSTRUMENTAL; DISCRIMINATION AND
of responding to each new stimulus is an orderly func- GENERALIZATION; REINFORCEMENT
tion of its similarity to the positive training stimulus.
Bibliography
If the stimuli are more complex, such as pictures, Catania, A. C. (1979). Learning. Englewood Cliffs, NJ: Prentice-
this provides an opportunity for the study of concept Hall.
496 ORAL TRADITIONS
Domjan, M. P., and Burkhard, B. (1985). The principles of learning at least as well formed and strict as a college students
and behavior, 2nd edition. San Francisco: Brooks/Cole. knowledge of going to the dentists office or a fast-
Flaherty, C. F. (1985). Animal learning and cognition. New York:
Knopf.
food restaurant. The forms of thematic organization
Schwartz, B., and Reisberg, D. (1991). Learning and memory. New allow singers to expand or contract their story at will,
York: Norton. as is common in epic (Lord, 1960).
Skinner, B. F. (1938). The behavior of organisms. New York: Apple-
ton-Century. Poetics and music each add their own unique con-
W. K. Honig tribution. When two words in a ballad are linked by
Brent Alsop rhyme or alliteration, undergraduates have a higher
recall for them than when the poetics are broken.
Furthermore, when ballad singers perform the same
ballad twice, they are less likely to change poetically
ORAL TRADITIONS linked words (Wallace and Rubin, 1988). Some
genres, such as counting-out rhymes, have nearly all
Oral traditions depend on human memory for their their words poetically linked (Rubin, Ciobanu, and
preservation. Songs or stories from a tradition must Langston, 1997), whereas others have minimal poet-
be stored in one persons memory and passed to an- ics. Scientists know from the extensive research con-
other person who can also remember and retell them.
ducted by Douglas Nelson and Cathy McEvoy (Nel-
This process must occur over many generations. For
son, 1981) that rhyme cues function differently than
example, verses from versions of ballads collected in
meaning cues. It is as if rhyme cues a whole set, while
the 1600s in Great Britain are similar to versions col-
meaning cues, when available, single out the target.
lected since the 1980s in North Carolina. Most of the
Rhyme, as opposed to meaning cues, tends to work
words have changed, but the basic ideas and poetic
best with fast presentation rates, small set size, and
structures have not. Similarly, the counting-out
strong cue strengththree conditions that tend to be
rhyme Eenie meenie has remained stable, though
present in the small world of oral traditions. Thus,
with much less change, for since the end of the nine-
rhymes have their own peculiar properties, which
teenth century. Rote memorization is not occurring.
have been studied extensively and which are often
Rather there is evidence that poetic and meaning
rules are being transmitted. Oral traditions must, well suited for oral traditions. This trait is true even
therefore, have forms of organization and modes of with subjects not trained to attend to rhyme the way
transmission to decrease the changes that human users of many oral traditions are.
memory usually imposes on verbal material (Rubin, Oral traditions are rhythmic. Rhythm functions
1995). in at least four ways: (1) rhythm is a constraint, like
The major forms of organization that contribute others, limiting word choice to words with the correct
to stability of oral traditions include imagery, gist, number of syllables or stress pattern; (2) rhythm
rhyme, alliteration, rhythm, and music. creates slots that need to be filled, producing a de-
Imagery is perhaps the most powerful and wide- mand characteristic to recall and thereby favors
spread factor in mnemonic systems. As Allen Paivio changes within a rhythmic unit rather than errors of
(1971) points out, imagery is most effective for con- omission; (3) rhythm, like meaning, provides an orga-
crete (versus abstract), parallel-spatial (versus sequen- nization, allowing singers to select, substitute, add, or
tial), and dynamic (versus static) processing. Oral tra- delete whole rhythmic units and continue, and such
ditions predominantly consist of sequences of rhythmic units typically coincide with meaning units
concrete actions by active agents, not abstract princi- in oral traditions (Lord, 1960); and (4) rhythm em-
ples (Havelock, 1978). In ballads, for example, verses phasizes certain locations within lines, that facilitate
that contain concrete, imageable actions are recalled other constraints, such as the placing of rhyme and al-
better than ones that do not contain such actions literation on stressed syllables.
(Wallace and Rubin, 1988). Imagery, meaning, poetics, and music all provide
Meaning or thematic organization plays a large forms of organization or constraint. Once the proper-
role in adult oral traditions. The cognitive psycholo- ties of each form of organization are listed, it is easy
gists descriptions of such organization, including to add the constraints together to produce an impres-
schemas, scripts, story grammars, and causal chains sive total degree of constraint. However, more than
can all be used to quantify and describe thematic or- additive effects are found. For example, although a
ganization, although the rules for these vary from tra- rhyme or meaning cue by itself may not lead to recall
dition to tradition. For instance, common scripts in of the last word of a line, when combined they can be
epics include arming a hero or the heros horse, as- effective because there is often only one word that fits
sembling an army, and joining battle. The scripts are them both.
ORIENTING REFLEX HABITUATION 497
Besides interaction effects that limit word Wallace, W. T., and Rubin, D. C. (1988). The wreck of the old 97:
choices, the specific properties of the various forms of A real event remembered in song. In U. Neisser and E. Wino-
grad, eds., Remembering reconsidered: Ecological and traditional
organization complement each other. For instance, approaches to the study of memory. Cambridge, UK: Cambridge
imagery leads to the original verbal stimulus being University Press.
transformed into a nonverbal, atemporal representa-
David C. Rubin
tion. When a verbal output is needed, the original
words and the order of presentation will not be avail-
able for retrieval and will be generated from the
image, resulting in changes in wording and the order
of ideas. Thematic organization, such as scripts, story ORIENTING REFLEX HABITUATION
grammars, and causal chains, however, function to The orienting reflex (OR) is a complex response of
preserve the temporal order lost by imagery. Even so, the organism to a novel stimulus. It was discovered by
in most models of memory, words are translated to Ivan Pavlov ([1927] 1960) as an interruption of ongo-
and from a more abstract representation that con- ing activity by presentation of an unexpected stimulus
tains none of the sound pattern, allowing for the pos- (external inhibition). This inhibition of the ongoing
sibility of translation errors. This remaining lack is activity, accompanied by somatic, vegetative, electro-
remedied by poetics and music, which preserve the encephalographic, humoral, and sensory manifesta-
sound pattern. tions, was termed the what-is-it reflex. The OR is a
Many strategies of transmission add to the stabili- set of components contributing to optimize the condi-
ty provided by the organizational constraints out- tions of stimulus perception. A sequence of ORs di-
lined. Songs in an oral tradition are recalled repeat- rected toward new aspects of the environment consti-
edly after they have been mastered, that is they tutes an exploratory behavior. The somatic
benefit from overlearning. Moreover, overlearning is components of the OR are represented by eye and
usually spaced over time, in some cases once a year head targeting movements, perking of ears, and sniff-
when the appropriate season arrives. Overlearning ing. The vasoconstriction of peripheral vessels and
and spaced practice are two of the most powerful fac- vasodilation of vessels of the head, heart rate deceler-
tors in maintaining material in memory for long peri- ation, and skin galvanic response (SGR) constitute
ods. In addition, there are social supports aiding sta- vegetative OR components. Positron Emission To-
bility. In many genres, only experts who are suited by mography has demonstrated enhancement of blood
interest and ability are the active transmitters. They supply in different brain areas during sensory stimu-
hear their songs from more than one source, which lation. The electroencephalographic manifestation of
allows better variants to replace inferior ones. Their OR is characterized by negative steady potential shift
audience, though it may not be able to supply alterna- that parallels a transition from slow-wave brain activi-
tives, can show approval or disapproval of what it ty to high-frequency oscillations, demonstrating an
hears. enhancement of the arousal level (Lindsley, 1961).
Humoral components of OR are represented by
(-endorphin and acetylcholine released within brain
Bibliography tissues. The sensory components of OR are expressed
Foley, J. M. (1988). The theory of oral composition: History and method- in a lowering of sensory thresholds and increase of fu-
ology. Bloomington: Indiana University Press. sion frequency.
Havelock, E. A. (1978). The Greek concept of justice: From its shadow
in Homer to its substance in Plato. Cambridge, MA: Harvard The repeated presentation of a stimulus results in
University Press. a gradual decrement of OR components, called habit-
Kelly, M. H., and Rubin, D. C. (1988). Natural rhythmic patterns uation. The process of habituation is stimulus-
in English verse: Evidence from child counting-out rhymes.
Journal of Memory and Language 27, 718740. selective. That selectivity can be demonstrated with
Lord, A. B. (1960). The singer of tales. Cambridge, MA: Harvard respect to elementary features (intensity, frequency,
University Press. color, location, duration) as well as to complex aspects
Nelson, D. L. (1981). Many are called but few are chosen: The in- of stimuli (shape, accord, heteromodal structure).
fluence of context on the effects of category size. In G. H.
The habituation of the OR is also semantically selec-
Bower, ed., The psychology of learning and motivation, Vol. 15.
New York: Academic Press. tive, indicating a high level of abstraction in the OR
Paivio, A. (1971). Imagery and verbal processes. New York: Holt, Rine- control. In the process of habituation of OR, a neuro-
hart and Winston. nal model of the presented stimulus is elaborated in
Rubin, D. C. (1995). Memory in oral traditions: The cognitive psychology the brain. Any change of stimulus parameters with re-
of epic, ballads, and counting-out rhymes. New York: Oxford Uni-
spect to the established neuronal model results in an
versity Press.
Rubin, D. C., Ciobanu, V., and Langston, W. (1997). Childrens elicitation of the OR. After a response to a novel stim-
memory for counting-out rhymes: A cross language compari- ulus, one sees the OR recover to a standard stimulus,
son. Psychonomic Bulletin & Review 4, 421424. a phenomenon called dishabituation. The OR is
498 ORIENTING REFLEX HABITUATION
evoked by a mismatch signal resulting from the com- perimental procedure, the OR is reestablished. The
parison of the presented stimulus with the established more difficult is the differentiation of signals, the
neuronal model. If the stimulus coincides with the greater the OR. Thus the magnitude and stability of
neuronal model, no OR is generated. The neuronal ORs depend on novelty, significance, and task diffi-
model can be regarded as a multidimensional, self- culty. Involuntary and voluntary ORs can be integrat-
adjustable filter shaped by a repeatedly presented ed within a common attentional process: A novel non-
stimulus. The magnitude of the OR depends on the signal stimulus triggering an involuntary OR followed
degree of noncoincidence of the stimulus with the by a voluntary OR constitutes sustained attention.
shape of the multidimensional filter. In accordance
The OR has its own reinforcement value and can
with the degree of spreading of excitation, local and
be used as a reinforcement in the elaboration of con-
generalized forms of ORs are distinguished. Short
ditioned ORs. The (-endorphin released by novel
and long duration of excitation constitutes a basis for
stimulus presentation plays a role of positive rein-
separation of phasic and tonic forms of ORs. In the
forcement in a search for novelty. The OR can con-
process of habituation, tonic and generalized forms of
tribute as an exploratory drive in selection of new
OR are transformed into phasic and local ones
combinations of memory traces during creative activi-
(Sokolov, 1963).
ty.
The habituation of the OR can be studied using
The OR at the neuronal level is represented by
event-related potentials (ERPs) represented by a se-
several populations of cells. The most important are
quence of positive (P1, P2, P3) and negative (N1, N2)
novelty detectors, represented by pyramidal hippo-
brain waves elicited by stimulus onset.
campal cells characterized by universally extended re-
The computer-based isolation of separate ERPs ceptive fields. Being activated by novel stimuli, these
evoked by rare stimuli demonstrates a partial habitua- cells demonstrate stimulus-selective habituation that
tion of vertex N1 that parallels the habituation of parallels the OR habituation at the macro level. Any
SGR. A novel stimulus results in an increase of N1 and change of input stimulus results in their spiking
evocation of SGR. Thus the stimulus deviating from again. Thus a multidimensional neuronal model of a
the neuronal model triggers a modality-nonspecific stimulus is formed at a single pyramidal neuron of the
negativity overlapping the stable part of N1 (Verba- hippocampus. The selectivity of the neuronal model
ten, 1988). The deviant stimuli following with short is determined by specific neocortical feature detectors
intervals among standard ones generate a modality- extracting different properties of the input signal in
specific mismatch negativity overlapping N1-P2 com- parallel. The feature detectors characterized by stable
ponents (Ntnen, 1990). The OR evoked by non- responses converge on novelty detectors through
signal stimuli is termed involuntary OR. It differs from plastic (modifiable) synapses. The plasticity of syn-
an OR evoked by signal stimuli, which is termed vol- apses on novelty detectors is dependent on hippo-
untary OR (Maltzman, 1985). The OR habituated to campal dentate granule cells. A set of excitations gen-
a nonsignal stimulus is recovered under the influence erated in selective feature detectors reach pyramidal
of verbal instruction announcing that the stimulus is and dentate cells in parallel. The dentate cells have
a target of the response. Such an enhancement of an synapses on pyramidal cells controlling the habitua-
OR due to verbal instruction is lacking in patients tion process. The synapses of feature detectors consti-
with frontal lobe lesions, whereas their ORs to nonsig- tute a map of features on a single novelty detector.
nal stimuli remain intact (Luria, 1973).
The neuronal model is represented on such a fea-
The verbal instruction actualizes a memory trace ture map by a specific pattern of synapses depressed
of the target stimulus. The presented stimuli are by repeated stimulus presentations. The output sig-
matched against the memory trace. The match signal nals of novelty detectors are fed to activating brain-
is evident in brain ERPs as a processing negativity stem reticular formation neurons, generating an
overlapping N1-N2. The processing negativity is the arousal reaction. The rest of the hippocampal pyra-
greater the more closely the stimulus matches the midal neurons are sameness detectors characterized
memory trace, which is activated by verbal instruction by a background firing. A new stimulus results in an
(Ntnen, 1990). Similar enhancement of OR can be inhibition of their spiking. This inhibitory reaction is
observed in the process of elaboration of conditioned habituated by repeated stimulus presentation. The
reflexes. The nonsignal stimulus evoking no OR after inhibitory response is evoked again by any stimulus
habituation produces an OR again after its reinforce- change. The maximal firing rate is observed in same-
ment. During conditioned reflex stabilization, OR is ness neurons under familiar surroundings. The out-
gradually extinguished, but more slowly than in re- put signals from sameness detectors are directed to
sponse to a nonsignal stimulus. When a new nonrein- inactivating reticular formation neurons, inducing
forced differential stimulus is introduced into the ex- drowsiness and sleep. The selective habituation of the
ORIENTING REFLEX HABITUATION 499
Figure 1
pyramidal cell responses is based on the potentiation sensitization of activating units, novel stimulus results
of synapses of dentate cells on pyramidal neurons. in an electroencephalographic arousal correlated
Under such potentiation of dentate synapses, the py- with sensitization of feature detectors and external in-
ramidal cells stop responding to afferent stimuli. The hibition of ongoing activity at the level of the com-
injection of antibodies against hippocampal granule mand neurons. The repeated presentation of a stimu-
cells results in elimination of pyramidal cell habitua- lus switches on sameness neurons, which, with
tion (Vinogradova, 1970). participation of inactivating units, induce lowering of
the arousal level, expressed in drowsiness and sleep.
Such are the neuronal mechanisms of involuntary
OR habituation. The neuronal mechanisms of volun- Bibliography
tary OR are more complex (see Figure 1). The stimuli Lindsley, D. B. (1961). The reticular activating system and percep-
analyzed by feature detectors are recorded by memo- tual integration. In D. E. Sheer, ed., Electrical stimulation of the
brain. Austin: University of Texas Press.
ry units of the association cortex. The verbal instruc-
Luria, A. R. (1973). The frontal lobes and the regulation of behav-
tion, through semantic units with the participation of ior. In K. H. Pribram and A. R. Luria, eds., Psychophysiology of
frontal lobe mechanisms, selects a set of memory the frontal lobes. New York: Academic Press.
units as a template. The match signal generated by Maltzman, I. (1985). Some characteristics of orienting reflexes. Psy-
memory units of the template is recorded as process- chiatry 2, 913916.
Ntnen, R. (1990). The role of attention in auditory information
ing negativity. The match signal is addressed to nov- processing as revealed by event-related potentials and other
elty detectors, resulting in an enhancement of OR to brain measures of cognitive function. Behavioral and Brain Sci-
significant stimuli. Through novelty detectors and ences 13, 201288.
500 OSCILLATIONS, SYNCHRONY, AND NEURONAL CODES
Pavlov, I. P. (1927; reprint 1960). Conditioned reflexes: An investiga- to be signaled by the simultaneous responses of the
tion of the physiological activity of the cerebral cortex., trans. and ed. ensemble of cells responding to the components.
G. V. Anrep. New York: Dover.
Sokolov, E. N. (1963). Perception and conditioned reflex. Oxford:
Thus, in ensemble coding individual neurons can
Pergamon Press. contribute at different times to the representation of
(1975). The neuronal mechanisms of the orienting reflex. different objects by forming ensembles with varying
In E. N. Sokolov and O. S. Vinogradova, eds., Neuronal mecha- partners. A neuron tuned to a particular component
nisms of the orienting reflex. Hillsdale, NJ: Erlbaum. can then contribute to the representation of all ob-
Verbaten, M. N. (1988). A model for the orienting response and
its habituation. Psychophysiology 25, 487488.
jects containing this particular component and neu-
Vinogradova, O. S. (1970). Registration of information and the rons representing elementary features can be recom-
limbic system. In G. Horn and R. Hind, eds., Short-term bined in ever-changing constellations to represent
changes in neuronal activity and behavior. Cambridge, UK: Cam- novel objects.
bridge University Press.
E. N. Sokolov
How to Tag Responses as Related?
Numerous theoretical studies have addressed the
question how assemblies can organize themselves
OSCILLATIONS, SYNCHRONY, AND through cooperative interaction within associative
NEURONAL CODES neuronal networks. Here we focus on the problem of
how cells can be tagged as related when they are
Cognitive systems have to explore a huge combinato- grouped into an assembly. An unambiguous signature
rial space when searching for the consistent relations of relatedness is absolutely crucial for assembly codes
among features that define a perceptual object. Thus, because, in contrast to labeled line codes, the mean-
mechanisms are required that permit rapid analysis ing of responses changes with the interpretive con-
and representation of relations between the re- text, thus rendering false conjunctions deleterious for
sponses of neurons whose activity signals the presence object recognition. The required mechanism must as-
of particular features. A common and well- sure that the responses of the neurons that constitute
documented strategy for the binding of distributed an assembly are processed and evaluated together at
responses is the implementation of conjunction- subsequent processing stages and are not confounded
specific neurons that receive convergent input from with other, unrelated responses. In principle, such a
elementary feature detectors and therefore respond process can be achieved by raising jointly and selec-
selectively to the conjunctions of the respective fea- tively the saliency of the responses belonging to an as-
tures. This process is known as labeled line coding. sembly.
However, this coding strategy, if not complemented
by an additional binding mechanism, meets with a There are three ways to achieve this goal: the in-
number of problems. First, large numbers of conjunc- hibition and exclusion of responses from further pro-
tion units are required for the exhaustive representa- cessing, the enhancement of the discharge frequency
tion of the manifold intra- and cross-modal feature of the selected responses, and the inducement of pre-
constellations of real-world objects. Second, it is hard cise synchrony of the selected cells. All three mecha-
to see how novel objects and hence entirely new rela- nisms enhance the relative impact of the selected re-
tions among features can be recognized and repre- sponses and can therefore be used to tag them as
sented because this would require rapid reconfigura- related. Single-cell studies have provided robust evi-
tion of input connections to previously uncommitted dence that the first two mechanisms play a crucial role
cells. Third, unresolved problems arise with the rep- in the selection and grouping of responses.
resentation of the nested relations among the compo- The simultaneous organization of assemblies
nents of composite objects such as visual scenes or sharing common subsets of neurons is precluded by
sentences (Singer, 1999; von der Malsburg, 1999). the impossibility of knowing which of the shared neu-
A complementary strategy is needed, therefore, rons would belong to which assembly. Hence, the fre-
that permits a more flexible definition of relations quent overlapping of assemblies necessitates a tempo-
than can be achieved with hard-wired conjunction ral segregation through multiplexing. Processing
units. As proposed by Donald Hebb (1949) and later speed is then limited essentially by the rate at which
elaborated by others, that complementary strategy is different assemblies can follow one another and thus
assembly coding. The assumption is that only compo- by the temporal resolution of the mechanism that la-
nents of objectswhich may consist of rather com- bels responses as related. Through synchronization
plex conjunctions of elementary features but may be the temporal regrouping of spikesthe saliency of
common to different objectsare represented by in- responses can be modulated with higher temporal
dividual cells. The presence of the whole object seems resolution than with tonic changes in the firing rate.
OSCILLATIONS, SYNCHRONY, AND NEURONAL CODES 501
Synchronization exploits exclusively spatial and no frequency range are particularly well expressed when
additional temporal summation, and therefore this the brain is in an activated statewhen the EEG is de-
tagging mechanism can operate in principle with a synchronized and exhibits high power in the - and
temporal resolution at the level of individual spikes. -frequency range. Such EEG patterns are character-
Using synchronization as a complementary mecha- istic of the aroused, attentive brain, suggesting a role
nism for the definition of relations also permits the for synchronization in cognitive processes. This sug-
possible advantage of specification of relations inde- gestion is supported by numerous observations in
pendently of the firing rate. The discharge rates of both animals and human subjects that synchronous
neurons depends on numerous variables such as the oscillations in the -frequency range and their syn-
physical energy of stimuli or the match between stim- chronization become more prominent during states
ulus and receptive field properties, and it may not al- of focused attention or when subjects are engaged in
ways be obvious how these modulations of response cognitive tasks that put strong demands on feature
amplitude can be distinguished from those signaling binding or short-term memory functions (Tallon-
the relatedness of responses. Not all strong responses
Baudry and Bertrand, 1999; Engel, Fries, and Singer,
are necessarily related.
2001). Multielectrode recordings from awake cats and
monkeys trained to perform discrimination tasks in-
Synchrony as a Code for the Definition of dicate that attentional mechanisms enhance neuronal
Relations synchrony in anticipation of the expected task. There
is an increase in oscillatory activity in the gamma-
Some researchers have suggested that the cere-
frequency range that is associated with increased co-
bral cortex imposes a temporal micro-structure on
herence between the spontaneous discharges of
otherwise sustained responses and uses this temporal
patterning to express through synchronization the cells and the oscillations of the local field potential
degree of relatedness of the responses (Singer, 1999; (Roelfsema, Engel, Knig, and Singer, 1997).
Engel, Fries, and Singer, 2001). This suggestion en- These attentional effects appear to be selective
sued from the following discoveries: cortical neurons and confined to cortical areas and sites that need
often engage in synchronous oscillatory activity that to be engaged for the execution of the anticipated
is not stimulus-locked but caused by internal interac- tasks (Fries, Reynolds, Rorie, and Desimone,
tions (Gray and Singer, 1989); neurons distributed 2001).
both within and across cortical areas can synchronize
A close correlation between response synchroni-
their discharges within a millisecond; and synchroni-
zation and stimulus selection has been found in ex-
zation probability reflects common Gestalt-criteria of
perceptual grouping. periments on binocular rivalry that were performed
in strabismic animals (Fries, Schrder, Singer, and
If internally generated synchronization were to Engel, 2002). Because of experience-dependent
serve as a signature of relatedness, it would need to modifications of processing circuitry, perception in
meet several criteria: First, its precision should be in strabismic subjects always alternates between the two
the millisecond range to match the temporal windows
eyes. We have exploited this phenomenon of rivalry
for effective spatial summation and Hebbian modifi-
to investigate how neuronal responses that are select-
cations. Second, it must be possible to generate and
ed and perceived differ from those that are sup-
dissolve episodes of synchronous firing at a rate fast
pressed and excluded from supporting perception
enough to be compatible with known processing
(see Figure 1). A close and highly significant correla-
speed. Third, synchronized activity must be more ef-
fective than nonsynchronized activity in driving cells tion was observed between changes in the strength of
in target structures because it can only serve as a tag response synchronization and the outcome of rivalry.
of relatedness if it enhances the saliency of the syn- Cells mediating responses of the eye that won in in-
chronized responses. Fourth, there should be correla- terocular competition increased the synchronicity of
tions between the occurrence of synchronization pat- their responses upon presentation of the rivalrous
terns and perceptual or motor processes. These stimulus to the other, losing eye, whereas the reverse
predictions are supported by experimental evidence was true for cells driven by the eye that became sup-
(Gray, 1999; Singer, 1999). Here we shall review only pressed. Surprisingly, there were no consistent modi-
a selection of studies addressing the last postulate. fications of the amplitude of responses. It is only at
later processing stages that the poorly synchronized
responses to the suppressed stimuli fail to elicit
Attention and Response Selection suprathreshold responses and that cells respond only
Spike synchronization and the often concomitant to the selected stimulus (Leopold and Logothetis,
oscillatory patterning of responses in the - and - 1996).
Figure 1
Neuronal synchronization under conditions of binocular rivalry. (a) Using two mirrors, different patterns were
presented to the two eyes of strabismic cats. Panels (b-e) show normalized cross-correlograms for two pairs of
recording sites activated by the eye that won (b, c) and lost (d, e) in interocular competition, respectively. Insets above
the correlograms indicate stimulation conditions. Under monocular stimulation (b), cells driven by the winning eye
show a significant correlation, which is enhanced after introduction of the rivalrous stimulus to the other eye (c). The
reverse is the case for cells driven by the losing eye (compare conditions d and e). The white continuous line
superimposed on the correlograms represents a damped cosine function fitted to the data. RMA, relative modulation
amplitude of the center peak in the correlogram, computed as the ration of peak amplitude over offset of
correlogram modulation. This measure reflects the strength of synchrony.
Figure 2
A. Two superimposed gratings that differ in orientation and drift in different directions are perceived either as two independently
moving gratings (component motion) or as a single pattern drifting in the intermediate direction (pattern motion), depending on
whether the luminance conditions at the intersections are compatible with transparency. B. Predictions on the synchronization behavior
of neurons as a function of their receptive field configuration (left) and stimulation conditions (right). C. Changes in synchronization
behavior of two neurons recorded simultaneously from areas 18 and PMLS that were activated with a plaid stimulus under component
(upper graph) and pattern motion conditions (lower graph). The two neurons preferred gratings with orthogonal orientation (see
receptive field configuration, top, and tuning curves obtained with component and pattern, respectively) and synchronized their
responses only when activated with the pattern stimulus (compare cross-correlograms on the right).
Synchronization and Feature Binding some of the basic Gestalt criteria according to which
The hypothesis that internal synchronization of the visual system groups related features during scene
discharges groups responses for joint processing pre- segmentation. A series of studies provided evidence
dicts that synchronization probability should reflect that neurons distributed across different columns
within the same or different visual areas and even superimposed contours and thereby bias their associ-
across hemispheres synchronize their responses with ation with distinct surfaces.
almost no phase lag when activated with a single con-
tour but fire independently when stimulated simulta-
neously with two different contours. This pattern sug-
Conclusion
gests that synchronization results from a context- There is evidence that neuronal networks can
dependent selection and grouping process. The synchronize neuronal discharges with a precision in
probability and strength of response synchronization the millisecond range and appear to exploit this abili-
reflect indeed some of the elementary Gestalt-criteria ty for at least three purposes: first, for the precise sig-
that underly perceptual grouping. Stimulus configu- naling of temporal features across processing stages;
rations that comply with criteria such as continuity, second, for the selection of responses; and third, for
proximity, and similarity in the orientation domain, the definition of relations among distributed re-
colinearity, and common fate evoke synchronized responses with high temporal resolution. This selection
sponses with higher probability than configurations and binding mechanism is best for ensemble coding
that are devoid of groupable features (Singer, 1999). because it meets the requirement for flexible and
rapid binding of distributed responses in ever-
Researchers have observed an especially close changing constellations. Assembly coding, in turn,
correlation between neuronal synchrony and percep- appears necessary in order to cope with the represen-
tual grouping in experiments with plaid stimuli. tation of the astronomical number of possible rela-
These stimuli are well suited for the study of dynamic tions among features describing real-world objects.
binding mechanisms because minor changes of the
stimulus cause a binary switch in perceptual group- It appears, then, as though the cerebral cortex
applies two complementary coding strategies: an ex-
ing. Two superimposed gratings moving in different
plicit representation of features and their conjunc-
directions (plaid stimuli) may be perceived either as
tions in the tuned responses of individual specialized
two surfaces, one being transparent and sliding on
neurons or populations of such neurons and an im-
top of the other (component motion), or as a single
plicit representation of conjunctions of such explicitly
surface, consisting of crossed bars that moves in a di-
coded contents in dynamically associated assemblies.
rection intermediate to the component vectors (pat-
The first strategy seems to apply to the representation
tern motion). If this grouping of responses is initiated
of a limited set of features and some of their conjunc-
by selective synchronization, three predictions must
tions and is in all likelihood reserved for items that
hold (see Figure 2): First, neurons that prefer the di-
occur very frequently and/or are of particular behav-
rection of motion of one of the two gratings and have
ioral importance. The second strategy seems to be re-
colinearly aligned receptive fields should always syn-
served for the representation of novel objects and of
chronize their responses because they respond always
all those items for which an explicit representation
to contours that belong to the same surface. Second,
cannot be realized, either because the explicit repre-
neurons that are tuned to the respective motion di-
sentation would require too many neurons or because
rections of the two gratings should synchronize their
the contents to be represented are too infrequent to
responses in case of pattern motion because they then
warrant the implementation of specialized neurons.
respond to contours of the same surface, but they
should not synchronize in case of component motion See also: GUIDE TO THE ANATOMY OF THE BRAIN:
because their responses are then evoked by contours CEREBRAL CORTEX
belonging to different surfaces. Third, neurons pre-
ferring the direction of pattern motion should also Bibliography
synchronize only in the pattern and not in the compo- Castelo-Branco, M., Goebel, R., Neuenschwander, S., and Singer,
nent motion condition. W. (2000). Neural synchrony correlates with surface segrega-
tion rules. Nature 405, 685689.
Cross-correlation analysis of responses from cell Engel, A. K., Fries, P., and Singer, W. (2001). Dynamic predictions:
pairs distributed either within or across areas 18 and Oscillations and synchrony in top-down processing. Nature
the posterior medio-lateral suprasylvian sulcus Reviews Neuroscience 2, 704716.
Fries, P., Reynolds, J. H., Rorie, A., and Desimone, R. (2001). Mod-
(PMLS) of the cat visual cortex confirmed all three ulation of oscillatory neuronal synchronization by selective vi-
predictions. Cells synchronized their activity if they sual attention. Science 291, 1,5601,563.
responded to contours that are perceived as belong- Fries, P., Roelfsema, P. R., Engel, A. K., Knig, P., and Singer, W.
ing to the same surface (Castelo-Branco, Goebel, (1997). Synchronization of oscillatory response in visual cor-
Neuenschwander, and Singer, 2000) (see Figure 2C). tex correlates with perception in interocular rivalry. Proceed-
ings of the National Academy of Sciences of the United States of Amer-
Dynamic changes in synchronization could, thus, ica 94, 12,69912,704.
serve to encode in a context-dependent way the rela- Fries, P., Schrder, J.-H., Singer, W., and Engel, A. K. (2002). Os-
tions among the simultaneous responses to spatially cillatory neuronal synchronization in primary visual cortex as
a correlate of perceptual stimulus selection. Journal of Neuro- Roelfsema, P. R., Engel, A. K., Knig, P., and Singer, W. (1997).
science 22, 3,7393,754. Visuomotor integration is associated with zero time-lag syn-
Gray, C. M. (1999). The temporal correlation hypothesis of visual chronization among cortical areas. Nature 385, 157161.
feature integration: Still alive and well. Neuron 24, 3147. Singer, W. (1999). Neuronal synchrony: A versatile code for the
Gray, C. M., and Singer, W. (1989). Stimulus-specific neuronal os- definition of relations? Neuron 24, 4965.
cillations in orientation columns of cat visual cortex. Proceed- Tallon-Baudry, C., and Bertrand, O. (1999). Oscillatory gamma ac-
ings of the National Academy of Sciences of the United States of Amer- tivity in humans and its role in object representation. Trends
ica 86, 1,6981,702. in Cognitive Sciences 3, 151162.
Hebb, D. O. (1949). The Organization of Behavior. New York: John von der Malsburg, C. (1999). The what and why of binding: the
Wiley. modelers perspective. Neuron 24, 95104.
Leopold, D. A., and Logothetis, N. K. (1996). Activity changes in
early visual cortex reflect monkeys percepts during binocular
rivalry. Nature 379, 549553. Wolf Singer
P
PARALLEL DISTRIBUTED file, which we can later retrieve and reexamine. There
PROCESSING MODELS OF MEMORY are several problems with this view.
This article describes a class of computational models Memories are accessed by content. First of all, the nat-
that help us understand some of the most important ural way of accessing records in a computer is by their
characteristics of human memory. The computational address in the computer. However, what actually hap-
models are called parallel distributed processing (PDP) pens in human memory is that we access memories by
models because memories are stored and retrieved in their contents. Any description that uniquely identi-
a system consisting of a large number of simple com- fies a memory is likely to be sufficient for recall. Even
putational elements, all working at the same time and more interesting, each individual element of the de-
all contributing to the outcome. They are sometimes scription may be nearly useless by itself, if it applies
also called connectionist models because the knowledge to many memories; only the combination needs to be
that governs retrieval is stored in the strengths of the unique. Thus
connections among the elements.
He bet on sports. He played baseball.
The article begins with a common metaphor for
human memory, and shows why it fails to capture sev- is enough for many people to identify Pete Rose, even
eral key characteristics of memory that are captured through the cues about baseball and betting on sports
by the PDP approach. Then a brief statement of the would not generally be sufficient as cues individually,
general characteristics of PDP systems is given. Fol- since each matches too many memories.
lowing this, two specific models are presented that Memory fills in gaps. The computer-file metaphor
capture key characteristics of memory in slightly dif- also misses the fact that when we recall, we often fill
ferent ways. Strengths and weaknesses of the two ap- in information that could not have been part of the
proaches are considered, and a synthesis is presented. original record. Pieces of information that were not
The article ends with a brief discussion of the tech- part of the original experience intrude on our recol-
niques that have been developed for adjusting con- lections. Sometimes these intrusions are misleading,
nection strengths in PDP systems. but often enough they are in fact helpful reconstruc-
tions based on things we know about similar memo-
ries. For example, if we are told that someone has
Characteristics of Memory been shot by someone else from a distance of 300
A common metaphor for human memory might yards, we are likely to recall later that a rifle was used,
be called the computer file metaphor. On this met- even though this was not mentioned when we heard
aphor, we store a copy of an idea or experience in a about the original event.
507
508 PARALLEL DISTRIBUTED PROCESSING MODELS OF MEMORY
Memory generalizes over examples. A third crucial When we are remembering events, there is a unit for
characteristic of memory is that it allows us to form each event. Such models are called localist models. In
generalizations. If every apricot we see is orange, we the second type of model, cognitive units are not sep-
come to treat this as an inherent characteristic of apri- arately assigned to individual computing units. Rath-
cots. But if cars come in many different colors, we er, the representation of each cognitive unit is
come to treat the color as a freely varying property. thought of as a pattern of activation over an ensemble
So when we are asked to retrieve the common proper- of computing units. Alternative objects of thought are
ties of apricots, the color is a prominent element of represented by alternative patterns of activation. This
our recollection; but no color comes out when we are type of model is called a distributed model.
asked to retrieve the common properties of cars.
Proponents of the computer-file view of memory A Localist PDP Model of Memory
deal with these issues by adding special processes. Ac-
cess by content is done by laborious sequential search. McClelland (1981) presented a PDP model that
Reconstruction is done by applying inferential pro- illustrates the properties of access by content, filling
cesses to the retrieved record. Generalization occurs in of gaps, and generalization. The database for the
through a process of forming explicit records for the model is shown in Figure 1. The network is shown in
category (e.g., car or apricot). Figure 2.
In PDP systems, these three characteristics of The data base consists of descriptions of a group
memory are intrinsic to the operation of the memory of people who are members of two gangs, the Jets and
system. the Sharks. Each person has a name, and the list spec-
ifies the age, marital status, and education of each
person. Perusal of the list reveals that the Jets are, by
Characteristics of PDP Systems and large, younger and less well educated than the
A PDP system consists of a large number of neu- Sharks, and tend to be single rather than married.
ron-like computing elements called units. Each unit However, these tendencies are not absolute and, fur-
can take on an activation value between some maxi- thermore, there is no single Jet who has all of the
mum and minimum values, often 1 and 0. In such sys- properties that tend to be typical of Jets.
tems, the representation of something that we are The goal of the network is to allow retrieval of
currently thinking about is a pattern of activation over general and specific information about individuals in
the computing elements. Processing occurs by the the data base. The network consists of a unit for each
propagation of activation from one unit to another person (in the center of Figure 2) and a unit for each
via connections among the units. A connection may property (name, age, educational level, occupation,
be excitatory (positive-valued) or inhibitory (nega- gang) that a person can have. Units are grouped into
tive-valued). If the connection from one unit to an- pools by type as shown, so that all the name units are
other is excitatory, then the activation of the receiving in one pool, for instance. There is a bidirectional ex-
unit tends to increase whenever the sending unit is ac- citatory connection between each persons unit and
tive. If the connection is inhibitory, then the activa- the units for each of his properties; and there are
tion of the receiving unit tends to decrease. But note bidirectional inhibitory connections between units
that each unit may receive connections from many that can be thought of as incompatible alternatives.
other units. The actual change in activation, then, is Thus there is inhibition between the different
based on the net input, aggregated over all of the ex- occupation units, between the different age units,
citatory and inhibitory connections. and so on. There is also inhibition between the
In a system like this, the knowledge that governs different name units and between the units for differ-
processing is stored in the connections among the ent individuals.
units, for it is these connections that determine what In this network, units take on activation values be-
pattern will result from the presentation of an input. tween 1 and -0.2. The output is equal to the activa-
Learning occurs through adjustments of connection tion, unless the activation is less than 0; then there is
strengths. Memory storage is just a form of learning, no output. In the absence of input, the activations of
and also occurs by connection weight adjustment. all the units are set to a resting value of -0.1.
To make these ideas concrete, we now examine
two PDP models of memory. The models differ in a
crucial way. In the first, each individual computing el- Retrieval by Name
ement (henceforth called a unit) represents a separate Retrieval begins with the presentation of a probe,
cognitive unit, be it a feature (for example, the color in the form of externally supplied input to one or
of something), or a whole object, or the objects name. more of the property units. To retrieve the properties
PARALLEL DISTRIBUTED PROCESSING MODELS OF MEMORY 509
of Lance, for example, we need only turn on the name Figure 1

unit for Lance. The activation process is gradual and
builds up over time, eventually resulting in a stable
pattern that in this case represents the properties of
Lance. Activation spreads from the name unit to the
property units by way of the instance unit. Feedback
from activated properties tends to activate the in-
stance units for other individuals, but because of the
mutual inhibition, these activators are kept relatively
low.
Retrieval by Content
It should be clear how we can access an individual
by properties, as well as by name. As long as we pres-
ent a set of properties that uniquely matches a single
individual, retrieval of the rest of what is known of
properties of that individual is quite good. Other sim-
ilar individuals may become partially active, but the
correct person unit will dominate the person pool,
and the correct properties will be activated.
Filling in Gaps
Suppose that we delete the connection between
Lance and burglar. This creates a gap in the database.
However, the model can fill in this gap, in the follow-
ing way. As the other properties of Lance become ac-
tive, they in turn feed back activation to units for
other individuals similar to Lance. Because the in-
stance unit for Lance himself is not specifying any ac-
tivation for an occupation, the instance units for In summary, this simple model shows how re-
other, similar individuals conspire together to fill in trieval by content, filling in gaps, and generalization
the gap. In this case it turns out that there is a group are intrinsic to the process of retrieval in the PDP ap-
of individuals who are very similar to Lance and who proach to memory.
are all burglars. As a result, the network fills in burglar
for Lance as well. One may view this as an example
A Distributed PDP Model of Memory
of guilt by association. In this case, it so happens that
the model is correct in filling in burglar, but of course The second model to be considered is a distribut-
this kind of filling in is by no means guaranteed to be ed model. Many authors (e.g., Kohonen, 1977; An-
correct. Similarly, in human memory, our reconstruc- derson et al., 1977) have proposed variants of such
tions of past events often blend in the contents of models. The one shown in Figure 3 is from McClel-
other, similar events. land and Rumelhart (1985). The model is called dis-
tributed because there are no single units for individu-
als or for properties. Instead, the representation to be
Generalization stored is a distributed pattern over the entire set of
The model can be used to retrieve a generaliza- units. Similar memories are represented by similar
tion over a set of individuals who match a particular patterns, as before; but now each unit need not corre-
probe. For example, one can retrieve the typical spond to a specific feature or property, and there are
properties of Jets simply by turning on the Jet unit no separate units for the item as a whole. Again, the
and allowing the network to settle. The result is that knowledge is stored in the connections among the
the network activates 20s, junior high, and single units.
strongly. No name is strongly activated, and the three Methods for training such networks will be con-
occupations are all activated about equally, reflecting sidered in more detail below. Suffice it to note one
the fact that all three occur with equal frequency simple method, called the Hebbian method. Accord-
among the Jets. ing to this method, we increase the connection
510 PARALLEL DISTRIBUTED PROCESSING MODELS OF MEMORY
Figure 2
Some of the units and interconnections needed to represent the individuals shown in Figure 1. The units connected with double-
headed arrows are mutually excitatory. All the units within the same cloud are mutually inhibitory.
strength between two units if they are both active in There is a final important property of distributed
a particular pattern at the same time. memory models, and that is graceful degradation.
Distributed networks trained with this Hebbian The knowledge that governs the ability to reconstruct
learning rule exhibit many of the properties of local- each pattern is distributed throughout the network,
ist networks. They perform an operation, called pat- so if some of the connections are lost, it will not neces-
tern completion, that is similar to retrieval by content. sarily be catastrophic. In fact, the network can func-
In pattern completion, any part of the pattern can be tion quite well even when many of the units are de-
used as a cue to retrieve the rest of the pattern, al- stroyed, especially if it is relatively lightly loaded with
though there are limits to this that we will consider memories.
below. Because many memories are stored using the
same connection weights, they have a very strong ten-
dency to fill in gaps in one pattern with parts of other, A Synthesis
similar patterns. These models also generalize. When Each of the two models described above has some
similar patterns are stored, what is learned about one limitations. The localist model requires a special in-
pattern will tend to transfer to those parts that it has stance unit to be devoted to each memory trace; this
in common with the other. When a set of similar pat- is inefficient, especially when there is redundancy
terns is stored, what is common to all of them will across different memories in terms of what properties
build up as each example is learned; what is different tend to concur in the same memory. On the other
will cancel out. hand, the distributed model is limited because only
PARALLEL DISTRIBUTED PROCESSING MODELS OF MEMORY 511
a few distinct patterns can be stored in the direct con-

nections among the members of a set of units.
The best of both worlds can be obtained in a hy-
brid system, in which the various parts of the repre-
sentation of a memory are bound together by a set of
superordinate units, as in the localist model, but each
superordinate unit participates in the representation
of many different memories, as in the distributed
model.
Learning Rules for PDP Systems

Several of the learning rules for PDP systems are
reviewed in Rumelhart, Hinton, and McClelland
(1986). Here we consider two main classes, Hebbian
learning rules and error-correcting learning rules. We
have already mentioned the Hebbian learning rule,
which increases the strength of the connection be-
tween two units when both units are simultaneously
active. In a common variant, the strength of the con-
nection is decreased when one unit is active and the
other is inactive.
These Hebbian learning rules are limited in what
can be learned with them. Some of these limitations
are overcome by what are called error-correcting
learning rules. In such learning rules, the idea is that
the pattern to be learned is treated not only as input
but also as the target for learning. A pattern is pres-
ented, and the network is allowed to settle. Once it
has done so, the discrepancies between the resulting
Hertz, J., Krogh, A., and Palmer, R. (1990). Introduction to the theory
pattern and the input pattern are used to determine of neural computation. Redwood City, CA: Addison-Wesley.
what changes should be made in the connections. For Hinton, G. E., and Anderson J. A., eds. (1981). Parallel models of as-
example, if a unit is activated that should not be ac- sociative memory. Hillsdale, NJ: Erlbaum.
tive, the connection weights coming into that unit Kohonen, T. (1977). Associative memory: A system theoretical approach.
from other active units will be reduced. Several very New York: Springer-Verlag.
powerful learning procedures for adjusting connec- McClelland, J. L. (1981). Retrieving general and specific informa-
tion from stored knowledge of specifics. Paper presented at
tion weights that are based on the idea of reducing
the third annual meeting of the Cognitive Science Society.
the discrepancy between output and target have been Berkeley, CA.
developed in recent years. The best-known is the McClelland, J. L., and Rumelhart, D. E. (1985). Distributed memo-
back-propagation learning procedure (Rumelhart, ry and the representation of general and specific information.
Hinton, and Williams, 1986). Another important Journal of Experimental Psychology: General 114, 159188.
learning rule for PDP systems is the Boltzmann ma- Rumelhart, D. E., Hinton, G. E., and McClelland, J. L. (1986). A
chine learning rule (Ackley, Hinton, and Sejnowski, general framework for parallel distributed processing. In D.
E. Rumelhart, J. L. McClelland, and the PDP Research
1985). Both work well in training the hybrid systems
Group, eds., Parallel distributed processing: Explorations in the
described above. microstructures of cognition, Vol. 1. Cambridge, MA: MIT
Press.
See also: LOCALIZATION OF MEMORY TRACES;
Rumelhart, D. E., Hinton, G. E., and Williams, R. J. (1986).
NEURAL COMPUTATION; RECONSTRUCTIVE
Learning internal representations by error propagation.
MEMORY
In D. E. Rumelhart, J. L. McClelland, and the PDP Research
Bibliography Group, eds., Parallel distributed processing: Explorations in the
microstructures of cognition, Vol. 1. Cambridge, MA: MIT
Ackley, D. H., Hinton, G. E., and Sejnowski, T. J. (1985). A learn-
Press.
ing algorithm for Boltzmann machines. Cognitive Science 9,
147169. Rumelhart, D. E., McClelland, J. L., and the PDP Research Group.
Anderson, J. A., Silverstein, J. W., Ritz, S. A., and Jones, R. S. (1986). Parallel distributed processing: Explorations in the micro-
(1977). Distinctive features, categorical perception, and prob- structure of cognition, 2 vols. Cambridge, MA: MIT Press.
ability learning: Some applications of a neural model. Psycho-
logical Review 84, 413451. James L. McClelland
512 PASSIVE (INHIBITORY) AVOIDANCE, FEAR LEARNING
PASSIVE (INHIBITORY) AVOIDANCE, shock compartment, which at this point is not electri-
FEAR LEARNING fied, is measured. This measure (retention latency) is
used to infer the animals memory for the fearful ex-
Fear learning is the process of gathering information periencethe longer the retention latency, the better
about the internal and external environment in situa- the memory. A long retention latency indicates a sig-
tions that evoke fear. Fear learning is the first step to- nificant modification in the animals behavior, as it
ward creating memories for fearful events (fear mem- contrasts with the animals low initial entrance latency
ories), which are robust and represent a long-lasting displayed before the training.
record of the acquired information that is capable of
PA/IA testing has several features that make it a
modifying behavior when retrieved. Like other forms
valuable tool for investigating brain systems involved
of learning, fear learning (fear memory acquisition)
in memory acquisition, consolidation, and retrieval.
is followed first by memory consolidation, a period of
It is simple to administer and thus easy to replicate
time when memories are still labile and can be modu-
across laboratories, yet it is complex enough to en-
lated (enhanced or impaired), and then by memory
gage multiple brain regions and neurotransmitter
storage (McGaugh, 2000).
systems (Ambrogi Lorenzini et al., 1999; Izquierdo,
Because memories cannot be directly observed Medina, and Barros, 1999; McGaugh, Ferry, Vazdar-
and assayed, their existence is, by necessity, inferred janova, and Roozendaal, 2000). In addition, PA/IA is
from changes in behavior following an experience ethologically relevant in that it promotes fast learn-
(Cahill, McGaugh, and Weinberger, 2001). Thus, to ing, usually after one trial (in nature, an animal may
study learning and memory, subjects are interrogated not survive to pass on its genes if it does not learn to
by observing their performance in carefully designed recognize and avoid a predator after a single encoun-
behavioral tasks. In the study of fear learning and ter). One-trial learning provides a clear time stamp of
memory, the most widely used tasks are passive (in- when learning occurred. In combination with post-
hibitory) avoidance and fear conditioning. This entry training manipulations, which selectively affect mem-
will focus on the former task and will discuss how its ory consolidation but not acquisition or retrieval (Mc-
use has advanced the understanding of brain struc- Gaugh, 1989), PA/IA training provides researchers
tures and neuromodulatory systems involved in con- the ability to study the brain mechanisms involved in
solidating and storing memories for fearful events. consolidation of explicit-declarative memory for fear-
Although this presentation is based on findings ob- ful events.
tained in rats and mice, passive avoidance has also
been instrumental in understanding memory pro-
cesses in chicks (Rose, 2000) and humans (Cahill and What Kind of Memory Does the Retention
McGaugh, 1998). Latency Measure Represent?
Arguably, retention latency measured in PA/IA
may reflect only a procedural memory (i.e., entering
Passive (Inhibitory) Avoidance a place is aversive) (Wilensky, Schafe, and LeDoux,
Memory in a passive avoidance task is inferred 2000). However, during PA/IA animals could acquire
from the delay of a response that was readily made be- both explicit-declarative memories, such as those ac-
fore the training. Because delaying a response is an quired during Pavlovian contextual fear conditioning
active process, some investigators refer to the task as (I got shocked in this place), and implicit-
inhibitory, rather than passive, avoidance (Cahill and procedural memories that reflect a response-
McGaugh, 1998; Izquierdo, Medina, and Barros, reinforcement contingency. Because acquiring proce-
1999). Passive (inhibitory) avoidance, or PA/IA, ex- dural memories takes many more trials than acquir-
periments are conducted in a two-compartment being explicit-declarative memories (Packard and Mc-
havioral apparatus, where one compartment is de- Gaugh, 1996), measuring retention latencies after
signed to be naturally preferred by the animal (see one-trial PA/IA most likely reflects explicit-
Figure 1). During training, the animal is placed in the declarative fear memories. Moreover, when the two
less-preferred compartment and the latency to enter types of memory were experimentally pitted against
the preferred compartment is noted. Upon complete- each other in a modified avoidance task, animals un-
ly entering the preferred compartment, the animal ambiguously expressed explicit-declarative fear
receives one or more inescapable foot shocks of a memories. In this test, rats were trained on an active
specified intensity and duration. The information the avoidance task by being repeatedly placed in the pre-
animal gathers during the training is fear learning. At ferred compartment, each time shocked until they en-
a retention test, conducted hours, days, or months tered the less-preferred (light) compartment, and
later, the animal is returned to the previously less- then tested by being placed in the light compartment.
preferred compartment and the latency to enter the By showing high retention latencies, the rats indicat-
PASSIVE (INHIBITORY) AVOIDANCE, FEAR LEARNING 513
ed that they had learned where they had received foot Figure 1
shocks (explicit memory), rather than that they had
to perform a response (procedural memory) (Parent,
West, and McGaugh, 1994). Additionally, latencies to
enter a compartment where foot shock was experi-
enced can be successfully used to assess memory
for contextual fear conditioning, a task in which
receiving foot shocks is not contingent on any one
particular response (Vazdarjanova and McGaugh,
1998).
Factors Involved in Consolidation of

Memory for PA/IA
Stress hormones (adrenaline [or epinephrine]
and glucocorticoids), when injected after training at
the time when they are normally released by the adre-
nal glands following an emotional experience, can
modulate the memory strength for one-trial PA/IA
training. This modulation is dose-dependent (low to
medium levels of stress hormones are memory en-
hancing, whereas high levels are memory impairing)
and time-dependent (the effects are most pro-
nounced when the treatments are administered im- A schematic representation of two versions of the passive
mediately after training). Consistent with the idea (inhibitory) avoidance task (PA/IA) for rats and mice. A) In a
that these hormones are endogenously released as a step-through PA/IA task, the animal is placed in the light
result of an emotional experience, memory is im- compartment of the two-compartment apparatus. Being
paired by removing the adrenals, blocking peripheral nocturnal animals, rats and mice naturally prefer the dark
compartment, so they enter it quickly. The sliding door between
-adrenoceptors, or blocking the synthesis of gluco-
the two compartments is closed and the animal receives one or
corticoids. Although their effects on memory consoli-
more inescapable footshocks. At the retention test, the animal is
dation are similar, adrenaline and glucocorticoids placed in the light compartment (the sliding door is open) and
exert their effects via different pathways. Glucocorti- the latency to enter the dark compartment is recorded and used
coids are lipophilic and readily cross the blood-brain to infer the strength of the memory for the fearful experience.
barrier, thus directly affecting memory consolidation Higher footshock intensities, or more footshocks, lead to longer
in several brain regions by activating glucocorticoid retention latencies. B) In the step-down PA/IA task, the
receptors. When glucocorticoid receptor agonists are initially less-preferred compartment is a small platform raised
administered directly into the basolateral amygdala, above a grid floor (rats and mice are afraid of heights). When
hippocampus, or nucleus of the solitary tract (NTS), the animal steps down with all four paws, it receives one or
they enhance memory for PA/IA. The effects of gluco- more footshocks. The retention latency is the latency to step
down from the platform to the grid floor at the time of testing.
corticoids on memory appear to be mediated by the
basolateral amygdala, as lesions of this region or inac-
tivation of the amygdalas -adrenoceptors block the
memory-modulatory effects of peripheral and intra- cortex (Izquierdo, Medina, and Barros, 1999), nora-
hippocampal or intra-NTS glucocorticoid treatments drenergic activation of the basolateral amygdala is of
(Roozendaal, 2000). primary importance. Stress induces the release of
noradrenaline in this region, and activation of the
Unlike glucocorticoids, adrenaline does not read-
amygdalas -adrenoceptors reverses the memory-
ily pass the blood-brain barrier and its effects on
memory consolidation are mediated by - impairing effects of GABAergic (gamma-
adrenoceptors on the vagus nerve. Activation of the aminobutyric acid) and opioid agonists, while block-
vagus nerve stimulates noradrenergic neurons in the ing of the -adrenoceptors blocks the memory en-
NTS directly and noradrenergic neurons in the locus hancement produced by GABAergic and opioid an-
coeruleus indirectly. Thus, peripherally released ad- tagonists. The effectiveness of the basolateral
renaline leads to centrally released noradrenaline. Al- amygdalas -adrenoceptors themselves depends on
though noradrenaline can modulate memory consoli- the activation of glucocorticoid receptors in this re-
dation for PA/IA by acting directly on neurons in the gion (see Figure 2; also McGaugh, Ferry, Vazdar-
basolateral amygdala, hippocampus, and entorhinal janova, and Roozendaal, 2000).
514 PASSIVE (INHIBITORY) AVOIDANCE, FEAR LEARNING
Figure 2
A summary diagram of how stress hormones and other memory modulatory treatments interact in the basolateral amygdala to
influence memory consolidation of emotional events, such as PA/IA, in other brain regions. During stress, the adrenal medulla
releases adrenaline and the adrenal cortex releases corticosterone (cortisol in humans). Corticosterone freely passes the blood-brain
barrier and acts on multiple brain regions. In the basolateral amygdala it modulates processes occurring downstream of activation of
the -adrenoceptors. Adrenaline, which does not pass the blood-brain barrier, activates the vagus nerve by acting on peripheral -
adrenoceptors. Vagus nerve afferents activate directly noradrenergic neurons in the nucleus of the solitary tract to release
noradrenaline in the basolateral amygdala. Activation of -adrenoceptors in the basolateral amygdala enhances memory by
strengthening memory consolidation processes in target regions of the stria terminalis, and it supersedes the influences on memory
of treatments acting on the GABAergic and opioid receptors in the basolateral amygdala.
Adrenergic, glucocorticoid, and cholinergic amygdala are reversed by lesions of the stria termi-
memory modulatory treatments produced by either nalis, one of the two major input-output pathway of
systemic treatments or direct manipulations of the the amygdala. Thus, the amygdala appears to modu-
PASSIVE (INHIBITORY) AVOIDANCE, FEAR LEARNING 515
late memory storage in target areas of the stria termi- a distributed neural network across several brain re-
nalis (McGaugh, Ferry, Vazdarjanova, and Roozend- gions.
aal, 2000).
Memory for PA/IA is also modulated by glucose Temporal Pattern of Memory Consolidation
(the release of which is enhanced by adrenaline), va- for PA/IA
sopressin, and substance P, as well as cholinergic,
Studies of functional inactivation of selective
GABA-ergic, dopaminergic, serotonergic, and opioid
brain regions by anesthetics that block sodium chan-
agents (Gold, 1995; Koob et al., 1991; McGaugh,
nels, or by increasing inhibition through GABA-A re-
1989; Izquierdo, Medina, and Barros, 1999; Zhang,
ceptors, have shown that memory consolidation for
Berbos, and Wiley, 1996; Riekkinen, Kuitunen, and
PA/IA occurs over a long period of time and involves
Riekkinen, 1995).
sequentially a multitude of brain regions. Functional
All neuromodulatory treatments discussed thus integrity of the hippocampus, medial septum, baso-
far appear to affect memory storage not only for lateral amygdala, nucleus of the solitary tract, and
PA/IA but for emotional experiences in general, as several cortical regions (insular, anterior prefrontal
similar findings have been shown in other emotional- [Fr2], and posterior cingulated) is required during an
ly motivated tasks. Predictions based on the reported early phase of memory consolidation (immediately
findings in rats and mice have been supported by re- and up to 1.5 hours after training), whereas that of
sults obtained in human subjects (Cahill and Mc- the entorhinal and parietal cortecies is required dur-
Gaugh, 1998). ing a later phase (0.5 to 3 hours or 1 to 3 hours, re-
spectively). Finally, the functional integrity of brain
Researchers debate whether the described mech- regions that release neuromodulators in the hippo-
anisms of memory consolidation for PA/IA, notably campus, amygdala, and cortexsuch as the nucleus
the role of the basolateral amygdala in memory stor- basalis, substantia nigra, and parabrachial nucleus
age, are the same as those for fear conditioning (Fan- is required for a couple of days after training. The du-
selow and LeDoux, 1999; Cahill, Weinberger, Roo- ration of memory consolidation appears to depend
zendaal, and McGaugh, 1999; Maren, 1999; on the intensity of the trainingthe higher the foot-
Vazdarjanova, 2000). Although the basolateral amyg- shock stimulus, the quicker the consolidation (Am-
dala plays a modulatory role in memory consolidaton brogi Lorenzini et al., 1999; Bermdez-Rattoni, In-
of PA/IA, it is probably also the site of memory storage troini-Collison, and McGaugh, 1991; Izquierdo, Quil-
for fear conditioning. This conclusion is based mainly lfeldt, and Medina, 1997; Mello e Souza, Vianna, and
on the observation that pretraining or pretest lesions Izquierdo, 2000).
or inactivation (both general and NMDA-receptor
specific) of this brain region lead to decreased freez-
ing behavior, a typical measure used to assess memo- Conceptual Advances Facilitated by the
ry for fear conditioning. In addition, the basolateral Use of One-Trial PA/IA
amygdala develops LTP-like (long-term potentiation) The use of PA/IA and timed manipulations after
plasticity as a result of auditory fear conditioning. training has helped identify the neuromodulatory
However, both lines of evidence are also consistent systems and their interactions in brain regions, nota-
with a modulatory role of the basolateral amygdala in bly the basolateral amygdala, that modulate the con-
fear conditioning. Thus, if disruption of activity in solidation of explicit memories for fearful events.
this region leads to decreased strength of memories PA/IA studies also revealed that memory consolida-
for fear conditioning, animals will be less likely to dis- tion is a dynamic process that occurs over hours to
play freezing, a behavior that indicates high levels of days and depends sequentially on networks of neu-
fear, while other measures of memory may be still evi- rons distributed among several brain regions. The
dent. Consistent with this hypothesis are findings that shift over time in the dependence of memory consoli-
following contextual fear conditioning, rats with com- dation to cortical regions also indicates that, consis-
plete lesions of the basolateral amygdala do not show tent with early predictions (Gerard, 1961), the cortex
freezing, but do display training-induced avoidance appears to be the ultimate repository for explicit-
of the place where they had received foot shocks (Vaz- declarative long-term emotional memories. Finally,
darjanova and McGaugh, 1998). Furthermore, physi- because PA/IA is useful in assessing both short-term
ological plasticity following fear conditioning is not memory (developing within seconds and lasting min-
unique to the basolateral amygdala but is present in utes to hours) and long-term memory (developing
multiple brain regions (Mcintosh and Gonzalez- over time and lasting hours to years) (McGaugh,
Lima, 1998; Poremba and Gabriel, 2001), suggesting 2000) in the same subjects, it enabled the first demon-
that fear memories are likely encoded and stored in stration in mammals that the processes underlying
516 PAVLOV, IVAN
these two types of memory are independent. Fifteen McGaugh, J. L. (1989). Dissociating learning and performance:
different pharmacological manipulations that affect Drug and hormone enhancement of memory storage. Brain
Research Bulletin 23, 339345.
specific receptors or molecular cascades in the hippo- (2000). Neuroscience: Memorya century of consolida-
campus, entorhinal, and parietal cortex impair short- tion. Science 287 (5,451), 248251.
term memory without affecting long-term memory McGaugh, J. L., Ferry, B., Vazdarjanova, A., and Roozendaal, B.
(Izquierdo, Medina, and Barros, 1999). Thus, the de- (2000). Amygdala: Role in modulation of memory storage. In
velopment of long-term memory does not require in- John P. Aggleton, ed., The Amygdala: A functional analysis, 2nd
edition. London: Oxford University Press.
tact short-term memory. The conceptual advances af- Mcintosh, A., and Gonzalez-Lima, F. (1998). Large-scale functional
forded by the use of PA/IA underscore the importance connectivity in associative learning: Interactions of the rat au-
of this paradigm as a model system for studying emo- ditory, visual, and limbic systems. Journal of Neurophysiology
tional learning and memory processes. 80, 3,1483,162.
Mello e Souza, T., Vianna, M. R. M., and Izquierdo, I. (2000). In-
See also: ACTIVE AND PASSIVE AVOIDANCE volvement of the medial precentral prefrontal cortex in mem-
LEARNING: BEHAVORIAL PHENOMENA; NEURAL ory consolidation for inhibitory avoidance learning in rats.
SUBSTRATES OF AVOIDANCE LEARNING; NEURAL Pharmacology, Biochemistry, and Behavior 66 (3), 615622.
SUBSTRATES OF CLASSICAL CONDITIONING; Packard, M. G., and McGaugh, J. L. (1996). Inactivation of hippo-
campus or caudate nucleus with lidocaine differentially affects
expression of place and response learning. Neurobiology of
Bibliography
Parent, M. B., West, M., and McGaugh, J. L. (1994). Memory of
Ambrogi Lorenzini, C., Baldi, E., Bucherelli, C., Sacchetti, B., and rats with amygdala lesions induced thirty days after footshock-
Tassoni, G. (1999). Neural topography and chronology of motivated escape training reflects degree of original training.
memory consolidation: A review of functional inactivation Behavioral Neuroscience 108 (6), 1,0801,087.
findings. Neurobiology of Learning and Memory 71 (1), 118. Poremba, A., and Gabriel, M. (2001). Amygdala efferents initiate
Bermdez-Rattoni, F., Introini-Collison, I. B., and McGaugh, J. L. auditory thalamic discriminative training-induced neuronal
(1991). Reversible inactivation of the insular cortex by tetro- activity. Journal of Neuroscience 21 (1), 270278.
dotoxin produces retrograde and anterograde amnesia for in- Riekkinen Jr., P., Kuitunen, J., and Riekkinen, M. (1995). Effects
hibitory avoidance and spatial learning. Proceedings of the Na- of scopolamine infusions into the anterior and posterior cin-
tional Academy of Sciences of the United States of America 88, gulate on passive avoidance and water maze navigation. Brain
5,3795,382. Research 685 (12), 4654.
Cahill, L., and McGaugh, J. L. (1998). Mechanisms of emotional Roozendaal, B. (2000). Glucocorticoids and the regulation of mem-
arousal and lasting declarative memory. Trends in Neuroscience ory consolidation. Psychoneuroendocrinology 25 (3), 213238.
21 (7), 294299. Rose, S. P. (2000). Gods organism? The chick as a model system
Cahill, L., McGaugh, J. L., Weinberger, N. M. (2001). The neuro- for memory studies. Learning and Memory 7 (1), 117.
biology of learning and memory: Some reminders to remem- Vazdarjanova, A. (2000). Does the basolateral amygdala store
ber. Trends in Neurosciences 24 (10), 578581. memories for emotional events? Trends in Neurosciences 23 (8),
Cahill, L., Weinberger, N. M., Roozendaal, B., and McGaugh, J. L. 345.
(1999). Is the amygdala a locus of conditioned fear? Some Vazdarjanova, A., and McGaugh, J. L. (1998). Basolateral amygda-
questions and caveats. Neuron 23, 227228. la is not critical for cognitive memory of contextual fear condi-
Fanselow, M., and LeDoux, J. (1999). Why we think plasticity un- tioning. Proceedings of the National Academy of Sciences of the
derlying Pavlovian fear conditioning occurs in the basolateral United States of America 95, 15,003 15,007.
amygdala. Neuron 23, 229232. Wilensky, A. E., Schafe, G. E., and LeDoux, J. E. (2000). The amyg-
Gerard, R. W. (1961). The fixation of experience. In A. Fessard, R. dala modulates memory consolidation of fear-motivated in-
W. Gerard, and J. Konorski, eds., Brain mechanisms and learn- hibitory avoidance learning but not classical fear condition-
ing: A symposium. Springfield, IL: Thomas. ing. Journal of Neuroscience 20 (18), 7,0597,066.
Gold, P. E. (1995). Modulation of emotional and non-emotional Zhang, Z. J., Berbos, T. G., and Wiley, R. G. (1996). Loss of nucleus
memories: Same pharmacological systems, different basalis magnocellularis, but not septal, cholinergic neurons
neuroanatomical systems. In J. L. McGaugh, N. M. Weinber- correlates with passive avoidance impairment in rats treated
ger, and G. Lynch, eds., Brain and memory: Modulation and me- with 192-saporin. Neuroscience Letters 203 (3), 214218.
diation of neuroplasticity. New York: Oxford University Press.
Izquierdo, I., Medina, J. H., and Barros, D. M. (1999). Separate Almira Vazdarjanova
mechanisms for short- and long-term memory. Behavioural
Brain Research 103 (1), 111.
Izquierdo, I., Quillfeldt, J. A., and Medina, J. H. (1997). Sequential
role of hippocampus and amygdala, entorhinal cortex and
parietal cortex in formation and retrieval of memory for in-
hibitory avoidance in rats. European Journal of Neuroscience 9 PAVLOV, IVAN (18491936)
(4), 786793.
Koob, G. F. et al. (1991). Vasopressin and learning: Peripheral and The Russian physiologist Ivan Petrovich Pavlov is best
central mechanisms. In R. Frederickson, J. L. McGaugh, and known as the discoverer of the conditioned reflex.
D. L. Felten, eds., Peripheral signaling of the brain: Role in neu- The life and work of this Nobel laureate is encapsulat-
ral-immune interactions, learning and memory. Toronto: Hogrefe ed in his motto, Observation and observation! His
and Huber.
Maren, S. (1999). Long-term potentiation in the amygdala: A work had an enormous influence on psychology in
mechanism for emotional learning and memory. Trends in general and on the theory of learning and memory in
Neurosciences 22 (12), 561567. particular.
PAVLOV, IVAN 517
Early Life and Work

Pavlov evolved from a religious to a scientific
framework. His ancestry tracex to an illiterate eigh-
teenth-century serf known only by his first name,
Pavel (Anokhin, 1949). Pavels son gained emancipa-
tion and became a member of the clerical estate. Dur-
ing the next two generations, the family head rose
through the religious hierarchy from church sexton
to deacon. The deacon was able to provide a seminary
education for his sons, who became ordained priests.
Pavlovs father, Petr Dmitrievich, the youngest of
these sons, was a priest in Riazan, an ancient town
about 120 miles south of Moscow.
Petr Pavlov, as Windholz notes (1991), had a li-

brary of his own and transmitted to his son a love of
knowledge. His advice to his children was that any
book should be read at least twice in order not to miss
anything important and to recall it more accurately;
Ivan took this advice to heart throughout his scientific
career. But he could not accept his fathers position
on some fundamental issues, including religious ones.
I had heated arguments with my father, Ivan wrote Ivan Pavlov (Hulton-Deutsch Collection/Corbis-Bettmann)
later, which, because of my position, led to strong
words and ended in serious disagreements (Pavlov,
1952, p. 447). Windholz suggests that the main dence concerning those constructs be stated in an ob-
sources of discord were Pavlovs loss of faith by the jective or scientifically communicable way (Furedy,
time he entered the seminary in Riazan in 1864; he Heslegrave, and Scher, 1984, p. 182). Thus there are
left before completing his studies there. no Watsonian or Skinnerian restrictions on the nature
When, in 1870, Pavlov enrolled in natural sci- of theoretical concepts; the only restrictions pertain
ences in the Faculty of Physics and Mathematics at St. to the mode of evaluation of the inferences about
Petersburg University, his father refused to support those concepts.
him financially. Yet his family unwittingly supported Like Sechenov, Pavlov was a revolutionary think-
this important change by providing an intellectual er who was able to finesse his encounters with authori-
oasis during Pavlovs seminary period. Windholz
ty without surrendering on vital points or courting
notes that, unlike most seminary students, Pavlov
personal ruin. The ultimate testament of his political
lived in his parents home, which gave him consider-
skill was his ability to run an active laboratory during
able freedom to pursue his own intellectual interests
Stalins regime. By this time, he had achieved the pin-
by being able to avoid the discipline imposed upon
nacle of scientific status for his work on the physiology
seminarians living in the dormitory and enjoy[ing]
of digestion (the Nobel Prize, 1904), and had turned
uninhibited reading in a small room over the family
to the study of the psychic salivary (digestive) reflex
living quarters (Windholz, 1991, p. 58).
(i.e., Pavlovian conditioning). Yet Pavlov also showed
An important book that the young seminarian courage in challenging authority and asserting his in-
may have read in that small room was Ivan Sechenovs tellectual integrity. Horsley Gantt, a young American
Reflexes of the Brain (1866). Sechenov had had a brush scientist who was a visiting member of Pavlovs labora-
with the government censor, who forced him to tory in the 1920s, recounts that when, in 1926, the
change his original titleAn Attempt to Place Psychical minister of education and head of the department
Processes on a Physiological Basisto a less provocative that supported Pavlovs laboratory visited the site,
one (Koshotiants, 1945). The idea that the study of Pavlov refused even to meet with him, much less show
behavior can yield an objective account of subjective him the laboratory. His stated reason was that he dis-
processes is nevertheless evident in Sechenovs work. approved of the ministers recent book, The ABC of
This assumption was the basis Pavlovs method in psy- Communism (Gantt, 1991, p. 68). The survival and sub-
chology: the objective study of mental processes. In sequent success of Pavlovs laboratory in the wake of
this method, the only restriction on the levels of ex- this political insolence suggests that he knew just how
planatory constructs that are used [is] that the evi- far he could go in challenging authority.
518 PAVLOV, IVAN
One mark of Pavlovs fame is that his name is at- quires some three months of adaptation to the hold-
tached to key psychological ideas. The term Pavlov- ing harness before the food reliably (i.e., uncondi-
ian response usually refers to automatic, nonreflec- tionally) elicits salivation rather than competing
tive, reflexlike reactions. (The brainwashing struggling behavior.
activities of the Chinese and the North Koreans in the
The first set of systematic and extensive experi-
1950s were considered to be Pavlovian. Orthodox
mental explorations of the phenomenon of classical
Marxist writings in psychology during the same peri-
conditioning employed the human eyelid condition-
od in the Soviet Union paid lip service to Pavlovian
ing preparation. In this arrangement, an air puff to
principles).
the eye (or, in later versions, an infra-orbital shock)
The Pavlovian notion of differing strengths in the served as the US, a blink as the UR and CR, and (usu-
nervous system, a concept that arose from Pavlovs ally) a tone as the CS (conditional stimulus). In the
study of experimental neurosis occurring from the 1950s and 1960s many of the most rigorous of these
breakdown of conditional discrimination in dogs, is studies were reported in Journal of Experimental Psy-
of general interest for personality theory. In the West, chology, the leading experimental psychological jour-
Hans Eysencks work on extroversion and introver- nal of the period.
sion drew heavily on these Pavlovian concepts. The Most of the authors of these eyelid conditioning
theoretical ideas and concepts are complex and go reports subscribed to the then dominant stimulus-
well beyond the observable data that Watson (1913) response (S-R) approach to learning espoused by the-
argued were the only proper subject matter of psy- orists like C. L. Hull (Hull, 1943) and K. W. Spence
chology. (Spence, 1956). According to the S-R approach, only
stimulus-response associations were learned. This po-
sition contrasted with the position of the cognitive,
Classical Conditioning stimulus-stimulus (S-S) theorists, led by R. C. Tolman,
Pavlovs most important conceptual contribution, who argued that learning also involved both the ac-
however, was in experimental psychology. His most quisition of sign-signicate relationships such as cog-
influential concept is that of the conditional reflex, nitive maps as well as the teleological explanatory
which Pavlov examined in an experimental prepara- concept of purpose (e.g., Tolman, 1932).
tion now known as classical or Pavlovian condi-
tioning. In studying this form of learning, he pres- Applied to the phenomenon of classical condi-
ented canine subjects with food (the unconditional tioningwidely considered to be the simplest form of
stimulus or US) that unconditionally elicited saliva- learningthe S-R approach implies that condition-
tion (the unconditional response or UR) and with a ing is the learning of the association between a stimu-
bell (the conditional stimulus or CS) that originally lus (the CS) and a specific response (the CR). Accord-
did not elicit salivation. The CS, conditional on being ingly, most of these eyelid-conditioning experi-
paired repeatedly with the US, came to elicit saliva- menters (who produced a considerable volume of ex-
tiona response termed the conditional reflex or re- perimental evidence on the Pavlovian conditioning
sponse (CR). In Pavlovs view, the classical condition- phenomenon) were sensitive to independent vari-
ing preparation was contrary to subjective ables like the temporal contiguity between CS and US
psychology [which] held that saliva flowed because and to the dependent-variable distinction between re-
the dog wished to receive a choice bit of meat (Grig- sponses that were CRs and those that were not. The
orian, 1974, p. 433). This sort of cognitive and pur- contiguity independent variable was the period be-
posive formulation is akin to interpretations of the tween CS and US onsets, or the CS-US interval. The
Skinner-box bar press, in which the rat is said to press optimal CS-US interval was found to be slightly less
a bar in order to get the food. than a half second. In the eyelid-conditioning prepa-
ration, moreover, CS-US intervals of two seconds or
While the classical conditioning preparation had more produced no conditioning at all (i.e., no in-
been widely known in experimental psychology since crease in CS-elicited blinks as a function of paired
the early twenties, no body of systematic reports of CS-US trials), even though it could hardly be argued
Pavlovian dog-salivary conditioning was published in that the undergraduate subjects did not learn the
the journal literature of experimental psychology. (cognitive) relation between the tone and the puff
Pavlovs methods were based on single-case studies, that followed two seconds after the tone. Concerning
and the dependent-variable differences were report- the distinction between CRs and non-CRs, the CS-
ed quasi-anecdotally rather than with specified reli- elicited response latency measurement was also cru-
ability in terms of the rules of statistical inference. cial: shorter-latency blinks (occurring within 150 mil-
Furthermore, his preparation is extremely difficult to liseconds following CS onset) were found to decrease
work with. For example, the typical canine subject re- rather than increase as a function of repeated CS-US
PAVLOV, IVAN 519
pairings. Hence, as a routine matter of scoring the posive terms. According to R. A. Rescorla (the S-S
data, only longer-latency blinks (the frequency of oc- contingency approachs most eminent current ex-
currence of which did increase as a function of CS-US ponent) Pavlovian conditioning is now not just inter-
pairings) were classified as CRs. preted but described [authors emphasis] as the learn-
ing of relations among events so as to allow the
organism to represent its environment (Rescorla,
Influence on Other Researchers 1988, p. 151). It is interesting to compare this Tol-
In addition to its role in generating a body of em- man-like cognitive (relations between events) and
pirical knowledge about human classical eyelid condi- purposive (so as to allow) formulation with the
tioning in experimental psychology, Pavlovs CR con- subjective psychology account cited abovewhich
cept was also influential as a theoretical construct. Pavlov opposed and which is contrary to the mecha-
The originator of American S-R behaviorism, J. B. nistic accounts of conditioning proposed by S-R theo-
Watson (e.g., Watson, 1913) attempted to account for rists.
all psychological learning phenomena in terms of the The 1970s and 1980s also saw the near abandon-
Pavlovian CR, being attracted by the observable ment of the eyelid preparation as a means of studying
feature of this response, in contrast with unobserv- the Pavlovian (response) conditioning phenomenon.
able mental events. As Watson tackled more com- While a more reliable form of the preparationthe
plex psychological functions such as thinking, the rabbit nictitating membrane preparationwas devel-
observable status of his explanatory constructs be- oped in the 1960s, it has been used mainly as a tech-
came more dubious. For example, his explanation of nique to study the effects of physiological manipula-
thinking was that it was due to very slight movements tions rather than the phenomenon of conditioning
of the tongue, but when experimenters looked and itself. Consideration of the CS-US interval has essen-
could not observe these tongue movements, Watsoni- tially disappeared: indeed, in the currently dominant
an behaviorists retorted that the tongue movements Rescorla-Wagner (1972) model of Pavlovian condi-
were too small to be measured. tioning, the CS-US interval does not appear as a pa-
rameter. Considering the fact that, as indicated
The theoretical Pavlovian conditional response
above, the CS-US interval is crucial in preparations
concept reappeared in the guise of the fractional an-
such as eyelid conditioning, this omission may seem
ticipatory goal response in the theorizing of the S-R
strange. However, most of the evidence and experi-
learning theorist Hull (1935) and Spence (1956). This
mental work of current cognitive S-S Pavlovian condi-
hypothetical response was said to be learned through
tioners is based on preparations where conditioning
classical conditioning. It was invoked by S-R theorists
is not measured directly through assessing the CR but
from the 1930s until the 1960s to account, in S-R
indirectly through assessing the effect of the CS on
terms, for experimental results (mainly from latent
some instrumental indicator behavior, as in the con-
learning studies) that Tolman and his S-S learning
ditioned emotional response preparation. Such indi-
followers put forward as evidence for cognitive, sign-
cator-behavior preparations are more likely to sug-
significate, S-S learning. Like the earlier Watsonian
gest that Pavlovian conditioning, in line with what
slight-tongue-movements construct, the testable or
Pavlov characterized as subjective psychology, is
observable status of the fractional anticipatory goal
the learning of relations between events (i.e., cogni-
response was doubtful. It was even described by one
tive, S-S learning of CS/US contingency) rather than
Hull-Spence S-R proponent as incorporeal (Moltz,
the learning of responding (i.e., the CR) to the CS.
1957). A few years before this metaphysical qualifier
Still, the Pavlovian emphasis on using behavior as the
was issued for this hypothetical Pavlovian CR, propo-
objectively observed dependent variable has been re-
nents of Tolmans S-S expectancy-theory position
tained by both the S-R experimentalists and their
presciently complained that the construct was a deus
cognitive, S-S oriented successors.
ex machina used by S-R theorists to smuggle the
concept of cognition into their purportedly S-R ac-
counts (Meehl and MacCorquodale, 1953). The Pavlovian Society
The decline of the S-R approach in psychology Perhaps Pavlovs most important contribution to
and the paradigm shift to the cognitive, S-S ap- psychology was to encourage interdisciplinary re-
proach (Segal and Lachman, 1972) resulted in a radi- search on the organism as an integrated whole. This
cal change in psychologists theoretical perspective. legacy has been sustained by the Pavlovian Society, an
Consistent with the continuing cognitive revolution international organization founded in the United
that began in the late 1960s, Pavlovian conditioning, States in 1955 by Horsley Gantt, who, as noted above,
both as a phenomenon and as an explanatory con- worked in Pavlovs laboratory. The official journal of
cept, is conceived in Tolman-like cognitive and pur- this society is titled Integrative Behavioral and Physiolog-
520 PERCEPTION
ical Science. The focus interdisciplinary research on Steinmetz, J. (2001). An editorial: A continuing commitment to the
the integrated organism is central for the Pavlovian interdisciplinary research on the integrated organism. Inte-
grative Physiological and Behavioral Science 36, 34.
Society (Steinmetz, 2001). It is distinct from other Tolman, E. C. (1932). Purposive behavior in animals and men. New
contemporary scientific and scientific-professional York: Appleton-Century.
societies in its thoroughly apolitical and unusual oess- Watson, J. B. (1913). Psychology as the behaviorist views it. Psycho-
penn to discussion and debate (Furedy, 2001). The logical Review 20, 158177.
societys motto is Pavlovs Observation and observa- Windholz, G. (1991). I. P. Pavlov as a youth. Integrative Physiological
and Behavioral Science 26, 5167.
tion! To its members, the motto means that the is-
sues should be debated in the light of the observed ev- John J. Furedy
idence, not the political loyalties of the interlocutors.
Thus, in addition to the emphasis on interdisci-
plinary, integrated research, Pavlovs influence today
stands for open discussion and the resistance to politi-
PERCEPTION
cal pressures on scientific research. See: DISCRIMINATION AND GENERALIZATION
See also: CONDITIONING, CLASSICAL AND

INSTRUMENTAL
Bibliography PHARMACOLOGICAL TREATMENT

Anokhin, P. K. (1949). Ivan Petrovich Pavlov: Zhizn, deiatelnost i OF MEMORY DEFICITS
nauchnaia shkola. Moscow and Leningrad: Izdatelstvo
Akademii Nauk SSSR. Everyone experiences occasional memory lapses,
Furedy, J. J. (2001). An epistemologically arrogance community of which are often termed senior moments by those older
contending scholars: A pre-Socratic perspective on the past, than they care to be. While a frequent source of com-
present, and future of the Pavlovian Society. (Gantt memorial plaint, such events are generally not cause for worry.
Lecture, October, 2000 meeting of the Pavlovian Society, An- However, there are diseases that can cause much
napolis). Integrative Physiuological and Behavioral Science 36,
514. more profound and debilitating memory impair-
Furedy, J. J., Heslegrave, R. J., and Scher, H. (1984). Psychophy- ment. The most commonly recognized illness that
siological and physiological aspects of T-wave amplitude in causes memory loss is Alzheimers disease (AD). In
the objective study of behavior. Pavlovian Journal of Biological addition, other neurodegenerative diseases, such as
Science 19, 182194. Parkinsons, Huntingtons and Picks disease, include
Gantt, W. H. (1991). Ideas are the golden coins of science. Integra-
tive Physiological and Behavioral Science 26, 6873. memory problems among their symptoms. Vascular
Grigorian, N. A. (1974). Pavlov, Ivan Petrovich. In Dictionary of sci- disorders and stroke are common causes of memory
entific biography, Vol. 10. New York: Scribner. impairment in the elderly. There is also a rapidly
Hull, C. L. (1931). Goal attraction and directing ideas conceived growing population of otherwise healthy elderly peo-
as habit phenomena. Psychological Review 38, 487505. ple who develop memory problems that are severe
(1943). Principles of behavior: An introduction to behavior theory.
New York: Appleton-Century-Crofts. enough to interfere with their everyday lives. While
Koshotiants, K. S. (1950). I. M. Sechenov. Moscow: Izdatelstvo these conditions are most commonly seen in the aged,
Akademii Nauk SSSR. many young people with depression or schizophre-
MacCorquodale, K., and Meehl, P. (1953) Preliminary suggestions nia, or who suffer from AIDS, also experience im-
as to a formalization of expectancy theory. Psychological Review paired memory. All of these people could potentially
60, 5563.
Moltz, H. (1957). Latent extinction and the fractional anticipatory benefit from drug therapies that would relieve memo-
goal response. Psychological Review 64, 229241. ry deficits.
Pavlov, I. P. (1952). Ivan Petrovich Pavlov (avtobiografiia). In Pol- Many treatments for ailing memory, both phar-
noe sobranie sochinenii, Vol. 6. Moscow and Leningrad:
Akademii Nauk SSSR. macological and nonpharmacological have been de-
Rescorla, R. A. (1988). Pavlovian conditioning: Its not what you scribed. However, no drug has been discovered that
think it is. American Psychologist 43, 151160. is overwhelmingly effective in improving memory. A
Rescorla, R. A., and Wagner, A. R. (1972). A theory of Pavlovian major reason for this is that there are different types
conditioning: Variations in the effectiveness of reinforcement of memory, each of which engages, at least to some
eds., Classical conditioning: Current theory and research. New degree, different brain regions. Similarly, while the
York: Appleton-Century-Crofts. biochemical mechanisms required for memory for-
Sechenov, I. (1866). Refleksy golovnogo mozga. St. Petersburg: Tipo- mation are not completely understood, there is evi-
graphiia A. Golovachova. dence indicating that more than one system is in-
Segal, E. M., and Lachman, R. (1972). Complex behavior or higher volved. This suggests that different drugs will be
mental process: Is there a paradigm shift? American Psycholo-
gist 27, 4655. needed to treat the memory impairments that arise
Spence, K. W. (1956). Behavior theory and conditioning. New Haven, from different illnesses. In practice, the problem is
CT: Yale University Press. compounded by the possibility that different stages of
PHARMACOLOGICAL TREATMENT OF MEMORY DEFICITS 521
a disease may have different underlying causes, and tems. Overactivation of these latter systems induces
by situations where more than one disease is present. profound side effects, including nausea, vomiting, di-
For example, Alzheimers disease can be categorized arrhea and dizziness, and thus limiting the useful
as mild, moderate, or severe, and some patients with dose range of anticholinesterase drugs.
AD also have Parkinsons disease or vascular demen-
tia; each of these situations may require a different Cognex(r) (tacrine and tetrahydroacridine) be-
drug, or different amounts of the same drug, to be came the first FDA-approved treatment for AD in
treated effectively. Thus, in order to show that a drug 1993. Unfortunately, its mild memory enhancing
with only modest activity is effective, it is important properties were more than offset by the need to dose
to test the drug under controlled conditions that in- four times per day and problems with toxicity that re-
clude a uniform patient population. In addition, dou- quired regular monitoring of liver function. When Ar-
ble-blind trials comparing placebo against active drug icept(r) (donepezil) became available in 1997, its su-
are essential for objective evaluation of a drug candi- perior safety margin and longer half-life (once-a-day
date because placebo effects are relatively common in dosing) allowed it to quickly supplant Cognex(r) as
memory studies. the preferred therapy for AD. At the beginning of
2002, Aricept(r) remains the most-used agent in its
class. Two more recently available anticholinesterase
Prescription Drugs to Treat Memory compounds are Exelon(r) (rivastigmine; 1998) and
Impairment Reminyl(r) (galanthamine; 2000). In addition to in-
Drugs that are currently available to treat memo- hibiting acetylcholinesterase, Reminyl(r) is described
ry impairments can be divided into two broad classes, as acting on brain nicotinic cholinergic receptor to
according to whether a physicians prescription is enhance neurotransmitter release. It is not yet clear
necessary to obtain them. Prescription drugs have whether this action gives Reminyl(r) memory-
had to meet defined standards of safety and must be improving properties that are superior to other ace-
shown to be effective to treat a particular condition. tylcholinesterase inhibitors.
In the United States, the Food and Drug Administra-
tion (FDA) currently recognizes only AD and, more Two other prescription drugs are marketed for
recently (summer of 2001), mild cognitive impair- the treatment of memory impairments, but these
ment as indications for memory enhancing treat- agents have not been approved by the FDA and so are
ment. All currently marketed prescription drugs for not currently available in the United States. The first
memory improvement have shown statistically signifi- of these is Alcar(r) (acetyl-l-carnitine), which was first
cant results in AD patients. However, while the effects introduced in Italy in 1985; its mechanism of action
of these compounds were significant in controlled tri- seems, like that of the acetylcholinesterase inhibitors
als, in actual practice they have generally been mea- described above, also to involve enhancing choliner-
ger. The disappointing results of these drugs could be gic neurotransmission in the brain. However, the ef-
because their mechanism of action is not an effective fect appears to be achieved by elevating levels of nerve
way to enhance memory. More likely, the poor outgrowth factor (NGF), a neurotrophin that is critical to
come results from variability in the real-world patient maintaining the survival and normal functioning of
population. An additional problem is that currently cholinergic neurons. (NGF administration is being
available drugs do not treat the cause of the illness, explored as a therapy in its own right.) Akatinol(r)
so memory in AD patients is continually getting (memantine), first available in Germany in 1982, is
worse. It may be that existing drugs would be more different from all of the previously discussed drugs in
effective for treating people with less severe, or more that its primary action does not appear to involve cho-
slowly deteriorating, memory impairments. linergic neurons. Akatinol(r) interacts in a complex
Compounds that act to inhibit acetylcholinester- way with glutamate-mediated neurotransmission, the
ase dominate the group of currently available thera- most important excitatory system in the brain. Over-
peutics to treat memory impairment in AD. This en- activation of glutamate receptors, particularly those
zyme degrades acetylcholine, a neurotransmitter of the NMDA (n-methyl-d-aspartate) type, leads to
used by a population of neurons that degenerates in neuronal death. Akatinol(r) acts as an antagonist to
the early stages of the disease. The rationale for the prevent this overactivation, yet seems not to interfere
development of acetylcholinesterase inhibitors is that with normal synaptic function. At this point it is not
they would strengthen and prolong the weakened sig- possible to compare the efficacy of Alcar(r) or Aka-
nals sent by remaining cholinergic neurons in the tinol(r) to marketed anticholinesterase drugs, but ob-
brain of AD patients. A problem with this approach servations based upon the long existing period of
is that acetylcholine is also used as a neurotransmitter clinical experience suggest that neither drug will turn
in the neuromuscular and autonomic nervous sys- out to be a substantially better memory enhancer.
522 PHARMACOLOGICAL TREATMENT OF MEMORY DEFICITS
Nonprescription Drugs to Treat Memory reducing free radical damage. It is likely that vitamin
Impairment E will not be shown to be particularly effective at im-
proving memory; rather, its use seems better suited
In addition to prescription drugs, a large number
to protect against age- or disease-related memory im-
of over-the-counter agents are advertised as benefit-
pairments.
ing memory. Many nonprescription memory boosters
are formulations containing several compounds.
While it is not possible to discuss all of these com- Conclusion
pounds individually, some of them deserve mention.
In the 1990s, a significant breakthrough in the
Huperzine A is a potent acetylcholinesterase inhibitor
treatment of memory disorders occurred with FDA
that is derived from a particular type of club moss and
approval of several acetylcholinesterase inhibitors for
sold as a dietary supplement for memory loss and
symptomatic treatment of AD. Although the currently
mental impairment. While Huperzine A has been
marketed drugs are only modestly effective, they have
shown to be an effective memory enhancer, its side ef-
refocused attention on the problem of memory im-
fects are identical to other anticholinesterase drugs.
pairment and the possibility that it can be treated.
In contrast to the agents that will be described in the
This, in turn, has provided a strong stimulus for re-
remainder of this section, there are substantial risks
search to develop more useful pharmacological treat-
associated with the use of Huperzine A.
ments for memory impairment. Several promising
Extracts from leaves of the Ginkgo biloba tree have drugs are currently undergoing clinical trials, and it
been shown to improve memory in several studies of is likely that substantial advances in the treatment of
healthy volunteers and patients with dementia. How memory deficits will take place before 2010.
Ginkgo works is not entirely clear, but it seems to in-
See also: ALZHEIMERS DISEASE: BEHAVIORAL
volve a combination of effects including increasing ce- ASPECTS; ALZHEIMERS DISEASE: HUMAN
rebral blood flow and antioxidant action. The latter DISEASE AND THE GENETICALLY ENGINEERED
reduces or prevents neuronal damage caused by the ANIMAL MODELS; COGNITIVE ENHANCERS
generation of chemical-free radical species. The ac-
tive agents in Ginkgo extracts have not been com- Bibliography
pletely characterized, and the extraction procedure is Bai, D. L., Tang, X. C., and He, X. C. (2000). Huperzine A, a po-
not standardized, so it is likely that currently market- tential therapeutic agent for treatment of Alzheimers disease.
Current Medicinal Chemistry 7 (3), 355374.
ed products have varying effectiveness. Two com- Fioravanti, M., and Flicker, L. (2001). Efficacy of nicergoline in de-
pounds derived from the ergot fungus, hydergine mentia and other age associated forms of cognitive impair-
and nicergoline, share the enhancement of cerebral ment (Cochrane Review). Cochrane Database of Systematic Re-
circulation that has been described for Ginkgo ex- views 4, CD003159.
tracts. Both hydergine and nicergoline have shown Flicker, L., and Grimley Evans, G. (2001). Piracetam for dementia
or cognitive impairment (Cochrane Review). Cochrane Data-
some beneficial effects in treating memory impair- base of Systematic Reviews 2, CD001011.
ments, but the compounds have not been evaluated Gillis, J. C., Benefield, P., and McTavish, D. (1994). Idebenone. A
as rigorously as Ginkgo. review of its pharmacodynamic and pharmacokinetic proper-
ties, and therapeutic use in age-related cognitive disorders.
Compounds that enhance learning and memory Drugs and Aging 5 (2), 133152.
by an undefined mechanism are often termed Giovanello, K. S., and Verfaellie, M. (2001). Memory systems of the
nootropics. Although there are many agents that could brain: A cognitive neuropsychological analysis. Seminars in
Speech and Language 22 (2), 107116.
be called nootropic, the name is most often applied
Olin, J., Schneider, L., Novit, A., and Luczak, S. (2001). Hydergine
to piracetam and its relatives: aniracetam, oxira- for dementia (Cochrane Review). Cochrane Database of System-
cetam, and pramiracetam. Memory-enhancing effects atic Reviews 2, CD000359.
of piracetam in animals were first described in the late Perkins, A. J., Hendrie, H. C., Callahan, C. M., Gao, S., Unverzagt,
1960s. In the intervening years these drugs have been F. W., Xu, Y., Hall, K. S., and Hui, S. L. (1999). Association
of antioxidants with memory in a multiethnic elderly sample
used to treat many conditions with memory impair-
using the third national health and nutrition examination
ment, including AD, but no convincing improve- survey. American Journal of Epidemiology 150 (1), 3744.
ments have been described. Another nootropic, ide- Salvioli, G., and Neri, M. (1994). L-acetylcarnitine treatment of
benone, shares a similar history. mental decline in the elderly. Drugs Under Experimental and
Clinical Research 20 (4), 169176.
A final compound of note is vitamin E. Low serum Sano, M., Ernesto, C., Thomas, R. G., Klauber, M. R., Schafer, K.,
levels of vitamin E have been correlated with memory Grundman, M., Woodbury, P., Growdon, J., Cotman, C. W.,
Pfeiffer, E., Schneider, L. S., and Thal, L. J. (1997). A con-
impairments. In addition, a high dose of vitamin E
trolled trial of selegiline, alpha-tocopherol, or both as treat-
has been shown to be effective in slowing cognitive de- ment for Alzheimers disease. The Alzheimers Disease Coop-
terioration in AD patients. Vitamin E has antioxidant erative Study. New England Journal of Medicine 336 (17), 1,216
properties, so its beneficial properties are ascribed to 1,222.
PHOBIAS 523
Sramek, J. J., Veroff, A. E., and Cutler, N. R. (2001). The status of quire comparatively few; a satisfactory explanation of
ongoing trials for mild cognitive impairment. Expert Opinion phobias must accommodate this fact as well as the ap-
on Investigational Drugs 10 (4), 741752.
Wettstein, A. (2000). Cholinesterase inhibitors and Ginkgo ex-
pearance of phobias in a significant minority of the
tractsare they comparable in the treatment of dementia? human population. It has also become apparent that
Comparison of published placebo-controlled efficacy studies people are more resilient than most psychologists
of at least six months duration. Phytomedicine 6 (6), 393401. have implied. Phobic patients who behave coura-
Winblad, B., and Poritis, N. (1999). Memantine in severe demen- geously during the course of treatment and soldiers
tia: Results of the 9M-Best Study (Benefit and efficacy in se-
verely demented patients during treatment with memantine).
who perform dangerous acts are notable examples of
International Journal of Geriatric Psychiatry 14 (2), 135146. resilience.
Zurad, E. G. (2001). New treatments for Alzheimers disease: A re-
view. Drug Benefit Trends 13, 2740.
Gregory M. Rose Causes

The major features of the conditioning theory of
phobias are as follows. Fears are acquired by a process
of conditioning. Neutral stimuli that often are associ-
ated with a fear-producing or pain-producing state of
PHOBIAS affairs develop fearful qualities. They become condi-
Phobias are intense, persistent, unadaptive fears that tioned phobic stimuli. The strength of the phobia is
are irrational/excessive. They are commonly classified determined by the number of repetitions of the asso-
into three groups: complex phobias, including agora- ciation between the pain/fear experienced and the
phobia (fear of public places, travel); social phobias stimuli, and by the intensity of the pain or fear experi-
(fear of social situations/scrutiny); and circumscribed enced in the presence of the stimuli. Stimuli that re-
phobias, including intense fears of insects, animals, semble the fear-evoking ones also acquire phobic
heights, and enclosed spaces. properties; that is, they become secondary condi-
tioned stimuli. The likelihood of a phobias develop-
There are biological contributions to the devel-
ing is increased by confinement, by exposure to in-
opment of some phobias, but the main determinants
tensely painful or frightening situations, and by
appear to be learned. Three main pathways to the ac-
frequent associations between the new conditioned
quisition of phobias have been identified. The condi-
stimulus and the pain/fear. In an important exten-
tioning acquisition of a phobia results from exposure
sion, it has also been proposed that once objects or sit-
to a traumatic stimulation or from repeated expo-
uations acquire phobic qualities, they develop moti-
sures to aversive sensitizing conditions. The second
vating properties. A secondary fear drive emerges.
pathway is vicarious acquisition: direct or indirect ob-
Behavior that successfully reduces fear, notably avoid-
servations of people, or of other animals, displaying
ance behavior, will increase in strength.
fear. Among humans the transmission of fear-
inducing verbal information is the third pathway. For Supporting evidence for the theory was drawn
a considerable time, explanations of the acquisition of from six sources: research on the induction of fear in
phobias were dominated by the conditioning theory, laboratory animals, the development of anxiety states
which emphasized the importance of exposure to in combat soldiers, experiments on the induction of
traumatic stimulation; recognition that fears can be fear in a small number of children, clinical observa-
acquired vicariously and/or by the direct transmission tions (e.g., dental phobias), incidental findings from
of information has led to a fuller account of the causes the use of aversion therapy, and a few experiments on
of phobias. the effects of traumatic stimulation.
Important advances have been made in our abili- The strongest and most systematic evidence was
ty to reduce phobias. Under controlled conditions, it drawn from a multitude of experiments on laboratory
is now possible to produce substantial and lasting re- animals. It is easy to generate conditioned fear reac-
ductions of phobias within a few sessions. It requires tions in animals by exposing them to a conjunction of
greater effort and far more time to reduce the com- neutral and aversive stimuli, usually electric shock.
plex and intense phobias, such as agoraphobia, but These fear reactions can be intense and persistent.
even these respond moderately well to treatment pro- Phobias can result from traumatic experiences in
grams. There have been several attempts to explain combat. In clinical practice, it is not uncommon for
how and when these methods of fear reduction patients to give an account of the development of
achieve their effects, but each explanation has limita- their phobias that can be construed in conditioning
tions. terms, and sometimes they can date the onset of the
Despite the many opportunities and circum- phobias to a specific conditioning experience (e.g.,
stances in which phobias might develop, people ac- Lautch, 1971, on thirty-four cases of dental phobia).
524 PHOBIAS
Di Nardo et al. (1988) found that nearly two-thirds of lays between the events. Hence, if the food aversion
their subjects who were phobic toward dogs had expe- phenomenon provides support for a new or a revised
rienced a conditioning event in which a dog featured, conditioning theory of phobias, the temporal quali-
and in over half the dog had inflicted pain. It is im- ties of the conditioning processes must be deempha-
portant, however, that over two-thirds of a compara- sized.
ble group of subjects who were not frightened of dogs
reported that they, too, had experienced a condition- There are various arguments against acceptance
ing event, and that in over half of these instances the of the conditioning theory of phobias as a compre-
animal had inflicted pain. hensive explanation. People fail to acquire phobias in
what should be fear-conditioning situations, such as
These reports provide some support for the con- air raids. It is difficult to produce stable phobic or fear
ditioning theory but also illustrate the fact that condi- reactions in human subjects even under controlled
tioning experiences, even those of a painful nature, laboratory conditions. The theory rests on the unten-
do not necessarily give rise to a phobia or even to fear. able equipotentiality premise (Seligman, 1972). The
Here, as in other instances, there was less fear than distribution of fears and phobias in normal and neu-
an unqualified conditioning theory would lead us to rotic populations is difficult to reconcile with the the-
expect. Presumably the people who experienced con- ory. A significant number of people with phobias re-
ditioning events but failed to acquire a fear or phobia count histories that cannot be accommodated by the
were protected by a history of harmless contacts theory. We also know that fears and phobias can be
with dogs. The roles that phobic patients attribute to acquired vicariously, and that fears can be acquired by
direct and indirect experiences in generating their the reception of threatening verbal information.
phobias differ with the content of the phobia, and of Fears, and possibly phobias as well, can be acquired
course the accuracy of their reports cannot be as- even when the causal events are temporally separated
sured. In their analysis of 183 phobic patients, divid- (see Rachman, 1990).
ed into six groups. Ost and Hugdahl (1985) found a
range of attributions. For example, 88 percent of the
agoraphobic patients attributed the onset of the pho- Neoconditioning Concepts
bia to a conditioning experience, but only 50 percent The traditional insistence on the contiguity of the
of those who were frightened of animals attributed conditioned stimulus and the unconditioned stimulus
the onset of their phobias to such an experience. as a necessary condition for the establishment of a
Among the animal phobics, 40 percent traced the ori- conditioned response is mistaken. Rescorla has ob-
gin of the phobia to indirect experiences; such attri- served that although conditioning can sometimes be
bution was uncommon among the agoraphobics. slow, in fact most modern conditioning preparations
Although the importance of the phenomenon of routinely show rapid learning. One trial learning is
acquired food aversions was not made evident until not confined to flavor-aversion (1988, p. 154). Ap-
1966, it is sometimes used to buttress the condition- parently the associative span of animals is capable of
ing theory. The findings on this aversion also served bridging long temporal intervals (Mackintosh, 1983,
to prompt radical rethinking of the concept of condi- p. 172). However, the learning must be selective; oth-
tioning. Garcia and his colleagues were the first to erwise, animals would collect what Mackintosh has re-
demonstrate that strong and lasting aversive reac- ferred to as a useless clutter of irrelevant associa-
tions can be acquired with ease when the appropriate tions. According to Mackintosh, the functioning of
food stimulus is associated with illness, even if the ill- conditioning is to allow organisms to discover proba-
ness occurs many hours after eating (Garcia, Ervin, ble causes of events of significance.
and Koelling, 1966). Given that the genesis of food
Given this new view, that conditioning is far more
aversions is a form of conditioning, if we also allow an
flexible and wide-ranging than was previously sup-
equation between the acquisition of a food aversion
posed, many of the objections to the conditioning
and the acquisition of a fear, this phenomenon may
theory of fear and phobias are weakened or eliminat-
have a bearing on the effect of the phobias.
ed. Although the application of neoconditioning con-
If the acquisition of aversions is used to support cepts can shore up the conditioning theory, at the
the conditioning theory of phobias, it will have to take present stage the new view is too liberal. It lacks lim-
into account the temporal stretch of the phenome- its, and there is little left to disallow. In theory, almost
nonthat is, the delay that can intervene between the any stimulus can become a conditioned signal for
tasting of the food and the onset of the illness. Classi- fear; but in practice people develop comparatively
cal conditioning is expedited by temporal proximity few phobias, and those we do acquire are confined to
between the stimuli, but food aversions can be easily a limited range of stimuli. Phobias are not normally
and rapidly established even when there are long de- distributed.
PHOBIAS 525
Several sources of evidence suggest that phobias jected to rigorous testing. The demonstration of pre-
and fears can be acquired vicariously. Reports given paredness in the development of phobias among lab-
by phobic patients, wartime observations, correlations oratory monkeys encourages the belief that
between the phobias displayed by parents and chil- Seligmans theory retains considerable value. Mineka
dren, laboratory demonstrations of conditioned fears, (1988) has shown that the fears induced in monkeys
and research on animals have all provided some sup- in laboratory conditions are intense, vivid, and last-
port for this view. ing. The animals readily developed fears of snakes
but showed little or no propensity to develop fears of
Verbal information can also generate a fear, and
biologically insignificant stimuli such as flowers.
it is possible that in limited circumstances, it can even
induce a phobia. Clinical evidence, especially that ac-
cumulating on the nature of panic disorders, suggests Fear Reduction
that phobias can be generated by information that is
slightly or not at all threatening but is catastrophically Three powerful and dependable methods for re-
misinterpreted by the recipient as being threatening. ducing fear have been developed since the 1970s: de-
sensitization, flooding, and therapeutic modeling.
The common element in all three methods is the re-
Biological Determinants peated and/or prolonged exposure of the fearful per-
The nonrandom distribution of human phobias, son to the stimulus or situation that provokes the fear
the high incidence of phobias of snakes and spiders (the exposure method). The selection of the appro-
and the low incidence of fears/phobias of motor trav- priate fear-reducing technique depends on the na-
el, the remarkable speed with which certain objects ture of the phobia and the preference of the fearful
can be transformed into objects of fear, and the com- subject, but all three methods are reliably robust.
mon occurrence of irrational fears all point to the op- Fears of circumscribed stimuli, such as snakes or spi-
eration of nonlearned processes in the acquisition of ders, can be reduced fairly rapidly, even if they are in-
fears and phobias. The main explanations fall into tense and well established. The reduction or elimina-
two classes: Some human fears and phobias are in- tion of more complex fears, such as agoraphobia,
nately determined, or people are innately disposed requires greater effort and time. Despite these impor-
rapidly to acquire phobias of certain specifiable ob- tant practical advances, there still is no widely accept-
jects or situations. ed explanation for the effects of these techniques.
The most influential explanation is that set out by

See also: BEHAVIOR THERAPY; CONDITIONING,
Martin Seligman, who argued, The great majority of
CLASSICAL AND INSTRUMENTAL;
phobias are about objects of natural importance to OBSERVATIONAL LEARNING; TASTE AVERSION
the survival of the species . . . (human phobias are AND PREFERENCE LEARNING IN ANIMALS
largely restricted to objects that have threatened sur-
vival, potential predators, unfamiliar places, and the
dark) (1972, p. 450). He postulates that certain kinds Bibliography
Di Nardo, P. A., Guzy, L. T., Jenkins, J. A., Bak, R. M., Tomasi, S.
of fears are readily acquired because of an inherited
F., and Copland, M. (1988). Etiology and maintenance of dog
biological preparedness. These phobias are highly fears. Behaviour Research and Therapy 26, 245252.
prepared to be learned and, like other highly pre- Garcia, J., Ervin, F., and Koelling, R. (1966). Learning with pro-
pared relationships, they are selective and resistant longed delay of reinforcement. Psychonomic Science 5, 121
to extinction, and probably non-cognitive (1972, p. 122.
Lautch, H. (1971). Dental phobia. British Journal of Psychiatry 119,
455).
151158.
The main features of prepared phobias are that Mackintosh, N. J. (1983). Conditioning and associative learning. New
York: Oxford University Press.
they are very easily acquired (even by watered-down
Mineka, S. (1988). A primate model of phobic fears. In H. Eysenck
representations of the actual threat), selective, stable, and I. Martin, eds., Theoretical foundations of behaviour therapy.
biologically significant, and probably noncognitive. New York: Plenum Press.
After some encouraging early laboratory demonstra- Ost, L. G., and Hugdahl, K. (1985). Acquisition of blood and dental
tions of fear preparedness in human subjects, subse- phobia and anxiety response patterns in clinical patients. Be-
haviour Research and Therapy 23, 2734.
quent research was disappointing because the phe-
Rachman, S. J. (1990). Fear and courage, 2nd edition. New York: W.
nomenon appeared to be too fragile. The laboratory H. Freeman.
effects appeared to be weak, transient, and difficult to Rescorla, R. A. (1988). Pavlovian conditioning. American Psychologist
reproduce. The plausibility of the concept has been 43, 151160.
weakened but not seriously damaged, and more pow- Seligman, M. E. P. (1972). Phobias and preparedness. Behavior
Therapy 2, 307320.
erful stimuli and more appropriate measures of fear
responding are needed before the theory can be sub- Stanley J. Rachman
526 PHOTOGRAPHIC MEMORY
1990. The total oeuvre comprises over sixty books

and monographs plus nearly a thousand articles.
During his long life Piaget held professorships at
the University of Paris and at the Swiss universities of
Neuchtel, Lausanne, and Geneva. The chairs he
held were in psychology, sociology, and the history
and philosophy of science. His longest association was
with the University of Geneva. Among his many hon-
ors were over thirty honorary doctorates from major
universities (the first from Harvard, 1936) in a dozen
countries; numerous awards, including the Distin-
guished Scientific Contribution Award of the Ameri-
can Psychological Association (1970); and the presi-
dency of various scientific associations. For many
years he was director of the International Bureau of
Education. His wife, Valentine Chatenay Piaget, was
among his collaborators in the research for his first
few books, and especially in the study of their three
babies.
As an adolescent Piaget pursued malacology, the
study of mollusks, and reached a professional level,
publishing thirty-two papers in this field by 1916, his
twentieth year. He continued this line of work in natu-
ral history for the rest of his life.
In 1918 Piaget received his doctorate in natural
Jean Piaget (Psychology Archives, University of Akron)
science from the University of Neuchtel for a disser-
tation on the mollusks of the Valais, a region of Swit-
zerland. By that time he had begun to move toward
PHOTOGRAPHIC MEMORY the study of psychology, which he pursued in Zurich
See: EIDETIC IMAGERY and in Paris. In 1921 he published his first article on
child logic and thought, a subject that grew to domi-
nate his thinking throughout his later life.
PIAGET, JEAN (18961980) Egocentrism

Piagets first book in psychology, The Language
Together with Sigmund Freud and B. F. Skinner,
and Thought of the Child, appeared in 1923. In it he in-
Jean Piaget (18961980) was one of the three most in-
troduced his conception of egocentrism, interpreting
fluential psychologists of the twentieth century.
the world from ones own immediate perspective
Among developmental psychologists he has had no
without adequately taking into account the existence
equal or close second as to the volume, scope, and im-
of alternative perspectives. In three subsequent books
pact of his work. Yet he thought of his psychological
during the 1920s Piaget showed how this egocentrism
work primarily as a tool for the creation of a new sci-
pervades the childs mentality from about the age of
ence, genetic epistemologya new synthesis of logic,
5 to 10 in the domains of logic and reasoning, causal
philosophy, history of science, biology, and psycholo-
thinking, conceptions of the world, and (not pub-
gy.
lished until 1932) moral judgment.
All of these works can be construed as studies of
Life and Oeuvre learning in a wide sense, since through its interaction
Piaget was born on August 9, 1896, in Neuchtel, with the world, the childs intelligence develops: Mov-
Switzerland, and died in Geneva, on September 17, ing through several necessary stages, the child learns
1980. His father, Arthur Piaget, was a historian. to think more and more like an adult. The same can
Jeans first publication, a paragraph about sighting an be said of the trilogy Piaget wrote about his own three
albino sparrow, appeared in 1907, when he was 11 children in the first two years of life. These works,
years old. He was active until the end of his life, and using the method of naturalistic observation, delin-
posthumous monographs continued to appear until eate the major stages in (a) the development of active,
PIAGET, JEAN 527
intelligent, inventive exploration of the world; (b) the This phase was most fully expressed in Piagets
childs own activity in the construction of reality (the three-volume work Introduction lepistmologie gn-
permanent object, space, time, and causality); and (c) tique (1950). In 1956 he organized the Centre Inter-
the emergence of language and the symbolic or rep- national de lEpistmologie Gntique, an interdisci-
resentational function through play, dreams, and imi- plinary center for research and reflection on
tation. questions concerning the intersection of the natural
sciences, psychology (especially developmental), and
philosophy (especially epistemology). Rather than be-
Assimilation and Accommodation coming a philosopher, Piaget hoped to transform one
In the course of his work on infant development, branch of philosophy, epistemology, into a new sci-
Piaget introduced the twin concepts of assimilation ence. In 1957, the Centre turned its attention to the
and accommodation as tools for understanding cog- study of learningboth logical models of the learn-
nitive growth. The infant is born with a few basic re- ing process and the learning and development of log-
flexes. Through its own activity novelties arise (for ex- ical reasoning by the child. Piaget and his collabora-
ample, through the chance coincidence of events) tors published their theoretical and empirical
that are assimilated into these initial schemes, giving findings in four monographs (1959).
rise to changing schemes of action. These schemes
can assimilate external events or stimuli and, equally
important, they can assimilate each other, giving rise Equilibration Model
to new adaptive organizations. Paired with the pro- In the 1960s and 1970s, without dropping any of
cess of assimilation is that of accommodation, the way his previous concerns, Piaget turned his attention to
in which the set of schemes or cognitive organizations elaborating the equilibration model, an attempt to
must change in response to the new inputs (ali- specify the actual mechanisms by which intellectual
ments, as Piaget sometimes called them, emphasiz- growth and change come about.
ing the digestion metaphor).
Toward the end of the first year of life, as the
child repeats its actions in order to make interesting Alternative Theoretical Approaches
events recur (e.g., the noise of a rattle), it notices vari- Throughout his life Piaget was interested in the
ations in its own actions and their consequences. contrast between his own theoretical approach and
These variations and their consequences are, in their two others. He was quite drawn toward Gestalt psy-
turn, assimilated into existing schemes, and thus chology, especially its emphasis on holism and self-
these schemes grow. Piaget considered this analysis of regulating systems; but in the end he rejected it as re-
infant cognitive growth to be germane to his analysis lying too heavily on the analogy between perception
and descriptions of the growth of thought at other le- and thought, and being consequently nondevelop-
vels, including the history of science. mental. He was never drawn toward behaviorism (or
its antecedent, associationism); he objected to the lack
of any intrinsic structure in knowledge accrued as an
Stages of Development arbitrary collection of chance associations. He
In the 1930s and 1940s Piagets main focus was summed all this up in a favorite aphorism: Gestalt
on the stagewise progression of intelligence in infancy psychology speaks of structure without development;
and childhood. He elaborated his idea of three great behaviorism, of development without structure.
periods of intellectual growth: sensorimotor (02
years), preoperational (26 or 7 years), and concrete
operational (711 or 12 years). In the 1950s, this Method
model was expanded to include adolescent cognitive Piaget relied mainly on two related methods for
development in the period of formal operations. the exploration of the childs intellect. For his trilogy
on the origins of intellect in babies, and also for his
earlier Language and Thought in the Child, he relied on
Genetic Epistemology naturalistic observation. For most of his work, howev-
In the late 1940s Piaget displayed increasing pre- er, he stuck to the clinical method of extended in-
occupation with the further elaboration of ideas long teraction between child and investigator, with search-
held, the approach that has become known as genetic ing analysis of the protocols (i.e., of what the child
epistemology. Some of the most important components said and did in reaction to the problems posed by the
of this approach had been sketched in his religious adult). The clinical method evolved into something
prose poem La mission de lide (1915) and in his philo- approaching naturalistic observation in problem-
sophical novel Recherche (1918). solving situations. The problems were not construed
528 PIAGET, JEAN
as tasks having definite solutions to be sought by the of reversibilityif the water is poured back into the
child, but as occasions to provoke thought in the first container, it regains the original levelwill be a
child, thus permitting the experimenter to observe satisfying demonstration of conservation for the older
the childs way of thinking. child. The young child watching, or even pouring, the
Using different age groups, these methods per- liquid can actually see the level mounting or falling
mitted the study of the broad trajectory of cognitive but, because attention is centered on one dimension,
development. By avoiding narrower experimental does not yet grasp the idea of conservation that would
approaches, Piaget sacrificed the opportunity for lead to the coordination of changes of length and
what might be called microscopic analysis of the ef- width.
fects of specifiable variables on cognitive functioning. In other words, direct teaching or learning is rela-
What he gained was a better picture of child mentality tively ineffectual in modifying the growth of funda-
as a wholefirst as a system of beliefs and later as a mental cognitive categories and operations, because
structured group of operations. these depend on the protracted, often slow, develop-
ment of the knowing system as a whole through the
self-regulated activity of that knowing system.
Development, Learning, and Structure
Although the development of cognition as depict- Memory
ed by Piaget resembles what other psychologists
might call learning, there are a number of impor- From 1921, when he published his first empirical
tant differences. First, changes in performance are study of child development, until the 1960s, Piaget
seen as a function of developmental stage, not of re- did virtually no work on memory. A possible excep-
peated exposures and responses to the same stimulus. tion was his use of tasks involving memory but focus-
Second, the investigator analyzes broad strategic ing on other problems. For example, in 1923, in Lan-
changes in approach rather than the correctness and guage and Thought of the Child, Piaget studied the way
incorrectness of solutions or memories. Third, what in which a child who has just been told a story repeats
accrues over time is not a sum of associations but op- it to another child. This work could be considered a
erative structures; thus, for Piaget the older subject study of memory, but Piagets interest was in the rela-
does not necessarily know more than the younger tion between the first childs comprehension of the
but knows differently. story and the communication from the first child to
the second. He was aware of the involvement of mem-
For Piaget a structure is not the momentarily ory in this task but thought he could distinguish er-
given perceptual configuration of interest to Gestalt rors of memory from those of comprehension and
psychologists. A mental structure is a set of logicoma- communication. In his later work he took a very dif-
thematical operations or mental acts, permitting the ferent tack, emphasizing the effect of changes in com-
decomposition of wholes into parts and the recompo- prehension on memory.
sition of wholes from parts. To take a very simple ex-
ample, the idea of the permanent object entails the For the most part, during a very long period Pia-
recognition of the continued existence of an object as gets interest lay primarily in the general operations
it moves around in spacethe movements AB + BC and structures of mental activity rather than in the
AC, the movements AB + BA 0. Thus, the idea contents of experience or the stuff of thought. But in
of the permanent object is an embodiment of the a wider perspective Piaget believed that growth comes
group of displacements. about through interaction with the world, and that
this world must somehow be represented in the
childs mind (and consequently in Piagets theory).
Conservation of Matter About 1942, Piaget began a systematic study of what
In a key and famous illustration of the mentality he called the figurative aspects of thought
of the concrete operational child, Piaget discovered perception, imagery, and memorywas contrasted
that the young child does not understand that a given with the operative aspects of mental activity. By far
amount of matter remains the same under transfor- the largest part of this effort was centered on the
mations of shape; thus, if water is poured from a study of perception: some sixty experimental papers,
short, wide vessel into a long, thin tube, the child may brought together in The Mechanisms of Perception
believe that as the water level mounts higher and (1969). But he also studied mental imagery, which re-
higher, the amount of water increases. The weight of sulted in yet another volume, Mental Imagery in the
experimental evidence shows that the child is rela- Child: A Study of the Development of Imaginal Representa-
tively unaffected by repeated exposures to this event, tion (Piaget and Inhelder, 1971).
because it can always map the results of direct obser- The work on imagery led on to work on memory,
vations onto the preexisting schema. The argument resulting in the book Memory and Intelligence (Piaget,
PLACE CELLS 529
Inhelder, and Sinclair-De Zwart, 1973). Perhaps the Gruber, H. E., and Vonche, J. J. (1977). The essential Piaget. New
most striking finding of this work is that rather than York: Basic Books.
Inhelder, B., Sinclair, H., and Bovet, M. (1974). Learning and the
remaining stable or decaying, a memory can actually development of cognition. Cambridge, MA: Harvard University
improve with time because its evolving structure de- Press.
pends on the childs maturing operativity. For exam- Kuhn, D., ed. (1989). Human Development 32 (6), 325387.
ple, a young child shown a series of rods arranged Piaget, J. (1921). Essai sur quelques aspects du dveloppement de
from short to long may remember them 1 week later la notion de partie chez lenfant. Journal de psychologie normale
et pathologique 18, 449480.
as a dichotomy, short rods and long rods. But 6 (1961; reprint 1969). The mechanisms of perception, trans. G.
months later, reflecting the childs growing mastery N. Seagrim. New York: Basic Books.
of the scheme of seriation, the child may remember (1970). Piagets theory. In P. H. Mussen, ed., Carmichaels
the series as it was originally presented. In contrast manual of child psychology. New York: Wiley.
with his position in the 1920s, when he tried to sepa- (1967; reprint 1971). Biology and knowledge, trans. B. Walsh.
Chicago: University of Chicago Press.
rate memory from understanding, Piaget now con- Piaget, J., and Inhelder, B. (1966; reprint 1969). The psychology of
cluded, The structure of memory appears to be part- the child, trans. H. Weaver. New York: Basic Books.
ly dependent on the structure of the operations (1966; reprint 1971). Mental imagery in the child: A study of
(Piaget, 1970, p. 719). the development of imaginal representation, trans. P. A. Chilton.
New York: Basic Books.
Piaget, J., Inhelder, B., and Sinclair-De Zwart, H. (1968; reprint
1973). Memory and intelligence, trans. A. J. Pomerans. New
Collaborators York: Basic Books.
The work we call Piagets was really teamwork. Its
Howard E. Gruber
scope and volume are so vast that it cannot be imag-
ined without the skillful leadership necessary to gen-
erate enthusiasm and maintain a sense of direction.
Piaget had many collaborators, ranging from student
assistants to distinguished scientists and scholars in PLACE CELLS
various fields. Besides psychologists there were math- The hippocampus is a medial temporal lobe structure
ematicians, logicians, philosophers and historians of of critical importance for the encoding and retention
science, biologists, physicists, and linguists. Almost of episodic memory in general and spatial memory in
everyone he worked with called him patron (boss). particular. A milestone in the detection of these hip-
His longest collaboration (50 years), and the most im- pocampal functions was the discovery that most of the
portant, was with Brbel Inhelder, who began as his pyramidal cells in the hippocampus exhibit location-
student and became a distinguished scientist in her specific activity (see Figure 1) and that the activity of
own right, almost always working together or in close such place cells is influenced by the training history
proximityboth spatially and intellectuallywith of the animal. This chapter reviews place cells gov-
Piaget. erning mechanisms, their ensemble properties, and
their possible contribution to spatial memory.
Conclusion
Historical Landmarks
Since about 1970 there have been numerous criti-
cal studies of Piagets empirical findings and of his Our understanding of hippocampal place cells
theoretical approach. By about 1990, much of the rests on two important discoveries from the early
anti-Piagetian criticism had ebbed and had given way 1970s. First, James B. Ranck reported that hippocam-
to neo-Piagetian efforts to assimilate Piagets find- pal neurones fall into two functionally different class-
ings, correct some of his errors, and synthesize his es based on their firing patterns in freely moving rats.
work with newer developments in cognitive and social Complex-spike cells fired at low rates (normally < 1
psychology. Most of his empirical findings have been Hz), but often in bursts of two to seven spikes at 150
verified by studies in many countries. Perhaps his 200 Hz. These cells were likely to be pyramidal cells.
most important contribution to developmental psy- Theta cellsthe second classhad high spontaneous
chology was to reveal the child as a thinking being, firing rates (normally > 10 Hz) and were probably in-
and the childs intellect as growing through its own ef- hibitory interneurons.
forts in interaction with the physical and social world. The second discovery was the observation by
John OKeefe and Jonathan Dostrovsky that a major
See also: OBJECT CONCEPT, DEVELOPMENT OF
proportion of complex-spike cells in the rat hippo-
Bibliography campus had spatial correlates. These place cells
Chapman, M. (1988). Constructive evolution, origins and development fired when the rat was in a specific location (the place
of Piagets thought. New York: Cambridge University Press. field of the cell) but were nearly silent in other posi-
530 PLACE CELLS
Figure 1
Place fields of 10 simultaneously recorded pyramidal cells and one interneuron (cell 10). Each diagram shows the path of the animal
during foraging in a square open field (viewed from above). Locations of firing are shown for each cell. Each dot indicates the position
of a spike. Note that pyramidal cells fire at distinct locations, whereas the interneuron fires throughout the environment.
PLACE CELLS 531
tions (see Figure 1). Firing rates in the field were ments may give rise to very different place represen-
often ten to twenty times higher than the background tations, and subtle changes in sensory input may com-
rate. Place fields reflected location as such and not be- pletely change the spatial firing correlates of a set of
havior emitted at specific locations. Different cells hippocampal neurons. Third, not all locations in an
had place fields at different places, and collectively environment are represented by an equal number of
they covered the entire test environment (see Figure place cells. In enclosed chambers, place fields appear
1). Unlike the sensory cortices, firing correlates in the to be more common near edges and walls than in the
hippocampus were strongly nontopographic: place center; in environments with distinct reward loca-
fields of neighboring cells in the hippocampus were tions, a larger number of cells may have place fields
not closer than those of distant cells. at the goal location than at other locations. All these
observations suggest that the relation between the
These discoveries led to the suggestion that place structure of the environment and the firing fields of
cells form a distributed, maplike representation of hippocampal neurons is nonlinear.
the spatial environment that the animal can use for
efficient navigation (OKeefe and Nadel, 1978).
These findings and the accompanying theory have Place Cells, Spatial Memory, and Synaptic
strongly stimulated research on hippocampal func- Plasticity
tion. New technology has expanded the study of place
Place cells appear to be responsible for some of
cells to neuronal ensembles (Wilson and McNaugh-
the underlying memory computations of the hippo-
ton, 1993) and has the power to reveal how cognitive
campus (Moser and Paulsen, 2001; Eichenbaum,
functions arise from complex interactions in defined
2001). Several observations suggest that individual
neuronal networks.
hippocampal pyramidal cells express information re-
trieved from the animals memory. First, as long as a
What Factors Determine a Place Field? rat remains in its recording apparatus, there is often
no change in the location-specific firing of hippocam-
Hippocampal place cells respond to multiple pal place cells after the removal of surrounding land-
sources of sensory information (Best et al., 2001). marks or the switching off of lights. That this persis-
Under most conditions, distal visual landmarks exert tence of firing obtains even in cells that were
the strongest influence. When such landmarks are ro- originally under strong visual control suggests that
tated in concert, place fields frequently follow the recent experience influences firing patterns. Second,
landmarks. When the test environment is stretched or because a cells firing or not at a given location may
truncated, there is often a corresponding change in depend on where the animal comes from and where
the shape of the place field. Proximal landmarks such it is going next, it appears that recent memory can in-
as surfaces usually exert weaker control over the place fluence the activity of a place cell. Third, hippocam-
field. Place fields are also controlled by other sensory pal area CA1 contains cells that apparently respond
modalities. Many hippocampal neurons respond to when experience is incongruent with predictions
distinct odors and are influenced by kinesthetic and from memory, such as when a salient stimulus sud-
vestibular cues generated by the rats own movement. denly appears or disappears at a particular location.
These influences are particularly powerful when visu- The existence of these mismatch-responsive cells
al input is unavailable or less reliable. When a rat is means that CA1 cells may simultaneously receive in-
released from a closed start box at an unpredictable formation from the senses and from memory.
location, for example, place fields are controlled by
the amount of movement for the first few seconds, be- Once a place cell has developed a localized firing
fore external visual landmarks take over. pattern in a new environment, the place field remains
stable for weeks or more, as predicted if the cell con-
Place cells, however, do not passively mirror the tributes to a particular spatial memory. Some re-
sensory input that the animal receives. First, place searchers have suggested that the formation of place
fields develop slowly. When a rat enters a new envi- fields, like spatial memory, depends on long-term po-
ronment for the first time, place fields are weak and tentiation (LTP) in hippocampal excitatory synapses.
dispersed. Sharp and distinct place fields develop Blockade of the NMDA receptor abolishes both asso-
only after five to ten minutes, at a much slower rate ciative LTP and overnight stability of new place fields
than that at which sensory information passes to the in a new environment. However, the development of
hippocampus. Second, changes in place fields are not place fields is not disrupted, and new place fields can
predictable from the amount of change in sensory be maintained for at least one to two hours in the ab-
input. Place fields often remain in place after signifi- sence of NMDA receptors, suggesting that LTP-
cant landmarks are removed from the environment. independent mechanisms may be essential only for
At other times, two apparently identical environ- the long-term maintenance of place fields.
532 PLACE CELLS
Place Cell Ensembles nal cortex. The sharpest fields are in the hippocam-
pus proper; firing fields in the subiculum and
Memory operations are reflected in the firing
entorhinal cortex are more dispersed. Place-related
properties of individual hippocampal neurons; a
activity in the hippocampus may result from sequen-
more complete understanding of how place cells con-
tial processing along the trisynaptic circuit of the hip-
tribute to spatial memory requires insight into the
pocampus (dentate gyrus, CA3, CA1), or the relevant
constantly changing interaction between large num-
information is carried by the direct excitatory input
bers of place cells. Two phenomenaremapping and
from entorhinal cortex to each subfield. Disruption of
reactivationillustrate the dynamic organization of
the trisynaptic circuit by selective lesions in the den-
place cells at the ensemble level.
tate gyrus or CA3 does not abolish place fields in CA1,
Place fields of simultaneously recorded pyrami- suggesting that the direct input is sufficient for estab-
dal cells exhibit all-or-none remapping. The entire lishing and maintaining spatial activity in hippocam-
population of recorded neurons may adopt new firing pal pyramidal neurons (Moser and Paulsen, 2001).
correlates after a change in a single but defining fea-
If place cells participate in neuronal ensembles
ture of an environment, such as the conversion of a
that collectively contribute to memory of location and
square environment to a circular one while all other
other episodic information, these ensembles must
landmarks remain fixed. Some place cells start to fire
somehow be tied together. It has been suggested that
at new locations, others become silent, and previously
place cells are organized as continuous attractor net-
silent cells become active. The original map of place
works where neurons with firing fields at the current
fields is accurately reinstated when the original envi-
or nearby locations are mutually excited, whereas
ronment is restored. This pattern suggests a linkage
those with fields at other locations are inhibited in a
of place cells in functional ensembles that correspond
distance-dependent manner. Distance between place
to distinguishable test environments or test condi-
fields of two pyramidal neurons may be encoded by
tions. Each place cell is likely to participate in multi-
the strength of the connecting synapses. Some re-
ple ensembles that are active at different times.
searchers have suggested that the recurrent network
A second example of ensemble coding is the strik- of area CA3 has attractor properties, but recurrent
ing observation that cells with overlapping place networks in afferent structures such as the entorhinal
fields persist in correlated firing during sleep epi- cortex may have similar capacity. The CA1 lacks the
sodes subsequent to the behavioral session (Suther- internal excitatory connections needed to maintain
land and McNaughton, 2000). Not only the pattern ensemble structure.
of coactivity, but also temporal firing sequences re-
semble those recorded during preceding behavior.
This form of reactivation is temporally specific. It is Conclusion
strongest shortly after the behavioral session and de- Place cells carry strong signals that are expressed
cays with time. Reactivation is particularly associated reliably in large proportions of the hippocampal neu-
with sharp waves, which are bursts of synchronous ac- ronal population, and they are found in a part of the
tivity in hippocampal pyramidal cells during slow- brain that plays clear roles in specific memory opera-
wave sleep and awake rest. These bursts have the cations. With new powerful techniques that allow the
pacity to induce plasticity in downstream areas and study of neuronal computation at the ensemble level,
may be involved in the consolidation of long-term the study of place cells can contribute to our under-
memory in the neocortex (Buzsaki, 1989). Reactiva- standing of the workings of memory and cognition.
tion occurs during REM sleep, too.
There is also temporal organization of pyrami- Bibliography
dal-cell activity in the hippocampus. Hippocampal Best, P. J., White, A. M., and Minai, A. (2001). Spatial processing
networks display characteristic oscillatory activity in the brain: The activity of hippocampal place cells. Annual
during spatial learning, and the timing of spikes rela- Review of Neuroscience 24, 459486.
tive to oscillations in the theta and gamma frequency Buzsaki, G. (1989). Two-stage model of memory trace formation:
A role for noisy brain states. Neuroscience 31, 551570.
bands may carry significant information. Such oscilla-
Eichenbaum, H. (2001). A cortical-hippocampal system for declar-
tions occur during memory processing, but the exact ative memory. Nature Reviews Neuroscience 1, 4150.
significance of temporal firing patterns for memory Moser, E. I., and Paulsen, O. (2001). New excitement in cognitive
formation in the hippocampus is not yet understood. space: Between place cells and spatial memory. Current Opin-
ion in Neurobiology 11, 745751.
OKeefe J., and Dostrovsky, J. (1971). The hippocampus as a spa-
Place Fields and Hippocampal Circuitry tial map. Preliminary evidence from unit activity in the freely
moving rat. Brain Research 34, 171175.
Place cells exist in all subfields of the hippocam- OKeefe, J., and Nadel, L. (1978). The hippocampus as a cognitive
pus and in the dentate gyrus, subiculum, and entorhi- map. Oxford: Clarendon Press.
PLACE VERSUS RESPONSE LEARNING REVISITED IN THE BRAIN 533
Ranck, J.B., Jr. (1973). Studies on single neurons in dorsal hippo- The Plus-Maze Task and the Behaviorist
campal formation and septum in unrestrained rats. I. Behav-
ioral correlates and firing repertoires. Experimental Neurology
versus Cognitivist Debate
41, 461531. A significant empirical battle flared between S-R
Sutherland, G. R., and McNaughton, B.L. (2000). Memory trace behaviorist and cognitive-learning theorists, and in-
reactivation in hippocampal and neocortical neuronal ensem- volved the investigation of behavior in several learn-
bles. Current Opinion in Neurobiology 10, 180186.
ing tasks. One task that employed a plus-maze appa-
Wilson, M. A., and McNaughton, B. L. (1993). Dynamics of the
hippocampal ensemble code for space. Science 261, 1,055 ratus can illustrate these two approaches to
1,058. understanding what animals learn (Tolman, Ritchie,
and Kalish, 1946). The plus-maze is essentially two T-
Edvard I. Moser mazes arranged so that a goal box (e.g., east or west),
can be approached from one of two start boxes (e.g.,
north or south). In one version of the task, rats are
trained over trials to obtain food from a consistently
baited goal box (e.g., west), from the same start box
PLACE VERSUS RESPONSE (e.g., south). According to S-R learning theory, rats
LEARNING REVISITED can learn to approach the baited goal box by acquir-
ing a response tendency (i.e., a specific body turn at
IN THE BRAIN
the choice point). In contrast, according to cognitive
A chief concern of learning and memory researchers learning theory, rats trained in this task learn the
involves determining what is learned in a given situa- place or spatial location of the reinforcer, and this ex-
tion. Edward L. Thorndike (1933) was an early pro- pectation can guide an approach response to the bait-
ponent of the view that animals learn associations be- ed goal box.
tween stimuli and responses (i.e., S-R learning), and Both behaviorist and cognitive learning theories
in his influential law of effect, Thorndike essentially can adequately explain the acquisition of this version
proposed that S-R associations were strengthened by of the plus-maze task. However, a probe trial in which
reinforcement (a satisfying event) and weakened by trained rats are given a trial starting from the oppo-
nonreinforcement (an annoying event). In Thorn- site start box (e.g., north) can assess the type of infor-
dikes laboratory investigation of the learning abilities mation acquired. Rats with knowledge of the spatial
of several animal species, the learning curves that he location of the reinforcer should continue to ap-
observed were gradual; he therefore argued that proach the baited goal box on the probe trial (i.e.,
learning was not akin to intuition or the sudden illu- place learning), whereas rats that have learned a spe-
mination of a light bulb. Rather, learning appeared cific body turn should choose the opposite goal box
to be an incremental process of trial-and-error, re- on the probe trial (i.e., response learning). In their
sulting in the eventual acquisition of S-R associations. early research, Tolman and colleagues demonstrated
Thorndikes early S-R learning theory and the rise of that rats trained in a plus-maze to approach a goal
James B. Watsons psychological behaviorism influ- box from the same start point on each trial could in-
enced the subsequent work of the psychologist Clark deed display place learning when probe trial behavior
L. Hull, who introduced fairly elaborate mathemati- was later assessed. In addition, rats can acquire place
cal formulas in his description of a S-R habit learning learning in a version of the plus-maze task in which
theory (Hull, 1943). they are trained to approach the same goal arm (e.g.,
west) from two different starting points (e.g., north
Edward C. Tolman was an early proponent of a and south). In this version of the task, rats are re-
different theoretical approach to understanding quired to make a different body turn at the maze
learned behavior. Tolman (1932) argued that S-R choice point (i.e., left or right), depending on the
theory did not adequately explain all learning phe- start position; therefore response learning would be
nomena; he suggested instead that animals acquire ineffective for acquiring the task.
expectations about how various behaviors would lead Taken together, these demonstrations of place
to a desired goal. Tolman used terms such as inference, learning in the plus-maze suggest that S-R theories
intention, and purpose to explain learned behavior. His may not adequately explain all types of learning.
views on animal learning were perhaps most clearly However, subsequent research revealed that in addi-
defined in his hypothesis that animals form a cognition to place learning, rats could also use response
tive map of the environment in which spatial relation- learning in acquiring plus-maze behavior. For exam-
ships among multiple stimuli are represented in ple, under some experimental conditions, rats
memory and could be used to guide goal-directed be- trained to approach the same goal arm (e.g., west)
havior. from a consistent starting point (e.g., south), tended
534 PLACE VERSUS RESPONSE LEARNING REVISITED IN THE BRAIN
to display response learning on a subsequent probe quired the task using place information or had
trial. Moreover, rats can acquire response learning in learned a specific body-turn response. Prior to the
a version of the plus-maze task in which they are probe trial, rats received intrahippocampal or intra-
trained from two different starting points (e.g., north caudate infusions of a vehicle solution or lidocaine, an
and south) and are required to make a consistent anesthetic that produces a temporary inactivation of
body turn response (e.g., turn left) at the maze choice neural function in the affected brain region. On the
point. In an influential review of the plus-maze litera- probe trial, rats receiving vehicle infusions into the
ture, Restle suggested that the relative use of place hippocampus or caudate-putamen were predomi-
and response learning depends on various experi- nantly place learners. Lidocaine infused into the hip-
mental factors, most critically the availability of intra- pocampus blocked expression of place learning,
and extra-maze cues (Restle, 1957). For example, en- whereas similar infusions into the caudate-putamen
vironments in which various distal extra-maze cues did not. Therefore, the functional integrity of the hip-
are present favor place learning, whereas the use of pocampus but not caudate-putamen is necessary for
sparsely cued extra-maze environments favor re- the expression of place learning. Following extended
sponse learning. training in the plus-maze, rats given a second probe
trial switch from place learning to response learning
(Ritchie, Aeschliman, and Pierce, 1950; Hicks, 1964).
Neurobiology of Memory Organization and Therefore, in the study by Packard and McGaugh
the Behaviorist versus Cognitivist Debate (1996), the rats were trained for an additional seven
Although the plus-maze task appeared to hold days, given a second probe trial on the sixteenth day,
early promise for resolving the dispute between be- and again received intracerebral infusions of lido-
haviorists and cognitive theorists in favor of one theo- caine prior to the probe trial. On this second probe
retical viewpoint, the findings of various plus-maze trial rats receiving vehicle infusions into either the
studies clearly indicated that brain-intact rats are ca- hippocampus or caudate-putamen were predomi-
pable of both place and response learning. However, nantly response learners, providing evidence of a
contemporary behavioral neuroscience research sug- switch from place to response learning tendencies
gests a possible resolution of this debate. Research with extended training.
has revealed that mammalian memory is organized in
On the second probe trial, intrahippocampal in-
relatively independent brain systems that differ in the
fusions of lidocaine did not block the expression of
types of memory they mediate (Hirsh, 1974; OKeefe
response learning. However, rats receiving intracau-
and Nadel, 1978; Cohen and Squire, 1980; Mishkin
date infusions of lidocaine prior to the second probe
and Petri, 1984; Eichenbaum and Cohen, 2001).
trial exhibited place learning, demonstrating a block-
There is evidence that the hippocampus is part of a
ade of the expression of response learning. Taken to-
memory system that mediates cognitive memory,
gether, these findings demonstrate a double dissocia-
whereas the caudate-putamen is part of a memory sys-
tion between the roles of the hippocampus and
tem that mediates stimulus-response or habit memo-
caudate-putamen in place and response learning, re-
ry (e.g., Packard, Hirsh, and White, 1989; Fernandez-
Ruiz et al., 2001). The multiple-memory-systems hy- spectively. Moreover, when the shift from the use of
pothesis raises the interesting possibility that place place to response learning occurs, the hippocampus-
and response learning may in fact have distinct neural dependent place representation is not extinguished
substrates, suggesting that a neurobiologically based or forgotten. Rather, at a time in training in which an-
approach may help to address the differing view- imals predominantly use response learning, the place
points of S-R and cognitive-learning theorists (Mish- representation can be brought back into use or un-
kin and Petri, 1984). masked by a blockade of the caudate-putamen re-
sponse learning system.
The shift in the use of a hippocampus-dependent
Multiple Memory Systems and the Place place strategy to a caudate-dependent response strat-
versus Response Learning Debate egy suggests that in a learning task in which both
A plus-maze study was designed to differentiate memory systems can provide an adequate solution,
the mnemonic roles of the hippocampus and caudate- the hippocampal system mediates a rapid cognitive
putamen (Packard and McGaugh, 1996). In this study form of learning that initially guides behavior, where-
rats were trained in a daily session to obtain food from as the caudate-putamen mediates a more slowly de-
a consistently baited goal box and were allowed to ap- veloping S-R or habit form of learning that eventually
proach this maze arm from the same starting box on guides learned behavior. This raises the intriguing
each trial. Following seven days of training, rats were possibility that infusions of memory-enhancing drugs
given a probe trial to determine whether they had ac- into these two brain structures during early training
PREFRONTAL CORTEX AND MEMORY IN PRIMATES 535
might influence the time-course of this shift. In an ex- tional Academy of Sciences of the United States of America 98,
periment designed to address this possibility, rats re- 4,1964,201.
ceived posttraining intrahippocampal or intracau- Hull, C. L. (1943). Principles of behavior. New York: Appleton-
Century-Crofts.
date infusions of the amino acid neurotransmitter Mishkin, M., and Petri, H. L. (1984). Memories and habits: Some
glutamate during early time points in cross-maze implications for the analysis of learning and retention. In L.
training (Packard, 1999). As observed previously, rats R. Squire and N. Butters, eds., Neuropsychology of memory. New
receiving saline control injections predominantly dis- York: Guilford.
played place learning on an early (day eight) probe OKeefe, J., and Nadel, L. (1978). The hippocampus as a cognitive
map. Oxford: Oxford University Press.
trial, and response learning on a later (day sixteen) Packard, M. G. (1999). Glutamate infused posttraining into the
probe trial. However, rats receiving posttraining intr- hippocampus or caudate-putamen differentially strengthens
ahippocampal infusions of glutamate predominantly place and response learning. Proceedings of the National Acade-
displayed place learning on both the early and late my of Sciences of the United States of America 96, 12,88112,886.
probe trials, suggesting that infusion of glutamate Packard, M. G., Hirsh, R., and White, N. M. (1989). Differential ef-
fects of fornix and caudate nucleus lesions on two radial maze
into the hippocampus strengthened a place learning tasks: Evidence for multiple memory systems. Journal of
representation and prevented the shift to response Neuroscience 9, 1,4651,472.
learning that occurs with extended training. In con- Packard, M. G., and McGaugh, J. L. (1996). Inactivation of the hip-
trast, rats given posttraining glutamate infusions into pocampus or caudate nucleus with lidocaine differentially af-
the caudate-putamen predominantly displayed re- fects expression of place and response learning. Neurobiology
of Learning and Memory 65, 6572.
sponse learning on both the early and late probe tri- Restle, F. (1957). Discrimination of cues in mazes: A resolution of
als, suggesting that infusion of glutamate into the cau- the place versus response controversy. Psychological Review 64,
date-putamen accelerated the shift to response 217228.
learning that is normally observed with extended be- Ritchie, B. F., Aeschliman, B., and Pierce, P. (1950). Studies in spa-
havioral training. tial learning: VIII. Place performance and acquisition of place
dispositions. Journal of Comparative and Physiological Psychology
The findings from behavioral neuroscience re- 43, 7385.
search employing brain lesion and intracerebral drug Thorndike, E. L. (1933). A proof of the law of effect. Science 77,
infusion techniques provide a partial neurobiological 173175.
resolution of the place versus response learning de- York: Appleton-Century-Crofts.
bate. Whereas previous research clearly demonstrates
that brain-intact animals are capable of both place Mark G. Packard
and response learning (Restle, 1957), such behaviors
do not reflect a single learning and memory system.
Rather, distinct neuroanatomical substrates that in-
clude the hippocampus and caudate-putamen medi- PREFRONTAL CORTEX AND
ate the acquisition of place and response learning, re- MEMORY IN PRIMATES
spectively. Thus, four decades after the introduction
of the plus-maze task as a means of addressing the A primate needs its prefrontal cortex for behavior
fundamental question of what animals learn in a based on information accumulated before the mo-
given situation, the mammalian brain appears to have ment of action. The prefrontal cortex is especially im-
spoken in favor of the viewpoint that the historic de- portant if the information is new to the organism or
bate between S-R and cognitive learning theorists conflicts with prior cues or memories that call for dif-
may in part have been misguided in to impose a sin- ferent actions. Whether the information is new or old,
gle theoretical viewpoint on all types of learning. it must be retained in memory until the moment it
can inform an act. It is the prefrontal cortex that sup-
See also: GUIDE TO THE ANATOMY OF THE BRAIN: ports this short-term memory (also called working
HIPPOCAMPUS AND PARAHIPPOCAMPAL REGION memory) that subserves behavior.
Prefrontal memory is a matter not of content or
Bibliography duration but of context, the context of action. Short-
Blodgett, H. C., and McCutchan, K. (1947). Place versus response term memory is not the only function of the prefron-
learning in the T-maze. Journal of Experimental Psychology 37,
412422. tal cortex, nor is it exclusively the role of the prefron-
Cohen, N. J., and Eichenbaum, H. (1993). Memory, amnesia, and the tal cortex. But, insofar as action requires short-term
hippocampal system. Cambridge, MA: MIT Press. memory, it also needs the prefrontal cortex. It follows
Cohen, N. J., and Squire, L. R. (1980). Preserved learning and re- that this part of the cortex is essential for the con-
tention of pattern analyzing skill in amnesics: Dissociation of struction of sequential behaviors, especially if they are
Fernandez-Ruiz, J., Wang, J., Aigner, T. G., and Mishkin, M. novel or require choices. It is needed for the syntax
(2001). Visual habit formation in monkeys with neurotoxic le- of the action, including, of course, the syntax of the
sions of the ventrocaudal neostriatum. Proceedings of the Na- spoken language, particularly creative speech.
536 PREFRONTAL CORTEX AND MEMORY IN PRIMATES
The prefrontal cortex is the cortex of the pole of the cue must be retained. Furthermore, dorsolateral
the frontal lobe. It is that part of the cerebral cortex prefrontal lesions also impair performance of delay
to which the nucleus mediodorsalis of the thalamus tasks in which the sensory cue is not spatially defined,
projects. Phylogenetically, it is the neocortical region a correlation that has been demonstrated by local cor-
that undergoes the greatest and latest expansion (see tical cooling. The cryogenic depression of a large por-
Figure 1). It reaches maximum development in the tion of dorsolateral cortex (area 9), including the sul-
brain of the human, where it occupies almost one- cus principalis, induces a reversible deficit in
third of the totality of the neocortex. In the course of performance of delay tasks, whether the cue is visual
evolution, its dorsolateral aspect, that is, the cortex of and spatially defined (as in DR) or not (as in delayed
the external convexity of the frontal lobe, develops matching to sample). Furthermore, the cryogenic
relatively more than its medial and inferior aspects. deficit also affects delayed matching tasks in which
This is an important consideration because the dorso- the cue (sample) is perceived by active touch (hapti-
lateral prefrontal cortex supports mainly cognitive cally).
functions, whereas the orbitomedial prefrontal cortex Humans with dorsolateral prefrontal lesions also
mostly pertains to emotional and visceral functions. show impairments in delay tasks. Tasks in which the
Cytoarchitectonically, the prefrontal cortex of the material to be retained can be verbally encoded are
primate (area FD of Von Bonin and Bailey, 1947) in- affected more by lesions of the left than of the right
cludes areas 9, 10, 11, 12, and 13 of Brodmann prefrontal cortex. They are impaired most of all by bi-
(1909). It is one of the best-connected of all neocorti- lateral lesions. The human prefrontal syndrome usu-
cal regions; it is directly and reciprocally connected ally involves disorders of attention, planning, and
to the anterior and dorsal thalamus, the hypothala- language. Prefrontal patients have difficulty main-
mus, and limbic structures, especially the amygdala taining attention on internal cues or short-term re-
tention of mental material. Their planning is poor for
and the hippocampus. It sends profuse efferent fibers
both the short and the long term; it is as if they lacked
to the basal ganglia. Dorsolateral prefrontal areas
memory of the future in addition to memory of the
have rich reciprocal connections with many other
recent past. Both retrospective and prospective rep-
neocortical areas of the frontal lobe and of the tempo-
resentations are needed for the sequential construc-
ral and parietal lobes. The first clear indication of the
tion not only of external motor action but also of in-
involvement of the prefrontal cortex in short-term
ternal action, such as sequential logical thinking
memory was provided by Jacobsen in the early 1930s
hence the trouble the patients have in this kind of
(Jacobsen, 1935). He showed that monkeys with le-
activity, whether expressed in spoken language or
sions of the dorsolateral prefrontal cortex are im-
not. Speech is most impaired if it requires the bridg-
paired in the learning and performance of delayed
ing of long intervals (cross-temporal contingencies)
response (DR) and delayed alternation (DA) tasks.
between subjects and verbs, subjects and predicates,
These tasks fall within a general category of behavior-
or logically interdependent sentences. The trouble is
al tasksdelay tasksthat demand from the animal
extreme in lesions of Brocas area, which is a part of
the performance of motor acts in accord with sensory
the prefrontal cortex specialized in the most elemen-
information presented a few seconds or minutes earli-
tary aspects of linguistic syntax. Animal and human
er. In other words, delay tasks demand short-term
neuropsychology thus suggests that the prefrontal
memory for the logical and consequent bridging of
cortex is essential for bridging cross-temporal contin-
temporal gaps between perception and action, the
gencies of behavior, and that this is so at least in part
mediation of cross-temporal contingencies of behav-
because of its role in short-term memory.
ior. Primates deprived of substantial portions of dor-
solateral prefrontal cortex cannot properly perform Microelectrode recording in the monkey has cor-
delay tasks, regardless of the nature of the sensory in- roborated the role of the prefrontal cortex in short-
formation that guides them, especially if after a long term memory. Prefrontal neurons show sustained ac-
delay between sensory cue and motor response. tivation of firing during the delay periods of delay
tasks (see Figure 2). Statistical analysis and control ex-
There appears, however, to be some specificity of periments indicate that this activation is
prefrontal areas with regard to the type of sensory in-
a result of learning the task;
formation they help retain. Lesions of the cortex of
the sulcus principalis are most detrimental to perfor- dependent on the presence of a cross-temporal
mance of delay tasks with spatially defined sensory contingency between cue and motor response;
cues, such as DR and DA. However, time seems to related in some cells to the property of the cue on
override space on this matter. Those spatial tasks are which the response depends;
impaired only when a delay occurs between cue and related in some cells to the response that the ani-
response. The critical factor is the time during which mal has to execute at the end of the delay;
PREFRONTAL CORTEX AND MEMORY IN PRIMATES 537
Figure 1
538 PREFRONTAL CORTEX AND MEMORY IN PRIMATES
directly related to the efficacy with which the ani- ry that subserves action. Thus, it is critically important
mal performs the task. for sequential behaviors with cross-temporal contin-
gencies. In the human, such behaviors include the
All these indications are consistent with the as- spoken language and logical thought, which is a form
sumption that prefrontal neurons participate in corti- of sequential inner action. The memory function of
cal networks that, by their sustained activation, retain the prefrontal cortex is one of several functions that
sensory information as long as needed for prospective this cortex supports in cooperation with other asso-
action. ciative cortical areas. These functions are essential for
the syntax of actionfor organization of behavior in
The analysis of single-unit discharge in other
the time domain.
areas of association cortex (inferotemporal and poste-
rior parietal) during visual and haptic delay tasks has See also: GUIDE TO THE ANATOMY OF THE BRAIN;
revealed sustained neuronal activations in those WORKING MEMORY: HUMANS
areas. This observation, in addition to the study of the
effects of local cortical cooling on remote cortical cell Bibliography
discharge and delay-task performance, has led to the Bauer, R. H., and Fuster, J. M. (1976). Delayed-matching and de-
following inference: The prefrontal cortex exerts its layed-response deficit from cooling dorsolateral prefrontal
cortex in monkeys. Journal of Comparative Physiology and Psy-
role in short-term memory and preparation for action
chology 90, 293302.
through close functional interplay with areas of poste- Bonin, G. von, and Bailey, P. (1947). The neocortex of Macaca mulat-
rior association cortex with which it is intimately and ta. Urbana: University of Illinois Press.
reciprocally connected. Thus, the activated cortical Brodmann, K. (1909). Vergleichende Lokalisationslehre der Grosshirn-
network that retains sensory information during be- rinde in ihren Prinzipien dargestellt auf Grund des Zellenbaues.
Leipzig: Barth.
havior has several components in short-term memo-
Fuster, J. M. (1973). Unit activity in prefrontal cortex during de-
ry. Posterior areas are involved in the network inas- layed-response performance: Neuronal correlates of transient
much as the information falls within their special memory. Journal of Neurophysiology 36, 6178.
modality (vision, audition, somesthesis, and so on). (1985). The prefrontal cortex, mediator of cross-temporal
The prefrontal cortex is involved inasmuch as the in- contingencies. Human Neurobiology 4, 169179.
(1997). The prefrontal cortex, anatomy, physiology, and neuropsy-
formation calls for prospective goal-directed action.
chology of the frontal lobe. 3rd edition. Philadelphia: Lippincott-
The maintenance of activity in the network is assured Raven.
by the continuous circulation of excitatory impulses (2001). The prefrontal cortexan update: Time is of the
within and between all contributing areas, including essence. Neuron 2, 319333.
the prefrontal cortex if the outcome is to be a goal- Fuster, J. M., Bodner, M., and Kroger, J. K. (2000). Cross-modal
and cross-temporal association in neurons of frontal cortex.
directed action in accord with the sensory informa-
Nature 405, 347351.
tion. Goldman-Rakic, P. S. (1987). Circuitry of primate prefrontal cor-
tex and regulation of behavior by representational memory.
The prefrontal cortex of the primate is a large In F. Plum, ed., Handbook of physiology, Section 1: The Nervous
and phylogenetically new part of the cerebral cortex System, Vol. 5. Bethesda, MD: American Physiological Associa-
that performs a critical function in short-term memo- tion.
PRIMATES, VISUAL ATTENTION IN 539
Jacobsen, C. F. (1935). Functions of the frontal association area in Control of Attention

primates. Archives of Neurology and Psychiatry 33, 558569.
The selective filters that underlie selective atten-
Joaquin M. Fuster tion operate in space and on features of objects, se-
lecting locations and features of visual stimuli for
heightened processing. The neurobiology underlying
this process lends itself to separation into two parts:
the mechanisms that control attention and the mech-
PRIMATES, VISUAL ATTENTION IN anisms that are, in turn, influenced by attention. In
some cases, attention is controlled by features intrin-
The neural mechanisms for visual attention intervene sic to a stimulus. Some objects attract attention to
between sensation and action. Neither strictly sensory themselves, an automatic process that occurs because
nor motor, they constitute the neural filters that select of separation of figures from their background result-
certain aspects of visual information for greater atten- ing from differences in color, contrast, motion (pop-
tion and influence over action. Visual attention in- out). Yet, even after automatic figure-ground sepa-
cludes mechanisms for arousal, alerting, vigilance, ration, a typical visual scene will still contain many dif-
and selective attention for spatial location or visual ferent figures of equal salience that are nevertheless
features. Of these, the ones that most directly affect of different behavioral relevance. In this case, atten-
visual processing are selective attention to a location tion can be actively wielded to filter information in
in space or to a visual feature of an object. These latter the scene; a neural signal that reflects this active pro-
mechanisms help route selected visual inputs to visu- cess is a top-down signal that controls attention.
al-processing structures in the brain and also to motor
control systems, usually at the expense of nonselected Studies using functional imaging in human sub-
inputs. Behavioral examinations of visual processing, jects have identified several target structures that may
functional-imaging studies in humans, and single- provide the signals that control selective visual atten-
unit studies in the nonhuman primate indicate that tion in humans. The control system for spatially di-
selection mechanisms are in play. These experiments rected attention, identified by functional imaging, is
aim to understand the influence of attention on the both widely distributed and closely associated with the
processing of information and the neural signals that oculomotor system; it involves the posterior parietal
underlie selective attention. cortex, parts of the frontal cortex, the pulvinar nucle-
us of the thalamus, and possibly the superior col-
liculus. The close association with the oculomotor sys-
Behavioral Experiments tem accords with the principle that an important role
Behavioral experiments show that individuals for selective attention is the direction of the eyes to
cannot and do not process all the information present the behaviorally relevant stimuli in complicated visual
in complex visual scenes. It is not possible, for exam- scenes. Neurophysiology in nonhuman primates per-
ple, to recognize within the same instant more than forming attention tasks indicates a neural mechanism
one or two objects in a typical complex scene. When for the operation of attention in some of these struc-
subjects are asked to report different attributes of si- tures. In these structures, a given neuron responds to
multaneously presented images, they nearly always stimuli in a limited region of space (i.e., its receptive
perform worse than when asked to report the attri- field), and this response is enhanced whenever the
butes of the objects separately. In certain conditions, animal directs its attention into the region. This re-
major changes in scenes go undetected because the sponse enhancement may provide the top-down
capacity of the brains limited capacity to code and signal that potentiates the processing of visual stimuli
process complex information. But, when subjects are that appear in that location. When these signals are
told to attend to only one of two simultaneously pres- absent because of damage or dysfunction within any
ented objects, or to a particular location in a scene, of these structures, particularly if the damage is uni-
their performance returns to the levels expected for lateral, subjects suffer a variety of attentional disor-
only one object. Selective attention provides a mecha- ders. One such attentional disorder is extinction, a
nism for choosing which information is noticed and difficulty in detecting or perceiving objects within the
which is ignored, thus ensuring that unimportant in- affected portion of the visual field when a competing
formation does not crowd out important information. object is located in intact portions of the field.
Moreover, since visual acuity is poor outside of the
center of gaze, the eyes must be moved to scan and
process visual scenes. Selective attention provides a Influence of Attention
method for selecting single targets for the neural sys- The neural signals derived from these brain areas
tem that controls the eyes (oculomotor system). influence the processing of visual stimuli in the senso-
540 PRIMATES, VISUAL PERCEPTION AND MEMORY IN NONHUMAN
ry areas that process visual information. Functional Reynolds, J. H., and Desimone, R. (1999). The role of neural
imaging studies show that attention influences the mechanisms of attention in solving the binding problem. Neu-
ron 24, 1929, 11125.
processing of visual information in all visual areas,
from the earliest visual area, primary visual cortex to Robert Desimone
the last visual areas, in both the ventral and dorsal vi- Revised by Bharathi Jagadeesh
sual processing streams. In these imaging studies, the
instruction to attend to a stimulus can modulate the
response to visual stimuli, altering the response to a
visual stimulus by 10 to 30 percent. This effect is larg- PRIMATES, VISUAL PERCEPTION
er in later stages of visual processing, raising the pos- AND MEMORY IN NONHUMAN
sibility that the smaller modulations in primary visual Anatomical and physiological studies have revealed at
cortex result from feedback from higher areas. The least thirty separate areas with visual functions in the
results in human imaging replicate those in neuro- cortex of monkeys, and there could be even more vi-
physiological recordings of behaving primates, al- sual areas in the cortex of humans. One view of the
though the influence of attention on responses in division of labor among the visual areas holds that
early visual areas in primates has not been demon- there are two main processing systems: one devoted
strated in single-unit studies. But extrastriate visual to identifying what an object is and the other devoted
cortical areas that process information about object to identifying where an object is (Ungerleider and Mis-
features such as color and shape (such as Visual Area hkin, 1982). On this view, the occipitotemporal cor-
4), receive attentional control signals, and act as the tex is mainly the province of the former system,
local attentional filters. The properties of the neu- known as the ventral stream because of the idea that
rons responses to visual stimuli and the effect of at- visual information flows, in some sense, from occipital
tention reflect both behavior in the absence of atten- cortex forward and ventrally toward the temporal cor-
tion and the influence of attention on that behavior. tex (see Figure 1). The other system, known as the
When more than one stimulus is placed in the recep- dorsal stream because information is held to flow
tive field of a single neuron, the neurons activity re- from occipital to parietal cortex, seems to play a role
flects an averaging of the response to both of the stim- in either spatial perception, as noted above, or alter-
uli. natively, in the guidance of movements to spatial tar-
The presence of more than one stimulus in the gets (Milner and Goodale, 1996). This article will
receptive field interferes with that neurons ability to focus on the ventrally directed, occipitotemporal pro-
carry information about the stimulus. Behaviorally, cessing stream and its role in object perception and
this interference is mirrored in poor performance memory.
with simultaneously presented objects. But when an
animal is instructed to attend to a single stimulus Anatomy of the Occipitotemporal Pathway
within the receptive field of the neuron, the response for Object Identification
of the neuron will be determined almost solely by the
color, shape, or orientation of the attended stimulus. The occipitotemporal pathway begins with the
In presence of the top-down signal wielded by atten- projection from the striate cortex (the primary visual
tion, neurons communicate little information about cortex, V1) to the second and third visual areas, V2
the features of ignored stimuli, and the interference and V3, which in turn project to area V4 (see Figure
is filtered out of the signal. Thus, the neurons that 1). These prestriate visual areas are arranged in adja-
process the visual information are influenced by the cent cortical belts that nearly surround the striate cor-
control signals for attention to act as filters of infor- tex. The major output of area V4 is to a widespread
mation that select relevant information and discard region within the inferior temporal cortex, including
irrelevant information, providing a mechanism for area TEO posteriorly and area TE anteriorly. The in-
sorting out the complexity of the visual world. ferior temporal cortex is in receipt of highly pro-
cessed visual information arising not only from lower-
order visual areas but also from subcortical structures,
Bibliography including the thalamus and basal ganglia. Area TE is
Kanwisher, N., and Wojciulik, E. (2000). Visual attention: Insights often considered to be the last, or highest-order, visual
from brain imaging. Nature Reviews Neuroscience 1, 91100.
Kastner, S., and Ungerleider, L. G. (2000). Mechanisms of visual area in the cortical system for object identification be-
attention in the human cortex. Annual Review of Neuroscience cause its principal cortical outputs are to areas in the
23, 315341. temporal and frontal lobes that are not exclusively
Olson, C. R. (2001). Object-based vision and attention in primates. concerned with vision.
Current Opinion in Neurobiology 11, 171179.
Rensink, R. A. (2002). Change detection. Annual Review of Psycholo- Murray and Bussey (1999) have argued that per-
gy 53, 245277. irhinal cortex, a multimodal region that lies adjacent
PRIMATES, VISUAL PERCEPTION AND MEMORY IN NONHUMAN 541
to area TE, should be considered a ventral extension

of the ventral visual stream. The perirhinal cortex re-
ceives anatomical connections not only from area TE
but also from brain areas that process auditory, so-
matic sensory, and visuospatial information. Al-
though perirhinal cortex is not solely dedicated to the
processing of visual sensory inputs, it may neverthe-
less carry out a higher-order level of visual processing
than TE, perhaps by representing even more com-
plex conjunctions of stimulus features than area TE,
together with other kinds of sensory inputs. In this
way, the perirhinal cortex may bring together dispa-
rate pieces of information about objects, including
their associated attributes.
Neurons in the visual cortex see, or represent,

pieces of our visual world. The amount of visual space
that a neuron represents, or is responsive to, is the vi-
sual receptive field. In striate and prestriate areas the
representation in the cortex is maplike or visuotopic;
that is, nearby parts of visual space are represented
in nearby parts of a cortical area in a systematic map-
ping. The representations in these fields are also re-
stricted to the contralateral visual field. In area TEO
of inferior temporal cortex, this visuotopic organiza-
tion seems to be coarser than that in earlier fields; in
area TE it may be nonexistent. In area TE, in contrast
to other visual fields, neurons have large receptive
fields that often cross the midline. Thus, neurons in
area TE can see an object regardless of its position in
the visual field.
Also in the temporal cortex are at least two areas

involved in processing motion vision: the middle tem-
poral and middle-superior temporal areas, abbreviat-
ed as MT and MST, respectively (not illustrated).
These areas are in the depths of the superior tempo-
ral sulcus, near the temporal-parietal-occipital junc-
tion, and are often considered a third visual stream,
contrasting motion vision with the object vision of the
ventral stream and the spatial vision of the dorsal
stream.
Although much of the neural mechanism for ob-

ject identification can be viewed as a bottom-up pro-
Neuronal Properties in the Visual Areas of
cess, in which low-level inputs are transformed into a the Occipitotemporal Pathway
more useful representation through successive stages Given the nearly sequential anatomical route to
of processing, anatomical studies have shown that the inferior temporal cortex from V1, one would ex-
each of the feedforward projections between succes- pect many types of visual information relevant to ob-
sive pairs of areas in the occipitotemporal pathway is ject identification to be processed within each area
reciprocated by a feedback projection. Such projec- along the route. Indeed, physiological studies have
tions from higher-order processing stations back to shown that neurons in V1, V2, and V3 are sensitive to
lower-order ones could mediate some top-down as- one or (usually) more stimulus qualities that we per-
pects of visual processing, such as the influence of se- ceive, such as size, orientation, spatial frequency, tex-
lective attention. ture, color, direction of motion, and binocular dispar-
542 PRIMATES, VISUAL PERCEPTION AND MEMORY IN NONHUMAN
ity. Neurons in the motion pathway reflect precisely the part of the visual field represented in the dam-
the judgements reported by monkeys about the direc- aged area. The effects of selective lesions of V2 and
tion in which a field of light spots is moving on aver- V3 on visual perception and memory have not yet
age. Furthermore, exciting cells by passing electrical been investigated. As described below, the impair-
current through an electrode near them systematical- ments that follow lesions of areas V4, TEO, TE,
ly affects a monkeys perceptual judgement. For ex- and perirhinal cortex are consistent with a contribu-
ample, if a group of neurons represents upward motion from these areas to the object-identification pro-
tion, then stimulating these cells during a period cess.
when the light spots are tending to move to the left
will induce the illusion that they are moving slightly Lesions of Area V4
up as well as left. Finally, in the perirhinal cortex and
Several studies in monkeys have examined the
in the neighboring area TE, neurons respond less to
contribution of V4 lesions to color and form vision.
familiar than to novel visual stimuli. Thus, these neu-
rons carry signals about whether objects have been Although one reported that the lesion impaired color
seen before (recognition memory). In addition, neu- constancy but not hue or form discrimination, subse-
rons in this region also reflect the relative recency of quent studies found impairments in both hue and
viewing of objects. form discrimination, which is consistent with the no-
tion that V4 processes both color and form informa-
It is possible to view the areas along the occipi- tion and relays this information to the inferior tempo-
totemporal pathway as forming a processing hierar- ral cortex.
chy. As one moves along the pathway from V1 to per-
irhinal cortex, both neuronal response latencies and Lesions of Areas TEO, TE, and Perirhinal
average receptive field size increase steadily, which is Cortex
consistent with the notion that neuronal responses in
Damage to the inferior temporal cortex, includ-
later areas are built up from those in earlier areas. A
hierarchical model further predicts that the product ing areas TEO, TE, and the perirhinal cortex, leads
of visual processing will become more complex at to deficits in the identification of objects through vi-
each successive stage in the pathway. So far, the re- sion. Behavioral tasks designed to test visual abilities
sults of anatomical and physiological studies support in monkeys indicate that inferior temporal cortical
this idea. damage produces difficulties in discriminating and in
matching objects on the basis of vision. If the damage
Although early findings led some theorists to pro- is restricted to areas TE and TEO, objects can still be
pose that neurons in the anterior ventral temporal identified by touch. Because basic sensory capacities
cortex code specific objects within their large, bilater- such as acuity and color vision are largely intact after
al receptive fields, there is overwhelming evidence
inferior temporal cortical damage, the difficulty is
that neurons in this region, especially in area TE and
considered to be one of higher-order vision. The na-
perirhinal cortex, respond selectively to object fea-
ture of the impairment caused by inferior temporal
tures such as shape, color, and texture rather than to
cortex lesions is controversial. One idea is that dam-
specific objects. Thus, the neural code for object rec-
ognition and identification must be a population code age to this region results in a loss of the ability to rep-
based on stimulus features. The only exception ap- resent features (e.g. the color red), in such a way that
pears to be the small proportion of neurons selective the greater the damage, the fewer visual features that
for faces and, more rarely, hands. Why should the can be represented. Another idea is that it is not the
coding of faces be treated differently from the coding ability to represent features that is lost, but, rather,
of other objects? One possibility is that faces are ex- the ability to represent conjunctions of features. Both
tremely important to primates not only for the recog- these ideas emphasize the role of inferior temporal
nition of individuals in the social group but also for cortex in visual perception, that is, its role in repre-
social communication by facial expression. Thus, senting visual features. On either view, damage to in-
there may have been selective pressure to evolve spe- ferior temporal cortex will lead to disruption of
cialized neural mechanisms for the analysis of faces stored representations of features, resulting in a loss
and facial expression. of long-term visual memory. At the same time, this
damage would result in an impaired ability to repre-
sent features, a perceptual deficit. Thus, it appears
Effects of Lesions in the Visual Areas of the that neurons in the inferior temporal cortex contrib-
Occipitotemporal Pathway ute to both perception and memory and that percep-
Numerous studies have shown that lesions of V1 tual and mnemonic functions are anatomically insep-
produce a scotoma, or blind spot, that corresponds to arable in this part of the brain.
PRIMATES, VISUAL PERCEPTION AND MEMORY IN NONHUMAN 543
Visual Attention tures. Some accounts have questioned the memory

interpretation and have argued instead for a unified
Contributions of ventral stream areas to object
theory of the contribution of the inferior temporal
identification typically are examined by presenting
cortex to perception and memory.
objects against a blank background. In a naturalistic
visual scene, however, the visual system must identify
and select objects from a multitude of objects, a pro- Interactions of the Occipitotemporal
cess termed attention. Attention appears subject to at Pathway with the Limbic System
least two influences: the strength of the sensory sig-
In some instances, storage of visual memories
nals processed by the ventral stream areas and the
must occur through the interaction of the visual cor-
(learned) relevance of a particular object. Monkeys
tex with medial temporal-lobe limbic structures (see
with combined lesions of V4 and area TEO are defi-
Figure 1). It appears that the inferior temporal cortex
cient in directing attention to objects in the visual
is critical for storing knowledge about objects, the
field when there are multiple, distracting objects, and
hippocampus for knowledge about places and events,
their difficulty increases when the distracting stimuli
and the amygdala for linking object, event, or place
are made more salient by increasing their contrast
information with affective valence. Thus, the system
(De Weerd, Peralto, Desimone, and Ungerleider,
might be organized in a hierarchical fashion with the
1999).
inferior temporal cortex providing an initial stage of
processing involving object recognition, association,
Visual-Recognition Memory and identification. In later stages, already-processed
information about objects would be linked with events
In monkeys the delayed nonmatching-to-sample or affective valences by the hippocampus and amyg-
(DNMS) task has proved a reliable measure of visual dala, respectively. If so, we might expect that, to the
recognition memory for objects. On each trial the ani- extent that the later stages of processing require
mal is shown a sample object, which it displaces in knowledge about objects, damage to the inferior tem-
order to find a food reward underneath; then, about poral cortex, including the perirhinal cortex, would
ten seconds later, the animal is shown both the sam- disrupt storage of information linking objects with
ple and a novel object for choice. The monkey can ob- events and affective valence.
tain another food reward by choosing the novel ob-
ject. When the animal has learned the nonmatching See also: DISCRIMINATION AND GENERALIZATION
rule, the task is made more difficulteither by in-
creasing the delay between the sample presentation Bibliography
and the choice test or by increasing the number of Dean, P. (1982). Visual behavior in monkeys with inferotemporal
lesions. In D. J. Ingle, M. A. Goodale, and R. J. W. Mansfield,
items to be rememberedto tax the monkeys memo- eds., Analysis of visual behavior. Cambridge, MA: MIT Press.
ry. Desimone, R., and Ungerleider, L. G. (1989). Neural mechanisms
of visual perception in monkeys. In F. Boller and J. Grafman,
Lesions of area TE together with much of the ven-
eds., Handbook of neuropsychology, Vol. 2. Amsterdam: Elsevier.
trally adjacent perirhinal cortex lead to severe deficits De Weerd P., Peralto, M. R. III, Desimone, R., and Ungerleider,
on DNMS (Mishkin, 1982). More recent work sug- L. G. (1999). Loss of attentional stimulus selection after ex-
gests that area TE, at least dorsal portions of it, is not trastriate cortical lesions in macaques. Nature Neuroscience 2,
as important for visual-recognition memory as the 753758.
Gaffan, D., Harrison, S., and Gaffan, E. A. (1986). Visual identifica-
perirhinal cortex. Indeed, lesions restricted to the
tion following inferotemporal ablation in the monkey. Quar-
perirhinal cortex lead to more severe impairments in terly Journal of Experimental Psychology 38B, 530.
recognition memory than do lesions to any other part Milner, A. D., and Goodale, M. A. (1996). The visual brain in action.
of the inferior temporal cortex yet investigated. Oxford: Oxford University Press.
Mishkin, M. (1982). A memory system in the monkey. Philosophical
The multimodal nature of the sensory inputs to Transactions of the Royal Society of London B298, 8595.
this region is shown by the deficits in tactile and visu- Murray, E. A. (2000). Memory for objects in nonhuman primates.
al-recognition memory that result from lesions of the In M. Gazzaniga, ed., The Cognitive Neurosciences, 2nd edition.
perirhinal cortex. Monkeys with lesions of perirhinal
Murray, E. A., and Bussey, T. J. (1999). Perceptual-mnemonic
cortex can usually perform the nonmatching task if functions of the perirhinal cortex. Trends in Cognitive Sciences
the delay between the sample and the choice is short, 3, 142151.
but their performance falls nearly to chance levels Newsome, W. T. (1997). The King Solomon lectures in neuro-
when the delays are as long as a minute or two; hence, ethology. Deciding about motion: Linking perception to ac-
tion. Journal of Comparative Physiology A 181, 512.
the deficit seems to be one of memory. However, it is
Squire, L. R. (1987). Memory and brain. New York: Oxford Universi-
clear that the role of perirhinal cortex on visual pro- ty Press.
cessing extends beyond recognition memory and in- Tanaka, K. (1996). Inferotemporal cortex and object vision. Annual
cludes a role for representing conjunctions of fea- Review of Neuroscience 19, 109139.
544 PROCEDURAL LEARNING: Animals
Ungerleider, L. G., and Mishkin, M. (1982). Two cortical visual sys- dural memory involves the acquisition, consolidation,
tems. In D. J. Ingle, M. A. Goodale, and R. J. W. Mansfield, and retrieval of information acquired in tasks involv-
eds., Analysis of visual behavior. Cambridge, MA: MIT Press.
ing various types of Pavlovian classical conditioning
Leslie G. Ungerleider and instrumental/operant learning paradigms. This
Elisabeth A. Murray article describes evidence indicating that procedural
Revised by Elisabeth A. Murray memory involves several, largely separable neuroana-
tomical systems, including the cerebellum, basal gan-
glia, and amygdala.
PROCEDURAL LEARNING
Cerebellum and the Classically
[Procedural learning or procedural memory refers to reten- Conditioned Eyeblink Response
tion of motor skills. The distinction between procedural
memory and declarative memory is usually framed between In the early 1980s Richard Thompson and his
knowing how and knowing that, with the first phrase colleagues (McCormick et al., 1981) demonstrated a
referring to procedural knowledge (knowing how to do some- critical role for the cerebellum in the classically condi-
thing) and the second phrase to declarative knowledge tioned eyeblink response in the rabbit. In this para-
(knowing the nature of events or the world). People know digm animals receive pairings of an auditory condi-
how to ride a bicycle, how to type, or how to find their way tioned stimulus (CSa tone), and an unconditioned
by using a map. On the other hand, people know that Rome stimulus (UCSan air puff delivered to the eye). The
is the capital of Italy, that the attack on Pearl Harbor oc- naturally occurring rabbit eyeblink response is a de-
curred on December 7, 1941, and that Winston Churchill fensive mechanism initially emitted in response to un-
was prime minister of England. conditioned stimulus. However, following several
presentations of the tone-air puff contingency, the
The two entries on procedural learning that follow are animal learns to produce the eye blink (i.e., a condi-
devoted to research conducted in two different but overlap- tioned response) in the absence of the air puff. A se-
ping traditions, one with animal models and the other with ries of elegant studies employing irreversible and re-
human subjects. The former is concentrated on research on versible lesions, neuroanatomical tracing, and
the neural substrates of several conditioning paradigms used electrophysiological techniques have mapped out the
to study learning in ANIMALS. The work with HUMANS is neural circuitry underlying the processing of uncon-
concerned with how patients with various types of brain ditioned and conditioned stimuli in eyeblink condi-
damage learn new skills. Although the experimental tradi- tioning (Kim and Thompson, 1997). This circuitry in-
tions are different, the overarching aim of both is to charac- cludes an unconditioned stimulus pathway in which
terize the neural processes underlying procedural learning the dorsal accessory olive projects through the inferi-
and memory.] or cerebellar penduncle and a conditioned stimulus
pathway involving mossy-fiber projections to the cere-
bellum via the pontine nuclei. The interpositus nucle-
ANIMALS us of the cerebellum receives converging CS-US in-
Neurobehavioral research conducted primarily be- formation, and this deep cerebellar nuclei plays a
tween 1980 and 2000 employed dissociation method- critical role in both the acquisition and expression of
ology to provide compelling evidence for the exis- the classically conditioned eyeblink response. The
tence of multiple memory systems in the mammalian lack of impairment of performance of unconditioned
brain. Although evidence of neuroanatomical dissoci- eyeblink behavior following lesions of the interpositus
ations of the role of various brain structures in differ- nucleus provides critical evidence of a selective role
ent memory tasks is a critical step in advancing the for this nucleus in learned or conditioned eyeblink
multiple-memory systems hypothesis, a full under- behavior. Study of the role of cerebellar circuitry in
standing of memory organization involves elucida- the acquisition and expression of conditioned eye-
tion of the psychological operating principles that blink behavior represents perhaps the best example
distinguish different types of memory. One putative of a model approach to understanding the neuro-
form of memory present in the mammalian brain is biological bases of a form of procedural learning in
declarative memory and involves the acquisition, con- the mammalian brain.
solidation, and retrieval of cognitive information for
facts and events (Cohen and Squire, 1980; Eichen-
baum and Cohen, 2001). Declarative memory ap-
The Basal Ganglia and Stimulus-Response
pears to rely on a neuoranatomical system composed Habit Learning
of various structures in the medial temporal lobe, In a multiple-memory systems approach to mem-
most notably the hippocampus. In contrast, proce- ory organization, there is evidence that components
PROCEDURAL LEARNING: Animals 545
of the basal ganglia (specifically the caudate-putamen The Amygdala and Stimulus-Affect
or dorsal striatum) mediate the acquisition of tasks Conditioning
that involve the formation of stimulus-response habits
The mammalian amygdala has long been impli-
(Mishkin and Petri, 1984; Packard, Hirsh, and White,
cated in the neurobiology of emotional behavior; this
1989). In several studies, the role of the rodent basal function of the amygdala seems to extend to emotion-
ganglia in procedural/habit learning has been dissoci- al learning. Bruce S. Kapp and his colleagues provid-
ated from the role of the hippocampal system in cog- ed some of the early evidence of a role for the amyg-
nitive/declarative memory. An early study examining dala in conditioned fear learning (Kapp, Frysinger,
the selective role of the basal dorsal striatum in S-R Gallagher, and Haselton, 1979). Joseph E. LeDouxs
habit learning involved an experiment using two team has extensively investigated the role of the
food-rewarded, eight-arm radial maze tasks; a cogni- amygdala in a fear-conditioning paradigm in rats in
tive/declarative win-shift task that required rats to re- which brief electrical shock is paired with exposure to
member the arms they had visited earlier in a daily either a discrete auditory cue or a specific environ-
training session; and a procedural/S-R habit win-stay mental context (LeDoux, 1992; Fendt and Fanselow,
task that required rats to acquire a simultaneous 1999). In a series of anatomical, lesion, and pharma-
visual (light-dark) discrimination. Lesions of the cological studies these investigators have mapped out
dorsal striatum impair acquisition of the win-stay a potential neural circuit in which conditioned stimu-
task and do not affect acquisition of the win-shift task, lus (e.g., tone) and unconditioned stimulus (e.g.,
whereas lesions of the hippocampal system produce shock) information converge in the basolateral amyg-
the opposite dissociation (Packard, Hirsh, and White, dala. Projections from the basolateral nucleus to the
1989). central nucleus of the amygdala and subsequent
downstream projections to various brain-stem regions
An additional study used two water-maze tasks to allow for the expression of various autonomic and be-
investigate the selective role of the basal ganglia in havioral responses (e.g., freezing behavior, heart-rate
S-R memory (Packard and McGaugh, 1992). In these and blood-pressure changes) in response to discrete
tasks two rubber balls protruding above the water sur- and contextual conditioned stimuli. The auditory cor-
face served as cues. One ball (correct) was on top of tex is not necessary for the acquisition of discrete au-
a platform that could be used to escape from the ditory cue fear conditioning. Rather, auditory infor-
water, and the other ball (incorrect) was on top of a mation projected to the amygdala from the medial
thin rod and thus did not provide escape. The two geniculate nucleus of the thalamus is apparently suffi-
balls also differed in visual appearance (i.e., vertical cient to support this form of fear conditioning. These
versus horizontal black/white stripes). In a cognitive/ findings suggest a rapid-response procedural learn-
declarative version of the task, the correct platform ing system that can bypass neocortical involvement in
was in the same spatial location on every trial, but the cognitive assessment of threatening stimuli.
appearance of the ball varied. Therefore, this version Further evidence of a role for the amygdala in
of the task requires rats to learn to approach the cor- learned fear responses in the rat emerged from an in-
rect ball on the basis of spatial location, not visual pat- vestigation of the fear-potentiated startle response by
tern. In a procedural/S-R habit version of the task, the Michael Davis and his colleagues (e.g. Davis, 1992).
rats located the correct platform in different spatial In these experiments, rats initially receive pairings of
locations across trials, but the visual pattern was con- a light or tone with foot shock. During subsequent
sistent. Therefore, this task could be acquired by training, the normal startle response that results from
learning an approach response to the visual cue. Le- exposure to a loud noise is potentiated by concurrent
sions of the dorsal striatum impair acquisition of the presentation of the light or tone conditioned stimu-
S-R habit task without affecting acquisition of the spa- lus. Findings from lesion, pharmacological, and ana-
tial task (Packard and McGaugh, 1992). Other find- tomical studies converge to indicate a role for the
ings indicating that lesions of the rodent dorsal stria- amygdala in the acquisition and expression of fear-
tum impair two-way active avoidance behavior, potentiated startle.
simultaneous tactile discriminations, egocentric re- Aside from its involvement in aversively motivat-
sponse learning, and conditional visual and auditory ed learning, evidence indicates a role for the amygda-
conditioning also support a role for the dorsal stria- la in the acquisition and expression of appetitively
tum in S-R habit learning and memory (Kirkby and motivated stimulus-affect learning tasks. For exam-
Kimble, 1968; Colombo, Davis, and Volpe, 1989; ple, pretraining and posttraining basolateral amyg-
Reading, Dunnett, and Robbins, 1991; Adams, Kes- dala lesions in rats impair acquisition and expression
ner, and Raggozino, 2001; Kesner, Bolland, and of conditioned place preference behavior for both
Dakis, 1993). natural rewards such as food and sex (Cador, Rob-
546 PROCEDURAL LEARNING: Animals
bins, and Everitt, 1989) and addictive drugs such as Mangels, and Squire, 1996; Cohen and Squire, 1980;
amphetamines (Hiroi and White, 1991; Hsu, Johnsrude et al., 2000).
Schroeder, and Packard, 2002). In this task, rats are
confined in contrasting environments that alternate
See also: DECLARATIVE MEMORY; NEURAL
daily: on one day in an environmental context paired SUBSTRATES OF CLASSICAL CONDITIONING;
with natural or drug rewards and on the next day in NEURAL SUBSTRATES OF EMOTIONAL MEMORY
a second context that is not paired with the rewarding
treatment. On a reward-free test session given follow-
ing training, the amount of time spent in the two envi- Bibliography
ronments is measured, and rats demonstrate a reli- Adams, S., Kesner, R. P., and Ragozzino, M. E. (2001). Role of the
medial and lateral caudate-putamen in mediating an auditory
able conditioned place preference forthat is, spend conditional response association. Neurobiology of Learning and
longer amounts of time inthe environment previ- Memory 76, 106116.
ously paired with the rewarding stimulus. Investiga- Cador, M., Robbins, T. W., and Everitt, B. J. (1989). Involvement
tors interested in the neurobiological bases of addic- of the amygdala in stimulus-reward associations: Interaction
with the ventral striatum. Neuroscience 30, 7786.
tion have traditionally used this task. However,
Cahill, L., Weinberger, N. M., Roozendaal, B., and McGaugh, J. L.
animals display approach behavior to specific envi- (1999). Is the amygdala a locus of conditioned fear? Some
ronmental stimuli in a drug-free state on the test day, questions and caveats. Neuron 23, 227228.
and therefore expression of a conditioned place pref- Cohen, N. J., and Squire, L. R. (1980). Preserved learning and re-
erence requires memory for an association between tention of pattern analyzing skill in amnesics: Dissociation of
previously neutral stimuli and the rewarding affective
Colombo, P. J., Davis, H. P., and Volpe, B. T. (1989). Allocentric
consequences of the treatment. spatial and tactile memory impairments in rats with dorsal
caudate lesions are affected by preoperative training. Behav-
It should be noted that not all investigators agree ioral Neuroscience 103, 242250.
that the mnemonic function of the amygdala involves Davis, M. (1992). The role of the amygdala in conditioned fear. In
a long-term role in memory storage. James L. Mc- J. P. Aggleton, ed., The amygdala: Neurobiological aspects of emo-
Gaugh and his colleagues argue that amygdala dam- tion, memory, and mental dysfunction. New York: Wiley-Liss.
age impairs unconditioned fear responses (specifical- Fanselow, M. S., and LeDoux, J. E. (1999). Why we think plasticity
underlying Pavlovian fear conditioning occurs in the baso-
ly freezing behavior), and therefore suggest caution lateral amygdala. Neuron 23, 229232.
in proposing a role for the amygdala role in learned Fendt, M., and Fanselow, M. S. (1999). The neuroanatomical and
fear responses (Cahill, Weinberger, Roozendaal, and neurochemical basis of conditioned fear. Neuroscience and Bio-
McGaugh, 1999; Fanselow and LeDoux, 1999). In ad- behavioral Reviews 23, 743760.
dition, evidence indicates that the amygdala (specifi- Fernandez-Ruiz, J., Wang, J., Aigner, T. G., and Mishkin, M.
(2001). Visual habit formation in monkeys with neurotoxic le-
cally the basolateral nucleus) modulates memory pro- sions of the ventrocaudal neostriatum. Proceedings of the Na-
cesses occurring in other brain structures, including tional Academy of Sciences of the United States of America
declarative memory processes mediated by the hippo- 984,4,1964,201.
campus and stimulus-response habit learning mediat- Hiroi, N., and White, N. M. (1991). The lateral nucleus of the
amygdala mediates expression of the amphetamine condi-
ed by the basal ganglia (Packard, Cahill, and Mc-
tioned place preference. Journal of Neuroscience 11, 2,107
Gaugh, 1994). However, a modulatory role for the 2,176.
amygdala in some types of learning does not itself Johnsrude, I. S., Owen, A. M., White, N. M., Zhao, W. V., and Boh-
rule out a possible long-term role in memory storage bot, V. (2000). Impaired preference conditioning after anteri-
in stimulus-affect conditioning. There is a need for or temporal lobe resection in humans. Journal of Neuroscience
20, 2,6492,656.
further task-specific analysis of the potential short-
Kapp, B. S., Frysinger, R. C., Gallagher, M., and Haselton, J. R.
term versus long-term roles of the amygdala in mem- (1979). Amygdala central nucleus lesions: Effects on heart
ory and for consideration of the role of separate rate conditioning in the rabbit. Physiology and Behavior 23,
amygdala nuclei in different types of learning and 1,1091,117.
memory. Kesner, R. P., Bolland, B. L., and Dakis, M. (1993). Memory for
spatial locations, motor responses, and objects: Triple dissoci-
There has been significant progress in identifying ation among the hippocampus, caudate nucleus, and ex-
neuroanatomical components of separable procedur- trastriate visual cortex. Experimental Brain Research 93, 462
470.
al memory systems in the mammalian brain, includ- Kim, J. J., and Thompson, R. F. (1997). Cerebellar circuits and syn-
ing the cerebellum, basal ganglia, and amygdala. Al- aptic mechanisms involved in classical eyeblink conditioning.
though this article has focused on studies of lower Trends in Neuroscience 20, 177181.
animals, there is similar evidence for the role of these Kirkby, R. J., and Kimble, D. P. (1968). Avoidance and escape be-
various structures in some of these forms of procedur- havior following striatal lesions in the rat. Experimental Neurol-
ogy 20, 215227.
al memory in nonhuman primates (Zola-Morgan, Knowlton, B. J., Mangels, J. A., and Squire, L. R. (1996). A
Squire, and Mishkin, 1982; Fernandez-Ruiz, Wang, neostriatal habit learning system in humans. Science 273,
Aigner, and Mishkin, 2001;) and humans (Knowlton, 1,3991,402.
PROCEDURAL LEARNING: Humans 547
LeDoux, J. E. (1992). Emotion and the amygdala. In J. P. Aggleton, eventually takes on an automatic or reflexive quality.
ed., The amygdala: Neurobiological aspects of emotion, memory, and It is, however, long-lasting and reliable, as any bike
mental dysfunction. New York: Wiley-Liss.
McCormick, D. A., Lavond, D. G., Clark, G., Kettner, R. E., Rising,
rider knowseven after years of absence from a bicy-
C. E., and Thompson, R. F. (1981). The engram found? Role cle, one never loses the skill. While most declarative
of the cerebellum in classical conditioning of nictitating mem- learning is not impaired by rapid-eye-movement
brane and eylid responses. Bulletin of the Psychonomic Society 18, (REM) sleep deprivation, procedural learning is
103105. (Stickgold et al., 2001).
Mishkin, M., and Petri, H. L. (1984). Memories and habits: Some
implications for the analysis of learning and retention. In L. Skills, the procedures that allow us to function in
R. Squire and N. Butters, eds., Neuropsychology of memory. New the world, include motor, perceptual, and cognitive
York: Guilford. processes. Examples of learned skills are driving a car
Packard, M. G., Cahill, L., and McGaugh, J. L. (1994). Amygdala
modulation of hippocampal-dependent and caudate nucleus-
with a manual transmission (motor), a parents atten-
dependent memory processes. Proceedings of the National Acad- tiveness to his or her babys cry in a distant room (per-
emy of Sciences of the United States of America 91, 8,4778,481. ceptual), and increasing alacrity in solving a Rubiks
Packard, M. G., Hirsh, R, and White, N. M. (1989). Differential ef- Cube with practice (cognitive). Habits are a form of
fects of fornix and caudate nucleus lesions on two radial maze gradual, incremental learning, a settled pattern of re-
tasks: Evidence for multiple memory systems. Journal of
sponses toward repeated stimuli. An example of a
Packard, M. G., and McGaugh, J. L. (1992). Double dissociation of habit is a person regularly opening the refrigerator
fornix and caudate nucleus lesions on acquisition of two water door when he or she walks into the kitchen. Like
maze tasks: Further evidence for multiple memory systems. skills, habits allow us to function efficiently in the
Behavioral Neuroscience 106, 439446. world by responding to stimuli with minimal cogni-
Reading, P. J., Dunnett, S. B., and Robbins, T. W. (1991). Dissocia-
ble roles of the ventral, medial, and lateral striatum on the ac-
tive effort. While declarative memory can, in some
quisition and performance of a complex visual stimulus re- cases, enhance or hasten the acquisition of skills and
sponse habit. Behavioral Brain Research 45, 147161. habits, usually conscious awareness of learning is not
Zola-Morgan, S., Squire, L., and Mishkin, M. (1982). The necessary; once the information is acquired, it often
neuroanatomy of amnesia: Amygdala-hippocampus versus becomes difficult to verbalize it. This is why procedur-
temporal stem. Science 218, 1,3371,339.
al learning is called nondeclarative.
Mark G. Packard
The Case of Patient H.M.

HUMANS By far the most famous example of a patient who
has retained procedural learning in the absence of
In discussing long-term memory, scientists have declarative memory is the case of H.M., a patient with
found it useful to distinguish between several kinds of severe intractable epilepsy who underwent surgery as
memory that rely on different brain systems. The a last attempt at correcting his condition. Surgeons
major distinction is between declarative memory, removed most of his medial temporal lobe structures
which refers to the conscious memory of facts and bilaterally and left him devoid of declarative memory,
events, and nondeclarative memory, which refers to and therefore suffering from amnesia. Brenda Milner
nonconscious memory of skills, habits, or other (1962) later showed that he was capable of improving
modes of learning that proceed beneath the surface his performance on a mirror-drawing task in which
of conscious awareness. Declarative memory is what participants trace the outline of a figure with a stylus
most people call memory. It depends on the integrity (e.g., a star) while watching the reflection of their ef-
of the hippocampus and related structures of the me- forts in a mirror. While initially challenging, the task
dial temporal lobe. Nondeclarative memory affects becomes easier with practice. Despite not having any
our behavior without our explicit knowledge. It is a recollection of ever performing the task, H.M. be-
heterogeneous collection of nonconscious memory came more accurate across sessions, demonstrating
abilities that depend on various other structures with- that he had acquired this skill. This was the earliest
in the brain (Squire et al., 1993; see Figure 1). evidence that skill-learning can occur without declar-
One well-studied component of nondeclarative ative memory. There is evidence that not only motor-
memory is procedural memory. The difference be- skill learning (as tested by the mirror drawing task)
tween declarative memory and procedural memory is but also perceptual and cognitive-skill learning can
the difference between knowing that and knowing also proceed in the absence of declarative memory.
how. Procedural learning describes the formation of
skills and habits. It is the most primitive form of
learning, the first to develop in infancy (Tulving and Motor-Skill Learning
Schacter, 1990). Because it requires extensive prac- Motor-skill learning is tested primarily by three
tice, it is a slow and inflexible learning system that tasks: the mirror drawing described above, rotary
548 PROCEDURAL LEARNING: Humans
Figure 1
pursuit, and serial reaction time (SRT). In rotary pur- Perceptual-Skill Learning
suit, participants try to keep a hand-held stylus in In addition to having intact motor-skill learning,
contact with a nickel-sized metal disk that rotates on patients with amnesia also have intact perceptual-skill
a table. With practice, participants are able to main- learning. One of the most popular tests of perceptual-
tain this contact for longer periods of time. Suzanne skill learning is reading mirror-reversed text. While
Corkin (1968) has shown that patients with declara- people with declarative memory problems show im-
tive memory deficits (i.e., amnesics) are able to improved mirror-reading with practice, patients with
prove their performance on this task. In the SRT task, basal ganglia damage (e.g., people with Huntingtons
participants press the corresponding button when a disease) evince mildly impaired learning (Martone et
target item appears in one of four locations on a al., 1984).
screen. The appearance of the target follows a ten-to-
twelve trial sequence that the participants eventually
learn implicitly. SRT learning remains intact in am- Learning Cognitive Skills and Habits
nesics (Nissen and Bullemer, 1987). Cognitive-skill learning has also been tested
using several tasks, the most common of which are the
Furthermore, both the basal ganglia and the cer- tower tasks and the artificial grammar learning task.
ebellum play a role in motor-skill learning (Gabrieli, Habit learning in humans has been measured using
1998). The basal ganglia are a group of subcortical a probabilistic classification task. The tower tasks re-
structures that surround the thalamus and include the quire planning and problem solving: participants are
caudate, putamen, and globus pallidus. Neuroimag- asked to change the location of objects according to
ing techniques have shown that SRT skill learning ac- certain rules. For example, in the tower of Hanoi task,
tivates the basal ganglia (Karni et al., 1995). Lesions the subject is presented with three pegs with a set of
to the cerebellum, a structure at the base of the brain, disks on the leftmost peg. The goal is to move all the
can impair mirror-tracing skills (Sanes et al., 1990). disks to the rightmost peg, with the disks piled in
Many scientists believe that the acquisition of motor order from largest to smallest. The subject may move
skills depends on the basal ganglia, while the associa- a disk to an adjacent peg on each move and may not
tion of visual cues with motor actions depends on the put a larger disk on top of a smaller disk. The tower
cerebellum (Willingham et al., 1996). tasks have yielded mixed results: some researchers
PROCEDURAL LEARNING: Humans 549
have found normal learning in amnesic patients probabilistic classification task. Since the brain re-
under some circumstances (Cohen et al., 1985) but gions damaged by PD include the neostriatum (cau-
not others (Butters et al., 1985). It is likely that the date and putamen structures that are part of the basal
tower tasks draw on both procedural memory and de- ganglia) but not the medial temporal lobes, the ex-
clarative memory for the consequences of particular perimenters concluded that the neostriatum is re-
moves that one has already tried. sponsible for habit learning.
The artificial grammar task requires the abstrac- The ultimate output of the basal ganglia is the
tion of rules and regularities underlying seemingly frontal cortex. In turn, the frontal cortex sends a
random strings of letters. The rules allow only certain major projection to the neostriatum. This cortico-
letter strings to follow other letter strings. After view- striatal loop appears to be important for procedural
ing a series of these grammatical letter strings with- learning. This finding is supported by reports that pa-
out being told about the rules, people are able to clas- tients with striatal damage exhibit deficits in skill
sify new strings as either grammatical or learning as well as habit learning. Furthermore, a re-
nongrammatical fairly accurately. However, they cent neuroimaging study by Poldrack and colleagues
are not able to report much about the rules and gen- (2001) has shown the medial temporal lobe-based de-
erally feel that they had simply been guessing. Again, clarative memory system and the cortico-striatal
amnesic patients cannot remember individual train- based procedural memory system may have a recipro-
ing letter strings very well but perform as well as nor- cal relationship during learning. They administered
mal people when classifying new strings (Knowlton, the probabilistic classification (weather prediction)
Ramus, and Squire, 1992; Knowlton and Squire, task to healthy young people and designed two condi-
1996). tions, one emphasizing the declarative aspects of the
task and the other the nondeclarative aspects. The in-
The probabilistic classification task that has been vestigators found that the declarative version elicited
used to test habit learning is the weather prediction medial temporal lobe activity, while the nondeclara-
task, which involves a series of cues that are proba- tive version elicited activity in the basal ganglia and
bilistically associated with either a sunny or rainy other subcortical structures. Also, activity in the basal
outcome. On each trial participants try to guess which ganglia was negatively correlated with activity in the
outcome will occur for the cues presented. Because medial temporal lobes.
the cues are associated with a particular outcome on
only 60 to 90 percent of the trials, memorizing indi-
vidual trials is not as helpful as accruing knowledge Conclusion
across many trials. Patients with amnesia perform Procedural learning involves skill and habit
normally on the probabilistic classification task, even learning, both of which are spared in the abolition of
in the absence of conscious recollection of the train- declarative memory. While declarative memory de-
ing episode (Knowlton et al, 1994; 1996). pends on medial temporal lobe structures (e.g, the
hippocampus), skill and habit learning depend on the
basal ganglia. Other structures, such as the cerebel-
Neural Substrates of Habit Learning in lum, may also play a role in some forms of procedural
Humans learning. The development of new procedural learn-
In neuropsychology, the best evidence for a brain ing tasks that can be used with brain-injured patients
regions involvement in a cognitive process is a dou- or in neuroimaging studies will help elucidate the
ble dissociation, in which patients with a particular le- neural substrates of this form of learning.
sion are impaired at task A but intact at task B, while
patients with a different lesion show the opposite pat- See also: DECLARATIVE MEMORY
tern. In humans, such a double dissociation was found
by Knowlton and colleagues (1996) among amnesic Bibliography
patients and those with Parkinsons disease (PD). Butters, N., Wolfe, J., Martone, M., Granholm, E., and Cermak, L.
S. (1985). Memory disorders associated with Huntingtons
Both groups of patients were asked to perform the
disease: verbal recall, verbal recognition, and procedural
weather prediction task described above. Both groups memory. Neuropsychologia 6, 729744.
were also asked some multiple-choice questions de- Cohen, N. J., Eichembaum, H., Deacedo, B. S., and Corkin, S.
signed to investigate whether they remembered the (1985). Different memory systems underlying acquisition of
learning situation. Patients with amnesia were able to procedural and declarative knowledge. Annals of the New York
Academy of Sciences 444, 5471.
learn the classification task but showed almost no de-
Corkin, S. (1968). Acquisition of a motor skill after bilateral medial
clarative memory for the learning episode. In con- temporallobe excision. Neuropsychologia 6, 255265.
trast, PD patients were able to remember details of Gabrieli, J. D. E. (1998). Cognitive Neuroscience of human memo-
the learning episode but showed no learning in the ry. Annual Review of Psychology 49, 87115.
550 PROSE RETENTION
Karni, A., Meyer, G., Jezzard, P., Adams, M. M., Turner, R., and of presentation, this entry considers each of these fac-
Ungerleider, L. G. (1995). Functional MRI evidence for adult tors separately. In reality, however, each of these fac-
motor cortex plasticity during motor skill learning. Nature
377, 155158.
tors does not operate in isolation; instead, they work
Knowlton, B. J., Ramus, S. J., and Squire, L. R. (1992). Intact artifi- together to influence memory for prose.
cial grammar learning in amnesia: Dissociation of classifica-
tion learning and explicit memory for specific instances. Psy-
chological Science 3, 177179. Influences of Encoding and Reading
Knowlton, B. J., and Squire, L. R. (1996). Artificial grammar learn-
ing depends on implicit acquisition of both abstract and ex-
Processes
emplar-specific information. Journal of Experimental Psychology: One of the dominant themes in contemporary
Learning, Memory, and Cognition 22, 169181. memory research is that memory benefits to the ex-
Knowlton, B. J., Squire, L. R., and Gluck, M.A. (1994). Probabilistic
tent that the material is processed for meaning. The
classification in amnesia. Learning, Memory, and Cognition 21,
699710. same holds true for prose. For example, research has
Martone, M., Butters, N., Payne, M., Becker, J. T., and Sax, D. S. shown that readers who have been instructed to count
(1984). Dissociations between skill learning and verbal recog- the number of four-letter words in a passage do not
nition in amnesia and dementia. Archives of Neurology 41, 965 remember the passage as well as readers who have
970.
been instructed to rate the paragraphs for ambiguity
Milner, B. (1962). Les troubles de la mmoire accompagnant des
lesions hippocampiques bilaterales. In P. Passouant, ed., Phy- (Schallert, 1976). Similarly, if a text is written so that
siologie de lhippocampe. Paris: Centre de la Recherche Scienti- its topic is obfuscated, the degree to which meaning-
fique. ful processing can take place decreases, and this text
Nissen, M. J., and Bullemer, P. (1987). Attentional requirements is recalled less well than a text in which the topic is
of learning: Evidence from performance measures. Cognitive
clearly stated (e.g., through an informative title or by
Psychology 19, 132.
Poldrack, R. A., Clark. J, Pare-Blagoev, E. J., Shohamy, D., Creso reference to the topic throughout the text; Bransford
Moyano, J., Myers, C., and Gluck, M. A. (2001). Interactive and Johnson, 1972). The lesson here is straightfor-
memory systems in the human brain. Nature 414, 546550. ward: A prerequisite for remembering prose is ex-
Sanes, J., Dimitrov, B., and Hallett, M. (1990). Motor learning in tracting the meaning that is being conveyed.
patients with cerebellar dysfunction. Brain 113, 103120.
Squire, L. R., and Knowlton, B. J. (1995). Memory, hippocampus On the assumption that readers typically are ex-
and brain systems. In M. Gazzaniga, ed., The cognitive neuro- tracting meaning as they read, one can further ana-
sciences. Cambridge, MA: MIT Press.
Squire, L. R., Knowlton, B., and Musen, G. (1993). The structure
lyze the components of that meaning which are re-
and organization of memory. Annual Review of Psychology 44, tained. It is typically the case that people do not
453495. remember prose verbatim (except with concerted ef-
Stickgold, R., Hobson, J. A., Fosse, R., and Fosse M. (2001). Sleep, fort, and even then it is difficult; Sachs, 1967). In-
learning, and dreams: Off-line memory reprocessing. Science stead, prose retention is characterized by memory for
294, 1,0521,057.
Tulving, E., and Schacter, D. L. (1990). Priming and human mem-
the ideas that are captured by the particular words
ory systems. Science 247, 301306. used. Further, memory for the ideas in a text general-
Willingham, D. B., Koroshetz, W. J., and Peterson, E. W. (1996). ly conforms to the following function. The main or
Motor skills have diverse neural bases: Spared and impaired most important ideas are recalled best, followed by
skill acquisition in Huntingtons disease. Neuropsychology 10, ideas of intermediate importance, with unimportant
315321.
details usually recalled least well.
Indre V. Viskontas
Barbara J. Knowlton The readers point of view or frame of reference
partly determines the importance of the ideas con-
tained in the text. Because an ideas level of impor-
tance determines its memorability, very different pat-
terns of memory for the same text can be observed
PROSE RETENTION
across readers who adopt different points of view. For
As with memory for other kinds of verbal material, example, a prose passage about a household and its
memory for prose is a complex function of various daily activities will be remembered differently by a
factors. Some of the more important factors include reader who adopts the perspective of a prospective
1. the kind of reading or studying activity in which the home buyer compared with a reader who adopts
learner engages while processing a text; 2. the partic- the perspective of a burglar appraising particular
ular type of prose that is being remembered; 3. the properties for a clandestine visit (Fass and Schum-
kinds of information and events that intervene be- acher, 1981). The differences in recall for these two
tween the initial reading of the text and when remem- readers will directly reflect the fact that the particular
bering is attempted; and 4. the way in which memory ideas in the passage differ in importance, depending
is assessed (e.g., through recall, short-answer, multi- on the perspective adopted while reading the pas-
ple-choice, or true/false tests; see Figure 1). For clarity sage.
PROSE RETENTION 551
Figure 1
Consistent with findings for other types of verbal Prose Organization and Prose Type
material, good prose retention depends on reading
Generally, narrative prose (e.g., a story) is better
activities that engage attention toward understanding
remembered than expository prose (e.g., an essay or
the individual ideas in a text and how these ideas are
article). The superior memorability of narratives rela-
organized or interrelated. Such activities can include
tive to expositions is probably due in large part to
trying to explain the information to oneself, thinking
readers better ability to organize and interrelate the
about causes or results, relating the information to
ideas expressed in a narrative. There are a number
prior knowledge, and creating linkages between the
of differentiating factors between narratives and ex-
presented ideas. Numerous factors influence the de-
gree to which such reading activities will be engaged. pository prose that could underlie the extent to which
Not surprisingly, readers seem to become more en- the ideas expressed in the text are organized by the
gaged for text rated as more interesting than for less reader. Though it may ultimately turn out that not all
interesting text, with more interesting text being bet- of these factors are important in accounting specifi-
ter remembered (Son and Metcalf, 2000). Individual cally for the memory differences between narratives
readers vary as well, with some readers spontaneously and expository text, these factors do seem to play a
engaging in more active processing of the text than role in prose memory.
others, thereby increasing retention for the prose ma- One factor is that readers presumably activate rel-
terial. Individuals can be instructed or encouraged to evant structures of world knowledge when attempting
use selected reading or study strategies (e.g., self- to understand a text. These structures are usually
explanation, outlining) designed to stimulate greater
called schemata (singular, schema). In general,
comprehension and organization of the information
schemata are believed to influence the degree to
in the text. These reading strategies generally im-
which incoming text information is organized. Text
prove overall recall of factual (expository) text
content for which readers have a schema is believed
(McDaniel and Donnelly, 1996; Slater, Graves, and
to be better organized and elaborated, and thus
Piche, 1985).
better remembered, than text content for which
One difference between prose memory and mem- readers have a poorly articulated or sketchy schema.
ory for word lists is that prose allows special forms of For instance, researchers have shown that a narrative
organization that are not operative with simpler ma- about a fictitious baseball game is better recalled by
terials. The varied kinds of organizational factors that readers knowledgeable about baseball (i.e., readers
can be influential in prose memory are best revealed who have a baseball schema) than readers who are
by considering how the genre of the prose affects re- not knowledgeable about baseball (Spilich et al.,
membering. 1979).
552 PROSE RETENTION
In contrasting narratives and expository text, it is Remembering Prose over Time

arguably the case that narratives have content for The organizational processes outlined previously
which most of the readers in the culture have relevant exert greater influence as the interval between read-
schemata, whereas for expository texts, the content ing and attempted remembering increases. If recall
articulates with schemata that only a subset of readers is attempted within minutes after reading, then infor-
might possess. Narratives are about people or other mation that does not fit an initial perspective or sche-
animate objects, their goals and conflicts, and their ma can be recovered. If recall is delayed for at least
other activitiesconcepts for which readers have one day, however, then recall of information that
schemata. Narratives frequently entail events that, if does not fit the schema may drop substantially,
not directly experienced by the reader, relate to basic whereas information that fits the schema will still be
cultural values and goals (e.g., finding happiness, well recalled (Fass and Schumacher, 1981). Similarly,
wealth, and love). The influence of schemata in orga- in some cases study strategies designed to increase or-
nizing and guiding memory of narratives can be ob- ganizational processing of a text (e.g., outlining) will
served when readers try to recall narratives from a not improve memory for a text (relative to reading
culture with which they have little familiarity. Recall alone, with no studying) when recall is tested immedi-
of such narratives is characterized by the omission ately, but will increase recall when testing is delayed
and transformation of ideas and details in the narra- for several days (Einstein et al., 1990). This increas-
tive that do not fit with the conventional schema of ingly beneficial mnemonic effect of organization with
the readers culture (Bartlett, 1932). longer retention intervals is another aspect of prose
memory that parallels memory for simpler verbal ma-
Prose can also be organized by linking the ele-
terial.
ments in terms of causal relations. Causal relations
capture relations between events (described in a text) Another factor that affects memory for prose is
that involve motivation, psychological causation (e.g., information that the learner encounters between
greed causes certain behaviors), and physical causa- reading a text and trying to remember it. Such inter-
tion (e.g., rain causes things to become slippery). Un- vening information can have facilitative as well as in-
derstanding a narrative seems to involve the forma- terfering effects on retention. To illustrate both kinds
tion of a sequence of causal relations into a chain that of effects, consider a biography about a poet named
links the opening of the narrative to its outcome. In Susan. In the biography it is stated that Susans father
remembering a narrative, people tend to remember was a servant who died of diphtheria when Susan was
those parts of a narrative which fit into the causal five years old. After reading this biography, another
chain and forget the parts which are not incorporated biography is encountered in which Anns blacksmith
into it (Fletcher and Bloom, 1988). More generally, father dies of lung cancer when Ann is two years old.
better memory for narrative prose may occur because Research has shown that readers who are given the
causal sequences are more clearly defined in narra- two biographies and are then asked to recall the first
tives than in expository passages. biography, recall the theme of the first biography
(e.g., the main characters father died when she was
A third factor that contributes to prose organiza- young) better than readers who are not given the
tion is the readers knowledge about the structure of second biography. But readers given the second biog-
conventional text forms. That is, when reading a fairy raphy do not recall the specific details of the first biog-
tale, regardless of the particular content, one expects raphy (the fathers occupation, what he died of, and
to see certain components: a setting, an initiating when) as well as the readers not given the second bi-
event posing a problem, an attempt at its solution by ography (Bower, 1974).
the protagonist, and a consequence. This knowledge
presumably allows the reader to better organize the Perhaps one of the more interesting features of
information found in the passage. prose recall is that intervening information, or even
a long retention interval, can promote reconstruction
In contrast, expository prose does not seem to en- in remembering. Reconstruction refers to the finding
tail a consistent structure, thereby reducing the ex- that memory of a text can be distorted by the inclu-
tent to which the reader has a prestored structure that sion of information consistent with the theme of the
facilitates organization. Late-twentieth-century re- text but not actually mentioned in it. Reconstruction
search indicates, however, that if readers are explicit- also includes alteration of information that was in the
ly trained to notice and utilize some conventional ex- text to bring it more in line with information encoun-
pository forms (e.g., argumentative structures), then tered subsequent to the text or with readers schemata
memory for such material is improved. Further, this that were activated in comprehending the text. For
positive effect is obtained for older as well as younger instance, in one study (Spiro, 1980) participants read
adults (Meyer, Young, and Bartlett, 1990). a narrative about two college students who started
PROTEIN SYNTHESIS IN LONG-TERM MEMORY IN VERTEBRATES 553
dating seriously but disagreed about their desires to Bransford, J. D., and Johnson, M. K. (1972). Contextual prerequi-
have children. After the participants read the text, the sites for understanding: Some investigations of comprehen-
sion and recall. Journal of Verbal Learning and Verbal Behavior
experimenter casually mentioned that the students 11, 717726.
had ended up getting married. Several weeks later, Einstein, G. O., McDaniel, M. A., Owen, P. D., and Cot, N. C.
the participants attempted to recall the narrative. (1990). Encoding and recall of texts: The importance of mate-
Their recall included reconstructions and distortions rial appropriate processing. Journal of Memory and Language
consistent with the new information mentioned by the 29, 566581.
Fass, W., and Schumacher, G. M. (1981). Schema theory and prose
experimenter but not actually in the narrative (e.g., retention: Boundary conditions for encoding and retrieval ef-
The students underwent counseling to correct the fects. Discourse Processes 4, 1726.
major discrepancy). Fletcher, C. R., and Bloom, C. P. (1988). Causal reasoning in the
comprehension of simple narrative texts. Journal of Memory
Prose recall need not always be reconstructive. If and Language 27, 235244.
the reader expects a memory test on what has been Hasher, L., and Griffin, M. (1978). Reconstructive and reproduc-
read, then recall is more reproductive. That is, details tive processes in memory. Journal of Experimental Psychology:
are recalled more accurately, and there is very little Human Learning and Memory 4, 318330.
inclusion of extra information not actually in the text. McDaniel, M. A., and Donnelly, C. M. (1996). Learning with analo-
gy and elaborative interrogation. Journal of Educational Psy-
Also, in situations in which the readers initial inter- chology 88, 508519.
pretative schema is invalidated at the time of recall, McDaniel, M. A., and Kerwin, M. L. E. (1987). Long-term prose re-
recall becomes more reproductive (Hasher and Grif- tention: Is an organizational schema sufficient? Discourse Pro-
fin, 1978). cesses 10, 237252.
Meyer, B. J. F., Young, C. J., and Bartlett, B. J. (1989). Memory im-
proved: Reading and memory enhancement across the life span
Assessing Prose Retention through strategic text structures. Hillsdale, NJ: Erlbaum.
Sachs, J. (1967). Recognition memory for syntactic and semantic
In discussing prose retention, much of the focus aspects of connected discourse. Perception and Psychophysics 2,
has been on remembering as evidenced in recall. The 437442.
influence of many of the aforementioned factors Schallert, D. L. (1976). Improving memory for prose: The relation-
ship between depth of processing and context. Journal of Ver-
changes, however, if remembering is assessed with
bal Learning and Verbal Behavior 15, 621632.
recognition memory tests. In recognition tests, the Slater, W. H., Graves, M. F., and Piche, G. L. (1985). Effects of
learner is presented with facts stated in the text and structural organizers on ninth-grade students comprehen-
facts not stated in the text, and must determine which sion and recall of four patterns of expository text. Reading Re-
facts were presented and which were not. Common search Quarterly 20, 189202.
Son, L. K., and Metcalfe, J. (2000). Metacognitive and control strat-
kinds of recognition tests include true/false and multi-
egies in study-time allocation. Journal of Experimental Psycholo-
ple-choice tests. In general, if prose memory is tested gy: Learning, Memory, and Cognition 26, 204221.
with recognition tests, the robust influence of organi- Spilich, G. J., Vesonder, G. T., Chiesi, H. L., and Voss, J. F. (1979).
zational factors like those discussed earlier can be mit- Text processing of domain-related information for individu-
igated or eliminated. Thus, recognition is the same als with high and low domain knowledge. Journal of Verbal
Learning and Verbal Behavior 18, 275290.
regardless of whether the text is written so that its
Spiro, R. J. (1980). Accommodative reconstruction in prose recall.
overall theme can be clearly identified. Further, in- Journal of Verbal Learning and Verbal Behavior 19, 8495.
formation that does not fit an encoding schema is rec-
Mark A. McDaniel
ognized as well as information that does fit it, even if
memory is tested one week after the text was read
(McDaniel and Kerwin, 1987). Thus, as is the case
with simple verbal materials, organizational factors
PROTEIN SYNTHESIS IN LONG-TERM
play a more prominent role when prose retention is
measured with recall than when measured with recog- MEMORY IN VERTEBRATES
nition. Most modern theories assume that memories are
stored in the brain in the form of changed patterns
See also: CODING PROCESSES: LEVELS OF of synaptic connections within ensembles of neurons,
PROCESSING; CODING PROCESSES: although it remains debated whether such patterns
ORGANIZATION OF MEMORY; EXPERTS are stable once formed or subject to dynamic change.
MEMORIES; ORAL TRADITIONS;
Any such growth or reorganization of synapses re-
RECONSTRUCTIVE MEMORY
quires the synthesis of the molecules comprising
them, especially the proteins and lipids of the synap-
Bibliography
tic and dendritic membranes. The idea that memory
Bartlett, F. C. (1932). Remembering. Cambridge, UK: Cambridge
University Press. formation involves protein synthesis has been around
Bower, G. H. (1974). Selective facilitation and interference in re- for a long timecertainly since the days of Santiago
tention of prose. Journal of Educational Psychology 66, 18. Ramn y Cajal (Spanish histologist, 18521934) at
554 PROTEIN SYNTHESIS IN LONG-TERM MEMORY IN VERTEBRATES
the beginning of the twentieth centurybut serious using an inhibitor of RNA synthesis, 8-azaguanine.
experimental tests of the idea became possible only Rats were trained to swim a water maze and then in-
with techniques available beginning in the 1960s. The jected with the inhibitor. It had no effect on the per-
adult brain has one of the highest rates of protein syn- formance of animals that had already learned the
thesis of any body organ, and also shows the greatest maze, or on their ability to swim in general. However,
diversity of protein molecules. Some 30,000 different if the inhibitor was injected during the training trials,
proteins are synthesized in the primate brain, only a and the rats were tested the following day, they
small fraction of which have been characterized or as- showed impaired memory for the maze.
cribed functional roles. Many protein species are like-
ly to be involved in memory consolidation, and it is In the decades that followed, experimenters em-
unlikely that any of them are uniquely required; what ployed various antibiotics that interfere with particu-
provides the specificity for the memory trace is not lar steps in protein synthesis: puromycin, cyclohex-
the particular protein but the new pattern of synapses imide, acetoxycycloheximide, and anisomycin. The
it helps to construct. consensus observation, in a variety of appetitive and
Two general approaches, sometimes described as aversive paradigms and with several species including
interventive and correlative, have been used to identify rodents, birds, fish, molluscs and insects, is that con-
these proteins. In the first, the behavioral conse- centrations of antibiotic sufficient to inhibit more
quences of either blocking protein function or pre- than eighty percent (e.g., 80%) of all protein synthesis
venting protein synthesis from occurring during or in the brain for several hours are, perhaps surprising-
after training an animal on a particular task have ly, without effect on performance of already-learned
been studied. If protein synthesis is necessary for tasks or other aspects of behavior. However, if the in-
memory formation, animals in which the synthesis hibitors are injected within specific time windows rel-
has been blocked at the time of training should be ative to the time of training (or testing), they will pro-
amnesic for the task when subsequently tested. The duce amnesia in animals tested twenty-four hours or
second approach attempts to measure an increase in more later. Behavioral controls rule out the possibility
the synthesis or turnover of particular proteins as a that these amnestic effects are due to some form of
result of the training experience. Both approaches state dependency, and although the inhibitors have
have methodological pitfalls. Preventing protein ex- a variety of less specific biochemical effects (notably
pression or blocking function may have general ef- increasing the concentrations of intracellular amino
fects on behavior, sensory or motor performance, or acids, including several that are neurotransmitters
arousal, and therefore exert an effect on memory and can be neurotoxic), it is generally agreed that
nonspecifically. Similarly, increases in protein synthe- they do indeed exert their amnestic effect by prevent-
sis following training could be the consequence of ing the synthesis of proteins necessary for memory
nonspecific behaviors such as the motor activity in-
formation. More recently attention has turned to the
volved in carrying out the training task. Ruling out
blockade of specific proteins or protein classes. Glyco-
such alternative explanations requires devising care-
protein synthesis can be blocked with the metabolic
ful control experiments.
analogue 2-deoxygalactose. The expression of specif-
ic proteins can be prevented in mice or fruit flies by
Interventive Strategies transgenic techniques such as the use of inducible
knockouts, and in many species by antisense RNAs
Proteins are synthesized in a series of steps begin-
targeted against the specific transcribed RNA se-
ning with the copying (transcribing) of strands of
DNA into messenger RNA (mRNA), after which the quences coding for a particular protein. Appropriate
mRNA is translated into the sequence of amino acids antibodies can also functionally block proteins, espe-
that constitutes the protein. Many proteins, especially cially membrane proteins, whose structure makes
those of the cell and synaptic membranes, are further them accessible in vivo. Thus a monoclonal antibody
modified posttranslationally. For instance, to synthe- raised against a key synaptic membrane constituent,
size glycoproteins, which are key constituents of the the neural cell adhesion molecule N-CAM, a central
cell membrane, it is necessary to add sugar molecules player in the processes of cell-cell recognition, is am-
to the protein chains before transporting them to the nestic if injected into rats trained in a passive avoid-
cellular sites at which they will function. Inhibiting ance task some six to eight hours after training. An-
any one of the steps in this synthetic sequence from tisense molecules that prevent the synthesis of the
the DNA to the finished molecule may block the syn- amyloid precursor protein, APP, or N-CAM, are also
thesis of a protein or render it nonfunctional, and amnestic. Mouse knockouts lacking the cyclic AMP re-
hence interfere with memory formation. The first ex- sponse element binding protein (CREB) also show
periment of this sort was made in the mid-1960s, characteristic learning deficits.
PROTEIN SYNTHESIS IN LONG-TERM MEMORY IN VERTEBRATES 555
Correlative Studies been shown to be relevant in mammalian learning.

Other imaging techniques, such as in situ hybridiza-
Like the interventive approach, correlative
tion or immunocytochemistry, can define specific cel-
studies began in the 1960s. Because it is unlikely that
lular regions, such as dendritic spines, as showing in-
there will be measurable increases in the total amount
creased levels of particular messenger RNAs or
of proteins in general or of specific proteins in partic-
proteins.
ular during memory formation, the initial approach
was to measure the rate of protein synthesis during
training by the use of a radioactive precursor. A radio- Memory Consolidation and Memory Phases
actively labeled amino acid, injected into the blood-
stream, is taken into the brain and there becomes in- The demonstration of the time windows during
corporated into protein. The amount of radioactivity which protein synthesis inhibitors are amnestic has
found in the protein after a fixed time interval then been an important piece of evidence in developing
depends, among other factors (which need to be con- stage theories of time-dependent processes in memo-
trolled for), on the rate of synthesis of the protein. If ry formation. Thus it was early demonstrated that if
more radioactivity is found in brains of trained than protein synthesis inhibitors were administered
of control animals, this is assumed to indicate that the around the time of training in, for instance rodents
training procedure has resulted in enhanced synthe- or chicks, then animals could learn and show memory
sis (or turnover), and the behavioral question then be- retention for periods of an hour or so, after which a
comes that of ensuring that it is memory formation progressive amnesia set in. This was taken to indicate
rather than some other aspect of the task that has in- a distinction between short-term memory, not requir-
creased the synthesis. ing new protein synthesis, and long-term memory, for
which such synthesis was necessary. Giving the inhibi-
Some of the clearest evidence for enhanced pro- tor one to two hour posttraining no longer results in
tein synthesis using this approach came from studies amnesia and hence memory was said to be now stable
of early learning in the chick, especially imprinting and protein synthesis independent. However, this
and one-trial passive avoidance training. These tasks proved to be simplistic; a second time window four to
involve strong and biologically programmed learning six hours downstream of the training experience was
in an otherwise naive animal, and therefore maximize found in a number of paradigms during which memo-
the chance of finding changes. For example, training ry consolidation once more becomes sensitive to pro-
on the passive avoidance task results in a long lasting tein synthesis inhibitors.
increase in incorporation of radioactive amino acids The current working hypothesis is that the initial
into the proteins of specific brain regions. stages of memory formation involve transient synap-
To study such increases in more detail, it is neces- tic membrane events. These events include the phos-
sary to know which brain regions might be involved. phorylation of membrane proteins (including the
Here, an autoradiographic mapping technique can presynaptic protein known as B50 or GAP43) by a
be employed in which, after incorporation of the ra- membrane-bound enzyme, protein kinase C. This
dioactivity, the brains are sectioned and apposed to phosphorylation step activates a cycle of intracellular
X-ray film, and the specific regions showing training- second messengers, including calcium ions, and tran-
related increases identified by image analysis. Such scription factors, such as CREB, which in turn trigger
techniques have been used to show enhanced protein a genomic response in the neuronal cell nucleus. The
synthesis in the rat hippocampus during a variety of initial genomic response is to switch on a family of
learning paradigms. specific immediate early genes whose protein prod-
ucts include in particular c-fos, c-jun, and zif. These
The next task becomes that of identifying which proteins are expressed only during the early phases
of the many proteins are involved. Various approach- of neuronal plasticity, when neurons are growing or
es are possible. Subcellular fractionation can show in actively differentiating, and they have been shown to
which cellular compartments the new proteins are increase dramatically in concentration in a number of
most concentrated. In both rat and chick there are in- learning tasks, including brightness discrimination in
creases both in soluble proteins such as tubulin and the rat and passive avoidance in the chick, as well as
in synaptic membrane constituents such as N-CAM. hippocampal long-term potentiation. But c-fos and c-
In an unusual training task in goldfishin which the jun, although they are excellent markers to show
animal has a float attached to its belly that inverts it, where in the brain neural plasticity is occurring, are
so that it has to learn to swim correctly once morea themselves only intermediates; their production acts
different class of low-molecular weight soluble glyco- as a trigger for the activation of further genes (late
proteins, named ependymins, has been identified by genes) whose products include the proteins and gly-
similar techniques. This class of molecules has also coproteins already mentioned. It is the enhanced syn-
556 PSYCHOPHARMACOLOGY
thesis of these proteins, in the so-called second syn- pression of old but newly activated memories. These
thetic wave, four to six hours downstream of training, proteins are not, however, in themselves specific to
which makes consolidation once more sensitive to the the particular memory. What conveys specificity is the
general synthesis inhibitors as well as to specific an- pattern of neurons whose connections are modified
tisense or antibodies against, for instance, the families by the learning experience; the proteins are part of
of cell adhesion molecules such as N-CAM. These ad- the housekeeping processes involved in modifying
hesion molecules have sticky glycosylated extracellu- those connections. Identifying more precisely the
lar domains. Projecting both from pre- and postsyn- proteins involved, their locations, and their functions
aptic sides, they can bind together rather like a will help substantially in the development of theories
molecular form of Velcro, holding the synaptic junc- of memory formation, in addition to being of major
tion in a specific configuration. It is therefore as- intrinsic neurobiological interest, and of potential rel-
sumed that these are the ultimate effector proteins re- evance to the treatment of such conditions as Al-
quired for lasting remodelling of synapses. zheimers disease.
See also: MEMORY CONSOLIDATION: MOLECULAR
Consolidation and Reconsolidation AND CELLULAR PROCESSES; MEMORY
CONSOLIDATION: PROLONGED PROCESS OF
Recently even this conclusion concerning stable
REORGANIZATION
memory has been called into question. If an animal
is trained on a task, and some timehours to days Bibliography
latergiven a reminder by being exposed once more Davis, H. P., and Squire, L. R. (1984). Protein synthesis and memo-
to the original training situation, then the previously ry: A review. Psychological Bulletin 96, 518559.
stable memory again becomes labile and sensitive to DeZazzo, J., and Tully, T. (1995). Dissection of memory formation:
From behavioral pharmacology to molecular genetics. Trends
protein synthesis inhibitors. For some researchers, re-
in Neuroscience 18, 212217.
minder plus inhibitor results in lasting amnesia, im- Dudai, Y. (1989). Neurobiology of memory. Oxford: Oxford Universi-
plying that in some way reactivating a memory re- ty Press.
quires retraversing the same biochemical pathway Goelet, P., Castelluci, S., Schachner, S., and Kandel, E. R. (1986).
that was initially engaged. Others find that the effect The long and the short of long-term memorya molecular
framework. Nature 322, 19423.
of the inhibitors is transient, suggesting a temporary
Izquierdo, I., and Medina, J. H. (1997). Memory formation: The
blockade of access to an otherwise unimpaired memo- sequence of biochemical events in the hippocampus and its
ry. The phenomenon is intriguing but its interpreta- connection to activity in other brain structures. Neurobiology of
tion remains an arena of active debate. Learning and Memory 68, 285316.
Mileusnic, R., Lancashire, C., and Rose, S. P. R. (1999). Sequence
specific impairment of memory formation by NCAM antisen-
Conclusion se oligonucleotides. Learning and Memory 6, 120127.
Nader, K., Schafe, G. E., and LeDoux, J. E. (2000). Fear memories
Protein synthesis is universally necessary for the require protein synthesis in the amygdala for reconsolidation
formation of long-term memory, although some after retrieval. Nature 406, 722726.
forms of newly acquired memory may persist for long Radyushkin, K. A., and Anokhin, K. V. (1999). Recovery of memory
periods in the absence of such synthesis. The proteins in chicks after disruption during learning: The reversibility of
amnesia induced by protein synthesis inhibitors. Neuroscience
involved are synthesized in increased amounts in spe- and Behavioral Physiology 29 (1), 3136.
cific cells in the hours following a training experience Rose, S. P. R. (1992). The making of memory. London: Bantam.
and during the period of memory consolidation. Dif- (2000). Gods organism? The chick as a model system for
ferent proteins are synthesized at different times after memory studies. Learning and Memory 7, 117.
training in a time- and space-dependent sequence; in Steven P. R. Rose
the first phase they include members of the immedi-
ate-early family, such as c-fos and c-jun. Later stages
involve the synthesis of the microtubular protein tu-
bulin and glycoprotein components of synaptic and
dendritic membranes, including members of the cell PSYCHOPHARMACOLOGY
adhesion family. Others certainly remain to be identi- See: COGNITIVE ENHANCERS; DRUGS AND
fied. If their expression or function is prevented, am- MEMORY; PHARMACOLOGICAL TREATMENT
nesia results. Some may also be required for the ex- OF MEMORY DISORDERS
R
RAMN Y CAJAL, SANTIAGO was subject to great controversy. Shortly after the Ger-
(18521934) man botanist Jacob Schleiden, the German naturalist
Theodor Schwann, and the German pathologist Ru-
Santiago Ramn y Cajal, born in the small Spanish dolf Virchow proposed the cell theory in the 1830s,
town of Petilla de Aragn on May 1, 1852, was a major Joseph von Gerlach, Sr., and Otto Friedrich Karl Di-
figure in the history of neuroanatomy. As related in eters suggested that nerve tissue was special in the
his delightful autobiography, he was somewhat missense that nerve cells are not independent units but
chievous as a child and determined to become an art- instead form a continuous syncytium or reticular net.
ist, much to the consternation of his father, a respect- This concept was later refined by the Italian physician
ed local physician. Eventually, however, Ramn y Camillo Golgi, who concluded that the axons of nerve
Cajal entered the University of Zaragoza, and re- cells form a continuous reticular net, whereas their
ceived a degree in medicine in 1873. As a professor dendrites serve a nutritive role, much like the roots
of anatomy at Zaragoza, his interests were mostly in of a tree.
bacteriology until 1887, when he visited Madrid and
first saw through the microscope histological sections Using Golgis technique, Ramn y Cajal almost
of brain tissue treated with the Golgi method, which immediately arrived at the opposite conclusion, from
had been introduced in 1873. his examination first of the cerebellum, and then of
a wide variety of other sensory and motor systems. In
Although few workers had employed this tech- short, he proposed that neurons interact by way of
nique, Ramn y Cajal immediately saw that it offered contact rather than continuity. His work on both the
great hope in solving one of the most vexing and fun- mature and the developing nervous system (he dis-
damental problems in morphology: How do nerve covered the growth cone in 1890) provided the best
cells interact with each other? This realization galva- evidence for the neuron doctrine (that neurons are
nized and directed the rest of his scientific life, which independent units or cells, as in other tissues) until
was extremely productive in originality, scope, and the introduction of the electron microscope in the
accuracy. 1950s.
To place Ramn y Cajals work in historical per- Important as this evidence was, Ramn y Cajals
spective, one must recall that, while studies of the greatest conceptual achievement was the law of dy-
gross anatomy of the brain can be traced back as far namic polarity. Based on his analysis of Golgi-
as Greek philosopher Aristotle, the first real insight impregnated neurons in the retina and other sensory
into the disposition of its major fiber tracts was not systems, where the direction of information flow from
gained until the middle of the nineteenth century, the periphery to the central nervous system seems ob-
and even then the interpretation of this information vious, Ramn y Cajal concluded that, in general, the
557
558 RECONSTRUCTIVE MEMORY
age of eighty-two in 1934, he had become one of the

most famous and revered Spaniards of the twentieth
century. In addition to the neurohistological work
outlined herein, he wrote widely appreciated books
on color photography, advice to young investigators,
and well-known Spanish aphorisms.
Ramn y Cajals description of the organization
of cerebral cortical circuitry is unparalleled in depth
and breadth. As early as 1894 he advanced the hy-
pothesis that the remarkable intellectual growth seen
in people who engage in continuous mental exercise
is due to an enhanced elaboration of axon collaterals
and dendritic processes within cortical circuitry.
See also: GUIDE TO THE ANATOMY OF THE BRAIN:

NEURON; GUIDE TO THE ANATOMY OF THE
BRAIN: SYNAPSE
Bibliography
De Filipe, J., and E. G. Jones. (1988). Cajal on the cerebral cortex. New
Ramn y Cajal, S. (18991904). Textura del sistema nervioso del hom-
bre y de los vertebrados, 3 vols. Madrid: N. Moya.
(19091911). Histologie du systme nerveux de lhomme et des
vertbrs, trans. L. Azoulay. 2 vols. Paris: Norbert Maloine.
(1928). Degeneration and regeneration of the nervous system.
trans. and ed. R. M. May. 2 vols. London: Oxford University
Press.
(1989). Recollections of my life, trans. E. H. Craigie with J.
Santiago Ramn y Cajal (Library of Congress) Cano. Cambridge, MA: MIT Press.
(1990). New ideas on the structure of the nervous system in man
and vertebrates, trans. N. Swanson and L. W. Swanson. Cam-
dendrites and perikaryon of a neuron receive infor-
mation, whereas its axon transmits information. This Larry W. Swanson
brilliant generalization allowed him to lay out the
basic organization of circuitry throughout the ner-
vous system. This research was summarized in the
monumental three-volume work Textura del sistema RECONSTRUCTIVE MEMORY
nervioso del hombre y de los vertebrados, published be-
tween 1899 and 1904, then expanded and translated Subjectively, memory feels like a camera that faithful-
into the definitive French edition, Histologie du systme ly records and replays details of our past. In fact,
nerveux de lhomme et des vertbrs (19091911). This ac- memory is a reconstructive process prone to systemat-
count deals systematically with all parts of the mam- ic biases and errorsreliable at times,and unreliable
malian nervous system and many aspects of its devel- at others. Memories are a combination of new and old
opment, and provides a great deal of information on knowledge, personal beliefs, and ones own and oth-
the organization of the nervous system in fish, am- ers expectations. We blend these ingredients in form-
phibians, reptiles, and birds. ing a past that conforms to ones haphazardly accu-
rate view of oneself and the world.
Ramn y Cajal next examined in great detail the
histological changes that can be observed during the
degeneration and regeneration of neural tissue fol- Reconstructing the Past
lowing damage. The results of this work were summa- Traditionally, psychologists were interested in
rized in another monumental work that is still well the temporal retention of information. Since the early
worth reading, the two-volume Degeneration and Re- 1930s, many psychologists have shifted their focus
generation of the Nervous System, first published in from the quantity of memory to its accuracy (Koriat,
Spanish in 19131914. Goldsmith, and Pansky, 2000). The British psycholo-
Ramn y Cajal received many honors, including gist Frederic C. Bartlett (1932) conducted one of the
the Nobel Prize, which he shared with a contentious first systematic investigations of memory accuracy. He
Golgi in 1906. By the time Ramn y Cajal died, at the asked subjects to read a legend about Indian hunters
RECONSTRUCTIVE MEMORY 559
called The War of the Ghosts and then to retell it ple draw inferencesnot necessarily accurateabout
to another subject who had not read it. The second their present and past experiences.
subject then told the story to another subject, and so
Yet another way to demonstrate memorys at-
on, until ten subjects had heard it. The story involves
tempt at synthesis is to present subjects with succes-
two young Indian hunters who meet a group of men
sive, thematically related slides depicting common
in a canoe, who, in turn, invite the hunters to join
routines like going grocery shopping. One slide
them in battle upriver. One young Indian accepts and
shows a woman putting a box of items into her shop-
the other declines. During battle, the young Indian is
ping cart. The next slide shows several oranges on the
wounded and realizes that the men of the war party
ground. When subjects are asked later to recognize
are ghosts. He returns home, recounts his tale, and
slides that had previously been shown, they mistaken-
dies the next morning.
ly say that they saw a slide depicting the woman re-
Bartlett found that subjects retained the overall moving an orange from the bottom of a pile of or-
gist of the story but that they also revised the story, anges (Hannigan and Tippens-Reinitz, 2001). They
systematically omitting and modifying details. For ex- make this causal inference because people naturally
ample, subjects omitted mystical references, such as attempt to piece together the fragments of their past
ghosts, which are not part of Westerners worldview; in order to make memory as coherent as possible.
they embellished other details. In the original story,
the second Indian declined to join the party because
his relatives would not know his whereabouts. By the Misinformation
tenth retelling, one subject explained that this Indian Work on the misinformation effect further
refused because his elderly mother was dependent on demonstrates the ease with which accumulated infor-
him, a revision that manifests Western concepts of a mation skews memory (Loftus, 1979). In the misinfor-
sons responsibilities in general and perhaps that sub- mation effect, misleading information about an event
jects family ties in particular. Another common from ones past reduces the accuracy of the memory
change was that subjects tended to add a moral, possi- of an event. In one study, Elizabeth Loftus and col-
bly because stories in Western culture often have mor- leagues showed subjects a simulated automobile-
als. Bartlett concluded, Remembering . . . is an imag- pedestrian accident (Loftus, Miller, and Burns, 1978):
inative reconstruction, or construction, built out of a vehicle stops at an intersection, turns right, and
the relation of our attitude towards a whole mass of then hits a pedestrian. Half the subjects viewed a stop
organized past reactions or experiences (p. 213). sign at the intersection. After viewing the scene, these
subjects were asked a question that mentioned either
Bartletts study exemplifies how time and retell- a stop sign or a yield sign. In a final memory test,
ing distort the memory of stories. Another study con- these subjects were asked whether they saw a stop sign
ducted in the early 1930s using ambiguous drawings or a yield sign. When the subjects were asked a ques-
showed that what we are told that we are viewing easi- tion consistent with what they had seen, they chose
ly distorts visual material. If people are shown two cir- the correct sign 75 percent of the time. However,
cles and a line and are told that the picture represents when the question was inconsistent with what they
either glasses or dumbbells, subjects later drawings of had seen, they chose the correct sign only 41 percent
the original picture will assume the suggested appear- of the time. These investigators concluded that some
ance (Carmichael, Hogan, and Walter, 1932). subjects had initially encoded a stop sign in memory
but that the subsequent mention of a yield sign al-
There are many other studies that demonstrate
tered their memory.
the malleability of memory for words, stories, and pic-
tures. For example, Henry Roediger and Kathleen Does the new information alter the original mem-
McDermott (1995) altered a procedure originally de- ory trace, or does it coexist with the original informa-
veloped by James Deese in which people study lists of tion in memory (Ayers and Reder, 1999)? According
closely related words like bed, pillow, tired, and dream. to the altered-trace view, the original memory is
When later asked to recall studied words, subjects fre- changed permanently and is inaccessible to recollec-
quently claim that they saw other words like sleep that tion. According to the coexistence view, the original
were not presented but are related to those that were. information is still accessible with the right retrieval
Subjects often assert these false memories with a cues. The issue of memorys permanence remains a
high degree of confidence and detail (e.g., that a male fundamental, unresolved question in memory re-
as opposed to a female voice spoke the word). There search. But whatever your view about the underlying
is some debate about whether subjects generate the memory traces, it is clear that the memory reports of
word sleep while studying the word list or later, when subjects are changed, and many subjects appear to be-
asked to recall the entire word list. In either case, peo- lieve strongly in their misinformation memories.
560 RECONSTRUCTIVE MEMORY
Figure 1 Pickrell asked subjects to recall as much as possible

about four childhood event descriptions that a rela-
tive had provided. Three of these events were true,
and one was false: that the subject had been lost in a
shopping mall at the age of five for an extended time
and had been rescued by an elderly woman and re-
united with the family. In three suggestive interviews,
during which subjects were led to believe all the
events occurred, subjects remembered the real events
about 70 percent of the time and the false ones about
25 percent of the time (see Figure 1).
Imagination offers another way to implant false
memories. Subjects are asked to imagine in detail an
event that never occurred. Later, they are asked to
rate their confidence that the event truly happened.
The act of imagination typically causes subjects to in-
crease their confidence in the reality of these events.
Although some researchers argue that certain memo-
ries are highly resistant to suggestion and imagina-
tion, others have shown that it is even possible to in-
crease peoples confidence that they had witnessed
demonic possession as a child (Mazzoni, Loftus, and
Kirsch, 2001). These studies indicate that implanta-
tion of entirely false memories is possible.
Once implanted, the false memory is often barely
distinguishable from real ones. Research has shown
Although the evidence indicates that our memo- that there can be statistical differences between a
ries are malleable and easily manipulated, there are group of real memories and a group of false ones: For
circumstances in which memory is relatively resistant example, the real memories possessing more sensory
to change. For example, if people publicly state that detail (Heaps and Nash, 2001; Schooler, Gerhard,
they remember a detail, subsequent suggestions are and Loftus, 1986). But people can give detailed de-
less likely to induce a change of mind. There is also scriptions of their false memories that sometimes lead
resistance to changed recollection in the face of gross them and others to regard the memories as real.
disparities between clearly perceived details and con- Moreover, it is difficult to determine the truth or falsi-
tradicting misinformation. It is also possible to reduce ty of a single memory report.
misinformation effects by warning people about mis-
Another line of research aims to determine
leading messages or by requiring subjects to deter-
whether true and false memories elicit different brain
mine the precise source of the misinformationfor
activity. In such work, subjects read a list of closely re-
example, Did I see the flat tire in the film, or did I
lated words and later try to recognize whether or not
hear or read about it after I saw the film? (Loftus,
they had previously seen those words and other novel
1997). Thus, memory is reconstructive, and recon- but related words. During the recognition phase of
structions are susceptible tobut not powerless the experiment, subjects brain activity is monitored
againstsubsequent misleading information. by sophisticated neuroimaging tools like magnetic
resonance imaging (MRI) or event-related potentials
(ERPs). There is some preliminary evidence that
Implanted Memories neuroimaging may permit scientists to glimpse the
The previous examples demonstrate the disturb- neural signatures of true and false memories (Fabiani,
ing ease with which the details of memory can be ma- Stadler, and Wessels, 2000); however, more work is
nipulated. But it doesnt stop thereit is also possible needed to confirm the utility of this approach.
to implant entire false memories. People can be led
to believe that, as children, they were lost in a shop-
ping mall or that they had knocked over a punch bowl Evaluation and Attribution
at a wedding and spilled punch on the brides parents It is clear that memory can fail in a variety of ways.
(Hyman, Husband, and Billings, 1995; Loftus and However, the precise reason why memory fails is less
Pickrell, 1995). In a series of interviews, Loftus and clear. Larry Jacoby and others have shown that the
REHABILITATION OF MEMORY DISORDERS 561
manner in which people evaluate their present pro- Koriat, A., Goldsmith, M., and Pansky, A. (2000). Toward a psy-
cessing in light of the past may explain in part both chology of memory accuracy. Annual Review of Psychology 51,
481537.
how and why memory fails. For example, the ease Loftus, E. F. (1979). Eyewitness testimony. Cambridge, MA: Harvard
with which a memory comes to mind after exposure University Press.
to misinformation or after imagining the memory in (1997). Creating false memories. Scientific American 277, 70
question may rightly or wrongly lead the person to 75.
believe that the memory is real. In fact, unless there Loftus, E. F., Miller, D. G., and Burns, H. J. (1978). Semantic inte-
gration of verbal information into a visual memory. Journal of
is another, more likely, reason or source to explain Experimental Psychology 4, 1931.
why a memory or experience currently feels familiar, Loftus, E. F., and Pickrell, J. E. (1995). The formation of false
people will typically attribute feelings of familiarity to memories. Psychiatric Annals 25, 720725.
past experience (Jacoby, Kelley, and Dywan, 1989; Mazzoni, G. A. L., Loftus, E. F., and Kirsch, I. (2001). Changing
Whittlesea and Williams, 2001). Thus, it is possible to beliefs about implausible autobiographical events: A little
plausibility goes a long way. Journal of Experimental Psychology:
influence memory by changing the way in which the Applied 7, 5159.
present experience is processed, evaluated, and then Roediger, H. L., III., and McDermott, K. B. (1995). Creating false
attributed to the past. memories: Remembering words not presented in lists. Journal
of Experimental Psychology: Learning, Memory, and Cognition 21,
803814.
Conclusion Schooler, J. W., Gerhard, D., and Loftus, E. F. (1986). Qualities of
the unreal. Journal of Experimental Psychology: Learning Memory,
Far from a reliably faithful rendering of the past, and Cognition 12, 171181.
memory is a reconstruction that usually retains the Whittlesea, B. W. A., and Williams, L. D. (2001). The discrepancy-
gist but not the details of bygone experiences. The re- attribution hypothesis: II. Expectation, uncertainty, surprise,
counting of ones past, the exposure to misleading and feelings of familiarity. Journal of Experimental Psychology:
Learning, Memory, and Cognition 27, 1433.
postevent information and suggestion, integration of
thematically related material, and imagination are Elizabeth F. Loftus
several of the means by which memory is construct- Rick L. Leitner
edor misconstructed. Once reconstructed, the orig- Revised by Daniel M. Bernstein and Elizabeth F. Loftus
inal memory may prove elusive. Given the potential
fallibility of our recollections, it is surprising that
memory functions as well as it does.
REHABILITATION
See also: FALSE MEMORIES See: ELECTROCONVULSIVE THERAPY AND
MEMORY LOSS; HEAD INJURY;
Bibliography PHARMACOLOGICAL TREATMENT OF
Ayers, M. S., and Reder, L. M. (1999). A theoretical review of the MEMORY DEFICITS; REHABILITATION OF
misinformation effect: Predictions from an activation-based MEMORY DISORDERS
memory model. Psychonomic Bulletin & Review 5, 121.
Bartlett, F. C. (1932). Remembering: A study in experimental and social
psychology. New York: Macmillan.
Carmichael, L., Hogan, H. P., and Walter, A. A. (1932). An experi-
mental study of the effect of language on the reproduction of REHABILITATION OF MEMORY
visually perceived form. Journal of Experimental Psychology 15, DISORDERS
7386.
Fabiani, M., Stadler, M. A., and Wessels, P. M. (2000). True but not Memory impairment is one of the most pervasive and
false memories produce a sensory signature in human lateral- debilitating consequences of a range of neurological
ized brain potentials. Journal of Cognitive Neuroscience 12, 941 conditions, including closed-head injury, aneurysm,
949.
Hannigan, S. L., and Tippens-Reinitz, M. T. (2001). A demonstra- stroke, encephalitis, tumor, anoxia, Korsakoffs syn-
tion and comparison of two types of inference-based memory drome, and dementia. This inability to remember re-
errors. Journal of Experimental Psychology: Learning, Memory, cent experiences or to acquire any new long-term
and Cognition 27, 931940. memories is frequently resistant to rehabilitation. Re-
Heaps, C. M., and Nash, M. (2001). Comparing recollective expe- cent empirical and theoretical advances in psychology
rience in true and false autobiographical memories. Journal of
Experimental Psychology: Learning, Memory, and Cognition 27, and neuroscience, however, have broadened under-
920930. standing of the cognitive and neural mechanisms of
Hyman, I. E., Jr., Husband, T. H., and Billings, F. J. (1995). False memory and have provided a new perspective on
memories of childhood experiences. Applied Cognitive Psychol- memory rehabilitation. It is now apparent that mem-
ogy 9, 181197. ory is not a unitary entity; it can break down in a vari-
Jacoby, L. L., Kelley, C. M., and Dywan, J. (1989). Memory attribu-
tions. In H. L. Roediger III, and F. I. M. Craik, eds., Varieties ety of ways, and although some aspects of memory are
of memory and consciousness: Essays in honour of Endel Tulving. seriously compromised, others remain unaffected.
Hillsdale, NJ: Erlbaum. For rehabilitation to be successful, it must take ac-
562 REHABILITATION OF MEMORY DISORDERS
count of the spared as well as the impaired cognitive egies, mnemonic techniques can be used to help
and neural processes. Finding ways to tap into intact people acquire specific pieces of new information
processes to accomplish the kinds of memory tasks important in their daily lives, but they do not pro-
that are impaired, however, has not been easy, and so vide broad-based memory improvements because of
this approach to rehabilitation remains only one of their lack of applicability to new materials and con-
several. texts. Because mnemonic strategies are useful only
Approaches to rehabilitation fall into two broad for mildly impaired patients, these strategies likely
categories: those that attempt to remediate the un- rely on residual memory function (Glisky and Glisky,
derlying impairment and achieve broad improve- 2002).
ments in memory and those that intervene at a behav-
ioral level and try to achieve specific functional
outcomes. The second approach has been more fruit- External Aids and Environmental Supports
ful: patients have been able to take advantage of ways Compensating for memory impairments has
to bypass or compensate for their deficits and thus at-
often involved providing external aids for remember-
tain some level of independence in their everyday
ing, such as notebooks and diaries; bell timers and
lives. But general mnemonic improvements, which
alarm watches; environmental restructurings such as
might apply broadly across a range of memory situa-
labels and directions; and, most recently, microcom-
tions, have proved elusive.
puters, recorders, pagers, and electronic organizers.
Although the simplest of theselabels and signage
Practice and Rehearsal Techniques can be used effectively by even the most severely im-
Extensive practice or rehearsal is essential for the paired individuals, aids such as notebooks, comput-
acquisition and retention of information by memory- ers, and electronic organizers require training and
impaired individuals. The effects of practice, howev- extensive practice and may be beneficial only for
er, are highly specific and tend not to generalize be- mildly impaired individuals; more severely impaired
yond the trained materials. It is, therefore, important patients may be unable to learn how to use these de-
that practice be directed towards information that is vices. Barbara Wilson and her colleagues (2001) have
useful in everyday life. Simple repetitive exercise or reported considerable success with a simple paging
rote rehearsal, however, is unlikely to achieve long- device that requires little learning, can be used effec-
term retention. Instead, short periods of distributed tively by even very severely impaired patients, and
practice are more effective. One such method, spaced can be programmed to meet the needs of individual
retrieval, involves repeated practice at retrieving to- patients. The device provides people with needed in-
be-learned information at gradually increasing reten- formation at various times of the day, enabling them
tion intervals. This technique has been successful in to perform important daily functions that they would
teaching memory-impaired patients, including those otherwise be unable to remember.
with Alzheimers disease, various kinds of information
important in their daily lives, including name-face as-
sociations, locations of objects, and aspects of orienta- Vanishing Cues
tion (Camp and McKitrick, 1992). The method may
The method of vanishing cues was developed ex-
rely on residual memory function or tap into other
pressly to take advantage of the implicit learning and
preserved memory processes.
memory abilities that are spared in amnesic patients
in order to teach them useful information and skills.
Mnemonic Strategies This technique, which involves the gradual withdraw-
Teaching a variety of mnemonic strategies, inal of cues across learning trials, was thought to capital-
cluding visual imagery and verbal elaboration, has ize on patients preserved abilities to retrieve recently
achieved only spotty success with neurologically im- presented information in response to fragment cues.
paired patients. For many patients, the techniques Using this method, Glisky and colleagues (1994)
are too demanding and difficult. They are often effec- demonstrated that even people with serious memory
tive, however, for individuals with unilateral lesions impairments could acquire a variety of domain-
and material-specific deficits. People with left- specific skills and knowledge, including computer
hemisphere lesions and problems in verbal memory procedures, data entry, database management, and
may benefit from visual-imagery strategies, whereas word-processing. Patients learned this information
those with right-hemisphere lesions and difficulties in despite having little or no memory of the occasions of
visuo-spatial memory may find verbal methods, such learning and were later able to use it in the workplace
as first-letter cuing, more useful. Like rehearsal strat- or in the home.
REINFORCEMENT 563
Errorless Learning Grady, C. L., and Kapur, S. (1999). The use of neuroimaging in
neurorehabilitative research. In D. T. Stuss, G. Winocur, and
Wilson and colleagues (Evans et al., 2000) showed I. H. Robertson, eds., Cognitive Neurorehabilitation. Cambridge
that preventing errors during learning facilitates the UK: Cambridge University Press.
acquisition of new information in people with memo- Wilson, B. A., Emslie, H., Quirk, K., and Evans, J. (2001). Reducing
everyday memory and planning problems by means of a pag-
ry disorders. Errorless learning has been effective in
ing system: A randomized control crossover study. Journal of
teaching memory-impaired patients a range of useful Neurology, Neurosurgery, and Psychiatry 70, 477482.
information, including names, the use of a memory
Elizabeth L. Glisky
notebook and an electronic memory aid, and items of
general knowledge. Other methods, such as vanish-
ing cues and spaced retrieval in combination with er-
rorless learning, may prove beneficial, particularly for
very severely impaired patients, who have to rely on REINFORCEMENT
preserved implicit memory.
In its earliest technical usages, the term reinforcement
implied strengthening, echoing its colloquial usage.
Conclusion It has been applied to a broad range of phenomena
Current understanding of memory disorders fa- in learning, including the operant or instrumental be-
vors a rehabilitation strategy that focuses on achiev- havior studied by B. F. Skinner and the respondent
ing improved functioning rather than on treating un- or classical conditioning procedures of Ivan Pavlov.
derlying impairment. Because the benefits are The term, which also applies to certain types of ma-
specific and limited, it is important to ensure that re- chine learning procedures studied by computer scien-
habilitation is directed towards solving real-world tists and engineers, now refers mostly to cases in
problems. The finding of preserved-memory abilities which behavior has some consequences and, by virtue
in amnesic patients suggests that there may be alter- of these consequences, comes to occur more often.
native routes to usable memory if ways could be found The term reward, sometimes used as a nontechnical
to recruit these intact cognitive and neural structures. synonym, is not equivalent. For example, one can
Although relatively little is known about the reorgani- speak of delivering a reward even without evidence
zation of brain function after damage, recent neuroi- that the reward has an effect on behavior.
maging studies have suggested that such reorganiza- As a classical example of reinforcement, imagine
tion is possible and may have functional a rat in a chamber with a lever and a cup into which
consequences (Grady and Kapur, 1999). Finding ways food pellets can be delivered. If pressing the lever
to encourage reorganization and to tap into pre- does nothing, the rat presses only occasionally. If
served functions will be important tasks for future re- each press produces a food pellet, however, the rat
search in memory rehabilitation. presses the lever more often. The food is called a rein-
forcer, and the rats lever press is said to have been
See also: MNEMONIC DEVICES; PHARMACOLOGICAL reinforced. The response that increases must be the
TREATMENT OF MEMORY DEFICITS one that produced the consequence. For example, if
lever presses produce shock and only the rats jump-
Bibliography ing increases, it would be inappropriate to speak of
Baddeley, A. D., Wilson, B. A., and Watts, F. N. (1995). Handbook either lever pressing or jumping as reinforced. An
of memory disorders. Chichester, UK: John Wiley. ambiguity of usage is that reinforcement sometimes re-
Camp, C. J., and McKitrick, L. A. (1992). Memory interventions in fers to delivery of the reinforcer (the lever press was
Alzheimers-type dementia populations: Methodological and
theoretical issues. In R. L. West and J. D. Sinnott, eds., Every- reinforced means that lever presses produced food)
day memory and aging: Current research and methodology. New and sometimes to the resulting behavior change (the
York: Springer-Verlag. lever press was reinforced means that lever presses
Evans, J. J., Wilson, B. A., Schuri, U., Andrade, J., Baddeley, A., occurred more often because they produced food).
Bruna, O., Canavan, T., Della Sala, S., Green, R., Laaksonen, The sense is usually clear from the context.
R., Lorenzi, L., and Taussik, I. (2000). A comparison of er-
rorless and trial-and-error learning methods for teaching
individuals with acquired memory deficits. Neuropsychological The Operant Nature of Reinforcement
Rehabilitation 10, 67101.
Glisky, E. L., and Glisky, M. L. (2002). Learning and memory im- Edward Thorndikes experiments with animals in
pairments. In P. J. Eslinger, ed., Neuropsychological interven- problem boxes were early examples of studies of rein-
tions. New York: Guilford Press. forcement. Typically, a food-deprived animal was
Glisky, E. L., Schacter, D. L., and Butters, M. A. (1994). Domain-
specific learning and remediation of memory disorders. In M. placed inside a box with food available outside. Soon-
J. Riddoch and G. W. Humphreys, eds., Cognitive neuropsy- er or later, typically by chance, the animal operated
chology and cognitive rehabilitation. Hove, UK: Erlbaum. a device that released the door and freed it from the
564 REINFORCEMENT
box. With repetition it operated the device more and Positive and Negative Reinforcement
more rapidly upon being placed in the box. To de-
A variety of consequences can serve as reinforc-
scribe his findings, Thorndike formulated his law of
ers, ranging from those of obvious biological signifi-
effect. The law went through many revisions, but its
cance, such as food or water or a sexual partner, to
essence was that responses can be made more proba-
minor changes in things seen or heard or touched.
ble by some consequences and less probable by oth-
Consequences are not restricted merely to the pro-
ers; in language closer to Thorndikes, responses with
duction of stimuli. Responses can remove stimuli, as
satisfying effects are stamped in, whereas those with
when operating a switch turns off a light; responses
annoying effects are stamped out.
can prevent stimuli from occurring, as when unplug-
Behavior that is modified by its consequences is ging a lamp before repairing it eliminates the possi-
called operant behavior or, in older and less common bility of a shock; and responses can even change the
usages, instrumental behavior, which is behavior that consequences of other responses, as when replacing
operates on its environment. It is said to be emitted a burned-out light bulb makes the previously ineffec-
because it is primarily determined by its conse- tive response of operating the light switch an effective
quences; it does not require an eliciting stimulus, as response again. Each type of consequence may affect
in reflex relations (e.g., as when a puff of air to the eye later behavior.
produces a blink). Operants are classes of responses
A stimulus is a positive reinforcer if its presenta-
defined by their environmental effects rather than by
tion increases the likelihood of responses that pro-
their topography (their form or the particular muscle
duce it and a negative reinforcer if its removal in-
groups involved). For example, a rat might press a
creases the likelihood of responses that terminate or
lever with its left paw, its right paw, or both paws; it
prevent or postpone it. Negative reinforcers are
might even depress the lever by biting or sitting on
sometimes called aversive stimuli. Responding that
it. But similar movements at the other end of the
turns off an aversive stimulus (e.g., when a rat turns
chamber, far from the lever, would not be lever press-
off a loud noise by pressing a lever) is escape respond-
es no matter how closely they resembled the re-
ing. Responding that prevents or postpones an aver-
sponses that operated the lever.
sive stimulus (e.g., when a rats lever press prevents
Skinner significantly extended the analysis of re- or postpones the delivery of an electric shock) is
inforcement. One of his critical contributions was to avoidance responding.
clarify the difference between operant learning and
the classical or respondent learning that had been
studied by Pavlov (in Pavlovs experiments, the organ- Reinforcement and Punishment
isms responses had no effect on the stimuli that were Consequences can reduce responding instead of
presented). increasing it. For example, if shock is delivered when-
Consider the relation between a red traffic light ever a rat grooms its tail, the rat may groom its tail
and a drivers stepping on the brakes. The red light less often. Responding reduced by its consequences
sets the occasion for stepping on the brakes; this reis therefore punished. Punishment, simply the in-
sponse occurs at a red light and not at a green light verse of reinforcement, has had a more controversial
because it has different consequences in the presence history. Thorndikes early versions of the law of effect
of each stimulus. A stimulus that signals or sets the oc- argued that behavior could be stamped out by an-
casion for different consequences of responding is noyers as well as stamped in by satisfiers, but he later
discriminative. Three terms are involved: the discrim- withdrew the part that dealt with the stamping out of
inative stimulus, the response, and the consequences behavior.
of the response in the presence of that stimulus. A rats reinforced lever pressing can be reduced
For example, imagine a pigeon in a chamber with when presses are punished by shock, but the respond-
a feeder; above the feeder is a small disk or key that ing recovers to earlier levels once punishment is dis-
can be lit from behind with lights of different colors. continued. Because punishment suppresses respond-
Suppose that reinforcement of the pigeons key peck ing only temporarily, some have argued that it is
depends on the presence of a green or a red light. If ineffective. Yet reinforcement would also have to be
the peck produces food only when the key is green, judged ineffective by this criterion. In early studies
the pigeon will come to peck in the presence of green standards for the effectiveness of punishment were
but not red. Its pecking in the presence of green is a different from those for the effectiveness of reinforce-
discriminated operant. Discriminative stimuli corre- ment. Investigations tended to concentrate on the re-
spond to the stimuli colloquially called signals or covery of responding after punishment was discontin-
cues. They do not elicit responses; instead, they set ued rather than on the reduction of responding
the occasions on which responses have consequences. during punishment.
REINFORCEMENT 565
Punishment is effective, but that does not recom- usually is a list of numbers, together with a desired,
mend it. It is usually an undesirable method for elimi- or target, output specifying the response that the sys-
nating behavior. Thorndike and his successors were tem should produce when given that input. The ob-
probably right for the wrong reasons. For example, ject of the task is for the system to learn a response
one major problem is that aversive stimuli used as rule that produces the desired responses for the ex-
punishers have side effects, such as eliciting aggres- ample inputs and also for novel inputs it may receive
sive behavior. in the future.
A reinforcement learning system, in contrast,

Shaping and the Selection of Behavior learns from signals that evaluate the quality of the
Reinforcement does not establish connections or learning systems behavior without explicitly telling
associations between responses and stimuli. Rather, it the system what its behavior should be. Evaluative
changes the probability of classes of responses. Some feedback tells the learning system whether or not, and
responses become more likely and others less likely. possibly by how much, its behavior has improved or
In other words, some responses survive in the organ- deteriorated; or it provides a measure of the good-
isms behavior, whereas others do not. In this respect, ness of the behavior; or it just provides an indication
reinforcement works by selection. It operates on pop- of success or failure. Evaluative feedback does not di-
ulations of responses within the lifetime of the indirectly tell the learner what it should have done or how
vidual organism much as evolutionary selection oper- it should change its behavior. Instead of trying to
ates on populations of organisms over successive match a standard of correctness (i.e., match desired
generations. outputs), a reinforcement learning system tries to
The selective nature of reinforcement is best illus- maximize the goodness of behavior as indicated by
trated by shaping, a procedure that creates new re- evaluative feedback. To do this, it has to actively try
sponses through the successive reinforcement of alternatives, compare the resulting evaluations, and
other responses that more and more closely approxi- use some kind of selection mechanism to guide be-
mate it. For example, a rat may rarely if ever press a havior toward the better alternatives. Whereas super-
lever with a force that approximates its own body vised learning is instructional, reinforcement learn-
weight, but it can be induced to do so by selective rein- ing is selectional, like reinforcement in animal
forcement of stronger and stronger presses. Shaping learning.
begins with whatever behavior is available. At the
start, most presses will probably be weak; but with re- Reinforcement learning is useful from a compu-
inforcement of the strongest ones, the force of press- tational perspective because the evaluation compo-
ing will increase a little, and the criterion for rein- nent, sometimes called a critic, requires much less
forcement then moves up to the strongest of the new information and knowledge than the teacher in a
population of responses. This step in turn produces supervised learning system. It is possible to evaluate
another increase in force and another change in crite- a systems behavior without knowing what the correct
rion, and so on, until the rat is pressing with a force behavior would be. In fact, a critic does not even need
that might never have occurred in the absence of the access to the learning systems actions; it can evaluate
shaping procedure. Reinforcement has had many their consequences on some complex process. It does
useful applications, and shaping has figured in a sig- not need to know anything about the mechanism by
nificant way in many of them, including the training which the actions produce these consequences. A
of animals and the production of speech in autistic complication is that actions can have delayed as well
children. as immediate consequences; evaluative feedback gen-
erally evaluates the consequences of all of the systems
past behavior. Researchers have developed a number
Computational Reinforcement of methods enabling a system to learn efficiently
Learning through reinforcement is studied in under these conditions.
computer science and engineering, where the term
reinforcement learning is applied to learning algorithms Several technologies have benefited from appli-
that include some of the principles of reinforcement cations of reinforcement learning. Although inspired
observed in animal behavior. However, a different by reinforcement phenomena in animal behavior, re-
type of learning is most commonly studied in these search in computational reinforcement learning
fields. This is called supervised learning (or some- seeks mainly to create learning methods that can
times learning by example or learning with a make artificial systems more autonomous and adapt-
teacher) in which a system learns from a collection able; it does attempt to create accurate models of ani-
of examples, where each example consists of an input, mal learning.
566 REINFORCEMENT OR REWARD IN LEARNING: Anatomical Substrates
Conclusion the memory trace that mediates this form of learning. In this
situation, the reinforcing stimulus is viewed as a teaching
The principle of reinforcement emerged in ex-
input that directly elicits the motor response and, when
perimental psychology to account for increases in an
paired with a neutral stimulus, confirms the predictive rela-
animals responding that are due to its consequences.
tionship between the two stimuli. Multiple lines of evidence
Computer scientists and engineers have adapted this
have been brought to bear on the nature of this interaction
principle to create systems that improve their behav-
and its pathways. These studies have revealed that the dorsal
ior over time by learning from their own experiences.
olivary nucleus is the source of the reinforcing input to the
See also: ALGORITHMS, LEARNING; OPERANT cerebellar cortex in support of conditioned motor responses.
BEHAVIOR The section on ELECTRICAL SELF-STIMULATION fo-
Bibliography cuses on the use of electrical stimulation to identify brain
Estes, W. K. (1944). An experimental study of punishment. Psycho-
areas and pathways that mediate reinforcement. These
logical Monographs 57, 263. studies have revealed that brief trains of electrical stimula-
Pavlov, I. P. (1927). Conditioned reflexes, trans. G. V. Anrep. London: tion at specific sites serve as reinforcements for goal-directed
Oxford University Press. behaviors. This artificial activation of the reinforcement
Skinner, B. F. (1938). The behavior of organisms. New York: Apple- pathway is perceived as intensely rewarding, and can sup-
ton-Century-Crofts.
(1969). Contingencies of reinforcement. New York: Appleton-
port a variety of behaviors. Studies using this technique,
Century-Crofts. combined with other approaches, have been used to charac-
(1981). Selection by consequences. Science 213, 501504. terize a widespread brain system that supports reinforced be-
Sutton, R. S., and Barto, A. G. (1998). Reinforcement learning: An in- haviors. This pathway involves the mesolimbic dopamine
troduction. Cambridge, MA: MIT Press. system that arises in the ventral tegmental area and projects
Thorndike, E. L. (1898). Animal intelligence: An experimental
study of the associative processes in animals. Psychological Re-
to the forebrain in the nucleus accumbens. The outputs of
view Monograph Supplements 2, 109. this system strongly influence the extrapyramidal motor sys-
tem, as well as limbic cortical and subcortical areas involved
A. Charles Catania
in the assessment of stimulus contingencies and emotional
Revised by Andrew G. Barto
behavior.
Another brain system in which reinforcement learning
plays a critical role is the extrapyramidal motor system, and
REINFORCEMENT OR REWARD the STRIATUM is a major component of that system. Neurons
IN LEARNING in the striatum detect both the delivery of rewards and the
presentation of stimuli that predict rewards. Some cells fire
[Reinforcements and rewards drive learning. They can add at specific steps in a sequence of events leading to rewards.
effect to otherwise neutral percepts with which they coincide. In addition, the activity of striatal neurons can reflect gener-
They can alter the probability of behaviors that precede them, al states of expectation of rewards predicted by previous expe-
as Thorndike captured in his Law of Effect. How reinforcers rience. The striatum receives signals from the orbitofrontal
and rewards exert these effects is the topic considered in the cortex about reward preferences, and receives global rein-
following four sections. forcement signals from the mesolimbic dopamine system.
How the striatum integrates this information to make predic-
The first section considers the ANATOMICAL pathways
tions about future rewarding events constitutes a major area
of the brain that mediate reinforcement. This section takes
of study on reinforcement mechanisms.
the perspective that pathways for reinforcement are part of
a larger brain system that supports adaptive, goal-directed Together these sections provide an overview of different
behaviors, including feeding, drinking, and reproduction. aspects of reinforcement and reward as a major influence in
From this perspective, reinforcement-driven learning and the formation of memory traces that support acquired behav-
memory are major influences on the likelihood of production ioral responses.]
of goal-directed behaviors. The pathways that mediate rein-
forcement have been delineated by a combination of tradi-
tional anatomical tracing techniques and the use of electrical
stimulation to identify areas where neuronal activation sub- ANATOMICAL SUBSTRATES
stitutes for natural reinforcements. The relationship of reinforcement to behavior is a
A brain system in which reinforcement learning plays useful prologue to a discussion of the brain circuits
a strong role involves pathways through the cerebellum that that mediate reinforcement (see Figure 1). Reinforce-
mediate classical conditioned motor responses. As summa- ment and reward are parts of a larger system that me-
rized in CEREBELLUM, conditioned stimuli and reinforce- diates adaptive behavior that ensures the survival of
ment signals critical to classical motor conditioning con- the animal and the species. Adaptive behaviors, often
verge in the cerebellar cortex and deep nuclei, and compose called goal-directed or appetitive behaviors, include
REINFORCEMENT OR REWARD IN LEARNING: Anatomical Substrates 567
feeding, drinking, and reproduction. Goal-directed Figure 1

behavior can be divided into three phases: initiation,
procurement, and consummatory (Watts and Swan-
son, 2002). In the initiation phase, there is a need or
desire for the goal and a decision to fulfill it. For ex-
ample, when you realize you are hungry, you first
make the decision to eat. The consummatory phase
involves fulfillment of the goalin this case, eating.
In between is the procurement phase where the goal
is sought out. In reinforcement, a part of the procure-
ment phase, the goal or reward increases the proba-
bility of occurrence of the behavior used to obtain it.
There are many factors that influence reinforce-
ment: learning, memory, motivation, emotion, and
locomotion. There must be a memory of previous ex-
perience with the goal to adopt the best strategy to
obtain it. Learning occurs when the animal (or per-
son) adapts its previous experience to the current cir-
cumstances. This may involve a Pavlovian associative
mechanism, through which environmental stimuli
that signal reward availability become conditioned
stimuli, or instrumental learning, to determine which
actions are effective in obtaining the goal, or an inter- A schematic representation of the ventral striatal circuit
action between the two (Parkinson, Cardinal, and mediating adaptive motor responding for reward. The three
Everitt, 2000). Motivation is how hard one is willing major transmitters used in this circuit are indicated. The nucleus
to work to obtain a goal. Put another way, motivation accumbens (ACB) is viewed as a ventral part of a larger striatal
energizes the response needed to obtain the goal. system and a primary anatomical region for integrating
Emotion is very difficult to define precisely. However, glutamatergic inputs from the limbic cortex (a general term for
by introspection we know that events with a high emo- areas in the medial [e.g., prelimbic and anterior cingulate] and
tional impact profoundly affect our ability to learn lateral [e.g., insular] prefrontal cortices that share many
and remember. Once the associations have been structural and functional aspects), hippocampus, and amygdala
with the dopaminergic input originating in the ventral
formed and the strategies selected, any attempt to
tegmental area (VTA). Activation of the nucleus accumbens is
reach the goal requires the skeletomotor system. If
thought to inhibit the ventral pallidum (VP), the major output
the response requires locomotion, then it also re- structure of the ventral striatal circuit. Inhibition of the ventral
quires spatial or contextual learning to navigate pallidum would then disinhibit its downstream targets in the
through the environment to a remembered location mediodorsal nucleus of the thalamus (MD) and
of the reward. However, the response may require pedunculopontine tegmental nucleus (PPN) in the caudal
other forms of motor output instead, such as reaching brainstem. The mediodorsal nucleus provides feedback related
and grasping, or pressing a lever, as in the classic lab- to ventral striatal output via projections to the same areas of
oratory experiment. limbic cortex that project to the nucleus accumbens. The
pedunculopontine nucleus sends descending projections to the
reticular formation and spinal cord to initiate locomotion and
Historical Perspective other motor responses.
The first evidence that reinforcement is mediated

by dissociable brain pathways came from studies of in-
tracranial self-stimulation. This research demonstrat- the brain stem (Wise, 1998; Ikemoto and Panksepp,
ed that focal electrical stimulation of various sites in 1999). Subsequently, the emphasis shifted to the me-
the brain could be reinforcing events (Wise, 1998). solimbic dopamine system that arises in the ventral
Identifying sites involved in self-stimulation made it tegmental area and projects densely to the nucleus ac-
possible to ask questions about the systems with which cumbens. Research showed that systemic administra-
these sites interact and how they ultimately interact tion of dopamine antagonists, which block or inacti-
with motor systems to influence behavior. Mappings vate dopamine receptors, decrease rates of
of the most sensitive reinforcing sites implicated a va- spontaneous locomotion; rates of self-stimulation, re-
riety of structures in the limbic system, including the gardless of the site of stimulation; and rates of instru-
nucleus accumbens, limbic cortex, amygdala, hippo- mental responding for both natural (e.g., food or
campus, and dopaminergic pathways originating in water) and drug reward (e.g., cocaine or heroin)
568 REINFORCEMENT OR REWARD IN LEARNING: Anatomical Substrates
(Salamone, Cousins, and Snyder, 1997; Ikemoto and 1996). A second consequence is the realization that,
Panksepp, 1999). The latter observation is important like the dorsal striatum, most of the output of the nu-
because it implies that drug abuse involves the same cleus accumbens is to nigral- and pallidal-like struc-
reinforcement mechanisms in the brain as natural re- tures (Groenewegen, Wright, and Beijer, 1996). Pro-
wards (Wise, 1997). While these observations are com- jections to nigral-like structures are inputs to the
pelling, it should be noted that dopamine antagonists ventral tegmental area and to the substantia nigra
decrease responding but do not abolish it; therefore, that provide the dopaminergic input to the ventral
the learned response is intact. In addition, dopamine- and dorsal striata, respectively (Redgrave, Prescott,
depleted rats always choose a less preferred but freely and Gurney, 1999). The nucleus accumbens also proj-
available reward, whereas intact animals always do the ects to a ventral pallidum, whose output is organized
opposite. These observations suggest that dopamine in a manner similar to that of the globus pallidus of
is more involved in the arousal or motivational as- the dorsal basal ganglia; its major projections are to
pects of reward than in the informational aspects of the thalamus and brain stem (Groenewegen, Wright,
stimulus-reward association formation (Salamone, and Beijer, 1996).
Cousins, and Snyder, 1997; Parkinson, Cardinal, and
Everitt, 2000). The thalamic projection of the ventral pallidum
is to the mediodorsal nucleus. This nucleus is recipro-
The nucleus accumbens is a major site at which cally connected with the limbic cortex. Thus, the same
dopamine exerts its effects, in part because dopamine cortical regions that project to the nucleus accumbens
levels in the nucleus accumbens increase in response receive feedback related to nucleus accumbens out-
to the presentation of natural reward, the self- put. This pallidothalamocortical loop may have a cog-
administration of most drugs of abuse, or in response nitive function through its influence on the planning
to conditioned stimuli that predict their availability of subsequent actions (Floresco, Braaksma, and Phil-
(Di Chiara, 1998). Moreover, microinjection of dop- lips, 1999). These limbic cortical regions also project
amine antagonists directly into this structure had to the (pre) motor cortex, providing yet another ave-
much the same effect on the above responses as did nue through which the ventral striatal system can in-
systemic administration (Salamone, Cousins, and fluence the extrapyramidal motor system (Groenewe-
Snyder, 1997; Ikemoto and Panksepp, 1999). gen and Uylings, 2000).
The projection from the ventral pallidum to the
The Nucleus Accumbens brain stem is primarily to the pedunculopontine teg-
A major advance in understanding the circuits mental nucleus, an interdigitated population of neu-
that mediate reinforcement came with the realization rons with ascending and descending projections
that the nucleus accumbens was the ventral compo- (Winn, Brown, and Inglis, 1997). The region of de-
nent of a larger striatal system. The dorsal striatum scending projections is sometimes called the mesen-
is part of the basal ganglia or extrapyramidal motor cephalic locomotor region because of its projections
system and has been studied extensively for its role in to the spinal cord and to reticulospinal neurons and
initiating or coordinating movements of individual because electrical stimulation of this region elicits lo-
joints and limbs (Mink). The advantage of viewing the comotion that varies with the intensity of stimulation.
nucleus accumbens as part of this system is that it gen- Thus, low-intensity stimulation elicits walking, where-
erates predictions of its anatomical organization and as increasing levels elicit first a trot and then a gallop
provides insights into the functional organization of in four-legged animals. The region of ascending pro-
behavior. Thus, the close relationship between the jections is sometimes called the midbrain extrapy-
nucleus accumbens, or ventral striatum, and motor ramidal area because of its projections to the dorsal
systems was fully appreciated only after recognition of basal ganglia, which include an input to the dop-
its striatal nature. aminergic neurons of the substantia nigra.
Anatomically, this organization has several conse- Thus, the ventral striatal circuit has two major
quences. One is the appreciation of the predomi- routes to gain access to the motor system and influ-
nance of inputs from the cortical mantle. Whereas the ence behavioral output: directly, via projections to the
cortical inputs to the dorsal striatum arise primarily brain stem, and indirectly, via interactions with the
from somatosensory and motor-cortex and other iso- basal ganglia. Influences on the dorsal basal ganglia
cortcal (i.e., neocortex, or six-layered cortex) regions, primarily involve modulation of its dopaminergic
the inputs to the ventral striatum originate in allocor- input but also include those mediated by the more ex-
tical (a general term for cortex with less than six lay- tensive projections of the midbrain extrapyramidal
ers) regions such as the limbic cortex, amygdala, and area and those transmitted via the thalamocortical
hippocampus (Groenewegen, Wright, and Beijer, loop.
REINFORCEMENT OR REWARD IN LEARNING: Anatomical Substrates 569
Cortical Contributions to Reinforcement Information Flow Through the Nucleus

The major inputs to the nucleus accumbens Accumbens
(other than dopamine) originate in the limbic cortex, Information flow through the nucleus accumbens
amygdala, and hippocampus. In addition to dense, appears to operate on the principle of disinhibition.
convergent inputs to the nucleus accumbens, each of The GABA neurons in the ventral pallidum have high
these structures is reciprocally connected with the rates of spontaneous activity that tonically inhibit
other (Groenewegen and Uylings, 2000). Although a their downstream targets in the mediodorsal nucleus
selective role in reinforcement has been shown for and brainstem (Mogenson et al., 1993). The projec-
each, these interconnections illustrate the interde- tion neurons in the nucleus accumbens are GABAer-
pendent nature of the various aspects of reinforce- gic but have very low rates of spontaneous activity.
ment. For example, learning about the contingency Thus, glutamatergic inputs from the allocortex excite
of reward, to some degree depends on the context in nucleus accumbens neurons, causing them to fire and
which it is learned, whereas the reverse holds true for inhibit the pallidal neurons, which, in turn, disinhibit
contextual learning. the projections from the thalamus and brain stem. An
important aspect of this disinhibitory process is that
The functions attributed to each of these regions it permits or releases behavior rather than initiating
have been based primarily on lesion studies, wherein it directly. This process is consistent with the idea that
a nucleus or a region is inactivated to allow an assess- the command is for a general direction or plan of be-
ment of its contribution to behavior. From these havior, rather than dictating the individual compo-
studies, the limbic cortex has been implicated in nents.
forming the neural representations that underlie as-
sessment of contingency (determining the relation Accordingly, the nucleus accumbens may serve as
between response and outcome) and goal status (the a selection filter, first matching the information about
current value of the reward) (Balleine and Dickinson, the nature of the stimulus and the response require-
1998). The amygdala has been implicated in the pro- ments from the allocortex with the motivational, or
cess whereby affective value is attributed to events. arousing, effects mediated by dopaminergic projec-
The amygdala is actually composed of many nuclei, tions from the brain stem and then selecting the ap-
each implicated in its own set of emotional responses propriate response strategy (Redgrave, Prescott, and
(Aggleton, 1992). In terms of appetitive behavior, the Gurney, 1999). The ventral striatal circuit is well-
basolateral nucleus of the amygdala has received the suited to this task. The projection neurons in the nu-
most extensive study. The basolateral nucleus medi- cleus accumbens are bistablethey exhibit two rela-
ates the ability of conditioned stimuli to influence be- tively stable states: a hyperpolarized down state that
havior (Parkinson, Cardinal, and Everitt, 2000). This results in low levels of spontaneous activity and a
is especially relevant in the context of drug abuse, more depolarized up state in which synaptic inputs
wherein conditioned stimuli that predict drug avail- easily trigger bursts of action potentials (ODonnell et
ability can elicit craving and precipitate relapse, even al., 1999).
after long periods of abstinence (Shalev, Grimm, and
Shaham, 2002). The hippocampus has long been im- The cellular function of dopamine varies with the
plicated in spatial and contextual learning. Lesions of state of the nucleus accumbens neuron such that dop-
this structure decrease exploratory locomotion and amine may act as an excitatory, or facilitory, neuro-
impair the ability of animals to navigate through the transmitter when the neuron is in the up state or may
environment while searching for food and other re- exert an inhibitory effect when they are in the down
wards (Mogenson et al., 1993). state (ODonnell et al., 1999). Such a mechanism
would increase the output of the most active nucleus
Thus, the allocortical inputs to nucleus accum- accumbens neurons and suppress activity in the re-
bens are collectively involved in assessing the rele- maining neurons. The effect would create a winner-
vance of primary and conditioned stimuli and their take-all system where only one functionally related
relation to the required response. These are essential ensemble of neurons is active at a time. This can be
contributions to Pavlovian, instrumental, and contex- conceptualized as different patterns of allocortical ac-
tual learning. Therefore, these inputs to the nucleus tivity selecting or disinhibiting different channels that
accumbens transmit the informational content rele- are maintained as they pass through the nucleus ac-
vant to reinforcement learning (Parkinson, Cardinal, cumbens circuit on their way to the motor system to
and Everitt, 2000). influence specific sets of behaviors.
570 REINFORCEMENT OR REWARD IN LEARNING: Cerebellum
Conclusion ODonnell, P., Greene, J., Pabello, N., Lewis, B. L., and Grace, A.
A. (1999). Modulation of cell firing in the nucleus accumbens.
We can now more precisely define reinforcement Annals of the New York Academy of Sciences 877, 157175.
as a process that selectively potentiates goal-directed Parkinson, J. A., Cardinal, R. N., and Everitt, B. J. (2000). Limbic
behavior, the form and direction of which depends cortical-ventral striatal systems underlying appetitive condi-
tioning. Progress in Brain Research 126, 263285.
upon learned associations relating the significance of
Redgrave, P., Prescott, T. J., and Gurney, K. (1999). The basal gan-
predictive stimuli and environmental conditions. In glia: A vertebrate solution to the selection problem? Neuro-
terms of the neural circuits mediating reinforcement science 89, 1,0091,023.
(see Figure 1), information about stimulus-reward as- Salamone, J. D., Cousins, M. S., and Snyder, B. J. (1997). Behavior-
sociations is generated in the allocortex, where each al functions of nucleus accumbens dopamine: Empirical and
conceptual problems with the anhedonia hypothesis. Neuro-
part mediates a specific set of functions: The limbic
science and Biobehavioral Reviews 21, 341359.
cortex evaluates goal status and its relation to predic- Shalev, U., Grimm, J. W., and Shaham, Y. (2002). Neurobiology of
tive cues, the amygdala generates the emotional value relapse to heroin and cocaine seeking: A review. Pharmacologi-
attached to primary reward and conditioned stimuli, cal Reviews 54, 142.
whereas the hippocampus generates contextual rep- Watts, A. G., and Swanson, L. W. (2002). Anatomy of motivational
systems. In H. Pashler and R. Gallistell, eds., Stevenss handbook
resentations of the environment. In the nucleus ac-
of experimental psychology. New York: John Wiley.
cumbens, these inputs are integrated with the arous- Winn, P., Brown, V. J., and Inglis, W. L. (1997). On the relation-
ing or motivational input subserved by dopaminergic ships between the striatum and the pedunculopontine teg-
neurons in the midbrain. The nucleus accumbens mental nucleus. Critical Reviews in Neurobiology 11, 241261.
then selects the appropriate channel before transmit- Wise, R. A. (1997). Drug self-administration viewed as ingestive be-
haviour. Appetite 28, 15.
ting its output to motor systems. Thus, reinforcement
(1998). Drug-activation of brain reward pathways. Drug and
occurs when the nucleus accumbens integrates infor- Alcohol Dependence 51, 1322.
mational aspects of stimulus-reward associations with
motivational arousal to select and initiate the behav- Richard H. Thompson
ioral strategy best suited to obtaining the goal.
See also: REINFORCEMENT

CEREBELLUM
Researchers have extensively studied the neural sub-
Bibliography
strates mediating reinforcement during classical con-
Aggleton, J. P. (1992). The amygdala: Neurobiological aspects of emo-
tion, memory, and mental dysfunction. New York: Wiley-Liss. ditioning of discrete motor responses. In a seminal
Balleine, B. W., and Dickinson, A. (1998). Goal-directed instru- Science (1986) paper, Richard F. Thompson proposed
mental action: contingency and incentive learning and their a cerebellar model of associative learning in which the
cortical substrates. Neuropharmacology 37, 407419. convergence of conditioned stimulus (CS) and uncon-
Di Chiara, G. (1998). A motivational learning hypothesis of the role
ditioned stimulus (UCS) information at cerebellar
of mesolimbic dopamine in compulsive drug use. Journal of
Psychopharmacology 12, 5467. cortex and the cerebellar deep nuclei form the basis
Floresco, S. B., Braaksma, D. N., and Phillips, A. G. (1999). Tha- of the memory trace in classical conditioning. One of
lamic-cortical-striatal circuitry subserves working memory the central tenets of this modelwhich is based on
during delayed responding on a radial arm maze. Journal of electrophysiological, lesion, and stimulation datais
that reinforcement in the form of the UCS is con-
Groenewegen, H. J., and Uylings, H. B. (2000). The prefrontal cor-
tex and the integration of sensory, limbic and autonomic in- veyed to the cerebellum via climbing fibers originat-
formation. Progress in Brain Research 126, 328. ing in the inferior olive.
Groenewegen, H. J., Wright, C. I., and Beijer, A. V. (1996). The nu-
cleus accumbens: Gateway for limbic structures to reach the
motor system? Progress in Brain Research 107, 485511. The Nature of Reinforcement
Ikemoto, S., and Panksepp, J. (1999). The role of nucleus accum-
bens dopamine in motivated behavior: A unifying interpreta- In its simplest and most widely accepted defini-
tion with special reference to reward-seeking. Brain Research tion, the term reinforcement applies to any stimulus or
Reviews 31, 641. consequence that increases the likelihood that the im-
Mink, J. W. (1996). The basal ganglia: Focused selection and inhi-
bition of competing motor programs. Progress in Neurobiology mediately antecedent behavior will occur again in an
50, 381425. operant learning task. As regards classical condition-
Mogenson, G. J., Brudzynski, S. M., Wu, M., Yang, C. R., and Yim, ing, however, the term is often applied to the UCS,
C. C. Y. (1993). From motivation to action: A review of dop- which elicits a reflex (unconditioned response, or
aminergic regulation of limbicnucleus accumbensventral UCR) and, when paired with a neutral CS, strength-
pallidumpedunculopontine nucleus circuitries involved in
limbic motor integration. In P. W. Kalivas and C. D. Barnes, ens or confirms the predictive relationship between
eds., Limbic motor circuits and neuropsychiatry. Boca Raton, FL: the two stimuli. Conceptually, the UCS may be
CRC Press. thought of as a teaching input.
REINFORCEMENT OR REWARD IN LEARNING: Cerebellum 571
Electrophysiology of the Inferior Olive Inferior Olive Lesions

Rescorla and Wagner have proposed a behavioral If the inferior olive is the source of reinforcing
model of classical conditioning (1972; Rescorla, input to the cerebellum, then destruction of this nu-
1988) in which associative strength is greatest early in cleus should produce behavioral phenomena that are
training and declines with each stimulus pairing. If identical to those observed when the exteroceptive
that model is true, then neural activity within the rein- UCS is omitted from training trials. Lesions of the
forcement pathway ought to decrease across training. rostromedial portions of the inferior olive block CR
In 1991 Sears and Steinmetz tested this hypothesis acquisition in rabbit training using paired-tone (CS)
and reported that UCS-evoked activity in the rabbit and air-puff (UCS) presentations in a standard delay-
dorsal accessory olive was large during the initial conditioning procedure (McCormick, Steinmetz, and
training trials but decreased concomitantly with the Thompson, 1985). Further, in experienced animals,
development of conditioned responses (CRs) later lesions of this nucleus gradually abolished learned re-
in training. On exclusively UCS trials or on a paired sponses in a manner that was similar to that observed
trial in which the rabbit failed to give a CR, evoked in intact control animals undergoing extinction (i.e.,
activity within the olive was again apparent, sug- exclusively CS trials).
gesting that olivary activity may be an error-detection
signal. Sears and Steinmetz proposed that such a
Electrical Stimulation of the Inferior Olive
partial reinforcement mechanism maintains con-
tinued CR expression. Single-unit recordings of Substitution of an exteroceptive UCS with electri-
Purkinje cells conducted by Foy and Thompson in cal stimulation of the inferior olive or climbing fiber
1986 indicated that the decrement in olivary ac- afferents to the cerebellum provides, perhaps, the
tivity with continued training was reflected in cere- most stringent test of whether this structure is the
bellar cortical activity. At the beginning of train- locus of reinforcement during classical conditioning.
ing, 61 percent of 118 studied neurons displayed Electrical stimulation must be able to recreate all of
complex spikes in response to UCS onset. At the the behavioral phenomena normally associated with
end of training, only 27 percent of the Purkinje peripheral UCS presentations during conditioning.
cells displayed complex spikes evoked by UCS Stimulation of the inferior olive produces a vari-
onset. ety of movements, depending on the location of the
stimulating electrode. These movements may include
The foregoing data indicate that once the associa- eye blinks; movement of the head, neck, facial mus-
tion between the CS and UCS has been formed, no cles; or limbs. Pairings of a tone CS and electrical
further activity in the UCS pathway is necessary ex- stimulation of the dorsal accessory nucleus of the infe-
cept in the case of a performance error. A significant rior olive as the UCS produces normal rates of eye-
body of behavioral data supports this view. For exam- blink conditioning (Mauk, Steinmetz, and Thomp-
ple, Kamin (1968) has demonstrated that once a son, 1986). Furthermore, the range of interstimulus
CS-UCS association has been formed, insertion of intervals (ISI) with UCS electrical stimulation was
a second CS immediately after the first yields no identical to that of peripheral UCSs. Conditioning
behavioral learning to the new CS. Kamin calls was maximal when the CS-UCS interval was held at
this process blocking. No new information is added 150 or 250 msec. Shorter intervals of 50 msec pre-
by the second CS, so the animal essentially ignores vented acquisition.
it. If continued activity were to occur in the UCS
pathway, then the animal should attach the same
associative properties to the second CS. Kim, Krupa, Electrical Stimulation of Cerebellar White
and Thompson (1998) demonstrated exactly this Matter
phenomenon by injecting the GABA antagonist One criticism leveled against the preceding ex-
picrotoxin into the olivary complex after acquisition periments is that eye blinks elicited by stimulation of
of the first CS-UCS (tone-air puff) association. The the inferior olive as the UCS might reflect antidromic
antagonist prevented the normal diminution of oli- activation of spinal trigeminal neurons and spread of
vary activity that occurs with training. Subsequent activation to mossy-fiber collaterals that project to
training with both the original and a second CS rabbit lobule HVI (site of facial map) (Moore and
(light) was accompanied by evoked complex spike Blazis, 1989). To explore this issue, Swain, Shinkman,
activity in the cerebellar cortex to both CSs. Exclu- Nordholm, and Thompson performed a parametric
sively CS test trials indicated that the animal acquired investigation of cerebellar white-matter stimulation in
comparable associations between both CSs and the 1992. They chose as the site of stimulation the white
UCS. matter immediately underlying cerebellar lobule HVI
572 REINFORCEMENT OR REWARD IN LEARNING: Cerebellum
because it is remote from brain stem and spinal-reflex experiment, an impromptu experiment was conduct-
pathways. Stimulation at this site would activate ed. While the animal was in its home cage, if a whistle
climbing fibers and other cerebellar afferents. In at about 1kHz (CS-tone frequency) was presented, the
their study, white-matter stimulation elicited eye rabbit responded with a conditioned lip movement.
blinks as well as movements of the face or neck and, If a whistle of a different pitch was presented, the rab-
when paired with a tone, CS produced learning com- bit did not respond. These observations suggest that
parable to that seen with peripheral UCSs or with oli- conditioning was specific to the tone CS and not to
vary stimulation. No significant differences were context and that the CR exhibited a stimulus general-
noted in learning rate for the various movements ization gradient to the frequency of the whistle.
evoked by stimulation. When switched to CS alone
presentations, the animals extinguished rapidly and
upon reinstatement of paired training showed reac- Conclusion
quisition with substantial savings. Similar findings Destruction of the olive or its efferent climbing fi-
have been reported by Gormezano and colleagues bers blocks learning in nave animals and extinguish-
(1983) in paradigms using a peripheral UCS. The es it in the experienced ones. Physiological records
correspondence between learning in their experi- indicate that olivary neurons respond strongly to UCS
ment and conditioning with a peripheral UCS was representation at the beginning of training but subside
markable even upon examination of the small details. as learning occurs. Stimulation or suppression of ac-
For example, researchers found that the percentage tivity within the olivary climbing-fiber system can rec-
of CRs on the last day of reacquisition training was reate a host of behavioral phenomena, including con-
smaller than that of the last day of acquisition train- ditioned inhibition, the UCS preexposure effect, UR
ing. While this difference was not significant, it is con- augmentation, and blocking. Together these data in-
sistent with reports by several investigators that the dicate that the inferior olive and its efferent axons,
CS may acquire inhibitory properties during extinc- the climbing fibers, are the neural pathway that con-
tion training that become evident upon retraining. veys information about reinforcement to the cerebel-
lum.
The experiment included control rabbits that re-
ceived either randomly or explicitly unpaired presen-
See also: CONDITIONING, CELLULAR AND NETWORK
tations of the tone CS and white-matter stimulation
SCHEMES FOR HIGHER-ORDER FEATURES OF;
as the UCS. Animals that received explicitly unpaired GUIDE TO THE ANATOMY OF THE BRAIN:
presentations of the conditioning stimuli were pro- CEREBELLUM; KAMINS BLOCKING EFFECT:
foundly impaired when they were subsequently NEURONAL SUBSTRATES; REINFORCEMENT
switched to a paired CS-UCS training procedure. Pre-
vious behavioral work by Rescorla (1969) has demon- Bibliography
strated that the explicitly unpaired control procedure Foy, M. R., and Thompson, R. F. (1986). Single unit analysis of
may result in the CS acquiring inhibitory properties Purkinje cell discharge in classically conditioned and un-
such that subsequent acquisition training is retarded. trained rabbits. Society for Neuroscience Abstracts 12, 518.
Gormezano, I., Kehoe, E. J., and Marshall, B. S. (1983). Twenty
A subsequent study by the group (Swain et al., years of classical conditioning research with the rabbit. In J.
1999) found that exposure to as few as 108 exclusively M. Sprague and A. N. Epstein, eds., Progress in psychobiology
UCS trials was sufficient to produce a potent UCS and physiological psychology, Vol. 10. New York: Academic Press.
Kamin, L. J. (1968). Attention-like processes in classical condition-
preexposure effect. Animals exposed to the UCS
ing. In M. R. Jones, ed., Miami symposium on the prediction of be-
prior to training typically required more than 600 tri- havior: Aversive stimulation. Miami: University of Miami Press.
als to learn. Animals that received no preexposure Kim, J. J., Krupa, D. J., and Thompson, R. F. (1998). Inhibitory
learned at a normal rate (100200 trials). Exposure cerebelloolivary projections and blocking effect in classical
to UCSs of fixed duration also promoted an increase conditioning. Science 279, 570573.
Mauk, M. D., Steinmetz, J. E., and Thompson, R. F. (1986). Classi-
in the amplitude and a decrease in the latency of the
cal conditioning using stimulation of the inferior olive as the
stimulus-evoked reflex as trial presentations prog- unbconditioned stimulus. Proceedings of the National Academy of
ressed. There have been similar findings for UCS pre- Sciences of the United States of America 83, 5,3495,353.
exposure and reflex augmentation with exteroceptive McCormick, D. A., Steinmetz, J. E., and Thompson, R. F. (1985).
UCSs (Mis and Moore, 1973). Lesions of the inferior olivary complex cause extinction of the
classically conditioned eyelid response. Brain Research 359,
The authors (Swain et al., 1992) also reported an- 120130.
ecdotal observations that further support the hypoth- Mis, R. W., and Moore, J. W. (1973). Effects of preacquisition US
exposure on classical conditioning of the rabbits nictitating
esis that cerebellar white-matter stimulation as a UCS
membrane response. Learning and Motivation 4, 108114.
results in normal behavioral learning. After an animal Moore, J. W., and Blazis, D. E. J. (1989). Stimulation of a classically
that exhibited an ipsilateral lip movement to cerebel- conditioned response: A cerebellar network implementation
lar stimulation had been trained and completed the of the Sutton-Barto-Desmond model. In J. H. Byrne and W.
REINFORCEMENT OR REWARD IN LEARNING: Electrical Self-Stimulation, Brain 573
O. Berry, eds., Neural models of plasticity. New York: Academic Researchers have linked the rewarding effect of elec-
Press. trical stimulation to the mechanisms governing
Rescorla, R. A. (1969). Pavlovian conditioned inhibition. Psychologi- choice between competing goals and activities.
cal Bulletin 72, 7794.
(1988). Behavioral studies of Pavlovian conditioning. Annu-
al Review of Neuroscience 11, 329352.
Rescorla, R. A., and Wagner, A. R. (1972). A theory of Pavlovian The Relationship Between BSR and
conditioning: Variations in the effectiveness of reinforcement Natural Rewards
eds., Classical conditioning II: Current research theory. New York: Researchers explored the relationship between
Appleton-Century-Crofts. BSR and gustatory rewards in an extensive series of
Sears, L. L. and Steinmetz, J. E. (1991). Dorsal accessory inferior experiments whereby, under various physiological
olive activity diminishes during acquisition of the rabbit classi- states, rats had to choose between lateral hypothalam-
cally conditioned eyelid response. Brain Research 545, 114
122.
ic (LH) stimulation varying in strength and either
Swain, R. A., Shinkman, P. G., Nordholm, A. F., and Thompson, gustatory reward alone or a compound reward con-
R. F. (1992). Cerebellar stimulation as an unconditioned stim- sisting of a gustatory reward plus fixed BSR. The re-
ulus in classical conditioning. Behavioral Neuroscience 106, searchers estimated preference by varying the
739750. strength of the stimulation triggered by one spout
Swain, R. A., Shinkman, P. G., Thompson, J. K., Grethe, J. S., and
Thompson, R. F. (1999). Essential neuronal pathways for re-
while holding constant the reward triggered by a sec-
flex and conditioned response initiation in an intracerebellar ond spout. They measured the stimulation strength
stimulation paradigm and the impact of unconditioned stim- required to produce isopreference. The experimental
ulus preexposure on learning rate. Neurobiology of Learning protocol sought to minimize the difference between
and Memory 71, 167193. the gustatory reward and the BSR. Thus, the gustato-
Thompson, R. F. (1986). The neurobiology of learning and memo-
ry. Science 233, 941947.
ry reward was intraorally infused in small volumes via
a catheter, while an intragastric cannula minimized
Rodney A. Swain the accumulation of gustatory reward in the gut to
simulate the instantaneous and insatiating nature of
BSR.
ELECTRICAL SELF-STIMULATION, BRAIN Initially, when given a choice between a gustatory
For physiological psychology, the discovery that elec- stimulus (an intraoral infusion of sucrose) and LH
trical stimulation of certain brain regions is so power- stimulation, the rats choose the sucrose infusion if the
fully rewarding that laboratory rats eagerly self- brain stimulation was weak; but they shifted their
administer it was an earthquake, the shock waves of preference as the LH stimulation became stronger.
which rippled through the popular press in hyperbol- With sucrose availability, rats eschew suprathreshold
ic recountings. It may prove the key to human be- levels of brain stimulation in favor of the sucrose re-
havior, trumpeted a Montreal newspaper. Some reward. Clearly, the availability of sucrose alters the
ports even went so far as to fuel fears that brain preference for BSR when the two rewards are evaluat-
stimulation reward (BSR) could be used as an agent ed in a common system of measurement wherein the
for social control. However amusing in retrospect, larger reward is selected. In a subsequent experi-
much of this hype was an understandable reflection ment, researchers pitted LH stimulation against a
of the amazement, shared by scientists and laypeople compound reward consisting of sucrose and an equi-
alike, of the powerful and immediate impact of elec- preferred train of brain stimulation to determine
trical stimulation on behavior. The stimulation elec- whether the results of the common evaluation can
trode, injecting a meaningful signal into the neural summate. Indeed, the rats assigned a higher value to
circuitry mediating goal-directed behavior, can turn the compound reward than to its sucrose component
the average lab rat into a craven voluptuary, willing alone. The strength of LH stimulation required to
to press a lever to the point of starvation or exhaus- balance the compound reward exceeded the stimula-
tion. tion strength required to balance the sucrose reward
alone. The finding that BSR and sucrose reward can
Even though the initial hype has subsided (Tal-
be combined lends further support to the idea of eval-
war et al., 2002), BSR has become a model system for
uation in a common system of measurement.
the study of positive reinforcement. Because organ-
isms acquire new responses to obtain electrical stimu- The competition and summation experiments
lation, the effect qualifies as reinforcement according confirm the hypothesis that BSR and natural rewards
to the law of effect. Early evidence suggested that have something important in common: they are eval-
brain stimulation reward differs in some respects uated on a common currency scale. As McFarland and
from conventional rewards such as food and water, Sibley pointed out (1975), orderly choice between
whereas later research emphasized their similarities. mutually exclusive behaviours must be based on a
574 REINFORCEMENT OR REWARD IN LEARNING: Electrical Self-Stimulation, Brain
common currency; the influences contributing to the stimulation has been the focus of intensive research.
choice of each activity seem to converge on a be- For decades researchers sought the quantitative char-
havioural final common path (p. 265). There is con- acteristics of the first-stage neurons through psycho-
siderable evidence that thresholds for BSR remain re- physical techniques. Behavioral measurements of re-
markably stable across a variety of states that are covery from refractoriness, collision block, and
associated with changes in choice of goal object. anodal hyperpolarization block have led to the idea
Prominent among these states are changes in energy that the directly stimulated neurons subserving the
balance or nutrient requirements. For example, nei- powerfully rewarding effect produced by electrical
ther the increase in salt appetite that accompanies so- stimulation of the MFB originate in the basal fore-
dium depletion nor the suppression of appetite that brain and give rise to fine, myelinated axons that de-
accompanies postingestive feedback is associated with scend through the MFB toward the tegmentum. Fur-
substantial changes in BSR. These findings suggest ther research has supported this idea. Excitotoxic
that BSR mimics a global currency of goal evaluation. lesions of a region in the lateral basal forebrain that
In contrast to these findings, other research has includes the lateral preoptic area, the substantia in-
shown that BSR is sensitive to changes in energy bal- nominata, and parts of the bed nucleus of the stria
ance when the electrode is in a specific region of the terminalis increase the threshold for self-stimulation
LH (perifornical LH). Moreover, leptin, a hormone of more caudal sites along the MFB. Fos im-
that communicates the state of the fat stores to the munostaining has shown that rewarding stimulation
brain, exerts opposite influences on the rewarding ef- of the MFB activates neurons in the region of the ef-
fect of stimulating food restriction-sensitive and -in- fective lesions.
sensitive sites. These results suggest that functionally The Role of Dopamine Neurons
distinct pathways, differentially responsive to the
BSR is similar in certain respects to the self-
state of the organism, exist within the neural substrate
administration of psychomotor stimulants. Both are
for BSR. One way of interpreting the contrast be-
self-administered in a compulsive fashion, both may
tween the results obtained at different stimulation
activate the same brain circuitry, and both are dop-
sites is to propose that stimulation at the restriction-
amine-dependent. Microdialysis and voltametry
insensitive sites mimics a global currency, whereas
studies indicate that both psychomotor stimulant re-
stimulation at restriction-sensitive sites mimics a local
ward and BSR elevate dopamine levels in the nucleus
one, a currency related to long-term energy stores.
accumbens. But the most compelling evidence sug-
gesting dopaminergic involvement in BSR comes
The Neural Circuitry Subserving Reward: from pharmacological studies. Dopamine receptor
blockers increase self-stimulation thresholds, whereas
Lessons from Studies on BSR
dopamine agonists have the opposite effect.
The Directly Stimulated Stage Researchers believed that direct activation of
Electrical stimulation provides the most effective MFB dopamine fibers initiated rewarding signals in
means of introducing a reward signal into the brain. this region until psychophysical and electrophy-
The strength of the electrically induced rewarding ef- siological findings ruled out this possibility. Dop-
fect and the ease with which it can be controlled ren- amine neurons are not myelinated, and their thresh-
der electrical stimulation a powerful tool for studying olds of activation are above the level of the
the reward substrate. Consider the task of linking stimulation parameters commonly used in BSR
BSR to the activity of identified neurons. The first studies. The question then arises: What place does the
step toward accomplishing this task is to identify the dopamine system occupy in the neural substrate of
neurons in the immediate vicinity of the electrode tip the rewarding effect?
whose direct activation gives rise to the rewarding ef- According to the simplest hypothesis, the dop-
fect of the stimulation. Finding these cells and tracing amine neurons are in series with the first-stage ones
their inputs and outputs is likely to shed light on the and thus carry the reward signal. An alternative hy-
mechanisms in the brain that underlie the behavioral pothesis is that the dopamine neurons do not actually
effects of positive reinforcement. relay the reward signal but rather play a permissive
The initial strategy employed to investigate the role at some stage of the BSR substrate. One example
BSR system was to map the brain sites that support of a permissive role would be to gate signal processing
self-stimulation. Stimulation of numerous regions in the pathway responsible for the rewarding effect.
evokes BSR. An especially effective site is the medial Perhaps increased activity in the dopamine neurons
forebrain bundle (MFB), which allows easy shaping of enhances transmission of the reward signal, whereas
self-stimulation behavior. Accordingly, finding the di- decreased activity reduces or even blocks its transmis-
rectly stimulated stage (first stage) of MFB self- sion.
REINFORCEMENT OR REWARD IN LEARNING: Striatum 575
The permissive-role hypothesis finds substantial Carr, K. D. (1996). Feeding, drug abuse, and the sensitization of
corroboration in a microdialysis study designed to reward by metabolic need. Neurochemistry Research 21, 1,455
1,467.
measure the levels of dopamine released in the nucle-
Fulton, S., Woodside, B., and Shizgal, P. (2000). Modulation of
us accumbens during self-stimulation for equi- brain reward circuitry by leptin. Science 287, 125128.
rewarding pulses that consisted of different combina- Gallistel, C. R., Shizgal, P., and Yeomans, J. S. (1981). A portrait
tions of stimulus parameters of the electrical stimula- of the substrate for self-stimulation. Psychological Review 88,
tion. Some had argued that if the reward signal 228273.
travels along dopamine neurons, the release of dop- Garris, P. A., Kilpatrick, M., Bunin, M. A., Michael, D., Walker, Q.
amine should not depend on the stimulus parameters D., and Wightman, R. M. (1999). Dissociation of dopamine
release in the nucleus accumbens from intracranial self-
because equi-rewarding stimuli should produce a
stimulation. Nature 398, 6769.
constant output in all neural stages carrying the re- McFarland, D. J., and Sibley, R. M. (1975). The behavioural final
ward signal, irrespective of the spatio-temporal na- common path. Philosophical Transactions of the Royal Society of
ture of the signal. However, the results showed that London B 270, 265293.
the magnitude of dopamine release differed between Miliaressis, E., Emond, C., and Merali, Z. (1991). Re-evaluation of
sets of self-stimulation parameters that, nevertheless, the role of dopamine in intracranial self-stimulation using in
produced the same rewarding effect. The finding that vivo microdialysis. Behavioural Brain Research 46, 4348.
Olds, J., and Milner, P. M. (1954). Positive reinforcement pro-
the rewarding effect of the stimulation is not mir-
duced by electrical stimulation of septal area and other re-
rored in the magnitude of dopamine release as mea- gions of rat brain. Journal of Comparative and Physiological Psy-
sured by microdialysis contravenes the hypothesis chology 47, 419427.
that the reward signal elicited by MFB stimulation is Reynolds, J. N. J., Hyland, B. I., and Wickens J. R. (2001). A
relayed by the mesolimbic dopamine neurons. cellular mechanism of reward-related learning. Nature 413,
6770.
Another study measuring dopamine release from Shizgal, P. (1997). Neural basis of utility estimation. Current Opin-
dopamine terminals in the nucleus accumbens by ion in Neurobiology 7, 198208.
voltametry provided additional evidence that dop- (1999). On the neural computation of utility: Implications
amine is not a neural substrate for reward per se. The from studies of brain stimulation reward. In D. Kahneman, E.
study showed that although dopamine release corre- Diener, and N. Shwarz, eds., Well-being: The foundations of he-
lated with the ability of animals to learn the self- donic psychology. New York: Russell Sage Foundation.
Talwar, S. K., Xu, S., Hawley, E. S., Weiss, S. A., Moxon, K. A., and
stimulation behavior, extracellular dopamine was Chapin, J. K. (2002). Rat navigation guided by remote con-
nonetheless absent during self-stimulation itself. trol. Nature 417, 3738.
These results are consistent with two hypotheses: that Wise, R. A. (1996). Addictive drugs and brain stimulation reward.
dopamine is related to the novelty and predictability Annual Review of Neuroscience 19, 319340.
of rewards and that dopamine neurons provide teach-
ing signals for appetitive learning and reinforcement. Andreas Arvanitogiannis
Dopamine could thus mediate plasticity changes at
sites in the brain that subserve the learning of the self-
stimulation behavior. Indeed, there appears to be a
correlation between dopamine-dependent potentia- STRIATUM
tion in corticostriatal circuits and the rate of acquisi-
The role of rewards in the survival and well-being of
tion of lever-pressing behavior.
biological agents ranges from the control of vegeta-
Future studies might clarify the relationship be- tive functions to the organization of voluntary, goal-
tween the first stage of the BSR substrate and dop- directed behavior. The brain extracts the reward in-
amine neurons while shedding light on circuitry information from a large variety of stimuli and events
volved in both the processing of reward information for appropriate use in the control of behavior. Recent
and the reinforcement learning. studies revealed that neurons in a limited number of
See also: GUIDE TO THE ANATOMY OF THE BRAIN brain structures carry specific signals about past and
future rewards. The neurophysiological study of re-
Bibliography ward processing within the framework of goal-
Arvanitogiannis, A., Flores, C., Pfaus, J. G., and Shizgal, P. (1996). directed behavior may contribute to a basic under-
Increased ipsilateral expression of Fos following lateral hypo-
thalamic self-stimulation. Brain Research 720, 148154.
standing of mechanisms of drug abuse and could thus
Arvanitogiannis, A., Tzschentke, T. M., Riscaldino, L., Wise, R. A., have a strong medical and social impact. This article
and Shizgal, P. (2000). Fos expression following self- concerns the reward signals in the striatum and de-
stimulation of the medial prefrontal cortex. Behavioural Brain scribes how its neurons detect rewards, learn to pre-
Research 107, 123132.
Arvanitogiannis, A., Waraczynski, M., and Shizgal, P. (1996). Ef-
dict future rewards from past experience, and use re-
fects of excitotoxic lesions of the basal forebrain on MFB self- ward information for learning, choosing, preparing
stimulation. Physiology and Behavior 59, 795806. and executing goal-directed behavior.
576 REINFORCEMENT OR REWARD IN LEARNING: Striatum
Behavioral Functions of Rewards parate and initate movements and process informa-
tion about rewarding outcomes (Schultz, 2000).
Biological agents need to acquire nutrional sub-
stances from the environment and interact with sexu-
al partners for reproduction. The brain controls the Neurophysiology of Reward Mechanisms in
contact with foods, fluids, and sexual partners and the Striatum
mediates the adaptation of behavior to novel or
changed situations. Neuronal mechanisms not only Neurons in the striatum detect the delivery of re-
detect rewards but also predict them on the basis of wards and discriminate among them. Other striatal
representations formed by past experience. Through neurons detect conditioned, reward-predicting visual
stimuli and discriminate reward-predicting from
these mechanisms rewards serve as goals for volun-
nonpredicting stimuli. Many natural situations in-
tary and intentional forms of behavior.
volve sequences of individual movements that lead to
Rewards have several basic functions. A popular a final reward. Neurons in the ventral striatum re-
view holds that rewards induce subjective feelings of spond differentially to sequential cues at different
pleasure and contribute to positive emotions. Re- steps away from the reward. It appears that these neu-
wards act as positive reinforcers by increasing the fre- rons report the positions of individual stimuli within
quency and intensity of behavior leading to the acqui- a behavioral sequence and thus signal the progress to-
sition of goal objects in classical and instrumental wards the reward.
conditioning procedures. The learning function is A well learned reward-predicting stimulus evokes
based on the discrepancy between the prediction and a state of expectation. Striatal neurons are active dur-
occurrence of rewards (reward-prediction error). ing several seconds of expectation of predictable re-
Rewards are goals that elicit approach and con- wards. Their activity follows a reward-predicting stim-
summatory behavior. Reward objects have positive ulus and persists for several seconds until the reward
motivational value because they elicit effortful behav- is delivered. Such neurons discriminate between dif-
ioral reactions. The values arise either through innate ferent future rewards and other, nonrewarding pre-
mechanisms or, in most cases, learning. In this way re- dictable task events, such as movement-eliciting stim-
wards help to establish value systems for behavior and uli or instructions cues. The activity may reflect a
serve as key references for behavioral decisions. Dif- neuronal representation of reward established
ferences in perceived reward values in individual through previous experience. Reward-detecting and
agents and situations may help to explain variations reward-expecting neurons are found about twice as
in behavioral choices. often in the ventral striatum as in the caudate and pu-
tamen. These findings may provide neurophysiologi-
cal correlates for the known motivational functions of
A Case for the Striatum in Reward the ventral striatum.
Mechanisms Expectations change with experience. Animals
The basal ganglia are composed of the striatum expect reward initially on all trials when learning to
(caudate nucleus, putamen, and ventral striatum, in- discriminate rewarded from nonrewarded stimuli.
cluding nucleus accumbens), dopamine neurons of Similarly, neurons in the striatum show reward expec-
the substantia nigra and ventral tegmental area tation activity during initial trials with novel stimuli,
(groups A8, A9, and A10), and a number of other and this activity is progressively restricted to reward-
structures. Current evidence suggests that the basal ed rather than unrewarded trials. These adaptations
ganglia are important for control of voluntary, goal- are reflected in the activity of neurons in the striatum.
directed behavior and the processing of rewarding Expected rewards may serve as goals for volun-
outcomes. Human diseases involving these structures tary behavior if information about the reward is pres-
lead to deficits in voluntary behavior and movements ent while behavioral reactions toward the reward are
(Parkinsonism, schizophrenia, and Huntingtons cho- being prepared and executed (Dickinson and Ball-
rea). Lesioning and psychopharmacological and elec- eine, 1994). Neuronal mechanisms in the striatum
trical self-stimulation experiments strongly indicate may integrate reward information into processes me-
that dopamine neurons and the ventral striatum serve diating the behavior leading to the reward. Some
prime motivational functions (Robbins and Everitt neurons in the anterior striatum show sustained activ-
1996). Major addictive substances, such as cocaine ity for a few seconds during the preparation of a
and heroin, increase the dopamine concentration in movement. These activations occur much more com-
the ventral striatum, and animals try to receive injec- monly in rewarded than unrewarded trials and vary
tions of dopamine or opiates into the ventral striatum with the type of reward expected for the movement.
(Wise, 1996). Single neurons in the basal ganglia pre- Thus both the future reward and the movement to-
REINFORCEMENT OR REWARD IN LEARNING: Striatum 577
ward the reward are represented by these neurons. Less is known about reward processing in other
The reward-dependent activity may be a way in which brain structures. Neurons in the amygdala, which
the expected reward is represented by neurons and projects to the ventral striatum, react to the occur-
can influence neuronal processes underlying the be- rence of rewards and discriminate between different
havior toward that reward. liquid and food rewards. Neurons in the dorsolateral
prefrontal cortex show reward-discriminating activity
during the preparation of movements and may thus
Reward Information Arriving at the be involved in the organisation of behavior directed
Striatum towards rewarding goals (Watanabe, 1996).
The striatum is a part of a limited network of
brain structures involved in the processing of reward
Conclusion
information. Although it is difficult at the moment to
assess the different functions of each of these struc- The striatum processes reward information in
tures, we can describe the neuronal activities in some different ways. Neurons in the striatum detect and
of the structures sending information to the striatum. discriminate between among rewards and may play a
role in assessing the nature and identity of individual
The orbitofrontal cortex, the ventral part of the rewards. However, the striatum not only detects and
frontal lobe, is a part of the limbic system involved in analyses past events, but it also constructs and dynam-
motivation and emotions. Some of its neurons project ically modifies predictions based on past experience.
to the striatum, in particular its ventral part, includ- Striatal neurons respond to learned stimuli that pre-
ing nucleus accumbens. Orbitofrontal neurons dis- dict rewards and show sustained activity during peri-
criminate between different rewards on the basis of ods pr reward expectations. They even take guesses
the subjects preferences (Tremblay and Schultz, about future rewards and adapt their activity to expe-
1999). For example, a neuron is more active when a rience.
preferred rather than a nonpreferred reward is ex-
pected. But when the initially nonpreferred reward Once we understand more about how the brain
occurs in trials alternating with an even less preferred treats rewards, we can investigate how reward infor-
reward, the nonpreferred award will become the pre- mation produces motivated behavior. Neurons in
ferred one and the neuron will be activated predomi- structures that control behavior seem to incorporate
nantly by this reward. These neurons do not code re- information about upcoming rewards when coding
wards on the basis of their physical properties. reward-seeking behavior. Future research may per-
Rather, they code the relative preference of the re- mit us to understand how such activity may lead to
ward. Neurons coding the relative preference for re- choices and decisions incorporating both the cost and
wards might provide important information to neu- the benefit of behavior. We should also investigate
ronal mechanism in the frontal lobe underlying goal- neuronal mechanisms behind higher, more cognitive
directed behavioral choices. rewards typical for human behavior. Such research
would help us to understand further how brains con-
Dopamine neurons in the substantia nigra and trol important characteristics of voluntary and moti-
the ventral tegmental area project to the dorsal and vated behavior of complex organisms, possibly to the
ventral striatum, respectively. Most dopamine neu- extent that individual differences in brain function
rons show short, phasic activations following the pre- explain variations in basic traits of personality.
sentation of liquid and food rewards and conditioned
visual and auditory reward-predicting stimuli. How-
Bibliography
ever, dopamine neurons do not respond to rewards
Dickinson, A., and Balleine, B. (1994). Motivational control of
unconditionally but code them relative to what is pre- goal-directed action. Animal Learning and Behavior 22, 118.
dicted, suggesting that they code a reward-prediction Montague, P. R., Dayan, P., and Sejnowski, T. J. (1996). A frame-
error (Waelti et al., 2001). The dopamine response work for mesencephalic dopamine systems based on predic-
may be a global reinforcement signal sent in parallel tive Hebbian learning. Journal of Neuroscience 16, 1,9361,947.
Robbins, T. W., and Everitt, B. J. (1996). Neurobehavioral mecha-
to all neurons in the striatum. Learning theory sug-
nisms of reward and motivation. Current Opinion in Neuro-
gests that prediction errors play a crucial role in biology 6, 228236.
learning and that the phasic dopamine response may Robinson, T. E., and Berridge, K. C. (1993). The neural basis for
be an ideal teaching signal for approach learning. drug craving: An incentive-sensitization theory of addiction.
This signal strongly resembles the teaching signal Brain Research Reviews 18, 247291.
Schultz, W. (2000). Multiple reward systems in the brain. Nature Re-
used by effective temporal-difference reinforcement
views Neuroscience 1, 199207.
models (Montague et al., 1996). Indeed, artificial Waelti, P., Dickinson, A., and Schultz, W. (2001). Dopamine re-
neuronal networks that use this type of teaching sig- sponses comply with basic assumptions of formal learning
nal learn to play world-class backgammon. theory. Nature 412, 4348.
578 REPETITION AND LEARNING
Watanabe, M. (1996). Reward expectancy in primate prefrontal the first exposure to a stimulus or situation, and the
neurons. Nature 382, 629632. amount learned in each subsequent exposure contin-
Wise, R. A. (1996). Neurobiology of addiction. Current Opinion in
Neurobiology 6, 243251.
ually declines until further improvement is too small
to be detected. The rate of learning is negatively re-
Wolfram Schultz lated to the amount already learned. Hintzman and
Curran (1995) have shown that people can register
the occurrence of a repeated stimulus while failing to
learn more about its specific details. First impressions
REPETITION AND LEARNING of a repeated stimulus are particularly important, as
people may show little evidence for having noticed
Sayings such as Practice makes perfect illustrate the
subtle changes that are introduced to a stimulus after
well-known fact that repetition improves learning.
its first presentation (DiGirolamo and Hintzman,
This was discussed by numerous ancient and medi-
1997).
eval thinkers and was demonstrated empirically by
Hermann Ebbinghaus, the first researcher to carry
out a prolonged series of experiments on human Why Does Repetition Improve Learning?
memory. In a classic 1885 book, Ebbinghaus showed
that retention of information improves as a function Anderson and Schooler (1991) have pointed out
of the number of times the information has been stud- that the sensitivity of learning to repetition is evi-
ied. Since the time of Ebbinghaus, countless investi- dence for its efficiency and adaptiveness because the
gators have used repetition to study learning and frequency with which information has been used in
memory. the past is a very good predictor of whether it will be
needed in the future. Still, although repetition has
Although experimenters typically find a consis-
been intensively studied, the mechanisms underlying
tent relationship between repetition and learning,
its effects are still poorly understood. Moreover, there
numerous authors (Guthrie, 1935) have pointed out
is no reason to believe that a single explanation could
that this does not necessarily mean that the learning
apply to all situations where repetition facilitates
process itself has to be either gradual or continuous.
learning.
Most learning situations contain a number of smaller
facets or subproblems that must be mastered before Of particular interest to many researchers has
learning is complete. It is possible that each of these been the effect of repeated study on human memory,
subproblems is mastered suddenly, perhaps through and the two dominant explanations of these repeti-
insight. However, the subproblems are learned at dif- tion effects were both discussed by Ward (1893). One
ferent times, with more and more of them mastered class of explanations (called a functional approach by
as the number of trials increases. This analysis pro- Ward but more commonly known as strength theory
poses that a gradual improvement in learning as a rein twenty-first-century scientific circles) claims that
sult of repetition may reflect the accumulation of sub- there is a single location in memory storage that cor-
problems that have been mastered in a sudden responds to an event. Every time the event is repeat-
fashion. Distinguishing between a truly continuous ed, that location (known as the memory trace) in-
learning process and the accumulation of small, sud- creases in effectiveness or strength. It is also assumed
den insights is difficult. A common assumption is that that stronger traces are easier to retrieve from memo-
learning may be either gradual or sudden, depending ry than are weaker traces. Repetition thus improves
on the background of the learner and the nature of learning by increasing the strength of a single memo-
the information to be learned. For example, Harry ry trace. A second class of explanation for the effects
Harlow (1949) showed that learning to novel situa- of repetition on memory was called an atomistic ap-
tions may occur slowly and continuously but may approach by Ward but is now known as multiple-trace
pear in sudden flashes of insight when the organism theory. This approach assumes that every occurrence
has had experience in a number of similar situations. of an event is a unique episode. Every time an event
Thus, although the amount of learning may appear occurs, a separate, independent memory trace is
to grow gradually and continuously as a result of repe- formed. This trace contains information about the
tition, determination of whether subcomponents of time and situation in which that occurrence hap-
the task are learned gradually or suddenly is more pened. The more times an event occurs, the more
difficult and requires careful analysis. traces of that event are placed in memory. According
Although the total amount learned increases as a to this multiple-trace theory, repetition improves
function of repetition, the amount learned on each learning because finding at least one trace of an event
trial will not be constant. Repetition effects exhibit becomes easier when there are more traces of that
negative acceleration: The most learning occurs in event in memory.
REPETITION AND LEARNING 579
A fundamental difference between these two ac- frequency judgments are seen as being based on a
counts concerns the representation of the individual count of individual traces, each carrying information
occurrences of a repeated item. The strength theory about its time of formation.
claims that each occurrence of an event strengthens Subsequent studies have found further evidence
a single memory trace. Since each occurrence has the in favor of a multiple-trace theory. For example,
same effect, the specific details of individual occur- when some words are presented visually and others
rences are lost. In contrast, the multiple-trace theory auditorily, participants are able to give separate fre-
claims that every occurrence produces its own trace. quency judgments for each kind of presentation. Also,
The individuality of specific occurrences is main- they are able to judge how often a word followed an-
tained. other word on a list. Such findings suggest that the in-
Experiments distinguishing between these two dividual identities of the occurrences of a repeated
accounts have often required participants to remem- event are maintained in memory, as assumed by the
ber a list of words. A word on the list may occur once multiple-trace theory (Greene, 1992).
or a varying number of times. After seeing the list, Studies such as these suggest that a multiple-trace
participants are shown the list items again and asked theory is necessary to account for the effects of repeti-
to make a judgment regarding how often each item tion on memory. They do not show that such a theory
occurred on the list. Even when they do not expect to is sufficient to account for all the effects of repetition.
be tested on the frequencies of the items, people are The question of whether repetition has other effects
typically able to perform this task with considerable in addition to the creation of multiple memory traces
(but not perfect) accuracy. However, strength theory has not been resolved, although there is some evi-
and multiple-trace theory make different proposals as dence that repeated events are remembered better
to how participants are able to make judgments about than would be expected on the basis of memory for
the frequency of occurrence of list items. Strength specific presentations (Watkins and LeCompte,
theory claims that participants retrieve the memory 1991). Moreover, there is little evidence that would
trace corresponding to a test item and evaluate that allow one to determine whether the multiple-trace
traces strength. They then use the strength to make approach can be applied to all of the situations in
a judgment of frequency. For example, if a memory which learning is improved by repetition.
trace is very strong, participants will guess that the
item occurred many times on the list. If a memory
trace is weak, they may decide that the item occurred When Is Repetition Ineffective in
once (or possibly not at all) on the list. In contrast, the Increasing Learning?
multiple-trace theory claims that participants make Although the emphasis in this entry has necessar-
judgments of frequency by retrieving as many traces ily been on the mechanisms through which repetition
as possible of that item occurring in the context of the improves learning, one should not assume that repe-
list. They then base their judgments on a count of the tition alone is always sufficient. For example, consider
traces they found. a common coin, such as the American penny. Al-
Numerous experiments have investigated wheth- though people have seen such coins countless times,
er a frequency judgment is based on a single trace or as Nickerson and Adams (1979) showed, people can
on the retrieval of many different traces. For exam- have quite poor memory for the details of a penny.
ple, Hintzman and Block (1971) showed participants They are often unable to remember exactly where
two lists of words, five minutes apart. Some words oc- such features as the date and the words In God We
curred on both lists. Each word occurred zero, two, or Trust are located. There is no need for people to at-
five times on List 1 and zero, two, or five times on List tend to these features of a penny because pennies can
2. After seeing both lists, participants were asked to easily be distinguished from other coins on the basis
estimate frequency of occurrence separately for each of their size and color. This suggests that attention to
list. They were quite accurate at this task; their esti- an event may be necessary before repetition of that
mates were chiefly influenced by the frequency of the event leads to noticeable improvements in memory.
item on the list being judged and were influenced lit- The generality of this claim has been established by
tle by the items frequency on the other list. Such a studies demonstrating poor memory for other cur-
finding is difficult for a strength theory to explain: If rencies, for the details of telephone dials, and for the
judgments of frequency were based simply on the messages of common advertisements.
overall strength of the trace of the word, people Additional examples of ineffective repetition
would not be able to make separate estimates for the have come from experiments on rote rehearsal (Glen-
frequency of an item on two lists. However, a multi- berg, Smith, and Green, 1977; Rundus, 1977). In
ple-trace theory would predict this finding because these studies, participants read repeated words aloud
580 RESCORLA-WAGNER MODEL
over and over. An unexpected memory test on the Rundus, D. (1977). Maintenance rehearsal and single-level pro-
words is later given. Memory performance is usually cessing. Journal of Verbal Learning and Verbal Behavior 16, 665
681.
only slightly affected by the number of times that a Ward, J. (1893). Assimilation and association. Mind 2, 347362.
person read each word aloud. On the other hand, if Watkins, M. J., and LeCompte, D. C. (1991). The inadequacy of re-
people are encouraged to carry out more active, ef- call as a basis for frequency knowledge. Journal of Experimental
fortful processing on the words, memory improves Psychology: Learning, Memory, and Cognition 17, 1,1611,176.
dramatically as study time is increased. Robert L. Greene
One situation in which repetition impairs memo-
ry is when people have to recall a short series of digits
or letters in order. Recall is impaired if one of the
items is repeated in the series. This phenomenon, RESCORLA-WAGNER MODEL
known as the Ranschburg effect, was introduced into
See: KAMINS BLOCKING EFFECT: NEURONAL
the modern psychological literature by Crowder and SUBSTRATES; LEARNING THEORY: CURRENT
Melton (1965). Critical to understanding this nega- STATUS; MATHEMATICAL LEARNING THEORY;
tive effect of repetition is the fact that people have to SOMETIMES OPPONENT PROCESS (SOP)
remember that an item was repeated and the loca- MODEL, IN CONDITIONING
tions of each occurrence in the series. The Ransch-
burg effect occurs because recall of the first occur-
rence of the repeated item inhibits accurate recall of
the second occurrence (Greene, 2001). RETRIEVAL PROCESSES IN MEMORY
Thus, repetition need not lead to improved The processes of learning and memory are often sub-
learning. Rather, repetition leads to increased oppor- divided into stages of encoding (initial learning of in-
tunities for learning to occur. Whether learning takes formation), storage (maintaining information over
place will depend on the type of information that has time), and retrieval (using stored information). Pro-
to be remembered and the amount and nature of processes of encoding establish some representation of
cessing that a person carries out. experience in the nervous system, which is referred to
See also: DISTRIBUTED PRACTICE EFFECTS as an engram or memory trace. Memory traces cer-
tainly have physiological underpinnings, but cogni-
Bibliography tive psychologists use the construct as an abstraction
Anderson, J. R., and Schooler, L. J. (1991). Reflections of the envi- to refer to the changed state of the cognitive system
ronment in memory. Psychological Science 2, 396408. before and after some experience. Retrieval processes
Crowder, R. G., and Melton, A. W. (1965). The Ranschburg phe-
nomenon: Failures of immediate recall correlated with repeti- refer to the means of accessing stored information
tion of elements within a stimulus. Psychonomic Science 2, 295 and can be affected by a variety of factors.
296. Retrieval is the key process in the act of remem-
DiGirolamo, G. J., and Hintzman, D. L. (1997). First impressions
are lasting impressions: A primacy effect in memory for repe-
bering (Roediger, 2000). Most experiences of life are
titions. Psychonomic Bulletin & Review 4, 121124. encoded and stored (at least briefly) but will never be
Ebbinghaus, H. (1885; reprint 1964). Memory: A contribution to ex- retrieved and thus will have no real consequence for
perimental psychology. New York: Dover. the individual. Encoding and storage are cheap, in
Glenberg, A., Smith, S. M., and Green, C. (1977). Type 1 rehearsal: the sense that all events leave their traces. Retrieval
Maintenance and more. Journal of Verbal Learning and Verbal
Behavior 16, 339359.
is the process by which latent information is actual-
Greene, R. L. (1992). Human memory: Paradigms and paradoxes. Hil- ized in ongoing behavior. A person may or may not
lsdale, NJ: Erlbaum. realize that past events are being retrieved and are af-
(2001). Repetition effects in immediate memory in the ab- fecting current behavior. Explicit retrieval is referred
sence of repetition. In H. L. Roediger III, J. S. Nairne, I. to as remembering; when previous behavior affects
Neath, and A. M. Surprenant, eds., The nature of remembering:
Essays in honor of Robert G. Crowder. Washington, DC: Ameri-
ongoing performance outside of awareness, psycholo-
can Psychological Association. gists refer to it as priming of the behavior. The dis-
Guthrie, E. R. (1935). The psychology of learning. New York: Harper. tinction corresponds roughly to the contrast between
Harlow, H. F. (1949). The formation of learning sets. Psychological explicit and implicit memory processes (McDermott,
Review 56, 5165. 2000).
Hintzman, D. L., and Block, R. A. (1971). Repetition and memory:
Evidence for a multiple-trace hypothesis. Journal of Experimen- The division of processes into those affecting en-
tal Psychology 88, 297306. coding, storage, or retrieval seems simple in concept
Hintzman, D. L., and Curran, T. (1995). When encoding fails: In- but cannot be completely defended in practice. One
structions, feedback, and registration without learning. Memo-
ry & Cognition 23, 213226. reason is that it is impossible to distinguish cleanly be-
Nickerson, R. S., and Adams, M. J. (1979). Long-term memory for tween encoding and storage processes, because the
a common object. Cognitive Psychology 11, 287307. two are inextricably connected. When does initial
RETRIEVAL PROCESSES IN MEMORY 581
learning (encoding) end and maintenance of infor- rushed). After a first test, they would be given a sec-
mation over time (storage) begin? There is no clear ond and then a third test under identical conditions
answer to this question; suggestions provided by some without intervening study of the material. An almost
theorists to cut this Gordian knot are relatively arbi- universal finding in such experiments is that people
trary. However, separation between the bundle of will recall items on the second and third tests that they
processes referred to as encoding and storage, on the did not recall on the earlier tests, a phenomenon
one hand, and those involving retrieval, on the other, called reminiscence (Ballard, 1913). Of course, some
can be accomplished more directly. pictures recalled on the first test might be forgotten
on later tests, but surprisingly the reminiscence or re-
The general logic of this separation is to hold covery between tests often outweighs the inter-test
conditions of encoding and storage constant and to forgetting. When total recall improves over tests, this
manipulate only conditions of retrieval. For example, phenomenon is called hypermnesia (Erdelyi and
two groups of people could be presented with materi- Becker, 1974). Whereas reminiscence (recall of items
al (lists of words or sets of stories) to remember and on a later test that could not be recalled on an earlier
could be treated identically until the time they are test) almost always occurs in experiments, hypermne-
tested. Then one group of people might simply be sia is observed more rarely and usually under free re-
given a blank sheet of paper and asked to recall all call conditions (i.e., without retrieval cues). Under
that they can of the material. Imagine that they recall certain conditions the phenomenon is quite reliable,
40 percent of the materials under these free recall so the challenge is to specify the necessary conditions
conditions (so called because they are given no exter- for its observation. One idea is that relational process-
nal cues to aid recall and are free to recall material in ing (associating the materials with one another) pro-
any order). This measure might be thought to reflect tects against forgetting across repeated tests whereas
the amount of information that people have encoded processing of individual items (providing features
and storedthe amount they knowbut this conclu- that distinguish them from others) leads to recovery
sion would ignore the possibility that the bottleneck of new items across tests (Burns, 1993). Whichever is
in performance is at the retrieval stage. Perhaps the the caseand some argue for a hybrid theorythe
people really have encoded and stored much more, phenomena of reminiscence and hypermnesia point
but have simply failed to retrieve the extra material. up again that a single test of retention provides a
This possibility can be examined by testing another faulty assessment of the amount of information stored
group of subjects who are given retrieval cues to in memory (Roediger and Challis, 1989).
prompt recall of the material. Often appropriate cues
can produce great benefits relative to free recall
(Mantyl, 1986). Testing with Retrieval Cues
The advantage of cued recall over free recall indi- The most popular method of studying retrieval
cates that more information is available (or is stored) processes is by manipulating the nature of the testing
in memory than is accessible (retrievable) on a partic- conditions, particularly the types of cues given to aid
ular test such as free recall (Tulving and Pearlstone, recollection. One precondition for studying explicit
1966). More generally, no test of memory provides a retrieval with cues is that the rememberer must be
perfect measure of information stored in memory; placed in what E. Tulving (1983) has called the re-
the retrieval processes involved in any test filter the trieval mode. That is, the individual must be attempt-
information. At best, people assess the information ing to retrieve from his or her past. For example, you
that can be produced under a particular set of retriev- might see a ladder in your environment as you walk
al conditions. Although no test or set of retrieval con- to work and it does not cue any memories; however,
ditions can ever provide a perfect window on the con- if you are asked, Tell me about an experience from
tents of memory, study of retrieval processes can your past that involves a ladder, the query will cata-
proceed meaningfully in many different ways. pult you into the retrieval mode and you will probably
come up with a relevant memory. In studies of cued
recall the retrieval mode is assumed by giving people
Repeated Testing explicit instructions to remember, which has led some
One straightforward way to study retrieval pro- theorists to ignore the concept (because it is a cons-
cesses is to test people repeatedly on the same materi- tant condition). However, retrieval mode is critical to
al, under the same or differing conditions. For exam- understanding remembering.
ple, people might study sixty pictures of easily named In a typical cued recall paradigm, people are
objects and then be tested on the names of those ob- given a list of words to remember that belong to com-
jects under conditions of free recall for seven minutes mon categories. The list might be composed of words
(that is, with a large amount of time so they are not such as hawk, crow, goose, woodpecker, desk, dresser,
582 RETRIEVAL PROCESSES IN MEMORY
couch, and footstool, representing birds and articles of gory names plus two items from the categories, recall
furniture. After receiving a long list with many words of the remaining items from the categories will be bet-
and categories, some people are tested under condi- ter when only category names are given as cues. Pro-
tions of free recall (recall the words in any order) and viding some items from the category in addition to
some under conditions of cued recall (the same in- the category names will recreate the learning situa-
struction, but now the names of the categories are tion better than just giving the category names, but in
provided as retrieval cues). The typical finding is that fact such item cues hurt recall.
people tested with category names as retrieval cues Explaining this retrieval inhibition has proved to
recall many more items than those tested without be a challenge; much research has established the va-
cues, showing again the disparity between informa- lidity of the finding and eliminated many artifactual
tion that is available in memory and that accessible on possibilities for the results (Nickerson, 1984). One in-
a particular test. The gains from cues are genuine and terpretation links the part-list cuing effect to a cue-
not due merely to guessing items belonging to the cat- overload principle, an idea embedded in the forego-
egories, because the items used are typically not the ing paragraphs but not named. The cue-overload
most likely to be guessed (robin and sparrow are avoid- principle states that a retrieval cue becomes less effec-
ed as study materials in favor of crow and woodpecker). tive as more events are subsumed under the cue (Wat-
What causes retrieval cues to be effective? A pri- kins, 1975) because each extra event makes the cue
mary consideration is what type of information was less distinctive with regard to any particular event. So,
learned and how it was encodedwhat information for example, a category name retrieval cue provides
is stored in memory. The general principle governing better recall of the studied members of a category if
retrieval of such stored information is called the en- two items were given in each category of the list rather
coding specificity principle: Retrieval cues are effec- than six items. In the case of part-list cues, it may be
tive to the extent that they help reinstate or recreate assumed that presentation of the category members
processes involved in original learning (Tulving, at test somehow adds to the number of items sub-
1983). The idea is that events are encoded in specific sumed under the category name cue and thereby re-
wayspeople retain specific coded features of their duces recall. Numerous observations accord with the
experiences that may comprise the memory traces of cue overload principles interpretation of the part-list
these experiences. Retrieval cues are then effective to cuing effect, although other theories exist as well.
the extent that they match or overlap the specific en- The cue overload principle complements the en-
coded features. In addition, the match between cues coding specificity principle in making sense of the
and traces must be distinctive for provoking a specific variable effectiveness of retrieval cues. Further, the
memory; the cue should specify one event and not cue-overload idea shows that both the compatibility
many events. A cue such as remember the lecture of cue to the encoded trace and the distinctiveness of
is ineffective in aiding recollection of a particular lec- the process matters. That is, if one has had many ex-
ture because it is too general. You have been to many periences that leave similar traces, a retrieval cue may
lectures. match too many of the traces to be effective.
Numerous laboratory experiments have con-
firmed the essence of the encoding specificity princi- State-Dependent Retrieval
ple. If people are in the retrieval mode, retrieval cues Alcohol and other drugs having a depressing ef-
that more precisely match, or recreate, the original fect on the central nervous system are known to im-
features of the learning experience (and not other ex- pair retention of information. The usual interpreta-
periences) promote better recall (see Roediger and tion is that alcohol interferes with the neural
Guynn, 1996, for numerous examples). This is not to processes that underlie encoding and storage of in-
say that all retrieval phenomena are well accounted formation, or the consolidation of information. This
for, because empirical problems exist. For example, is likely true, but may not represent the whole story
certain types of cues that seem as if they should be ef- of drug-induced amnesia. Retrieval factors are at
fective are not; in some cases, seemingly good re- work, too. Clinicians working with alcoholic patients
trieval cues actually hinder rather than help recall. have observed that the patient may, for instance, hide
One example is part-list cuing inhibition, wherein a paycheck while drunk and then not be able to re-
giving people part of a list of studied items hurts remember where it is hidden when sober. However, the
call relative to control conditions (Slamecka, 1968). next time the patient gets drunk, the check may be re-
For example, if people are given lists of words belong- covered. The phenomenon suggests that successful
ing to common categories (such as the aforemen- retrieval of memories may depend on matching the
tioned examples) and then at test time either are pharmacological states in which information is
given only category names as cues or are given cate- learned and used.
RETRIEVAL PROCESSES IN MEMORY 583
This phenomenon of state-dependent retrieval Table 1

(better recall when pharmacological states of learning
and testing match rather than mismatch) has been
verified in laboratory experiments. In one case (Eich,
Weingartner, Stillman, and Gillin, 1975) volunteer
students smoked marijuana or a placebo cigarette be-
fore being exposed to a categorized word list. Four
hours later the subjects again smoked either a mari-
juana cigarette or a placebo and then were tested on
the material, first by free recall and then by cued re-
call in which category names served as the retrieval
cues.
The results are shown in Table 1; first consider
the free recall results, where the number of words re-
called from the set of 48 is shown. If people were
sober both when they studied the words and when
they were tested on them, they performed best (11.5
words recalled). If they were under the influence of
Do state-dependent retrieval phenomena exist
marijuana at study but sober at test time, they recalled
with states other than drug states? The conditions
fewest (6.7). This condition represents the usual case
most often investigated are moods, in studies where
of drug-induced amnesia, when people experience
researchers induce happy or depressed moods in
events under the drug but are sober when tested. Is
people by various means prior to study or test of ma-
this effect due only to encoding and storage factors,
terial. The expected finding is that congruence of
or are retrieval factors at work, too? This question can
mood at study and test should produce better reten-
be answered by examining the last row: When people
tion than when moods mismatch. This result has been
were drugged at both study and test time, they re-
reported in some studies, but there have been numer-
called more words (10.5) than when they were
ous failures to replicate it and the reasons for this
drugged only during study time (6.7). Just as in the
state of affairs are not well understood at this point.
anecdote about the alcoholics related above, reten-
tion improved when the pharmacological state at test
time matched that at study time. Do not conclude Transfer-Appropriate Processing
from this experiment that drugs improve memory,
A viewpoint related to the encoding specificity hy-
because they usually do not. (When people learned
pothesis is transfer-appropriate processing, which
the information sober but were tested under marijua-
emphasizes that all retention tests can be considered
na, they performed worse than when tested sober.)
as cases of transfer of prior experience to the test situ-
Although state-dependent retrieval is a real phe- ation. Depending on the nature of the task used to as-
nomenon, it usually occurs only under free recall con- sess memory, some experiences will provide good
ditions, as can be verified by examining the cued re- transfer and others will provide poor transfer. Fur-
call results. The category names served as good ther, this approach emphasizes the relativity of mem-
retrieval cues, because cued recall was better than free ory tests: Some methods of learning may prove supe-
recall in all four conditions. However, the state- rior for one type of test but disastrous for another.
dependent retrieval effect (better recall in the drug- The phrase transfer-appropriate processing was first
drug study and test condition compared with the used to explain some puzzling results obtained in the
drug-sober condition) has vanished. These results are levels of processing tradition (Morris, Bransford, and
broadly consistent with the encoding specificity hy- Franks, 1977). Under many conditions, if people
pothesis. Under conditions of free recall, a persons study events while focusing on their meaning, they re-
pharmacological state can serve as a retrieval cue, and tain the events better later than if they had focused on
if the cues match between study and test conditions, other aspects of the events, such as what they look or
performance is enhanced. However, when powerful sound like, while studying them (Craik and Tulving,
external retrieval cues are provided, they overshadow 1975). For words, retention is better on many tests
the weak state cues and render them ineffective. after people have generated meaningful associations
This account explains the common finding that state- for the words (thus forcing attention to their mean-
dependent retrieval effects are rarely found on tests ing) than when rhyming words have been generated
employing cued recall (Eich, 1989). (causing attention to sounds or phonemes).
584 REWARD LEARNING
Most of the tests showing the superiority of mean- Bibliography

ingful encodings have been those, such as recall or Ballard, P. B. (1913). Oblivescence and reminiscence. British Jour-
recognition, that are thought to rely heavily on mean- nal of Psychology Monograph Supplements 1, 182.
ing (Roediger and Guynn, 1996). But suppose a test Burns, D. J. (1993). Item gains and losses during hypermnesic re-
call: Implications for the item-specificelational information
were given for the sound of words following study ex- distinction. Journal of Experimental Psychology: Learning, Memo-
periences encouraging attention to either the mean- ry, and Cognition 19, 163173.
ing or the sound of words; e.g., the word beagle was Craik, F. I. M., and Tulving, E. (1975). Depth of processing and the
studied in a long list and the retrieval cue is, Recall retention of words in episodic memory. Journal of Experimental
a word on the list that rhymed with legal. When such Psychology: General 104, 268294.
Eich, E. (1989). Theoretical issues in state-dependent memory. In
tests were constructed, the results came out largely as
H. L. Roediger, III, and F. I. M. Craik, eds., Varieties of memory
expected: Having people think about the rhyming as- and consciousness: Essays in honour of Endel Tulving. Hillsdale,
pects of words during study produced better perfor- NJ: Erlbaum.
mance on tests requiring knowledge of the sound of Eich, J. E., Weingartner, H., Stillman, R. C., and Gillin, J. C.
the words than did study experiences emphasizing (1975). State dependent accessibility of retrieval cues in the
the meaning of words (Morris, Bransford, and retention of a categorized list. Journal of Verbal Learning and
Verbal Behavior 14, 408417.
Franks, 1977). Therefore, the ways in which one
Erdelyi, M. H., and Becker, J. (1974). Hypermnesia for pictures:
studies events are not inherently good or bad for later Incremental memory for pictures but not for words in multi-
retention; instead, whether study strategies are good ple recall trials. Cognitive Psychology 6, 159171.
or bad depends on their relation to the nature of the Mantyl, T. (1986). Optimizing cue effectiveness: Recall of 500 and
test. Learning experiences transfer well or poorly de- 600 incidentally learned words. Journal of Experimental Psychol-
pending on the nature of the test and the type of ogy: Learning, Memory, and Cognition 12, 6671.
McDermott, K. B. (2000). Implicit memory. In A. E. Kazdin, ed.,
knowledge it requires. The same idea is inherent in The encyclopedia of psychology. New York: American Psychologi-
the encoding specificity principle. cal Association and Oxford University Press.
The concept of transfer-appropriate processing Morris, C. D., Bransford, J. D., and Franks, J. J. (1977). Levels of
processing versus transfer appropriate processing. Journal of
has been applied to several different problems. One
Verbal Learning and Verbal Behavior 16, 519533.
is the explanation of differences between explicit tests Nickerson, R. S. (1984). Retrieval inhibition from part-set cuing:
of memory (those in which people are told that their A persisting enigma in memory research. Memory & Cognition
memories are being tested) and implicit tests of mem- 12, 531552.
ory (those in which people are simply given a new task Roediger, H. L. (1990). Implicit memory: Retention without re-
and retention is measured by how prior experiences membering. American Psychology 45, 1,0431,056.
(2000). Why retrieval is the key process in understanding
transfer to the new task). Many implicit memory tests human memory. In E. Tulving, ed., Memory, consciousness, and
seem to involve perceptual components and to bene- the brain: The Tallinn conference. Philadelphia: Psychology
fit from appropriate perceptual processing, whereas Press.
many explicit tests depend upon meaningful process- Roediger, H. L., and Challis, B. H. (1989). Hypermnesia: In-
ing. Numerous experiments have confirmed that creased recall with repeated tests. In C. Izawa, ed., Current is-
sues in cognitive processes: The Tulane-Floweree symposium on cog-
these two broad areas of experience (perceptual, con-
nition. Hillsdale, NJ: Lawrence Erlbaum Associates.
ceptual) differentially affect certain tests in the pre- Roediger, H. L., and Guynn, M. J. (1996). Retrieval Processes. In
dicted manner (Roediger, 1990). E. L. Bjork and R. A. Bjork, eds., Memory: Handbook of percep-
tion and cognition. San Diego, CA: Academic Press.
Slamecka, N. J. (1968). An examination of trace storage in free re-
Related Topics call. Journal of Experimental Psychology 76, 504513.
Tulving, E. (1983). Elements of episodic memory. New York: Oxford
Retrieval processes play a role in all memory phe- University Press.
nomena, so the coverage in this entry has perforce Tulving, E., and Pearlstone, Z. (1966). Availability versus accessibil-
been selective. For example, the fact that distinctive ity of information in memory for words. Journal of Verbal
events are well remembered may be interpreted in Learning and Verbal Behavior 5, 381391.
terms of the cue overload principle; similarly, the in- Watkins, M. J. (1975). Inhibition in recall with extralist cues.
Journal of Verbal Learning and Verbal Behavior 14, 294303.
hibition from part-list cues may be related to the tip-
of-the-tongue phenomenon wherein people are Henry L. Roediger III
blocked from recalling well-known information by in- Michelle L. Meade
trusion of related information. All memories depend
not just on conditions of encoding and storage, but
also on myriad retrieval factors.
See also: CODING PROCESSES; DRUGS AND MEMORY; REWARD LEARNING

EMOTION, MOOD, AND MEMORY; IMPLICIT
MEMORY See: REINFORCEMENT OR REWARD IN LEARNING
RIBOT, THODULE 585
RIBOT, THODULE (18391916)

Thodule Armand Ribot (18391916) was born in
Guingamp, Brittany. After attending lyce in Saint-
Brieuc, he entered the Ecole Normale Supricure at
Paris in 1862. He received his degree in philosophy
in 1865, and until 1872 he taught philosophy in the
secondary schools of Vesoul and Laval. In 1870 Ribot
published his first work, La psychologie anglaise contem-
poraine. Seven years later he gave up teaching so that
he could concentrate on writing. He also attended
clinical courses in psychiatry given by Valentin Mag-
nan, Benjamin Ball, Jules Luys, Flix Voisin, and
Jean-Martin Charcot, then defended his thesis,
Lhrdit psychologique. In 1876 Ribot and Hip-
polyte Taine founded the journal Revue Philosophique,
which is still published.
In 1885, Ribot started a course in experimental
psychology at the Sorbonne; in 1888, through the in-
fluence of Ernest Renan, a chair of experimental and
comparative psychology was created at the Collge de
France that Ribot occupied until he retired in 1901.
He was elected to the Acadmie des Sciences Morales
et Politiques (Section of Philosophy) in 1899.
Ribot was responsible for creating in France sci-
entific psychology, rejecting a psychology that de- Thodule Ribot (Psychology Archives, University of Akron)
pended on spiritualism and introspection in favor of
one that depends on facts and must agree with known
physiological and biological data. Ribot was interest- ory, defined by a double capacity of conservation and
ed in pathological psychology because it enabled one of reproduction of certain states (for example, a skill);
to understand normal psychological mechanisms by the recognition and localization in the past that are
discovering the laws that govern facts. Influenced by carried by consciousness are exclusively psychological
Herbert Spencers evolutionism, Ribot described, as and do not constitute memory. Ribot applied his law
did Hughlings Jackson, the law of regression (or of to aphasias, which he regarded as partial amnesias.
dissolution) that controls pathological mental phe-
nomena, such as the amnesias. Importantly, Ribots Ribots influence was significant because it repre-
contributions were purely intellectual, the result of sented the beginnings of pathological psychology,
personal reflection upon events reported by others, which included neuropsychology. Two of his students
which he categorized and regrouped. He never tried influenced psychology: Pierre Janet, who succeeded
to construct models. His work was empirical and ra- him at the Collge de France, and Alfred Binet.
tional rather than experimental. Ribots biological concepts led the philosopher Henri
Bergson to write Matire et mmoire (1896).
Ribot is probably best known today for his law of
regression in the amnesias, Ribots Law. The law out- Bibliography
lines in a logical fashion the progressive dysfunction Centenaire de Th. Ribot. Jubil de la psychologie scientifique franaise
of memory in disease. First to be affected are recent 183918891939 (1939). Agen: Imprimerie Moderne.
memories. Second, personal memories disappear, Dugas, L. (1924). Le philosophe Thodule Ribot. Paris: Payot.
Gasser, J. (1988). La notion de mmoire organique dans loeuvre
going downward to the past. Third, things acquired
de T. Ribot. History and Philosophy in Life Sciences 10, 293313.
intellectually are lost bit by bit; last to disappear are Ribot, T. (1881). Les maladies de la mmoire. Paris: Baillire.
habits and emotional memories. Thus, Ribots Law (1883). Les maladies de la volont. Paris: Baillire.
refers to progressive amnesia as a temporal gradient (1885). Les maladies de la personnalit. Paris: Alcan.
going from the most recent to the oldest memories. (1889). La psychologie de lattention. Paris: Alcan.
(1896). La psychologie des sentiments. Paris: Alcan.
For Ribot this law implied that memory depends
(1897). Lvolution des ides gnrales. Paris: Alcan.
upon permanent modifications and organization of (1900). Limagination cratrice. Paris: Alcan.
neurons, and it is their disorganization that leads to (1905). La logique des sentiments. Paris: Alcan.
amnesia. Ribots Law considers only one type of mem- (1907). Essai sur les passions. Paris: Alcan.
586 RIGHT HEMISPHERE
(1910). Problmes de psychologie affective. Paris: Alcan. RIGHT HEMISPHERE

(1914). La vie inconsciente et les mouvements. Paris: Alcan.
See: KNOWLEDGE SYSTEMS AND MATERIAL-
Jean-Louis Signoret
SPECIFIC MEMORY DEFICITS
S
SAVANT SYNDROME The condition was first described in 1887 by J.
Langdon Down (better known for having described
Savant syndrome (formerly called idiot savant syn- Down syndrome). He was struck by the paradox of de-
drome) refers to an exceedingly rare but remarkable ficiency and superiority in a number of cases he saw
condition in which persons with severe mental handi- as superintendent of a hospital in England. At that
caps have some isolated but spectacular islands of ge- time in Britain the word idiot was an accepted legal
nius or brilliance that stand in stark, incongruous classification for a severe degree of mental retardation
contrast with the serious limitations of the overall and did not have the pejorative, comical connotation
handicaps. The mental handicap can be either autism the term now has. Down combined that word with the
or mental retardation. The skills, remarkable as they term savantknowledgeable personderived from
are, exist within a very narrow range of human abili- the French word savoir (to know), to denote these fas-
ties. They include music (usually piano); art (drawing cinating persons. The term is a misnomer in that al-
or sculpting); calendar calculating (the ability to give most all cases have an IQ of 40 or above, and thus
the day of the week for any past or future date); light- have moderate rather than the severe mental retarda-
ning calculating (the ability to add, multiply, subtract, tion that now-archaic legal term idiot once defined.
or divide complex numbers with lightning rapidity);
and mechanical or spatial skills including map memo- Superior memory is a trait all savants share. It is
rizing, visual measurement, unusual sensory discrimi- a particular type of memory, however: very deep but
nation such as enchanced sense of touch and smell, within a very narrow area; concrete and not richly as-
or perfect appreciation of time without knowledge of sociative; direct and nonsymbolic; nonemotional and
a clock. These skills, within these very narrow bands, seemingly automatic or unconscious. This memory
are always linked to a spectacular memory. without reckoning may be akin to what Mishkin,
Malamut, and Bachevalier (1984) refer to as habit
In some persons (talented savants) the skills are re-
memory, as opposed to cognitive memory; it relies
markable simply in contrast with the handicap; in
on a brain circuitry that is more primitive and lower
others (prodigious savants) the abilities are spectacular
than the later-developed, higher brain circuitry of
in contrast with the handicap and would be spectacu-
cognitive or associative memory. In the savant this un-
lar even if they occurred in normal persons. Savant
conscious memory presumably is relied upon as an al-
syndrome can be hereditary or it can be acquired fol-
ternative pathway to damaged higher-level cognitive
lowing central nervous system injury before, during,
memory circuitry.
or after birth. It occurs in males approximately six
times more frequently than in females. The skills can There has been no well-designed study of the
appear suddenly, without explanation, and can disap- prevalence of savant syndrome, but it has been re-
pear just as suddenly. ported to occur in 1 out of 2,000 in an institutional-
587
588 SCHIZOPHRENIA AND MEMORY
ized mentally retarded population and in as many as Memory Deficits in Schizophrenia

1 out of 10 autistic persons. Mental retardation and
Despite variable methods and the inherent heter-
autism are both developmental disabilities, but since
ogeneity of schizophrenia, a number of consistent
autism is so much less common than mental retarda-
findings appear throughout the literature (Aleman et
tion, the number of savants is generally evenly divid-
al., 1999). Here we highlight some of the patterns of
ed between those with autism and those with mental
memory deficits in schizophrenia:
retardation. In the movie Rain Man, the character
Raymond Babbit (played by Dustin Hoffman) is the While schizophrenia is associated with a wide
best-known portrayal of a savant, in that instance an range of cognitive difficulties, deficits in memory
autistic savant. It is important to remember, however, are particularly pronounced.
that not all autistic persons are savants and not all sa-
The memory impairment is selective, affecting
vants are autistic. The number of prodigious savants
primarily explicit memory. Implicit memory, in-
worldwide is estimated to be less than fifty as of 1991.
cluding perceptual priming and procedural
Talented savants are, of course, more common; but
memory, appears to be relatively intact.
the savant syndrome overall is still a rare condition.
Even within explicit memory, tests of recall (Tell
me everything you saw) are affected more than
Bibliography recognition (Did you see this before?). Patients
Mishkin, M., Malamut, B., and Bachevalier, J. (1984). Memories with schizophrenia have difficulty forming associ-
and habits: Two neural systems. In G. K. Lynch, J. L. Mc-
Gaugh, and N. M. Weinberger, eds., Neurobiology of learning
ations between items during the encoding (study)
and memory. New York: Guilford Press. period, and do not show the normal pattern of
Treffert, D. A. (1988). The idiot savant: A review of the syndrome. heightened recall after studying words that have
American Journal of Psychiatry 145 (5), 563572. common features (i.e., words that are easily asso-
(1989). Extraordinary people: Redefining the idiot savant. New ciable). Defective recall may therefore be attribut-
York: Harper & Row.
(1990). Extraordinary people: Understanding savant syndrome.
able to both poor organization of related items at
New York: Ballantine Books. encoding, and poor search strategies at retrieval.
Further examination of the explicit memory defi-
Darold A. Treffert
cit indicates a specific impairment in memory
that requires conscious recollection of the study
episode (i.e., episodic memory) as compared to
simple recognition or familiarity-based memory
processes. Patients with schizophrenia are less
SCHIZOPHRENIA AND MEMORY likely to state that they clearly remember an
Schizophrenia affects approximately 1 percent of the event, more often saying that they simply know
population worldwide. It typically involves hallucina- that the event has occurred. Word frequency
tions and delusions, also known as psychotic or posi- modulation, which leads to a greater degree of
tive symptoms. Markedly impaired social skills and conscious recollection in healthy adults, has no
cognitive deficits, also known as negative symptoms, such effect on patients with schizophrenia. Final-
are also core features. In fact, schizophrenia was origi- ly, while healthy subjects use both personally rele-
nally called dementia praecox (i.e., early onset de- vant and inter-item associations to recollect
mentia, to highlight the impairment of cognition. words, patients with schizophrenia are less likely
The cognitive domains most often affected by schizo- to do so, relying instead on bizarre and ineffectu-
phrenia include attention, memory, and language. al linkages.
These deficits are evident even before the onset of These deficits are not due solely to poor concen-
positive symptoms (during the so-called prodrome) tration, lack of motivation, distracting positive
and in untreated schizophrenic patients. It is there- symptoms, or a medication side effect.
fore unlikely that the cognitive features of schizophre-
nia result simply from chronic illness, institutionaliza- Patients with schizophrenia often suffer from se-
tion, or medication side effects. Most important, the lective impairment of memory that requires the con-
degree of cognitive decline is the best predictor of scious recollection of the study episode, but they re-
functional outcome: that is, how the patient will do in tain implicit (and other automatic) memory. Given
the community once the most severe psychotic fea- the recognized abnormality in creating effectual asso-
tures have abated. Here we will focus on abnormali- ciations, a primary role of impaired feature-
ties of memory, describing the type of memory im- binding seems to account for the episodic memory
pairment typical of schizophrenia and the cerebral impairment in schizophrenia. That is, patients with
abnormalities that may account for this impairment. schizophrenia do not appear to bind together the
SCHIZOPHRENIA AND MEMORY 589
multifold features of everyday events and therefore ulation (e.g., schizophrenic patients with poor memo-
find it hard to remember them in detail. Because the ry) and those that assess changes in cerebral activity
ability to place memories in the detailed context of during the performance of a task. There are several
the initial experience is critical for a continued sense studies of baseline cerebral activity during memory
of self and the maintenance of a personal history, ab- performance in schizophrenia, but they have yielded
normalities in this realm may relate to the aberrant discrepant results. Hence, the exact relation between
conscious experiences of schizophrenia. resting cerebral activity and memory performance in
schizophrenia remains unclear. Because measures of
baseline cerebral activity may not adequately repre-
Neuroimaging of Memory Function in sent the fleeting changes of neuronal activity associat-
Schizophrenia ed with mental processes, research has shifted to the
Uncovering the neural basis for these deficits in study of cerebral activity during the performance of
memory is an area of intensive research. Given the memory tasks.
role of the hippocampus and the prefrontal cortex These cognitive activation studies attempt to
(PFC) in normal memory, these regions are the pri- capture the neural signature of memory process-
mary focus of this pursuit. Scientists are using both ingto identify those structures involved at the time
structural neuroimaging techniques (e.g., magnetic of task performance. Researchers obtain images dur-
resonance imaging [MRI]) and functional neuroi- ing the performance of a memory task (encoding or
maging techniques (e.g., positron emission tomogra- retrieval) and during a baseline state. The subtraction
phy [PET] and functional magnetic resonance imag- of baseline activity from that demonstrated during
ing [fMRI]). In the following section we will briefly memory processing allows researchers to infer which
summarize the major findings as they apply to schizo- regions are activated during the memory. Thus it
phrenia (Weiss and Heckers, 2001). is the change in activity associated with task perfor-
mance, sometimes called recruitment, that is com-
monly reported. Using this type of neuroimaging
Structural Neuroimaging Studies paradigm, schizophrenic subjects (when compared to
The spatial resolution afforded by MRI allows in- controls) appear to show a relative lack of task-related
vestigators to correlate performance on cognitive activation in regions known to play a role in normal
tasks with the volume of specific brain regions, includ- memory (e.g., PFC and hippocampus).
ing the hippocampus and subdivisions of the PFC.
Some evidence suggests just such a link in schizophre-
Frontal Lobe Abnormalities
nia, with memory performance correlating with the
integrity of these two regions. Overall, however, it has Patients with schizophrenia commonly demon-
been difficult to link memory performance to cere- strate diminished frontal lobe activation (particularly
bral structure in schizophrenia, perhaps because the in the left hemisphere) during memory task perfor-
structural changes associated with schizophrenia are mance. Fletcher and colleagues (1998) found that
subtle. Indeed, the neuropathology of schizophrenia schizophrenic subjects recalled brief word lists (fewer
may not involve overt neuronal loss or marked tissue than four words) as well as controls but that they
atrophy but rather decreases in the number of only a showed a dramatic drop-off in performance with lon-
subset of neurons or deficient connections. These ger lists, a result that suggests a deficit in explicit
subtle changes may lead to disturbances of cognitive memory. As word-list lengths increased, both groups
performance without gross morphologic abnormality initially showed increasing left PFC activation in re-
and therefore would be undetectable with structural sponse. With longer word lists, however, the patients
neuroimaging. To evaluate the possibility of a func- were unable to sustain the degree of prefrontal activa-
tional abnormality within the frontal lobes, the hippo- tion displayed by the normal subjects. These results
campus, or both, researchers must use methods that indicate that schizophrenic subjects are unable to
allow visualization of cerebral activity. maintain necessary levels of PFC activation in re-
sponse to increasing task demands.
The relative lack of memory-task-related frontal
Functional Neuroimaging Studies activation is not a universal finding. Some investiga-
Functional neuroimaging proceeds from the tors have demonstrated task-related hyperactivity of
premise that focal neuronal activity relates either to the DLPFC in schizophrenic subjects, possibly indi-
regional cerebral blood flow (rCBF) or to the local de- cating impaired efficiency of DLPFC function during
gree of glucose metabolism. There are two major the performance of a memory task. Task-related hy-
classes of functional imaging paradigms: those that poactivity may depend on disease-specific character-
examine the resting pattern of activity in a target pop- istics, particularly the degree of negative symptoms.
590 SCHOOL LEARNING
Deficient frontal activity has been linked to dy- Conclusion

sexecutive features of memory; the inability to orga- The results reviewed here suggest an abnormality
nize information for optimal performance. While of hippocampal and prefrontal cortex function dur-
control subjects use the implicit semantic associations ing memory in schizophrenia. The pace of discovery
between words to facilitate encoding and support sub- in this area has been quite rapid. The next major
sequent retrieval, schizophrenic subjects do not use strides in the understanding of the neural basis of
this strategy. Instead, patients tend to use ineffectual schizophrenia await the advent of even better tech-
or bizarre associations, leading to poorer recall. The nology that would allow more precise temporal and
inability to use the normal associative mechanisms spatial assessment of neural functions.
that take advantage of inherent properties of the
items to be remembered correlates with a lack of acti- See also: BEHAVIOR THERAPY
vation in the inferior frontal lobes bilaterally. Difficul- Bibliography
ty in organizing encoding and retrieval strategies in Aleman, A., Hijman R., de Haan, E. H. F., and Kahn, R. S. (1999).
schizophrenia therefore appear to be associated with Memory impairment in schizophrenia: A meta-analysis. Amer-
lower memory task-related metabolism in regions of ican Journal of Psychiatry 156, 1,3581,366.
prefrontal cortex. Fletcher, P. C., McKenna, P. J., Frith, C. D., Grasby, P. M., Friston,
K. J., and Dolan, R. J. (1998). Brain activations in schizophre-
nia during a graded memory task studied with functional
neuroimaging. Archives of General Psychiatry 55, 1,0011,008.
Temporal-Lobe Abnormalities Heckers, S. (2001). Neuroimaging studies of the hippocampus in
schizophrenia. Hippocampus 11, 520528.
Recent studies in patients with schizophrenia
Heckers, S., Rauch, S. L., Goff, D., Savage, C. R., Schacter, D. L.,
demonstrate abnormal baselin levels of activity in the Fischman, A. J., and Alpert, N. M. (1998). Impaired recruit-
temporal lobe and memory task-related hypoactivity of ment of the hippocampus during conscious recollection in
the hippocampus. Heckers and colleagues (1998) schizophrenia. Nature Neuroscience 1, 1,3181,323.
Weiss, A. P., and Heckers, S. (2001). Neuroimaging of declarative
studied subjects after they learned words using a deep
memory in schizophrenia. Scandinavian Journal of Psychology
semantic encoding strategy or a shallow perceptual 42, 239250.
one. During scanning, subjects were asked to com-
Stephan Heckers
plete three-letter word stems of the target words. Nor-
Anthony P. Weiss
mal subjects recalled more words after deep semantic
encoding, which was associated with a recruitment of
the hippocampus; the attempt to retrieve the words
encoded using the shallow perceptual strategy result- SCHOOL LEARNING
ed in bilateral activation of the prefrontal cortex.
School learning is the acquisition of knowledge, sub-
Schizophrenic subjects recalled fewer words after
ject matter, information, understanding, and skill
deep semantic encoding, which was associated with an
from teaching. Research on learning and memory
impaired recruitment of the hippocampus. Increased
provides teachers with essential scientific knowledge
hippocampal activity at baseline and impaired re-
that is useful for understanding and improving school
cruitment during episodic memory retrieval might
learning.
represent the functional correlate of an abnormal
cortico-hippocampal interaction in schizophrenia.
These findings complement the studies of abnormal What Can Be Learned from Teachers?
frontal cortex activity in schizophrenia. A frontal- Schools and teachers have made much progress
hippocampal disconnection in schizophrenia may in- in the teaching of complex subject matter. Today
volve temporal-lobe circuits that play a role in the teachers teach difficult subjects, such as calculus in
processing of language, pathways to and from the high school, that were previously thought to be im-
hippocampal formation, or both. possible for students of that age to learn. Even the
learning of expert performance involves intensive
As with the findings of frontal hypoactivity, ab- work with teachers. See Ericsson and Charness (1994)
normalities in temporal lobe activation are not uni- and Ericsson, Krampe, and Tesch-Romer (1993) for
versal. Several studies, despite showing impairments discussions of expert performance.
of explicit memory in the schizophrenic cohort, have
failed to demonstrate functional abnormalities in the
temporal lobe. It is unclear whether this is due to dif- Ancient and Modern Conceptions of
ferences in experimental design, image acquisition, Learning
disease characteristics, or other as yet unknown fac- Modern conceptions of learning originated in an-
tors. cient Greece and Rome. Since Socrates tried to teach
SCHOOL LEARNING 591
a slave boy the Pythagorean theorem in the Meno by and can learn to control and to monitor their own
having him construct the theorem from innate ideas learning (metacognition).
rather than by telling him the theorem, teachers and
researchers have studied how students acquire, re- Most such research focuses on how students learn
member, and use knowledge. Platos emphasis on when they work individually with teachers, mentors,
constructive learning resembles modern conceptions and coaches. However, recent research on situative
of school learning that emphasize the importance of learning emphasizes social context, viewing learning
having students construct meaning by relating new as the product of systems of group interaction (Gr-
concepts to familiar ideas called schemata or to mem- eeno, Collins, and Resnick, 1996). Research on group
ories of experience called scripts. learning in schools has begun to identify the condi-
tions best suited to conceptual learning by coopera-
Aristotle wrote that students learn information by tive groups. Cohen (1994) finds that cooperative
forming relations or associations between new infor-
groups interact well in tasks that require cooperation,
mation and previously learned information or con-
present loosely structured problems, provide pre-
cepts that are stored in memory as images. He be-
training in group work, and impose little direct teach-
lieved that subject matter and other abstract
er supervison. Cooperative group learning can play
information are logically organized in memory into
an important role in achievement in concept learn-
hierarchies that consist of classes, or general con-
cepts, divided into species, or smaller groups. His ing. For example, Kourilsky and Wittrock (1992)
ideas about hierarchically organized abstract verbal found that by working in cooperative groups, high
information resemble modern understanding of the school students increased their learning of concepts
organization of semantic memory (Kintsch, 1980). in economics by 30 percent when they generated rela-
Aristotles ideas about associations underlie much of tions between their experience and the abstract sub-
modern thinking about how to enhance learning, reject matter.
tention, and understanding in schools by teaching
students to associate new information and prior
knowledge through the use of imagery and verbal Memory
processes. The information students have organized into
schemata in their semantic (abstract and verbal)
Aristotle influenced school learning in ancient
memory and into scripts in their episodic (concrete
times, when memory-training systems (mnemonics)
and imaginal) memory influences their learning and
many of them similar to present-day techniques
comprehension. In schools, teaching procedures
were taught to students, orators, and statesmen (Cic-
focus on building relations between these semantic
ero, 1967). In the Middle Ages, Aristotles ideas about
and episodic types of information in memory and the
forming associations between new and familiar con-
cepts influenced the widespread use of statues, paint- new information to be learned. Mining these relations
ings, mosaics, and murals to relate various abstract boosts memory or comprehension or both.
moral concepts to familiar contexts. Modern concep-
tionsof school learning as association building and
knowledge acquisition as the construction of hierar-
Learning to Remember Facts
chically organized relations between new and old in- Factual informationdates, names, places, and
formationstill reflect these ancient ideas of Plato so onis often difficult to remember because it seems
and Aristotle. to bear little relation to other information of interest
or importance to students. Memory for factual infor-
mation improves when learners construct relations,
How Students Learn Individually and in even arbitrary ones, between their semantic or epi-
Groups sodic memories and the facts to be learned. Mnemon-
Recent research in educational psychology, cog- ic systems are built on that principle. For example,
nitive psychology, and neuropsychology helps us to lists of words can be remembered in order by forming
understand how memory, knowledge acquisition, at- images between the words on the list and mnemonic
tention, learning strategies, and metacognition (self- words organized into a series (e.g., one is a bun, two
control of learning) influence school learning. This is a shoe). Capitals of states can be remembered by
research shows the importance of getting students to forming colorful or comic associations between the
attend to the information to be learned, to relate it ac- name of the state and its capital city (Levin, 1985).
tively to their knowledge and experience, and to use The objective of this procedure is to make the learn-
it to solve problems. The research also shows that stu- ing of facts meaningful in some way, much as Aristotle
dents can learn how to learn (learning strategies) and Cicero did with mnemonic devices.
592 SCHOOL LEARNING
Acquiring Knowledge in School Attention

Knowledge acquisition requires understanding Research in neuropsychology and in cognitive
and comprehension of what one has learned. Re- psychology shows the importance of attention in
search on reading comprehension and science learn- learning in schools. Long-term, voluntary attention
ing shows the importance of schemata (Rumelhart, (i.e., ability to focus ones thoughts on themes or con-
1980) and student preconceptions (Wittrock, 1994) in cepts rather than classroom distractions) is especially
determining what students will understand and how important. Mentally retarded or learning-disabled
their comprehension can be improved. For example, children and some hyperactive children have trouble
when students read a passage from a new perspective, maintaining needed levels of attention in school.
they can comprehend it differently. These children often can be taught to control their
Elementary school students preconceptions voluntary attention by learning self-talk strategies
about concepts in science influence their comprehen- that emphasize the relevant learning task. Research
sion in school. Students who have unscientific concep- on divided attention shows that learners can process
tions of electricity often have difficulty in learning two different tasks, especially familiar ones, at the
how it functions. Students in elementary school and same time if each of the tasks comes from a different
in high school also differ in their understanding of sensory mode. But when both tasks simultaneously
the composition of matter, such as gases. Benson, enter the same sensory register, e. g. the ears, only
Wittrock, and Baur (1993) found that many students, one message will receive much attention (Anderson,
even in high school, do not understand that gases are 1990).
made of particles (molecules). It is important that In classrooms teachers can apply knowledge from
teachers take into account such student preconcep- research on attention to understand how objectives
tions. and questions enhance learning. Objectives and ques-
Research on the teaching of comprehension and tions direct student attention in productive ways de-
the acquisition of knowledge in schools shows that signed to stimulate learning. Clark and Wittrock
students in schools acquire knowledge and increase (2000) discuss methods to facilitate attention. These
their comprehension when they generate meanings methods include increasing motivation to attend,
for information, such as subject matter, stories, and avoiding divided attention, directing attention by
concepts (Wittrock, 1998). The generation of summa- cueing important concepts, and providing advance
ries, pictures, inferences, applications, and examples organizers (coherent introductions).
is an effective way to organize information and to re-
late it to experience and knowledge stored in memo-
ry. Comprehension and understanding function best Learning Strategies and Metacognition
when students construct two types of relations: among Learning strategies are procedures for construct-
the parts of the subject matter and between the sub- ing relations across concepts and between new infor-
ject matter and the learners knowledge and experi- mation and experience and knowledge (e.g., con-
ence (Wittrock, 1990). Both verbal processes (e.g., structing summaries, inferences, and pictures).
summaries and inferences) and imagery (pictures, Metacognition is awareness of and control over ones
graphs, and diagrams) enhance comprehension when thoughts (e.g., planning to use a comprehension
students use them to construct these two types of rela- strategy). Learning strategies and metacognition
tions. often produce large increases in comprehension and
Concise summaries that contain appropriate ver- knowledge acquisition when students learn to plan
bal and visual components (words and pictures), as and to monitor their thinking and studying, to con-
opposed to mere written summaries, improve expla- trol their attention, and to use comprehension strate-
nations and foster comprehension (Mayer et al., gies such as analogy building and summarizing.
1996). Student generation of summaries or of analo- Studies on teaching thinking skills, learning strate-
gies as they read a text also can significantly increase gies, and metacognition show large gainsaround 50
comprehension without increasing the time needed percentin comprehension when students use learn-
to learn (Wittrock and Alesandrini, 1990). As with the ing strategies or metacognition in school (Mayer and
facilitation of memory of factual information, the stu- Wittrock, 1996; Lambert and McCombs, 1998).
dents building of relations between memory and new
information enhances school learning. However, with
comprehension and understanding, these relations Conclusion
are less arbitrary, more organized, more meaningful, Modern research on learning and memory has
and more integrated with schemata and scripts (Gra- greatly enhanced the understanding of school learn-
bowski, 1996). ing and its improvement. Unfortunately, the power
SECOND MESSENGER SYSTEMS 593
and the utility of these research findings has only Plato. Meno. Loeb Classical Library. Cambridge, MA: Harvard Uni-
begun to affect teaching in the schools. There is a use- versity Press.
Rumelhart, D. E. (1980). Schemata: The building blocks of cogni-
ful convergence between the ancient writings of Plato tion. In R. J. Spiro, B. C. Bruce, and W. F. Brewer, eds., Theo-
and Aristotle about constructive learning and associa- retical issues in reading comprehension. Hillsdale, NJ: Erlbaum.
tion building and modern scientific research on Wittrock, M. C. (1990). Generative processes of comprehension.
learning and knowledge acquisition; both the ancient Educational Psychologist 24, 345376.
and modern precepts indicate that school learning in- (1994). Generative science teaching. In P. J. Fensham, R.
F. Gunstone, and R. T. White, eds., The content of science: A con-
volves the construction of meaning for new informa- structivist approach to its teaching and learning. London: Falmer
tion by attending to it, organizing it, and relating it Press.
to ones knowledge and experience. (1998). Cognition and subject matter learning. In B. L. Mc-
Combs and N. Lambert, eds., How students learn. Washington,
DC: American Psychological Association.
See also: ARISTOTLE; MNEMONIC DEVICES; PROSE Wittrock, M. C., and Alesandrini, K. (1990). Generation of summa-
RETENTION ries and analogies and analytic and holistic abilities. American
Educational Research Journal 27, 489502.
Bibliography M. C. Wittrock
Anderson, J. R. (1990). Cognitive psychology and its implications. New
York: W. H. Freeman and Company.
Aristotle (1928). The works of Aristotle. Vol. 1, ed. W. D. Ross. Ox-
ford: Clarendon Press.
(1964). On memory and recollection. In On the soul (De SECOND MESSENGER SYSTEMS
anima); Parva naturalia; and on breath, trans. W. S. Hett. Cam-
bridge, MA: Harvard University Press. All cells of the body respond to their environment.
Benson, D. L., Wittrock, M. C., and Baur, M. E. (1993). Students For most cell types, the responses can be both general
preconceptions of the nature of gases. Journal of Research in and be related to the cells particular function. An ex-
Science Teaching 30, 587597. ample of a general response is the regulation of sugar
Cicero. (1967). De oratore, ed. E. W. Sutton, Cambridge, MA: Har- utilization: When sugar is plentiful, glucose is poly-
vard University Press.
Clark, R., and Wittrock, M. C. (2000). Psychological principles in- merized to glycogen, a storage form. When sugar is
training. In S. Tobias and J. D. Fletcher, eds., Training and re- scarce, the biochemistry of the cells will be adjusted
training. New York: Macmillan. so that glycogen is broken down to make sugar avail-
Cohen, E. G. (1994). Restructuring the classroom: Conditions of able. Stimulation of gland cells to release their secre-
productive small groups. Review of Educational Research 64, tion is a specialized response. In neurons the regulat-
135.
Ericsson, K. A., and Charness, N. (1994). Expert performance, its ed specialized functions include properties of ion
structure and acquisition. American Psychologist 49, 725747. channels; availability of synaptic vesicles for exocyto-
Ericsson, K. A., Krampe, R. T., and Tesch-Romer, C. (1993). The sis (a process called mobilization); and responsiveness
role of deliberate practice in the acquisition of expert perfor- of postsynaptic receptors. These functions can be al-
mance. Psychological Review 100, 363406. tered transiently (in a time frame of minutes to hours)
Grabowski, B. L. (1996). Generative learning. In D. H. Jonassen,
ed., Handbook for Research on Educational Communications and or for much longer periods, if not permanently.
Technology. New York: Macmillan. Short-term changes are produced by modifying exist-
Greeno, J. G., Collins, A. M., and Resnick, L. B. (1996). Cognition ing proteins in the cell, modifications that are rapidly
and learning. In D. C. Berliner and R. C. Calfee, eds., Hand- reversed when the second-messenger is removed.
book of educational psychology. New York: Simon & Schuster. Long-term changes result when the second-
Kintsch, W. (1980). Semantic memory: A tutorial. In R. S. Nicker-
son, ed., Attention and performance, Vol. 8. Hillsdale, NJ: Erl- messenger pathway changes transcription, ultimately
baum. producing a change in gene expression and the pro-
Kourilsky, M., and Wittrock, M. C. (1992). Generative teaching: An duction of new proteins.
enhancement strategy for the learning of economics in coop-
erative groups. American Educational Research Journal 29, 861 Many of these responses, both general and spe-
876. cialized, are produced by signal transduction, a pro-
Lambert, N. M., and McCombs. B. L., eds. (1998). How students cess in which an extracellular stimulus activates a spe-
learn. Washington, DC: American Psychological Association. cific receptor on the surface of the responding cell. As
Levin, J. R. (1985). Educational applications of mnemonic pic- a result, the receptor initiates changes in the bio-
tures: Possibilities beyond your wildest imagination. In A. A.
Sheikh and K. S. Sheikh, eds., Imagery in education. Farming- chemical state of the cell through the production of
dale, NY: Baywood. a substance that, in turn, alters the cells responsive-
Mayer, R. E., Bover, W., Bryman, A., Mars., R., and Tapangco, L. ness. This modulating substance is called a second
(1996). When less is more: Meaningful learning from visual messenger because it is evoked by an environmental
and verbal summaries of science textbook lessons. Journal of cue (the first messenger). The process is called sig-
Educational Psychology 88, 6473.
Mayer, R. E., and Wittrock, M. C. (1996). Problem-solving transfer. nal transduction because the external stimulus is re-
In D. C.. Berliner and R. C. Calfee, eds., Handbook of education- coded (transduced) into the change in biochemical
al psychology. New York: Simon & Schuster. state. Although ionized calcium (Ca2+) is often
594 SECOND MESSENGER SYSTEMS
thought of as a second messenger, it may be more in- When the subunit binds GTP because of recep-
structive to call it a primary regulator. Along with tor activation, it dissociates from the portion of the
membrane potential, it is the ultimate governor of ex- complex. Nevertheless, it retains its association with
citability and synaptic transmission in nerve cells. the inner surface of the cells external membrane. De-
pending on the particular combination of receptor
Historically, formulation of the concepts and and G-protein, the dissociated subunit either acti-
principles for understanding second messenger sys- vates or inhibits the next component of the second
tems arose from studies of two physiological process- messenger system, which can be called a primary ef-
es: regulation of sugar metabolism in muscle and fat fector. If the subunit is stimulatory, its interaction
cells (adipocytes), and the conversion of light energy with the primary effector results in synthesis of a sec-
into nerve impulses in the rod cells of the retina. In ond messenger.
the 1950s, Sutherland and Rall (1957) discovered the
first second messenger, cyclic adenosine monophos- The cAMP system is the best-understood intracel-
phate (cAMP), and Sutherland and his coworkers lular signaling pathway. Binding of a neurotransmit-
identified the enzymatic pathway by which cAMP is tersuch as norepinephrine to the 1-adrenergic re-
synthesized and showed how it brings about the ceptoractivates a stimulatory G-protein (Gs), which
changes in sugar metabolism. At about the same time, promotes the synthesis of cAMP from adenosine
Wald (1959) investigated how rhodopsin converts the triphosphate (ATP) by adenylyl cyclase. This primary
energy of a photon into a chemical signal; the enzy- effector enzyme, which is a twelve-membrane-
matic and electrophysiological events in phototrans- spanning protein, operates only when it is associated
duction were thus identified. We now know that both with an s subunit with bound GTP. The s-cyclase
of these processes are mediated by sets of proteins en- complex also acts as a GTPase, an enzymatic activity
coded by genes that are closely related phylogeneti- that hydrolyzes bound GTP to GDP and Pi. When
cally and that operate by similar mechanisms. GTP is replaced by GDP, the s subunit dissociates
from the cyclase and reassociates with the receptor, to
be activated again. Some receptorsfor example, the
Production of Second Messenger Molecules muscarinic acetylcholine receptorinteract with an
Most second messenger systems follow a similar inhibitory G-protein (Gi). The i subunit, when acti-
molecular strategy. The pathway is initiated when an vated by the muscarinic receptor, associates with ad-
external cuemost often a neurotransmitter, a hor- enylyl cyclase to block the synthesis of cAMP. The op-
mone, an odorant, a physical stimulus such as light, posing actions of norepinephrine and acetylcholine
or a mechanical forceactivates a receptor. These re- on the cyclase through Gs and Gi represent one form
ceptors, which are single polypeptide chains, have of second messenger interaction called cross-talk.
seven membrane-spanning (serpentine) domains. An important functional aspect of signal trans-
Although these molecules have several domains ex- duction pathways, inherent in the arithmetic of the
posed on the external surface of the responding cell, relationships among these components (receptor/G-
the actual binding site for the first messenger is slight- protein/primary effector), is amplification of the ex-
ly buried within the membrane. Some of the receptor ternal stimulus. A neuron has far fewer receptor mol-
is also situated on the intracellular side of the mem- ecules than G-proteins and primary effectors. Ampli-
brane; a part of this intracellular domain associates fication occurs when the relatively few stimulated
with another protein, called a G-protein because it receptors activate many primary effectors. Moreover,
binds a guanosine nucleotide (GDP or GTP). since primary effectors are enzymes that produce the
The receptors that usually mediate signal trans- second messenger catalytically, this step amplifies the
duction are closely related members of a large gene signal even further.
family that, in addition to rhodopsin, contains adren-
ergic receptors, muscarinic acetylcholine receptors,
and receptors for serotonin, dopamine, histamine, What Second Messengers Do
and all neuropeptides. The trimeric G-proteins be- Second messengers activate secondary effectors,
long to another large gene family. They are made up enzymes that are protein kinases in most instances.
of three subunits (). The subunit binds GDP Typically these enzymes, which catalyze the transfer
when associated with an inactive receptora receptor of the terminal () phosphoryl group of ATP to hy-
that has not been activated by an external stimulus; droxyl groups of serine or threonine residues in pro-
it binds GTP when the receptor is activated. More teins, are multifunctional: They can phosphorylate
than a dozen isoforms of subunits have been identi- many different protein substrates. Secondary effector
fied, each of which interacts with different receptors. enzymes include the cAMP-dependent protein ki-
There are fewer types of the and subunits. nases (PKA), protein kinase C (PKC), and the
SECOND MESSENGER SYSTEMS 595
Ca2+/calmodulin-dependent protein kinases. Phos- leased, are then free to phosphorylate protein sub-
phorylation of substrate proteins, which can be called strates. The free C subunits remain enzymatically ac-
secondary regulators, changes their properties either tive until the concentration of cAMP within the cell
to stimulate or to inhibit their function. These falls, which results in the dissociation of the cAMP
changes in the activity of substrate proteins by phos- bound to the R subunits. Lacking cAMP, R subunits
phorylation are the means by which the responses to reassociate with C subunits, thereby inhibiting the en-
the environmental stimulus are produced. For exam- zyme.
ple, phosphorylation of a channel protein can alter
the flux of ions into the neuron to raise or lower the
membrane potential. Another example is the phos- Inositol Polyphosphates, Diacylglycerol
phorylation of transcription activators, which pro- and Arachidonic acid
motes their binding to DNA. An example important Many receptors activate phospholipase C (PLC)
to the neurobiology of memory is the cAMP response through other G-proteins. PLC catalyzes the hydroly-
element binding protein (CREB). When phosphoryl- sis of phospholipids in the external membrane of the
ated by cAMP, CREB joins with other proteins to responding cell. Phospholipids consist of a glycerol
form a complex that binds to the promoter region of moiety esterified at the first (sn 1 position) and second
genes whose transcription is stimulated by cAMP. (sn 2) hydroxyl groups to fatty acids, and, at its third,
Whether a substrate can be phosphorylated de- to a diester of phosphoric acid and one of four special
pends on the amino acid sequences around the serine alcohols (inositol in phosphatidylinositol, PI; choline
and threonine residues in the substrate protein: Each in PC; ethanolamine in PE; and serine in PS). In the
type of kinase prefers special sequences. Often more PI of nervous tissue, the fatty acid at the sn 1 position
than one type of protein kinase can phosphorylate is usually stearic (an eighteen-carbon saturated fatty
the same protein. The phosphorylations then occur acid); the second hydroxyl is usually esterified to ar-
at different sites in the substrate molecule, however. achidonic acid, an unsaturated twenty-carbon fatty
acid. The PLC activated in this second messenger sys-
A serine/threonine protein kinase can exist in sev- tem is a diesterase that hydrolyzes PI to an inositol
eral isoforms in the same neuron. The isoforms of phosphate and diacylglycerol (DAG). Inositol is an
each type of kinase are closely related, some encoded unsaturated six-membered cyclic polyalcohol that, in
by the same genes with diversity produced by alterna- addition to the phosphoryl linkage, can be phospho-
tive RNA splicing and others encoded by distinct but rylated at hydroxyl groups on the other five carbons.
closely related genes. Most protein kinases are related Most often the fourth and fifth hydroxyl groups are
phylogenetically, including tyrosine-specific kinases phosphorylated in the inositol moiety of nerve cells
that phosphorylate proteins on the hydroxyl group of (PIP2). When hydrolyzed by PLC, it is therefore called
tyrosine residues. Protein kinases that can phospho- inositol 1,4,5-trisphosphate (IP3), a water-soluble sec-
rylate a variety of substrates are called multifunction- ond messenger. Other inositol polyphosphates exist,
al. There are also dedicated kinases that phosphoryl- but it is not yet certain whether they, too, act as sec-
ate only a special protein substrate (for example, the ond messengers. Diacylglycerol, which is soluble only
-adrenergic receptor kinase) or kinases that are in lipid, remains within the membrane and also serves
present in only a few kinds of neurons (for example, as a second messenger. PLA2, which hydrolyzes PI at
the cGMP-dependent protein kinase). the sn 2 position to release arachidonic acid, also is ac-
PKA was the first protein kinase to be described. tivated by a G-protein-coupled receptor for histamine
This enzyme consists of two regulatory (R) subunits and other neurotransmitters. The arachidonic acid
and two catalytic (C) subunits. (There are two major produced by receptor-mediated activation of PLA2
types of R subunits, RI and RII, and several isoforms can serve as a second messenger itself or it can be con-
of C.) The PKAs illustrate how second messenger ki- verted into many metabolites, some of which alter
nases are regulated: The common mechanism of acti- synaptic transmission and neuronal excitability.
vation is through binding of the second messenger to PKC, another multifunctional enzyme, is activat-
release the catalytic center from inhibition. With PKA, ed by DAG. (Phorbol esters, tumor-promoting sub-
the inactive (inhibited) kinase is activated when the stances from plants, act as potent pharmacological
concentration of cAMP is elevated within the cell ac- analogues of DAG.) There now are fourteen isoforms
cording to the reaction of PKC known. All require the presence of membrane
lipid, notably PS, to be active. The PKCs fall into four
R2C2 + 4 cAMP + 2(R - 2cAMP) + 2C. groups (classical, novel, atypical and -like), whose
enzymatic properties reflect the presence or absence
The two R subunits each bind two molecules of cAMP of various functional domains. For example classical
and dissociate from the C subunits, which, when re- PKCs which require CaM2+-ion for enzymatic activity,
596 SECOND MESSENGER SYSTEMS
have a C2 consensus sequence that is responsible for stream kinase (a MAP kinase kinase, or MAPKK) is a
binding Ca2+. Other PKCs lack the sequence and are distinctive enzyme that phosphorylates the MAP ki-
active independently of Ca2+. The functional signifi- nase at two adjacent, amino acid residues, one a thre-
cance of this variety of isoforms is not yet known, but onine, the other a tyrosine.
structural diversity is presumed to cause differences The activated MAP kinase is imported into the
in substrate specificity and subcellular localization. nucleus, where it phosphorylates (and activates) spe-
Unlike PKA, the regulatory and catalytic regions of cific transcription factors. Thus, even though MAP ki-
the PKCs are both parts of a single polypeptide chain. nases do phosphorylate proteins in the cytoplasm (for
The catalytic part is masked by the regulatory do- example, p38 kinase phosphorylates and activates
main. When a lipid activator and membrane (and PLA2), their most characteristic function appears to
Ca2+ for the classical forms) bind to the regulatory be the second-messenger mediation of changes in
domain of the enzyme, its conformation changes, ex- gene expression.
posing the catalytic part of the kinase for action. The
dependence of the PKC isoforms on Ca2+ is an im-
portant instance of the complexity of regulation by Degradation of Second Messengers
second messengers. In the pathway involving PLC, cAMP and cGMP are rapidly degraded by several
both DAG and IP3 are formed. The function of IP3 as different phosphodiesterases. There are many phosp-
a second messenger is to bind an intracellular recep- hodiesterase inhibitors that prolong the action of
tor that is located on the cytoplasmic surface of the these cyclic nucleotides, including caffeine and theo-
endoplasmic reticulum. When this receptor is activat- phylline, which occur naturally. Some of these degra-
ed, stored Ca2+ is released, thereby raising the intra- dative enzymes can also be secondary effectorsfor
cellular concentration of the free ion. example, in rod cells, where receptor-activated G-
Ca2+ is required for the action of many enzymes proteins activate a cGMP-phosphodiesterase. The
in neurons, either as the free ion or often complexed second messengers derived from membrane pho-
to calmodulin, a small protein that can bind four Ca2+ spholipids are inactivated by being reincorporated
ions. In addition to the major forms of PKC, Ca2+ is into the membrane. IP3 is dephosphorylated by sever-
required by some forms of adenylyl cyclase, guanylyl al phosphatases, one of which is blocked by Li+ ion.
cyclase, PLC, phospholipase A2 (PLA2), 5- (This inhibition may be important for the effective-
lipoxygenase, some protein phosphatases, and one ness of Li+ in the treatment of bipolar depression.)
other kinase, Ca2+/calmodulin-dependent protein ki- Although individual protein substrates phospho-
nase II. This multifunctional kinase, like the PKCs, is rylated by the various protein kinases are discussed
a single polypeptide chain containing both regulatory elsewhere, it is important to point out here that they
and catalytic domains. Unlike either of the other ki- exist only transiently because of the rapidity with
nases, however, it exists in the cell as a complex of sev- which protein phosphatases remove phosphoryl
eral individual kinase molecules, the exact number groups from the proteins. Phosphoryl groups from
varying with the type of cell and with the location proteins phosphorylated at serine/threonine residues
within the neuron. This enzyme is quite abundant but are removed by one class of protein phosphatases,
is highly concentrated in dendritic spines of neurons phosphate on tyrosine residues, by another. With
where it is the most abundant protein in the postsyn- MAP kinases, the phosphoryl groups on both serine/
aptic density. threonine and tyrosine are removed by a specific class
of phosphatase.
MAP Kinases
There are three families of MAP (mitogen activat- Variations of Signal Transduction Pathways
ed protein) kinases: kinases of the ERK type, p38 Although the typical second messenger system
MAP kinases, and JUN kinases. These enzymes are all consists of all of the components described, there are
activated through cascades of kinase kinases, each examples in which one or more of them is absent. In
one specific to the MAP kinase to be activated. The some instances the receptor-activated G-protein itself
first kinase kinase in a cascade typically is activated by acts directly on a secondary regulatorfor example,
the receptor-mediated mobilization of a small (MW in heart muscle cells, where the G-protein modulates
20,00040,000) G-protein (Ras, Raf, Rho, Ran, which an ion channel without any intervening constituents.
are distantly related to the -subunits of the trimeric Some ion channels are gated directly by cyclic nucleo-
G-proteins.) This mobilization, which initiates the sides, thereby circumventing the need for protein
second-messenger pathway, results in the phospho- phosphorylation. Guanylyl cyclase can be activated by
rylation of the first upstream kinase, which then nitric oxide (NO), a short-lived gas that passes
goes on to phosphorylate the next. Finally the last up- through membranes, bypassing a receptor/G-protein
SEMANTIC MEMORY 597
complex. NO is formed in neurons from the amino Wald, G. (1959). Life and light. Scientific American 201, 92108.
acid arginine as a consequence of the activation of the Zang, X., and Majerus, P. W. (1998). Phosphatidylinositol signaling
reactions. Seminars in Cell Development and Biology 9, 153160.
enzyme NO synthetase (which requires reduced nico-
tinamide adenine dinucleotide phosphate and Ca2+ James H. Schwartz
ions) through a second messenger pathway. NO is un-
usual because it does not act as a second messenger
in the neuron in which it is synthesized, but in a
neighboring nerve cell. It is possible that arachidonic
acid and its metabolites also may act transcellularly as SEMANTIC MEMORY
second messengers; in this way they also would bypass [Semantic memory is our acquired knowledge about the ev-
receptor, G-protein, and primary effector enzyme. Fi- eryday world. The following two sections characterize our
nally, some secondary effector enzymes are linked di- current understanding about semantic memory from the cog-
rectly to their own special receptor through the mem- nitive and the neurbiological perspectives. In both sections,
brane; important examples of this kind of signaling two common themes in research on semantic memory are dis-
are the so-called receptor tyrosine protein kinases. cussed: whether semantic memory is fundamentally distinct
from memory for personal experiences (episodic memory) and
See also: APLYSIA: CLASSICAL CONDITIONING AND
OPERANT CONDITIONING; APLYSIA: MOLECULAR
how to characterize the organization of semantic knowledge
BASIS OF LONG-TERM SENSITIZATION; GENETIC in terms of its cognitive dimensions and its representation in
SUBSTRATES OF MEMORY: HIPPOCAMPUS; the brain.
HORMONES AND MEMORY; INVERTEBRATE Semantic memory is often distinguished from episodic
LEARNING; LONG-TERM DEPRESSION IN THE memory as what one knows independent of what one re-
CEREBELLUM, HIPPOCAMPUS, AND NEOCORTEX;
members, that is, consciously recollects learning in a partic-
LONG-TERM POTENTIATION: SIGNAL
ular personal episode. Endel Tulving, who introduced the
TRANSDUCTION MECHANISMS AND EARLY
EVENTS; NEUROTRANSMITTER SYSTEMS AND
distinction between semantic and episodic memory, has ar-
MEMORY gued that semantic memory is a distinct memory system, inde-
pendent from that of episodic memory. Yet, all of our con-
Bibliography scious learning experiences are initially a part of episodic
Caron, M. G., and Lefkowitz, R. J. (1993). Catecholamine recep- memory, consistent with a common alternative view that se-
tors: Structure, function, and regulation. Recent Progress in mantic memory is knowledge abstracted from the information
Hormone Research 48, 277290. common to many episodes. One central question in research
Cobb, M. H. (1999). MAP kinase pathways. Progress in Biophysics and on semantic memory is whether and how episodic memory
Molecular Biology 71, 479500.
Dekker, L. V., Palmer, R. H., and Parker, P. J. (1995). The protein
and semantic memory are mediated by different neural struc-
kinase C related gene families. Current Opinion in Structural Bi- tures or brain systems. Several lines of evidence suggest se-
ology 5, 396402. mantic memory is mediated by cerebral cortical areas without
Greengard, P. (2001). The neurobiology of slow synaptic transmis- involvement of the hippocampus, whereas episodic memory
sion. Science 294, 1,0241,030. relies critically on hippocampal function.
Kennedy, M. B. (2000). Signal-processing machines at the post-
synaptic density. Science 290, 750754. As yet we have only a preliminary understanding of the
Nishizuka, Y. (1995). Protein kinase C and lipid signaling for sus- cognitive dimensions of semantic memory. COGNITIVE
tained cellular responses. FASEB Journal 9, 484496. EFFECTS outlines research suggesting three principal cogni-
Schulman, H., and Hyman, S. E. (1999). Intracellular signaling. In
M. J. Zigmond, F. E. Bloom, S. C. Landis, and L. R. Squire, tive dimensions: categorization of objects by their features,
eds., Fundamental neuroscience. San Diego: Academic Press. judgments about relationships between concepts, and the ca-
Schwartz, J. H. (2001). The many dimensions of cAMP signaling. pacity to combine multiple elementary concepts into distinct
Proceedings of the National Academy of Sciences of the United States and more complex concepts.
of America 98, 13,48313,484.
Siegelbaum, S. A., Schwartz, J. H., and Kandel, E. R. (2000). Modu- It is generally viewed that semantic memory is mediated
lation of synaptic transmission: Second messengers. In E. R. by the cerebral cortex, as best demonstrated in cases of seman-
Kandel, J. H. Schwartz, and T. M. Jessell, eds., Principles of tic dementia, a selective loss of semantic memory, consequent
Neuroscience, 4th edition. New York: McGraw-Hill. to cortical deterioration. However, there is considerable con-
Soderling, T. R. (2000). CaM-kinase: Modulators of synaptic plas-
ticity. Current Opinion in Neurobiology 10, 375380. troversy over whether separate areas of the cerebral cortex
Sutherland, E., and Rall, T. W. (1957). Isolation of cyclic AMP. process specific domains of semantic memory, and about the
Journal of the American Chemical Society 79, 3,608. nature of the functional distinctions. There have now been
Takai, Y., Takuya, S., and Takahashi, M. (2001). Small GTP bind- several observations of deficits in specific domains of seman-
ing proteins. Physiology Review 81, 153208. tic knowledge following damage to particular areas of the
Taylor, S. S., Buechler, J. A., and Yonemoto, W. (1990). cAMP
dependent protein kinase: Framework for a diverse family of cerebral cortex, as well as observations of activation of par-
regulatory enzymes. Annual Review of Biochemistry 59, 971 ticular cortical areas during naming in specific domains of
1,005. objects. It is not clear whether these selectivities reflect specif-
598 SEMANTIC MEMORY: Cognitive Effects
ic categories of semantic knowledge or distinct types of cogni- are good examples of a category are more readily ver-
tive or perceptual processing associated with particular cate- ified as members than are poor examples. For exam-
gories of information. NEUROBIOLOGICAL PERSPECTIVE ple, people decide that Robins are birds more rap-
outlines this issue.] idly than Chickens are birds.
The typicality effect has its counterpart for false
statements (the relatedness effect: Smith, Shoben,
COGNITIVE EFFECTS and Rips, 1974). These are easier to disconfirm if sub-
ject and predicate are unrelated. For example, A
Researchers originally conceptualized semantic
goose is a mammal is more difficult to disconfirm
memory as human knowledge of language (Tulving,
than A goose is a tool.
1972). The term now generally refers to ones every-
day knowledge of the world, as contrasted with epi- There are many proposed explanations of the
sodic memory, ones knowledge of personal experi- typicality effect. It has been ascribed to computing
ence. Semantic memory includes such facts as, similarity to a prototype or to multiple members of a
Robins are birds, Chairs have legs, and Fire- category, and to feature overlap between an item and
works are dangerous. Semantic memory has been a category description. Other factors such as frequen-
thought by some researchers (Tulving and Schacter, cy (Barsalou, 1985; Heit and Barsalou, 1996) appear
1990) to be a functionally separate memory system, to moderate typicality, although twenty-first-century
distinct from episodic memory. Neuroimaging tech- evidence does not strongly support an earlier claim
niques implicate differential brain region activity in that familiarity is one of these (Hampton, 1997).
semantic and episodic tasks (Wheeler, Suss, and Tulv- While typicality is often discussed in terms of similari-
ing, 1997; Knowlton, 1998). There is some suggestion ty of semantic memory representations, there is a
in scalp-based recordings of electrical brain activity growing acknowledgment among the scientific com-
following stimulus presentation (event related poten- munity that the notion of similarity is itself a complex
tials) that a wave of negative amplitude may accompa- construct requiring some basis for comparison
ny. Nevertheless, the interdependence of semantic (Medin, Goldstone, and Gentner, 1993). For exam-
and episodic information (McKoon, Ratcliff, and ple, similarity may sometimes involve abstract deci-
Dell, 1986; Shoben and Ross, 1986) urges caution in sions based on common goals or instrumentalities.
accepting the claim that these are separate systems. Typical items to bring to school (books, a calculator,
Indeed, the best-known computational model of a pen) may be similar by virtue of accomplishing re-
memory continues to treat the two as part of the same lated goals, but will be unlikely to share much physical
(declarative) memory system (Anderson, and Lebiere, resemblance.
1998). Context effects are common in semantic memory
Semantic and episodic recall may be distin- research. For example, early findings indicated that
guished by whether one believes one knows some- the size of the typicality effect varies as a function of
thing, or whether one personally remembers it the proportion of related false statements among the
(Knowlton and Squire, 1995). Mark Wheeler and col- test stimuli (McCloskey and Glucksberg, 1979). Other
leagues (1997) proposed that semantic memory is as- studies have shown that the same categories may ex-
sociated with the absence of the phenomenological hibit different typicality and similarity relationships
experience of remembering, a distinction that may in different contexts. Thus, what is a typical bird in
prove more relevant than one based on content. southern Florida may depend in part upon whether
Thus, semantic memory may include autobiographi- the informant is Hispanic (Schwanenflugel and Rey,
cal content that is known, but not remembered. 1986). Similarly, North Americans will alter their
judgments of the typicality of various birds when
Research in semantic memory can be divided into
asked to take the perspective of a South American
three categories concerning human knowledge of
(Barsalou, 1987). These effects extend to similarity.
concepts: First, one knows that concepts belong to
Black and grey will be more similar than grey and
various categories. Second, one knows that concepts
white in the context of cloud color, but not in the con-
have certain properties and bear certain relations to
text of hair color (Medin and Shoben, 1988).
other concepts. Third, one knows that concepts can
be combined with other concepts. Other effects are more controversial (see Chang,
1986, for a review). For example, it has often been ar-
gued that there is a category size effect in semantic
Categorical Knowledge memory in which it is harder to confirm membership
Categorical knowledge is the most studied area in in a large category than it is in a smaller one. Howev-
semantic memory. One of the robust findings in this er, reversals of this finding do occur. The finding may
area is the typicality effect (Rosch, 1975). Items that in part be confounded with the fact that larger catego-
SEMANTIC MEMORY: Cognitive Effects 599
ries tend to be at a higher level of abstraction than further apart on some dimension are more readily
smaller categories (for example, tool and imple- discriminated than objects that are closer together
ment). (Moyer, 1973). For example, it is easier to determine
that Desks are larger than strawberries than Desks
are larger than dogs.
Theoretical Concerns
A second finding, the congruity effect, is that
The question of how categorical information is judgments are easier when the question matches the
represented in memory remains unresolved. Some magnitude of the objects, small or large (Banks,
early theorists suggested that semantic memory could 1977). If one asks which is larger, a persons response
be viewed as a hierarchical tree or semantic network will be faster with large than with small things. But if
(Collins and Quillian, 1969; Holyoak and Glass, one asks which is smaller, a persons response will be
1975) in which concepts such as robin and canary faster with small things.
would be represented at the bottom of the tree with
links from both extending up to bird. Others es- A third finding is the bowed serial position effect:
poused a componential approach in which concepts Holding symbolic distance constant, objects of inter-
were represented as sets of properties or semantic mediate magnitude are more difficult to discriminate
features. Semantic decisions were made by a compari- than objects of extreme magnitude (Shoben, Cech,
son of these features or properties (Smith et al., 1974; Schwanenflugel, and Sailor, 1989). For example, it is
McCloskey and Glucksberg, 1979). Yet other repre- more difficult for a person to select the larger of
sentations and mechanisms have been proposed to wolf and pig than to select the larger of either
account for categorical knowledge. A partial listing of toad and snail or of bull and elephant.
representations includes prototypes, images, exem- Like decisions about categorical information,
plars, schemas, situational models, and hyperdimen- judgments of relative magnitude are subject to con-
sional vectors that represent semantic similarity by text effects. Whether pairs such as rabbit-beaver are
tracking the contexts in which different concepts may discriminated more readily under the instruction to
be found. None of these approaches by itself is likely select the smaller or the larger item will depend on
to account for all of the available data, and different the presence of large pairs. If there are no larger
semantic units subject to different processes may be pairs, then rabbit and beaver suddenly behave as
required (Brewer, 1993). For example, production if they are large items (Cech and Shoben, 2001).
frequency (the frequency with which an item is gener- Moreover, in appropriate contexts, multiple congru-
ated as an example of a category) and typicality rat- ity effects may ensue, each corresponding to a differ-
ings appear to provide different contributions to clas- ent group of items.
sification times. The suggestion has been made that
the former relies on retrieval of network links whereas There are also effects of categorization difficulty.
the latter involves a similarity computation operating Some items are difficult to classify as large or small,
over features (Hampton, 1997). Sophisticated tech- and others are relatively easy. In general, items of ex-
niques suggest that retrieval of information occurs treme magnitude are more readily classified than
continuously in time as opposed to in discrete quanta items of intermediate magnitude. This finding helps
(Kounios, 1997), but whether such results are incon- explain the serial position effect. When items across
sistent with network models, as researchers some- the magnitude scale are equated for categorization
times claim, will be sensitive to the specific structural difficulty, no bowed serial position effect occurs (Sho-
and processing assumptions adopted by such models. ben and Wilson, 1998).
Finally, there are reference point effects. If peo-
ple are asked to determine which of two objects is
Judgments of Relative Magnitude closer to a third, the task is harder if the items are far
Although there have been some direct investiga- from the reference point (Holyoak, 1978). Thus, for
tions of how people process information about prop- example, it is easier to determine whether 4 or 5 is
erties of concepts, most of the work on this question closer to 3 than whether 6 or 7 is closer to 3.
has concerned how people make judgments of rela-
tive magnitude such as, Are rabbits larger than
mice? Theories of Relational Information
Many theories have been proposed to account for
relational judgments, although most have problems
Empirical Findings explaining at least some of the findings. Priming or
The oldest and most robust finding in the litera- expectancy-based theories claim that asking which is
ture is the symbolic distance effect. Objects that are larger or smaller will bias people toward the specified
600 SEMANTIC MEMORY: Cognitive Effects
size. However, there appear to be congruity effects numerical comparisons does hold out the promise of
that are not due to such priming; moreover, these helping scientists determine how rich semantic repre-
models wrongly predict that congruity should occur sentations are.
early in processing, rather than late (Cech, 1995).
Traditionally, theories fall into two broad camps:
those like the discrete code model (Banks, 1977), Conceptual Combinations
which posit comparison of discrete, propositional in- A problem in categorization concerns how con-
formation, and analogical models like the reference cepts are combined. Some of these combinations ap-
point model (Holyoak, 1978), which posit direct com- pear superficially simple, while others are clearly
parison of stored analogical magnitude information. more complex. For example, red ball can readily be
Models used by twenty-first-century researchers in- paraphrased as a ball that is red, but malarial mos-
clude the recoding model (Cech and Shoben, 2001), quitoes must instead be paraphrased as mosquitoes
evidence accrual models (Petrusic, 1992; Birnbaum that cause malaria. Red ball uses a predicating ad-
and Jou, 1990), and the connectionist model of Leth- jective, red, in which a property of the modifier is
Steensen and Marley (2002). ascribed to the head noun, while malarial mosqui-
The recoding model claims that people use range toes employs the nonpredicating adjective malari-
and co-occurrence information to cluster items into al. Phrases involving nonpredicating adjectives are
groups. People assign relative, contrastive magnitude sometimes more difficult to comprehend (Murphy,
1990), although faster comprehension times have
codes to items within a group (this is large, that is
also been reported for nonpredicating nominal com-
small) to facilitate comparison. Comparison thus
pounds (Gagn, 2000). In nonpredicating com-
necessarily relies on attentional and categorization
pounds, the interpretation appears to rely on estab-
processes. However, in evidence accrual models, item
lishing some relationship between the two
magnitudes are sampled until there is sufficient evi-
components (a relational interpretation).
dence regarding relative size. Finally, the Leth-
Steensen and Marley model also includes an evidence Most research on conceptual combinations has
accrual component, but enables responding on the used predicating adjectives. One fundamental ques-
basis of positional information (whether an item is tion is how people determine membership in com-
fourth or third largest, for example). This latter bined categories. For example, A cardinal is a red
model employs learned associations and effectively bird is clearly true, because a cardinal is clearly both
combines aspects of the other approaches. William a red thing and a bird. At the same time, A female
Petrusic has suggested that models under which peo- cardinal is a red bird is only partly true, because fe-
ple operate may differ: Some people seem to act in male cardinals are only somewhat red. Such concerns
accord with the recoding model, and others seem to might lead researchers to propose a min rule: An ex-
use evidence accrual. emplar is a member of a combined category only to
the minimum degree that it is a member of the cate-
A domain of particular interest concerns numero-
gory defined by the adjective and the category speci-
sity judgments. Researchers in this area study the rel-
fied by the noun. Despite the intuitive appeal of such
ative merits of the single-format and the multiple-
a rule, it cannot handle many cases: Guppies are per-
format assumptions (Blankenberger and Vorberg,
haps the paradigmatic example of the category pet
1997). The single-format assumption asserts that dif-
fish, but they are good examples of neither pets nor
ferent number representations (4, four, IV)
fish. Empirical judgments of subjects show clear viola-
project onto the same semantic representations,
tions of the min rule (Smith and Osherson, 1984).
whereas the multiformat assumption claims multiple
semantic representations for numbers (Dehaene, In natural language corpora, conceptual combi-
1997). Some numerical comparison studies examine nations that involve an initial predicating noun are
these positions by focusing on the brain regions active less frequent than combinations that require a rela-
during the comparison of different number types, tional interpretation, although there is disagreement
and others study how format influences the distance about the exact disproportion in the two types. Most
effect. For example, is the average time required to contemporary work in conceptual combination cen-
determine the larger of the pairs two-four and ters around the dual-process model (Wisniewski,
2-4 the same as the time required to compare 1997) and the CARIN model (Gagn and Shoben,
two-4 and 2-four? The answer appears to be no: 1997). In the dual-process model, relational interpre-
Different distance effects are found with mixed- tations require a process of scenario construction in
format numbers (Vorberg and Blankenberger, 1993). which specified roles may be assigned the two con-
Although it is premature to favor the multiple-format cepts. Property interpretations, in contrast, arise
assumption from results such as these, the domain of from a process of alignment in which alignable differ-
SEMANTIC MEMORY: Cognitive Effects 601
ences (differing values on a common dimension) help See also: CONCEPTS AND CATEGORIES, LEARNING
drive the decision regarding which property to move OF; DRUGS AND MEMORY; SEMANTIC MEMORY:
to the complex conceptual schema of the head noun. NEUROBIOLOGICAL PERSPECTIVE
An alignable difference between dog and man,
Bibliography
for example, is that each locomotes using legs, al-
Anderson, J. R., and Lebiere, C. (1998). The atomic components of
though a dog moves on four legs. Thus, a potential thought. Mahwah, NJ: Erlbaum.
interpretation of dog man is a man who travels on Banks, W. P. (1977). Encoding and processing of symbolic informa-
all fours, whereas a potential interpretation of man tion in comparative judgments. In G. H. Bower, ed., The psy-
dog is a dog walking upright. By way of contrast, a chology of learning and motivation, Vol. 11. New York: Academic
Press.
relational interpretation of dog man might include
Barsalou, L. W. (1985). Ideals, central tendency, and frequency of
some scenario in which an individual prefers dogs to instantiation as determinants of graded structure in catego-
cats. This model has recently been extended to in- ries. Journal of Experimental Psychology: Learning, Memory, and
clude nonalignable differences that display spatial Cognition 11 (4), 629654.
correspondences (Wisniewski and Middleton, 2002): (1987). The instability of graded structure: Implications for
the nature of concepts. In U. Neisser, ed., Concepts and concep-
Kangaroos and dogs do not have a readily retrievable tual development: Ecological and intellectual factors in categoriza-
common dimension on which pouch can serve as a tion. Cambridge, UK: Cambridge University Press.
contrasting value that helps distinguish the two, but Birnbaum, M. H., and Jou, J. (1990). A theory of comparative re-
the presence of equivalent spatial areas on dogs and sponse times and difference judgments. Cognitive Psychology
22 (2), 184210.
kangaroos may allow the interpretation that a kan-
Blankenberger, S., and Vorberg, D. (1997). The single-format as-
garoo dog is a dog with a pouch on the lower front sumption in arithmetic fact retrieval. Journal of Experimental
of its body. Psychology: Learning, Memory, and Cognition 23 (3), 721738.
Brewer, W. F. (1993). What are concepts? Issues of representation
CARIN does not rely on an alignment process, or and ontology. In G. V. Nakamura, R. M. Taraban, and D. L.
a privileged attempt to modify the representation of Medin, eds., The psychology of learning and motivation, Vol. 29:
the head noun. Instead, the model posits that the rep- Categorization by humans and machines. San Diego, CA: Academ-
ic Press.
ertoire of relations associated with the modifier will Cech, C. G. (1995). Is congruity due to encoding? Journal of Experi-
guide finding an appropriate interpretation. For ex- mental Psychology: Learning, Memory, and Cognition 21 (5),
ample, because mountain has a high frequency of 1,2751,288.
relations involving location, mountain man will Cech, C. G., and Shoben, E. J. (2001). Categorization processes in
mental comparisons. Journal of Experimental Psychology: Learn-
likely be interpreted as a man who lives on a moun-
ing, Memory, and Cognition 27 (3), 800816.
tain. Accordingly, the relative rarity of property inter- Chang, T. M. (1986). Semantic memory: Facts and models. Psycho-
pretations ought to result in long interpretation logical Bulletin 99 (2), 199220.
times. Collins, A., and Quillian, M. R. (1969). Retrieval time from seman-
tic memory. Journal of Verbal Learning and Verbal Behavior 8,
The interpretation process for conceptual combi- 240247.
nations has also been modeled as a constraint satisfac- Costello, F. J., and Keane, M. T. (2000). Efficient creativity: Con-
straint-guided conceptual combination. Cognitive Science 24
tion process that requires simultaneous solutions to (2), 299349.
the constraints of diagnosticity, plausibility, and in- Dehaene, S. (1997). The number sense: How the mind creates mathemat-
formativeness (Costello and Keane, 2000). Moreover, ics. New York: Oxford University Press.
like categorization and magnitude comparisons, con- Gagn, C. L. (2000). Relation-based combinations versus property-
based combinations: A test of the CARIN theory and the dual-
ceptual combinations are also subject to context ef-
process theory of conceptual combination. Journal of Memory
fects (Gerrig and Bortfeld, 1999). The work on con- and Language 42 (3), 365389.
ceptual combinations illustrates the reliance on a Gagn, C. L., and Shoben, E. J. (1997). Influence of thematic rela-
much richer and more complex semantic memory tions on the comprehension of modifier-noun combinations.
than researchers assumed in the early work on net- Journal of Experimental Psychology: Learning, Memory, and Cogni-
tion 23 (1), 7187.
work and feature models. Gerrig, R. J., and Bortfeld, H. (1999). Sense creation in and out of
discourse contexts. Journal of Memory and Language 41 (4),
457468.
Broader Issues Hampton, J. A. (1997). Associative and similarity-based processes
in categorization decisions. Memory & Cognition 25 (5), 625
This review has necessarily been selective, with 640.
many topics omitted. In closing, however, it should be Heit, E., and Barsalou, L. W. (1996). The instantiation principle in
noted that ones knowledge of the world will influence natural categories. Memory 4 (4), 413451.
most of the cognitive things that one does. Solving a Holyoak, K. J. (1978). Comparative judgments with numerical ref-
erence points. Cognitive Psychology 10 (2), 203243.
problem, following directions, or simply reading in- Holyoak, K. J., and Glass, A. L. (1975). The role of contradictions
volves ones semantic memory. Semantic memory is and counterexamples in the rejection of false sentences. Jour-
fundamental to cognition. nal of Verbal Learning and Verbal Behavior 14 (2), 215239.
602 SEMANTIC MEMORY: Neurobiological Perspective
Knowlton, B. J. (1998). The relationship between remembering Wisniewski, E. J., and Middleton, E. L. (2002). Of bucket bowls and
and knowing: A cognitive neuroscience perspective. Acta Psy- coffee cup bowls: Spatial alignment in conceptual combina-
chologica 98 (23), 253265. tion. Journal of Memory and Language 46 (1), 123.
Knowlton, B. J., and Squire, L. R. (1995). Remembering and know- Edward J. Shoben
ing: Two different expressions of declarative memory. Journal Revised by Claude G. Cech
of Experimental Psychology: Learning, Memory, and Cognition 21
(3), 699710.
Kounios, J. (1996). On the continuity of thought and the represen-
tation of knowledge: Electrophysiological and behavioral NEUROBIOLOGICAL PERSPECTIVE
time-course measures reveal levels of structure in semantic
memory. Psychonomic Bulletin & Review 3 (3), 265286.
Consider the following questions: Do acorns have
Leth-Steensen, C., and Marley, A. A. J. (2000). A model of response stems? Do you tune a guitar by tightening and loosen-
time effects in symbolic comparison. Psychological Review 107 ing the strings? Do deer have white noses? If you were
(1), 62100. able to answer these questions correctly (yes, yes, no),
McCloskey, M., and Glucksberg, S. (1979). Decision processes in then you were using semantic memory. Much of se-
verifying category membership statements: Implications for mantic memory relates to the brain systems that en-
models of semantic memory. Cognitive Psychology 11 (1), 137.
able us to store and retrieve semantic knowledge.
McKoon, G., Ratcliff, R., and Dell, G. S. (1986). A critical evalua-
tion of the semantic-episodic distinction. Journal of Experimen-
This article explores the neurobiology of this process.
tal Psychology: Learning, Memory, and Cognition 12 (2), 295306.
Medin, D. L., Goldstone, R. L., and Gentner, D. (1993). Respects Multiple Long-Term Memory Systems
for similarity. Psychological Review 100 (2), 278.
Medin, D. L., and Shoben, E. J. (1988). Context and structure in When you remember that acorns have stems, you
conceptual combination. Cognitive Psychology 20 (2), 158190. are using your semantic memory. When you remem-
Moyer, R. S. (1973). Comparing objects in memory: Evidence sug- ber the time you gathered acorns with your father in
gesting an internal psychophysics. Perception and Psychophysics fifth grade, you are using your episodic memory. In
13 (2), 180184.
both cases you are retrieving information from long-
Murphy, G. L. (1990). Noun phrase interpretation and conceptual
combination. Journal of Memory and Language 29 (3), 259288. term memory. Are there, in fact, two separate memo-
Petrusic, W. M. (1992). Semantic congruity effects and theories of ry systems for these two types of memories?
the comparison process. Journal of Experimental Psychology: In 1972, the psychologist Endel Tulving first ar-
Human Perception and Performance 18 (4), 962986. gued that there are two separate memory systems.
Rosch, E. H. (1975). Cognitive representations of semantic catego-
Tulving distinguished between memories that have
ries. Journal of Experimental Psychology: General 104 (3), 192
233.
an autobiographical reference (i.e., episodic memo-
Schwanenflugel, P. J., and Rey, M. (1986). The relationship be- ries) and memories that do not (i.e., semantic memo-
tween category typicality and concept familiarity: Evidence ries). Cognitive psychologists have disagreed about
from Spanish- and English-speaking monolinguals. Memory & whether episodic and semantic memories reflect two
Cognition 14 (2), 150163. distinct systems or whether they are part of a common
Shoben, E. J., Cech, C. G., Schwanenflugel, P. J., and Sailor, K. M. long-term memory system. However, neuropsy-
(1989). Serial position effects in comparative judgments. Jour-
nal of Experimental Psychology: Human Perception and Perfor-
chologists have provided compelling evidence that
mance 15 (2), 273286. there are two distinct systems: one for our knowledge
Shoben, E. J., and Ross, B. H. (1986). The crucial role of dissocia- of events and one for our knowledge of facts. This evi-
tions. Behavioral and Brain Sciences 9 (3), 568571. dence comes from patients who have selective impair-
Shoben, E. J., and Wilson, T. L. (1998). Categorization in judgments of semantic memory.
ments of relative magnitude. Journal of Memory and Language
Impairments if semantic memory correlate with
38 (1), 94111.
Smith, E. E., and Osherson, D. N. (1984). Conceptual combination several etiologies of brain damage. The most com-
with prototype concepts. Cognitive Science 12 (4), 485527. mon cause of semantic-memory impairments is Al-
Smith, E. E., Shoben, E. J., and Rips, L. J. (1974). Structure and zheimers disease, a progressive dementia that causes
process in semantic memory: A featural model for semantic widespread cortical degeneration. Semantic memory
decisions. Psychological Review 81 (3), 214241. impairments also attend acute illnesses such as stroke
Tulving, E. (1972). Episodic and semantic memory. In E. Tulving
or infection. These conditions, however, commonly
and W. Donaldson, eds., Organization of memory. New York:
Academic Press.
involve both episodic and semantic memory impair-
Tulving, E., and Schacter, D. L. (1990). Priming and human mem- ments. Yet there is one condition, semantic dementia,
ory systems. Science 247, 301306. in which patients exhibit a progressive loss of seman-
Vorberg, D., and Blankenberger, S. (1993). Mentale reprsenta- tic memory (Hodges, Patterson, Oxbury, and Fun-
tion von zahlen. Sprache and Kognition 12 (2), 98114. nell, 1992). By definition, patients with semantic de-
Wheeler, M. A., Suss, D. T., and Tulving, E. (1997). Toward a theo- mentia have an impairment in semantic memory but
ry of episodic memory: The frontal lobes and autonoetic con-
normal episodic memory.
sciousness. Psychological Bulletin 121 (3), 331354.
Wisniewski, E. J. (1997). When concepts combine. Psychonomic Bul- One of the earliest symptoms of semantic demen-
letin & Review 4 (2), 167183. tia is anomia, an impairment in the ability to find
SEMANTIC MEMORY: Neurobiological Perspective 603
words when speaking or writing. Patients with this brain of a healthy subject who is performing a cogni-
disorder also have problems with semantic memory tive task (e.g., functional magnetic-resonance imag-
tasks that do not require language at all. The pyra- ing). Several neuroimaging studies have tested the
mids and palm trees test is a common test of nonver- existence of a unitary semantic store. In one such
bal knowledge that requires the patient to identify se- study (Vandenberghe et al., 1996), brain imaging was
mantic relationships between pictures (Howard and conducted during either a verbal semantic task (word
Patterson, 1992). For example, the patient might be matching) or a visual semantic task (similar to the pyr-
asked to indicate which picture belongs with a pyra- amids and palm trees test). The activation of many
mid: a palm tree or a fir tree. This task does not re- common regions of the brain in these tests is consis-
quire language comprehension or production (i.e., tent with the theory that there is a unitary semantic
the patient is simply pointing at pairs of pictures), but memory store for both verbal and nonverbal retrieval.
it does require semantic memory. Patients with se-
mantic dementia are unable to answer simple ques- Categories of Semantic Memory
tions of this type. The selective loss of semantic mem-
ory in semantic dementia may indicate idea that For decades cognitive psychologists have been
semantic memory and episodic memory are two dis- debating the two issues discussed above: whether se-
tinct long term memory systems. mantic memory is distinct from episodic memory and
whether there is a unitary semantic store. The evi-
In all of the etiologies of brain damage associated dence reviewed above simply fuels the controversy
with semantic memory loss, there is widespread and with evidence from neurobiology. There is, however,
often bilateral damage to the temporal lobes. Seman- another question about semantic memory that origi-
tic dementia may result from a variant of Alzheimers nated with neuropsychologists: Are there separate
disease in which degeneration is initially restricted to stores of semantic memory for different taxonomic
the temporal lobes. Semantic dementia may also re- categories, such as animals and tools?
sult from Picks disease, another degenerative disease
that is associated with temporal lobe atrophy. Picks This might seem like a strange hypothesis, but
disease does not affect the medial temporal lobes, in- the evidence for it is very compelling. Imagine ad-
cluding the hippocampus, which is a part of the brain ministering a simple picture-naming task to a brain-
associated with episodic memory. damaged patient. You show the patient pictures of all
sorts of objects, such as a trellis, a compass, and an
abacus, and the patient can name them with little dif-
Separate or Unitary Semantic Stores ficulty. However, when you show the patient a picture
of something as simple as a duck or a bee, the patient
When Tulving first defined semantic memory, he
struggles and fails to come up with the name. This
described it as a mental thesaurus that supported
very pattern of behavior has been reported in dozens
language use. Since that time, many investigators of
of patients, most of them afflicted with herpes sim-
semantic memory have broadened their definitions of
plex encephalitis (a brain infection), who have seman-
semantic memory to include both verbal and nonver-
tic memory impairments restricted to categories of
bal knowledge. Clearly, semantic memory can be ac-
natural items (Warrington and Shallice, 1984).
cessed from both verbal labels and visual depictions.
Many psychologists, however, have questioned One hypothesis for this peculiar pattern of im-
whether there is a unitary semantic memory store or pairments is that natural items (e.g., plants and ani-
whether there are separate stores for verbal and non- mals) are more difficult to identify because they are
verbal memory. less familiar or more visually complex (Funnell and
Sheridan, 1992). Carefully controlled studies have
We have already seen one piece of evidence for
largely ruled out this supposition, however (Farah,
a unitary semantic memory store: As mentioned
McMullen, and Meyer, 1991). There have been a few
above, patients with semantic dementia have impair-
cases of the reverse pattern: patients with impaired
ments of both verbal and nonverbal tests of semantic
knowledge of man-made objects but with relatively
knowledge. Curiously, a few patients have deficits that
well-preserved knowledge of natural objects (War-
seem restricted to a single test modality (e.g., Beau-
rington and McCarthy, 1987). Although rare, this dis-
vois, 1982). But it is unclear whether this pattern re-
order provides stronger evidence for the hypothesis
flects a true loss of semantic memory or merely an im-
that there are separate memory stores for natural ob-
pairment in accessing semantic memory from one
jects and human-made objects. Some researchers
kind of input.
(Caramazza and Shelton, 1998) have argued that sep-
Another method for investigating questions arate stores have developed for categories for which
about semantic memory is to noninvasively measure there is some evolutionary significance (e.g., plants,
regional changes in metabolism or blood flow in the animals, and tools).
604 SEMANTIC MEMORY: Neurobiological Perspective
Domains of Semantic Memory There are other sources of evidence that semantic
memory is organized according to domain. Recall
In the preceding section, the evidence for sepa-
that patients with semantic dementia have global defi-
rate semantic memory stores for different categories
cits in semantic memory when tested either verbally
seemed clear-cut. However, many investigators be-
or visually. However, there is one domain in which
lieve that there is another, more plausible, interpreta-
these patients show preserved semantic knowledge:
tion of these category-specific deficits. The following
Patients with semantic dementia often continue to
two questions illustrate an important difference be-
demonstrate normal use of objects long after they fail
tween our knowledge of natural objects and our
to identify these objects correctly on other tests of se-
knowledge of human-made objects: How would you
mantic memory. This pattern of behavior stands in
define tiger so that someone could distinguish it from
stark contrast to another disorder, ideational apraxia,
similar concepts (e.g., leopards, lions, and bobcats)?
an impairment in object use in patients who test nor-
You would probably mention something about the mally for traditional semantic memory (Ochipa,
color and size of tigers. In contrast, how would you Rothi, and Heilman, 1989). Thus, there may be a se-
define a clock (as distinct from a radio or a lamp)? In mantic memory store for object use that is distinct
this case, you are probably less likely to talk about from other domains of semantic memory.
color or size or even shape, because clocks come in all
varieties. Instead, you would probably talk about the
function of clocks. Hence the defining attributes of Conclusion
natural items are more likely to be visual than those Neurobiological studies of semantic memory
of human-made items. have addressed some fundamental questions of se-
This distinction leads to another hypothesis mantic memory and have raised a few new questions
as well. Disorders that selectively impair semantic
about a distinction within semantic memory: there
memory provide evidence that semantic memory is a
may be different memory stores for different kinds,
memory system that is distinct from episodic memo-
or domains, of semantic knowledge. Knowledge of
ry. Neuroimaging studies have indicated that verbal
any concept may be distributed across a variety of dis-
and nonverbal semantic retrieval depends on a com-
tinct memory stores that are tied to sensorimotor sys-
mon memory store. Within this semantic store, how-
tems (Allport, 1985). According to this hypothesis,
ever, there appear to be distinctions based either on
patients who appear to have a deficit pertaining to
taxonomic category or type of knowledge. Neuroi-
natural objects in fact have a visual-knowledge deficit
maging and neuropsychological studies support the
that is most evident when tested on items defined by
theory that semantic memory is distributed across dif-
visual attributes (i.e., natural objects). The difficulty
ferent sensorimotor systems, such as object appear-
of distinguishing absolutely between object category ance and object use.
and knowledge domain makes it is hard to experi-
mentally distinguish between these two hypotheses. See also: EPISODIC MEMORY; SEMANTIC MEMORY:
Some investigators have pointed to interesting excep- COGNITIVE ASPECTS
tions to the natural/man-made distinction that are
more consistent with a visual/functional dichotomy. Bibliography
For example, patients described as having a deficit Allport, D. A. (1985). Distributed memory, modular subsystems
pertaining to natural objects may also have impair- and dysphasia. In S. K. Newman and R. Epstein, eds., Current
ments with musical instruments, which are largely de- perspectives in dysphasia. Edinburgh: Churchill Livingstone.
Beauvois, M. F. (1982). Optic aphasia: A process of interaction be-
fined by visual attributes (Warrington and Shallice, tween vision and language. Philosophical Transactions of the
1984). Royal Society of London, 298, 3547.
Caramazza, A., and Shelton, J. R. (1998). Domain-specific knowl-
Neuroimaging studies have also been used to test edge systems in the brain the animate-inanimate distinction.
this hypothesis (e.g., Thompson-Schill, Aguirre, Journal of Cognitive Neuroscience 10, 134.
DEsposito, and Farah, 1999). For example, when Farah, M. J., McMullen, P. A., and Meyer, M. M. (1991). Can rec-
subjects are asked to name natural objects (e.g., ani- ognition of living things be selectively impaired? Neuropsy-
chologia, 29, 185193.
mals), brain activation is observed in areas associated Funnell, E., and Sheridan, J. (1992). Categories of knowledge? Un-
with vision; in contrast, when subjects are asked to familiar aspects of living and nonliving things. Cognitive
name human-made objects (e.g., tools), brain activa- Neuropsychology 9, 135153.
tion is observed in areas concerned with motor con- Hodges, J. R., Patterson, K., Oxbury, S., and Funnell, E. (1992). Se-
trol (Martin et al., 1995). Results of this sort have mantic dementia. Progressive fluent aphasia with temporal
lobe atrophy. Brain 115, 1,7831,806.
been taken as evidence that the fundamental distinc- Howard, D., and Patterson, K. (1992). Pyraminds and palm trees: A
tion within semantic memory is one of domain of test of semantic access from pictures and words. Bury St. Edmunds:
knowledge and not taxonomic category. Thames Valley Test Company.
SEMON, RICHARD 605
Martin, A., Haxby, J. V., Lalonde, F. M., Wiggs, C. L., and Ungerl-
eider, L. G. (1995). Discrete cortical regions associated with
knowledge of color and knowledge of action. Science 270, 102
105.
Ochipa, C., Rothi, L. J., and Heilman, K. M. (1989). Ideational
apraxia: A deficit in tool selection and use. Annals of Neurology
25, 190193.
ThompsonSchill, S. L., Aguirre, G. K., DEsposito, M., and Farah,
M. J. (1999). A neural basis for category and modality specific-
ity of semantic knowledge. Neuropsychologia 37, 671676.
Tulving, E. (1972). Episodic and semantic memory. In E. Tulving
and W. Donaldson, eds., Organization of memory. New York:
Academic Press.
Vandenberghe, R., Price, C., Wise, R., Josephs, O., and Frac-
kowiak, R. S. (1996). Functional anatomy of a common se-
mantic system for words and pictures. Nature 383, 254256.
Warrington, E. K., and McCarthy, R. A. (1987). Categories of
knowledge. Further fractionations and an attempted integra-
tion. Brain 110, 1,2731,296.
Warrington, E. K., and Shallice, T. (1984). Category specific se-
mantic impairments. Brain 107, 829854.
Edward J. Shoben
Revised by Sharon L. Thompson-Schill
SEMON, RICHARD (18591918)

Richard Wolfgang Semon is a relatively unknown but
nevertheless important figure in the history of re-
search on learning and memory. Although little no-
ticed by both his contemporaries and memory re-
searchers today, Semon anticipated numerous Richard Semon (Source unknown)
modern theories and, perhaps ironically, created one
of the best known terms in the memory literature, en-
gram.
Semon was born in Berlin on August 22, 1859. a major expedition to Australia in search of the miss-
His father, Simon, was a stockbroker, and the Semon ing link. The expedition was responsible for the dis-
family became part of the upper echelon of Berlin covery of 207 new species and twenty-four new
Jewish society during Richards childhood. Simons genera. When he returned from Australia in 1893,
severe losses in the stock market crash of 1873 im- Semon continued his research at Jena until 1897,
posed a much humbler lifestyle on the family. when his life changed dramatically. He became in-
Semons older brother, Felix, left Germany after re- volved with Maria Krehl, then wife of an eminent pro-
ceiving a medical degree and practiced in England, fessor of pathology at Jena, Ludolph Krehl. The ensu-
where he became a historic pioneer of clinical and sci- ing scandal in Jena led to Semons resignation. He
entific laryngology.
and Maria moved to Munich, where he began work-
As a child, Richard Semon expressed a strong ining as a private scholar, and the pair eventually mar-
terest in biology and zoology. He attended the Uni- ried.
versity of Jenaa major European center of biologi-
cal researchand received a doctorate in zoological Semon wrote two major books on memory during
studies in 1883 and a medical degree in 1886. While the next twenty years: Die mneme (1904), translated as
at Jena, Semon was influenced heavily by the famous The Mneme (1921), and Die mnemischen Empfindungen
evolutionary biologist Ernst Haeckel, whose monistic (1909), translated as Mnemic Psychology (1923). His
philosophy stressed the importance of attempting to work attracted little attention, and Semons acute
unify diverse biological phenomena within a single dismay about his lack of recognition is evident in
set of theoretical principles. letters written to his colleague and ally, the Swiss
Semons career as an evolutionary biologist devel- psychiatrist August Forel (Schacter, 1982). Depressed
oped rapidly during the 1890s. Shortly after assum- over the neglect of his work, troubled by Germanys
ing an associate professorship at Jena in 1891, he led role in World War I, and shattered by his wifes
606 SEMON, RICHARD
death from cancer, Semon took his own life on De- engraphy (Schacter, 2001; Schacter, Eich, and Tulv-
cember 27, 1918. ing, 1978). By contrast, Semon developed a detailed
theory of ecphoric processes and argued that succss-
ful ecphory requires that the conditions prevailing at
Theory of Memory the time of engraphy (i.e., encoding) are partially re-
A full appreciation of Semons ideas about human instated at the time of ecphory. He laid great empha-
memory requires a look at the biological context from sis on this latter idea, elevating it to a Law of Ec-
which they emerged. His first book, Die mneme, em- phory. This concern with the relation between
bedded human memory a more global theory that conditions of engraphy and ecphory anticipated rath-
broadened the construct of memory to include more er closely such modern notions as the encoding-
than simple remembering of facts, events, and the specificity principle and transfer-appropriate pro-
like. Semon argued for viewing heredity and reprocessing.
duction as forms of memory that preserved the effects
of experience across generations. He referred to the Semon also developed novel ideas about the ben-
fundamental process that subserved both heredity eficial effects of repetition on memory. In contrast to
and everyday memory with a term of his own creation, the then widely accepted idea that repetition of a
Mneme. According to Semon, Mneme is a funda- stimulus improves memory by strengthening the pre-
mental organic plasticity that allows the preservation existing engram of that stimulus, Semon argued that
of effects of experience; it is Mneme which in the or- each repetition of a stimulus creates a unique, con-
ganic world links the past and present in a living text-specific engram; at the time of ecphory, the mul-
bond (1921, p.12). tiple, separate engrams are combined by a resonance
process that Semon termed homophony. This multiple-
Semon distinguished among three aspects of the engram approach to repetition effects, with its strong
mnemic process that he believed are crucial to the emphasis on ecphoric processes, anticipated a num-
analysis of both everyday memory and of hereditary ber of recently influential conceptualizations, such as
memory, and he described them with additional the multiple-trace model developed by Hintzman
terms of his own invention in order to avoid the po- and colleagues (Schacter et al., 1978).
tentially misleading connotations of ordinary lan-
guage: engraphy, engram, and ecphory. Engraphy refers Notwithstanding the prescience of many of
to the encoding of information into memory; engram Semons ideas, his contemporaries ignored his contri-
refers to the change in the nervous systemthe butions. This neglect may be due to several factors:
memory tracethat preserves the effects of experi- his theoretical emphasis on ecphoric processes at a
ence; and ecphory refers to a retrieval process, or the time when few were interested, his social isolation as
influences which awaken the mnemic trace or engram a private scholar without institutional affiliation, and
out of its latent state into one of manifested activity his discredited Lamarckian approach to hereditary
(Semon, 1921, p.12). In attempting to apply these memory. Curiously, the one construct developed by
constructs to the analysis of hereditary memory Semon that appropirated by subsequent research-
that is, to understanding how the experiences of one ersthe engramdid not represent a novel contri-
organism could somehow influence its progeny bution and was one of the less interesting parts of his
Semon encountered a variety of biological phenome- otherwise innovative theoretical approach.
na that led him to place great emphasis on ecphory
as a crucial determinant of memory.
See also: EPISODIC MEMORY; REPETITION AND
Semons speculative ideas on hereditary memory LEARNING; RETRIEVAL PROCESSES IN MEMORY
met with severe criticism because they relied heavily
on the discredited doctrine of the inheritance of ac-
quired characteristics, which had been developed by Bibliography
the French biologist Lamarck (Schacter, 2001). Nev- Schacter, D. L. (2001). Forgotten ideas, neglected pioneers: Richard
Semon and the story of memory. Philadelphia: Psychology Press.
ertheless, the concern with ecphoric processes that Schacter, D. L., Eich, J. E., and Tulving, E. (1978). Richard
emerged from this analysis enabled Semon to devel- Semons theory of memory. Journal of Verbal Learning and Ver-
op new perspectives on human memory that elaborat- bal Behavior 17, 721743.
ed without any reference to hereditary phenomena in Semon, R. (1904). Die Mneme. Leipzig: W. Engelmann.
his second book, Die mnemischen empfindungen. At the (1909). Die mnemischen empfindungen. Leipzig: W. Engel-
mann, Leipzig.
time that Semon wrote this book, memory researchers
(1921). The mneme. London: Allen and Unwin.
paid almost no attention to the ecphoric or retrieval (1923). Mnemic psychology. London: Allen and Unwin.
stage of memory; they were caught up almost entirely
with processes occurring at the time of encoding or Daniel L. Schacter
SENSORY MEMORY 607
SENILITY task was to adjust the timing of two auditory clicks to

coincide with the apparent onset and offset of the dis-
See: AGING AND MEMORY IN HUMANS; play. The duration of the display varied from 100 to
ALZHEIMERS DISEASE: BEHAVIORAL ASPECTS;
1,000 milliseconds. By turning a knob, subjects could
ALZHEIMERS DISEASE: HUMAN DISEASE AND
THE GENETICALLY ENGINEERED ANIMAL control the occurrences of two clicks relative to the vi-
MODELS; PHARMACOLOGICAL TREATMENT sual exposure of the display. For a given objective du-
OF MEMORY DEFICITS ration, the mean onset adjustment can be subtracted
from the mean offset adjustment to arrive at an esti-
mate of how long the display seemed to last. Haber
and Standing found that these subjective durations
were longer than the objective durations. This is con-
SENSORY MEMORY sistent with the suggestion that some form of visual
Sensory memory is an agency of information storage storage follows the termination of the external dis-
that not only carries the mark of the sense modality play. Various procedures have been used to examine
in which the information originally arrivedimagery subjective persistence and have arrived at estimates of
is the more general term for thatbut also carries about 100 to 200 milliseconds.
traces of the sensory processing that was engaged by Max Coltheart distinguished two sorts of memo-
the experience. Sensory memory is the brains de- ry. One type, visible persistence, refers to the subjec-
tailed record of a sensory experience. Thus, we can tive experience that the stimulus remains available to
generate a visual image of an object without actually the visual system after stimulus offset, much in the
seeing it, but we cannot thereby have a sensory mem- manner of an afterimage. A second type, termed icon-
ory of it. Although auditory and visual verbal stimuli ic memory by Coltheart, refers to the formal availabil-
have received the most attention, there are other ity of information from the stimulus as measured in
forms of sensory memory (e.g., for nonverbal shapes, Sperlings partial-report technique. The main sup-
touch, and smell). port for Colthearts distinction between visible persis-
tence and iconic memory is that the two obey differ-
Visual Sensory Memory ent empirical laws. Experiments on iconic memory
(for example, the study by Sperling) show essentially
Research on visual sensory memory has focused
no effect of initial stimulus duration within a reason-
on two phenomena: iconic memory and subjective
able range during the first few hundred milliseconds.
persistence, descriptions of which follow below.
Likewise, in iconic memory experiments, the effect of
Iconic Memory stimulus luminance on performance is either positive
A single monograph by George Sperling, The In- or negligible. When techniques measuring visible
formation Available in Brief Visual Presentations (1960), persistencesubjective durationare used, however,
abruptly brought both the concept and the methods both display duration and luminance show an inverse
of visual sensory memory to modern attention. The effect on the length of persistence. That is, brighter
subjects in Sperlings experiment saw twelve letters and briefer displays seem to last longer than dimmer
(three rows of four) in a brief flash. In a whole-report and longer ones.
control condition, the subject was asked to report all
twelve of the letters presented; in the partial-report
Auditory Sensory Memory
conditions, a tone indicated which row was to be re-
ported, the pitch of the tone corresponding to the Different aspects of auditory sensory memory
row tested (high, medium, and low tones for first, sec- have been clarified through work on precategorical
ond, and third rows, respectively). The results showed acoustic storage and on recognition masking, dis-
that subjects had about nine letters available to the vi- cussed below in turn.
sual system if the tone indicating which row to report
Precategorical Acoustic Storage
sounded just as the display went off. People could re-
port an average of three out of the four letters on any Robert G. Crowder and John Morton proposed
row. However, partial-report scores dropped to half in 1969 that auditory sensory (that is, precategorical)
that figure, almost exactly the level of whole report, memory lies behind the consistent advantage of audi-
if the cue tone was delayed by one second. tory over visual presentation in serial, immediate re-
call situations. They suggested that following a spo-
Subjective Persistence ken stream of characters or words, people have access
Ralph N. Haber and Lionel Standing briefly not only to the interpretations they have made of
showed subjects a three-by-three array much like these items (categorical memory) but also to the actu-
those used in Sperlings experiments. However, the al sounds of the most recent item or items. This is why
608 SENSORY MEMORY
in modality comparisons the auditory presentation onds, roughly a logarithmic step longer, and under-
resulted in superior performance, but only for the last lies the suffix and modality comparisons.
few positions in the list. Presentation of an extra item
called a stimulus suffix, posing no additional load on
memory, erased most or all of this auditory advan- Developments since 1990
tage. Subsequent experiments showed that the mean- Research since 1990 has addressed the mecha-
ing of this suffix item had no effect on its tendency to nisms of sensory memory. In the previous edition of
reduce performance on the recency portion of an au- this volume, storage and proceduralist views of senso-
ditory list. However, differences between the list to be ry memory were compared. The storage view suggests
remembered and the redundant suffix had a large ef- that there are dedicated storage repositories in the
fect if they were changes in physical properties, such brain for sensory information, whereas the procedu-
as spatial location or voice quality (male versus fe- ralist view states instead that retention is a natural
male). This sensitivity to physical attributes, along consequence of the information processing that was
with the insensitivity to conceptual attributes, would originally aroused by the experience in question.
be expected of a precategorical memory store. There are still puzzles that remain to be sorted out for
each view. If there are dedicated storage repositories,
The modality-suffix findings on immediate mem- they must be complex enough to explain why sensory
ory were confidently attributed to precategorical memory of a stimulus seems to be influenced by the
acoustic storage until experiments by Kathryn T. context of preceding stimuli in that modality. If re-
Spoehr and William J. Corin (1978) and by Ruth tention is a consequence of processing, though, it
Campbell and Barbara Dodd (1980) showed that the must be complex enough to explain why there can be
original hypothesis had been too simple. These au- brain damage that interferes with the memory for
thors demonstrated that silent lipreading and related short lists of spoken words while leaving the ability to
procedures produced results in immediate memory perceive spoken words intact (as discussed, for exam-
that were almost indistinguishable from auditory pre- ple, by Alan D. Baddeley and Robert H. Logie).
sentation and were readily distinguishable from visual
presentation. The truth may lie in between these views. In a
1995 book, Attention and Memory: An Integrated Frame-
Recognition Masking work, Nelson Cowan argued that there actually are
In 1972 Dominic W. Massaro delivered to sub- short and long sensory stores in all modalities, not
jects one of two possible pure tones, twenty millisec- just the auditory modality. If so, it may be that a
onds long and pitched at either 770 or 870 Hz. The proceduralist view is more suitable for the short store,
main task was to identify which of the two tones had which is intricately tied to perception and is experi-
been presented. After this target, and at delays of enced as continuing sensation, than for the long
from 0 to 500 milliseconds, a masking tone (820 Hz) store, which is experienced as memory.
was presented. In general, presentation of the mask- The contradiction between visual information
ing tone reduced subjects abilities to identify correct- persistence and subjective persistence has been ad-
ly or to recognize which of the two tones had come be- dressed, for example by Dominic W. Massaro and
fore, especially if the mask came within about 250 Geoffrey R. Loftus (1996), with the idea that both
milliseconds of the target. The logic of this experi- could result from a single underlying process, with
ment is that if the original target tone had been fully properties that seem to match Cowans (1995) short
processed before the mask arrived, there would have store. The change in the intensity of the process over
been no decrement in its identification. But if the tar- time would determine the subjective experience of
get was still being processed when the mask arrived, the iconic image, whereas the accumulation or inte-
there must have been a sensory trace of it still avail- gration of this process over time would determine the
able somewhere in the auditory system. Comparable available information about the visual stimulus. In
experiments with speech have given much the same 1987, Cowan proposed something similar for sounds.
result.
Sensory memory is interesting as a bridge be-
From a detailed review of results and models of tween what we experience and what we remember. A
auditory integration and auditory persistence, Nelson simple view in which sensory memory fades inevitably
Cowan (1984) distinguished two types of auditory in a few seconds, like a fizzling sparkler, has proved
sensory memory: short and long. The short auditory to be too simplistic. Sensory memory rides upon per-
store is believed to have a useful life of about 250 mil- ceptual processing but then seems to outlast it in a
liseconds and is represented in the experiments on weakened form. Some residue even seems perma-
recognition masking and related techniques. The nent, as in the memory that allows recognition of the
long auditory store may last as long as two to ten sec- voices of ones close friends.
SERIAL ORGANIZATION 609
See also: MODALITY EFFECTS SERIAL ORGANIZATION

A critical form of memory organization, and one peo-
Bibliography ple frequently use, is retention of events in the tempo-
Baddeley, A. D., and Logie, R. H. (1999). Working memory: The ral order in which they occurred. Consider, for exam-
multiple-component model. In A. Miyake and P. Shah, eds., ple, your memory for the events that occurred last
Models of working memory: Mechanisms of active maintenance and summer. If someone asked you what you did during
executive control. Cambridge, UK: Cambridge University Press.
your summer vacation, most likely you would discuss
Campbell, R., and Dodd, B. (1980). Hearing by eye. Quarterly Jour-
nal of Experimental Psychology 32, 8599.
the events in the sequence in which they occurred, be-
Coltheart, M. (1980). Iconic memory and visible persistence. Per- ginning with those that occurred at the start of the
ception and Psychophysics 27, 183228. summer and concluding with those that occurred at
Cowan, N. (1984). On short and long auditory stores. Psychological the summers end. Alternatively, you could report to-
Bulletin 96, 341370. gether all the parties you attended and report as an-
(1987). Auditory sensory storage in relation to the growth
other group all the times you went hiking or swim-
of sensation and acoustic information extraction. Journal of
Experimental Psychology: Human Perception and Performance 13, ming. However, retention in terms of temporal
204215. sequence, or serial order, is most common.
(1995). Attention and memory: An integrated framework. Oxford
Psychology Series, No. 26. New York: Oxford University
Press. Definitions and Distinctions
Cowan, N., Saults, S., and Nugent, L. (2001). The ravages of abso-
lute and relative amounts of time on memory. In H. L. To study the retention of serial order in the labo-
Roediger III, J. S. Nairne, I. Neath, and A. Surprenant, eds., ratory, the information pertaining to temporal se-
The nature of remembering: Essays in honor of Robert G. Crowder. quence must be distinguished and isolated from other
Washington, DC: American Psychological Association. types of related information. The relevant distinc-
Crowder, R. G. (1976). Principles of learning and memory. Hillsdale,
tions can be made clear by considering the following
NJ : Erlbaum.
(1978). Sensory memory systems. In E. C. Carterette and hypothetical situation: Imagine a waiter in a restau-
M. P. Friedman, eds., Handbook of perception, Vol. 9. New York: rant who is taking dinner orders from the people sit-
Academic Press. ting around a table. Usually in such a situation the in-
(1983). Iconic memory. Philosophical Transactions of the Royal dividuals make their requests in a temporal sequence
Society of London B302, 283294.
that follows the spatial arrangement of the seats
(1989). Modularity and dissociations of memory systems. In
H. L. Roediger III and F. I. M. Craik, eds., Varieties of memory
around the table. However, in the present situation
and consciousness: Essays in honor of Endel Tulving, pp. 271294. this ordinary practice is not observed. Instead, the
Hillsdale, NJ: Erlbaum. waiter takes the requests in an order determined by
Crowder, R, G., and Morton, J. (1969). Precategorical acoustic stor- the individuals ages and genders, starting with the
age (PAS). Perception and Psychophysics 5, 365373. oldest woman and ending with the youngest man.
Dixon, P., and Di Lollo, V. (1994). Beyond visible persistence: An
This situation is illustrated in Figure 1. The first order
alternative account of temporal integration and segregation
in visual processing. Cognitive Psychology 26, 3363. is for ham, the second for liver, the third for steak,
Efron, R. (1970). The minimum duration of a perception. Neuro- and the fourth for chicken. The temporal sequence of
psychologia 8, 5763. the requests is thus ham, liver, steak, and chicken, a
Haber, R. N., and Standing, L. (1970). Direct estimates of the ap- sequence that does not correspond to the spatial ar-
parent duration of a flash. Canadian Journal of Psychology 24,
rangement around the table. Hence, the temporal
216229.
Kolers, P. A., and Roediger, H. L. III (1984). Procedures of mind. and spatial orders are not the same. When the waiter
Journal of Verbal Learning and Verbal Behavior 23, 425449. returns to deliver the dinners, he serves the first per-
Massaro, D. W. (1972). Preperceptual images, processing time, and son liver, the second turkey, the third steak, and the
perceptual units in auditory perception. Psychological Review last chicken. The waiter thus makes two mistakes. In
79, 124145.
the case of the turkey, he brings a dinner requested
Massaro, D. W., and Loftus, G. R. (1996). Sensory and perceptual
storage: Data and theory. In E. L. Bjork and R. A. Bjork, eds.,
by nobody, and in the case of the liver, he gives a din-
Memory. San Diego: Academic Press. ner ordered by one person to another. The first type
Sakitt, B. (1976). Iconic memory. Psychological Review 83, 257276. of mistake is an item error because the identity of the
Sperling, G. (1960). The information available in brief visual pre- dinner item is incorrect. The second type is an order
sentations. Psychological Monographs 74 (498). error because a correct item is brought but is placed
Spoehr, K. T., and Corin, W. J. (1978). The stimulus suffix effect
in the wrong position in the temporal sequence. For
as a memory coding phenomenon. Memory & Cognition 6,
583589. a discussion of laboratory methods used to distinguish
between the retention of item, temporal order, and
Robert G. Crowder spatial order information, see Alice F. Healy et al.
Revised by Nelson Cowan (1991).
610 SERIAL ORGANIZATION
Figure 1
Arrangement of table when waiter takes dinner orders (A) and when dinner is served (B) in hypothetical restaurant situation.
Another important distinction that must be made across a number of successive trials. On each trial the
is the difference between the order of the items and list is presented and the subject tries to recall it. With
their positions in the sequence. (Some researchers the method of anticipation, the subjects are not re-
refer to this same distinction as involving relative ver- quired to recite the entire list at one time. Rather,
sus absolute positions.) This distinction can be clari- each list item is presented in turn, and the subjects are
fied by considering a related hypothetical situation required to anticipate (i.e., recall) each item before it
with the same waiter and diners. In this case, the wait- is presented, in response to the item immediately pre-
er returns to take the requests for dessert. The first ceding it on the list. A correct response is scored
person orders ice cream, the second nothing, the whenever the subject correctly anticipates an item,
third cake, and the last pudding. The waiter correctly and the subject receives feedback (i.e., the subject is
brings the first person ice cream and the last pudding, told the next item in the sequence) regardless of
but he gives the cake to the second person instead of whether a correct response is made. Usually the ex-
the third. The waiter, therefore, correctly remembers perimental trials are continued until the subjects are
the order of dessertscake between ice cream and able to anticipate every item with no errors. At that
puddingbut he confuses the second and third tem- point, the investigator counts the number of correct
poral positions. responses made at each position in the list, and it be-
comes evident that the items in the different ordinal
positions in the list are not learned with the same
Empirical Results
ease. Rather, items at the beginning and end of the
Although a number of techniques have been used list yield more correct responses than those in the
to study retention of serial order (see Cowan, Saults, middle. The point of maximum difficulty is somewhat
Elliott, and Moreno, 2002, for a novel set of tech- beyond the center of the list.
niques), two procedures have been most popular.
Much of the early work on this topic used serial learn- Although the total number of correct responses
ing with the method of anticipation, and many subse- on the list may decrease when the items are more dif-
quent studies used short-term serial recall with the ficult, as when nonsense syllables are used instead of
distractor paradigm. In both methods, the results of words or when the rate of list presentation is faster,
primary interest have focused on the serial position the serial position curve remains constant across such
curve, which reveals the proportion of correct re- changes in the learning situation when it is plotted as
sponses as a function of the position of each item on the proportion of the total number of correct re-
the list. sponses made at each position in the list. The con-
stancy of the serial position curve when plotted in this
Serial Learning Method of Anticipation manner is known as the Hunter-McCrary law (Mc-
In serial learning, subjects attempt to learn an or- Crary and Hunter, 1953). A typical serial position
dered list of items (often nonsense syllables or words) curve for an eight-item list is shown in Figure 2, which
SERIAL ORGANIZATION 611
presents data reported by Bennett B. Murdock (1960) Figure 2

in an important article relating the serial position
function to results of experiments in domains of psy-
chology outside of verbal learning. The serial position
function is described as bow shaped because it resem-
bles a bow used in archery. The large relative advan-
tage for the items from the beginning of the list is
known as the primacy effect, and the smaller relative
advantage for the items from the end of the list is
known as the recency effect.
Short-Term Serial Recall with the

Distractor Paradigm
On a trial in the distractor paradigm used to
study serial recall over a short time interval, subjects
are given a short list of items to remember (typically
three to five letters), then are required to participate
in an interpolated distractor task that is meant to pre-
vent them from rehearsing the list (e.g., they may be
told to count backward from a random number), and
finally they are asked to recall the list of items accord-
ing to the order of presentation. The duration of the
distractor task, or the length of the retention interval,
varies from trial to trial but is usually quite short (no
longer than twenty seconds). Also, the list of items to Typical serial position curve for an eight-item list learned under
be recalled changes from one trial to the next. A cor- the serial method of anticipation. The data are plotted as the
rect response is scored whenever the subject recalls an proportion of the total number of correct responses made at
item that was shown and places it in the ordinal posi- each serial position of the list.
tion in which it occurred on the list. The time course
of forgetting the serial list is revealed by comparing
the proportion of correct responses at each retention Classic Models
interval. The resulting retention function is usually Although researchers have proposed many mod-
very steep; forgetting is very rapid in this paradigm. els of serial order retention, two simple opposing
A plot of the proportion of trials on which correct re- models dominated the early research on this topic.
sponses are made at each ordinal position in the list Both models include associative mechanisms as the
reveals a serial position curve that is usually bow basis for retaining serial order information. Accord-
shaped and nearly symmetrical; the primacy effect is ing to the associative chaining model, item-to-item as-
approximately equal in magnitude to the recency ef- sociations are constructed so that the first item in a se-
fect. Typical serial position curves for three different rial list is linked to the second item as a stimulus-
retention intervals are shown in Figure 3, which pres- response pair, the second item is linked in the same
ents data reported by Healy (1974). In Healys study, way to the third item, and so on to form an associative
order information was isolated from item information stimulus-response chain of items (Crowder, 1968).
because the same four items were shown on every trial For example, given the list of dinner orders ham,
of the experiment; the subjects knew the identity liver, steak, and chicken in the hypothetical restau-
of the items in advance and had only to reconstruct rant example discussed earlier, ham would be associ-
ated with liver, liver with steak, and steak with chick-
the order in which they were shown on a particular
en. The second model to account for serial order
trial.
retention involves positional associations. By this ac-
count, each item on the list is associated with the ordi-
nal number corresponding to its serial position in the
Theoretical Models list (Young, Hakes, and Hicks, 1967). In the example
A number of theoretical models have been pro- described earlier, ham would be associated with the
posed to account for serial organization. Classic mod- number 1, liver with 2, steak with 3, and chicken with
els have included simple associative mechanisms, 4. Experimental evidence has refuted both of these
which were elaborated by contemporary models. simple explanations for serial order retention.
612 SERIAL ORGANIZATION
Figure 3 phan Lewandowsky and Bennett B. Murdock (1989)

and Murdock (1995).
In contrast with TODAM, a second influential
model of serial order retention emphasizes positional
information. According to this perturbation model,
the representation of order information derives from
the representation of position information, so that
subjects can recall items in the temporal sequence ex-
actly to the extent that the positional information
they have stored in memory can adequately prescribe
the order. Associations are included in the perturba-
tion model. However, instead of associative bonds
linking successive items on the list, there are associa-
tive bonds between each item and a single control ele-
ment, which represents some aspect of the current
context in which the list is presented. For example,
given two successively presented items X and Y, rath-
er than an association of the form X-Y, the perturba-
tion model includes associations of the form X-C-Y,
where C represents the control element. This new as-
sociative mechanism has allowed for an elegant de-
scription of both short-term and long-term serial
order recall. For lucid discussions of the perturbation
model, see William K. Estes (1972, 1997).
Late-twentieth-century models have either re-
fined the associative mechanisms (Henson, 1998) or
Typical serial position curves for a four-item list recalled under
introduced new nonassociative mechanisms for ex-
the distractor paradigm. The data are plotted as the proportion
plaining serial recall. For example, the primacy
of trials on which correct responses were made at each retention
interval and serial position of the list.
model of Michael P. A. Page and Dennis Norris
(1998) is centered on the notion of a primacy gradient
of activations reflecting the strength with which the
start of the list is associated with each successive list
Contemporary Models
item. According to this model, there is a repeating
Despite the problems with the simple associative cycle in which the item with the greatest activation is
models, two more complex contemporary models selected for recall and then suppressed.
have been proposed that can be viewed as extensions
of the earlier models. Both models have derived sup- Clearly much progress has been made in under-
port from a wide range of experimental investigations standing the processes underlying serial organization
and observations, including the pervasive serial posi- in memory, but there is still considerable controversy
tion functions. According to the theory of distributed among researchers because the processes appear to
be more complex than originally envisioned.
associative memory (TODAM), serial order informa-
tion is represented in memory as a series of pairwise See also: MEMORY SPAN
associations linking successively presented items.
TODAM resembles the traditional chaining approach Bibliography
except that the items in the list and their associations Cowan, N., Saults, J. S., Elliott, E. M., and Moreno, M. V. (2002).
are stored together in memory. Each item is repre- Deconfounding serial recall. Journal of Memory and Language
sented as a list of features or a vector of numbers. 46, 153177.
Crowder, R. G. (1968). Evidence for the chaining hypothesis of se-
Rather than simple links connecting the two items in
rial verbal learning. Journal of Experimental Psychology 76, 497-
a pair, the items are connected by means of convolu- 500.
tion, which is a mathematical operation that merges Estes, W. K. (1972). An associative basis for coding and organiza-
the separate vectors for the items into a single com- tion in memory. In A. W. Melton and E. Martin, eds., Coding
posite vector. The vectors for all the items in the list processes in human memory. Washington, DC: Winston.
and for all the pairwise associations are added to a (1997). Processes of memory loss, recovery, and distortion.
Psychological Review 104, 148169.
common memory vector. The mathematical details of Healy, A. F. (1974). Separating item from order information in
TODAM and its ability to account for data from many short-term memory. Journal of Verbal Learning and Verbal Be-
serial order tasks are described in the work of Ste- havior 13, 644655.
SEX DIFFERENCES IN LEARNING 613
Healy, A. F., Cunningham, T. F., Gesi, A. T., Till, R. E., and Sexual Dimorphism in Learning Abilities
Bourne, L. E. (1991). Comparing short-term recall of item,
temporal, and spatial information in children and adults. In Although the majority of data pertaining to the
W. E. Hockley and S. Lewandowsky, eds., Relating theory and effects of gonadal steroid hormones on learning abili-
data: Essays on human memory, in honor of Bennet B. Murdock. ties have been gathered in rodents, there is growing
Hillsdale, NJ: Erlbaum. evidence for similar effects in primates, including hu-
Henson, R. N. A. (1998). Short-term memory for serial order: The mans. In rodents (Beatty, 1979), mature male rats
start-end model. Cognitive Psychology 36, 73-137.
make fewer errors than females in learning complex
Lewandowsky, S., and Murdock, B. B. (1989). Memory for serial
order. Psychological Review 96, 2557. mazes, and are faster learners in appetitive learning.
McCrary, J. W., and Hunter, W. S. (1953). Serial position curves in Conversely, female rats outperform males in discrimi-
verbal learning. Science 117, 131-134. nation reversal learning and in the acquisition of ac-
Murdock, B. B. (1960). The distinctiveness of stimuli. Psychological tive avoidance. This pattern of male and female per-
Review 67, 1631. formance in avoidance learning can be reversed by
(1995). Developing TODAM: Three models for serial-order treating ovariectomized female rats with androgens
information. Memory & Cognition 23, 631645.
Page, M. P. A., and Norris, D. (1998). The primacy model: A new
(testosterone) and male rats with antiandrogens from
model of immediate serial recall. Psychological Review 105, birth to adulthood. Finally, lesions of brain areas
761781. known to be functionally involved in avoidance learn-
Young, R. K., Hakes, D. T., and Hicks, R. Y. (1967). Ordinal posi- ing, such as the basal ganglia, do not impair perfor-
tion number as a cue in serial learning. Journal of Experimental mance of males and females equally. For example,
Psychology, 73, 427438. small lesions of the globus pallidus impair acquisition
of active avoidance in males but not in females. This
Alice F. Healy
sex difference in the effects of pallidal lesions can be
reversed by androgenization of female rats and femi-
nization of male rats. These data in rodents suggest
that androgens are, at least in part, responsible for
the sex differences in learning abilities.
SEX DIFFERENCES IN LEARNING
In nonhuman primates (Mitchell, 1977), adult fe-
Findings in the field of neuroendocrinology, which
male rhesus monkeys outperform males in spatial
explores the functional relationships between hor- memory, as measured by the delayed-response task.
mones and the brain, have indicated that the release Similarly, adult female chimpanzees exhibit signifi-
of testicular (androgens) and ovarian (estrogens) hor- cantly superior short-term stimulus memory, as mea-
mones during critical periods of brain development sured by the delayed matching-to-sample task. Final-
exert a profound effect on the genesis and survival of ly, although male and female rhesus monkeys
neurons in specific brain areas, resulting in sex differ- perform similarly on tasks of object memory and ex-
ences in reproductive behaviors. Sex differences are ecutive function, young males outperform young fe-
not restricted to the reproductive sphere, however. males on a spatial memory task (Lacreuse et al.,
They are found in a broad range of nonreproductive 1999). This superior level of spatial ability in young
behaviors such as aggression, locomotor activity, play males declined sharply with age, and, at older age,
behavior, and learning abilities. While the debate on males do not perform significantly better than fe-
whether sex differences in cognition are determined males. These studies in adult primates did not exam-
by biological or sociocultural factors is long lasting, ine whether the sex differences in learning abilities
increasing experimental evidence indicates that sex could be reversed by perinatal gonadal steroid hor-
differences could be determined by genetic or hormone treatments, but such evidence has been ob-
monal factors. Support for this conclusion is of four tained in studies of infant primates.
kinds. First, sex differences have been reported in Acquisition of an object discrimination reversal
various learning abilities that depend on neural struc- task, known to depend on the integrity of the orbital
tures other than those involved in reproductive func- prefrontal cortex in the adult monkey, is significantly
tions. Second, there are sex differences in the mor- more rapid in male infant monkeys than in females
phology of neural structures mediating learning (Clark and Goldman-Rakic, 1989). Postnatal injec-
abilities. Third, lesions of neural structures underly- tions of testosterone propionate in the females en-
ing learning abilities can affect males and females dif- hances their performance to the level of the normal
ferently. Finally, sex differences in learning abilities, infant males. When orbital prefrontal cortex is re-
brain morphology, and/or in the effects of brain le- moved in infancy, intact male monkeys and androge-
sions can be reversed by manipulations of gonadal nized female monkeys are as impaired as adult mon-
steroid hormones in animals or by normal fluctuation keys with the same lesions, whereas untreated infant
of sex hormones across the life span in humans. females do not differ from untreated age-matched fe-
614 SEX DIFFERENCES IN LEARNING
males. The data suggest that the orbital prefrontal tion than at other points in the cycle. Conversely,
cortex matures earlier in male than in female mon- higher levels of estrogen during the cycle are associat-
keys. Conversely, acquisition of a concurrent visual ed with better performance on many tasks in which
discrimination learning task, known to depend on the women typically excel. When postmenopausal women
integrity of the inferior temporal cortex, is signifi- are tested either when receiving their estrogen thera-
cantly more rapid in female infant monkeys than in py or when they are off their medication for at least
males (Bachevalier and Hagger, 1991) and this sex four days, performance on fine motor and spatial
difference is positively correlated in three-month-old tasks tends to be faster and more accurate during the
male animals with circulating levels of testosterone testosterone treatment compared with the off-
(the higher the level of testosterone, the poorer the treatment phase (Hampson and Kimura, 1992). Fi-
score), but not with estradiol levels. Neonatal orchiec- nally, estrogen deficiency in menopausal women is as-
tomy, which reduced plasma testosterone levels, has- sociated with memory impairments that are reversible
tens performance on visual discrimination learning in by estrogen replacement therapy, whereas testoster-
male infant monkeys, whereas treatment of andro- one replacement therapy in elderly men enhances
gens (dihydrotestosterone) in neonatally ovariecto- spatial performance (Janowsky et al., 1994).
mized female infant monkeys delays their perfor- The double dissociation found in infant monkeys
mance. Finally, early postnatal inferior temporal with the object discrimination reversal and concur-
cortex lesions affect performance of female but not of rent discrimination tasks have been replicated in very
male infant monkeys, though male and female adults young children using almost identical cognitive tasks
with the same lesions are impaired equally. The data and nonverbal procedures (Overman et al., 1997).
suggest that this temporal cortical area is functionally Boys under the age of twenty-nine months significant-
more mature in female infant monkeys than in males. ly outperform girls on the object reversal task, but
Thus, gonadal steroid hormones appear to play an in- girls outperform boys on the concurrent discrimina-
ductive role in the postnatal differentiation of cortical tion task. Given the close parallel in learning behavior
mechanisms. in human infants and infant monkeys, it is reasonable
For humans, many studies have reported that to propose that the gender differences are mediated
women excel in verbal abilities, perceptual speed, ar- by similar biological mechanisms in both species.
ticulation, and fine motor skills, whereas men gener- Therefore, in children, as in infant monkeys, there
ally excel in tasks measuring visuospatial abilities, may be a more rapid maturation of orbital prefrontal
particularly those requiring mental rotation of objects circuits in boys and a more rapid maturation of inferi-
and imagining what an object would look like from a or temporal circuits in girls.
different vantage point (Halpern, 1992). Although di-
rect manipulation of gonadal hormones is not possi-
ble in humans, the organizational and activational ef-
Sex Differences in Brain Areas Related to
fects of the hormones on cognitive functions have Learning Abilities
usually been inferred from the studies of different Although no direct correlation has been estab-
populations of individuals. They include boys and lished between sex differences in learning abilities
girls suffering from long-standing prenatal hormonal and morphology of brain areas, sex differences in nu-
anomalies or that have been exposed to exogenous merous neural structures related to learning abilities
hormones in utero, women during regular fluctua- are well documented in rodents (Beatty, 1979;
tions of estrogen and progesterone throughout the Juraska, 1991). In the limbic system (bed nucleus of
menstrual cycle or postmenopausal women receiving the stria terminalis and hippocampus), the number
hormone replacement therapy, and elderly individu- and volume of neurons differ in male and female rats.
als showing a decline in cognitive functions. Thus, These morphological differences are reversed after
girls with congenital adrenal hyperplasia (CAH) who postnatal treatment with gonadal steroid hormones.
are genetically masculinized and prenatally exposed The rate of neonatal cell proliferation has been shown
to excessive androgens show significant enhancement to be slower in male than in female rats, indicating a
of visuospatial ability as compared to unaffected fe- delayed maturation of the neocortex in males com-
males. Also, boys exposed prenatally with DES (a syn- pared with that of females. Similarly, the neurons in
thetic estrogen) show poorer performance on several the somatosensory cortex of young male rats are larg-
spatial tasks than males who suffer from low testoster- er than those of females, reflecting a cortical imma-
one due to post puberty pathology (Reisnich et al., turity and, possibly, a less developed synaptic network
1991). In addition, when performance on several vi- in males than in females. In research conducted in the
suospatial tasks is compared at different phases of the 1990s, sex differences in humans have also been
menstrual cycle, women perform significantly more found for functional asymmetry (Voyer, 1996), brain
poorly during the estrogen surge just prior to ovula- morphology, metabolism, weigh, volume, and neo-
SEX DIFFERENCES IN LEARNING 615
cortical neuron number (Gur et al., 1995; Pakkenberg differences in structure and function are likely to be
and Gundersen, 1997; Hampson and Kimura, 1992). a pervasive characteristic of brain organization and
mediated by gonadal steroid hormones. Neverthe-
less, the challenge is to precisely specify biological
Mechanisms of Action mechanisms by which these differences occur and to
The question of precisely how gonadal hormones take into consideration the circular interactions be-
exert their organizational and activational influence tween biological factors and socioenvironmental fac-
on brain areas related to learning abilities remains to tors. Ultimately, the understanding of cognitive sex
be answered. There is, however, indirect evidence re- differences will necessarily depend upon converging
garding the mechanism of their action on brain areas evidence from many different disciplines, including
mediating learning abilities (Luine and McEwen, endocrinology, animal and human behavior, neuro-
1985). For example, gonadal steroid hormones are biology, electrophysiology, and brain imaging.
known to act via intracellular receptors located in lim-
bic structures and some parts of the neocortex. In the See also: DISCRIMINATION AND GENERALIZATION;
developing rhesus monkey, androgen metabolism NEURAL SUBSTRATES OF AVOIDANCE LEARNING
has been observed in all cortical areas; this activity de-
clines from prenatal to early postnatal life. Also, the Bibliography
presence of sex differences in neurochemical concen- Bachevalier, J., and Hagger, C. (1991). Sex differences in the de-
trations and regulatory processes suggests an influ- velopment of learning abilities in primates. Psychoneuroen-
ence of gonadal steroid hormones on the differentia- docrinology 16, 179190.
Beatty, W. W. (1979). Gonadal hormones and sex differences in
tion of neurochemical features of neurons. Finally, nonreproductive behaviors in rodents: Organizational and
gonadal steroid hormones stimulate neurite out- activational influences. Hormones and Behavior 12, 112163.
growth during the sensitive period of brain differenti- Clark, A. S., and Goldman-Rakic, P. S. (1989). Gonadal hormones
ation, presumably by increasing the competitive ad- influence the emergence of cortical function in nonhuman
vantage of neurons to make connections with other primates. Behavioral Neuroscience 103, 1,2871,295.
Gur, R. C., et al. (1995). Sex differences in regional glucose metab-
neurons. The perinatal androgen surge seen in infant olism during a resting state. Science 267, 528531.
males could therefore affect the rate of brain matura- Halpern, D. F. (2000). Sex differences in cognitive abilities, 3rd edition.
tion by influencing neuronal connectivity at the corti- Mahwah, NJ: Erlbaum.
cal level. Hampson, E., and Kimura, D. (1992). Sex differences and hormon-
al influences on cognitive function in humans. In J. B. Becker,
There are also several possible mechanisms by S. M. Breedlove, and D. Crews, eds, Behavioral Endocrinology.
which gonadal steroids may exert more transient, or Cambridge, MA: MIT Press.
activational, effects on central nervous system during Janowsky, J. S., Oviatt, S. K., and Orwoll, E. S. (1994). Testosterone
adulthood. Estradiol as well as other gonadal hor- influences spatial cognition in older men. Behavioral Neuro-
science 108, 325332.
mones can regulate the concentrations of specific en- Juraska, J. M. (1991). Sex differences in cognitive regions of the
zymes involved in neurotransmitter synthesis and rat brain. Psychoneuroendocrinology 16, 105119.
breakdown (McEwen et al., 1984). This action may Lacreuse, A., et al. (1999). Spatial cognition in rhesus monkeys:
offer a way by which a single hormone can simulta- Male superiority declines with age. Hormones and Behavior 36,
neously exert different effects on different behavioral 7076.
Luine, V. N., and McEwen, B. S. (1985). Steroid hormone recep-
systems. tors in brain and pituitary. In N. Adler, D. Pfaff, and R. W.
Goy, eds., Handbook of behavioral neurobiology, Vol. 7: Reproduc-
tion. New York: Plenum Press.
Conclusion McEwen, B. C., et al. (1984). Towards a neurochemical basis of ste-
roid hormone action. In L. Martini, and W. F. Ganong, eds.,
Although sex differences in cognitive and learn-
Frontiers in Neuroendocrinology, Vol. 8. New York: Raven Press.
ing abilities are presently widely acknowledged, the Mitchell, G. (1977). A note on sex differences in learning or moti-
basis for these differences remains controversial. The vation in nonhuman primates. Laboratory Primate Newsletter
data reviewed in this entry suggest that androgens or- 16, 15.
ganize the brain pre- and perinatally for all sexually Overman, W. H., Bachevalier, J., Schumann, E., and McDonough-
Ryan, P. (1997). Sexually dimorphic brain-behavior develop-
dimorphic behaviors, including problem-solving be-
ment: A comparative perspective. In N. A. Krasnegor, G. R.
haviors, and this appears to be true in mammals, non- Lyon, and P. S. Goldman-Rakic, eds., Development of the pre-
human primates, and humans. Moreover, the pattern frontal cortex: Evolution, neurobiology, and behavior. Baltimore,
of variation in learning abilities documented over the MD: Brookes Publishing Company.
menstrual cycle and during aging processes raises the Pakkenberg, B., and Gundersen, H. J. (1997). Neocortical neuron
number in humans: Effects of sex and age. Journal of Compara-
possibility that sex differences in cognitive abilities in
tive Neurology 384, 312320.
humans may also at least be partly due to an activa- Reinisch, J. M., Ziemba-Davis, M., and Sanders, S. A. (1991). Hor-
tional influence of sex hormones on the brain monal contributions to sexually dimorphic behavioral devel-
throughout life. Thus, it is becoming clear that sex opment in humans. Psychoneuroendocrinology 16, 213278.
616 SKINNER, B. F.
Voyer, D. (1996). On the magnitude of laterality effects and sex dif- ect to train pigeons to guide missiles. While at the
ferences in functional literalities. Laterality 1, 5183. University of Minnesota, he married Yvonne (Eve)
Jocelyne B. Bachevalier Blue, with whom he had two children, Julie and Debo-
rah. In 1945 he moved to Indiana University, where
he remained until 1947, when he returned to Har-
vard University. During that same year, he delivered
SKINNER, B. F. (19041990) his William James Lectures on Verbal Behavior,
which evolved into his book on that subject in 1957.
The American psychologist B. F. Skinner was re-
nowned for his pioneering work in behaviorism. Born As he himself implied, Skinner held on to the
on March 20, 1904, in Susquehanna, Pennsylvania, concept of reflex beyond its usefulness when he
Burrhus Frederic Skinner was the older son of Grace wrote his book The Behavior of Organisms in 1938. Not
Madge Burrhus Skinner and William Arthur Skinner, long after that, he gave up the concept because oper-
an attorney with some political aspirations. Skinners ant behavior not elicited but emitted; he thus ceased
younger brother died suddenly of a cerebral aneu- to be a stimulus-response psychologist. This means
rysm at the age of sixteen. Skinner did his undergrad- that Skinner did not conceive of human beings, or
uate work at Hamilton College in Clinton, New York, any organisms, as automatons waiting to have some
where he majored in English. During the summer be- behavior elicited. Rather, he viewed them as emitting
fore his senior year, he studied at the Bread Loaf behavior upon which the environment acts by select-
School of English at Middlebury, Vermont. There he ing some of it through the provision of consequences.
met Robert Frost, who asked Skinner to send him Also important in this context is the concept of classes
some of his work. Frosts comments encouraged Skin- of behavior and classes of stimuliSkinner referred
ner to try writing, at first in his parents home and to this as the generic nature of stimulus and response
later in New York Citys Greenwich Village. He dis- (1935). Even though behavior analysis, a term now used
covered that I had nothing important to say (Skin- to describe Skinners concepts of learning, refers to
ner, 1970, p. 7). He then turned to psychology and classes, not some hyperspecified atomistic stimulus
graduate work at Harvard University. and response, uninformed people still characterize
Skinners approach incorrectly as atomistic.
Several factors drew Skinner to psychology. First,
his biology teacher directed him to Jacques Loebs The difficulty of eliciting operant behavior neces-
Physiology of the Brain and Comparative Psychology sitated the invention of a special procedure to pro-
(1900) and Pavlovs Conditioned Reflexes (1927). Then duce new behaviorhence the concept of shap-
the writings of Bertrand Russell, in The Dial, a literary ing. Skinners approach to learning emphasized the
magazine, and in the book Philosophy (1927), which he three-part reinforcement contingency. Behavior oc-
read while writing in Greenwich Village, led him to J. curring on particular occasions and followed by cer-
B. Watsons Behaviorism (1924). Harvards depart- tain consequences (reinforcers) will be strengthened
ment of psychology did not strengthen his interest in by those consequences; that is, other members of the
behaviorism, but Fred S. Keller, then a graduate stu- same response class will have a higher probability of
dent in the department, did. Skinner described Keller occurring on similar occasions. There are positive
as a sophisticated behaviorist in every sense of the and negative reinforcers. The former strengthens the
word (1970, p. 9) and his thesis as having only the behavior that produces it and the latter strengthens
vaguest of Harvard connections (1970, p. 10). It in- the behavior that avoids or eliminates it. Reinforcers
cluded his study of eating rate in the rat (which came are also divided into unconditioned (primary) and
to be the response rate of later work), two brief papers conditioned (secondary). The former act as reinforc-
on the reflex and drive, and his paper on the concept ers without any learning history, whereas the latter act
of the reflex in psychology. That concept was based as reinforcers because of their association with the un-
on an operational analysis in which he insisted on de- conditioned reinforcers. Skinner distinguished rein-
fining it as an observed correlation of stimulus and re- forcers from punishing stimuli, which weaken the be-
sponse. He used the equation R = f (S, A), where R havior they produce.
stood for reflex strength, S for stimulus, and A for any Skinners concept of operant behavior has gener-
condition affecting reflex strength, such as drive, ated many experiments, including those on schedules
which was specified in terms of the deprivation opera- of reinforcement in which the different intermittent
tion (Skinner, 1977). patterns of reinforcement give rise to characteristic
After receiving his Ph.D., Skinner served as a ju- patterns of response rates (Ferster and Skinner,
nior fellow in the Harvard Society of Fellows for three 1957). The concept of intermittent reinforcement was
years; then he moved to the University of Minnesota significant in a variety of ways, notably in its resem-
where, during World War II, he embarked on a proj- blance to the conditions of the natural environment,
SKINNER, B. F. 617
which brought basic learning research closer to the

real world. The number of different kinds of inter-
mittent schedules that can be generated is limited
only by the experimenters imagination, but they gen-
erally fall into two broad classes: one in which rein-
forcement depends on the frequency or type of be-
havior and the other in which it depends on the
occurrence of a response following the passage of a
certain interval or various intervals.
Intermittent schedules of reinforcement pro-
duced behavior that is particularly resistant to extinc-
tion and thus gave rise to the study of maintenance
of behavior, to which other learning approaches gave
scant attention. Maintenance of behavior is like mem-
ory, a concept Skinner avoided. Instead of viewing re-
call as searching a storehouse of memory, he con-
sidered the conditions, both external and response-
produced, that increase the probability of
responses (Skinner, 1974, pp. 109210). Interest-
ingly, Skinner did not limit his work to basic research.
With respect to memory, he wrote a charming and in-
formative book (Skinner and Vaughn, 1983) outlin-
ing a program of self-management in old age.
Skinners book on verbal behavior (1957) ap-
B. F. Skinner (The Library of Congress)
peared in the same year as his work on intermittent
reinforcement (Ferster and Skinner, 1957). He con-
sidered the former to be his most important contribu-
tion to psychology and viewed verbal behavior as he and Heron, 1937), thereby initiating an area still
did other behavior, not as standing for something else practiced and useful; and, as already mentioned, he
(Skinner, 1945) but as constituting the subject matter applied it to old age.
of interest. In contrast with methodological behavior- Skinners first excursion into the study of culture
ists, who must restrict their studies to currently mea- and how to improve took the form of a novel, Walden
surable phenomena, Skinner the radical behaviorist Two (Skinner, 1948). He returned to that theme in
was able to extend his analysis to private events that Science and Human Behavior (Skinner, 1953) and in Be-
cannot yet be measured. In his book on verbal behav- yond Freedom and Dignity (Skinner, 1971). He always
ior and later in his Contingencies of Reinforcement (Skin- remained close to the principles of behavior analysis
ner, 1969), Skinner explicitly recognized that not all that he had discovered in his basic research. Skinner
behavior is produced through conditioning; rule- was undoubtedly one of the most influential psycholo-
governed behavior is produced not through exposure gists of the twentieth century. His systematization of
to the actual contingencies of reinforcement but to a behavior was never limited to learning as such. Rath-
verbal description of those contingencies. In one of er, he and others applied his approach to all areas of
his last papers, Skinner (1990) suggested that such psychology. A reconsideration of his basic papers,
rule-governed behavior might, as knowledge by de- complete with comments by his supporters and crit-
scription, postpone the destruction of the earth. icsalong with his response to those comments
appeared in Catania and Harnad (1984).
Skinner applied his principles of behavior to
many areas of functioning. In education, he invented B. F. Skinner died in 1990.
programmed instruction, a form of learning in which See also: BEHAVIORISM; OPERANT BEHAVIOR;
students always make the correct response, thus PAVLOV, IVAN; WATSON, JOHN B.
having their correct responses immediately rein-
forced (Skinner, 1954a; 1968). He used the methods Bibliography
of shaping and stimulus fading to make that possible. Catania, A. C., and Harnad, S., eds. (1984). Canonical papers of B.
F. Skinner. The Behavioral and Brain Sciences 7, 473724.
In abnormal psychology, he first talked about behav-
Ferster, C. F., and Skinner, B. F. (1957). Schedules of reinforcement.
ior modification by applying reinforcement to psy- New York: Appleton-Century-Crofts.
chotic patients behavior (Skinner, 1954b). He ap- Loeb, J. (1900). Physiology of the brain and comparative psychology.
plied behavior analysis to the study of drugs (Skinner New York: Putnam.
618 SLEEP AND MEMORY CONSOLIDATION
Pavlov, I. (1927). Conditioned reflexes. London: Oxford University studies have provided converging evidence on the im-
Press. portant role sleep plays in the off-line reprocessing of
Russell, B. (1927). Philosophy. New York: W. W. Norton.
Skinner, B. F. (1935). The generic nature of the concepts of stimu-
waking memories. Whether dreaming plays a similar
lus and response. Journal of General Psychology 12, 4065. role remains unclear.
(1938). The behavior of organisms. New York Appleton-
Century-Crofts. Sleeps complex role in memory consolidation
(1945). The operational analysis of psychological terms. makes our understanding of that role more difficult.
Psychological Review 52, 270277. Multiple memory systems exist within the brain store
(1948). Walden two. New York: Macmillan. and process different types of information in separate
(1953). Science and human behavior. New York: Macmillan.
anatomical regions. For example, episodic memory
(1954a). The science of learning and the art of teaching.
Harvard Educational Review 24, 8697. recall in humans is dependent on the hippocampus,
(1954b). A new method for the experimental analysis of the whereas access to procedural memories is not. Nu-
behavior of psychotic patients. Journal of Nervous and Mental merous mechanisms can contribute to memory con-
Diseases 120, 403406. solidation. Consolidation can refer to the simple
(1957). Verbal behavior. New York: Appleton-Century-
strengthening of a memory, its movement from one
Crofts.
(1968). The technology of teaching. New York: Appleton- memory system to another, or its functional linking
Century-Crofts. to other, associated memories. As a result, sleeps con-
(1969). Contingencies of reinforcement. New York: Appleton- tribution to memory consolidation depends on the
Century-Crofts. precise memory system involved and the form of con-
(1970). B. F. Skinner: An autobiography. In P. B. Dews, ed.,
solidation being considered. Sleep represents a simi-
Festschrift for B. F. Skinner. New York: Appleton-Century-
Crofts. larly complex phenomenon because it consists of a se-
(1971). Beyond freedom and dignity. New York: Alfred A. ries of brain states with different neurophysiological
Knopf. and neurochemical properties. This deceptively sim-
(1974). About behaviorism. New York. Alfred A. Knopf. ple question of sleeps role in memory reprocessing
(1976). Particulars of my life. New York: Alfred A Knopf.
must be expanded to address the contribution of spe-
(1977). The experimental analysis of operant behavior. In
R. W. Rieber and K. Salzinger, eds., The roots of American cific stages of sleep to various mechanisms involved
psychology: Historical influences and implications for the fu- in the consolidation of several distinct forms of mem-
ture. Annals of the New York Academy of Sciences 291, 374385. ory.
(1979). The shaping of a behaviorist. New York: Alfred A.
Knopf.
(1983). A matter of consequences. New York: Alfred A. Knopf.
(1990). To know the future. The Behavior Analyst 13, 103
The REM Sleep Cycle
106. Sleep in mammals follows a rhythmic pattern of
Skinner, B. F., and Heron, W. T. (1937). Effects of caffeine and alternating REM and nonREM (NREM) sleep. In hu-
benzedrine upon conditioning and extinction. Psychological
Record 1, 340346.
mans, this cycle has a period of approximately ninety
Skinner, B. F., and Vaughn, M. E. (1983). Enjoy old age. New York: minutes, and continues throughout the night (see
W. W. Norton. Figure 1, top panel). NREM sleep is divided into four
Watson, J. B. (1924). Behaviorism. New York: W. W. Norton. stages, ranging from very light (stage I) sleep onset
Kurt Salzinger sleep to deep slow wave sleep (SWS; stages III and IV),
so named because of its characteristic electroencepha-
lographic (EEG) pattern (see Figure 1, second panel).
Throughout the night this ninety-minute period re-
mains relatively constant, but a slow shift from a pre-
SLEEP AND MEMORY ponderance of SWS early in the night toward a pre-
CONSOLIDATION ponderance of REM sleep at the end of the night
occurs.
More than two hundred years have passed since
David Hartley (English psychologist and philosopher, Several physiological parameters vary across the
17051757) first proposed that dreaming might alter REM cycle. As sleep progresses from stage I to stage
the strength of associative links between memories, IV, EEG patterns show progressively slower and
and more than one hundred years since Sigmund higher amplitude waves, whereas REM sleep shows a
Freud (Austrian neurologist and founder of psycho- high frequency low amplitude EEG pattern. Thus,
analysis, 18561939) suggested that dreaming served REM sleep has been given the alternate name of para-
to process traumatic memories. But it has only been doxical sleep, based on the similarity between its EEG
since 1953, with the discovery of rapid eye movement pattern and the pattern seen in waking. These distinct
(REM) sleep by Eugene Aserinsky and Nathaniel EEG waveforms seen in different sleep stages may
Kleitman, that studies of sleeps role in memory pro- contribute differentially to various aspects of memory
cessing began in earnest. Since then, a wide range of consolidation. Tonic muscle activity, measured by the
SLEEP AND MEMORY CONSOLIDATION 619
electromyogram (EMG), decreases with descent

through light NREM sleep into SWS, but is at its low-
est level in REM sleep (see Figure 1, third panel).
During REM, a descending spinal path from the
brainstem bulbar reticular formation actively inhibits
the voluntary muscles. The resulting atonia produces
a functional paralysis during REM sleep, which is nec-
essary to prevent the physical acting out of dreams.
Spontaneous eye movements, recorded in the elec-
trooculogram (EOG) also show distinct stage-
dependent patterns (see Figure 1, bottom panel).
While NREM sleep stages II through IV show little or
no eye movements, both sleep onset (stage I) and
REM sleep show stereotypical patterns of movements.
During sleep onset, slow rolling eye movements, last-
ing one to three seconds are observed, whereas dur-
ing REM sleep phasic bursts of rapid eye movements
are seen. These bursts correlate with times of peak
dream recall.
The shift from NREM to REM sleep is accompa-
nied by an increase in release of acetylcholine in the
brain and a simultaneous near-cessation of release of
norepinephrine and serotonin. Brain imaging studies
show that most brain regions become less active dur-
ing NREM sleep; several distinct regions, including
the anterior cingulate and medial orbitofrontal cor-
tices and the amygdala become more active in REM
sleep. Together, these changes are thought to control
the variations in memory reprocessing and dreaming
seen across the sleep cycle.
Human Procedural Skill Consolidation

The clearest evidence of the important role sleep
plays in human memory reprocessing comes from
studies of the consolidation of procedural learning.
Posttraining sleep can improve both perceptual and
motor skill learning.
In visual texture discrimination tasks, training
only leads to improved performance after a nights
sleep (see Figure 2). Subjects trained and then re-
tested the same day show no improvement, whereas
subjects retested after a nights sleep show significant
improvement. Similarly, subjects that are sleep de-
prived the night after training and then retested two
days later, also show no improvement. Thus, sleep the
night after training appears critical for the consolida-
tion of this learning. Those stages of sleep that are
consolidation requiring SWS followed by REM sleep
most important remain unclear. Studies have found
occurs.
that selectively decreasing either REM or SWS can
block improvement, and overnight improvement cor- In contrast to these findings with perceptual skill
relates both with the amount of SWS early in the night tasks, motor skill tasks can show dependence on stage
and the amount of REM sleep late in the night. These II NREM sleep. Individuals show a twenty percent in-
findings in humans match findings in rats, suggesting crease in speed on a finger-tapping task after a nights
that a two-step process of sleep-dependent memory sleep, correlating with the amount of stage II sleep
620 SLEEP AND MEMORY CONSOLIDATION
terfere with the retention or consolidation of more

complex declarative memories, such as recall of lists
of words grouped into categories, or the acquisition
of such complex skills at BASIC computer program-
ming, foreign languages, or Morse code. It is, howev-
er, unclear whether it is the specifically declarative
portion of such learning or more subtle, nondeclara-
tive components that are being affected. Neverthe-
less, these results suggest that some aspects of com-
plex declarative memories are supported by sleep.
Animal Memory Consolidation

Because the classification of a memory as declara-
tive requires a verbal statement of recall, it is impossi-
ble to know whether animals possess such memories.
However, many forms of animal learning are clearly
impaired by subsequent REM sleep deprivation, indi-
cating that the role of sleep in memory consolidation
is not uniquely human. In rats, posttraining REM de-
privation impairs both aversive and appetitive tasks,
although simpler tasks may not share this property.
Thus, simple shock avoidance training is unaffected
by subsequent REM sleep deprivation, whereas a
more complex, shuttle box avoidance task is REM
sleep dependent. Memory consolidation appears to
be sensitive to REM deprivation only during specific
REM windows (Carlyle Smith, 1985), often occurring
hours to days after the initial training. These REM
windows are further characterized by increased REM
sleep following training. During periods when REM-
dependent memory consolidation is occurring, the
brain appears to produce more REM sleep.
In humans, declarative memories are dependent
on the hippocampus for their encoding and initial re-
call, whereas procedural skill learning is largely inde-
pendent of this structure. In rats, spatial learning
tasks are hippocampally mediated, and it is possible
to look at the role of sleep in consolidating hippocam-
pally mediated memories in rats by comparing spatial
and non-spatial tasks. Surprisingly, both types of
tasks are found to be sleep dependent; posttraining
REM deprivation impairs performance on both the
Morris water maze and the eight-arm radial arm
obtained, especially late in the night. These differ-
maze. Thus, these tasks may correspond to the com-
ences between perceptual and motor skill learning
plex declarative memory tasks in humans, which show
may reflect the fact that different brain regions are in-
a similar REM dependency.
volved in each of these forms of learning.
Memory Reactivation in Sleep

Human Declarative Memory Consolidation Studies showing that patterns of brain activation
Less convincing evidence exists for sleeps role in seen during learning are reactivated during sleep
the consolidation of declarative memories. Sleep de- provide additional support for sleep dependent
privation, particularly REM sleep deprivation, has lit- memory processing. Evidence for this comes from
tle or no effect on the retention of simple declarative both animal and human experiments. Recordings
memories such as paired word associates, but may in- from the rat hippocampus show the most direct evi-
SOCIAL MEMORY PROCESSES 621
dence, revealing that the rat hippocampus, during See also: MEMORY CONSOLIDATION: MOLECULAR
both REM and NREM sleep, activates patterns of AND CELLULAR PROCESSES; MEMORY
neuronal activity that mimic patterns seen earlier CONSOLIDATION: PROLONGED PROCESS OF
when the rat was navigating a maze. The patterns are REORGANIZATION; MOTOR SKILL LEARNING
sufficiently complex that one can visualize the virtual Bibliography
maze running activity of the sleeping rat. Interesting- Aserinsky, E., and Kleitman, N. (1953). Regularly occurring peri-
ly, this repetition is seen during NREM sleep only in ods of ocular motility and concomitant phenomena during
the first thirty minutes after maze running, while the sleep. Science 118, 361375.
REM reactivation is seen twenty-four hours later. Freud, S. (1900). The interpretation of dreams. New York: Basic
Books.
Thus, as has been suggested from behavioral studies, Hartley, D. (1791). Observations on man, his frame, his duty and his ex-
memory processing might occur first during NREM pectations. London: Johnson.
sleep and only subsequently during REM sleep. Simi- Hennevin, E., Hars, B., Maho, C., and Bloch, V. (1995). Processing
lar patterns of reactivation have been seen in the neo- of learned information in paradoxical sleep: Relevance for
memory. Behavioural Brain Research 69, 125135.
cortex. Unfortunately, no behavioral studies exist to
Peigneux, P., Laureys, S., Delbeuck, X., and Maquet, P. (2001).
show that this replay of patterns of either hippocam- Sleeping brain, learning brain. The role of sleep for memory
pal or neocortical neuronal activity during sleep are systems. Neuroreport 12 (18), A111124.
actually associated with memory consolidation. Siegel, J. M. (2001). The REM sleep-memory consolidation hy-
pothesis. Science 294, 1,0581,063.
Less direct evidence of the reactivation of neuro- Smith, C. (1985). Sleep states and learning: A review of the animal
nal ensembles coding memory traces is found in literature. Neuroscience and Biobehavioral Reviews 9,157168.
(1995). Sleep states and memory processes. Behavioural
dream reports collected during the sleep onset period Brain Research 69, 137145.
following intense video game play. Sixty to ninety Stickgold, R. (1998). Sleep: Off-line memory reprocessing. Trends
percent of subjects that played either the video game in Cognitive Sciences 2 (12), 484492.
Tetris, or the arcade game Alpine Racer II, reported Stickgold, R., Hobson, J. A., Fosse, R., and Fosse, M. (2001). Sleep,
learning and dreams: Off-line memory reprocessing. Science
dreamlike images from the game when awakened 294, 1,0521,057.
during the first few minutes of sleep following several Vertes, R. P., and Eastman, K. E. (2000). The case against memory
hours of intensive game play. In most cases, the im- consolidation in REM sleep. Behavioral and Brain Sciences 23,
ages were accurate copies of game elements, suggest- 867876.
ing that memory traces were being reactivated during Robert Stickgold
the sleep onset period. Subjects playing Alpine Racer
reported both visual and kinesthetic imagery, indicat-
ing that coordinated multimodal replay is occurring.
Surprisingly, the reactivation is not hippocampally
mediated, since amnesic patients with extensive dam- SOCIAL MEMORY PROCESSES
age to both hippocampi produce similar Tetris images Studies of social memory support the trite adage,
despite being unable to identify the source of the im- Two heads are better than one. But there is more
ages or to recall playing the game. to this story. When recalling meaningful materials
such as stories, word lists, and criminal acts, groups
Some researchers continue to question sleeps
remember more than individuals. But in the recall of
role in memory consolidation, but the findings re-
meaningless materials such as nonsense syllables,
viewed here point toward an important and complex
group and individual recall do not differ. However,
role for sleep in the off-line reprocessing of learning
even with meaningful materials, collaborative memo-
and memory. The evidence is clearest for the role of ry typically falls short of performance predicted by
REM sleep in the consolidation of procedural learn- combining individual output (Clark and Stephenson,
ing. The possible roles of deep sleep in consolidating 1989). To determine if social interaction influences
procedural memories and of stage II NREM sleep in group recall, students of social memory have turned
consolidating motor skill learning are less clear, as is to comparisons between collaborative and nominal
sleeps role in consolidating and integrating declara- groups (see Figure 1).
tive memories. There is mixed evidence of roles for
both SWS and REM sleep in these processes. Taken
together with evidence for patterns of neuronal re- Collaboration in Recall
play during REM, NREM, and even sleep onset, a pic- In 1997, Weldon and Bellinger advocated testing
ture begins to emerge in which each stage of sleep nominal groups. In nominal groups participants actu-
makes a unique contribution to off-line memory really recall separately, and the sum of their nonover-
processing. Further work is needed to permit the un- lapping output is calculated. For example, if Tom,
equivocal identification of these contributions. Susan, and Hugh are members of a three-person
622 SOCIAL MEMORY PROCESSES
Figure 1 ration enhanced the recall of individual subjects but

reduced the output of the group as a whole.
Interpretations of Collaborative Inhibition

Contrary to popular intuition, collaborative inhi-
bition does not result from social loafing. (Social loaf-
ing refers to motor tasks such as tug-of-war in which
a person working with others exerts less effort than
a person working alone.) Weldon, Blair, and Huebsch
(2000) tested the effect of motivational variables on
collaborative inhibition. When they paid collaborative
groups to recall more words, collaborative-group re-
call still fell short of nominal-group recall. Basden,
Basden, Bryner, and Thomas (1997) proposed an al-
ternative interpretation of collaborative inhibition.
They argued that individual retrieval strategies are
disrupted during collaborative recall. For example,
suppose Susans retrieval strategy involved recalling
puma, lion, and tiger, in that order. Hearing Tom re-
call puma and Hugh recall orange leads Susan toward
recalling fruits and away from recalling lion and tiger,
which she would have normally retrieved after recall-
ing puma.
Several lines of evidence support the strategy-
disruption interpretation of collaborative inhibition.
For one, organization as measured by clustering (the
nominal group and both Tom and Susan recall the
tendency to recall exemplars of taxonomic categories
word puma, that word would be counted as being re-
together) is greater in the recall protocols of nominal
called only once. The advantage of using nominal
subjects than those of collaborative groups, and forc-
groups is that the effect of social interaction on group ing groups to organize their recall by category elimi-
productivity can be determined; the recall of n per- nates collaborative inhibition (Basden, Bryner, and
sons recalling collaboratively is compared to the re- Thomas, 1997). Furthermore, providing retrieval
call of n persons recalling individually. From the re- cues at the time of the test reduces or eliminates col-
sults of experiments that tested recall in three-person laborative inhibition, as does having group members
nominal and collaborative groups, Weldon and Bel- study items in the same rather than in different or-
linger as well as Basden, Basden, Bryner, and Thomas ders (Finlay, Hitch, and Meudell, 2000). Alternative
(1997) concluded that nominal groups recall more interpretations are that waiting ones turn to recall
material than collaborative groups, a phenomenon and/or coordinating ones efforts with others may
they label collaborative inhibition. Thus, social inter- block production (Weldon, Blair, and Huebsch, 2000)
action hurts rather than helps recall. To rephrase the and that the specificity of retrieval cues is greater for
adage: Two heads apart are better than two heads to- individual recall than for collaborative recallthat is,
gether. the encoding context originally experienced by the
individual is diluted during collaborative tests (An-
There is an interesting addendum to this re- dersson and Rnnberg, 1996).
search. Basden, Henry, and Basden (2001) gave both
collaborative and nominal groups a second individual
recall test. Understandably, the average recall of col- Errors and Social Contagion
laborating subjects was higher than that of nominal Group recall is usually more complete and more
subjects on this final individual test. After all, collabo- accurate than individual recall when groups discuss
rating subjects have the benefit of hearing the recall the events to which they were exposed and strive to
of others in their group on the initial recall test. When reach consensus before recording their joint recall
the nonoverlapping recall of the subjects in former (e.g., Yarmey and Morris, 1998). When recall is col-
nominal or collaborative groups was examined, for- lected without prior discussion and ensuing arrival at
mer nominal group subjects outperformed former consensus, collaborative groups may produce more
collaborative group subjects. In other words, collabo- errors than nominal groups. Basden, Basden, Thom-
SOCIAL MEMORY PROCESSES 623
as, and Soupasith (1998) tested false memory in each group size was greater for younger than for
groups by omitting the most common examples (e.g., older participants. In a second experiment, Dixon
apple), from lists made up of common taxonomic cate- and Gould tested story recall in older and younger
gories (e.g., fruit.) Collaborative groups falsely re- married couples. Somewhat surprisingly, older cou-
called more critical omitted items than did nominal ples did not differ from younger couples. Older cou-
groups. Furthermore, members of collaborative ples appear to profit more from transactive memory
groups who had not recalled a given critical item than do younger couples.
often did so on a subsequent individual-recall test, in-
dicating that memory errors may spread from one
group member to another.
Collaboration in Recognition Tests
In tests of recognition memory, groups outper-
Roediger, Meade, and Bergman (2001) studied
form individuals. As before, Two heads are better
the spread of memory errors within groups. After
than one. However, as with recall tests, it is impor-
briefly studying an everyday scene, such as a bath-
tant to know if group members effectively share infor-
room, two undergraduates alternated in recalling ob-
mation. Clark, Hori, Putnam, and Martin (2000)
jects from that scene. Unbeknownst to the true subject
found that both two- and three-person groups pro-
in the experiment, the other undergraduate was a
duced hits more often than individuals but that col-
confederate of the experimenter. The confederate re-
laboration did not reduce false-alarm rates. Clark et
called a nonpresented critical objecta toothbrush,
al. argued that a recall-to-reject strategy underlies
for examplein the course of recalling objects that
group superiority in hits. When a group member can
were actually present in the scene. On a subsequent
recall an items occurrence or the circumstances sur-
individual recall test, true subjects often falsely re-
rounding the items occurrence, he/she may convince
called the toothbrush, the critical object. Roediger
others that the item was actually presented. Accord-
and colleagues referred to the spread of false infor-
ing to Clark et al., collaboration facilitates recognition
mation from person to person as social contagion. In
performance beyond levels expected from simple
subsequent research, social contagion was induced by
rules such as majority wins or follow the leader.
an implied presencea bogus confederate, for exam-
As in collaborative-recall tests, performance on col-
ple. Social contagion is greater when the presentation
laborative-recognition tests may be influenced by the
is brief, when false information is introduced more
responses of others in the group. Schneider and Wat-
than once, and when a live rather than a simulated
kins (1996) found that true subjects who made their
confederate provides the misinformation.
recognition choices after false responses were given
by the experimenters confederate often conformed
Characteristics of Group Members to the confederates choices.
People who are familiar with the memory abilities
of others in their group may show less collaborative Collective Memory
inhibition than people who are unfamiliar with others As illustrated by collaborative memory of long-
in the group. Collaborative inhibition is less evident married couples, memory for events may be distribut-
in dyads composed of friends than in dyads comed among the members of a group (Wegner et al.,
posed of nonfriends (Andersson and Rnnberg, 1985). Acquiring a complete account of an eventa
1996). Johannsen, Andersson, and Rnnberg (2000) collective memorymay require obtaining contribu-
found that prospective memory in older couples was tions from all members of the group. To study memo-
influenced by reliance on one anothers memory. ry distribution in groups, Weldon (2000) proposed a
Couples who reported using transactive memory social-network analysis of collective memory. Howev-
(Wegner, Giuliano, and Hertel, 1985)knowledge of er, the concept of collective memory goes beyond the
their partners memoryshowed less collaborative idea that memory for events is dispersed among
inhibition than older couples who did not. To illus- members of a group. Memory both emerges from and
trate transactive memory, a husband may not attempt supports social interaction (Halbwachs, 1980). For ex-
to remember proper names that he knows his wife will ample, the growth of autobiographical memory in
remember, so he instead concentrates on remember- children depends upon collaborative recall of parent
ing dates and times. Dixon and Gould (1998) studied and child (Farrant and Reese, 2000); researchers
story recall in young and old participants (older than have studied collective memories of cultures by ob-
sixty-five) tested either individually, in two-person taining memory reports from people of different na-
groups, or in four-person groups. Collaboration had tionalities. A full understanding of collective memory
similar effects on the recall of younger and older par- may require interdisciplinary efforts.
ticipants. As group size increased, recall increased as
much for older as for younger participants. Recall at See also: COLLECTIVE MEMORY
624 SOMETIMES OPPONENT PROCESS (SOP) MODEL, IN CONDITIONING
Bibliography SOMETIMES OPPONENT PROCESS

Andersson, J., and Rnnberg, J. (1996). Collaboration and memo- (SOP) MODEL, IN CONDITIONING
ry: Effects of dyadic retrieval on different memory tasks. Ap-
plied Cognitive Psychology 10, 171181. Sometimes opponent process (SOP) is an associative,
Basden, B. H., Basden, D. R., Bryner, S., and Thomas, Robert L. real-time, quantitative theory of Pavlovian condition-
III (1997). A comparison of group and individual remembering. As such, it describes basic principles from which
ing: Does collaboration disrupt retrieval strategies? Journal of
the behavioral regularities of Pavlonian conditioning
Experimental Psychology: Learning, Memory, and Cognition 23,
1,1761,189.
can be deduced, and it makes predictions about yet-
Basden, B. H., Basden, D. R., Thomas, R. L., III, and Souphasith, to-be observed Pavlovian phenomena. It specifies
S. (1998). Memory distortions in collaborative recall. Current rules for stimulus representation, how learning oc-
Psychology 16, 225246. curs, and how learning that cannot be observed di-
Basden, B. H., Henry, S., and Basden, D. R. (2001). Costs and ben- rectly in translated into performance. This article
efits of collaborative remembering. Applied Cognitive Psychology does not present the equations that describe these
14, 497507.
principles, but they are available in related articles
Clark, N. K., and Stephenson, G. M. (1989). Group remembering.
In P. B. Paulus, ed., Psychology of Group Influence, 3rd edition.
(Mazur and Wagner, 1982; Wagner, 1981).
Clark, S. E., Hori, A., Putnam, A., and Martin, T. P. (2000).
Group collaboration in recognition memory. Journal of SOPs Basic Principles
Experimental Psychology: Learning, Memory, and Cognition 26, Theories of learning assume that experiences are
1,5781,588. recorded in a theoretical memory system. They de-
Dixon, R. A., and Gould, O. N. (1998). Younger and older adults
scribe how that memory system is conceptualized,
collaborating on retelling everyday stories. Applied Develop-
mental Science 2, 160171.
how experiences come to be represented, and how
Farrant, K., and Reese, E. (2000). Maternal style and childrens memories affect behavior. As for any theory, these
participation in reminiscing: Stepping stones in childrens au- basic principles are the assumptions the theory makes
tobiographical memory development. Journal of Cognition and from which the predictions will follow. A good theory
Development 1, 193225. strives to explain the observable phenomena through
Finlay, F., Hitch, G. J., and Meudell, P. R. (2000). Mutual inhibi- a priori assumptions that are as few and as simple as
tion in collaborative recall: Evidence for a retrieval-based ac-
possible.
count. Journal of Experimental Psychology: Learning, Memory, and
Cognition 26, 1,5561,567. SOP assumes that experiences activate corre-
Halbwachs, M. (1950; reprint 1980). The collective memory, trans. F. sponding theoretical representations in the memory
J. Ditter, Jr., and V. Y. Ditter. New York: Harper and Row. system. For example, a cat hears the sound of a can
Johansson, O., Andersson, J., and Rnnberg, J. (2000). Do elderly opener and then is fed; these events may activate cor-
couples have a better prospective memory than other elderly
responding sound and food representations. The cat
people when they collaborate? Applied Cognitive Psychology 14,
121133. does not have to learn how to represent these incom-
Roediger, H. L., Meade, M. L., and Bergman, E. T. (2001). Social ing events. What it does learn is how to associate one
contagion of memory. Psychological Bulletin and Review 8, 365 event with another; that is, the cat may learn to asso-
371. ciate the sound of the can opener with the food, as a
Schneider, D. M., and Watkins, M. J. (1996). Response conformity function of experiencing that sound and the food to-
in recognition testing. Psychonomic Bulletin & Review 3, 481 gether in time and space. As is common in psycholog-
483.
ical theories, SOP assumes that memoriesor, repre-
Wegner, D. M., Giuliano, T., and Hertel, P. T. (1985). Cognitive
interdependence in close relationships. In W. Ickes, ed., Com- sentationsbecome associated with each other as a
patible and incompatible. New York: Springer-Verlag. function of their temporal and spatial contiguity, and
Weldon, M. S. (2000). Remembering as a social process. In Medin, that what is learned are changes in connections (asso-
D. L., ed., Psychology of Learning and Motivation,Vol. 40. San ciations) among representations. SOP further pro-
Diego: Academic Press. poses that observable behavior reflects which repre-
Weldon, M. S., and Bellinger, K. D. (1997). Collective memory: sentations are being processed, and how strong their
Collaborative and individual processes in remembering. Jour-
processing is.
23, 1,1601,175. The central principle of SOP is that stimulus rep-
Weldon, M. S., Blair, C., and Huebsch, P. D. (2000). Group remem- resentation consists of a large but finite set of theoret-
bering: Does social loafing underlie collaborative inhibition? ical elements that are in one of three dynamic states.
Journal of Experimental Psychology: Learning, Memory, and Cogni-
This is illustrated in Figure 1a. To use the cat-food ex-
tion 26,1,5681,577.
Yarmey, A. D., and Morris, S. (1998). The effects of discussion on ample: When the cat encounters the food, elements
eyewitness memory. Journal of Applied Social Psychology 28, in the corresponding cat-food representation are acti-
1,6371,648. vated. The course of activation is determined: there
is a brief period of focal processing during which the
Barbara H. Basden elements are in a state designated A1. These elements
SOMETIMES OPPONENT PROCESS (SOP) MODEL, IN CONDITIONING 625
do not stay long in this focal activation state, however, Figure 1

but decay passively into a secondary, more peripheral
state designated A2. Eventually, when the food is no
longer present, the elements passively decay further
into inactivity (I). How long an element might be in
each dynamic state depends on three parameters: p1,
which reflects the perceived salience of the incoming
stimulus and is assumed to increase monotonically
with the intensity of the stimulus; pd1, which deter-
mines how likely an element is to go from A1 to A2;
and pd2, which determines how likely an element is to
go from A2 to I. It is assumed that pd1pd2.
Stimulus Representation
Figure 2 depicts the application of these rules for
elemental dynamics to the instance of exposure to a
stimulus that occurs in real time. The figure shows the
proportion of total stimulus representation elements
that are active across time. In each moment in which
the stimulus is presented, p1 of the inactive elements
are activated to the A1 state. However, pd1 of those
elements in A1 also decay to A2. The result is an A1
function that shows an increase followed by a decrease
and relative flattening, until the stimulus is withdrawn
and there is decay to inactivity. The form of the A1
function, which is SOPs stimulus representation, is
similar to that which has been observed in psycho-
physical experiments with human subjects who are
asked to estimate the perceived intensity of a stimulus
(e.g., light, sound, or smell) across time. They report
an initial growth in the intensity of sensation followed
by a decreasean adaptation to the stimulusbefore
it is turned off (Marks, 1974).
Rules for Learning

In SOP, learningthe acquisition of associations
between stimulus representationsis determined by
the conjunction of active A1 and A2 elements. If the (a) A depiction of a node in the memory system of SOP,
presumed to be activated by a peripheral stimulus in a
sound of the can opener is considered a conditioned
Pavlovian conditioning situation. The connected circles within
stimulus (CS), and the food an unconditioned stimu-
each node represent the three activity states in which nodal
lus (US), then an excitatory association between that elements may reside, and the connecting paths show the
CS and US will come about to the extent that the allowable state transitions. (b) The dynamic states (circles) and
elements that represent the CS and the elements the allowable course of activation (paths) for an SOP node that
that represent the US are currently in the A1 state. is activated by another, associated node.
The likelihood of such concurrent CSA1 and USA1
activation increases with the degree of contiguity
of the two events. An excitatory association is one
that may activate the representations that are linked; tween a CS and US is established, a CS may activate
hearing the opener now may elicit a memory of its associated US representation to its A2 processing
food. state, i.e., USA2. This is shown in Figure 1b as the di-
SOP captures the difference between the experi- rect link from I to A2. The likelihood of such activa-
ence of an event and its memory by assuming that as- tion is determined by the parameter p2, which is a
sociatively activated representations reach only the function of the strength of the CS trace and the
A2 processing level. Thus, once an association be- strength of the association.
626 SOMETIMES OPPONENT PROCESS (SOP) MODEL, IN CONDITIONING
Figure 2
The proportion of total elements that are in A1 and A2 across time with the presentation of a peripheral stimulus.
Associative strength can also be inhibitory, where ymptotic, and an initial extinction trial (Figure 3b).
an inhibitory association may make it harder for one On Acquisition, Trial 1, there is a considerable over-
representation to activate another to which it is lap of CSA1 and USA1, along with smaller overlap of
linked. Inhibitory associative strength grows when the CSA1 and USA2. (There is overlap of CSA2 and
elements of one representation are in A1 and ele- USA1and USA2 also, but the associations thus formed
ments of another are in A2. An effective way to pro- are presumed to have little behavioral effect.) The
duce an inhibitory CS-US association is to present the contiguous CSA1/USA1 processing produces a relative-
US first and then the CS. This should be effective to ly large increase in V+, and the smaller CSA1 and USA2
the extent that it ensures the concurrent processing overlap produces relatively little V-, resulting in a rel-
of CSA1 with USA2. Excitatory tendencies (V+) and in- atively large gain in VCS-US. After many pairings
hibitory tendencies (V-) are assumed to summate in a (Acquisition, Trial 50), the onset of the CS elicits
single associative connection (V) that may have a net USA2 processing as a function of the associative con-
excitatory or net inhibitory strength. nection that is established; that is, the CS now re-
sults in a CR (recall that the CR reflects USA2 pro-
Rules for Performance cessing). There also is a reduction in USA1 when
the US is presented because of the effect of the CS
The performance rule in SOP follows from its as-
to put elements into USA2, making these elements
sumptions about dynamic states and the nature of the
unavailable for activation to USA1. This reduction
learned associations: a conditioned response (CR) is
in USA1 processing can be seen in a reduced un-
functionally the same as the behavior that is elicited
by the A2 processing state of the stimulus with which conditioned response (UR) to the US, where the
it is associated. The behavior that is identified as a primary UR is assumed to reflect USA1 processing.
CR, therefore, can be elicited by presenting the asso- Finally, because the US is not present until the
ciated US and looking for its secondary (A2) re- end of the CS period, there also is considerable
sponse. overlap of CSA1 and USA2, which results in an in-
crement in V-. At asymptote, the V+ and V- cancel
each other out, and there is no net change in associa-
Acquisition and Extinction tive strength. The negatively accelerated course of
Figure 3a shows the expected change in net asso- learning, the elicitation of a CR by a CS, and the so-
ciative strength and the corresponding stimulus called conditioned diminution of the UR to that US,
traces and increments in V+ and V- for a first CS-US are fundamental behavioral characteristics of Pavlov-
pairing, a fiftieth CS-US pairing when learning is as- ian conditioning.
SOMETIMES OPPONENT PROCESS (SOP) MODEL, IN CONDITIONING 627
Figure 3
(a) Predicted VCS-US for a series of acquisition trials followed by a series of extinction trials. (b) Simulated activity in CS and US nodes
and resulting changes in V+CS-US, V-CS-US, and VCS-US are shown for the first acquisition trial, the fiftieth acquisition trial, and the first
extinction trial.
Extinction occurs (Extinction, Trial 1) with CS- Cue-Competition Effects

alone trials. Now, because the CS has the capacity to SOP explains cue-competition effects in Pavlov-
elicit USA2 processing, there is an increment in V- and ian conditioning with the same principles as those
no counteracting increment in V+, since the US is not that explain acquisition and extinction. One powerful
presented. With continued presentations of the CS cue competition effect is blocking, a condition in
alone, there is a gradual decrease in VCS-US and less which learning a CS-US association fails in spite of the
and less of a tendency for the CR to be elicited. The good temporal contiguity of the two stimuli. For ex-
latter also is a behavioral characteristic of Pavlovian ample, suppose that the cat of our previous example
conditioning. had previously learned to associate the turning on of
628 SOURCE MONITORING
the kitchen light with food and that only then did we BLOCKING EFFECT: NEURONAL SUBSTRATES;
present the sound of the can opener and the light to- LEARNING THEORY: A HISTORY; LEARNING
gether preceding the food. Now the cat would be un- THEORY: CURRENT STATUS; MATHEMATICAL
likely to learn to make a conditioned response to the LEARNING THEORY; NEURAL COMPUTATION:
sound. It appears that previously learning to associate APPROACHES TO LEARNING
the light with food makes the normally effective pair- Bibliography
ing of sound and food ineffective for association. Brandon, S. E., Vogel, E. H., and Wagner, A. R. (2002). Computa-
The blocking effect can be understood with refer- tional theories of classical conditioning. In J. W. Moore, ed.,
A neuroscientists guide to classical conditioning. London: Cam-
ence to the functions in Figure 3 for Acquisition Trial
bridge University Press.
50. Here, the previously trained CS (the light) regu- Brandon, S. E., and Wagner, A. R. (1998). Occasion setting: Influ-
larly elicits a CR, reflecting the USA2 processing elicit- ences of conditioned emotional responses and configural
ed by that CS. The addition of the novel CS (the cues. In N. Schmajuk and P. Holland, eds., Occasion setting: As-
sound), will result in the overlap of CSA1-sound with the sociative learning and cognition in animals. Washington, DC:
American Psychological Association
same USA1 and subsequent USA2 processing that is
Marks, L. E. (1964). Sensory processes. New York: Academic Press.
elicited already by the conjunction of CSA1-light and Mazur, J. E., and Wagner, A. R. (1982). An episodic model of asso-
the US. That is, there will be an initial increment in ciative learning, in M. L. Commons, R. J. Herrnstein, and A.
V-, followed by a comparable increment in V+, that R. Wagner, eds., Quantitative analyses of behavior, Vol. 3: Acqui-
will accrue equally for both CSs. Since V- and V+ bal- sition. Cambridge, MA: Ballinger.
Wagner, A. R. (1981). SOP: A model of automatic memory process-
ance out, there is no net increment in either associa-
ing in animal behavior. In N. E. Spear and R. R. Miller, eds.,
tion. The CSsound US association, starting at 0, thereby Information processing in animals: Memory mechanisms. Hillsdale,
fails to acquire any V. That is, the cat will not learn NJ: Erlbaum.
a sound-food association. Wagner, A. R., and Brandon, S. E. (1989). Evolution of a structured
connectionist model of Pavlovian conditioning (AESOP). In S.
B. Klein and R. R. Mowrer, eds., Contemporay learning theories:
Conclusion Pavlovian conditioning and the status of traditional learning theory.
SOP is able to explain much of what is known (2001). A componential theory of Pavlovian conditioning.
about Pavlovian conditioning, including cue- In R. R. Mowrer and S. B. Klein, eds., Handbook of contemporary
competition effects. By rationalizing excitatory and Learning Theory. Mahway, NJ: Erlbaum.
inhibitory learning in terms of A1/A1 and A1/A2 con- Wagner, A. R., and Donegan, N. H. (1989). Some relationships be-
tween a computational model (SOP) and an essential neural
junctions in time, it allows us to understand how the circuit for Pavlovian (rabbit eyeblink) conditioning. In R. D.
outcome of a conditioning trial depends on the order Hawkins and G. H. Bower, eds., The psychology lf learning and
of the CS and US and the difference between simple motivation, Vol. 23: Computational models of learning in simple
conditioning and conditioned inhibition training. By neural systems. Orlando, FL: Academic Press.
rationalizing priming effects in terms of A1 and A2 ef- Susan E. Brandon
fects over time, it allows us to understand blocking (as Allan R. Wagner
shown above), as well as short-term habituation and
pre-trial CS and US exposure effects. By rationalizing
the relationship of a CR and UR as the CR reflecting
only USA2 processing, without the USA1 processing
additionally elicited by the US, we can understand SOURCE MONITORING
why the CR sometimes mimics the UR and sometimes Source monitoring refers to cognitive processes in-
does not. volved in making attributions about the origins of
Extensions of the SOP model have been devel- mental experiences; for example, attributing a men-
oped to increase its theoretical power, to allow for an tal experience to something dreamed, something
understanding of occasion setting (Brandon and imagined, or a perceived event. The concept of
Wagner, 1998), CR timing (Brandon, Vogel, and source memory overlaps with, but is more general
Wagner, 2002), and various differences in Pavlovian than, the idea of memory for context. Source moni-
conditioning involving skeletal versus emotional re- toring is an important aspect of everyday cognition,
sponses (Wagner and Brandon, 1989). Wagner and for example, in deciding whether one took ones
Donegan (1989) have further indicated how it may re- medication or just thought about taking it, read about
late to the known neural circuitry for eyeblink condi- a space alien invasion in a tabloid or a news magazine,
tioning. or really saw the defendant at the crime scene with a
knife or just heard about the knife later. Errors in
See also: ALGORITHMS, LEARNING; CONDITIONING, source monitoring range from the trivial (telling a
CELLULAR AND NETWORK SCHEMES FOR joke to the same person you heard it from) to the
HIGHER-ORDER FEATURES OF; KAMINS egregious (mistaking a memory of a dream of being
SOURCE MONITORING 629
sexually abused for a memory of a real event from talking to Paul at a party might be contradicted by re-
childhood). trieving additional information that places Frank out
of town at the time of the party. Similar distinctions
Marcia Johnson and her colleagues (1993) have
between relatively automatic and more controlled
detailed a theoretical framework for understanding
processes of remembering have been made by L.
the cognitive processes and factors that influence
Hasher and R. Zacks (1979), L. Jacoby and C. Kelley
source memory. According to the source monitoring
(1989), and other researchers. Both heuristic and sys-
framework, any given mental experience typically
tematic source attributions are affected by a remem-
does not include a single feature or tag or label speci-
berers biases, goals, agendas, and meta-memory be-
fying what it is (e.g., a memory of a dream, an imagi-
liefs. For example, one will usually engage more
nation, a perception). Rather, people attribute some
systematic source monitoring processes if the cost of
mental experiences to memories based on the experi-
a mistake is high, but engage only relatively automat-
ences features. Events have many features (objects,
ic, heuristic processes for most everyday remember-
location, people, color, taste, emotions, ongoing
ing.
thoughts), some of which are encoded in memory; a
few or many of these features may be brought to mind
(reactivated) after only a few minutes or years later. Historical Context
What a person calls that later mental experience de-
pends on what features it includes and on the persons The concept of reality monitoring was introduced
beliefs about the differences between mental experi- in the early 1980s by M. K. Johnson and C. L. Raye
ences from different sources. For example, people (1981) to explain how memories for internal events
usually expect memories for events (sometimes called (e.g., thoughts, imaginations) are discriminated from
episodic memory) to contain details reflecting such memories for external, perceived events, and why
aspects as the who, what, when, where, why, and how they are sometimes confused. This concept was sub-
of the event. A mental experience that does not have sumed by the more general source monitoring frame-
such details might be attributed to, for example, in- work in the early 1990s. The theoretical ideas incor-
ference, general knowledge, or belief, depending on porated in the source monitoring framework were
the particular features it does have. proposed to help organize and understand diverse
findings and guide additional research. For example,
Different types of encoding processes (e.g., see- studies in the 1950s and 1960s showed that people
ing, hearing, thinking, dreaming) and different types falsely recall (Deese, 1959) or falsely recognize (Un-
of events (e.g., movie, telephone call) tend to produce derwood, 1965) associates of presented words: Hear-
memorial representations that are characteristically ing thread, haystack, sharp, and so on, can lead people
different from each other. For example, memories of to misremember hearing needle, presumably because
imagined events typically have less vivid perceptual, they thought of needle during study and later mistake
temporal, and spatial information than perceived the thought for an actual presentation of the word. In
events and more information about intentional cogni- the 1970s, M. K. Johnson and J. D. Bransford and col-
tive operations (e.g., actively generating images while leagues (1973) showed that people falsely recognize
thinking). Therefore, if a mental experience had sub- ideas that were only implied in sentences. For exam-
stantial perceptual detail, one would tend to attribute ple, after hearing, The man dropped the delicate glass
it to a perceived event (e.g., something one saw). pitcher on the floor people often remember hear-
However, there is variability among memories from ing, The man broke the delicate glass pitcher on the
any particular source, and the distributions of fea- floor. The 1970s and 1980s produced many studies
tures from different sources overlap. For example, showing that peoples memory for experiences tends
some dreams are more vivid or plausible than some to be shaped by their expectations or schemas (see
waking events. Thus, remembering always involves Alba and Hasher, 1983, for a review). For example,
evaluating the quality and quantity of activated char- W. F. Brewer and J. C. Treyens (1981) showed that
acteristics in light of expectations about typical char- people who had briefly waited in an office were likely
acteristics of mental experiences from various to falsely remember items such as books, which were
sources. not in the office but might be expected to be, and to
Source attributions are often made rapidly and not remember unexpected items that were there (e.g.,
without deliberation based on heuristic judgments a skull). E. F. Loftus and colleagues (1978) showed
about activated features. However, source monitoring that information introduced when people were ques-
sometimes entails more systematic processes that are tioned about an event was later sometimes
typically slower and more deliberate, including re- (mis)remembered as part of the original event.
trieving additional information, extended reasoning, Such findings illustrate that people confuse infor-
and so on. For example, a vivid memory of Frank mation from different sources. For example, as part
630 SOURCE MONITORING
of their normal comprehension processes, people Brain Regions Involved in Source

think of related information during encoding or re- Monitoring
membering (or both) and misattribute this informa-
Neuroimaging data (e.g., from functional mag-
tion to the actual event. Other times, they confuse
netic resonance imaging) together with neuropsy-
what they saw with what they heard or read, or con-
chological studies of brain damaged patients with am-
fuse two similar experiences. Yet, sometimes memory
nesia indicate that the hippocampus plays a central
is quite accurate. The source monitoring framework
role in the binding of features into complex represen-
specifies the factors that influence the likelihood that
tationsa process critical for later source monitoring.
memory will be accurate or distorted.
Profound disruptions in source monitoring, such as
delusions, hallucinations, and confabulations, can
Factors Affecting Source Monitoring arise from damage to frontal brain regions, indicating
Source monitoring depends on the type, amount, that these regions are critical for source monitoring.
and quality of activated information, the extent to Neuroimaging studies that show activation of frontal
which the active information helps uniquely specify regions during source monitoring in healthy individ-
the source, the judgment processes engaged, the uals provide converging evidence. Children and
weights assigned to different features, and the criteria older adults have more difficulty with source monitor-
used when making the source attribution. Neither the ing than do college-aged adults, particularly as the
activated features (representations) nor the processes similarity of the sources increases. Researchers have
that act on them are perfect, and thus errors occur. suggested that such findings may reflect the relatively
A basic tenet of the source monitoring framework is late maturation of frontal functions in children and
that inaccurate source monitoring (sometimes called the increased probability of pathology in frontal re-
source confusions, source misattributions, source er- gions associated with aging. One goal of current
rors, source amnesia, source forgetting, memory dis- neuroimaging work is to more clearly delineate the
tortions, or false memories) and accurate source mon- brain circuits underlying the encoding, revival, and
itoring arise via the same mechanisms. evaluation of memories.
Anything that disrupts the encoding, consolida-
tion, or retention of the features of events will nega- Bibliography
tively affect source monitoring. For example, at en- Alba, J. W., and Hasher, L. (1983). Is memory schematic? Psycholog-
coding, divided attention or focusing on ones own ical Bulletin 93, 203231.
Belli, R. F., and Loftus, E. F. (1994). Recovered memories of child-
emotions rather than event details can increase hood abuse: A source monitoring perspective. In S. J. Lynn
source monitoring errors, presumably because useful and J. W. Rhue, eds., Dissociation: Clinical and theoretical per-
source-specifying information fails to be, or is weakly, spectives. New York: Guilford Press.
bound to other features of the event. Errors increase Brewer, W. F., and Treyens, J. C. (1981). Role of schemata in mem-
when the diagnosticity of available source informa- ory for places. Cognitive Psychology 13, 207230.
Deese, J. (1959). On the prediction of occurrence of particular ver-
tion is reduced, for example, when semantic or per- bal intrusions in immediate recall. Journal of Experimental Psy-
ceptual similarity between events from different chology 58, 1722.
sources is increased. Errors also increase when more Gardiner, J. M., and Richardson-Klavehn, A. (2000). Remember-
lax criteria are used to evaluate mental experiences, ing and knowing. In E. Tulving and F. I. M. Craik, eds., The
features are weighted inappropriately, attention is di- Oxford handbook of memory. New York: Oxford University Press.
Hasher, L., and Zacks, R. T. (1979). Automatic and effortful pro-
vided at test, or the time that is available to make a cesses in memory. Journal of Experimental Psychology: General
source judgment is limited. Individual motives and 108, 356388.
the social/cultural context can influence all of these Hintzman, D. L. (2000). Memory judgments. In E. Tulving and F.I.
factors. M. Craik, eds., The Oxford handbook of memory. New York: Ox-
The general view that remembering is not a sim- Jacoby, L. L., Kelley, C. M., and Dywan, J. (1989). Memory attribu-
ple matter of retrieving memory traces but rather tions. In H. L. Roediger, III, and F. I. M. Craik, eds., Varieties
a subjective experience with phenomenal qualities of memory and consciousness: Essays in honour of Endel Tulving.
that differ in important ways has generated new inter-
Johnson, M. K. (1997). Source monitoring and memory distortion.
est in assessing the subjective qualities of memories. Philosophical Transactions of the Royal Society of London 352,
One approach asks people to distinguish between 1,7331,745.
items they know and items they remember; another uses Johnson, M. K., Bransford, J. D., and Solomon, S. K. (1973). Mem-
memory characteristics questionnaires to elicit more ory for tacit implications of sentences. Journal of Experimental
detailed ratings of features of memories. For exam-
Johnson, M. K., Hashtroudi, S., and Lindsay, D. S. (1993). Source
ple, such studies have shown that, on average, false monitoring. Psychological Bulletin 114, 328.
memories tend to be rated as having less perceptual Johnson, M. K., Hayes, S. M., DEsposito, M., and Raye, C. L.
detail than true memories. (2000). Confabulation. In J. Grafman and F. Boller, eds.,
SPATIAL LEARNING: Animals 631
Handbook of neuropsychology, 2nd edition. Amsterdam, Nether- on sites containing the other untasted and so more
lands: Elservier Science. appealing fooda result that implies that birds know
Johnson, M. K., and Raye, C. L. (1981). Reality monitoring. Psycho-
where they have hidden particular food items. By
(1998). False memories and confabulation. Trends in Cogni- using food that rapidly rots, the same experimenters
tive Sciences 2, 137145. showed that scrub jays also know when they made the
Loftus, E. F., Miller, D. G., and Burns, H. J. (1978). Semantic inte- cache. Lastly, birds remember that they have emptied
gration of verbal information into a visual memory. Journal of a cache, and they avoid revisiting empty sites. Other
Experimental Psychology: Human Learning and Memory 4, 1931.
resources that animals remember include watering
Mitchell, K. J., and Johnson, M. K. (2000). Source monitoring: At-
tributing mental experiences. In E. Tulving and F. I. M. sites, nests, places where mates are to be encountered,
Craik, eds., The Oxford handbook of memory. New York: Oxford shelters, and bolt holes. Spatial knowledge is closely
University Press. integrated with other forms of knowledge that may
Roberts, K. P., and Blades, M., eds. (2000). Childrens source monitor- influence what spatial information is stored, when it
ing. Mahwah, NJ: Erlbaum.
Roediger, H. L., III (1996). Memory illusions. Journal of Memory and
is retrieved, and how long it is remembered.
Language 35, 76100. Some animals have evolved set procedures for ac-
Schacter, D. L., Norman, K. A., and Koutstaal, W. (1998). The cog-
nitive neuroscience of constructive memory. Annual Review of
quiring needed navigational information. Indigo
Psychology 49, 289318. buntings learn the constellations of stars around the
Spencer, W. D., and Raz, N. (1995). Differential effects of aging on North Star, and they use that constellation for guid-
memory for content and context: A meta-analysis. Psychology ance in their migration. The birds are preprogram-
and Aging 10, 527539. med to learn the unique pattern of stationary stars
Underwood, B. J. (1965). False recognition produced by implicit
verbal responses. Journal of Experimental Psychology 70, 122
that lies close to the Earths axis of rotation and so do
129. not move across the night sky. Wasps and bees per-
form highly structured learning or orientation flights
Marcia K. Johnson
when first leaving a new feeding site or their nest. The
Karen J. Mitchell
flight is designed to emphasise landmarks that are
close to the goal and that can thus provide precise
navigational information. In a single such flight, they
SPACED TRAINING learn enough about the arrangement of landmarks to
be able to return there. Rats explore a new environ-
See: DISTRIBUTED PRACTICE EFFECTS ment and reexplore a familiar one when changes
occur. During exploration they learn the layout of the
environment without the benefit of any immediate re-
ward. In one experiment to demonstrate such learn-
SPATIAL LEARNING: ANIMALS ing, rats explored a T-maze with two visually distinct
Resources that animals need are usually distributed goal boxes and with all extra-maze cues screened off.
patchily within their home range, and many animals After the rats had explored the empty maze several
learn where they are and how to reach them. Stuart times, they were placed singly in one of the goal boxes
Altman describes how one troop of baboons respond- and allowed to eat there. When re-placed at the en-
ed to ripe berries on an isolated bush in the center of trance, most rats returned directly to the box where
their home range as a sign of their availability else- they had been fed. A control group fed in one goal
where and trekked to a remote patch of bushes bear- box without prior exploration of that particular maze
ing the same fruit. showed no tendency to return to the same box. Explo-
ration allows the rats to learn the paths to different
Some animals cache food when it is abundant, re-
recognized locations that only later come to be associ-
member the precise locations of the caches for long
ated with a valued resource.
periods, and return to empty the caches when food is
scarce. Clarks nutcracker provides a dramatic exam- It is remarkable that a wide array of animals, from
ple. The birds collect tens of thousands of pine seeds insects to primates, acquire and store the same two
in the autumn for recovery during the subsequent distinct types of spatial information. One kind is de-
winter and spring. Scrub jays caching food in the lab- rived from dead reckoning, also known as path inte-
oratory remember not only where they have cached gration. An animal leaving its home base continuous-
it but also which items they have stored in which loca- ly monitors the direction and distance of the path that
tions. Proof came from an experiment by Nicola Clay- it takes and uses this path-derived information to
ton and Anthony Dickinson in which birds were given keep an updated record of its direction and distance
two different foods to cache. Before the birds were al- from home. Consequently, it is always able to take a
lowed to recover either food, they were prefed with direct path home, even after a circuitous outward
one. Prefeeding caused the birds to focus their search journey. On finding a good source of food, both in-
632 SPATIAL LEARNING: Animals
sects and mammals store the path integration coordi- landmarks. Consequently the distant panorama can
nates of the food site. Equipped with these stored co- be recognized more reliably than the local landmarks.
ordinates, they can later return to that site by Insects exploit this piece of ecological geometry and
subtracting the coordinates of the site from their cur- use their memory of the panorama to prime the recall
rent coordinates as given by path integration. of local landmarks, which can thus be identified when
The second kind of stored spatial information viewed from unusual viewpoints or under unusual
does not have positional coordinates attached to it. It lighting conditions.
comes from memories of visual landmarks that can To reach a desired goal animals either follow fa-
indicate a site or a route or signal what kind of action miliar paths or, more rarely, plan a new route from
the animal ought to perform there. It is still unclear an arbitrary starting point to a goal that is not visible
whether there exists in any animal an intimate con- from the starting point. Evidence for such route plan-
nection between memories of landmarks and their ning comes from a study by Charles Menzel and col-
positional coordinates. In insects the available evi- leagues on a young Bonobo chimpanzee, Kanzi.
dence suggests that memories of landmarks and of Kanzi could be told through lexigrams which out of
positional coordinates are kept separate. several goals in a familiar wooded area he should
Some insects, birds, and possibly mammals learn head for. The chimpanzee was led to an arbitrary
the position of a site in terms of the distance and di- starting point and then given a lexigram signaling a
rection of one or more visual landmarks from it. They familiar location that was out of sight and that har-
then return to the site by moving until their current bored food or some other desirable item. On all occa-
distances and directions from those landmarks match sions Kanzi led a human companion to the correct lo-
the stored ones. Precision in locating the site is en- cation. Sometimes Kanzi took direct trails, and
hanced by learning the distances and directions of sometimes he followed a more indirect path. There
several landmarks and by emphasizing information is as yet little understanding of the behavioral mecha-
that comes from landmarks that are nearer to the site. nisms underlying route planning.
It is likely that mammals also learn the spatial rela- Although ants and bees can reach a familiar feed-
tionships between landmarks, but experimental cor- ing site solely by means of path integration, they no-
roboration has proved elusive. tice and approach prominent objects on the route
The richness of animals memories of the ar- and rapidly come to learn the appearance of these
rangement of landmarks is illustrated by an experi- landmarks. The landmarks are used to segment the
ment by David Brodbeck on chickadees. It showed route, and insects learn vectors of the appropriate di-
that these birds learn a caching site in terms of several rection and distance to take them from one landmark
potential retrieval cues. Birds were trained to find to the next. The division of a route into sections in-
seeds in one wooden block in an array of four differ- creases the insects chances of reaching its goal. First,
ently decorated blocks that were attached to a wall in prominent beacons that attract the insect at the end
a large aviary. The site is thus specified by room cues, of each segment make it easy to correct for inaccurate
position in the array, and the appearance of the bait- vectors or crosswinds that deflect the insect from its
ed block. On test trials the array was shifted as a proper path. Second, the precision with which the in-
group, so that one block was moved to the location oc- sect is delivered to its goal depends on the accuracy
cupied by the baited block, and the other blocks in the of the vector from the final beacon rather than on a
array were rearranged. Birds in these tests looked for vector covering the whole path from start to finish.
their seeds following a consistent order. They first Again we see that animals are naturally biased to learn
searched at the site that was specified by the land- features of their environment that improve naviga-
marks in the room, then at the site defined by positional accuracy. Spatial learning is strongly guided by
tion within the array of blocks, and lastly in the block built-in mechanisms and strategies that predispose
that was correctly decorated. Birds had thus learned animals to learn about navigationally useful features
all these properties of the caching site and were guid- of their environment, often in anticipation of their
ed by them in a set hierarchy. need for later guidance.
A similar example of a predisposition to learn

See also: FORAGING; MIGRATION, NAVIGATION, AND
multiple features of the surroundings of a significant HOMING; PLACE CELLS; SPATIAL MEMORY
site comes from insects. Honeybees learn both the
local landmarks that pinpoint a site and the panoram-
Bibliography
ic context within which the local landmarks are set. As
Healy, S. D., ed. (1998). Spatial representation in animals. Oxford:
a bee flies within the vicinity of the site, the appear- Oxford University Press.
ance of the distant panorama changes less with the Pearce, J. M. (1997). Animal learning and cognition, 2nd edition.
bees movements than does the appearance of local Hove, UK: Psychology Press.
SPATIAL MEMORY 633
Shettleworth, S. J. (1998). Cognition, evolution, and behavior. New Figure 1

Lynn Nadel
Revised by Thomas S. Collett
SPATIAL MEMORY
Spatial memory pertains to the spatial structure of
outdoor spaces, buildings, rooms, and maps; it in-
cludes knowledge of where objects are, of routes from
place to place, and of distances and directions be-
tween locations. The evolutionary success of humans
has depended, in part, on abilities to navigate in unfa-
miliar territory, to locate sources of food and water,
and to be able to return to those sources and to home
at a later time. In contemporary societies, people rely
on their spatial memories for activities as mundane as
reaching out in the morning darkness to shut off an
alarm and as consequential as escaping from an office
building during a raging fire. The spatial memories
of insects, rodents, and nonhuman primates have
been investigated extensively (Gallistel, 1990), but
this chapter will focus on aspects of human spatial
memory.
An example of the types of layouts used to investigate the
orientation dependence of spatial memory. Objects are placed
Route versus Survey Knowledge on the floor of a large room. Participants memorize the layout
from one or more study views, and then make judgments of
Large-scale environments cannot be viewed in relative direction using their memories (e.g., Imagine you are
their entirety from a single vantage point and there- standing at the clock and facing the shoe. Point to the skillet).
fore can only be learned via navigation. When people The arrow indicates a possible viewing position.
learn a large-scale environment without the aid of a
map, their knowledge initially consists of routes from
place to place. With sufficient experience, people may at the figure: (a) Imagine you are standing at the book
learn the straight-line (Euclidean) distances and di- and facing the wood. Point to the clock. (b) Imagine
rections between locations and perhaps gain maplike you are standing at the wood and facing the shoe.
or survey knowledge of the environment (Thorndyke Point to the book. Problem (a) is easier than (b), even
and Hayes Roth, 1982). At one time researchers
though the pointing direction is the same. This phe-
thought that route knowledge had to precede survey
nomenon is an example of orientation dependence,
knowledge (Siegel and White, 1975). Recent studies
and it is observed in many tasks and in memories of
have shown that the type of knowledge acquired de-
maps, rooms, and large-scale environments, even
pends on the goals of the learner, irrespective of the
learning experience. For example, people perform when several views are learned (Shelton and McNa-
better on survey knowledge tasks (e.g., drawing a mara, 2001). Although there may be situations in
map) when their goal is to learn the overall layout of which spatial memories are orientation independent,
an environment than when their goal is to learn such that familiar and unfamiliar views are retrieved
routes, regardless of whether they learn the environ- and recognized equally efficiently (Evans and Pezdek,
ment from a map or from navigation (Taylor, Naylor, 1980), these situations are the exception.
and Chechile, 1999).
Orientation dependence has typically taken the
form of better performance on familiar views and per-
Orientation Dependence spectives, as in (a), than on unfamiliar views and per-
An important property of spatial memories is that spectives, as in (b), but other patterns have also been
information is easier to retrieve from some perspec- observed (Mou and McNamara, 2002). An important
tives than from others. Commit Figure 1 to memory. consequence of the orientation dependence of spatial
Then make the following judgments without looking memories is that people sometimes get lost when re-
634 SPATIAL MEMORY
turning from a destination. The environment looks Hierarchical Representations

different coming and going.
Spatial knowledge is stored in the brain hierar-
chically (Stevens and Coupe, 1978). When people
Spatial Reference Systems learn the locations of objects in an environment,
they group the objects into ever larger clusters. For
The concept of location is inherently relative. example, a hierarchical representation of an office
One cannot describe or specify the location of an ob- may specify that the telephone and coffee cup are
ject without providing a frame of reference. The loca-
on the desk, that the desk is next to the chair, and
tion of a chair in a classroom, for example, can be
that the desk and chair are in the office. Objects
specified in terms of the room itself (e.g., the chair is
are grouped on the basis of their properties (e.g., the
in the corner by the door), other chairs in the room
functional relation between a chair and a desk), as-
(e.g., the chair is in the first row and second column),
or an observer (e.g., the chair is in front of me). The pects of the environment (e.g., barriers), and even or-
primate brain represents spatial information using ganizational strategies unique to a particular person
many types of spatial reference systems (Anderson, (McNamara, 1986; McNamara, Hardy, and Hirtle,
Snyder, Bradley, and Xiang, 1997). 1989).
Spatial reference systems fall into two categories: One way to interpret these findings is that people
egocentric reference systems, which specify location represent spatial relations in locally defined reference
and orientation with respect to the observer (as in, the systems and these reference systems are then related
chair is in front of me); and environmental reference to one another in higher-order reference systems
systems, which define spatial relations with respect to (Poucet, 1993). For example, locations of objects in a
elements of the environment, such as the perceived room might be represented by a reference system
direction of gravity, landmarks, or the floor, ceiling, local to the room. Such a reference system could then
and walls of a room (as in, the chair is in the corner serve as an object in a reference system defining the
by the door). spatial relations among the rooms on the same floor
Human spatial memories rely primarily on envi- of a house; these could then serve as objects in a ref-
ronmental reference systems (Shelton and McNa- erence system relating floors of the house to one an-
mara, 2001). When people learn a map, locations of other.
objects in a room, or even larger spaces, they select
a reference system in the environment for represent-
ing its spatial structure. The particular reference sys- Spatial Memory and the Brain
tem that is selected depends on the persons experi- Functional neuroimaging using PET (positron
ences, the structure of the environment itself, and emission tomography) and fMRI (functional magnet-
spatial and nonspatial properties of objects in the en- ic resonance imaging) are proving to be powerful
vironment. This process is similar to determining the methods for investigating areas of the brain involved
top of a figure or an object (Rock, 1973; Tversky, in spatial learning and navigation. This research is
1981); in effect, conceptual north (and perhaps,
new, but consistent findings are emerging. The para-
south, east, and west) is assigned to the layout, creat-
hippocampal gyrus seems to be involved in a variety
ing privileged directions in memory of the environ-
of spatial tasks (Epstein and Kanwisher, 1998). The
ment (conceptual north need not correspond to true
hippocampus, predominantly on the right, is associ-
or magnetic north or any other cardinal direction).
ated with the recollection of familiar routes (Maguire,
Retrieval of spatial relations is more efficient in direc-
tions aligned with this reference axis than in other di- Frackowiak, and Frith, 1997). Dorsal areas of the
rections. In Figure 1, for instance, an observer view- brain (e.g., the posterior parietal cortex) seem to be
ing the space from the position of the arrow might use recruited for processing the locations of objects,
the axes defined by the walls of the room and the in- whereas ventral areas (e.g., the lingual and fusiform
trinsic structure of the layout (the objects can be gyri) seem to be recruited for processing the appear-
grouped into rows and columns parallel to the walls) ance of objects and scenes (Aguirre and DEsposito,
to construct a reference system for remembering the 1997). This dorsal-ventral dissociation is analogous to
locations of objects in the room. Egocentric spatial re- that observed in vision (Mishkin, Ungerleider, and
lations must be represented at the perceptual level for Macko, 1983). Recent evidence indicates that route
the purpose of guiding action in the environment and survey learning recruit a common cortical net-
(Sholl and Nolin, 1997). It is an open question wheth- work in the brain and that survey learning recruits a
er egocentric spatial relations are represented in subset of the route-learning areas (Mellet et al., 2000;
long-term spatial memories. Shelton and Gabrieli, 2002).
SPATIAL MEMORY 635
Figure 2
Testing situation used to investigate spatial reference systems. The participant viewed a layout of objects on a table, and then turned
halfway around to reconstruct the layout on another table. Tenejapan participants preserved the orientation of the layout with respect
to the environment, as illustrated.
Language and Culture Gender Differences

Spatial memories are not insulated from lan- There is a long history of research on gender dif-
guage and culture. The Mayan language Tzeltal uses ferences in spatial ability (Maccoby and Jacklin, 1974)
an environmental reference system to describe the but relatively few studies have examined spatial learn-
spatial relations among objects. This reference system ing and memory of room-sized and larger spaces.
corresponds to cardinal directions established by Those studies have shown that males perform better
the mountainous terrain in which the speakers live than females or that the genders do not differ (Mon-
(e.g., downhill = north, uphill = south, across = tello, Lovelace, Golledge, and Self, 1999). Recent ex-
east or west). Tzeltal does not have the egocen- periments have documented an intriguing dissocia-
tric spatial terms left and right. When shown an tion between females and males in spatial abilities.
array of objects on a table and asked to reconstruct Females perform better than males on tasks that re-
the array on another table behind them after quire memory of the locations and identities of ob-
turning 180 degrees (see Figure 2), Tzeltal speakers jects, whereas males perform better than females on
reconstruct the array preserving its orientation with tasks that require keeping track of orientation in
respect to the environmental reference system, not large-scale environments (Montello et al., 1999;
with respect to their egocentric perspective (Levin- Silverman et al., 2000). (There is, of course, large
son, 1996). In contrast, speakers of Dutch (and pre- variability within genders, with some men performing
sumably speakers of most Western languages) pre- better than some women on object-location tasks, and
serve egocentric, not environmental, spatial relations some women performing better than some men on
(e.g., plate to the left in the reconstructed array). orientation tasks.) How can this dissociation be ex-
Almost superhuman navigational abilities have been plained? According to the hunter-gatherer theory,
documented in several cultures, including those spatial sex differences arise from a sexual division of
of the Australian Aborigines (Lewis, 1976) and Pulu- labor between hunting and gathering during human
wat Islanders (Gladwin, 1970); experiments have evolution (Silverman and Eals, 1992). Tracking and
shown that Aboriginal children perform substantially killing animals, a predominately male activity, re-
better than white children on tests requiring the chil- quired monitoring ones location and orientation
dren to remember the locations of objects (Kearins, over large distances, whereas foraging for edible
1981). plants, a predominately female activity, required
636 SPENCE, KENNETH
memory of the locations of plants and other immobile Mellet, E., Bricogne, S., Tzourio-Mazoyer, N., Gham, O., Petit, L.,
foods and then relocating them in subsequent grow- Zago, L., Etard, O., Berthoz, A., Mazoyer, B., and Denis, M.
(2000). Neural correlates of topographic mental exploration:
ing seasons. The impact of route versus survey perspective learning.
NeuroImage 12, 588600.
Mishkin, M., Ungerleider, L. G., and Macko, K. A. (1983). Object
Conclusion vision and spatial vision: Two cortical pathways. Trends in
Neurosciences 6, 414417.
Spatial memories of large environments are com-
Montello, D. R., Lovelace, K. L., Golledge, R. G., and Self, C. M.
posed of route knowledge and survey knowledge. (1999). Sex-related differences and similarities in geographic
Route knowledge is usually acquired before survey and environmental spatial abilities. Annals of the Association of
knowledge, but the learners goals affect what is American Geographers 89, 515534.
learned. Some perspectives of a familiar environment Mou, W., and McNamara, T. P. (2002). Intrinsic frames of refer-
ence in spatial memory. Journal of Experimental Psychology:
can be retrieved more efficiently than others. This
Learning, Memory, and Cognition 28, 162170
orientation dependence arises because people store Poucet, B. (1993). Spatial cognitive maps in animals: New hypothe-
knowledge about location and orientation in terms of ses on their structure and neural mechanisms. Psychological
reference directions or axes. The use of different ref- Review 100, 163182.
erence directions or axes in different locales may ex- Rock, I. (1973). Orientation and form. New York: Academic Press.
Shelton, A. L., and Gabrieli, J. D. E. (2002). Neural correlates of
plain why spatial memories are hierarchical. Emerg-
encoding space from route and survey perspective. Journal of
ing evidence indicates that different areas of the brain Neuroscience 22, 2,7112,717
specialize in representing and processing specific Shelton, A. L., and McNamara, T. P. (2001). Systems of spatial ref-
kinds of spatial memories. Gender, language, and erence in human memory. Cognitive Psychology 43, 274310.
culture influence the development and pattern of Sholl, M. J., and Nolin, T. L. (1997). Orientation specificity in rep-
resentations of place. Journal of Experimental Psychology: Learn-
spatial capabilities.
ing, Memory, and Cognition 23, 1,4941,507.
Siegel, A. W., and White, S. H. (1975). The development of spatial
See also: SPATIAL LEARNING: ANIMALS
representations of large-scale environments. In H. W. Reese,
ed., Advances in child development and behavior. San Diego: Aca-
Bibliography
demic Press.
Aguirre, G. K., and DEsposito, M. (1997). Environmental knowl- Silverman, I., Choi, J., Mackewn, A., Fisher, M., Moro, J., and Ol-
edge is subserved by separable dorsal/ventral neural areas. shansky, E. (2000). Evolved mechanisms underlying wayfind-
Journal of Neuroscience 17, 2,5122,518. ing: Further studies on the hunter-gatherer theory of spatial
Anderson, R. A., Snyder, L. H., Bradley, D. C., and Xiang, J. sex differences. Evolution and Human Behavior 21, 201213.
(1997). Multimodal representation of space in the posterior Silverman, I., and Eals, M. (1992). Sex differences in spatial abili-
parietal cortex and its use in planning movements. Annual Re- ties: Evolutionary theory and data. In J. H. Barkow, L. Cos-
view of Neuroscience 20, 303330. mides et al., eds., The adapted mind: Evolutionary psychology and
Epstein, R., and Kanwisher, N. (1998). A cortical representation of the generation of culture. New York: Oxford University Press.
the local visual environment. Nature 392, 598601. Stevens, A., and Coupe, P. (1978). Distortions in judged spatial re-
Evans, G. W., and Pezdek, K. (1980). Cognitive mapping: Knowl- lations. Cognitive Psychology 10, 422437.
edge of real-world distance and location information. Journal Taylor, H. A., Naylor, S. J., and Chechile, N. A. (1999). Goal-
of Experimental Psychology: Human Learning and Memory 6, 13 specific influences on the representation of spatial perspec-
24. tive. Memory & Cognition 27, 309319.
Gallistel, C. R. (1990). The organization of learning. Cambridge, MA: Thorndyke, P. W., and Hayes Roth, B. (1982). Differences in spa-
MIT Press. tial knowledge acquired from maps and navigation. Cognitive
Gladwin, T. (1970). East is a big bird. Cambridge, MA: Harvard Uni- Psychology 14, 560589.
versity Press. Tversky, B. (1981). Distortions in memory for maps. Cognitive Psy-
Kearins, J. M. (1981). Visual spatial memory in Australian Aborigi- chology 13, 407433.
nal children of desert regions. Cognitive Psychology 13, 434
460. Timothy P. McNamara
Levinson, S. C. (1996). Frames of reference and Molyneauxs ques-
tion: Crosslinguistic evidence. In P. Bloom, M. A. Peterson,
L. Nadel, and M. F. Garrett, eds., Language and space. Cam-
Lewis, D. (1976). Observations on route finding and spatial orien-
tation among the Aboriginal peoples of the western desert re- SPENCE, KENNETH (19071967)
gion of central Australia. Oceania 46, 249282. Kenneth W. Spence (19071967) played a major role
Maccoby, E. E., and Jacklin, C. N. (1974). The psychology of sex differ-
ences. Stanford, CA: Stanford University Press. in psychology from the early 1930s until his untimely
Maguire, E. A., Frackowiak, R. S. J., and Frith, C. D. (1997). Recall- death. His impact is illustrated by the fact that from
ing routes around London: Activation of the right hippocam- 1962 to 1967, he was the most cited author in a survey
pus in taxi drivers. Journal of Neuroscience 17, 7,1037,110. of the fourteen most prestigious psychological jour-
McNamara, T. P. (1986). Mental representations of spatial rela- nals (Myers, 1970). Spences influence resulted from
tions. Cognitive Psychology 18, 87121.
McNamara, T. P., Hardy, J. K., and Hirtle, S. C. (1989). Subjective achievements as experimentalist, theorist, methodol-
hierarchies in spatial memory. Journal of Experimental Psycholo- ogist, and teacher. In all of these roles, he operated
gy: Learning, Memory, and Cognition 15, 211227. as a natural science psychologist, one who believed
SPENCE, KENNETH 637
that the science of psychology can employ the same

methods of empirical inquiry and theory construction
as physics, chemistry, and biology. In essence, he was
asserting that psychology, in principle, is capable of
producing a body of reliable scientific knowledge. To
achieve this goal he deemed it necessary to conceptu-
alize psychology as the science of behavior, not of the
mind. That is, the basic observations of the science of
psychology are the behavior of organisms, not the di-
rect examination of conscious experience. This meth-
odological position, Behaviorism, was initially ex-
pressed, in a radical form, by John B. Watson
(Kendler, 1987), but since has matured into a more
sophisticated version known as neobehaviorism.
Early Life
Spence was born May 6, 1907, in Chicago. When
he was 4, Western Electric transferred his father, an
electrical engineer, to Montreal. He majored in psy-
chology at McGill University, receiving a B.A. in 1929
and an M.A. in 1930. His Ph.D. was granted in 1933
by Yale University, where he served as a research as-
sistant to Robert M. Yerkes, under whose direction he
completed a dissertation on the visual acuity of chim-
panzees. The dominant intellectual influence during
his Yale days evolved from the inspirational ideas of
Clark L. Hull, who set as his goal the formulation of Kenneth Spence (Psychology Archives, University of Akron)
a theoretical interpretation of behavior that emulated
the conceptual structure of Newtonian physics.
Florida, where he published his classical theory of an-
Hulls general approach was shaped by both Ivan imal discrimination learning (Spence, 1936), a con-
Pavlov and Edward Thorndike. Pavlovian condition- ception that is still influential (Kendler, 1992). In
ing, for Hull, was the simplest form of learning, and 1937 he moved to the University of Virginia as an as-
hence principles of conditioning could provide the sistant professor of psychology. The following year he
premises from which more complex forms of behav- was appointed associate professor at the University of
ior could be deduced (explained). Thorndikes law of Iowa, where, in 1942, he became professor and head
effect represented, for Hull, another fundamental of the Department of Psychology, a position he occu-
principle of behavior; rewards, technically known as pied until 1964, when he moved to the University of
reinforcements, are necessary for the formation and Texas.
strengthening of a connection between a situation
and behavior, or what became to be known as a stimu-
lus-response association (e.g., in Pavlovian conditioning Spences Work
the sound of a tone became connected with salivation
Kenneth Spences efforts in theoretical psycholo-
because it was reinforced by food).
gy are characterized by a consistent direction and a
While a graduate student, Spence prepared a close relationship between facts and theory. He had
paper (Spence, 1932) that illustrated Hulls general (a) an ability to express ideas in a lucid prose that re-
strategy. Based on assumptions about the effect of de- sulted from clear thinking and hard work; (b) a flair
layed reinforcement in conditioning, Spence predict- for designing clever experimental tests of competing
ed that in a complex maze, one with several successive hypotheses; (c) a knack for analyzing data so that
choice points, entrances into blind alleys would be their theoretical implications would become fully ap-
eliminated in a backward order, that is, the blinds are parent; (d) an ingenious talent for theoretically inte-
more difficult the farther they are from the goal. After grating a range of experimental results with the
completing his graduate work at Yale, Spence accept- added, and unusual, feature of being able to offer
ed a National Research Council fellowship to the Yale possible alternative explanations, even some from
Laboratories of Primate Biology at Orange Park, competing theories; and (e) a special aptitude, in
638 SPENCE, KENNETH
spite of limited mathematical training, to coordinate was similar to that of rats, in that simple stimulus-
his theoretical interpretation with a quantitative response associations were formed but the acquisition
model that structured an empirical problem so that of symbolic skills introduced a more complex pattern
further theoretical clarification and empirical devel- of stimulation and associative connections that en-
opment could occur. hanced relational responding. This led to a two-stage
theory of discrimination learning (Kendler and
One excellent example of Spences style of theo-
Kendler, 1962, 1975) in which the lower stage (single-
retical and empirical structuring is exhibited in his in-
unit), based upon Spences continuity model, hypoth-
fluential analysis of discrimination learning of ani-
esized that the responses of subhuman animals were
mals (Spence, 1936), which contained one of the first
directly linked (associated) to stimuli (e.g., black,
mathematical simulations (without the aid of a com-
white), whereas higher-level functioning (mediational
puter!) of a theory. The model, which seeks to identify
model) suggested that incoming stimulation is trans-
the psychological processes involved in how animals
formed into some internal conceptual representation
learn to choose between two stimuli when one is fol-
(e.g., brightness) that guides subsequent behavior.
lowed by reward and the other is not, postulates two
This model, which had its origins in Spences discrim-
basic theoretical mechanisms that Hull had employed
ination learning theory, could account for develop-
in interpreting conditioning phenomena: excitation
mental changes in the discrimination learning exhib-
and inhibition. When the animal is reinforced for se-
ited by humans.
lecting one stimulus, that habit (stimulus-response as-
sociation) is gradually strengthened; when a response From 1950 to the end of his life, classical (Pavlov-
to the other cue is not reinforced, that habit is gradu- ian) eyeblink conditioning with human subjects occu-
ally weakened. When the difference in strengths be- pied Spences interest. His efforts revealed basic prin-
tween the two competing habits reaches a certain ciples of habit formation and the interaction effects
value, the subject consistently chooses the reinforced between habits and drives. With the collaboration of
stimulus. Spence had no illusion that his formulation Janet A. Taylor, who later became his wife, the role
represented a complete interpretation of the discrim- of anxiety as a motivational mechanism was empiri-
ination learning process. He readily acknowledged cally and theoretically analyzed (e.g., Spence and
that certain fundamental processes, such as complex Taylor, 1951; Spence, 1956). Spence (1966) was able,
perceptual mechanisms, were ignored but insisted, as by clever experimental manipulations, to get human
a matter of strategy, that discrimination learning, as subjects to respond either in a simple associative man-
well as all psychological problems, must be broken ner analogous to subhuman behavior or in a cognitive
down into bare essentials to be investigated fruitfully. (mediational) manner. This was another example of
Reducing discrimination learning to the analysis of his effort, as well as that of several of his students, to
competing habits in a simple experimental situation gain an experimental grip on evolutionary processes
was a fruitful strategy for dealing with fundamental in behavior theory.
principles of behavior.
Spences theory of discrimination learning had
many ramifications, including the formulation of an
Groundbreaking Psychologist
ingenious stimulus-response explanation (Spence, In 1955, Spence was invited to deliver the Silli-
1937) of the transposition phenomenon, the tenden- man Lectures at Yale, a prestigious series that until
cy for nonverbal organisms to transfer a relational then had been given by distinguished physical and bi-
choice from one problem to a subsequent one (e.g., ological scientists such as Ernest Rutherford, Enrico
after learning to choose a medium gray in preference Fermi, and Charles Sherrington. The lectures, pub-
to a lighter one, a rat will probably select a darker lished under the title Behavior Theory and Conditioning
gray rather than the previously rewarded medium (1956), can be characterized as a realistic and reason-
gray). In addition, his theory initiated a crucial con- able theoretical interpretation of a wide range of ex-
troversy as to whether discrimination learning was a perimental data. Unlike many of the theoretical mes-
continuous or noncontinuous process, occurring siahs who have dotted the psychological landscape
gradually or suddenly (Kendler, 1987). with grandiose theories, Spences formulation was
never far removed from experimental evidence. In
Although Spences discrimination learning theo- essence, Spence was an experimental psychologists
ry was designed for nonverbal organisms, he sought theorist because his hypotheses were clearly testable.
to discover how the acquisition of symbolic skills
would influence discrimination learning. He encour- Spences skills as an experimentalist and theorist
aged a doctoral student, Margaret Kuenne (1946), to were matched by his talents as a methodologist. With
investigate this problem; she found that in a transpo- the collaboration of the philosopher of science Gustav
sition problem the behavior of inarticulate youngsters Bergmann, Spence contributed to the clarification of
SPINAL PLASTICITY 639
the meaning of psychological concepts and the logic piricist who, while appreciating the immaturity of the
of psychological measurement (Bergmann and science of psychology, was nevertheless able to ad-
Spence, 1941, 1944). His greatest contribution as a vance it with fruitful conceptions of learning and mo-
methodologist was his clarification of issues in theo- tivation.
retical psychology, particularly in relation to the com-
parative analysis of competing learning theories. The See also: DISCRIMINATION AND GENERALIZATION;
EVOLUTION AND LEARNING
competing theory to the Hull-Spence model that in-
terested Spence the most was Edward Tolmans cog- Bibliography
nitive theory of animal learning. This formulation Bergmann, G., and Spence, K. W. (1941). Operationism and theory
generated much confusion among both behaviorists construction. Psychological Review 48, 114.
and antibehaviorists; the former could not reconcile (1944). The logic of psychological measurement. Psychologi-
cal Review 51,124.
Tolmans hypothesized mentalistic processes with
Kendler, H. H. (1967). Kenneth W. Spence: 19071967. Psychologi-
methodological behaviorism, and the latter denied cal Review 74, 335341.
that Tolman could be labeled a behaviorist if he em- (1984). Evolutions or revolutions? In K. M. B. Lagerspetz
ployed phenomenological experience as a metaphor and P. Niemi, eds., Psychology in the 1990s. Amsterdam: North
for his theoretical constructs. Spence clearly distin- Holland.
(1987). Historical foundations of modern psychology. Pacific
guished between the strategies employed for concep-
Grove, CA: Brooks/Cole.
tualizing theoretical processesmechanistic, phe- Kendler, H. H., and Kendler, T. S. (1962). Vertical and horizontal
nomenological, mathematical, or some otherand processes in problem solving. Psychological Review 69, 116.
the theorys deductive, empirical consequences. Un- (1975). From discrimination learning to cognitive develop-
fortunately this distinction between a theorists think- ment: A neobehavioristic odyssey. In W. K. Estes, ed., Hand-
book of learning and cognitive processes, Vol. 1. Hillsdale, NJ: Erl-
ing style and the explicit theory with its deductive em- baum.
pirical implications was not fully appreciated at the Kendler, T. S. (1992). Levels of cognitive development. Hillsdale, NJ:
time of the cognitive revolution, and as a result need- Erlbaum.
less disputation and misunderstanding were encour- Kuenne, M. R. (1946). Experimental investigation of the relation
aged. Many segments of the psychological communi- of language to transposition behavior in young children. Jour-
nal of Experimental Psychology 36, 471490.
ty failed to appreciate the linkage between Myers, C. R. (1970). Journal citations and scientific eminence in
neobehaviorism in general, and Spences theoretical contemporary psychology. American Psychologist 25, 1,041
efforts in particular, in attempting to understand the 1,048.
relationship between associative and cognitive pro- Spence, K. W. (1932). The order of eliminating blinds in maze
learning by the rat. Journal of Comparative Psychology 14, 927.
cesses (Kendler, 1984).
(1936). The nature of discrimination learning in animals.
Spences contributions to psychology cannot be Psychological Review 43, 427429.
(1937). The differential response in animals to stimuli vary-
limited to his own publications. His inspired teaching
ing within a single dimension. Psychological Review 44, 430
must also be considered. Many of his seventy-five doc- 444.
toral students, too many to mention, made solid con- (1956). Behavior theory and conditioning. New Haven, CT:
tributions to the science of psychology: All of his doc- Yale University Press.
toral students carry with them some of Spences ideas (1966). Cognitive and drive factors in the extinction of the
conditioned eye blink in human subjects. Psychological Review
and commitments and a desire to achieve a level of
73, 445451.
quality in their own work that would be acceptable to Spence, K. W., and Taylor, J. A. (1951). Anxiety and strength of the
their Professor (Kendler, 1967, p. 341). UCS as determiners of the amount of eyelid conditioning.
In addition to being the first and only psycholo-
gist to give the Silliman Lectures, Spence received Howard H. Kendler
many other honors, including membership in the Na-
tional Academy of Sciences, the Howard Crosby War-
ren Medal of the Society of Experimental Psycholo-
gists in 1953, and the American Psychological
SPINAL PLASTICITY
Associations Scientific Contribution Award in 1956, Spinal plasticity refers to short- or long-term alter-
the first year that it was presented. But perhaps the ations of the excitability of the spines neural path-
greatest honor Spence aspired to was a place in the ways. Although the mammalian spinal cord is a
history of psychology. Although it is difficult to pene- unique part of the central nervous system, it is most
trate the haze of the future, especially in relation to commonly identified as a transmitter of information
psychology, a discipline that is conceptualized in to and from the brain and as a repository of various
many antithetical ways, it is likely that Spence will be reflex functions that allow an automatic response to
remembered as a clear and sophisticated exponent of external stimuli. But spinal cord pathways are dynam-
natural science psychology and as a theorist and em- ic and continuously changing systems, not static,
640 SPINAL PLASTICITY
hard-wired entities simply transmitting information sensory nerve. Inflammation of a peripheral area
from the body to the brain and back. such as a knee joint can also produce increases in spi-
Spinal reflexes appear to be hard-wired function- nal reflex excitability that are similar to fixation. Fixa-
al circuits whose excitability temporarily varies with tion seems similar to long-term potentiation (LTP),
descending activity from the brain or with repeated which may be a part of learning and memory in intact
sensory input. In the 1930s, however, work began to animals. Fixation and other long-term excitability in-
show that the spinal reflex pathways might be altered creases are likely due to alterations in interneuron
in ways that have many characteristics of learning and cell membrane receptors produced by upregulation
memory in the intact mammal. of genes controlling membrane function.
The fourth stage of nonassociative excitability al-

Nonassociative Excitability Changes terations may be permanent. With prolonged stimu-
lus input into the reflex paths of the cord, inhibitory
Researchers have long been aware of memory-
interneurons regulating excitability balance may be
like but temporary alterations of spinal reflex excit-
destroyed. In addition, some evidence points to an in-
ability. Short-term decreases in excitability were stud-
crease in excitatory synapse formation in these cir-
ied in the early 1900s and termed reflex fatigue.
cumstances. The loss of inhibitory control and in-
However, since about 1975, researchers have learned
creased numbers of excitatory synapses would lead to
about both decreases and increases in spinal reflex
a permanent excitability increase.
excitability; they now know that there are essentially
four overlapping phases of spinal reflex alterations,
mainly representing increased excitability caused by
Associative Changes
sensory inputs to the cord.
Initial research on learning in the spinal cord
The most rapidly developing and rapidly lost
began with attempts to determine the simplest part of
changes are habituation and sensitization. Habitua-
the nervous system that could support learning. Re-
tion is a decrease in spinal reflex excitability caused
searchers used classical conditioning operations, fol-
by repeated stimulation; it results in decreased re-
lowing a conditioned stimulus with an unconditioned
sponse to a sustained stimulus. Once the stimulus is
removed, excitability returns to normal in seconds or stimulus to the hind limb of an anesthetized, spinally
minutes. Sensitization (sometimes known as windup) transected subject (usually a dog). Although the early
is an increased excitability to a sustained stimulus that studies were inconclusive, later studies using well-
is also rapidly lost, usually within 90 to 120 seconds established control group technology proved that spi-
after stimulus cessation. Either habituation or sensiti- nal reflex excitability could be altered by classical con-
zation may occur with as little as one or two stimulus ditioning procedures. Spinal conditioning shows
applications; both seem to be due to altered neuro- many features of classical conditioning in the intact
transmitter release from either incoming sensory animal, although the spinal reflex system apparently
nerve terminals or from interneurons within the re- cannot learn to respond differentially. The learned
flex pathways. changes do not spontaneously decay but show the ex-
tinction decreases typical of classical conditioning.
The second stage of excitability alteration is
Reflex excitability alterations induced by classical
termed long-term sensitization and occurs with lon-
conditioning procedures occur in the interneurons of
ger and/or more intense stimulus inputs. Long-term
the reflex pathways rather than in the initial synapses
sensitization, once established, decreases for hours or
of the sensory fibers into the cord or in the motoneu-
even for a day with no further stimulation to the re-
rons. In addition to excitability alterations arising
flex circuit. This excitability increase is likely due to
from classical conditioning, spinal reflexes respond
an alteration of cell membrane receptor sensitivity of
to instrumental conditioning operations, which can
the secondary interneurons of the reflex pathways.
drive spinal reflex excitability either up or down, de-
The third stage of excitability alterations is spinal pending on the conditioning situation.
fixation. Here, a fairly intense, thirty-to-fifty-minute
stimulus to the spinal cord can increase the excitabili- While most studies of classical and instrumental
ty of the activated reflex pathways for as long as sever- learning in the spinal cord have used painful or noci-
al days and is essentially a memory trace in the spinal ceptive stimuli, a similar alteration of spinal reflex ex-
cord. The excitability increase may be strong enough citability has been demonstrated in intact monkeys
to produce continuing motoneuron output after re- taught to increase leg muscle tone over many days.
moval of the initiating stimulus. Fixation can be pro- The change was gradual but also enduring, occurring
duced by stimulation or lesions of various brain rein the spinal reflexes but involving nonnociceptive in-
gions, or by stimulation of peripheral skin or a puts.
STRESS AND MEMORY 641
Conclusion STIMULUS CONTROL

The major spinal reflex excitability changes See: DISCRIMINATION AND GENERALIZATION
shown by associative and nonassociative procedures
indicate that spinal reflexes play a vital role in sensory
information processing. The major impact of spinal
reflex excitability changes may be in chronic pain. In- STRESS AND MEMORY
creased spinal-pathway excitability is probably a fac- Stressful and emotional events can promote or impair
tor in many of the most intractable chronic-pain syn- the acquisition of new memories, depending on the
dromes. Understanding the cellular basis of spinal type of stressful experience, the type of learning
reflex and spinal pathway excitability alterations even the sex of the animal. Stress is the external con-
might offer breakthroughs in the treatment of some dition that places demands on the organism, and the
of these debilitating conditions. In addition, the in- stress response is the organisms adaptive response to
creased understanding of spinal pathway plasticity is the stressor, typically measured as changes in perfor-
leading to advances in the rehabilitation of spinal mance or physiological or biochemical states.
cord injuries and in the technological means of in-
Most adaptive responses are crucial to an organ-
creasing the mobility of patients with incomplete or
isms capacity for survival and are easily reconcilable
even complete spinal transections.
with theories of natural selection. For instance, the re-
lease of glucocorticoids from the adrenal glands di-
See also: CONDITIONING, CLASSICAL AND rects glucose to the brain and musculature in prepa-
INSTRUMENTAL; HABITUATION AND ration for fight or flight, thereby eliminating
SENSITIZATION IN VERTEBRATES; NEURAL unnecessary processing of ongoing vegetative func-
SUBSTRATES OF CLASSICAL CONDITIONING: tions such as digestion. These physiological responses
DISCRETE BEHAVIORAL RESPONSES to stress in turn affect cognitive processes such as
learning and memory, thereby allowing the animal to
Bibliography prepare for or avoid subsequent sources of stress. The
Coderre, T. J. (2001). Noxious stimulus-induced plasticity in spinal interaction between stress and memory is a complex
cord dorsal horn: Evidence and insights on mechanisms ob- one that does not always serve the best interests of the
tained using the formalin test. In M. M. Patterson and J. W. animal.
Grau, eds., Spinal cord plasticity: Alterations in reflex function.
Boston: Kluwer Academic Press.
At first it seems reasonable to assume that there
Groves, P., and Thompson, R. (1970). Habituation: A dual-process is a direct correlation between degree of stress and
theory. Psychological Review 77(5) 419450. degree of detriment. In reality, animals (including
Patterson, M. M. (1976). Mechanisms of classical conditioning and humans) perform optimally at moderate levels of de-
fixation in spinal mammals. In A. H. Reisen and R. F. Thomp- mand, and performance is compromised at the ex-
son, eds., Advances in psychobiology. New York: Wiley.
tremes. Hence, the relationship between stress and
(2001). Spinal conditioning: The first seventy years. In J. E.
Steinmetz, M. A. Gluck, and G. D. Solomon, eds., Model systems
performance is an inverted U-shaped function: Per-
and the neurobiology of learning. Mahwah, NJ: Erlbaum. formance is impaired equally at high levels of stress
(2001). Spinal fixation: Long-term alterations in spinal re- and at low levels (boredom or drowsiness); moderate
flex excitability. In M. M. Patterson and J. W. Grau, eds., Spi- levels, performance is enhanced. For example, expo-
nal cord plasticity: Alterations in reflex function. Boston: Kluwer sure to a moderate stressor such as background noise
Academic Publishers.
can facilitate performance of a prolonged vigilance
M. M. Patterson, and J. W. Grau, eds. (2001). Spinal cord plasticity.
Boston: Kluwer Academic Publishers. task. Substituting arousal for stress, Donald Hebb
Wernig, A., Nanassy, A. et al. (2001). Laufband (treadmill) therapy (1955) suggested that continuing neural activity could
in incomplete para- and tetraplegia. In M. M. Patterson and provide a physiological means of describing such self-
J. W. Grau, eds., Spinal cord plasticity: Alterations in reflex func- motivating phenomena as curiosity and exploration
tion. Boston: Kluwer Academic Publishers.
and their obvious contribution to learning. Moreover,
Willis, W. D. (2001). Mechanisms of central sensitization of noci-
ceptive dorsal horn neurons. In M. M. Patterson and J. W.
he emphasized that without such an arousal founda-
Grau, eds., Spinal cord plasticity: Alterations in reflex function. tion, no learning would occur. This relationship, in
Boston: Kluwer Academic Publishers. combination with the inverted-U relationship pro-
Wolpaw, J. R. (2001). Spinal cord plasticity in the acquisition of a posed between stress and performance, predicts a lin-
simple motor skill. In M. M. Patterson and J. W. Grau, eds., ear relationship between stress and arousal. Others,
Spinal cord plasticity: Alterations in reflex function. Boston: Klu-
wer Academic Publishers.
however, have reported an upright U-shaped correla-
tion between stress and arousal, which forecasts a
Michael M. Patterson monotonic relationship between stress and perfor-
Michael J. Bartelt mance. Still others report that stress and arousal are
Revised by Michael M. Patterson independent variables.
642 STRESS AND MEMORY
Exactly how stressful experience will affect mem- of stress and learning and, in fact, depends on indi-
ory formation depends, presumably, on the signifi- vidual sex, species, and strain differences as well as
cance of the information to be remembered and the nature and difficulty of the task. For example,
therefore on the type of learning that is being tapped. prior exposure to inescapable stresses such as tail
Under most circumstances, an animal will remember shocks or swim stress facilitates classical conditioning
an aversive event for most of its life. But subsequent of the eyeblink response in male rats, whereas expo-
experience can change these memories. Autonomic sure to the same stresses impairs conditioning in fe-
tasks entailing a high degree of preparedness are less male rats. Recent studies in humans have suggested
likely to be adversely affected than tasks requiring that males and females also differ in the ways in which
high cognitive capacity and concentration. These two they deal with stress; males tend to isolate themselves,
extremes can be viewed as the ends of a continuum, whereas females tend to befriend others and seek
with stress impairing performance as task difficulty comfort. These differences can affect what types of
increases. This relationship is exemplified by the in- learning opportunities arise after stressful experi-
tense training and practice required of highly skilled ence.
professionals such as military personnel, who must
perform optimally under high levels of stress and un- One of the first scientists to define stress and its
certainty. adaptive utility and underlying mechanisms was the
eminent physiologist Hans Seyle. Dr. Seyle defined
The degree to which stress impinges on perfor-
the stress response as the nonspecific response of the
mance is highly susceptible to individual differences.
body to any demand placed upon it. In general, Seyle
Although there is a limit to the attentive capacity that
can be allocated to a particular task, this capacity is was referring to activation of the hypothalamic-
not fixed and will vary from individual to individual pituitary-adrenal (HPA) axis. In this axis, hypotha-
and from day to day. On the evidence of human self- lamic peptides activate adrenocorticotropin (ACTH)
report scales, high-anxiety subjects tend to perform secretion from the anterior pituitary. ACTH induces
optimally on easy tasks, and low-anxiety subjects per- the release of glucocorticoids from the adrenal cor-
form optimally under more demanding circum- tex, and glucocorticoids, in turn, inhibit the release
stances; within tasks, low-anxiety subjects perform of pituitary ACTH. Glucocorticoids are released pe-
better during early stages of learning, and high anxi- ripherally into the blood upon most stressful encoun-
ety facilitates the later stages. Other variables include ters and are potentially damaging, especially in high
social factors, age, and disease. But the most signifi- concentrations and/or chronic conditions. There are
cant is past memories. The animal must not only cal- numerous reports that they can impair later learning;
culate the imbalance between the perceived demand however, there are also reports that they can enhance
and the ability to cope with that demand but must also it. As with exposure to a stressful event, the degree
incorporate this information into its previous experi- and direction of the effect depends on the amount of
ence. glucocorticoids that are released and the time of ex-
posure.
In the early 1960s J. B. Overmier and M. E. P.
Seligman noticed that dogs exposed to inescapable In addition to glucocorticoids, many other stress-
shock were later impaired in their ability to perform related neuromodulators can affect mnemonic pro-
a task where escape was possible. They suggested that cesses. The catecholamines have been implicated in
the animals became helpless after learning that the learning, as are the dopaminergic systems. Along with
aversive event and the ability to cope with that event ACTH, -endorphin is released from the pituitary in
were not contingent; they termed the phenomenon response to stress. Peptides such as the opioids (the
learned helplessness. Because the secondary charac- enkephalins), vasopressin, and neuropeptide Y can
teristics that accompany helplessnessweight loss, be affected by exposure to a stressful experience and
sleep disturbances, decreased activity, and so on can thereby influence memory processes. One of
resemble characteristics of depressed humans, the more ubiquitous substrates affected by stress is the
phenomenon evolved into one of the first experimen- glutamatergic system, including the corresponding
tal models of depression. It has helped to demon- receptors, n-methyl-d-aspartate (NMDA) and -
strate stress-induced effects on immune function, amino-3-methylsoxazole-4-propionic acid (AMPA).
ulcer development, tumor growth, analgesia, aggres- Glutamate is the primary excitatory neurotransmitter
sion, and status within a dominance hierarchy. But in the brain, and activation of its receptors is critical
impairment in performance remains the most nota- to many types of learning. The dramatic extent to
ble result.
which stress affects this system affords ample oppor-
However intuitively appealing, this paradigm of tunity for stressful experiences to affect processes of
helplessness is constrained by more general theories learning and memory.
STRUCTURAL CHANGES AND MEMORY 643
Exposure to acute or chronic stressor experiences HIPPOCAMPUS; NEURAL SUBSTRATES OF

induces a variety of effects on neuronal plasticity, af- CLASSICAL CONDITIONING: DISCRETE
fecting synaptic morphology, receptor affinity and BEHAVIORAL RESPONSES; NEURAL SUBSTRATES
number, gene expression, and electrophysiological OF EMOTIONAL MEMORY
responsiveness. For example, exposure to an acute Bibliography
stressor of inescapable tail shocks induces immediate Blanchard, D. C., and Blanchard, R. J. (1988). Ethoexperimental
early gene (IEG) expression, impairs long-term po- approaches to the biology of emotion. Annual Review of Psy-
tentiation (LTP), and can enhance the density of den- chology 39, 4368.
dritic spines; research has identified all of these fac- Cox, T. (1980). Stress. Baltimore: University Park Press.
tors as possible biological substrates for learning in Lupien, S. J., and Lepage, M. (2001). Stress, memory, and the hip-
pocampus: Cant live with it, cant live without it. Behavioral
the mammalian brain. Many of these effects unfold in Brain Research 14, 137158.
the hippocampal formation, a part of the brain that Shors, T. J. (1998). Sex and stress effects on learning and memory:
plays a critical role in some types of memory forma- For better or for worse. The Neuroscientist 4, 353364.
tion. In addition, the amygdala is a critically factor in Taylor, S. E., Klein, L. C., Lewis, B. P., Gruenewald, T. L., Gurang,
many defensive and affective responses to stress or R. A., and Updegraff, J. A. (2000). Biobehavioral responses to
stress in females: Tend-and-befriend, not fight-or-flight. Psy-
danger. This limbic structure integrates information chological Review 107, 411429.
from numerous sources and sensory systems and par-
ticipates in the formation of memories about those Tracey J. Shors
experiences that are necessary for ensuring an appro-
priate response to future danger.
Many of the foregoing neuromodulatory systems
are colocalized, and the brain regions are highly in- STROKE
terconnected, making it difficult to interfere experi-
mentally with one without affecting another. Further- See: AMNESIA, ORGANIC; FRONTAL LOBES AND
more, given the wide diversity of stressors and the EPISODIC MEMORY; KNOWLEDGE SYSTEMS
AND MATERIAL-SPECIFIC MEMORY DEFICITS;
inherent variance in individual responsiveness to the
REHABILITATION OF MEMORY DISORDERS
same stressor, it is probable that specific but overlap-
ping circuits contribute to stress effects on learning.
Finally, these stress-induced responses are not neces- SYNAPSE
sarily detrimental to the physical and psychological
See: GUIDE TO THE ANATOMY OF THE BRAIN
well-being of the organism; in the aftermath of the
stressful encounter, they contribute to the reestablish-
ment of homeostasis and an appropriate consolida- STRUCTURAL CHANGES
tion of the experience. AND MEMORY
See also: HORMONES AND MEMORY; LEARNED See: MORPHOLOGICAL BASIS OF LEARNING AND
HELPLESSNESS; NEURAL COMPUTATION: MEMORY
T
TASTE AVERSION AND PREFERENCE In the mid-1960s taste aversion learning caught
LEARNING IN ANIMALS the attention of experimental psychologists. They ob-
served that when an animal drinks a tasty solution
Historically taste aversion learning arose as a prob- marked by a bright-noisy signal, and is later injected
lem in evolutionary biology. The English naturalist with a mild toxin, the animal will develop an aversion
Charles Darwin was puzzled by an incongruity: Some to the taste but not to the bright-noisy signal. Con-
tender caterpillars were brightly colored and exposed versely, if the animal is mildly shocked on the feet, it
themselves so that they caught the eye of every pass- will avoid the bright-noisy signal but not the tasty so-
ing bird. Such behavior appeared maladaptive. Years lution. Second, and more important, the shock must
later, the English anthropologist and naturalist Al- be applied immediately after the signal for effective
fred Russell Wallace suggested that brightly colored learning, but the toxin injection can be delayed for
butterfly larva probably tasted bitter and might be several hours after the consumption of the tasty solu-
poisonous; therefore the colors served to deter birds tion. Moreover, a one trial situation, that is, a single
and other predators. Subsequent research supported pairing of the taste with the toxin injection, is suffi-
Wallaces hypothesis. Consumption of the colorful in- cient to elicit a strong aversion to that particular taste.
sects causes gastric nausea and emesis, and after one These factors, known as (1) selective association and
or two trials birds and other predators learn to avoid (2) long-delay learning, are the major behavioral
them. As larva, these insects feed on plants that characteristics of taste aversion learning. This type of
evolved the bitter toxins as a defense against herbi- behavioral paradigm, that came to be known later as
vores; the insects turned that defense to their own ad- conditioned taste aversion (CTA), can tolerate an in-
vantage. terstimulus interval of up to six to eight hours be-
tween the taste (the conditioned stimulus, CS) and the
Taste aversion learning proved to be widespread malaise inducing agent (the unconditioned stimulus,
in phylogeny and ontogeny. Taste-toxin conditioned US) during the training session. After consumption,
aversions have been observed in snails, insects, fish, the internal representation of the taste would proba-
frogs, salamanders, lizards, snakes, domestic and wild bly be encoded in an on-hold position over hours
birds, and in mammals, ranging from fetal and neo- before a decision is being made of whether the food
nate rats, to young children and adult humans. Even is safe or not. This long interstimulus interval enables
protozoans reject bitter, the natural taste of plant poi- the dissociation in time of neuronal events that gener-
sons. The ubiquity of the phenomenon indicates that ate the memory of the sensory stimulus from those
this mechanism to protect the gut must have evolved that subserve the association of the memory of the CS
many millions of years ago. with the US, that is, the negative reinforcer.
645
646 TASTE AVERSION AND PREFERENCE LEARNING IN ANIMALS
Since the 1980s, the ecological paradigm of CTA, and in the processing the detection of taste unfamil-
by virtue of its aforementioned experimental advan- iarity (see below); the amygdala as the region respon-
tages, has been adopted by several research groups in sible for the evaluation of the hedonic value of the
the study of the behavioral, pharmacological, cellular, taste as well as the expression of CTA; and the PBN
and molecular aspects of learning and memory in as the locus of the CS-US association.
mammals. In the laboratory, a routine CTA protocol
An important element in CTA is the novelty, or
is composed of the following steps: (1) the pre-
unfamiliarity, of the taste stimulus. Sampling of any
conditioning session (three to four days), in which rats
kind of a novel tastant by the rat, even in the absence
learn to drink water from the liquid container (usually
of the negative reinforcer, will result in the formation
the liquid is supplied in glass pipettes); (2) the condi-
of a memory to that specific taste. However, if these
tioning session (a single day), where rats sample the
same animals are then subjected to CTA training
taste (usually a solution of saccharin, but many other
tastes can be used) and around thirty to sixty minutes (now in the presence of the malaise-inducing com-
later administered with the transient malaise- pound), they will show a poorer aversion to the taste
inducing agent (an intraperitoneal injection of a solu- compared to animals that were not pre-exposed to it.
tion of lithium chloride is used as the standard US, The CS in CTA is most effective in rendering a strong
but the US can range from rotation and irradiation aversion response if it is unfamiliar to the organism
to drugs and poisons); and (3) the testing session, at the time of conditioning, A major question is, How
which can vary between one and six days. Rejection does the brain know when a taste is familiar or unfa-
of the conditioned taste can be easily monitored by miliar? Taste novelty detection is expected to require
measuring the amount of the taste consumed by the some type of fast internal comparator that matches
animal on the day of the test (a single-bottle test), and the on-line (sensory) information with off-line (mem-
comparing this volume to that consumed on the day ory) information. A potential candidate for this com-
of conditioning. Another way of quantifying aversion parator is a corticothalamo-brainstem system. The
to a specific taste is to calculate a so-called aversion thalamus may compare the on-line sensory informa-
index, based on the total amount of water consumed tion coming from the brainstem with previous taste
compared to the taste in a multiple-choice situation. memory representations retained in the IC, and when
a mismatch is identified (if the on-line taste informa-
The neurobiological mechanisms of the CS-US tion is novel), it triggers the behavioral response on
association in CTA are known only in broad outline. the one hand, and initiates memory encoding in the
In the rat, the processing of gustatory information be- IC on the other.
gins with transduction of chemical stimuli which
reach the oral cavity. The taste receptors send affer- By using the advantage of a single conditioning
ents via the facial and glossopharyngeal nerves to the trial, investigators have examined the role of several
nucleus of the solitary tract in the brainstem. The re- biochemical and molecular processes involved in the
ceptors in the viscera also send vagal fibers converg- discrete phases of acquisition, consolidation, and ex-
ing to the same nucleus. The blood carries absorbed tinction of CTA memory. For example, the microin-
food products to the area postrema, where blood fusion of a protein synthesis inhibitor into the IC
monitors report to the solitari nucleus. The CS-US blocks CTA learning when administered before the
routes then proceed to the parabrachial nucleus exposure to the taste, but not before the testing trial,
(PBN) located in the pons, the more anterior region indicating that the synthesis of new proteins in the
of the brainstem. Neurophysiological experiments in- cortex is a critical step for the formation, but not for
dicate that a complex series of looping circuits inter- the retrieval, of the taste memory. Along this line of
connect these nuclei in the brainstem with higher experimentation, researchers have found that the cel-
brain areas such as the gustatory cortex, located in the lular and molecular mechanisms that subserve CTA
insular cortex (IC). Although the IC is unnecessary are similar to those that subserve other forms of
for simple reflex responses to gustatory stimulus, ana- learning. These specific molecular devices (switch-
tomical and metabolic lesion experiments have shown es) are turned on/off in different brain regions dur-
that the IC is required for the retention of learned ing CTA (e.g., activation of cholinergic receptors and
taste aversion. The amygdala, another region inter- phosphorylation of glutamate NMDA receptors in the
connected with the PBN and the IC, is believed to IC, and activation of specific intracellular signaling
play an important role is assessing the hedonic value cascades together with modulation of gene expres-
of the consumed taste (i.e., the emotional aspect of sion in the IC and in the central nucleus of the amyg-
the taste learning experience). All in all, the behavior- dala). These essential molecular entities are also dif-
al, anatomical, and pharmacological data accumulat- ferentially activated in the brain according to the
ed to date suggest that the IC is the area of the brain stimulus dimension and context. For example, mus-
involved in the encoding of the memory of the taste carinic receptors and activation of members of the mi-
THORNDIKE, EDWARD 647
togen-activated protein kinase (MAPK) signaling cas-

cade are necessary for the acquisition of CTA to a
novel taste but not to a familiar one.
As mentioned, CTA results in a robust learning,
but yet, this behavior is very plastic. The aversive
memory can last for several months without signifi-
cant decay. However, if after conditioning animals are
subsequently exposed to the taste in the absence of
the negative reinforcer (as in a test situation), and
provided that they sample even a tiny amount of the
taste, the aversive memory will commence to decay.
This phenomenon, first described by the Russian
physiologist Ivan Pavlov as experimental extinction,
does not result in the erasure of the original aversive
memory, but rather reflects a relearning process in
which now the new CS-NoUS association comes to
control behavior.
See also: FOOD AVERSION AND PREFERENCE

LEARNING IN HUMANS
Bibliography
Berman, D. E., and Dudai, Y. (2001). Memory extinction, learning
anew, and learning the new: Dissociations in the molecular
machinery of learning in cortex. Science 291, 2,4172,419.
Berman, D. E., Hazvi, S., Neduva, V., and Dudai, Y. (2000). The
role of identified neurotransmitter systems in the response of
insular cortex to unfamiliar taste: Activation of ERK1-2 and Edward Thorndike (Psychology Archives, University of Akron)
formation of a memory trace. Journal of Neuroscience 20,
7,0177,023.
Berman, D. E., Hazvi, S., Rosenblum, K., Seger, R., and Dudai, Y.
(1998). Specific and differential activation of mitogen-
THORNDIKE, EDWARD (18741949)
activated protein kinase cascades by unfamiliar taste in the in- Edward Lee Thorndike was born on August 31, 1874,
sular cortex of the behaving rat. Journal of Neuroscience 18,
10,03710,044. in Williamsburg, Massachusetts. He died on August 9,
Bures, J., Bermudez-Rattoni, F., and Yamamoto, T. (1998). Condi- 1949, in Montrose, New York. Thorndike proceeded
tioned taste aversion: Memory of a special kind. New York: Oxford very rapidly through his graduate education. After re-
University Press. ceiving a B.A. at Wesleyan University in 1895, he
Bures, J., Buresova, O., and Krivanek, J. (1988). Brain and behavior:
Paradigms for research in neural mechanisms. New York: Wiley.
transferred to Harvard University, where he received
Garcia, J., Ervin, F. R., and Koeling, R. A. (1966). Learning with a second B.A. in 1896 and his M.A. the following year.
prolonged delay of reinforcement. Psychonomic Science 5, 121 In 1898, Thorndike completed his Ph.D. at Columbia
122. University, where he spent virtually his entire aca-
Garcia, J., Kimmeldorf, D. J., and Koelling, R. A. (1955). Condi-
demic career as a professor at Teachers College
tioned aversion to saccharin resulting from exposure to
gamma radiation. Science 122, 157158. (18991940). While this article will not dwell on
Garcia, J., McGowan, B. K., Ervin, F. R., and Koelling, R. A. (1968). Thorndikes achievements in the applied area that
Cues: Their relative effectiveness as a function of the reinforc- came to be called educational psychology, it should be
er. Science 160, 794795. noted that he created that field and developed it dur-
Lamprecht, R., and Dudai, Y. (2000). The amygdala in condi-
tioned taste aversion: Its there, but where. In J. Aggleton, ed.,
ing his entire career at Columbia.
The amygdala: A functional analysis, 2nd edition. Oxford: Ox-
Thorndikes experimental studies of learning in
Rosenblum, K., Berman, D. E., Hazvi, S., Lamprecht, R., and monkeys brought him his first position at Teachers
Dudai, Y. (1997). NMDA receptors and the tyrosine phospho- College as an instructor in genetic psychology. The
rylation of its 2B subunit in taste learning in the rat insular dean of Teachers College, James E. Russell, hired
cortex. The Journal of Neuroscience 17, 5,1295,135.
Thorndike because he thought those studies to be a
Schafe, G. E., Sollars, S. I., and Bernstein, I. L. (1995). The CS-US
interval and taste aversion learning: A brief look. Behavioral pretty good stepping stone to a study of the nature
Neuroscience 109, 799802. and behavior of children (Current Biography, 1941, p.
Yamamoto, Y., et al. (1994). Neural substrates for conditioned taste 857). The ingenuity of Thorndikes animal experi-
aversion in the rat. Behavioral Brain Research 65, 123137. ments had made a very favorable impression on Wil-
Diego E. Berman liam James at Harvard, as well as on his major profes-
648 THORNDIKE, EDWARD
sors at Columbia, each of whom would have strongly Thorndike injected into the concept of associationism
endorsed him for the position in Teachers College. concerned the nature of the association: It was not
Thorndikes contributions to the fledgling science of ideas that animals associated, it was movements (R) in
psychology are of the highest importance and come a given situation (S) that led to satisfying conse-
down to us well into the end of the twentieth century. quences. Thorndike spoke of neuronal connections in
the brain and hypothesized that the synaptic links be-
tween them were gradually made more traversable by
The Study of Animal Intelligence use and by satisfactory consequences, whereas disuse
To Thorndike goes the credit for putting what or annoying consequences made the neuronal con-
was to become the experimental psychology of animal nections underlying other S-R associations less tra-
learning (the study of animal intelligence) on a sound versable. This was Thorndikes neurological account
laboratory and theoretical footing. Thorndikes of trial-and-error learning, later reintroduced and ex-
studies (1911), which began around 1896 at Harvard panded in a significant way by Donald O. Hebb
and continued at Columbia University, were much (1949).
more controlled than those of any of his predecessors. The foregoing represents Thorndikes bequest to
He employed a variety of vertebrate species (fish, what became general behavior (learning) theory in
chicks, dogs, cats, monkeys) and typically placed the field of psychological science in the 1940s. Today,
them in a problem situation, such as a puzzle box or general behavior theory no longer holds such a cen-
a maze, from which they had to escape in order to get tral place in psychological science.
food and/or join companions. He observed the num-
ber of errors or latency to escape across trials and
generally published quantitative information on the Contributions to Animal Psychology
behavior of his experimental subjects. He attempted Thorndike also made three influential contribu-
to control the life history or extra-experimental expe- tions to comparative (animal) psychology. First, by
riences of his animals and also kept the problem situ- the introduction of the puzzle box and other standard
ation standard. testing situations, and the careful quantification of his
observations, he set the comparative psychology of
learning on an objective course from which it has
General Learning Theory rarely deviated. Thorndike also incidentally helped
From his experimental observations, Thorndike pave the way for the methodological and theoretical
proposed a general theory of learning that held that Behaviorism of John B. Watson a decade later, but it
the animal learned the association between an act and is clear that a number of other significant intellectual
a situation on the basis of the success of the act in threads were tending in that same direction around
bringing about a satisfying state of affairs. He pro- (and even before) the turn of the century: Ivan Sec-
posed that it is through the Law of Effect that acts that henovs reflexes of the brain (1863), Ivan Pavlovs
bring about a satisfying state of affairs are gradually highly quantitative conditioning procedures (Yerkes
stamped-in the nervous system and those that lead and Morgulis, 1909), Jacques Loebs (1912) physico-
to an annoying or discomforting state of affairs are chemical reductionism and his tropistic theory of psy-
gradually stamped-out. Thorndike believed the an- chology (movements are forced by external stimu-
imal had to behave actively to learn in the problem li), and especially H. S. Jenningss (1906) objective
situation and that the animals particular movements experimental approach to behavioral adaptation in
in the particular situation are validated or made in- single-celled paramecia, among other protozoan and
valid depending on whether they lead to satisfying or lower metazoan organisms. In fact, with the publica-
annoying consequences. Learning was thus the grad- tion of Watsons Behavior: An Introduction to Compara-
ual association of particular movements in particular tive Psychology (1914) and Psychology from the Standpoint
situations leading to certain ends. Thorndike be- of a Behaviorist (1919), virtually all of the general psy-
lieved that the actual connections between nerve cells chology became objective in methodology. Psycholo-
in the brain underlying situation-response (S-R) asso- gy became the study of behavior (instead of the
ciations were strengthened by reward (satisfaction) mind), with the conditioned response (reflex) as its
and weakened by punishment (annoyance). primary unit of analysis and conditioning as its tool.
To better appreciate Thorndikes theoretical con- (Not all would agree that psychology should be limit-
tribution, it is necessary to briefly recall the history of ed to the study of reflexes or conditioning, so cogni-
thought about learning. The dominant theme of the tive psychology has become quite popular in the late
psychological process involved in learning has always 1900s.)
been some sort of associationism, so that was not Second, Thorndike postulated that all learning
Thorndikes particular contribution. The change that involved the Law of Exercise (use and disuse) and the
THORNDIKE, EDWARD 649
Law of Effect. Therefore, from the comparative- behavior when it is fully developed (p. 126). He also
psychological viewpoint, according to Thorndike, an- predicted the relevance of the ratio to the slow initial
imals differed merely in the delicacy, number, com- or primary learning of higher vertebrates (especially
plexity, and permanence of associations. On these primates), in which the sensory projections are small
measures, dogs were somewhat more intelligent than relative to the size of the association areas: If the sen-
cats, dogs and cats exceeded fish, monkeys exceeded sory projection is small, association cortex large, the
dogs and cats, and humankind exceeded monkeys. [environmental] control will take longer; the period
(However, a real gauge of intelligence or learning of primary learning, that is, will be long (p. 124).
abilityor a genuine application of a gaugewas still Finally, the larger association area of the higher ver-
lacking.) According to Thorndike, nonhuman ani- tebrates (birds and mammals) would account for their
mals did not have ideas: Homo sapiens had ideas but greater efficiency at maturity. For Hebb, the learning
these were derived from learning via exercise and ef- capacity of higher species at maturity is not merely
fect. the capacity for a greater number of associations
(Thorndike); it also reflects an emancipation from di-
These issues (S-R association and the roles of be-
rect control by the stimulus of the moment from the
havioral activity and reinforcement) have remained
immediate environment (i.e., an evolutionary differ-
prominent to this day in general learning (behavior)
ence in central or psychological mediation).
theory, which one might define somewhat mischie-
vously as the noncomparative approach to animal in- In recent years, an important tome has appeared
telligenceif we understand by comparative the to put some meat on the bare bones of Thorndikes
search for species differences as well as similarities not and Hebbs speculations on the increase in the size of
only in behavioral adjustment per se but also in the the brain in the vertebrate lineage: Harry Jerisons
psychological processes mediating these adjustments. Evolution of the Brain and Intelligence (1973). There has
General behavior theory holds that principles of con- indeed been a progressive enlargement of the brain
ditioning are applicable across all vertebrate species, even when the general increase in body size in higher
including humans, and that cognitive psychology will versus lower vertebrates is calculated in the equation.
one day be explicable in terms of principles of condi- Birds and mammals stand out conspicuously in en-
tioning. The power or influence of conditioning theo- cephalization quotient when compared with lower
ry has waned considerably in the final decades of the vertebrates of the same body size: Birds and mam-
1900s, whereas cognitive psychology is in the ascen- mals have extra neurons when their brain size is
dancy. compared to the size it ought to be, given a high cor-
relation between brain and body size across species
Third, one of Thorndikes least appreciated con-
(not, however, across individuals within a species).
tributions to comparative psychology is embodied in
The conventional explanation for the increase in
his notion that as we ascend the vertebrate series of
brain size in birds and mammals is that they had in-
animals, we likely have the possibility for the learning
vaded new niches in which there was an adaptive ad-
of more associations more quickly and lastingly be-
vantage for enlarged brains (Jerison, 1973, p. 16).
cause the trend is for the brain to be larger and thus
to have more connections (synapses) in it. Thorndike With Jerisons elegant statistical formulas we at
does not give us his authority for this generalization last have a metric or gauge (however gross it may be)
about the evolution of the brain, but likely he was fol- for the evolution of the brain. But what of the evolu-
lowing Herbert Spencer and the writings of more con- tion of intelligence? Is it valid to continue to ask
temporaneous, neurologically well-versed writers whether intelligence (learning ability) has evolved in
such as the Herrick brothers (Clarence Luther, the the usual sense that transcends ecological niches and
originator of psychobiology, and his younger brother ecological considerations? If so, can it be meaningful-
C. Judson). In any event, this grand generalization rely and validly measured in the laboratory? What is the
surfaced, with new trappings, in Hebbs (1949) con- psychological gauge? These issues bedevil us to the
cept of the A/S ratio: the ratio of association area to present day with supporters on both sides of the con-
sensory projection area of the brain. Although there troversy.
are no exact figures available, the ratio of association [Some material for this entry was excerpted, with per-
to primary sensory areas increases rather remarkably mission, from G. Gottlieb (1979), Comparative psychology
from the lower vertebrates (fishes, amphibians, rep- and ethology, in E. Hearst, ed.,The first century of ex-
tiles) to the higher ones (birds and mammals). perimental psychology. Hillsdale, NJ: Erlbaum.]
By inference, Hebb (1949) assumed this ratio to
be relevant to the greater speed with which the Bibliography
lower species can learn to respond selectively to the Bitterman, M. E. (1965). Phyletic differences in learning. American
environment, and to the comparative simplicity of the Psychologist 20, 396410.
650 TIP-OF-THE-TONGUE PHENOMENON
Gottlieb, G. (1984). Evolutionary trends and evolutionary origins: mation available about the target word, yet cannot re-
Relevance to theory in comparative psychology. Psychological trieve the word. Also, subjects sometimes report an al-
Review 91, 448456.
Hebb, D. O. (1949). The organization of behavior. New York: Wiley.
ternate word that comes to mind and seems to
Hodos, W., and Campbell, C. B. G. (1969). Scala naturae: Why block the retrieval of the target word. These block-
there is no theory in comparative psychology. Psychological Re- ers often share semantic (i.e., meaning) or phono-
view 76, 337350. logical (i.e., sound) features with the target word (e.g.,
Jennings, H. S. (1906). The behavior of lower organisms. New York: abide for the word abdicate). Diary studies of naturally
Macmillan.
Jerison, H. (1973). Evolution of the brain and intelligence. New York: occurring TOTs have revealed that most often TOTs
Academic Press. are resolved spontaneously, such that the target word
Loeb, J. (1912; reprint 1964). The mechanistic conception of life, ed. seems to simply pop into mind after previous re-
D. Fleming. Cambridge, MA: Harvard University Press. trieval attempts have been abandoned (Burke, MacK-
Sechenov, I. M. (1863; reprint 1965). Reflexes of the brain, trans. S.
ay, Worthley, and Wade, 1991; Heine, Ober, and
Belsky. Cambridge, MA: MIT Press.
Thorndike, E. L. (1911; reprint 1965). Animal intelligence. New Shenaut, 1999).
York: Hafner.
There have been two competing theoretical ex-
Watson, J. B. (1914). Behavior: An introduction to comparative psycholo-
gy. New York: Henry Holt. planations for why TOT states occur during memory
(1919). Psychology from the standpoint of a behaviorist. Philadel- retrieval. The blocking hypothesis states that TOTs
phia: Lippincott. are caused by an alternate, more accessible word that
Yerkes, R. M., and Morgulis, S. (1909). The method of Pawlow in first comes to mind that then serves to block or inhibit
animal psychology. Psychological Bulletin 6, 257273.
the retrieval of the correct target word. Support for
Gilbert Gottlieb the blocking explanation comes from the experimen-
tal finding that presenting a phonologically related
cue word with a definition for a target word resulted
in more TOTs than when an unrelated cue word was
TIP-OF-THE-TONGUE PHENOMENON presented (Jones, 1989; Jones and Langford, 1987).
The tip-of-the-tongue (TOT) phenomenon refers to Similar results have been found by presenting cues
the experience of feeling confident that one knows an that share orthography (letters) with the target word
answer, yet is unable to produce the word. For exam- (Smith and Tindell, 1997). Also, it appears that TOTs
ple, in conversation or writing most people have had are more difficult to resolve when an alternate word
the occasional experience of trying, but failing to re- has come to mind than when there is no alternate
trieve someones name or a word from memory. This word (Burke et al., 1991), possibly suggesting that the
type of memory retrieval has been referred to as a tip- alternate word blocks the target word and thus causes
of-the-tongue (TOT) state because one experiences the TOT.
the frustrating feeling that the retrieval of the word On the other hand, the incomplete activation hy-
is imminent and on the tip of the tongue. Although pothesis states that blocker words are merely the
psychologists have long been aware of this phenome- consequence, not the cause of TOTs. According to
non, Roger Brown and David McNeill (1966) con- this explanation, TOTs are caused by the weak or in-
ducted one of the first experimental studies of TOT complete activation of the target word. Semantic in-
states. In this study, they attempted to experimentally formation about the target word may be activated, but
induce TOT states in college students by presenting the corresponding phonological representation may
definitions of relatively rare words (e.g., to give up the be only partially activated. A more accessible, phono-
throne). The subjects task was to name the word for logically similar alternate word may become activated
the definition (e.g., abdicate). Brown and McNeill and at first appear to block the target word. Based
found that they could induce a TOT state on approxi- on the incomplete activation hypothesis, one would
mately 10 percent of trials. expect that providing a phonologically related cue
Subsequent research on TOT states has been con- word with the definition should actually facilitate tar-
ducted by experimentally inducing TOT states in the get retrieval, rather than produce a TOT state. Meyer
laboratory or asking subjects to keep diaries of every- and Bock (1992) provided initial evidence for the in-
day occurrences of TOTs (Brown, 1991). These complete activation hypothesis. In three experiments
studies have yielded varied characteristics of TOT that were designed to address these competing expla-
states. For example, TOTs are more likely to occur nations (blocking versus incomplete activation),
for infrequently used words. Interestingly, when ex- Meyer and Bock reported (1) semantically and pho-
periencing a TOT state, subjects are often able to ac- nologically related cue words facilitated rather than
curately retrieve information about the non-recalled hindered target word retrieval (contrary to Joness
word (e.g., the first letter of the word or the number findings); (2) phonological cues facilitated retrieval
of syllables). Thus, subjects seem to have partial infor- more than semantic cues; and (3) these related cues
TOLMAN, EDWARD C. 651
facilitated retrieval even after an initial unsuccessful Brown, A. S., and Nix, L. A. (1996). Age-related changes in the tip-
target retrieval attempt. Subsequent studies also have of-the-tongue experience. American Journal of Psychology 109,
7991.
provided evidence that processing phonologically re- Brown, R., and McNeill, D. (1966). The tip of the tongue phe-
lated words decreases TOT states and increases cor- nomenon. Journal of Verbal Learning and Verbal Behavior 5,
rect target responses (James and Burke, 2000), lend- 325337.
ing further support to the incomplete activation Burke, D., MacKay, D. G., Worthley, J., and Wade, E. (1991). On
explanation of TOTs. the tip of the tongue: What causes word finding failures in
young and older adults? Journal of Memory and Language 30,
Although most people experience TOTs occa- 542579.
sionally, there is evidence that TOTs may increase Heine, M. K., Ober, B. A., and Shenaut, G. K. (1999). Naturally oc-
with age. TOT experiences represent a common curring and experimentally induced tip-of-the-tongue expe-
riences in three adult age groups. Psychology and Aging 14,
memory complaint of older adults and thus have been 445457.
extensively studied in the older adult population. In- James, L. E., and Burke, D. M. (2000). Phonological priming ef-
deed, both laboratory and diary studies indicate that fects on word retrieval and tip-of-the-tongue experiences in
older adults report more TOT experiences and have young and older adults. Journal of Experimental Psychology:
less partial information available about target words Learning, Memory, and Cognition 26, 1,3781,391.
Jones, G. V. (1989). Back to Woodworth: Role of interlopers in the
than younger adults (Brown and Nix, 1996; Burke et tip-of-the-tongue phenomenon. Memory & Cognition 17, 69
al., 1991; Heine et al., 1999). The literature has been 76.
somewhat inconsistent regarding whether older Jones, G. V., and Langford, S. (1987). Phonological blocking in the
adults report more or fewer alternate words (i.e., tip of the tongue state. Cognition 26, 115122.
blockers) that come to mind while in a TOT state. Maril, A., Wagner, A. D., and Schacter, D. L. (2001). On the tip of
the tongue: An event-related fMRI study of semantic retrieval
There appears to be no age differences in terms of the failure and cognitive conflict. Neuron 31, 653660.
percentage of TOTs resolved or the time to resolu- Meyer, A. S., and Bock, K. (1992). The tip-of-the-tongue phenome-
tion. As might be expected, theoretical accounts for non: Blocking or partial activation. Memory & Cognition 20,
the increase in TOTs as a function of old age have fo- 715726.
cused on age-related deficits in word-retrieval due to Smith, S. M., and Tindell, D. R. (1997). Memory blocks in word
fragment completion caused by involuntary retrieval of ortho-
(1) the interfering effects of related words that act as graphically related primes. Journal of Experimental Psychology:
blockers that inhibit the retrieval of the target words; Learning, Memory, and Cognition 23, 355370.
or (2) incomplete activation of the target due to de-
graded connections between the semantic representa- Janet M. Duchek
tion and the phonological representation of the word Jessica M. Logan
(Burke et al., 1991; James and Burke, 2000).
There also has been an interest in the neural cor-
relates of TOTs as a reflection of memory retrieval
failure. Using neuroimaging techniques (event- TOLMAN, EDWARD C. (18861959)
related fMRI), Maril, Wagner, and Schacter (2001)
scanned subjects while answering general knowledge The American psychologist Edward Chace Tolman
questions. Subjects indicated whether they knew the was a forerunner of modern cognitive psychology; he
answer (successful retrieval), did not know the answer showed that animals in learning mazes acquire orga-
(unsuccessful retrieval), or were in a TOT state. The nized spatial and temporal information about the
results indicated that there was a selective response in maze and about the consequences of various alterna-
anterior cingulate-prefrontal cortices of the brain tive behaviors. In developing this approach, he was
during a TOT state, relative to successful and unsuc- combating the dominant views of his time, which em-
cessful retrievals. This is interesting because this area phasized the acquisition of conditioned reflexes rath-
of the brain has been associated with the control and er than knowledge about environmental events. Al-
resolution of cognitive conflict. Thus, these neuroi- though several short biographies or reviews of
maging results seem to correspond with the behavior- Tolmans contributions are available (Crutchfield,
al experience of TOT states. 1961; Crutchfield et al., 1960; Hilgard, 1980; Innes,
1999, 2000; McFarland, 1993; Ritchie, 1964; Tol-
The experimental study of TOTs has provided man, 1952), it is especially appropriate that one be in-
psychologists with valuable information on the pro- cluded in an encyclopedia of learning and memory
cess of memory retrieval. Neuroimaging data may because workers in this field today are using ideas that
further delineate the neural underpinnings of this ev- were initiated and often developed by Tolman, al-
eryday memory phenomenon. though they do not necessarily recognize the source.
Bibliography Tolmans concepts and findings have helped to shape
Brown, A. S. (1991). A review of the tip-of-the-tongue experience. modern understanding of learning, memory, and
Psychological Bulletin 109, 204223. cognition.
652 TOLMAN, EDWARD C.
of Giessen in Germany, where he studied with Kurt

Koffka, one of the founders of Gestalt psychology.
In 1915 Tolman married Kathleen Drew and re-
ceived his Ph.D. He then spent three years as an in-
structor at Northwestern University before accepting
a position at the University of California at Berkeley
in 1918. Except for brief periods, Tolman spent the
rest of his life at Berkeley, where he had a distin-
guished scientific career and was an intellectual lead-
er in the university community.
Early Experiments in Animal Learning

The line of research that occupied most of Tol-
mans life started when, on arriving in Berkeley, he
found, he later wrote, that it was up to me to suggest
a new course. Remembering Yerkes course and Wat-
sons textbook, I proposed Comparative Psychology,
and it was this that finally launched me down the be-
haviorist slope (1952, p. 329). This slope may have
been behaviorist, but it was of a new and unusual kind
that reflected Tolmans education at Harvard.
In his early experiments and papers, Tolman fo-
cused on the the rats behavior in the maze to the ex-
clusion of other types of apparatuses because it gave
Edward C. Tolman (Psychology Archives, University of Akron)
opportunities for observing the animals solution to
problems in space, in getting from here to there. He
believed that when a rat ran from the start of a maze
Early Life to the goal, its behavior reflected a purposegetting
Tolman was born in Newton, Massachusetts, on to the goal in order to get somethingand knowledge
April 14, 1886, into a prosperous family that valued about the spatial layout. In referring to such knowl-
hard work, high thinking, and social responsibility. edge, Tolman used terms such as sign-gestalt-
After high school he attended the Massachusetts In- expectation, which referred to his assumption that if, in
stitute of Technology, where his father served on the the presence of a certain sign (that is, the events at the
board of trustees. In his autobiography Tolman com- start box and on into the maze), the rat behaved in
ments, I went to MIT not because I wanted to be an a particular way, it would achieve certain goals. The
engineer but because I had been good at mathematics term gestalt referred to Tolmans assumption that the
and physics in high school and because of family pres- rat was acquiring a cognitive map that would allow
sure. After graduating from Technology (in electro- it to use its organized information in getting to the
chemistry), I became more certain of my own wants goal.
and transferred to Harvard for graduate work in phi- In Tolmans early writings, including his major
losophy and psychology (1952, p. 323). book, Purposive Behavior in Animals and Men (1932), he
Among the experiences at Harvard that Tolman maintained the neorealist argument that knowledge
mentions as having influenced his later life were and purpose could be directly observed in the behav-
Ralph Barton Perrys course in ethics, which, he ior of the rat in the maze. But by 1932 he was also
wrote, laid the basis for my later interest in motiva- working with a different idea: that knowledge and
tion and indeed gave me the main concepts (rein- purpose were inferences from behavior rather than
forced by a reading of McDougalls Social Psychology characteristics of behavior. These inferences Tolman
as part of the requirement of the course) which I have came to call intervening variables to convey the
retained ever since; . . . Holts seminar in epistemolo- idea that knowledge and purpose intervene between
gy in which I was introduced to and excited by the the stimulus and behavior and guide the behavior
New Realism; and Yerkes course in comparative, (Tolman, 1938). In his autobiography Tolman (1952)
using Watsons Behavior: An Introduction to Comparative takes the position that such intervening variables not
Psychology, which was just out, as a text (p. 325). Tol- only serve as summary statements that bring together
man also spent the summer of 1912 at the University data but also refer to real, presumably causal events.
TOLMAN, EDWARD C. 653
Latent Learning Experiments Although Tolman, like his contemporaries,

thought mostly in terms of the plasticity of behavior,
Tolman and his students conducted a vigorous,
he did not ignore genetic influences. In fact, in 1924
broad program of research on learning and problem
he was the first to apply the technique of selective
solving in rats that served both to test his ideas and
breeding to the study of genetics of behavior, obtain-
to change them in the light of new data. Two lines of
ing maze-bright and maze-dull strains of rats.
research will be mentioned briefly here. The first, la-
His student Robert Tryon then carried out a success-
tent learning experiments, showed that rats learn
ful program of selective breeding for maze ability
about the layout of a complex maze even though, in
over several generations. This was replicated in other
the absence of reward, they show little or no evidence laboratories and extended to other kinds of behavior.
of such learning. When, after some trials, they are first This clear evidence for the influence of genes on be-
rewarded in the goal box, they show almost error-free haviors was important in holding a place for behavior
behavior on the next trial. These latent learning ex- genetics during the period when environmentalism
periments demonstrated several points. First, learn- was dominant (McClearn and Foch, 1988).
ing is different from performance and is occurring
All of Tolmans research showed a remarkably co-
even when there is no clear evidence for it. Current
herent but nevertheless broad-ranging character. Al-
reviews show that research of this sort continues to
though he dissented from the animal-learning ortho-
grow and prove fruitful. Second, the latent learning
doxy from the 1930s through 1950s, Tolmans
experiments showed that rats gain organized knowl-
position had become a dominant one in animal learn-
edge of the maze that transcended the conceptual
ing by the 1980s and 1990s.
framework of stimulus-response. Third, animals learn
about rewards. This conclusion was inconsistent with
the dominant view of the era: that rewards determine Later Accomplishments
which behaviors are learned. Tolmans conclusion is Tolman received many honors, including elec-
consistent with much later research in Pavlovian con- tion to the Society of Experimental Psychologists, the
ditioning (Rescorla, 1978). National Academy of Sciences, the American Philo-
sophical Society, and the American Academy of Arts
and Sciences. He was an honorary fellow of the British
Groundbreaking Research Psychological Society and was awarded honorary de-
A second line of research, closely related to the grees by a number of universities. Tolman was presi-
first, directed a variety of cleverly constructed experi- dent of the American Psychological Association in
ments to the problem of whether the animal could use 1937, president of the Society for the Psychological
its knowledge of the maze to make inferences about Study of Social Issues in 1940, and vice president of
what to do in new situations. Tolmans team guided the American Association for the Advancement of Sci-
rats to the goal along a circuitous route for a number ence in 1942. The Fourteenth International Congress
of trial, then deprived it of that route, and then ex- of Psychology was scheduled to be held in the United
posed it to a variety of alternatives, one of which States in 1954, and Tolman was to be its president.
would lead more directly to the goal. The results When it became apparent that the United States, be-
showed that the animal was able to use its knowledge cause of its anticommunist policy, was likely to refuse
about the spatial arrangements in the room to make admission to many participants from abroad, the
the appropriate inference and take the direct route. venue was changed to Canada, and Tolman became
Other research by Tolman and his students aimed at copresident along with Canadian psychologist Ed-
control processes such as selective testing of alterna- ward A. Bott.
tive possible solutions (hypotheses and vicarious In 1949, Tolman took a leadership role in the
trial and error). Berkeley facultys resistance to the imposition of a loy-
alty oath by the university. Prevented from teaching,
At a time when learning theorists were still trying he spent the academic year of 19491950 away from
to establish the theory of learning, Tolman (1949) Berkeley. The nonsigners finally won their case in
published an article entitled There Is More than court in 1953, gaining recognition of tenure at the
One Kind of Learning. In it he proposed that some university, and Tolmans professorship was restored.
of the basic disputes about learning might be resolved
if investigators agreed that there are a number of See also: LEARNING THEORY: A HISTORY
kinds of learning: The theory and laws appropriate
to one kind may well be different to those appropriate Bibliography
to other kinds (p. 144). Some of the types of learning Crutchfield, R. S. (1961). Edward Chace Tolman. American Journal
that Tolman proposed are still under investigation. of Psychology 74, 135141.
654 TOLMAN, EDWARD C.
Crutchfield, R. S., Krech, D., and Tryon, R. C. (1960). Edward Ritchie, B. F. (1964). Edward Chace Tolman. Biographical Memoirs,
Chace Tolman: A life of scientific and social purpose. Science National Academy of Sciences, Vol. 37. New York: Columbia Uni-
131 714716. versity Press.
Hilgard, E. R. (1980). Edward Chace Tolman. Dictionary of American Tolman, E. C. (1920). Instinct and purpose. Psychological Review 27,
Biography, Supp. 6. New York: Scribners. 217233.
Innis, N. K. (1999). Edward Chace Tolman. In J. A. Garraty and (1924). The inheritance of maze-learning ability in rats.
M. C. Carnes, eds., American national biography, Vol. 21. New Journal of Comparative Psychology 4, 118.
York: Oxford University Press. (1932). Purposive behavior in animals and men. New York:
(2000). Edward Chace Tolman. In A. E. Kazdin, ed., Ency- Century.
clopedia of psychology, Vol. 8. Washington, DC: American Psy- (1938). The determiners of behavior at a choice point. Psy-
chological Association. chological Review 45, 141.
McClearn, G. E., and Foch, T. T. (1988). Behavioral genetics. In (1949). There is more than one kind of learning. Psychologi-
R. C. Atkinson, R. J. Herrnstein, G. Lindzey, and R. D. Luce, cal Review 27, 217233.
eds., Stevens handbook of experimental psychology, 2nd edition, (1952). Autobiography. In E. G. Boring et al., eds., A history
Vol. 1. New York: Wiley.
of psychology in autobiography, Vol. 4. Worcester, MA: Clark
McFarland, D. (1993). Animal behaviour: Psychobiology, ethology,
University Press.
and evolution. New York: Wiley.
Rescorla, R. A. (1978). Some implications of a cognitive perspective
on Pavlovian conditioning. In S. H. Hulse, H. Fowler, and W.
K. Honig, eds., Cognitive processes in animal behavior. Hillsdale, Mark R. Rosenzweig
NJ: Erlbaum. Donald A. Riley
U
UNDERWOOD, BENTON (19151994) phasized the minds structure and had given birth to
a scientific psychology, James stressed the minds ac-
Benton J. Underwood was one of the preeminent tivity. His ideas, combined with those of British phi-
leaders in the postwar development of research on losophers and the American spirit of doing and act-
the acquisition and retention of verbal materials ing, created a movement called American
(Keppel, 1997, p. 469). He studied verbal learning Functionalism. A student of Underwoods once com-
and memory in the 1940s at the University of Missou- mented: To my mind, the term functionalist and the
ri, then later at the University of Iowa, under such im- name Underwood are synonymous (Freund, 1998,
portant figures as Arthur W. Melton (Missouri), John p. 318).
A. McGeoch, and Kenneth W. Spence (both at Iowa).
In 1946, Underwood took a teaching position at The functionalist perspective within the newly de-
Northwestern University in Evanston, Illinois, where veloping field of psychology approached problems of
he remained until his retirement in 1983. Over four learning and memory in associationistic terms. How
decades he did groundbreaking work in associative are associations between verbal items learned and re-
learning, verbal discrimination, transfer of training, membered? As Underwood commented, I am an in-
distribution of practice, interference and forgetting, curable associationist (1982, p. 8). It seemed obvious
and the composition of memory. to him that in innumerable instances associations
played a dominant role in our learning and retention
of verbal material: When did Christopher Columbus
Antecedents discover the new world? Who is the current vice presi-
Hermann Ebbinghaus carried out the first sys- dent? Clearly, to answer such questions we must learn
tematic study of verbal learning and memory in 1885. an association between initially unrelated facts (e.g.,
Acting as both experimenter and subject, he learned between a name and date). The experimental task
lists of nonsense syllables (e.g., cak, roq) and then test- given to subjects typically involved paired-associate
ed his retention of them. His efforts provide our first learning: the presentation of pairs of items to a sub-
picture of an empirically generated forgetting curve. ject charged with learning an association between the
Ebbinghauss curve revealed substantial forgetting in members of each pair to facilitate the recall of the sec-
the hours immediately following original learning. ond item upon presentation of the first.
Why was there so much forgetting so soon after learn- Another important characteristic of the function-
ing? This puzzle Underwood later helped to solve. alist perspective was its use of stimulus-response (S-R)
At the start of the twentieth century, William language borrowed from behaviorism, which domi-
Jamess work was the chief influence on American nated American psychology until the 1950s. Paired-
psychology. Unlike German psychologists, who em- associate learning, for example, was described as
655
656 UNDERWOOD, BENTON
petition. On this view, forgetting occurs when new

learning competes with the original learning. In
1959, Underwood provided evidence for an addition-
al mechanism, that of unlearning. He showed that
original learning actually became unavailable for re-
call as new learning proceeded.
Nevertheless, additional research suggested that
the more powerful factor producing forgetting was
proactive inhibition. Proactive inhibition occurs when
information learned before the target material inter-
feres with retention (e.g., an old telephone number
intrudes on your attempt to recall a more recent num-
ber). Because memory researchers often had their
subjects learn and remember many lists, there was an
opportunity to examine forgetting as a function of the
number of prior lists acquired by subjects in the learn-
ing laboratory. In 1957, in a classic piece of detective
work, Underwood showed that retention of a single
list of items after twenty-four hours decreased as a
function of the number of prior lists that had been
learned (see Figure 1). Retention of a single list with-
out prior learning was about 75 percent, not just 25
percent, which had been a conclusion drawn from Eb-
binghauss forgetting curve. Ebbinghaus was affected
by the proactive inhibition resulting from learning
and recalling many, many lists. In later years Under-
wood and his students demonstrated how proactive
interference operates even in very short-term memo-
ry situations and provide further accounts of possible
Benton Underwood (Northwestern University Archives) sources of proactive interference in long-term reten-
tion of verbal material.
An associationistic perspective led Underwood to
learning an association between a stimulus (the first
conclude that associations of common words will like-
member of the pair) and a response (the second
ly occur implicitly during a learning task. An adult
member). Principles of classical conditioning were in- learning a list of words such as sugar, bitter, and candy
corporated into explanations of human learning and will likely implicitly think of the word sweet. In 1965,
retention of verbal material. Forgetting of verbal ma- Underwood showed that these implicit associative
terial, for example, was discussed in terms of extinc- responses, or IARs, played a role in producing false
tion and spontaneous recovery of associations, a lan- recognition. For example, learners who exper-
guage perfectly in tune with that of the behaviorists. ience words related to sweet, such as the above,
Other salient characteristics of the functionalist orien- later come to falsely state that they had actually expe-
tation to psychology include an insistence on modest rienced the word sweet. Underwood demonstrated a
theory building, a close interplay between theory and role for IARs in many different learning and memory
data, and an interest in analyzing phenomena in tasks.
terms of their component processes or parts.
Attributes of Memory
Learning and Memory Consider a simple learning task. A list of word
In 1932 McGeoch argued that forgetting was due pairs is presented. Your task is simply to remember
to the intervention of events that occured after the ac- which word in each pair is the right one (for example,
quisition of information and was not simply a matter as designated by the experimenter by underlining).
of decay or disuse. Forgetting caused by a learners When the word pairs are shown again (with no identi-
activities in the interval between original learning and fication of right items) can you remember which is the
a test of retention is called retroactive inhibition. Its right one? Of course you can, at least after a few learn-
primary mechanism was thought to be response coming trials. But how do you do it? In 1966, Underwood
UNDERWOOD, BENTON 657
and his students presented a frequency theory of ver- Figure 1

bal discrimination learning. It is a beautiful example
of the functionalist approach. The theory is modest
in scope (accounting mainly for verbal discrimination
decisions), closely tied to the data (making clear,
easily tested predictions), and permits analysis of a
larger phenomenon (verbal discrimination) into its
constituent processes. The theory posits that sub-
jects make decisions in verbal discrimination tasks
based on a subjective frequency differential be-
tween right and wrong members. This situational
frequency differential is built up through representa-
tional responses (a perceptual processing of the item),
rehearsal responses, and implicit associative re-
sponses.
This theory has survived many tests and has been
extended to other situations, accounting, for exam-
ple, for our ability to perform on a recognition mem-
ory test. Consider what sometimes happens on a mul- Percent recall of a single list as a function of the number of
tiple-choice exam. Initially, one alternative looks previous lists learned in the experiment.
right (familiar), but as we consider the other alterna-
tives, our confidence decreases. Frequency theory
suggests that the initial verbal discrimination is based See also: EBBINGHAUS, HERMANN
on a subjective difference in frequency that arises
from our prior study of the right item; but that differ-
Bibliography
ential is lost as we raise the frequency (familiarity) of
Barnes, J. M., and Underwood, B. J. (1959). Fate of first-list asso-
other alternatives. ciations in transfer theory. Journal of Experimental Psychology
Underwoods thinking about frequency as an at- 58, 97105.
tribute of memory led to a consideration of the com- Duncan, C. P., Sechrest, L., and Melton, A. W., eds. (1972). Human
memory: Festschrift in honor of Benton J. Underwood. New York:
position of memory. Or, as Underwood asked in Appleton-Century-Crofts.
1969, Of what does a memory consist? (p. 559). His Ekstrand, B. R., Wallace, W. P., and Underwood, B. J. (1966). A fre-
answer was that memory, our record of an event, is a quency theory of verbal discrimination learning. Psychological
collection of attributes. There is no corpus or body of Review 73, 566578.
memory that can be recalled directly. Memory attri- Freund, J. S. (1998). Benton J. Underwood: A tribute of memories.
In G. A. Kimble and M. Wertheimer, eds., Portraits of pioneers
butes both aid discrimination (e.g., frequency) and re-
in psychology, Vol. 3. Mahweh, NJ: Erlbaum.
trieval (e.g., IARs). Many everyday experiences pro- Keppel, G. (1997). Benton J. Underwood (19151994). American
vide evidence of such a conceptualization. Tip-of-the Psychologist 52, 469470.
tongue experiences, for example, sometimes involve McGeoch, J. A. (1932). Forgetting and the law of disuse. Psychologi-
an ability to remember letters of a word (orthographic cal Review 39, 352370.
attribute) when we cant remember the whole word. Postman, L. (1972). The experimental analysis of verbal learning
and memory: Evolution and innovation. In C. P. Duncan, L.
Many students have had the experience of being able
Sechrest, and A. W. Melton, eds., Human memory: Festschrift in
to remember where on a textbook page an answer was honor of Benton J. Underwood. New York: Appleton-Century-
studied but not being able to remember the item itself Crofts.
(spatial attribute). Underwoods attribute theory Underwood, B. J. (1949). Experimental psychology. New York: Apple-
helps to explain learners performance in many tasks ton-Century-Crofts.
and, along with other multidimensional theories, has (1957). Interference and forgetting. Psychological Review 64.
(1957). Psychological research. New York: Appleton-Century-
enhanced our understanding of memory. Crofts.
Underwood published more than 200 scientific (1965). False recognition produced by implicit verbal re-
articles, books, chapters, and monographs, and his sponses. Journal of Experimental Psychology 70, 122129.
(1969). Attributes of memory. Psychological Review 76, 559
work elicited a plethora of awards and honors, includ-
573.
ing election, in 1970, to the National Academy of Sci- (1982). Studies in learning and memory: Selected papers. New
ences. He was a distinguished teacher and superb York: Praeger.
methodologist whose research and textbooks in- Zechmeister, E. B., and Nyberg, S. E. (1982). Human memory: An in-
spired the work of countless students and colleagues. troduction to research and theory. Belmont, CA: Wadsworth.
Eugene B. Zechmeister
V
VESTIBULO-OCULAR REFLEX (VOR) aptation) (Gonshor and Melvill-Jones, 1974; Ito et al.,
PLASTICITY 1974; Robinson, 1976). Since the flocculus, an evolu-
tionarily old part of the cerebellum, sends inhibitory
The vestibulo-ocular reflex (VOR) is driven by head axons of Purkinje cells directly to the VOR relay neu-
movement and moves the eyes in the direction oppo- rons (Fukuda et al., 1972), the flocculus may be the
site to the head movement, automatically stabilizing center for the adaptation of VOR/OKR (Ito, 1982).
vision relative to space. The VOR is mediated by a The compensation and adaptation in the VOR/OKR
trineuronal arc composed of the primary vestibular provide simple model systems representing learning
neurons, relay neurons in vestibular nuclei, and mo- capabilities of neuronal circuits (Ito, 1984, 1998; Ray-
toneurons for the extraocular muscles. The optokine- mond et al., 1996).
tic eye-movement response (OKR) is another ocular
reflex induced by optokinetic stimuli generated by
movements of the visual field relative to the head. Neuronal Circuit for the VOR and OKR
VOR and OKR share the same vestibular nuclear neu- The VOR contains a number of component re-
rons and oculomotor neurons, and synergistically sta- flexes. They arise from the three semicircular canals
bilize vision. The VOR/OKR pathways are attached (horizontal, anterior, and posterior) and two otolith
with commissural connections between the bilateral organs (utricle and saccule) in each labyrinth and act
vestibular nuclei, internuclear connections between on six extraocular muscles (medial and lateral rectus,
motor nuclei, and other connections with the tegmen- superior and inferior rectus, superior and inferior
tum and the cerebellum. oblique) in each eye. In experiments, the VOR is usu-
ally divisible into horizontal, vertical, or torsional
Two types of plasticity have been recognized in
components through the application of yaw, pitch,
the VOR/OKR. The first is gradual recovery of the
roll, linear motion or static tilt to the head. When
heavy nystagmus induced by unilateral labyrinthecto-
moving freely, the VOR is mediated via the concerted
my (vestibular compensation) (Magnus, 1924; Vidal
operation of numerous parallel pathways linking the
et al., 1998). A lesion to the cerebellum retards the
ten sensors in two labyrinths with the twelve muscles
onset of recovery, and, when made after recovery, it
in two eyes (Ezure and Graf, 1984).
transiently removes the compensation. Hence, the
cerebellum may play a role in initiating this type of In the horizontal VOR, horizontal semicircular
plasticity (Courjon et al., 1982). The second is gradual canals are stimulated by ipsilateral head rotations that
changes of the VOR and OKR gain (amplitude ratio result in transmission of neural signals via the prima-
of eye rotation versus head or visual-field rotation) in ry vestibular nerve fibers to relay cells of the VOR in
altered visuovestibular environments (VOR/PKR ad- the vestibular nuclei. The VOR relay cells, in turn,
659
660 VESTIBULO-OCULAR REFLEX (VOR) PLASTICITY
transmit either excitatory or inhibitory signals to mo- cline in the horizontal slow-phase velocity, these cere-
toneurons in the abducens and oculomotor nuclei. bellar regions may control the velocity storage
Rotation of the head to one side induces contraction (Solomon and Cohen, 1994). The same system is re-
of the ipsilateral medial rectus and contralateral later- sponsible for optokinetic after-nystagmus (OKAN).
al rectus muscles, and relaxation of the ipsilateral lat-
eral rectus and contralateral medial rectus muscles, so In the VOR, the head velocity signals generated
that both eyes move in a direction opposite to that of by the labyrinth are converted to eye position signals
the head rotation. Optokinetic signals arising from for the extraocular muscles. In the OKR, the retinal
the retina are sent to VOR relay neurons, inducing slip velocity signals are also integrated to eye position
the OKR. signals. Hence, the VOR/OKR pathways must contain
a neural integrator mechanism (Robinson, 1975).
The integrator for the horizontal VOR/OKR may in-
Cerebellar Connections to VOR/OKR volve the nucleus prepositus hypoglossi and/or the
Pathways commissural inhibitory projections between bilateral
vestibular nuclei (Arnold and Robinson, 1997). The
The flocculus in rabbits and cats has five to six
integrator for the vertical VOR is provided by the in-
major folia and in rats only one. The classic anatomy
terstitial nucleus of Cajal in the midbrain (Fukushima
of the monkey cerebellum shows ten folia in the floc-
et al., 1992). The velocity storage and neural integra-
culus, but recent studies on neuronal connectivity
tor may represent functions of the same neuronal cir-
have revealed that the rostrally located five folia be-
cuit, but their identity remains unclear.
long to the ventral paraflocculus and not to the floc-
culus (Gerrits and Voogd, 1989). Purkinje cells, which
are localized to a narrow zone (H zone, about one mm
across) extending across the folia of the flocculus,
Compensation for Unilateral
supply inhibitory synapses to the horizontal VOR- Labyrinthectomy
relay neurons. Other zones related to the vertical or The behavioral recovery from unilateral labyrin-
torsional VOR flank the H zone (Nagao et al., 1985; thectomy in rats was accompanied by asymmetric ex-
Sato and Kawasaki, 1991; Van der Steen et al., 1994). pression of isoforms of protein kinase C (PKC , ,
Physiological experiments detect these zones by local and in the flocculonodular lobe, with a regionally se-
electrical stimulation, which induces distinct horizon- lective increase in the number of PKC-
tal, vertical, or rotatory eye movements. The flocculus immunopositive Purkinje cells contralateral to the le-
receives vestibular, optokinetic, and neck afferent sig- sion (Goto et al., 1997). This asymmetry occurred
nals via mossy fibers, and optokinetic and vestibular within six hours after the labyrinthectomy and was re-
signals via climbing fibers (Ito, 1984). solved to the control, symmetric pattern within twen-
The nodulus and uvula, other phylogenically old ty-four hours. The compensation was retarded in rats
parts of the cerebellum, receive vestibular mossy fiber after intracerebroventricular application of PKC in-
inputs through certain vestibular nuclear neurons, hibitors (Balaban et al., 1999). Since PKC is required
which receive Purkinje cell axons from the nodulus for induction of long-term depression (LTD) in Pur-
and uvula (Xiong and Matushita, 2000). The nodulus kinje cells, these observations suggests that LTD plays
and uvula receive optokinetic signals via climbing fi- a role in vestibular compensation. But because slow
bers. recovery of compensation still occurs after lesioning
the cerebellum, plastic changes should occur also out-
side the cerebellum. Reserachers have found an in-
Velocity Storage and Neural Integrator crease in GABAergic neurons in cat vestibular nuclei
Head rotation in darkness at a constant velocity after unilateral labyrinthectomy (Tighilet and La-
induces nystagmus, which consists of alternating slow cour, 2001); a neuronal network simulation suggests
phases representing the VOR and quick phases reset- a change in the commissural inhibitory connections
ting the eye position. Under this condition impulses (Graham and Dutia, 2001).
evoked in the first-order vestibular afferents decline
rapidly, within five seconds. Yet the slow-phase veloci-
ty of the nystagmus decays slowly, with a time constant VOR/OKR Adaptation
of about twenty seconds. There must be a brain-stem In laboratory experiments, sinusoidal or velocity-
mechanism, a velocity storage that stores the initial step, whole-body rotation in darkness induces the
head velocity as transduced by the canals and main- horizontal VOR; the VOR gain is the ratio of the at-
tains it despite the decay in the firing rate of the canal tained eye velocity vs. the applied head velocity. The
afferents (Raphan et al., 1979). Since electrical stimu- sinusoidal rotation is convenient for measuring the
lation of the nodulus and uvula induces a rapid de- gain and phase of the VOR separately, whereas veloci-
VESTIBULO-OCULAR REFLEX (VOR) PLASTICITY 661
ty steps enable us to separate components of VOR re- study on the monkey vertical VOR (Hirata and High-
sponses, which arise with different latencies. stein, 2001) suggests that adaptive changes occur in
both the flocculus and nonflocculus pathways.
The VOR exhibits marked adaptive changes in
gain under sustained mismatching between move-
ment of the head in space and movement of the visual Neuronal Mechanisms of VOR Adaptation
surroundings. Long-term visuovestibular mismatch- The major signals of Purkinje cells are simple
ing of days or months can be generated using prism spikes, elicited by mossy-fiber inputs. In rabbits, ipsi-
or lens goggles (Gonshor and Melvill-Jones, 1974; lateral horizontal head rotation causes either an in-
Robinson, 1976). However, for short-term mismatch- crease (in-phase type) or decrease (out-of-phase type)
ing of one to four hours, it is convenient to use the in the simple spike discharges from H-zone Purkinje
combined rotation of the turntable on which the ani- cells. During sustained visuovestibular mismatching,
mal is mounted and a screen representing the visual the simple spike modulation of H-zone Purkinje cells
surroundings (Ito et al., 1974; Nagao, 1989). Both the becomes predominantly out-of-phase or in-phase,
wearing of Dove prism goggles, which reverse the corresponding to the increase or decrease in the VOR
right-left axis of the visual field, and the in-phase ro- gain, respectively (Ito, 1984, 1998). Complex spikes
tation of the turntable and screen in the same direc- representing climbing fiber signals are evoked by op-
tion cause adaptive reduction of horizontal VOR gain. tokinetic stimuli. Researchers have assumed that
Both the wearing of magnifying lenses (2x) and the these complex spikes encode retinal slips, but a study
out-of-phase rotation of the turntable and screen in by Frens and colleagues suggests that they encode the
opposite directions cause an adaptive increase in hor- eye-movement performance error rather than retinal
izontal VOR gain. The OKR gain adaptively increases slip (2001). The changes in simple spike-response
during sustained rotation of the visual field around a patterns are supposed to result from LTD induced by
stationary animal. error signals of climbing fibers and to cause the VOR
The view that the flocculus is the center for the gain changes (Ito 1982, 1998). Nevertheless, studies
VOR/OKR adaptation gains support from the obser- of the monkey flocculus/paraflocculus suggest that
vation that the adaptation no longer occurs after le- VOR adaptation is the cause, not the consequence, of
sioning the flocculus, interruption of the climbing the simple spike modulation in Purkinje cells that re-
fiber input to the interruption of the climbing fiber flect eye velocity changes (Miles and Lisberger, 1981;
input to the flocculus, or deprivation of NO that is re- Hirata and Highstein, 2001).
quired for induction of LTD, as tested in various ani-
mals. Subdural application of a NO scavenger to the Comments
rabbit and monkey flocculus blocked VOR adaptation
(Nagao and Ito, 1991). Injection of a NO scavenger Although there is a scholarly consensus about the
or a NOS inhibitor into the goldfish cerebellum (Li important roles of the cerebellum in the VOR plastici-
et al., 1995) inhibited adaptive increase in VOR gain. ty, its mechanisms remain unclear, partly because of
NO synthase (NOS)-deficient mutant mice lacked the use of different animal species and different types
OKR adaptation (Katoh et al., 2000). Induction of of VOR. The sampling of Purkinje cells from both the
LTD also requires activation of PKC; accordingly, flocculus and ventral paraflocculus also complicate in-
transgenic mice that selectively express the pseudo- terpretation because these areas have substantially
substrate PKC inhibitor, PKC [19-31], in Purkinje different input and output connections (Gerrits and
cells lacked VOR adaptation (De Zeeuw et al., 1998). Voogd, 1989; Nagao et al., 1997a, b). It should be
noted that these areas have different functional as-
Researchers have studied the effects of the acute signments: The flocculus plays a role in the control of
removal of flocculus functions after the development VOR and OKR, whereas the ventral paraflocculus
of the VOR adaptation to determine whether or not plays roles in ocular following movement (Shidara et
the once-established VOR adaptation is retained in al., 1993) and in the smooth pursuit of a moving tar-
the flocculus. Microdialysis of lidocaine into the gold- get (Stone and Lisberger, 1990). These complications
fish cerebellum abolished both induction and reten- pose the need for further research.
tion of the VOR adaptation (McElligot et al., 1998),
suggesting that the memory for the VOR adaptation See also: GUIDE TO THE ANATOMY OF THE BRAIN:
is retained, probably entirely, in the cerebellum. In CEREBELLUM; LONG-TERM DEPRESSION IN THE
another experiment on goldfish, however, no less CEREBELLUM, HIPPOCAMPUS, AND NEOCORTEX
than 30 percent of the altered VOR gain was retained Bibliography
after ablation of the cerebellum (Pastor et al., 1994), Arnold, D. B., and Robinson, D. A. (1997). The oculomotor inte-
a result that favors the view that the memory is partly grator: Testing of a neural network model. Experimental Brain
stored outside the cerebellum. A system identification Research 113, 5774.
662 VESTIBULO-OCULAR REFLEX (VOR) PLASTICITY
Balaban, C. D., Preilino, M., and Romero, G. G. (1999). Protein ki- Miles, F. A., and Lisberger, S. G. (1981). Plasticity in the vesti-
nase C inhibition blocks the early appearance of vestibular buloocular reflex: A new hypothesis. Annual Review of Neuro-
compensation. Brain Research 845, 97101. science 4, 273299.
Courjon, J. H., Flandrin, J. M., Jeannerod, M., and Schmid, R. Nagao, S. (1989). Behavior of floccular Purkinje cells correlated
(1982). The role of the flocculus in vestibular compensation with adaptation of vestibuloocular reflex in pigmented rab-
after hemilabyrinthectomy. Brain Research 239, 251257. bits. Experimental Brain Research 77, 531540.
De Zeeuw, C. I., Hansel, C., Bian, F., Koekkoek, S. K. E., Van Alp- Nagao, S., and Ito, M. (1991). Subdural application of hemoglobin
hen, A. M., Linden, D. J., and Oberdick. J. (1998). Expression to the cerebellum blocks vestibuloocular reflex adaptation.
of a protein kinase C inhibitor in Purkinje cells blocks cerebel- NeuroReport 2, 193196.
lar LTD and adaptation of the vestibulo-ocular reflex. Neuron Nagao, S., Ito, M., and Karachot, L. (1985). Eye field in the cerebel-
20, 495508. lar flocculus of pigmented rabbit determined with local elec-
Ezure, K., and Graf, W. (1984). A quantitative analysis of the spatial trical stimulation. Neuroscience Research 3, 3951.
organization of the vestibulo-ocular reflexes in lateral- and Nagao, S., Kitamura, T., Nakamura, N., Hiramatsu, T., and Yama-
frontal-eyed animals. II. Neuronal networks underlying vesti- da, J. (1997a). Differences of the primate flocculus and ventral
bulo-oculomotor coordination. Neuroscience 12, 95109 paraflocculus in the mossy and climbing fiber input organiza-
Frens, M. A., Mathoera, A. L., and van der Steen, J. (2001). Floccu- tion. Journal of Comparative Neurology 382, 480498.
lar complex spike response to transparent retinal slip. Neuron (1997b). Location of efferent terminals of the primate floc-
30, 796801. culus and ventral paraflocculus revealed by anterograde axo-
Fukuda, J., Highstein, S. M., and Ito, M. (1972). Cerebellar inhibi- nal transport methods. Neuroscience Research 27, 257269.
tory control of the vestibulo-ocular reflex investigated in rab- Pastor A. M., De Cruz, R. R., and Baker, R. (1994). Cerebellar role
bits IIIrd nucleus. Experimental Brain Research 14, 511526. in adaptation of the goldfish vestibuloocular reflex. Journal of
Fukushima, K., Kaneko, C. R. S., and Fuchs, A. F. (1992). The neu- Neurophysiology 72, 1,3831,394.
ronal substrate of integration in the oculomotor system. Prog- Raphan, T., Matsuo, V., and Cohen, B. (1979). Velocity storage in
ress in Neurobiology 39, 609639. the vestibulo-ocular reflex arc (VOR). Experimental Brain Re-
Gerrits, N. M., and Voogd, J. (1989). The topographical organiza- search 35, 229248.
tion of climbing and mossy fiber afferents in the flocculus and Raymond, J. L., Lisberger, S. G., and Mauk, M. D. (1996). The cer-
the ventral paraflocculus in rabbit, cat and monkey. Experi- ebellum: A neuronal learning machine? Science 272, 1,126
mental Brain Research Series 17, 2629. 1,131.
Gonshor, A., and Melvill-Jones, G. M. (1974). Extreme vestibulo- Robinson, D. A. (1975). Oculomotor control signal. In G. Lennerst-
ocular adaptation induced by prolonged optical reversal of vi- rand and P. Bach-y-Rita, eds., Basic mechanisms of ocular motility
sion. Journal of Physiology, London 256, 381414. and their clinical implications. Pergamon: Oxford.
Goto, M. M., Romero, G. G., and Balaban, C. D. (1997). Transient (1976). Adaptive gain control of vestibuloocular reflex by
changes in flocculonodular lobe protein kinase C expression the cerebellum. Journal of Neurophysiology 39, 954969.
during vestibular compensation. Journal of Neuroscience 17, Sato, Y., and Kawasaki, T. (1991). Identification of the Purkinje
4,3674,381. cell/ climbing fiber zone and its target neurons responsible for
Graham, B. P., and Dutia M. B. (2001). Cellular basis of vestibular eye movement control by the cerebellar flocculus. Brain Re-
compensation: Analysis and modeling of the role of the com- search Review 16, 3964.
missural inhibitory system. Experimental Brain Research 137,
Shidara, M., Kawano, M., Gomi, H., and Kawato, M. (1993). In-
387396.
verse-dynamics model eye movement control by Purkinje cells
Hirata, Y., and Highstein, S. M. (2001). Acute adaptation of the
in the cerebellum. Nature 365, 5052.
vestibuloocular reflex: signal processing by fluccular and ven-
Solomon, D. and Cohen, B. (1994). Stimulation of the nodulus and
tral parafloccular Purkinje cells. Journal of Neurophysiology 85,
uvula discharges velocity storage in the vestibulo-ocular re-
2,2672,288.
flex. Experimental Brain Research 102, 5768.
Ito, M. (1982). Cerebellar control of the vestibulo-ocular reflex-
Stone, L. S., and Lisberger, S. G. (1990). Visual responses of Pur-
around the flocculus hypothesis. Annual Review of Neuroscience
kinje cells in the cerebellar flocculus during smoothpursuit
5, 275296.
eye movements in monkeys. I. Simple spikes. Journal of Neuro-
(1984). The Cerebellum and Neural Control. Raven Press: New
physiology 63, 1,2411,261.
York.
Tighilet, B., and Lacour, M. (2001). Gamma amino butyric acid
(1998). Cerebellar learning in the vestibulo-ocular reflex.
Trends in Cognitive Science 2, 313321. (GABA) immunoreactivity in the vestibular nuclei of normal
Ito, M., Shiida, N. Yagi, N., and Yamamoto, M. (1974). The cere- and uynilateral vestibular neurectoimized cats. European Jour-
bellar modification of rabbits horizontal vestibulo-ocular re- nal of Neuroscience 13, 2,2552,267.
flex induced by sustained head rotation combined with visual Van der Steen, J., Simpson, J. I., and Tan, J. (1994). Functional and
stimulation. Proceedings of Japan Academy 50, 8589. anatomic organization of three-dimensional eye movements
Katoh, A., Kitazawa, H., Itohara, S., and Nagao, S. (2000). Inhibi- in rabbit cerebellar flocculus. Journal of Neurophysiology 72, 31
tion of nitric oxide synthesis and gene-knockout of neuronal 46.
nitric oxide synthase impaired adaptation of mouse optokine- Vidal, P.-P., de Waele, C., Vibert, N., and Muhlethaler, M. (1998).
tic response eye movements. Learning and Memory 7, 220226. Vestibular compensation revisited. Otolaryngology-Head and
Li, J., Smith S. S., and McElligott, J. G. (1995). Cerebellar nitric Neck Surgery, 119, 3442.
oxide is necessary for vestibulo-ocular reflex adaptation, a Xiong, G., and Matushita, M. (2000). Connections of Purkinje cell
sensorimotor model of learning. Journal of Neurophysiology 74, axons of lobule X with vestibulocerebellar neurons projecting
489494. to lobule X or IX in the cat. Experimental Brain Research 133,
Magnus, R. (1924). Korperstellung. Berlin: Springer. 219228.
McElligot, J. D., Beeton, P., and Polk, J. (1998). Effect of cerebellar
inactivation by lidocaine microdialysis on the vestibuloocular
reflex in goldfish. Journal of Neurophysiology 79, 1,2861,294. Masao Ito
VISUAL MEMORY, BRIGHTNESS AND FLUX IN 663
VISUAL MEMORY neocortex cannot discriminate visual patterns

(Lavond and Dewberry, 1980). However, rudimenta-
See: MODALITY EFFECTS; PRIMATES, VISUAL ry visual functions remain in that decorticated rats
PERCEPTION AND MEMORY IN NONHUMAN;
can see moving objects, can detect the deep side of a
VISUAL MEMORY, BRIGHTNESS AND FLUX IN
visual cliff, and can learn or relearn the brightness
discrimination.
Importantly, the results of brightness discrimina-
tion training illustrate the property of behavioral re-
VISUAL MEMORY, BRIGHTNESS covery of function following brain injury. Lashley
AND FLUX IN trained rats on a brightness discrimination and then
The study of simple visual discriminations reveals systematically removed fractions of the cerebral neo-
fundamental properties of learning and memory in cortex. After removal of visual neocortex there was no
the nervous system. Physical measures of the light evidence for retention of the brightness discrimina-
source include energy emitted (flux) or reflected (re- tion. With continued training, however, the rats reac-
flectance) from the stimulus per unit area. Heinrich quired the discrimination, taking as many trials as ini-
Klver (1942) called it brightness if the total amount of tially required to learn. One possible conclusion
light was measured over the whole stimulus source would be that the lesion destroyed the memory and,
(including contours and edges), and density of luminous once it was removed, a new memory could be reestab-
flux if measured over a unit area of the stimulus. He lished with the same effort. However, this is a some-
concluded that visually decorticated monkeys could what fortuitous result in that it is an outcome of the
solve a luminous-flux problem but not a brightness training criterion used, because posteriorly decorti-
problem. cated subjects perform better when trained with less
stringent criteria and worse when trained to succes-
Using this definition, a brightness discrimination sively more stringent criteria (Spear and Braun,
includes both light intensity and light contrast on 1969). This indicates that normal learning and learn-
edges that contribute to pattern perception. A flux ing by posteriorly decorticated subjects are by differ-
discrimination, on the other hand, pertains to different mechanisms.
ences in light intensity per unit area. This intensity
can be for the whole stimulus (total flux) or for parts The recovery of brightness discriminations is con-
of the stimulus (local flux). For example, a horizontal- sistent with Lashleys previous observations on maze
versus-vertical-stripes pattern problem can be equat- learning (1929) where he found equipotentiality (all
ed for total flux and for total brightness (same num- parts of the neocortex have the same capacity for sup-
ber of contrasting contours) and yet the edges of the porting memory) and mass action (large areas of neo-
stimulus cards have differences in local flux. Discrimi- cortex contribute to the memory). There is one cave-
nation tasks for intensity that are used for testing ani- at, however, in that he confined these properties to
mals, whether black and white cards with edges or visual cortex for visual discriminations. Lashley did
black and white alleys with an edge that creates a con- not think that the anterior neocortex participated in
trast (the wall separating the alleys), normally are not visual discriminations as it does for maze learning.
purely flux discriminations but are brightness dis- However, more recent work supports the generaliza-
criminations that include elements of pattern dis- tion of visual function to the entire neocortex (Cloud,
crimination. Discriminations more purely restricted Meyer, and Meyer, 1982).
to flux can be achieved by use of contact lenses that Bauer and Cooper (1964) suggested that memory
diffuse light and obscure edges, by successive discrim- was not stored in the neocortex at all, but that visual
ination problems in which both alleys being lit means cortex was necessary for seeing the stimulus as a pat-
to go one way or both alleys being dark means to go tern discrimination (a brightness discrimination).
the other way, or by a shuttle box in which the subject They suggested that learning after a visual cortical le-
sits in the middle of the apparatus, one alley is lit, and sion was by using a different stimulus feature (i.e., a
the alley 180 degrees away is not. The distinguishing flux discrimination). However, there is evidence that
feature between a flux and a brightness discrimina- both brightness and flux are learned simultaneously.
tion is that the latter has contours created by contrast- Meyer and Meyer showed that a neural stimulant fa-
ing levels of flux within the same stimulus. cilitates recovery of the flux discrimination but not a
In the simplest visual discrimination a subject dis- pattern discrimination (see Meyer and Meyer, 1977,
tinguishes between a black and a white stimulus card. for review). They suggest that the role of the neo-
With this task Karl Lashley (1935) studied the role of cortex is to add context in facilitating access to sub-
cerebral neocortex in his search for the physical man- cortically established engrams rather than to act as a
ifestation of memory, for the engram. Rats without any store for memories. LeVere and Morlock supported
664 VISUAL MEMORY, BRIGHTNESS AND FLUX IN
this conclusion using a behavioral interference test, a Bibliography

simultaneous brightness discrimination (lit versus Bauer, J. H., and Cooper, R. M. (1964). Effects of posterior cortical
dark alleys, 1973), and a successive flux discrimina- lesions on performance of a brightness discrimination task.
Journal of Comparative and Physiological Psychology 58, 8493.
tion (both alleys lit means go one way, both dark Cloud, M. D., Meyer, D. R., and Meyer, P. M. (1982). Induction of
means go the other way, 1974). Rats were initially recoveries from injuries to the cortex: Dissociation of equipo-
trained to one habit, then had the visual neocortex re- tential and regionally specific mechanisms. Physiological Psy-
moved. If the cortical lesion actually destroyed the chology 10, 6673.
Hunter, W. S. (1930). A consideration of Lashleys theory of the
memory, then it should not matter whether the sub-
equipotentiality of cerebral action. Journal of Genetic Psychology
jects were trained to the same habit or to the opposite 3, 455468.
habit (go the other way). LeVere and Morlock found Klver, H. (1942). Functional significance of the geniculo-striate
that training to the opposite habit took substantially system. In H. Klver, ed., Visual mechanisms. Lancaster, PA:
Jaques Cattell Press.
longer to learn, indicating that the old memory still
Lashley, K. S. (1929). Brain mechanisms and intelligence: A quantitative
existed and interfered with learning the opposite study of injuries to the brain. Chicago: University of Chicago
habit. Press.
(1935). The mechanism of vision: XII. Nervous structures
As was true in Lashleys time, no one has yet local- concerned in the acquisition and retention of habits based on
ized the memory for a flux discrimination. Lesions of reactions to light. Comparative Psychology Monographs 11, 43
79.
visual subcortical structures (superior colliculus, later-
Lavond, D. G., and Dewberry, R. G. (1980). Visual form perception
al geniculate, pretectal area, accessory optic nuclei) or is a function of the visual cortex: II. The rotated horizontal-
of the limbic system (septum, hippocampus, amygda- vertical and oblique-stripes pattern problems. Physiological
la), in combination with visual decortication, do not Psychology 8, 18.
LeVere, T. E., and Morlock, G. W. (1973). Nature of visual recovery
prevent relearning. The fault probably lies within
following posterior neodecortication in the hooded rat. Jour-
Walter Hunters criticisms (1930) that there is not nal of Comparative and Physiological Psychology 83, 6267.
enough experimental control over the instrumental Meyer, D. R., and Meyer, P. M. (1977). Dynamics and bases of re-
training task used in this research. The more general coveries of functions after injuries to the cerebral cortex. Phys-
question is whether the neocortex is involved in any iological Psychology 5, 133165.
Spear, P. D., and Braun, J. J. (1969). Nonequivalence of normal
memory. Squire (1987) reviews the best evidence for and posteriorly neodecorticated rats on two brightness dis-
cortical memory, which can also be interpreted as crimination problems. Journal of Comparative and Physiological
suggesting that the cortex is necessary for perception Psychology 67, 235239.
of the stimuli but not for memory itself. Clearly, these Squire, L. R. (1987). Memory and the brain. New York: Oxford Uni-
versity Press.
are issues that are not resolved and continue to be of
interest. David Lavond
W
WATSON, JOHN B. (18781958) about and, despite taking a variety of philosophy
courses to fulfill a minor-area requirement, he later
John Broadus Watson (18781958), the founder of
confessed that only some of the British empiricists
behaviorism, was born January 9, 1878, near Green- (who emphasized past experience and principles of
ville, South Carolina. He spent his preadolescent association as the crucial sources of human knowl-
years in a farm community, where he acquired nu- edge) aroused his interest. Typically for the turn of
merous manual skills and an affectionate familiarity the century, psychology was part of the philosophy
with the behavior of many animals. At about the time department, and Watson soon gravitated toward
his father deserted the family, the Watsons moved James R. Angell as his major professor. Angell was ex-
into the cotton-mill town of Greenville, which his perimentally oriented and a leader of the burgeoning
mother thought would provide a better educational school of functionalism, which tolerated differing
and religious atmosphere for the children. Watson conceptions of the field of psychology but stressed the
later characterized himself as a mediocre student and role of evolutionary factors, environmental adapta-
a lazy, rebellious teenager (with a couple of arrests to tion, objectivity, and practical matters. This outlook
brag about). Nevertheless, he managed to persuade contrasted with that of experimental introspectionists
officials at Furman University in Greenville to admit (e.g., the structuralists), who used human observers
him. An average student at Furman from 1894 to reporting on their private conscious experience, with-
1899, Watson graduated with an A.M. degree; only out regard for biological or practical implications.
philosophy and psychology had interested him at all.
His mothers death in 1900 removed any remaining Watson felt uncomfortable when asked to intro-
pressure to pursue a career in theology; by then, in spect in the standard ways, and he did not produce
any case, he had become antagonistic to established consistent reports under those conditions; but he said
religion. Gordon Moore, his professor in philosophy he felt at home with animals. Working under Angell
and psychology, had attended and favorably de- and Henry Donaldson (who along with Jacques Loeb,
scribed the University of Chicago, so Watson wrote to an extremely mechanistic and materialistic biologist,
its president about his ambitions to attend a real uni- handled Watsons other minor area, neurology), he
versity and amount to something professionally. studied possible correlations between problem-
Persuasive once again, he started graduate work there solving skills and the degree of medullation (myelina-
in 1900. tion) in the brains of white rats at various ages. After
three years of intense dedication to university duties
Watson had expected to concentrate on philoso- and various odd jobs that he took to support him-
phy, with the eminent John Dewey as his mentor. selfoverwork that presumably caused the relatively
However, he never knew what Dewey was talking brief breakdown he suffered during his final yearin
665
666 WATSON, JOHN B.
boratory work is particularly noteworthy because,

somewhat ironically, B. F. Skinner later assessed it as
Watsons best research, and the ethologist Konrad
Lorenz falsely concluded that if J. B. Watson had
only once reared a young bird in isolation, he would
never have stressed conditioning as much as he did.
As early as 19031904 Watson confided to some
Chicago colleagues his growing belief that psychology
could become an objective and practical science only
if it rid itself of unverifiable, unreliable introspective
methods and focused instead on the study of observ-
able behaviorevents that could be recorded by an
outsiderrather than on inferred, private states of
consciousness or experience. Associates like Angell
argued that his suggestion might be appropriate for
animal research but would hardly be satisfactory for
human beings. Another 10 years passed before Wat-
son publicly proposed such ideas as the main bases
for the approach he called behaviorism.
In 1908 Watson became full professor of experi-
mental and comparative psychology at Johns Hop-
kins University in Baltimore. He continued his ani-
mal research, and soon assumed the leadership of the
Johns Hopkins psychology program and the editor-
ship of several important journals in experimental
John B. Watson (Library of Congress)
psychology. With the encouragement and stimulation
of Knight Dunlap and Karl Lashley, he began to con-
centrate on developing his behavioristic psychology,
1903 Watson received the first Ph.D. in psychology to first presented to a large audience in a landmark Psy-
be awarded by Chicago. His dissertation, Animal Edu- chological Review article in 1913. In a radical redefini-
cation, was published in the same year. tion of psychology, Watson claimed that his field, ani-
mal learning and behaviorwhich had generally
been relegated to a minor position in psychology or
Early Career had not been viewed as part of psychology at allwas
Watson remained at Chicago until 1908, first as the one truly objective, scientific area of psychology.
Angells assistant and then as an instructor. Even Furthermore, he maintained that the techniques used
though he taught his students about orthodox intro- in the animal laboratory could be profitably, objec-
spective methods with human observers, his own re- tively, and practically applied to human beings; the
search involved only animals. With Harvey Carr he goal of psychology was to predict and control behav-
carried out influential work on the sensory basis of ior, not to analyze consciousness into its elements or
maze learning in rats (neither vision nor audition nor to study vague functions or processes like percep-
smell was presumably crucial; rather, what was impor- tion, imagery, and volition. According to Watson, psy-
tant was feedback stimulation from the animals own chology had not yet emancipated itself from philoso-
movements: kinesthesis or the muscle sense); with phy and religion, which it must do to become a true
Robert Yerkes he began studies of color vision that sciencethe science of behavior, of stimulus (S) and
eventually involved several nonhuman species; and response (R: movements and secretions).
he failed to find good evidence for learning by imita- Historians of psychology have had no difficulty
tion in monkeys. In addition, Watson spent the first tracing possible antecedents for virtually all of Wat-
of several summers on an island near Florida, observ- sons specific ideas and arguments. Among others,
ing the natural, instinctive behavior of birds (noddy they have cited views of philosophers (empiricists-
terns and sooty terns), some of which he isolated at associationists, materialists, positivists, pragmatists),
birth. His bird studies were thoughtful and creative; biologists (evolutionary theorists, naturalists, objectiv-
besides homing behavior, he investigated what today ists, reflexologists), and early psychologists (nonmen-
we would call Imprinting, instinctive drift, territoriali- talistic students of animal and human sensation,
ty, and egg, mate, and nest recognition. This nonla- learning, memory, and intelligenceas well as func-
WATSON, JOHN B. 667
tionalists like Angell). However, the direct influence Immediately after the war, Watson worked with a
on Watson of most of these views is unclear. In any graduate student, Rosalie Rayner, on his most famous
event, his approach was original because of how it single study. It originated from his claim that emo-
combined a variety of emphases, dissatisfactions, and tional behavior in human infants was based on three
opinions in a unique, revolutionary way. He offered fundamental types of unlearned, well-defined stimu-
a straightforward, bold program that was easy to un- lus-response (S-R) patterns: fear, rage, and love.
derstand (and easy to attack). More complex emotional reactions, to specific objects
Generally favorable opinions about Watsons ap- and situations, arose through associative learning and
proach (as well as his established reputation as a re- transferand supposedly could not be attributed to
searcher, administrator, and editor) led to his elec- hereditary predispositions. Primarily by means of
tion as president of the American Psychological Pavlovian procedures adopted directly from animal
Association (APA) 2 years after the publication of his research, 11-month-old Albert B. was conditioned to
behaviorist manifesto. Many psychologists correctly fear a white rat by associating presentations of the rat
believed that institutional and societal support for in- with a very loud noise. Soon the mere sight of the rat
dependent departments of psychology and new re- caused Albert to whimper, cry, and move as far away
search facilities would be increased by redefining psy- as he could. This fear reaction transferred to other
chology along practical and objective lines like those furry objects, like a rabbit or a Santa Claus mask. Un-
offered by Watson. fortunately, Albert left the nursery too soon for Wat-
son to attempt to eliminate the childs newly acquired
habits. A few years later, Mary Cover Jones, whose re-
Human Learning Research search at Columbia University was unofficially super-
In his APA presidential address (1915) Watson vised by Watson, compared various methods for re-
described research with both animals and humans, moving childrens fears of animals. Some treatments
but for the first time in his career he stressed the lat- worked better than others. This research, along with
ter. The talk offered a specific positive alternative to Watsons and Joness comments about its practical im-
the techniques for studying human psychology that plications, marks the beginning of the fields of behav-
he had condemned in print two years before. Such an ior modification and behavior therapy.
extension of his approach would presumably help
convert to behaviorism those psychologists who be- Watson denied any significant initiating or medi-
lieved that animal studies could not be of great signif- ating role for the brain, and he would not consider
icance for human affairs. The new method was essen- possible cognitive processes intervening between the
tially the conditioned-reflex procedure of Ivan Pavlov external S and the subjects R. His approach was thus
and Vladimir Bekhterev, which Watson had only re- peripheralistic in its focus on movements and secre-
cently begun to examine and appreciate. (Previously tions, and not on changes in the central nervous sys-
he had stressed the associationist laws of frequency tem. He worried that serious consideration of the ex-
and recency; he frowned on Edward L. Thorndikes istence of such intervening, unobservable processes
law of effect because the notion of strengthening or would be subjective and unscientific; in any case it was
weakening S-R bonds by means of subsequent satis- unnecessary for behavioral prediction and control.
faction or discomfort seemed subjective to him, al- But Watson did include implicit or covert behavior
though it is the forerunner of Skinners law of operant and verbal reports within his behaviorism. For ex-
reinforcement.) From his own studies with human be- ample, he viewed thinking as basically silent speech,
ings Watson illustrated a variety of Pavlovian condi- talking to yourself, that was potentially measurable by
tioning phenomena that seemed relevant for every- means of sensitive recording instruments attached to
day human behavior. He boasted, We give no more appropriate muscles (of the lips, tongue, larynx)a
instruction to our human subjects than we give to our general idea, not really original with Watson, that
animal subjects. stimulated much research. Also, a persons regular,
overt utterances could be objectively recorded as a
Except for a minor study with rats, the rest of
form of behavior. Still, Watson was accused of making
Watsons academic career (suddenly aborted within 5
an alarming concession: of retaining introspection
years) involved work with humans, especially young
under another guise, the verbal report.
infants in the Phipps Psychiatric Clinic directed by
Adolf Meyer. There was one brief interruption, when In 1920, while engrossed in his work with infants
Watson served in the army during World War I and other experiments involving adult human learn-
(19171918) as a psychologist concerned mainly with ing, Watson was faced with divorce proceedings initi-
aviation skills. Despite his irritation with the military ated by his wife, who had discovered his love affair
establishment, Watsons views on the technological with Rayner. The participants were so well known (the
potential of psychology were bolstered. Rayner family was politically and socially prominent
668 WATSON, JOHN B.
in Maryland) that the case became a local and nation- was becoming more urbanized and seemed to recog-
al sensation. Although Watson had probably believed nize the need for an effective technology of behavior
that he was too important a figure at Johns Hopkins (for example, in education and retraining). Interest-
and in American psychology to lose his job over such ingly, behaviorism never gained strong support in
a personal matter, he was forced to resign from the Europe, perhaps because traditional values there
university in 1920. He never again held any official were more intellectual, philosophical, and abstract;
academic position. He and Rayner were married as democratic, practical ideals were not so prevalent.
soon as the divorce was final. Watsons popular book Psychological Care of Infant
and Child (1928), dedicated to the first mother who
brings up a happy child, had a definite influence on
From Science to Advertising American child-rearing practices in the 1930s. Some
Resilient and self-reliant, Watson began an en- writers have described Watson as the Dr. Spock of his
tirely new career at the J. Walter Thompson Agency, day, but unlike Spock he maintained that the up-
viewed by its president, Stanley Resor, as a university bringing of children should be quite objective and
of advertising. Watson started at the bottom, survey- routinized, with minimal affection and sentimentali-
ing the demand for different kinds of rubber boots ty. His own children said that he was all business,
along the Mississippi River and acting as a salesman believing that tenderness would have a harmful effect
in Macys department store to observe consumer reac- on their independence and emotional control. In
tions. He eventually became a vice president and was Watsons autobiographical sketch (1936) he apolo-
directly involved in many campaigns for specific gized for the infant-care book, admitting that he had
products. He favored emotional over rational appeals insufficient knowledge to write it. He did not, howev-
but contributed no strikingly novel methods to the er, retract any of its specific advice.
field of advertising, as some writers have claimed. Fi-
Different varieties of behaviorism had emerged
nancially successful compared with his academic
almost as soon as Watson proposed his own brand,
years, he asserted, It can be just as thrilling to watch
but in the 1930s to 1960s more sophisticated neobe-
the growth of a sales curve of a new product as to
haviorists (e.g., Edwin Guthrie, Clark Hull, B. F.
watch the learning curve of animals or men.
Skinner, and Edward Tolman) flourished during the
After his dismissal from Johns Hopkins, Watson so-called golden age of learning theory. These per-
continued to write and lecture about behaviorism, but sons and their current impact are discussed elsewhere
the books, radio broadcasts, and magazine articles in this volume, along with views of contemporary cog-
were directed mainly at a popular audience. Aside nitive psychologists, who generally reject many of be-
from Freud, he was probably the psychologist best haviorisms assumptions and emphasesbut not its
known to the American public in the first half of the objective methodology.
twentieth century. Unfortunately, his views became Rosalie Watsons death in 1936 left her husband
progressively more simplistic, dogmatic, brash, and depressed for a long time. Although he worked at an
extreme. Still, his book Behaviorism (1924), though advertising firm for another decade, he preferred the
hastily written, was favorably received; a New York isolation of his rural Connecticut home and farm,
Times reviewer said it marked a new epoch in the in- part of which he had built himself, to social and intel-
tellectual history of man, and the New York Herald- lectual activities. The APA presented Watson with a
Tribune declared that perhaps this is the most impor- special award in 1957, the year before his death on
tant book ever written. Even Bertrand Russell said September 25, 1958, and almost 40 years after he left
it was massively impressive. academia. He was honored as the initiator of a revo-
In this and later writings Watson repudiated his lution in psychological thought and a person whose
earlier acceptance of the existence of certain human work was a vital determinant of the form and sub-
instincts and instead presented an extremely environ- stance of modern psychology.
mentalist, learning-based point of view. A widely cited See also: BEHAVIORISM; CONDITIONING, CLASSICAL
passage, usually quoted without some qualifications AND INSTRUMENTAL; GUTHRIE, EDWIN R.;
that he did add, claimed that with the right kind of HULL, CLARK L.; LEARNING THEORY: A
early experience and training, one could make any HISTORY; LEARNING THEORY: CURRENT
healthy infant into a doctor, lawyer, artist . . . even STATUS; PAVLOV, IVAN; SKINNER, B. F.;
beggar-man and thief, regardless of the talents . . . THORNDIKE, EDWARD; TOLMAN, EDWARD C.
abilities, vocations, and race of his ancestors. Such a Bibliography
democratic view, combined with Watsons optimistic Boakes, R. A. (1984). From Darwin to behaviourism: Psychology and the
vision of psychologys general role in transforming minds of animals. Cambridge, UK: Cambridge University
society, was attractive to the American public, which Press.
WORKING MEMORY: Animals 669
Buckley, K. W. (1989). Mechanical man: John Broadus Watson and the than 30,000 pine seeds buried in some 3,000 cache
beginnings of behaviorism. New York: Guilford Press. sites in the forest (Vander Wall, 1982). It must re-
Cohen, D. (1979). J. B. Watson: The founder of behaviourism. London:
Routledge and Kegan Paul.
member site locations after snow covers the forest and
Harrell, W., and Harrison, R. (1938). The rise and fall of behavior- even burrow at an angle through the snow to arrive
ism. Journal of General Psychology 18, 367421. at the location. They use new cache sites each year
ODonnell, J. M. (1985). The origins of behaviorism: American psycholo- and must overcome potential confusion (i.e., interfer-
gy, 18701920. New York: New York University Press. ence) over current and previous sites.
logical Review 20, 158177. The recall of cache sites for six months certainly
(1914). Behavior: An introduction to comparative psychology. qualifies as long-term memory (LTM). But the bor-
New York: Henry Holt.
(1919). Psychology from the standpoint of a behaviorist. Philadel-
derline between long- and short-term memory (STM)
phia: Lippincott. is far from clear; in fact, the distinction between STM
(1924). Behaviorism. New York: W. W. Norton. and LTM may have outlived its usefulness for behav-
(1928). Psychological care of infant and child. New York: W. W. ioral analyses of memory (Crowder, 1993). Perhaps a
Norton. more useful distinction is that between working and
(1936). John Broadus Watson (autobiographical sketch). In
C. Murchison, ed., A history of psychology in autobiography, Vol.
reference memory (Honig, 1978). These terms are
3, pp. 271281. Worcester, MA: Clark University Press. free of the theoretical baggage carried by the terms
STM/LTM or the similar primary/secondary memory.
Eliot Hearst
In the nutcracker example, reference memory would
be the skill-learning involved in storing and retriev-
ing seeds (Balda and Kamil, 1992). Working memory
would apply to this years cache sites.
WORKING MEMORY Remarkable though the capacity and duration of
nutcracker may be, the underlying processes and
[Working memory refers to the ability to hold a small amount mechanisms are what compel the attention of re-
of information in mind and work with it, often in the face searchers who seek to understand how memory
of distraction. One aspect of working memory is the straight- works. Such insight requires laboratory investigations
forward ability to retain information over short intervals that manipulate independent variables critical to
or short-term memorya topic studied both in nonhuman memory. Such work is proceeding briskly along two
animals and in humans. The paradigms used are necessarily avenues of research: neurobiological and behavioral.
different in the two cases, but the intent is to study how infor-
mation is retained over short time periods. Working memory Investigations of neural mechanisms of memory
in humans is believed to involve verbal and nonverbal pro- use simple, quickalmost instantaneouslearning
cesses, perhaps with two types of capacity for working memo- procedures. A repeated application of a stimulus
ry. Working memory capacity in humans seems to be related (e.g., tactile) to an organism (e.g., sea slug) can pro-
to general intelligence; some propose that the capacity to hold duce habituationa diminished response (e.g., si-
information in mind in spite of interference may underlie the phon withdrawal)or sensitizationa heightened
concept of general intelligence. response. This is a primitive form of learning: modifi-
cation of a response through experience. Since all
The two sections here review work on the topic of work-
learning depends on memory, tests of learning are
ing memory both in ANIMALS and in HUMANS. Various spe-
also tests of memory. Neuroplasticity studies have
cies of animal show remarkable mnemonic abilities that are
shown changes in cellular mechanisms, membrane
quite specialized. However, when experiments are conducted
channels, second-messengers, and protein syntheses
that compare as directly as possible animals with humans,
related to memory (Squire and Kandel, 2001).
often many features of performance (such as serial position
curves) are similar. The first section covers such research Another nearly instantaneous learning procedure
with animal subjects; the second provides information con- is fear conditioning, a pairing of two stimuli (classical
ducted in both a behavioral and a neuroimaging tradition or Pavlovian conditioning). Rats learn to freeze (re-
with human subjects.] main immobile) when encountering a stimulus paired
with electric shock. Fear conditioning has shown that
dorsal hippocampal lesions disrupt recent memories
ANIMALS of the conditioning context (e.g., test chamber) but
not those of the conditioned stimuli (e.g., tones)
Animals have good memories. There are anecdotal
themselves (Anagnostaras et al., 2001).
reports of dogs greeting their masters after years of
absence and of bears and elephants squaring old Other neurobiological memory procedures re-
scores after many years. Birds, too, can have remark- quire instrumental responses to escape unpleasant
able recall. The Clarks nutcracker recovers more situations or to obtain rewards. Rats readily learn the
670 WORKING MEMORY: Animals
location of a hidden platform in order to rest from Two experiments with rhesus monkeys show simi-
swimming in a Morris water maze. Water maze studies larities to human memory processing and expand
have shown that LTP, NMDA receptors, and the dor- upon these findings. In a visual-memory experiment,
sal (but not ventral) hippocampus are related to monkeys saw lists of four different travel-slide pic-
memory (Morris et al., 1986). Rats also readily learn tures. After a delay (0, 1, 2, 10, 20, or 30 seconds later)
the locations of food in radial-arm mazes. They are following each list, the monkeys indicated whether a
very good at remembering, for example, which arms test picture presented in a different location was or
have not been visited on a particular trial and thus was not in the list (Wright et al., 1985). Figure 1
still contain rewards. Four-arm plus mazes have (upper portion) shows the changes in the monkeys
played a key role in revealing the mnemonic compo- four-item serial-position functions for three of the re-
nents of the place cells of rats (OKeefe and Speak- tention delays. Like humans, the monkeys recency ef-
man, 1987). Strategies used to solve water and radial- fect dissipated with retention delay. A new finding was
arm maze tasks occasionally make interpretation of the gradual appearance of the primacy effect with
results complex, and such strategies (i.e., reference delay, and this finding was made possible by using
memory) can confound measures of working memo- much shorter lists than those used with humans. The
ry. Other memory results raise questions about the primacy and recency effects were separately manipu-
universality of the hippocampus in rats place memo- lated and dissociated by retention interval, and the
ry. In delayed nonmatching to sample (DNMS), mon- benchmark U-shaped function emerged as a transi-
keys choose a stimulus that does not match a previ- tional function. Similar changes have been shown for
ously seen sample. DNMS in monkeys depends less humans (e.g., using four- or five-item lists of snow-
on the hippocampus than place memory in rats. flake or kaleidoscope patterns), capuchin monkeys,
Some researchers are trying to determine whether and pigeons (Korsnes, 1995; Neath, 1993; Wright,
the hippocampus will prove necessary for general 1999a; Wright et al., 1985). One issue was whether
relational memory (Cohen and Eichenbaum, 1994) SPFs for other types of memory (e.g., auditory) would
or just place memory. A newer, more sensitive proce- change with delay like those for visual memory.
dure for identifying the role of the hippocampus in
nonplace memory is a habituation procedure called In an auditory memory experiment, monkeys
preferential viewing (Alvarado and Bachevalier, heard lists of four natural/environmental sounds.
2000). Monkeys (or children) view and habituate to a After a delay (0, 1, 2, 10, 20, or 30 seconds later) fol-
picture, then are tested with that picture presented lowing each list, they had to touch one of two side
with a novel picture. The subject viewing mainly the speakers to indicate whether or not a test sound was
novel picture shows good memory of habituation. in the list (Wright, 1998). As in the visual memory ex-
periment, researchers drew from a large pool of
Investigations of behavioral mechanisms of ani- sounds (520 in this case), each one unique to each
mal memory have traditionally used single-memory days trials. The most striking feature of the auditory
items on each trial. Single-memory-item tests may be memory SPFs (see Figure 1) is that their shape is op-
well suited to neurobiological investigations (e.g., posite to that of the visual memory results. Initially,
treatment versus normal groups) but by themselves there was no recency effect. The recency effect ap-
do not reveal much about memory mechanisms (Shet- peared to lengthen with delay, and the primacy effect
tleworth, 1998, p. 257; Olson et al., 1995; Kamil, dissipated. Thus, evidence about how one memory
1988). Explorations of human memory mechanisms system works may not reveal much about the workings
have used list-memory procedures for a long time of other memory systems: vision, auditory, taste,
(Ebbinghaus, 1902). Memory is typically best for
smell, and touch.
items at the beginning (primacy effect) and at the end
(recency effect) of each list, producing a U-shaped se- Another finding is the counterintuitive one that
rial-position function (SPF). Delaying recall (or recog- memoryspecifically, visual primacy and auditory
nition) eliminates the recency effect. This selective recentycan improve with time. Thus, mechanisms
elimination supported the hypothesis that the recen- other than memory decay must be responsible for the
cy effect represented STM (Glanzer and Cunitz, SPFs.To explore possible mechanisms, researchers
1966), and helped stimulate the cognitive revolution manipulated interference among auditory list items.
of the 1960s. The degree to which animals share simi- The first list items were shown to interfere (proactive-
lar STM processes with humans became testable only ly, i.e., forward in time) with the monkeys memory of
in the 1990s because of the extreme difficulty of train- the last list items (Wright, 1999b). Diminishing or
ing animals in list-memory tasks. Nevertheless, the eliminating this interference greatly enhanced mem-
list memory of rats, pigeons, dolphins, squirrel mon- ory for the last items, a result that ruled out lack of
keys, capuchin monkeys, rhesus monkeys, and chim- consolidation as an explanation of poor memory of
panzees has been tested. the last list items. With long delays, the last list items
WORKING MEMORY: Animals 671
Figure 1
Average serial position functions for different groups of monkeys tested in a 4-item visual memory task (top) or a 4-item auditory
memory task (bottom). The parameter delay is the retention interval between list and test. Serial position 1 is the first list item.
Unfilled points (Diff) are from trials in which the test matched no list item.
interfered (retroactively, i.e., backward in time) with See also: APLYSIA: MOLECULAR BASIS OF LONG-TERM
the monkeys memory of the first list items. Diminish- SENSITIZATION; NEURAL SUBSTRATES OF
ing or eliminating this interference greatly improved EMOTIONAL MEMORY
memory of the first item, a result that ruled out for-
getting (i.e., memory decay) as an explanation for Bibliography
poor memory of the first list items. Alvarado, M. C., and Bachevalier, J. (2000). Revisiting the matura-
tion of medial temporal lobe memory functions in primates.
Researchers face the challenge of closing the gaps Learning and Memory 7, 244256.
between neurobiological and behavioral approaches Anagnostaras, S. G., Gale, G. D., and Fanselow, M. S. (2001). Hip-
to animal and human memory. This interdisciplinary pocampus and contextual fear conditioning: Recent contro-
versies and advances. Hippocampus 11, 817.
approach will require a blend of molecular, cellular, Balda, R. P., and Kamil, A. C. (1992). Long-term spatial memory
neurochemical, anatomical, imaging, behavioral, and in Clarks nutcracker, Nucifraga columbiana. Animal Behaviour
computational techniques. 44,761769.
672 WORKING MEMORY: Humans
Cohen, N. J., and Eichenbaum, H. (1994). Memory, amnesia, and the initial calculations while performing other opera-
hippocampal system. Cambridge, MA: MIT Press. tions, storing interim results for which you would
Crowder, R. G. (1993). Short-term memory: Where do we stand?
Memory & Cognition 21, 142145.
have no need after arriving at the answer and the de-
Ebbinghaus, H. E. (1902). Grundzuge der Psychologie. Leipzig: Von tails of which would likely escape you upon subse-
Veit. quent inquiry. The temporary storage and manipula-
Glanzer, M., and Cunitz, A. R. (1966). Two storage mechanisms in tion of information required to perform this and
free recall. Journal of Verbal Learning and Verbal Behavior 5, many similar tasks is working memory.
351360.
Honig, W. K. (1978). Studies of working memory in the pigeon. In
S. H. Hulse, H. Fowler, and W. K. Honig, eds., Cognitive pro-
cesses in animal behavior. Hillsdale, NJ: Erlbaum. Short-Term versus Working Memory
Kamil, A. C. (1988). A synthetic approach to the study of animal Researchers have long accepted this idea of a
intelligence. In D. W. Leger, ed., Nebraska Symposium on Moti-
form of separate, temporary memory storageone
vation, Vol. 35. Lincoln: University of Nebraska Press.
Korsnes, M. S. (1995). Retention intervals and serial list memory. quite distinct from the system of long-term storage
Perceptual and Motor Skills 80, 723731. for many years. In the nineteenth century William
Morris, R. G. M., Anderson, E., Lynch, G. S., and Baudry, M. James (1890) proposed the term primary memory for a
(1986). Selective impairment of learning and blockade of form of temporary storage that was at least partly re-
long-term potentiation by an N-methyl-D-aspartate receptor
sponsible for the experience of what he termed the
antagonist, AP5. Nature 319, 774776.
Neath, I. (1993). Distinctiveness and serial position effects in rec- specious present, the experience that time is contin-
ognition. Memory & Cognition 21, 689698. uous, extending beyond the consciousness of the cur-
OKeefe, J., and Speakman, A. (1987). Single unit activity in the rat rently encountered microsecond.
hippocampus during a spatial memory task. Experimental
Brain Research 68, 127. Despite the theorizing of James, the twentieth
Olson, D. J., Kamil, A. C., Balda, R. P., and Mims, P. J. (1995). Per- century saw little interest in the mental capacity for
formance of four-seed caching corvid species in operant tests short-term storage until the 1950s, which saw the rise
of nonspatial and spatial memory. Journal of Comparative Psy- of a cognitive revolution, a surge of interest in the
chology 109,173181.
Shettleworth, S. J. (1998). Cognition, evolution, and behavior. New
information-processing approach to the analysis of
York: Oxford University Press. human cognition that used the newly developed digi-
Squire, L. R., and Kandel, E. R. (1998). Memory from mind to mole- tal computer as a basis for new theories of the work-
cules. New York: W. H. Freeman and Co. ings of the mind. This sea change, coupled with some
Vander Wall, S. B. (1982). An experimental analysis of cache recov- of the practical problems that had arisen from the at-
ery in Clarks nutcracker. Animal Behaviour 30, 8494.
Wright, A. A. (1998). Auditory list memory in rhesus monkeys. Psy-
tempt to apply psychology to wartime issues, led to re-
chological Science 9, 9198. newed interest in both attention and short-term stor-
(1999a). Visual list memory in capuchin monkeys (Cebus age, thanks in good measure to the work of
apella). Journal of Comparative Psychology 113, 7480. Broadbent (1958).
(1999b). Auditory list memory and interference in mon-
keys. Journal of Experimental Psychology: Animal Behavior Process- By the late 1960s a general consensus emerged in
es 25, 284296. favor of the separation of memory into at least two
Wright, A. A., Santiago, H. C., Sands, S. F., Kendrick, D. F., and systems: a short-term, limited-capacity store that can
Cook, R. G. (1985). Memory processing of serial lists by pi-
hold information for a matter of seconds and that
geons, monkeys, and people. Science 229, 287289.
feeds a far more capacious and durable long-term
Anthony A. Wright store. The dominant model of this period was that
proposed by Atkinson and Shiffrin, in which informa-
tion passes through a series of brief sensory registers
that are part of the processes of perception and then
HUMANS moves into a limited-capacity short-term store. The
Working memory is a system that provides short-term latter acts as a working memory, with a series of strate-
storage of information used in complex cognitive ac- gies or controls that organize and maintain incoming
tivities such as planning, reasoning, problem-solving, material to optimize learning and subsequent recall.
and language. This mental blackboard is a temporary This model gives a good account of the available data
workspace that makes possible the examination, ma- and is still portrayed as the dominant view of memory
nipulation, and tranformation of internally repre- in many introductory textbooks.
sented information during these cognitive activities Nevertheless, despite its attractive simplicity, the
and also allows its subsequent erasure. model soon began to encounter problems. One diffi-
Suppose you ordered three bottles of mineral culty was its learning assumption: the longer an item
water at $1.80 a bottle and gave the waiter ten dollars. is held in short-term storage (STS), the greater is its
How much change would you expect? Working this probability of transfer into long-term memory. Evi-
out would likely involve retaining the results of your dence conflicting with this assumption came from
WORKING MEMORY: Humans 673
data that initially seemed to support the two-store Figure 1

model. Patients suffering from amnesia following
brain damage appeared to show drastic impairment
of long-term learning but retained their short-term
store (Baddeley and Warrington, 1970). Conversely,
Shallice and Warrington (1970) identified other pa-
tients showing exactly the opposite pattern: normal
long-term learning but problems on short-term mem-
ory tasks. The double dissociation found in such pa-
The structure of working memory proposed by Baddeley and
tients conformed to the Atkinson and Shiffrin model
Hitch. The boxes labeled phonological loop, central executive,
quite nicely, except for one crucial point: If, as the
and visuospatial scratchpad represent the proposed components
model proposed, short-term storage was a crucial or subsystems of working memory.
working memory that made learning and long-term
storage possible, then how could patients with a gross
disruption of the short-term store learn and recall The phonological-loop component seems to be
normally? A defective short-term store should have made up of two subcomponents, a store that will hold
led to massive problems in learning, memory, and auditory-verbal memory traces for about two seconds
general cognition if this system indeed served as a and a subvocal rehearsal process. This process can
working memory. both maintain the items within the store by recycling
them subvocally and register visually presented items
in the store by means of subvocalization. The short-
The Baddeley and Hitch Model term phonological storage system seems specially
Baddeley and Hitch (1974) tried to clarify the adapted to language learning since it provides critical
problems in the Atkinson and Shiffrin model by at- mechanisms for keeping novel phonological strings
tempting to simulate in normal subjects the deficits active and in the correct serial order that allows their
suffered by patients with short-term memory prob- binding to appropriate meanings. Evidence support-
lems. They attempted to disrupt short-term memory ing this hypothesis comes from studies of vocabulary
function by requiring the subject to perform a memo- acquisition in children, in whom the functioning of
ry-span task while attempting to learn, reason, or the phonological store predicts individual differences
comprehend. The two-store model should predict in language-learning skills (Baddeley, Gathercole,
that requiring the subject to maintain a string of, say, and Papagno, 1998).
six digits would nearly fill the limited-capacity short- The visuospatial sketchpad may represent a sepa-
term store. If the latter acts as a crucial working mem- rate system specializes in the maintenance of infor-
ory, then the subjects capacity for other tasks should mation using visuospatial rather than linguistic codes
be almost totally disrupted. The pattern of results ob- (i.e., mental images). There is evidence that it may
tained across a range of activities was quite consistent. constitute two subsystems, one concerned with spatial
While there was indeed some impairment, even a con- information and the relative location of objects and
current load of six numbers produced only modest the with pattern information. Patients with a disrup-
disruption, even in the capacity to perform quite de- tion to the operation of the spatial system may have
manding reasoning tasks. Baddeley and Hitch (1974) difficulty in finding their way around but have no
concluded that these results were inconsistent with problem in recalling or using information concerning
the Atkinson and Shiffrin model (1968), and pro- the visual characteristics of objects, such as the color
posed an alternative multicomponent model to re- of a banana or the shape of a dachshunds ears. Other
place the unitary short-term store. patients show the opposite pattern of deficits (Farah,
The Baddeley and Hitch working memory model 1988). Some research suggests that rehearsal of infor-
consisted of three components: an attentional control mation in the sketch pad might be similar to shifting
system they termed the central executive and two sub- spatial attention, and as such appears to be linked to
sidiary slave systems, the phonological loop which the brain system that controls eye movements (Awh
maintains and manipulates speech-based informa- and Jonides, 2001).
tion, and the visuospatial sketchpad which provides The central executive, the most complex and
a temporary storage system for visuospatial informa- least understood component of working memory, at
tion (see Figure 1). This model still remains as the first seemed to operate along the lines of the model
most widely accepted account of the function and or- of attentional control proposed by Norman and Shal-
ganization of working memory, and as such merits lice (1986). This model assumes that continual activity
further description. is controlled in two major ways: by the running off of
existing programs or scripts, or by the intervention of term memorythe strengthening of (synaptic) con-
the supervisory attentional system (SAS). The latter is nections between neurons.
capable of interrupting continual semiautomatic pro-
A primary focus of both the animal and human
grams when they reach an impasse or when longer-
brain imaging studies of working memory has been
term goals demand a departure from the continual
the prefrontal cortex. For example, when humans
activity. A secondary function proposed for the cen-
perform working memory tasks, there is a reliable
tral executive was to coordinate and integrate the op-
sustained increase in activity within the prefrontal
eration of the two buffer systems in tasks requiring ei-
cortex during maintenance periods, with the magni-
ther simultaneous or alternating storage of both
tude of increase related to working memory load.
verbal and visuospatial information.
This finding has been demonstrated using a number
The influential Baddeley and Hitch model of simple task paradigms, including the N-back
spawned a great deal of research in the 1970s and (Cohen et al., 1997) and Sternberg item-recognition
1980s, much of it geared toward the study of the task (Jha and McCarthy, 2000). Both tasks are de-
properties of the two slave storage systems, especially layed-response paradigms in which the information
that of the phonological loop. A second theme of re- to be stored in working memory must be matched
search on working memory during this period was an against a subsequent probe item, changes from trial
examination of its role in individual differences in to trial, and can vary in the number of items that must
higher-level cognitive abilities. In 1980, Daneman be held and in the delay period of maintenance. In
and Carpenter suggested that verbal working memo- the N-back a continuous sequence of items is pres-
ry capacity could predict abilities in a range of general ented one at a time, each requiring a determination
language skills, such as reading comprehension and whether it matches an item presented a specified
verbal SAT scores (Daneman and Carpenter, 1980). number (N) of trials back (e.g., in a three-back condi-
Verbal working memory capacity was measured using tion the last A in the sequence ABCA would be consid-
a task called the reading span, which determines how ered a match). Thus, the task requires that a previous
many words the subject can retain in short-term item must be maintained in its appropriate sequence
memory while simultaneously performing interven- and stored over intervening items. The Sternberg
ing language-processing tasks (reading sentences out task is similar but typically presents the memory set
loud). Subsequent studies have validated and replicat- simultaneously followed by an unfilled delay interval.
ed the original findings using a variety of similar span Figure 2 shows data on the time course of prefrontal
tasks and cognitive predictors. Many theorists now cortex activity as a function of working-memory load
believe that working-memory capacity might serve as in both the N-back and Sternberg tasks.
a strong analogue to the notion of a domain-general
intelligence factor (Engle, Tuholski, Laughlin, and Other brain-imaging research on working memo-
Conway, 1999). Other studies have supported the no- ry has provided converging support for the Baddeley
tion of the Baddeley and Hitch model that there are and Hitch model by suggesting that there is a
two separate working-memory capacities, with one neuroanatomical segregation of verbal and visuospa-
predicting verbal abilities and the other predicting vi- tial short-term storage, of storage and rehearsal in
suospatial ones (Shah and Miyake, 1996). verbal working memory, and of maintenance and ma-
nipulation (e.g., central executive) functions (Smith
and Jonides, 1999). These findings remain controver-
Mechanisms of Working Memory sial because not all studies have observed such distinc-
tions. One possibility is that the components of the
Since 1990, the focus of research on working
model do not cleanly map on to the brain, especially
memory has shifted to a more thorough study of its
as regards the relationship between the maintenance
critical psychological and biological mechanisms:
and control functions of working memory.
How do the components of working memory work
and how does the brain implement them? This trend As a consequence, attention has turned toward
has been driven by the rise of cognitive neuroscience gaining a better understanding the nature of the ex-
as a discipline and also by the development of new executive-control functions needed for successful work-
perimental methods such as human brain imaging for ing memory during complex cognition. A first critical
studying working memory. For example, nonhuman question is whether separable control functions play
primate research has suggested that active mainte- different roles within working memory. A number of
nance of information through the sustained firing of possible control operations have been suggested,
neuronal populations provides a cellular mechanism such as the following: shifting or switching attention
of working memory (Goldman-Rakic, 1995). This between the mental sets needed to perform different
short-term storage mechanism contrasts sharply with tasks; inhibition or suppression of inappropriate re-
the cellular mechanisms that seem to underlie long- sponse tendencies or irrelevant information; moni-
WORKING MEMORY: Humans 675
Figure 2
(Left) Activation of prefrontal cortex (dorsolateral region) during performance of the N-back task. The upper panel refers to the
experimental design of the study, in which four brain imaging scans were acquired during the course of each trial. Because of the long
delay interval (10 seconds) between trials, activity reflecting active maintenance should be present during the later scans of each trial
(scans 3 and 4). The lower panel shows the average activity level in this prefrontal cortex region across four different N-back
conditions (0-back through 3-back) during the course of a trial. The greater activity for 2-back and 3-back conditions relative to 0-back
and 1-back conditions reflects sensitivity to working memory load. The constant level of activity across the trial suggests a sustained
response reflecting active maintenance. Adapted from Cohen, J. D., Perstein, W. M., Braver, T. S., Nystrom, L. E., Noll, D. C., Jonides,
J., and Smith, E. E. (1997). Temporal dynamics of brain activation during a working memory task. Nature 386, 604608. (Right)
Activation of prefrontal cortex (same dorsolateral region) during performance of Sternberg task using faces as stimulus items.
Activation is plotted from the time of the presentation of the memory set (S1) to the time of the subsequent probe item (S2), which
occurred 27 seconds later. The increased activation for the 3-face condition reflects a sensitivity to working memory load. The
increased activity (relative to the pretrial baseline) throughout the delay interval suggests that the response reflects active maintenance.
Adapted from Jha, A. P., and McCarthy, G. (2000). The influence of memory load on delay-interval in a working-memory task: An
event-related functional MRI study. Journal of Cognitive Neuroscience 12, 90105.
toring and updating the contents of working memo- such as that of Baddeley and Hitch suggest a strict
ry; temporal tagging or contextual coding of segregation between maintenance and control func-
incoming information; and planning or sequencing tions. Yet other models suggest that these functions
intended actions (Smith and Jonides, 1999). Some are more tightly intertwined, with active maintenance
evidence suggests the statistical independence of of goal representations serving as a primary means of
these functions (Miyake et al., 2000); but it is control (OReilly, Braver, and Cohen, 1999). One
unclear whether these functions are instantiated view argues for a distinction between two types of
by truly separable systems or might be further maintenance mechanisms: goal-related information
reducible to a more fundamental, unitary set of maintained in the focus of attention that is robustly
mechanisms. protected from interference and transiently activated
A second central issue is the exact relationship (via spreading associative processes) information
between the mechanisms supporting active mainte- from long-term memory that decays over brief inter-
nance and those that control this process. Models vals (Cowan, 1995). The first mechanism seems more
domain-general and may reflect the functions of the Baddeley, A. D., and Hitch, G. J. (1974). Working memory. In G.
prefrontal cortex; the second mechanism likely re- Bower, ed., The psychology of learning and motivation, Vol. 8.
New York: Academic Press.
flects more domain-specific maintenance operations
Baddeley, A. D., and Warrington, E. K. (1970). Amnesia and the
that are widely distributed across different brain re- distinction between long- and short-term memory. Journal of
gions. However, the two sources of working memory Verbal Learning and Verbal Behaviour 9, 176189.
seem to be interdependent, with actively maintained Broadbent, D. E. (1958). Perception and communication. Oxford: Per-
goal information biasing the time course of activated gamon.
long-term memory (through refreshment or suppres- Cohen, J. D., Perstein, W. M., Braver, T. S., Nystrom, L. E., Noll,
D. C., Jonides, J., and Smith, E. E. (1997). Temporal dynam-
sion). Conversely, spreading activation of long-term ics of brain activation during a working memory task. Nature
memory elements might drive the activation or up- 386, 604608.
dating of goal representations. Cowan, N. (1995). Attention and memory. Oxford: Oxford University
Press.
Many studies have suggested that the individual
Daneman, M., and Carpenter, P. A. (1980). Individual differences
differences in working-memory capacity most strong- in working memory and reading. Journal of Verbal Learning
ly related to skill in complex cognitive tasks (and con- and Verbal Behavior 19, 450466.
structs of fluid intelligence) are those pertaining to Engle, R. W., Tuholski, S. W., Laughlin, J. E., and Conway, A. R.
goal-maintenance mechanisms rather than to the ac- A. (1999). Working memory, short-term memory, and general
tivation of long-term memory elements (Kane, Bleck- fluid intelligence: A latent-variable approach. Journal of Exper-
imental Psychology: General 128, 309331.
ley, Conway, and Engle, 2001). Thus, the notion of Farah, M. J. (1988). Is visual imagery really visual? Overlooked evi-
working-memory capacity may be a misnomer, refer- dence from neuropsychology. Psychological Review 95, 307
ring not to the number of items that can be main- 317.
tained simultaneously but rather to the efficacy with Goldman-Rakic, P. S. (1995). Cellular basis of working memory.
which an individual can represent and maintain even Neuron 14, 477485.
James, W. (1890). Principles of psychology. New York: Holt, Rinehart,
a single goal within the focus of attention in the face
and Winston.
of interference and distraction. Jha, A. P., and McCarthy, G. (2000). The influence of memory load
There is much still to be understood about the na- on delay-interval in a working-memory task: An event-related
functional MRI study. Journal of Cognitive Neuroscience 12, 90
ture and mechanisms of working memory. Progress
105.
in understanding working memory will likely come Kane, M. J., Bleckley, M. K., Conway, A. R. A., and Engle, R. W.
from the further specification of mechanisms that (2001). A controlled-attention view of WM capacity. Journal of
arise both from direct comparisons between models Experimental Psychology: General 130, 169183.
and from their implementation in explicit computa- Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Ho-
werter, A., and Wager, T. D. (2000). The unity and diversity
tional formalisms (Miyake and Shah, 1999). Never-
of executive functions and their contributions to complex
theless, the general concept of a working memory sys- frontal lobe tasks: A latent variable analysis. Cognitive Psy-
tem that provides temporary storage for a wide range chology 41, 49100.
of cognitive activities has proved to be a useful one, Miyake, A., and Shah, P., eds. (1999). Models of working memory:
and is likely to remain so. While theoretical models Mechanisms of active maintenance and executive control. New
of working memory are likely to change and become York: Cambridge University Press.
Norman, D. A., and Shallice, T. (1986). Attention to action: Willed
more complex, the study of working memory is likely and automatic control of behavior. In R. J. Davidson, G. E.
to significantly inform such diverse cognitive domains Schwartz, and D. Shapiro, eds., Consciousness and self-
as attention, reasoning, language, individual differ- regulation, Vol. 4. New York: Plenum Press.
ence, and the executive control of behavior. OReilly, R. C., Braver, T. S., and Cohen, J. D. (1999). A biological-
ly based computational model of working memory. In A. Mi-
See also: JAMES, WILLIAM; LONG-TERM yake and P. Shah, eds., Models of working memory: Mechanisms
POTENTIATION; PREFRONTAL CORTEX AND of active maintenance and executive control. New York: Cam-
MEMORY IN PRIMATES; SENSORY MEMORY bridge University Press.
Shah, P., and Miyake, A. (1996). The separability of working mem-
Bibliography ory resources for spatial thinking and languge processing: An
Atkinson, R. C., and Shiffrin, R. M. (1968). Human memory: A individual differences approach. Journal of Experimental Psy-
proposed system and its control processes. In K. W. Spence, chology 125, 427.
ed., The psychology of learning and motivation: Advances in re- Shallice, T., and Warrington, E. K. (1970). Independent function-
search and theory. New York: Academic Press. ing of verbal memory stores: A neuropsychological study.
Awh, E., and Jonides, J. (2001). Overlapping mechanisms of atten- Quarterly Journal of Experimental Psychology 22, 261273.
tion and spatial working memory. Trends in Cognitive Sciences Smith, E. E., and Jonides, J. (1999). Storage and executive process-
5, 119126. es in the frontal lobes. Science 283, 1,6571,661.
Baddeley, A., Gathercole, S., and Papagno, C. (1998). The phono-
logical loop as a language learning device. Psychological Review Alan D. Baddeley
105, 158173. Revised by Todd S. Braver
GENERAL INDEX
Page references to entire articles are Acquired distinctiveness, in Adrenal medulla, tyrosine
in boldface. Illustrations are indicat- discrimination, 114 hydroxylase modulation, 5
ed by page references in italics. Per- ACTH (adrenocorticotropin) Adrenocorticotropin (ACTH), 233,
sons are indexed only when there is memory effects, 233 476, 642
a substantial reference to them or a memory modulation, 476 -Adrenoreceptor antagonists,
substantial quotation by them. In in stress response, 642 injection into amygdala, 234
references to research work by three Actin -Adrenoreceptors, epinephrines
or more authors, as a rule only the filaments, in synapse structure, effect on, 232
first author is indexed. 215 Affective responses. See Emotion,
in neurons, 205 mood, and memory; Emotions
Actinomycin-D, for transcription Africanized honeybees, 262
A prevention, 167 Age
Ablation, in hypnotic age regression, Action potentials eidetic imagery and, 131
240241 back-propagating, in LTP signal infant memory and, 254, 255
Abraham, W. C., 341 transduction, 351352 See also Aging and memory in
Abstractions, comparative learning dendritic, and long-term animals; Aging and memory
studies, 9192 potentiation, 352, 354 in humans
Accessory olfactory bulb (AOB), 210 generated by neurons, 204 Age regression, hypnotic, 240241
Access to memory. See Cues, Activation view, of implicit memory, Aging and memory in animals, 710
retrieval; Recall; Retrieval 245 eyeblink classical conditioning,
processes in memory Active avoidance learning. See 79, 8
Accommodation and assimilation, Avoidance learning, active and spatial learning and memory, 9
Piaget on, 527 passive transgenic mouse models of
Acetylcholine Activity-dependent regulation of Alzheimers disease, 910
activity-dependent regulation of neurotransmitter synthesis, 46 Aging and memory in humans,
synthesis of, 4 acetylcholine, 4 1013
decrease of, in Alzheimers catecholamines, 46 cross-sectional methods of age
disease, 18, 85 defined, 4 comparison, 10
decrease of, in cognitive See also Protein synthesis episodic memory changes,
impairment, 476 Adaptation aftereffects, in visual 1112
as excitatory neuromodulator, cortex, 421 indirect memory tests of, 11
189190 Adaptive behavior, 566567 prospective memory failure, 13
memory and, 476 Adenylyl cyclase semantic memory changes, 11
neural computation and, 444 in Aplysia sensitization, 34 short-term and working memory
pregnenolone enhancement of, in memory consolidation, 367 changes, 1213
86 Ader, R., 122 underlying changes affecting,
visual cortical plasticity and, 428 ADHD (attentional deficits/ 1011
Acetylcholinesterase inhibitors. See hyperactivity disorder), 320 Agnosia
Cholinesterase inhibitors Adolescents, development of associative, 306
Acetylcoenzyme A, 4 memory in, 7174 hypnotic, 240
Acetyl-l-carnitine (Alcar), 521 Adrenal hyperplasia, congenital, modality-specific associative,
Acoustic startle reflex. See Startle learning and, 614 306307
reflex Adrenalin. See Epinephrine selective, 307
677
678 GENERAL INDEX
Agoraphobia, 523, 524 Amnesia temporally graded, 368, 369

Ai (chimpanzee), 315 double-dissociation in, 549, 673 370
Aitken, Professor (mnemonist), 396 posttraumatic, after head injury, transient global anemia and, 32
Akatinol (memantine), 521 228 Amniocentesis, Down syndrome
Albert B. (Watson research subject), transient global, 3132 diagnosis, 383
667 Amnesia, anterograde AMPA/quisqualate (QA) antagonists,
Alcar (acetyl-l-carnitine), 521 Alzheimers disease, 18 176
Alcohol, memory effects, 118, 120 declarative memory deficits, 106 AMPA receptors
Alcoholism in electroconvulsive therapy, AMPA/QA receptors, 176, 177
aversion therapy, 61 132 C. elegans (Caenorhabditis elegans),
social insect model of, 262 temporal lobe, 414 288, 289, 290
Alex (parrot), 92 transient global anemia and, 32 enhanced conductance
Algorithms, learning, 1417 Amnesia, functional, 2325 hypothesis, 354
applications of, 14 defined, 23 flip and flop elements, 350
error backpropagation, 16 dissociation of implicit and kinase-dependent regulation of,
gradient descent algorithm, 16 explicit memory in, 245246 353354
Hebbian learning, 1415 explicit and implicit memory in, long-term depression signal
least mean square algorithm, 24 transduction, 336, 338
1516 neuropsychology of, 25 long-term potentiation and,
reinforcement learning nonpathological, 24 340, 341, 350, 353354
algorithms, 1617 pathological (dissociative phosphorylation of, in LTD,
statistical learning, 15 disorders), 2324 338
posthypnotic, 24, 240 phosphorylation of, in LTP, 354
temporal difference algorithms,
17 posttraumatic, 2425 See also Non-NMDA receptors
psychological basis of, 2425 Amphetamine, as CNS stimulant, 84,
unsupervised learning
Amnesia, infantile, 2627 120
algorithms, 16
hypnosis and attempts at Amphibians and reptiles, brain
Alkon, D. L., 278
memory recovery, 241 anatomy, 445446
Allergy, to foods, 148
infant memory research and,
Alpha responses, 102 Amygdala, 166168, 183187, 342
257
Alpha-secretase, 22 345
language development and, 26
Alzheimer, Alois, 17 anatomy, 183187, 184
lower boundary for long-term
Alzheimers disease, 1723 in avoidance learning, 451, 453,
recall of early experiences, 26,
behavioral aspects, 1719 467468, 515
418419
Ch4 neurons of basal forebrain basolateral complex of, 333, 343
as nonpathological amnesia, 24
and susceptibility to, 190 beta-adrenoreceptor antagonists
theoretical explanations, 52
cholinergic neuron loss, 476 injected into, 234
See also Autobiographical
clinical-pathological features of, memory blocking and fear conditioning,
1920 303, 304
Amnesia, organic, 2831
diagnosis of, 19 anatomic loci of memory and, central nucleus, 166, 187
in Down syndrome, 383 29 connections with other brain
drug therapy for, 521 declarative memory in, 2930, regions, 185, 186, 187
familial, mutant genes/proteins 106 extended, 185
implicated in, 20 dissociation of implicit and external capsule of, 343
genetically engineered animal explicit memory in, 244 fear conditioning role, 166168,
models of, 1922 dual memory and hippocampal 184, 303, 304, 333, 334, 467
memory and learning damage, 414 468, 545546
impairment, 1819, 111, 520 etiologies of, 2829 fear-potentiated startle, 463464
neuromodulatory system defects, nondeclarative memory in, 30, fear response, 462
444 547 functional anatomy, 185, 186,
neuron vulnerability in, 2122 permanence of, 28 187
prevalence of, 1718 remote memory in, 3031 functions, 183187
semantic dementia in, 602, 603 rodents and nonhuman gene expression and learning,
transgenic mouse models, 910, primates, 107 166168, 167
2122 Amnesia, retrograde glutamate receptors, 464
American Sign Language, 315, 399 declarative memory deficits, 106 intrinsic structure, 184, 184185
Amino acids, as neurotransmitters, 6 in electroconvulsive therapy, lateral nucleus, 166
2-Amino-3-hydroxy-5-methyl-4- 132 learned bradycardia, 455456,
isoxazolepropionate (AMPA) hippocampal, 107, 471 457
receptors. See AMPA receptors posttraumatic amnesia vs., 228 long-term potentiation, 342
Ammons horn, 202, 214 Ribots law of regression and, 30 345, 467, 515
GENERAL INDEX 679
memory trace localization and, procedural learning, 544546 classical conditioning, 3334,
333 sleep and memory 98, 99
neuroanatomy and consolidation, 620 CREB and memory, 44, 367
neurophysiology, 342343 spatial learning, 631632 CS-US correlation in
neurons in, 181, 185 taste aversion and preference conditioning, 99
nuclei, 184, 184185 learning, 645647 development of processes
olfactory input, 210 working memory, 669671 underlying learning, 3741,
opiate antagonist injections in, See also Invertebrate learning; 38
234 Primates; specific animal forms of learning and
perirhinal cortex connections, names developmental timetables, 37
214 Anisomycin (protein-synthesis 39, 38
pre- and parasubicular inhibitor), 280, 366 gill-withdrawal reflex, 98
connections, 214 Anna O. (hypnosis patient), 155 heterosynaptic modulation, 43
in reinforcement, 569 Anomia, semantic dementia and, immediate early genes, 44
stimulus-affect memory and, 602603 life cycle, 37
415416, 545546 A-not-B error, 480481 molecular basis of long-term
synaptic plasticity and fear Antennal lobe, bees, 276 sensitization, 4144
conditioning memory, 334 Anterior commissure, 201 morphology and learning, 402,
-Amyloid (A) Anterior olfactory nucleus (AON), 403
A peptide vaccination for 210 neural mechanisms of aversive
reduced deposition, 9 Anterograde amnesia. See Amnesia, classical conditioning, 3334,
accumulation of, in Alzheimers anterograde 34
disease, 20, 110 Anticholinergic agents. See operant conditioning, 3436
transgenic models of Cholinesterase inhibitors plasticity in sensory neurons,
amyloidogenesis, 21 Anticipation method, serial learning, 100
Amyloid precursor protein (APP) 610611, 611 postsynaptic neuron
APPswe gene mutation, 20 Antisense oligodeoxynucleotides, modulation, 43
beta-amyloid accumulation and, injection into amygdala, in taste presynaptic sensory neuron
20 aversion learning, 168 modulation, 4243
mouse models, 21 Ants, learning in, 260261 second messengers and protein
mutation of, in familial Anxiety kinase cascades, 4344
Alzheimers disease, 20 mood-congruent memory and, second-order conditioning, 98
Amyloid precursor protein (APP)- 133 siphon gill withdrawal reflexes,
cleaving enzymes rumination about personal 33, 37, 3940, 42, 98, 401
BACE1 -/- mouse model, 21 concerns, 134 tail-shock induced inhibition of
BACE1 cleaving enzyme, 20, 22 ApAF (Aplysia Activating Factor), 367 siphon withdrawal, 3940
BACE2 cleaving enzyme, 22 L-AP4 (1-2-amino-4- tail siphon withdrawal reflex,
neuron damage due to, 2122 phosphonobutanoic acid), 176 42, 401
Anagenesis, theories of, 260 AP5 (D-2-amino-5- time factors in structural
Analogical reasoning, animal phosphonovalerate), 347, 348 change, 402
learning, 92 D-AP5 (amino-5- APP. See Amyloid precursor protein
Analogous traits, 259 phosphonopentanoate), 177 (APP)
Analogy, in concept learning, 9697 ApCAM (neuronal cell adhesion Appearance and reality, source
Anatomy. See Brain anatomy molecule in Aplysia), 43 monitoring, 629
Anderson, J. R., 97 ApC/EBP (CCATT enhancer-binding Appetitive behavior. See Adaptive
Anderson, M. C., 271 protein), 367 behavior
Androgen (testosterone) Aperture problem, 442, 442 Appetitive operant conditioning, in
birdsong learning and, 6869 Apes. See Chimpanzees; Monkeys Aplysia, 35, 35
memory effects, 234 and apes Apraxia, ideational, 604
spatial learning and, 614 Aphasia ApUCH (Aplysia ubiquitin C-terminal
Angell, James R., 238, 363, 665, 666 Lurias studies, 357 hydrolase), 44
Animals modality-specific, 307 APV glutamate antagonist, 344, 345
aging and memory, 710 Apis mellifera. See Bees Arachidonic acid, as retrograde
avoidance learning, 13 Aplysia, 3344 signal, 349, 352
comparative cognition, 8894 behavioral dissociation of Aricept (donepezil), for Alzheimers
dietary specialists vs. generalists, dishabituation, sensitization, disease, 521
147148 and inhibition, 40 Aristotle, 45, 4547
dual memory theories, 414415 cellular analogues of learning memory theories of, 4647
evolution and learning, 137140 and behavioral learning, 39 school learning and, 591
fear of, 524, 667 cellular correlates of long-term Arthropoda
foraging, 149151, 261 sensitization, 42 types of insects, 258
680 GENERAL INDEX
Arthropoda (continued) Associative networks, 82 Australian High Avoidance and

See also Insect learning Associative recall, defined, 439 Australian Low Avoidance rat
Artificial grammar (test), 549 Astereognosis, 306 strains, 3
Artificial intelligence, explanation- Astrocytes Autism, savant syndrome in, 587,
based learning, 96 environmental enrichment and, 588
Artificial neural network systems 407 Autoassociation, hippocampus in,
competitive learning in, 435 as glial cell, 181 439, 441
principles of, 433 Atkinson, R. C., 672673 Autobiographical memory, 5154
Aserinsky, Eugene, 618 Atropine, blocking of naloxone, 234 age of earliest memory, 26, 52
Assembly coding, 500 Attention, 4851 age-related differences, 1112
Assimilation and accommodation, central, 48, 50 brain injury or psychiatric
Piaget on, 527 controlled, working memory illness effects, 5354
Associationism and, 266 defined, 51
Hull and, 236 control of, 539, 674 depression and, 134
Thorndike on, 648649 defined, 48 emergence of, 2627
Underwood on, 656657 dj vu and, 109 infantile amnesia and, 52
Association regions, in cerebral divided, 48 lifespan retrieval curve in, 52,
cortex lobes, 200 EEG patterns, 501 53
Associations forms of, 48 parent-child conversations about
by contiguity, 220 forms of memory and, 49 past, 27
doctrine of, Guthries personality and, 53
implicit memory and, 5051
modification of, 219 relation to self, 5152
limitations, 48
Ebbinghauss studies, 126 See also Amnesia, infantile;
long-term memory and, 50
mnemonic devices and, 393 Episodic memory
mood impairment and, 134
organization of memory and, 82 Autonoetic consciousness, 136
neuronal synchronization, 501
selective, 645 Autoshaping, classical conditioning,
observational learning and, 482,
serial organization and, 611 101
483
612 Aversion
relation to memory, 4851
Associative agnosia, modality- aversive conditioning in
same sense vs. different input
specific, 306307 Drosophila melanogaster, 260
modalities, 48
Associative connections, long-term food aversion and preference
in school learning, 592
potentiation, 340 learning in humans, 147149,
Associative learning selectivity in, 539 524
Aplysia, 3336 sensory memory and, 49 uncontrollable events, in
bees, 261, 262, 273276 short-term memory and, 4950 learned helplessness, 319
biological substrates, 101102 visual, in primates, 539540 See also Taste aversion and
C. elegans (Caenorhabditis elegans), See also Repetition and learning preference learning in
290 Attentional deficits/hyperactivity animals
contingency theory and, 102 disorder (ADHD), 320 Aversion stimulus. See Negative
Diptera, 260 Attention and Memory: An Integrated reinforcer
Drosophila, 293 Framework (Cowan), 608 Aversion therapy, for alcoholism and
early handling of animals and, Attractor states, stored memories aberrant sexual behavior, 61
122 and, 443444 Avoidance learning, active and
frontal lobes in, 159160 Auditory associative agnosia, 306 passive, 13
Hermissenda, 277280 Auditory cortex active free-operant procedure, 1
Limax, 281285 conditioned stimulus amygdala in, 451, 453, 467468,
operants inapplicability in, 103 transmission, 455 515
prefrontal cortex function, 161 learning and, 422 behavioral phenomena, 13
spinal plasticity, 640 mechanisms, 425 defined, 451
synaptogenesis in, 406 neocortical plasticity and, 422 discrete-trial procedure, 1
See also Conditioning, classical; 425 early handling of animals and,
Conditioning, instrumental receptive field plasticity, 422, 122
(operant); Morphological 424, 424425 genetic control of, 23
basis of learning and memory, Auditory imprinting, 248 instrumental conditioning, 103
in invertebrates; Auditory memory. See Modality one-trial passive-avoidance
Morphological basis of effects; Precategorical acoustic training, 2
learning and memory, in storage; Sensory memory passive (inhibitory) avoidance,
vertebrates; Nonassociative Auditory processing deficits, 512516
learning learning disabilities related to, Pavlovian conditioning in, 2
Associative memory. See Secondary 322, 323 punishment procedure, 1
memory (long-term memory) Austin (chimpanzee), 315 rodents, 23
GENERAL INDEX 681
See also Conditioning, classical; BDNF (brain-derived neurotrophic cognitive behavior therapy vs.,
Reinforcement factor), birdsong learning and, 69 61
Avoidance learning, neural Bees, 273276 fear-reduction procedures,
substrates of, 451454 associative learning, 273276 5960
active, 451453 blocking, 261, 274 operant conditioning
information processing, 451, classical conditioning, 274 procedures, 6061
452 foraging, 261 Beliefs, memories as, 299
inhibitory, 453 memory dynamics and Benzer, S., 293
limbic system, 451453, 452 localization, 274276, 275 Benzodiazepines
motor system, 451453, 452 navigation in, 392 dissociation of implicit and
Axons neural morphology and explicit memory, 244
amygdalar nuclei, 185 learning, 401 GABA effects, 118, 120
anatomy, 206, 207, 208, 208 proboscis extension Bergson, Henri, 585
209 conditioning, 261 Bernheim, Hippolyte, 155
cerebral cortex, 201 risk sensitivity and choice Bernstein, Nicholas, 357
hillock of, 217 behavior, 262 Best frequency, cellular, defined,
initial segment, 208 spatial learning, 632 422
myelin sheath, 209 Behavior Beta-adrenoreceptor antagonists,
Ayllon, T., 60, 61 as criterion for learning injection into amygdala, 234
Azrin, T., 60, 61 disabilities, 320 Beta-adrenoreceptors, epinephrines
Lorenzs view, 356 effect on, 232
Behavior: An Introduction to Beta-amyloid. See Amyloid (A)
B Comparative Psychology (Watson), Beta-endorphin
Babbit, Raymond (movie character), 652 antagonists of, 234
588 Behavior, operant. See Operant memory effects, 233
BACE1 -/- mice, 21 behavior Beyond Freedom and Dignity (Skinner),
BACE1 cleaving enzyme, 20, 22 Behavioral contrast, 495 617
BACE2 cleaving enzyme, 22 Behavioral control, learned Bias, in life experience recall, 418
Baddeley, A. D., 264, 673675 helplessness and, 319 Bignami, G., 2
Bailey, D. J., 167 Behavioral phenomena Binding theory, 312, 313
Bain, Alexander, 325326 avoidance learning, 13 Binet, Alfred, 395, 585
Bammer, G., 3 discrete, in classical Binocularity, development of, 426
Barnes, C. A., 9 conditioning, 458460 427
Bartlett, Frederic, 5556, 56 long-term potentiation and, Binocular neurons, properties of,
on collective memory, 87 345, 346348 426
on false memory, 145 Behavior analysis (school of Birds
on reconstructive memory, 558 behaviorism), 5859, 616 episodic memory, 9293
559 Behaviorism, 5759 evolution of learning in, 139
schemata concept, 417 goal-oriented, 652 foraging, 150151
Basal forebrain, 187190 Hulls views, 57, 533 imprinting, 247250
amygdalar connections, 187 neobehaviorism and, 57 migration, 390391
anatomy, 187190 origin of theories of, 5759 navigation, 391392
cholinergic projections, 187, Skinners views, 57, 58, 533, recall in, 669
188 616617 spatial learning, 151, 631, 632
Basal ganglia, 191193 Tolmans views, 5758, 533, 652 See also Birdsong learning; Food
anatomy, 191193 Watson as founder of, 57, 533, caching/storing
direct and indirect pathways, 666 Birdsong learning, 6169
192 Behaviorism (Watson), 616, 668 auditory memory in, 63
double-inhibition circuit, 191, Behavior modification. See Behavior basic functions of songs, 6162,
191 therapy; Operant behavior 62
in motor skills learning, 548 The Behavior of Organisms (Skinner), brain nuclei size and, 139
movement release pathway, 191, 493, 616 hormonal effects and sex
192 A Behavior System: An Introduction to differences, 6769, 68
projections to brain areas, 191, Behavior Theory Concerning the interaction with neighboring
191 Individual Organism (Hull), 236 birds, 63
in reward, 576 Behavior theory. See Learning isolation and deafness effects,
in stimulus-response learning, theory 61, 62, 63
544545 Behavior Theory and Conditioning monitored learning in, 434435
Basolateral nucleus, in learned (Spence), 638 motor pathway for song, 6465,
bradycardia, 455456 Behavior therapy, 5961 66
BCM rule, 388 aversion therapy, 61 Oscine species learning, 62
682 GENERAL INDEX
Birdsong learning (continued) Brain anatomy, 179218 Bush, R. R., 360361

overproduction of songs, 63, 64, amygdala, 183187, 184 Butcher on the bus phenomenon,
65 basal forebrain, 187190 12
sensitive period and timing, 62 basal ganglia, 191193 Butterflies, mimicry, 139
63, 138 cerebellum, 193199
song development, 63, 64, 65 cerebral cortex, 200202
song-selective auditory chemoarchitecture, 182183
C
C. elegans. See Caenorhabditis elegans
pathways, 6567, 66 cytoarchitecture, 180181 Ca++. See Calcium ions
song system, 64, 66, 6769, 68, functional systems, 183 CA/CAM kinase II (calcium-
138 gross anatomy, 180 calmodulin-dependent protein
subsongs, 63, 64 hippocampus and kinase). See Calcium-calmodulin-
syntax of bird songs, 62, 63 parahippocampal region, dependent protein kinase (CA/
Bjork, E. L., 271 202204 CAM kinase II)
Bjork, R. A., 271 in learning, 566570, 567 Cache sites. See Food caching/storing
Blindness mammalian vs. reptile, 445446 Caenorhabditis elegans, 287291
imagery and concreteness neurons, 204210 advantages as model system,
effects, 79 olfactory cortex, 210213 287288, 288
inattentional, dj vu and, 109 overview, 179183 associative learning, 290
Blocking, 301304 perirhinal cortex and associated eat-4 gene, 289290
amygdala and fear conditioning, cortical areas, 213214 learning in, 287291
303, 304 synapses, 215218 long-term memory, 290
bees, 261, 274 See also specific anatomic area short-term memory, 288290,
cerebellum and eyeblink Brain damage 289
conditioning, 302303, 303 amnesia after, 28 Caffeine, as CNS stimulant, 84
classical conditioning, 76, 98 autobiographical memory Cajal, Santiago Ramn y. See Ramn
99, 100, 301304 effects, 5354 y Cajal, Santiago
invertebrates, 261 dissociation of implicit and Calcineurin
Kamins effect, 76, 261, 301 explicit memory in, 245246 hippocampal learning, 173
304, 360 eidetic imagery as marker of, long-term depression induction,
law of habit growth and, 360 131 338
paradigm of, 301, 302t minimal, 320 LTP induction regulation, 354
Rescorla-Wagner model/rule, Brain-derived neurotrophic factor Calcium
302, 329330 (BDNF), birdsong learning, 69 influx into cerebellar Purkinje
septohippocampal system in, Brain Mechanisms and Intelligence cells, 197
304 (Lashley), 231, 318, 332 long-term depression induction,
striatum and dopaminergic Brain stem (hindbrain), 180, 187 336337
midbrain neurons, 303 Brain stimulation reward (BSR), long-term potentiation
theory of, 302 573575 induction, 352
Blowflies (Phormia regina), 260 directly stimulated stage, 574 Calcium-calmodulin-dependent
Bonobos, language studies in, 316, dopamine neurons, 574575 protein kinase (CA/CAM kinase II)
316317 mapping of, 567 CREB activation by, 174, 367
Borszcz, G. S., 225 natural rewards and, 573574 hippocampal long-term
Bott, Edward A., 653 neural circuitry, 574575 potentiation and, 172173
Bouton, Mark E., 331 Brainwashing, Pavlovian response in, long-term potentiation
Boutons 518 induction, 352
multiple synaptic, 407 Breuer, Josef, 155 phosphorylation of AMPA
presynaptic, 215 Brief Contributions to the Psychophysics receptors, 354
synaptic junctions, 206, 209 of Color Sensations (Mller), 412 second messenger activation of,
synaptic vesicles, 209 Brightness discrimination, 663664 595, 596
Bradycardia, learned, central Broca, diagonal band of, 187188 tyrosine hydroxylase
structures in, 455457, 456 Brocas area phosphorylation, 5
Brain functions, 200 Calcium-calmodulin kinase IV
birdsong learning and, 6469 lesions and speech impairment, (CaMKIV), 367
evolution and, 649 536 Calcium channel-dependent long-
maturation of, and long-term Brock, L., 128 term potentiation, 344
recall of early experiences, 26 Brodmann, K., 182 Calcium-dependent phosphatases, in
See also Brain anatomy; Brain Brush, F. R., 3 LTP induction regulation, 354
damage; Brain stimulation BSR. See Brain stimulation reward Calcium ions
reward (BSR); Central (BSR) long-term depression induction,
nervous system; Neural Buller, A., 128 337
headings; specific regions Bumblebees, 138, 140 NMDA receptors, 177
GENERAL INDEX 683
as primary regulators, 593594 See also Glial cells; Neurons; in neural control, 437438, 438
second messenger systems, 596 Purkinje cells; Pyramidal cells neuronal network, 436, 437
stress-related metaplasticity, 389 Cell theory, proposal of, 557 nictitating membrane/eyelid
cAMP. See Cyclic adenosine Cellular analogues of learning, in classical conditioning, 458
monophosphate (cAMP); Protein Aplysia, 39 459
kinases, cAMP-dependent (PKA) Cellular and network schemes for reinforcement or reward in
Campeau, S., 166 higher-order features of classical learning, 570572
Canaries, and song learning, 62, 67 conditioning, 98100 vestibulo-ocular reflex plasticity,
3-(2-Carboxypiperazin-4-yl)propyl-1- Cellular correlates of long-term 170, 659660
phosphate (CPP), 172173, 177 sensitization, in Aplysia, 42 white matter stimulation, 571
Carew, Thomas J., 33 Central excitatory state (CES) 572
Carey, S., 96 Eccless work on, 127 wiring diagram (synaptic
CARIN model, 600601 flies, 260 organization), 195, 197, 197
Carpenter, P. A., 252, 265, 266 Central inhibitory state, Eccless Cerebellum as a Neuronal Machine
Carr, Harvey, 238, 362, 363, 666 work on, 127 (Eccles), 129
Ca+2-stimulated cyclase, in Aplysia Central nervous system Cerebral cortex, 200202
sensitization, 34 maturation of, and long-term abnormal long-term
Catecholamines recall of early experiences, 26 potentiation, 368
activity-dependent regulation of stimulant drugs, 8485 anatomy, 200202
synthesis of, 46 structural organization of, 434 anomalies of, dyslexia and, 322
groups of catecholaminergic See also Brain; Brain anatomy; 323
neurons, 182 Neural headings; Synapses cholinergic projections from
memory modulation, 476 Central pattern generator basal forebrain, 190
Categorical knowledge, 598599 interneurons, Tritonia, 291 connections with brain areas,
Categorization. See Concepts and Cerebellar cortex, eyeblink classical 200201
categories, learning of conditioning and, 169 cortical areas, 200, 200
Category-specific knowledge loss, Cerebellar-like structure of fish, gross anatomy, 180
307308 long-term depression and, 335, hippocampus as component of,
Caudate nucleus 338 202
in active avoidance learning, Cerebellar vermis layers and cell types, 201, 201
451 in learned bradycardia, 455 lobes of, 200, 200
cognitive-affective disorders 456, 457
long-term memory and, 333
associated with, 192 in long-term habituation, 225
memory consolidation and, 368
in forebrain, 180 Cerebellum, 169171, 193199
neuron structure, 181
multiple memory systems and, anatomy, 193199, 436
perirhinal cortex connections,
415, 534 complex motor-skill learning,
213
neuropsychiatric disorders 170171
postsynaptic muscarinic and
associated with, 192193 cytoarchitecture, 181
nicotinic receptors, 189
as striatal component, 192 eyeblink classical conditioning
subcortical structures, 180, 181
Caudolateral ventral hyperstriatum and, 78, 169, 197198, 198,
See also Neocortex
(clHV), in birdsong learning, 67 199, 302303, 303, 415, 544,
Causal relations, in prose retention, 571572 Cerebral ventricles, 180
552 feedforward mechanism in, 198 Cerebrovascular disease, dementia
CCAAT/enhancer binding protein genetic substrates of memory, related to, 110
(C/EBP), 44, 367 169171 CES (Central excitatory state), 127,
CCK (cholecystokinin), 234 hindbrain location of, 180 260
C/EBP (CCATT enhancer-binding inferior olive, 571 c-fos, 6, 166
protein), 367 learning, sequence of events in, cGMP (cyclic guanosine
Celestial rotation, in migratory 197199, 199 monophosphate)
orientation, 390 lesions of, 193 degradation, 596
Cell adhesion molecule, neuronal location in vertebrates, 194, 197 long-term depression induction,
(ApCAM), in Aplysia, 43 long-term depression in, 170, 337338
Cell assemblies, of Hebb, 231 335, 336, 436437 Chain associations, 611, 612
Cell body (soma), of neurons, 205 Marrs theory of, 197 Chandelier cell, 207, 208
Cell proliferation, neurogenesis, 471 microcomplex, 436437 Chantek (orangutan), 315
Cells motor functions of, 193, 195 Charcot, Jean, 155
best frequency, 422 in motor skills learning, 548 Chase, W. G., 140, 141
orienting reflex, 498 mutant mouse models in Chemoarchitecture of brain, 182
progenitor, 471 classical eyeblink 183
Schwann cells, 181 conditioning, 169170 Chen, C., 171
stellate cells, 185 neural computation, 436438 Chen, L., 169
684 GENERAL INDEX
Chess masters superior memory, Climbing fibers Cognitive enhancers, 8486

140143, 141, 267 cerebellar cortex, 169, 436 cholinergic agents, 85
Chickadee family (Paridae), 139 cerebellar long-term depression, CNS stimulants, 8485
Childhood 170, 335 mechanisms of, 120
hypnotic age regression and, cerebellum, 169170, 196, 197 over-the-counter agents, 8586,
240241 Closed head injury. See Head injury 522
infant memory and childhood CNS. See Central nervous system Cognitive interview, 417
IQ, 257 Cocaine, Freud and, 155 Cognitive maps
Children Cockroaches, 259 ants, 261
food aversion and preference Coding processes, 7783 bees, 261, 273
learning, 148149 aging effects, 12 Clarks nutcrackers and
grammar acquisition, 312314 assembly coding, 500 landmark navigation, 9091
learning disabilities in, 320323 associative networks, 8182, 500 Cognitive psychology
Children, development of memory behavioral responses, 459 connectionist models, 441
in, 7174 coding redundancy, 79 revolution in, 236
eyewitness testimony and, 73 deficient processing, and Cognitive Psychology and Emotional
knowledge and memory, 72, 72 distributed practice effect, Disorders (Williams), 133
memory strategies, 7172 116117 Cohen, M., 131
origins and early development, definition of encoding, 373 Collaboration
7273 dual coding and retrieval, dj in group recall, 621622, 622
parent-child conversations about vu and, 109 inhibition in, 622
past, 27 encoding-variability theory of recognition tests and, 623
working memory and, 7374 distributed practice effect, 117
Collective memory, 8688
See also Amnesia, infantile; ensemble, 532
group recall and, 623
Infancy, memory in episodic memory tasks, 136
history vs., 88
Chimpanzees false memory and, 146147
individual memory vs., 8788
language learning, 315317, imagery, 7779
strong and distributed accounts
316 incidental vs. intentional, 81
of, 8687
spatial learning, 632 infant visual recognition
Colliculi, inferior and superior, 180,
Cholecystokinin (CCK), 234 memory, 254
181, 455
Choline, in acetylcholine synthesis, 4 intelligence and memory, 264,
Color perception
Choline acetyltransferase (CAT), 4 264265, 267
integration in, 444
Cholinergic agents, as cognitive levels of processing, 8081
enhancers, 85 Mller on, 412
line coding, 500
Cholinergic system neural computation, 442
linguistic, 399
basal forebrain, 187190 Commissure, anterior, 201
material to be remembered, 81
blocking and, 304 multiplexing, 500 Comparative cognition, 8894
Cholinesterase inhibitors organization of memory, 8283 abstract concept learning, 9192
for Alzheimers disease, 111, orienting tasks, 8081 Africanized honeybees, 262
120, 521 place cells, 532 analogical reasoning, 92
for Parkinsons disease, 118 prose retention and, 550551 criticisms of research in, 9394
physostigmine as, 85 semantic encoding, functional episodic memory, 9293
Chomsky, N., 313 dissociation and, 244245 evolution and learning studies,
Chorionic villus sampling, Down semantic networks and features, 138139
syndrome diagnosis and, 383 8283 future research, 93
Christal, R., 264 Semons theories, 606 Harlows studies, 227228
Christie, F. M., 50 source monitoring, 629630 insect learning, 258259
Cingulate cortical neurons, in active specificity principle, 582 interval timing, 89, 89, 90
avoidance learning, 451 storage vs., 580581 landmark navigation and
Circumscribed phobias, 523 synchronization, 500501 cognitive maps, 9091
Cladogenesis, defined, 258 Codon formation, 440 Lorenzs studies, 355356
Clarks nutcracker, 9091, 93, 151, Cognition metacognition, 9293
631, 669 semantic memory and, 598601 metaplasticity in information
Classical conditioning. See Tolmans view of behaviorism, processing, 387388
Conditioning, classical 57 number scale learning, 8990
Claustrum, entorhinal cortex input, See also Comparative cognition; Complex phobias, 523
214 Learning Component-processes approach,
clHV (caudolateral ventral Cognitive behavior therapy, 61 implicit memory, 246
hyperstriatum), in birdsong Cognitive deficits, in schizophrenia, Compound stimulus trials, 330
learning, 67 588 Computational reinforcement, 565
GENERAL INDEX 685
Computers See also Unconditioned response blocking in, 76, 9899, 100,
Hulls theories and, 235236, (UR) 301304
237 Conditioned stimulus (CS) C. elegans (Caenorhabditis elegans)
information processing systems, alpha response effects, 102 learning, 290
433 ambiguous, interpretation of, classical-instrumental transfer
neural system simulation, 435 330331 procedures, 101
reinforcement learning, 565 Aplysia siphon-gill withdrawal conditioned response timing, 76
Concepts and categories, learning reflex, 33 contingency in, 99
of, 9597 auditory, 422 control methodology, 102
abstract concepts, animal blocking effect and, 98, 100, cue competition (blocking), 76,
learning of, 9192 301302, 302t, 330 627628
acquisition of concepts, 9697 in classical conditioning, 7476, defense conditioning, 101
animal learning of, 9192 101, 454 defined, 454
Aristotle on stored/ correlation between discrimination and
representations/resemblance, unconditioned response and generalization, 113115
4647 (contingency), 99 evaluative, 148
associative networks in, 82 excitatory classical conditioning, excitatory, 7475
categorical knowledge, 598599 7475 eyeblink classical conditioning,
classical vs. probabilistic view, 95 eyeblink classical conditioning, 79, 8, 169170
combinations of, 600601 7, 169 Hermissenda, 277280
defining features, 95 inhibitory classical conditioning, higher-order, 98100
definition of concept, 95 75 infant memory tests, 255, 257
grounding in perception and overshadowing of less salient inhibitory, 75
action, 83 CS, 76 instrumental conditioning vs.,
memory organization and, pathways, 455 101
8283 phobias as, 523524 Limax, 281285
in operant conditioning, 495 neural substrates, 454465
preexposure, in classical
prototype view, 95 paradigms, 101
conditioning, 7576
reasons for human learning of reward conditioning, 101
preexposure, in insects, 261
concepts, 97 second-order, 98, 99, 99100
262
semantic networks in, 8283 sometimes opponent process
proboscis-extension reflex, in
structure of human concepts, model, 624628
bees, 274
9596 stimulus preexposure, 7576
pseudo-conditioned and
theory view, 9596 stimulus-stimulus paradigm, 101
sensitized response, 102
Concreteness effect, 78, 79 truly random control and, 102
reinforcement and reward, 571
Conditional-reflex method. See visceral responses, 455
second-order conditioning, 98
Conditioning, classical; Pavlovian See also Associative learning;
conditioning taste aversion, 646 Conditioned response (CR);
Conditioned and Unconditioned unpaired control procedure, Conditioned stimulus (CS);
Reflexes and Inhibition (Hebb), 102 Habituation; Pavlov, Ivan;
230 See also Interstimulus intervals Pavlovian conditioning;
Conditioned emotional response (ISI); Stimulus-response (S-R) Rescorla-Wagner model/rule
(CER), latent inhibition studies, learning; Unconditioned Conditioning, classical and
123 stimulus (US) instrumental. See Conditioning,
Conditioned inhibition, 75, 113, 310 Conditioned stimulus (CS)/ classical; Conditioning,
Conditioned Reflexes (Pavlov), 616 Unconditioned stimulus (US) instrumental (operant); Operant
Conditioned Reflexes and Neuron association, blocking effect and, behavior; Pavlovian conditioning
Organization (Konorski), 309 301302, 302t Conditioning, instrumental (operant)
Conditioned reinforcers. See Conditioned suppression, 101 Aplysia, 3436
Reinforcement Conditioning, aversive, Drosophila behavior therapy, 6061
Conditioned response (CR) melanogaster, 260 classical conditioning vs., 101
Aplysia siphon-gill withdrawal Conditioning, cellular and network classical-instrumental transfer
reflex, 33 schemes for higher-order features procedures, 101
cardiovascular, 454457 of classical, 98100 control methodology, 103
in classical conditioning, 74, 76, Conditioning, classical, 7477, 100 discriminative conditioning
101 102 paradigms, 103
defined, 7 Aplysia, 3334 Konorski-Miller studies, 309
infant foot-kick response, 255 auditory conditioned stimulus, nonassociative response, 103
Pavlovian fear conditioning, 461 422 paradigms, 103
timing, in classical conditioning, bees, 274 reinforcement methods, 493
76 biological substrates, 101102 494, 563564
686 GENERAL INDEX
Conditioning (continued) Corticosteroid receptors. See food choices in humans, 148

reinforcement schedules, 494 Glucocorticoid receptors 149
495 Corticosterone. See Glucocorticoids spatial learning and, 635
yoked-control design, 103 Corticotropin-releasing factor (CRF) Cunitz, A. R., 49
See also Operant behavior; memory effects, 233 Cuthill, I. C., 150
Thorndike, Edward as neuropeptide, 182183 Cyclic adenosine monophosphate
Confabulation, frontal lobe damage, Cortisol. See Glucocorticoids (cAMP)
161 Courts of law, hypnotic recovery of Aplysia sensitization, 34, 4344
Confidence judgments, retrospective, memory and, 241242 degradation, 596
385386 Cowan, Nelson, 608 Drosophila gene mutation, 138
Configural approach, in learning Cpg15, ocular dominance plasticity Drosophila olfactory memory
theory, 330 and, 420 formation, 293294
Connectionism, 362 CPP (3-(2-Carboxypiperazin-4- modulation of Aplysia
Connectionist networks yl)propyl-1-phosphate) presynaptic neuron, 42
models, 441 blocking of hippocampal- ocular dominance plasticity, 420
Rescorla-Wagner model as dependent learning, 172173 as second messenger, 43, 402,
special case of, 362 as NMDA receptor antagonist, 594
Connection weights, Konorskis rule 177 signaling pathway, hippocampal
for changing, 310 learning and, 173
CR. See Conditioned response (CR)
Consciousness Cyclic adenosine monophosphate-
Craik, F. I. M., 12, 80
autonoetic, 136 dependent protein kinase. See
Cranney, J., 225
declarative memory and, 106 Protein kinases, cAMP-dependent
Crayfish, neuromuscular junction
107 (PKA)
ultrastructure, 402403
diminished, dj vu and, 109 Cyclic AMP response element
CREB. See Cyclic AMP response
Consolidation. See Amnesia, organic; binding (CREB) protein
element binding (CREB) protein
Memory consolidation activation of, 367
Cre-recombinase, gene deletion amygdalar learning, 166167,
Constancy, in insect pollination, 140 method, 173
Contast, law of, 219 168
Crimes. See Eyewitness testimony Aplysia memory, 44, 367
Context Cross-cultural psychology
C. elegans (Caenorhabditis elegans) downstream targets, in memory
Bartletts contribution, 55 consolidation, 367
learning, 290
Lurias contribution, 357 memory consolidation, 366367
of conditioning and extinction,
Crow, T., 278 phosphorylation of, 5
330331
Crowder, Robert G., 398 second messenger activation,
infant memory specificity and,
CS. See Conditioned stimulus (CS) 595
255
CSP24 protein, 280 transcription and long-term
Context effect, 598, 628
Contiguity, law of association by, 220 Cue-dependency theory of memory, 173174
Contingencies of Reinforcement forgetting, 153154, 153t Cyclic AMP response element
(Skinner), 617 Cues, retrieval binding transcriptional activators
Contingency childhood memory (CREB1, CREB2), 44
correlation between CS and US, development, 72, 72 Cyclic guanosine monophosphate
99 cued recall test, 364 (cGMP)
positive vs. negative, 102 extra-list cue, 364 degradation, 596
Contingency hypothesis, classical intra-list cue, 364 long-term depression induction,
conditioning, 102 memory rehabilitation, 562 337338
Contingent relationship, mnemonic devices, 393, 397 Cysternae, of endoplasmic reticulum,
instrumental conditioning, 103 natural settings, 418419 205
Contrast, behavioral, 495 overload principle, 271, 582 subsurface, 205
part-list cuing, interference and, Cytoarchitecture of brain, 180181
Control processes
271 Cytokeratin, of neurons, 205
metacognition, 386387
testing with, 581582 Cytoskeleton, neuronal, 205
retrieval, 386387
self-paced learning and, 386 vanishing, 562
Conway, M. A., 52, 88 Cuisine, in human food preference D
Coombs, J., 128 and selection, 149 DAG (diacylglycerol), 337, 595
Cooperative groups, 591 Cultural-familial mental retardation, Dale, H., 128
Corpus callosum 382 Dallenbach, K. M., 152
cortical communication via, 201 Cultural referencing. See Schemata Daneman, M., 252, 265, 266
forebrain location, 180 Culture and society D-AP5 (amino-5-
Corpus striatum. See Striatum collective memory, 8688 phosphonopentanoate), as NMDA
Cortical dementia, 110111, 111t culturally transmitted behavior receptor antagonist, 177
Cortical lobes. See Neocortex in animals, 139140 Darwin, Charles, 140, 238, 645
GENERAL INDEX 687
Davies, A., 48 long-term memory in, 111 Piaget stages of, 527
Deafness, birdsong learning effects, memory function in, 110111, Dewey, John, 665
62, 62, 63, 67 111t De Zeeuw, C. I., 170, 171
Decay theory, of interference and mental retardation vs., 110 Diacylglycerol (DAG), 337, 595
forgetting, 268269 mild cognitive impairment vs., Diagonal band of Broca, 187188
Declarative memory, 105107 112 Diary studies, 418
amnesia and, 24, 2930, 106, pathophysiology, 110 Diencephalon, 29, 159
369 presenile, 110 Diet and nutrition
brain mechanisms of, 107 recent research developments, dietary specialists vs. generalists,
consciousness and, 106107 111112 147148
defined, 415, 547 semantic, 602603 food aversion and preference
dementia and, 111 senile, 110 learning in humans, 147149
dissociation from implicit short-term memory in, 111 See also Food aversion and
memory, 244245 subcortical, 110111, 111t preference learning in
episodic and semantic memory See also Alzheimers disease humans; Food caching/storing
as subcategories, 136 Dementia praecox. See Dieters, Otto Friedrich Karl, 557
memory trace localization and, Schizophrenia Differential-interference contrast
333 Dempster, F. N., 116 (DIC) optics, 343
nature of, 105106 Dendrites Digit span
neuroanatomy, 544 action potentials of, 352, 354 intelligence tests and, 267
procedural memory vs., 105, anatomy, 205206 measure of recall, 364
333, 544 apical, of pyramidal cells, 209 memory testing and, 380
retention latency, 513 basal region, pyramidal cells, Diptera learning, 260
sleep and consolidation, 620 206, 209 See also Drosophila melanogaster
tests of, for amnesia, 2930, learning and changes in, 406 (fruit fly)
106, 548 pre- and postsynaptic, 205
Direct-access retrieval
Deese, James, 559 Dendritic spines
content-addressable operation
Defense conditioning, classical, 101 anatomy, 205, 206
in, 377
Defense mechanisms, 156 backpropagating, long-term
item information retrieval, 374
Defensive siphon withdrawal reflex potentiation and, 344
375, 376
(SWR), 389 constricted neck, 217
long-term memory retrieval,
Deferred imitation, infant memory electrical property changes, in
376
testing, 255256, 256 LTD, 350
short-term vs. long-term
Deficient processing theory of long-term potentiation and,
memory, 373374
distributed practice effect, 116 340, 350
time course of short-term
117 structure and function, 217218
memory retrieval, 376, 377
Deficits, memory. See Amnesia synaptic functions, 216, 217
Discontinuation of reinforcement
headings; Forgetting Denenberg, V. H., 122
(extinction), 330, 493
Degeneration and Regeneration of the Dentate gyrus, 202
Discrete-trial procedure
Nervous System (Ramn y Cajal), Depersonalization, 23
558 Depression (emotional) avoidance learning, 1
Degradation, graceful, 510 electroconvulsive therapy for, behavior modification, 494
Dj vu, 108109 132 Discrimination and generalization,
Delaney, P. F., 267 experimental models, 641 113115
Delayed response (DR)/delayed learned helplessness and, 320 acquired distinctiveness in, 114
alternation (DA) tasks mood-congruent memory and, defined, 113
frontal lobe damage, 159 133 gradients of, 495
Hunters application of, 237, mood-impaired memory in, 134 as memory characteristic, 508
239 stress and, 389 occasion setting paradigms, 113
Delay procedure, excitatory classical Depression, long-term. See Long- operant behavior, 495
conditioning, 74 term depression (LTD) parallel distributed processing
Delirium, dementia vs., 110 Derealization, 23 model, 509
De memoria et reminiscentia (Aristotle), Descartes, Ren, 129 pattern learning and, 113114
46 Descent of Man (Darwin), 238 perceptual learning and, 114
Dementia, 109112 Desensitization (behavioral therapy) Spences studies, 637638
behavioral changes in, 109110 systematic, 59, 60 time discrimination in, 114115
cortical, 110111, 111t Dethier, V. G., 260 visual, 663664
delirium vs., 110 Developmental delay, prefrontal Discrimination procedure, classical
diagnostic criteria, refined, 112 cortex lesions, 536 conditioning, 75
diseases associated with, 110 Developmental processes Discriminative approach procedure,
drug therapy, 111 object concept, 479481 101
688 GENERAL INDEX
Discriminative instrumental Donepezil (Aricept), for Alzheimers basal ganglia drug effects, 193
conditioning paradigms, 103 disease, 521 cognitive enhancers, 8486, 120,
Discriminative neuronal activity, in Donoghue, J., 347 522
active avoidance learning, 451, Doob, L. W., 131 dissociations between implicit
453 Dopamine and explicit memory, 244
Dishabituation activity-dependent regulation of memory impairment by, 119
Aplysia, 3739, 38, 40 synthesis of, 46 120
C. elegans (Caenorhabditis elegans), in blocking, 303 pharmacologic treatment of
288 in brain stimulation reward, memory deficits, 520522
cellular analogue, in Aplysia, 39 574575 in research on memory effects,
defined, 223 functions of, 182 118119
habituation-dishabituation in neural computation, 444 state-dependent recall, 582583,
technique, infant visual olfactory tubercle and, 210 583t
recognition memory, 254 reinforcement and reward, 567 types of drugs affecting
parametric features of, 223224 568, 569, 577 memory, 118, 119t
Disinhibition in substantia nigra neurons, 303 See also Cholinesterase
basal ganglia movement release- in ventral tegmental neurons, inhibitors; specific drug
inhibition function, 191 192, 303 names
dementia, 109110 Dorsolateral nucleus of thalamus Dual coding model, imagery, 78,
Disparity detectors (DLM), in birdsong learning, 64, 7879
defined, 426 66 Dual memory theories. See Multiple-
plasticity in, 427 Dorsomedial (DM) nucleus of memory systems
Dissociation between implicit and thalamus, in birdsong learning, 65 Dual process model, conceptual
explicit memory, 244245 Dostrovsky, Jonathan, 529 combinations, 600
functional, 244245 Double dissociation Dual-process theory of habituation,
pharmacological, 244 in amnesia, 549, 673 224
population, 244 in maze learning, 415 Dual task performance, 13, 4849
theories of, 245246 in Parkinsons disease, 549 Dyslexia, 322323
Dissociative disorders Down, J. Langdon, 382, 587
functional amnesia related to, Down syndrome, 382383 E
2324 Alzheimers disease in, 383 Early experience and learning, 121
See also Double dissociation behavioral characteristics, 383 123
Disterhoft, J. F., 8 development in, 384 handling effects, 121122
Distractor paradigm, in short-term recall ability in, 264265 latent inhibition and, 123
serial recall, 611, 612 Dreaming neural substrates of, 123
Distributed practice effects, 115117 brain activation in, 621 Ease-of-learning judgment, 385
deficient-processing accounts, Freud on, 156, 618 Ebbinghaus, Hermann, 124127,
116117 Drislane, F. W., 322 125
Ebbinghauss view, 126 Drosophila melanogaster (fruit fly), on distributed practice effects,
encoding-variability accounts, 292295 115
117 aversive conditioning, 260 forgetting curve, 152
importance, 116 behavioral plasticity, 293 influence of, 126127
infant visual recognition behavioral screens for memory life of, 124
memory, 255 mutants, 293, 295 memory studies of, 125126,
massed repetitions vs., 115116 biochemistry of, 293294 379380, 412, 578
multiprocess accounts, 117 brain anatomy, 294295 on verbal learning, 655
study-phase retrieval accounts, central excitatory state (CES), Eccles, John, 127, 127130
116 260 education of, 127
See also Repetition and learning genetic variations, 137138, 293 electrical-chemical synaptic
DLM (dorsolateral nucleus of morphology and learning, 401 transmission controversy, 128
thalamus), in birdsong learning, 403 leadership of, 130
64, 66 neurogenetics of memory, 292 on mind-brain problem, 129
DM (dorsomedial nucleus of 295 130
thalamus), in birdsong learning, neuron physiology in, 294 scientific discoveries of, 127
65 olfactory memory formation, 129, 129
Dogs 293295 Eccles, R., 128
avoidance learning, 2 Drug abuse, conditioned stimuli in, Echoic memory, 49, 397398
fear of, 524 569 Ecological memory. See Natural
Dollard, J., 236, 327 Drugs and memory, 118120 settings, memory in
Domoic acid, 119 Alzheimers disease therapy, Ecphory, 606
Donaldson, Henry, 665 111, 521 Ecstasy (MDMA), 119
GENERAL INDEX 689
ECT (electroconvulsive therapy) and lateral, 203, 210 Konorskis implementation of,
memory loss, 132133 medial, 203 310
Education in parahippocampal circuits, Error-correction learning, 433, 565
special, 321 203 Errorless learning, 563
See also Knowledge; Prose projections to cortical areas, 213 Escape behavior
retention; School learning subcortical connections, 214 early handling of animals and,
Effect. See Law of effect (Thorndike) Environmental enrichment 121122
The Ego and the Id (Freud), 156 changes in nonneural elements, instrumental conditioning, 103
Egocentrism, Piaget on, 527538 405406 Tritonia, 292
Eidetic imagery, 130132 neurogenesis and, 471 Essentials of Behavior (Hull), 236
age effects, 131 synaptic change and, 405406 Estes, W. K., 220
cross-cultural studies and, 131 Ependymal cells, 181 Estrogen
as marker for brain damage, Ependymins, 555 birdsong learning and, 69
131 Epinephrine for cognitive deficits, 476477
photographic memory vs., 130, activity-dependent regulation of learning and, 614
131 synthesis of, 46 memory effects, 234
Electrical self-stimulation of brain. functions of, 182 Ethanol, memory effects, 118, 120
See Brain stimulation reward memory consolidation, 513 Ethical issues, Skinners
(BSR) memory effects, 232, 234 reinforcement principles and, 58
Electric shock, aversion therapy, 61 Episodic memory, 135137 The Ethics of Helping People
Electroconvulsive therapy and aging effects, 1112 (Skinner), 58
memory loss, 132133, 370 Alzheimers disease and, 111
Evaluative conditioning. See
Electroencephalography (EEG) animal cognition and, 9293
Conditioning, classical
infant visual recognition autobiographical memory linked
Event-related potentials (ERP)
memory, 254 to, 51
imitative learning in infants,
neuronal synchronization, 501 defining features, 136
256
sleep and, 618619, 619 dementia and, 111
infant visual recognition
Electronic synapse (gap junction), false, 145147
memory, 254
204, 206, 217 feeling of familiarity and, 136
orienting reflex habituation, 498
Ellis, N., 265 frontal lobes and, 157162
Everyday memory. See Natural
Emotion, mood, and memory, 133 functional imaging of, 306
settings, memory in; Semantic
135 Huntingtons disease and, 111
memory
Emotional memory meanings assigned to, 135
Evolution and learning, 137140
amygdala in, 545 natural setting and life
neural substrates of, 466468 brain evolution and, 139
experiences, 418
Emotional reactivity, early handling organic amnesia and, 2930 comparative cognition and,
and, 122, 123 semantic memory vs., 597, 598 138139
Emotions support from declarative genetic variation in learning,
amygdalar control of, 185 memory system, 105 137138
conditioned emotional response system of, 136 learning effects on, 139140
(CER), 123 task and system relationship, reproductive success and, 138
uncontrollable events, in 136137 Thorndikes views, 649
learned helplessness, 319 tasks, 135136 Evolution of the Brain and Intelligence
Encoding. See Coding processes temporally graded retrograde (Jerison), 649
Encoding specificity principle, 582 amnesia and, 370 Excitatory classical conditioning,
Endbrain (telencephalon), 191, 445 See also Autobiographical 7475
449 memory Excitatory modulation, in active
Endoplasmic reticulum Epistemology, genetic, 527 avoidance learning, 451452
rough, 205 EPSP. See Excitatory postsynaptic Excitatory postsynaptic potential
smooth, 205 potential (EPSP) (EPSP)
-Endorphin Equilibration model, 527 Eccless discovery of, 128129,
antagonists of, 234 Equipotentiality, Lashleys notion of, 129
memory effects, 233 318 Hermissenda, 278
Engle, R. W., 265, 266 Erickson, E. H., 53 Limax learning, 283
Engram. See Memory traces; Ericsson, K. A., 267 long-term depression, 335, 337
Memory traces, localization of ERK. See Extracellular receptor NMDA receptor-mediated, 67
Enkephalins, memory effects, 233 activated protein kinase (ERK) non-NMDA receptor-mediated,
Entorhinal cortex Error backpropagation algorithms, 176
Alzheimers disease and, 18 16 Excitatory potential, in Hulls
cortical inputs, 213 Error-correcting learning rules stimulus generalization principle,
hippocampal connections, 214 Hebbian, 511 360
690 GENERAL INDEX
Excitatory synapses rabbit nictitating membrane, Feeding behavior

characteristics of, 215, 216 458459 Aplysia operant conditioning, 35,
location on dendritic spines, use of, 519 35
217 Eye-movement response, optokinetic See also Food aversion and
Executive dysfunction (OKR), 659661 preference learning in
Alzheimers disease, 18 Eyewitness testimony humans; Foraging; Taste
dementia, 110 children, 73 aversion and preference
Exelon (rivastigmine), for false memory and, 146, 147 learning in animals
Alzheimers disease, 521 McGeochs research, 363 Feeling of knowing. See
Exemplar-memory model, 362 unreliability of, 417 Metacognition
Exercise, adult neurogenesis and, Eysenck, M., 133 Feldon, J., 123
471 Fiber tracts, of forebrain, 180
Fight or flight, 641
Expectation, in Tolmans view of F Filial imprinting
behaviorism, 57 Face-name mnemonic, 394
Experience. See Experts memories; Face recognition, infants, 257 characteristics, 247
Schemata Facts, learning to remember, 591 neural mechanisms, 249
Experts memories, 140143 False memories, 145147 reversibility, 248
factors mediating superior construction of, 559 See also Imprinting
performance, 142143 external factors in, 146 Fixation, spinal, 640
meaningful relations and hypnotic age regression and, Flashbulb memories, 419
performance, 141, 142 240241 Flies. See Diptera learning;
organization of memory and, hypnotic memory recovery and, Drosophila melanogaster (fruit fly)
251, 267 241242 Floccus, 660, 661
specificity of, 140142, 141 imagination and, 560 Flooding, in phobia treatment, 59,
Explicit memory. See Declarative implanting of, 560 60
memory internal processes in, 145146 Flux discrimination, visual, 663664
Exploration practical implications of, 147 fMRI. See Magnetic resonance
See Migration, navigation, and processes in origin of, 146147 imaging, functional (fMRI)
homing source monitoring, 629 Folia, of cerebellum, 195, 195
. See Foraging suggestion and, 419 Food aversion and preference
Extinction Fear, 461462, 512516 learning in humans, 147149
behavioral, 493 amygdala in fear conditioning, dietary specialists vs. generalists,
representing new learning, 330 166168, 184, 303, 304, 333, 147148
Extracellular receptor activated 334, 467468, 545546 exposure and conditioning
protein kinase (ERK) synaptic plasticity and fear effects, 148
Aplysia sensitization, 44 conditioning memory, 334 knowledge of dangerous food
Hermissenda, 278 auditory conditioning, 467 in, 148
long-term potentiation behavior therapy for, 5960 phobias and, 524
induction, 352, 353 brain circuits, 462 rejection or acceptance of food,
tyrosine hydroxylase classical conditioning, 461, 466 148
phosphorylation, 5 467, 669 social influences on food
Eyeblink classical conditioning conditioning vs. passive choices, 148149
aging and memory in animals, avoidance, 515 taste-aversion learning, 148
79, 8 fight or flight, 641 See also Taste aversion and
Alzheimers disease and, 9 freezing, 461462 preference learning in
blocking effect and, 302303, instrumental learning, 467468 animals
303 learning, 512516 Food caching/storing
brain regions and circuitry, 460 startle reflex, 463465, 465 Clarks nutcrackers and
cerebellar cortex and, 169 See also Passive (inhibitory) landmark navigation, 9091
cerebellar learning, 197198, avoidance; Phobias comparative methods of
198, 199 Feature extraction, 440, 441 learning, 139
cerebellum and, 78, 169, 302 Fechner, Gustav, 411 evolution of learning and, 139
303, 303, 415, 544, 571572 Feedback (FB) mechanism foraging problems, 151
classical, defined, 169 birdsong learning, 6567 scrub jays and episodic memory,
memory trace localization and, negative, in blocking, 303 9293
333 olfactory neural computations, Food foraging. See Foraging
rats, 459 447448 Foot-contraction, Hermissenda, 277,
Rescorla-Wagner model and, Feedforward mechanism 279
302, 303 cerebellum, 198 Foraging, 149151
Eyelid conditioning olfactory neural computations, bees, 261
human studies, 518 446447, 448 cryptic prey, 150
GENERAL INDEX 691
nectar-feeding, 151 anomalies of, learning General behavior theory. See

optimal theory of, 149150 disabilities and, 323 Learning theory
patches of food, 150151 conditional association learning, Generalization. See Discrimination
prey selection, 150 159160 and generalization
See also Food caching/storing episodic memory and, 157162 General Psychology in Terms of
Forced-choice recognition tests, 365 localization of functions, 161 Behavior (Smith and Guthrie), 219
Forebrain (prosencephalon) 162 Generativity, Eriksonian notion of,
gross anatomy, 180 orbitofrontal cortex neurons, 53
olfactory cortex in, 210 577 Genetic epistemology, 527
See also Basal forebrain in schizophrenia, 589590 Genetics
Forel, August, 605 working-with memory, 157158, avoidance learning, 3
Forgetting, 152154 158 Drosophila memory, 292295
cue-dependency theory, 153 See also Prefrontal cortex in evolution of learning, 137
154, 153t Fronto-temporal dementia, 111 138
defined, 373 Frost, Robert, 616 memory consolidation, 367
early childhood events, 52 Fruit fly. See Drosophila melanogaster mental retardation and, 382
history of research on, 268269 (fruit fly) See also Protein synthesis
infant recall of forgotten events, Frustration and Aggression (Dollard), Genetic substrates of memory, 165
255 236 175
interference and, 145, 152153, Fugue, psychogenic, 23 amygdala, 166168
152t, 255, 268272, 656, 670 Functional amnesia. See Amnesia, cerebellum, 169171
671 functional hippocampus, 171175
Functional imaging. See Magnetic Geniculate nucleus, lateral, 322
McGeochs theory, 363
resonance imaging, functional Geniculate nucleus, medial, 322, 453
memory consolidation and, 369
(fMRI); Neuroimaging Geometric designs, memory testing
normal rates, 154
Functionalism, 655656 and, 380
trace decay, 152
Gerlach, Joseph von, Sr., 557
Forgetting curve, Ebbinghauss, 152
Gestalt psychology, Piaget on, 527
Forman, R. R., 259 G Gibbon, John, 331
Fornix, 204 GABA (gamma-aminobutyric acid) Gill withdrawal
c-fos, 6, 166 benzodiazepine effects, 118, 120 second-order conditioning, 98
Fractional anticipatory goal response cerebello-olivary GABAergic See also Siphon-gill withdrawal
(rG-sG), 235, 237 projection, in eyeblink reflex, in Aplysia
Franz, Shepherd, 318 conditioning, 302303 Ginkgo biloba
Free-association technique, 156 as inhibitory neurotransmitter, as cognitive enhancer, 85, 522
Free-operant procedures 185, 201 with ginseng, 8586
avoidance learning, 12 in memory, 476 Ginseng
behavior modification, 494 as primary synaptic as cognitive enhancer, 85
Free recall, modality effects in, 398, neurotransmitter, 215 with ginkgo biloba, 8586
399 Galaburda, A. M., 322 Giray, E. F., 131
Free recall test, 364 Galanthamine (Reminyl), for Glanzer, M., 49
Freezing behavior, 461462 Alzheimers disease, 521 Glenberg, A. M., 116
Frequency of occurrence estimation, Gallistel, Randy, 331 Glial cells
in frontal lobes damage, 159 Galton, Francis, 96, 418 environmental enrichment and,
Freud, Sigmund, 154157, 155 Gamma-aminobutyric acid (GABA). 407
on dreaming, 156, 618 See GABA (gamma-aminobutyric neurons vs., 204
education and early career, acid) overview, 181
154155 Gamma-secretase in familial removal of excess glutamate,
infantile amnesia explanations, Alzheimers disease, 20, 22 217
24, 26, 418 Gantt, Horsley, 517, 519 synaptic functions, 216, 217
psychoanalysis originated by, Gap junction (electronic synapse), Globus pallidus
155156 204, 206, 217 association with substantia nigra
studies of psychological Gaps, memory, filling in, 507, 509 and subthalamic nucleus,
processes, 156157 Garcia, J., 83 191192, 192
Frontal lobe damage Gardner, B. T., 315 basal ganglia connections, 191,
autobiographical memory loss Gardner, R. A., 315 191
in, 5354 Gartman, L. F., 117 in forebrain, 180
memory tasks, 158161 Geiselman, Edward, 417 neuron structure, 181
Frontal lobes, 157162 Gemmules, of dendrites, 205 Glucocorticoid receptors
anatomy and functions, 200, Gender. See Sex differences in in glucocorticoid memory
200 learning; Sexual imprinting effects, 232
692 GENERAL INDEX
Glucocorticoid receptors (continued) Grammar parametric features, 223224

hippocampal, 389 acquisition in children, 312314 procedural learning, 547
in hippocampus of nonhandled artificial, 549 spinal cord, 225
animals, 123 c-command relation, 312, 313 startle reflex, 225226
Glucocorticoids universal, theory of, 312314 Tritonia, 291292
in memory consolidation, 513 Grammatical knowledge, 311 vertebrates, 223226, 640
memory effects, 232233, 234 Granule cells, cerebellum, 195, 197 See also Sensitization
stress and, 642 Grasshoppers, 259 Habituation-dishabituation
Glucose, in memory consolidation, Gray matter, 204 technique, infant visual
515 Group learning, 591 recognition memory, 254255
Glutamate Group recall Haeckel, Ernst, 605
back-propagating dendritic collaborative inhibition, 622 Hammer, M., 276
action potentials and, 352 623 Handling, early. See Early
excess, removal by glia, 217 group member characteristics, experience and learning
as excitatory neurotransmitter, 623 Hannon, B., 265
185, 201 nominal vs. collaborative
Hardin, T. S., 134
groups, 621623, 622
in memory, 475476 Harlow, Harry F., 227, 227228, 578
recognition tests, 623
transmission via excitatory Hartley, David, 618
Groves, P. M., 224
synapses, 216 Hasselmo, Michael, 444
Growth factors, retinal lesions and,
Glutamate receptors, 175178 Head injury, 228229
421
characterization of, 175178 memory and learning disorders
Grundzge der Psychologie
fear-potentiated startle, 464 after, 228229
(Ebbinghaus), 124
genes in, 176177 Guanosine nucleotides, second posttraumatic vs. retrograde
G-protein-coupled, 178 messenger systems, 594 amnesia after, 228
imprinting and, 250 Guanylyl cyclase, 596 visual-spatial task deficits, 79
ionotropic, 176 Guessing Heart rate, classical conditioning
ion permeation, 177 in cued recall test, 364 and, 454455
long-term potentiation effects, forced-choice recognition test, Hebb, Donald, 229232, 230
174 365 on brain evolution, 649
metabotropic (mGluR), 176, Guide to anatomy of the brain. See early career, 230
178, 337, 338, 388 Brain anatomy on habit growth model
NMDA receptors, 177178 Guthrie, Edwin R., 218220, 219 limitations, 361
NMDA subtype, in birdsong honors, 218 influence on Olds, 485
learning, 67 on laws of association, 219220 Karl Lashley as mentor, 230
non-NMDA receptors, 176177 learning theory of, 219220, 231
Glutamatergic system, stress and, 326 later career of, 231232
642 learning algorithm of, 1415
Glycine, as primary synaptic localization of memory trace,
neurotransmitter, 215
H 332
H.M. (amnesia patient), 28, 30, 106,
Glycine binding site, NMDA neural model of idea and, 231
369, 471, 547
receptor, 177 on stress and memory, 641
Haber, R. B., 130
Goal-oriented behavior. See Adaptive on synaptic change, 405
Haber, R. N., 130
behavior; Motivation; Reward Hebbian processes
Habit growth, law of, 360
Golgi, Camillo, 215 Habits, basal ganglia involvement in, learning rules, 511
Golgi apparatus, neuronal soma, 205 193 neuroplasticity in, 435
Gonadal hormones Habit strength, Hulls measure of, in parallel distributed
birdsong learning and, 6769 359360 processing, 509510
learning and, 615 Habituation, 223226 in pattern completion, 439
memory effects, 234 Aplysia, 37, 38, 39 principle of locality, 433
See also Sex differences in C. elegans (Caenorhabditis elegans), somatosensory plasticity and,
learning 288290, 289 432
Gormezano, I., 458 cellular analogue, in Aplysia, 39 Hebbian synapse
Gorry, J. D., 188 defined, 223 long-term potentiation and, 340
G-protein-coupled receptors, dual-process theory of, 224 memory trace formation and,
glutamate receptors, 178 long-term, 225226 332
G-proteins, second messenger measurement of learning, 548 origin of theory of, 230, 231
systems, 594 549 Heinroth, Oskar, 356
Graceful degradation of memory, mechanisms of, 224225 Heisenberg, M., 294
510 neural substrates, 549 Helix (snail), 283284
Gradient descent algorithm, 16 orienting reflex, 497499 Helplessness, learned, 319320, 642
GENERAL INDEX 693
Hemispheres of brain. See Left eyeblink classical conditioning Hippocampus, Space, and Memory
hemisphere of brain and, 89 (Olton et al.), 489
Hereditary memory, 606 in fear behavior, 462 The Hippocampus as a Cognitive Map
Hermissenda, 277280 fornix, as connection to (OKeefe and Nadel), 414
associative learning in, 277280, hypothalamus, 204 Hirsh, R., 414
402 functions, 471472 Histamine, memory modulation, 476
cellular and synaptic plasticity, as gateway to memory, 470, Histologie du systme nerveux de
278, 279 472 lhomme et des vertbrs (Ramn y
memory consolidation general features, 202 Cajal), 558
mechanisms, 278, 280 genetic substrates of memory, Historia animalium (Aristotle), 47
Pavlovian conditioning of, 277 171175 Historicocultural psychology, 55, 357
278, 279 gradual memory consolidation Hitch, G. J., 264, 673675
Hertel, P. T., 134 in, 371 Hoff, M., 15
Heterosynaptic neurons, modulation inhibitory avoidance learning, Hoffman, Dustin, 588
of, in Aplysia, 43 453 Holst, Erich von, 355
Hewes, A. K., 264 learned bradycardia, 456457 Homing. See Migration, navigation,
The Higher Cortical Functions in Man long-term depression in, 335, and homing
(Luria), 357 337, 388 Homing pigeons, 392
Higher-order classical conditioning, long-term potentiation in, 172 Homology, defined, 258
98100 173, 388 Homophony, 606
experimental models, 98 LTP enhancement and learning, Honeybee. See Bees
theoretical models, 98, 99 174, 347 Hopfield, John, 444445
High vocal center (HVc) in birdsong memory consolidation and, Hormones, 232234
learning, 64, 66, 66, 68, 68, 69 367368, 369, 371, 372 birdsong learning and, 6769
Hilgard, E. R., 2 metaplasticity, 388 gonadal, 6769, 234, 615
Hindbrain (rhombencephalon), 180, multiple memory systems and, interaction among hormonal
187 414415, 534 systems, 234
Hintzman, D. L., 237 memory effects, 232234
neocortical interaction with,
Hippocampal amnesia. See Amnesia, neonatal disturbance of,
during memory consolidation,
organic behavioral effects, 121
367368, 371
Hippocampal damage stress hormones, 513, 514
neural computation, 439441,
blocking eliminated by, 304 See also specific hormone
440
deficient memory for targets or Horridge, G. A., 259
neurogenesis, 469, 470, 471
facts, 159 5-HT. See Serotonin
473
organic amnesia related to, 29 Hull, Clark L., 234237, 235
neuroimaging, in memory
recent memory impairment in, early career and honors, 234
consolidation, 371
333 235
neuron structure, 181
temporally-graded amnesia in, habit strength measure of, 359
nictitating membrane classical
368, 369, 370 360
conditioning, 458
Hippocampus, 171175, 202204 influence on Spence, 637
NMDA receptor function and
adult neurogenesis, 469, 470, learning theory of, 235237,
learning, 173
471473 327328
Alzheimers disease impairment perforant pathway, 203, 214 neobehaviorism, 57
of, 9, 10, 18 perirhinal cortex connections, on Pavlovian conditioned
anatomy, 202204 214 response, 519, 637
associative memory and, 439 place cells, 529532, 530 stimulus generalization principle
441 recent memory and, 333 of, 360
birds, size and role of, 139 in reinforcement, 569 Thorndikes influence on, 533,
blocking and, 304 in schizophrenia, 590 637
aCaMKII, LTP, and learning, signaling pathways and synaptic Hummingbirds, song learning in, 64
172173 plasticity, 173 Hunt, J. McV., 239
CA1 neurons, 203204 sleep and, 620621 Hunter, Walter S., 237239, 238,
CA3 neurons, 203 spatial learning and, 9, 472, 332
corticosteroid receptors, 389 489, 529, 531 Hunter-McCrary law, 610
declarative memory and, 107 transcription, translation, and Hunting behavior, predatory, 461
disruption of LTP and learning, long-term memory, 173174 Huntingtons disease
174175 trisynaptic organization, 203 basal ganglia involvement, 191
early experience and learning, voles, 139 dorsal striatum abnormality in,
123 wiring, 202204 192
entorhinal cortex connections, working memory and, 489491 episodic memory and, 111
214 See also Hippocampal damage memory impairment, 520
694 GENERAL INDEX
Huperzine A, 85, 522 Imagination, in false memory, 146, Indirect memory tests, in aging, 11
HVc (high vocal center) in birdsong 560 Individual differences in learning
learning, 64, 66, 66, 68, 68, 69 Imaging. See Neuroimaging and memory, 251253
Hyman, I. E., 146 IMHV (intermediate and medial avoidance learning, 23
Hymenoptera learning, 260261 part of hyperstriatum) ventrale, in capacity differences, 252
See also Bees imprinting, 249, 250 collective memory vs. individual
Hyperacusis, in Williams syndrome, Imitative learning, infants, 255256, memory, 8788
383 256, 257 knowledge differences, 251252
Hypermnesia, hypnotic, 240 Immediate early genes learning strategy differences,
Hyperphagia, in Prader-Willi Aplysia sensitization, 44 252
syndrome, 383 expression in amygdala, 166 retrieval speed differences, 252
Hyperstriatum ventrale, stress and, 643 253
intermediate and medial part of Implanted memories. See False Infancy, memory in, 254257
(IMHV), in imprinting, 249, 250 memories assessment methods, 7273
Hypnosis, 240242 Implicit memory, 243246 conditioning techniques for
age regression technique, 240 activation view, 245 evaluating, 255
241 amnesia and, 24 deferred imitation test of, 255
agnosia and, 240 attention and, 5051 early and rapid formation of,
amnesia after, 24, 240 component-processes approach 257
Freuds experiments with, 155 to, 246 intelligence quotient and, 257
156 declarative memory vs., 105, multiple memory systems in,
hypermnesia and, 240 333 257
memory recovery and, 241242 defined, 415, 547 parent-child conversations about
Hypoglossal nucleus, dissociation from explicit past, 27
tracheosyringeal portion of memory, 244245 visual recognition memory
(nXIIts), in birdsong learning, 64, experimental paradigm, 243 studies, 254255
65, 66 head injury and, 229 See also Children, development
Hypothalamic-pituitary-adrenal measures of, 365366 of memory in
(HPA) system, 122, 642 mood impairment and, 134 Infantile amnesia. See Amnesia,
Hypothalamic stimulation, 573574 multiple memory systems infantile
Hypothalamus theory, 245246 Infants
fear conditioning and memory, neuroimaging and, 246 early neonatal handling and
333 organic amnesia and, 30 learning, 121123
in forebrain, 180 repetition priming and, 243 object concept development,
fornix, as connection to 244 479481
hippocampus, 204 retention latency, 513 sexuality of, Freuds theory of,
paraventricular nucleus of, 181 taxonomy, 548 156
preoptic, amygdalar projections, tests of, in frontal lobe damage, See also Infancy, memory in
185, 186 160161 Information. See Knowledge
theoretical accounts, 245246 The Information Available in Brief
I transfer-appropriate processing Visual Presentations (Sperling), 607
Ibo (Nigeria), 131 theory, 246 Information storage, temporary
Iconic memory, 49, 607 Imprinting, 247250 mental. See Working memory
Idea, Hebbs neural model of, 231 auditory, 248 Infrared illumination-differential-
Ideational apraxia, 604 cerebral asymmetry and, 250 interference contrast (IR-DIC)
Idebenone, for memory conditions for studying, 247 optics, 343
enhancement, 522 248 Inhelder, Brbel, 529
Identity disorder, dissociative, 23 filial, 247 Inheritance. See Genetics
Idiot savant. See Savant syndrome learning and, 248 Inhibition
Image, mental rotation and scanning localization of neural changes, in Aplysia, 3940
of, 78 249, 249250 learning and memory, 656
Imagery, 7779 Lorenzs work on, 356 reproductive, in interference
dual coding model, 78, 7879 neural mechanisms of filial, 249 and forgetting, 269
eidetic, 130132 neuronal changes, 250 retroactive, 656
individual memory differences, predispositions in, 248249 Inhibition, conditioned, 75, 113, 310
252 reversibility, 248 Inhibition, proactive, 153, 159, 656,
memory coding process, 7779 sensitive periods, 248 670
in oral traditions, 496 sexual, 248 Inhibition, retroactive, in forgetting,
Simonidess method of loci and, Improvisation, in observational 153, 363
77, 78, 393, 396 learning, 483 Inhibitory avoidance. See Passive
Imagery effect, 78 Incentive motivation, 327, 328 (inhibitory) avoidance
GENERAL INDEX 695
Inhibitory-bursting neurons, Limax, output interference and part-list Tritonia, 291292

281282 cuing, 271 IQ tests. See Intelligence tests
Inhibitory classical conditioning, 75 retrieval and, 271272 Irion, Arthur L., 363
Inhibitory postsynaptic potential retroactive, 268, 269 Irwin, J. M., 269
(IPSP) theory, 152153, 152t, 269 Ischemia, dementia associated with,
Eccless discovery of, 128129, transfer in, 268 110
129 working memory and, 670671 Isolation, birdsong learning effects,
Hermissenda, 278 See also Forgetting 61, 62, 63
Inhibitory synapses Intermediate and medial part of Ito, M., 129, 169
characteristics of, 216, 216 hyperstriatum ventrale (IMHV), in
damping of excitation, 217 imprinting, 249, 250
Inositol polyphosphates, 595 Interneurons J
Inositol triphosphate (IP3) receptors, long-term potentiation on, 342 Jackson, Hughlings, 585
337 Jacobs, L. F., 139
olfactory cortex, 212
Jacobsen, C. F., 536
In Search of the Engram (Lashley), Tritonia Dorsal Swim
James, William, 297299, 298
413 Interneurons, 292
Ebbinghaus and, 125
Insect learning, 258262 Interocular disparities, defined, 426
education, 297
mimicry, 139 Interpretation and memorization.
Hunter and, 238
new developments in learning, See Coding processes
influence on psychology, 655
261262 Interpretation of Dreams (Freud), 156
on memories as beliefs, 299
spatial, 631632 Interstimulus intervals (ISI)
new psychology views, 297
value of studying, 258259 C. elegans (Caenorhabditis elegans),
298
See also Invertebrate learning; 288, 289, 290
opinion of Thorndikes work,
specific insect classical conditioning, 74, 76
647
Institute of Human Relations, 236 discrimination procedure, 76
parliamentary theory of
Instrumental behavior. See Operant dual ISI, 76
learning, 298299
behavior eyeblink classical conditioning, 7
on primary memory, 380381,
Instrumental conditioning. See ISI shift procedure, 76
672
Conditioning, instrumental Intertrial interval (ITI), avoidance
Thorndike and, 298299
(operant) learning, 1
Janet, Pierre, 24, 218, 585
Integrative Activity of the Brain Interval timing, in stimulus-response
Janowsky, J. S., 160
(Konorski), 310 (S-R) learning, 89, 89, 90
Jarrold, C., 264
Integrative Behavioral and Physiological Interview, cognitive, 417
Jay family (Corvidae)
Science (journal), 519520 Introduction lepistmologie gntique food cache behavior, 139
Intellectual disability. See Mental (Piaget), 527 scrub jays, and episodic
retardation Invertebrate learning, 273295 memory, 9293
Intelligence and memory, 263267 alcoholism model, 262 Jenkins, J. G., 152
encoding and, 264, 264265 Aplysia, 3336, 401 Jennings, Herbert S., 317
information processing role, bees, 261, 262, 273276, 392, Jerison, Harry, 649
265266 401, 632 Johnson, M. K., 146, 623
long-term knowledge and skills blocking in, 261 Johnson, N. F., 117
in, 266267 C. elegans (Caenorhabditis elegans), Johnston, John Black, 317
working memory and, 264, 264, 287291 Jon (amnesia patient), 30
265267, 266 comparative analysis in Jones, Mary Cover, 667
Intelligence tests Africanized honeybees, 262 Josselyn, S. A., 168
changes in, 267 Drosophila melanogaster, 292295, Jost, A., 115
infant memory and childhood 401402 Judgment of learning, 385
IQ, 257 Hermissenda, 277280, 402 Just, M., 252
mental retardation and, 381 insects, 258262
382 Limax, 281285
Intentional behavior. See Operant matching-to-sample learning, K
behavior 262 Kacelnik, A., 150
Interference and forgetting, 268 memory trace localization and, Kainate (KA) receptors, 176177,
272 333334 290
dynamics of, 269271, 270 morphological basis of, 401403 See also Non-NMDA receptors
in false memory, 145 new developments in learning, Kamin, L. J., blocking effect, 76,
history of research on, 268269 261262 261, 301304, 360
infant visual recognition preexposure to conditioned See also Blocking
memory, 255 stimulus, 261262 Kandel, Eric, 33
McGeoch and Underwood on, risk sensitivity and choice Kanzi (bonobo), 316, 316317
656 behavior, 262 Katz, B., 127
696 GENERAL INDEX
Keller, Fred S., 616 The Language and Thought of the Child behaviorist vs. cognitivist, 533
Keyword mnemonic, 394, 395 (Piaget), 526, 527, 528 534
Kinases. See Protein kinases Language learning, in humans, 311 birdsong learning, 6169
King Solomons Ring (Lorenz), 356 314 blocking of long-term
Kleitman, Nathaniel, 618 early and rapid memory potentiation and, 347348
Klima, Edward S., 399 formation in infants, 257 brain maps, 485486
Klver-Bucy syndrome, 185 grammar learning in children, cellular analogues, in Aplysia, 39,
Knowing, feeling of. See 312314 42
Metacognition Lurias contribution, 357 cerebellum in, 197199, 199
Knowledge, 306308 pragmatics and, 311 of concepts and categories,
acquisition in school, 592 recall of early memories and, 26 9597
categorical, 598599 specific language impairment early experience and, 121123
category-specific loss of, 307 and, 322323 errorless, 563
308 universal grammar theory, 312 evolution and, 137140
classification of memory 314, 313 food aversion and preferences,
systems, 306 Language learning, in nonhuman 147149
individual differences in, 251 primates, 314317 Hebbian rules, 511
252 bonobo studies, 316, 316317 imprinting and, 248
material-specific memory chimpanzee language-learning individual differences, 251253
deficits and, 306308 projects, 315 insects, 258262
in memory and intelligence, early studies, 315 invertebrates, 273295
266267 Project LANA, 315, 316 language, in humans, 311314
memory development in Lashley, Karl, 317, 317319 language, in nonhuman
children and, 72, 72 brightness discrimination primates, 314317
metaplasticity in information research, 663, 664 latent, 653
processing, 387388 education, 317318 long-term potentiation in, 388
modality-specific loss of, 306 localization of memory trace, metamemory judgments in, 385
307 318319, 332333, 413 morphological basis of,
See also Secondary memory as mentor of Donald Hebb, invertebrates, 404408
(long-term memory) 230231 motor skills, 409410
Knowledge representation, category- Latent inhibition in navigation, 391392
specific, 307308
in classical conditioning, 7576 observational, 482484
Knowledge theory. See Learning
early handling and, 123 in orientation, 390
theory
invertebrate learning, 261 place vs. response, 533535
Koelling, R. A., 83
Latent learning, 653 procedural, in animals, 544546
Koffka, Kurt, 652
Lateral olfactory tract (LOT), 210 procedural, in humans, 547549
Konorski, Jerzy, 309, 309310
212, 212 programmed instruction, 617
Korsakoffs syndrome, 29
Law of contrast, 219 receptive field plasticity and,
Krebs, J. R., 150
Law of disuse (Thorndike), 268, 269 422, 424, 424425
Krehl, Ludolph, 605
Law of dynamic polarity (Ramn y reinforcement or reward in,
Krehl, Maria, 605
Cajal), 557 566577
Kuffler, S., 127
Law of effect (Thorndike), 235, 326, repetition and, 578580
Kyllonen, P. C., 264
533, 564, 637, 648649 school setting, 590593
Law of exercise (Thorndike), 648 self-monitoring of, 385386
L Law of habit growth, 360 self-paced, control during, 386
Labyrinthectomy, unilateral, 660 Law of regression (Ribots), 30 sex differences, 613615
LANA Project, 315, 316 Learned helplessness, 319320, 642 sometimes opponent process,
Land, Edwin, on color perception, Learned response. See Conditioned 625626
442 response (CR); Conditioning, spatial, 392, 633636
Landmarks classical; Conditioning, supervised vs. unsupervised,
Clarks nutcrackers and instrumental (operant) 433434
landmark navigation, 9091 Learning taste aversion and preferences,
mental scanning of, 78 algorithms, 1417 645647
Language amygdalar gene expression taxonomy, 433434, 435
comprehension, imagery in, 79 effects, 166168, 167 See also Associative learning;
damage to speech areas, loss of Aplysia, underlying processes of, Learning theory; specific type
memory after, 333 3741 of learning
spatial learning and, 635 approaches to, 432435 Learning curves, Rescorla-Wagner
Williams syndrome and, 383, avoidance learning, 13, 451 model of Pavlovian conditioning,
384 454, 512516 329, 329
GENERAL INDEX 697
Learning disabilities, 320323 orienting tasks, 8081 Loeb, Jacques, 665

behavioral perspective, 320321 recall and processing, 80 Loewi, O., 128
defined, 320 Levine, S., 121, 122 Loftus, Elizabeth F., 146, 559
neurobiological correlates, 321 Levodopa, basal ganglia effects, 193 Loftus, Geoffrey R., 608
322 Lewy bodies Long-delay learning, taste aversion,
specific language impairment in dementia, 111 645
and dyslexia, 322323 in Parkinsons disease, 110 Long-term depression (LTD), 335
Learning mazes. See Maze learning Lexicon, learning of, 313314 338
Learning strategies Lexigrams, chimpanzee language- adult visual cortex, 420
individual differences, 252 learning, 315, 316 AMPA receptor phosphorylation
metacognition and, 592 Life events. See Autobiographical and inactivation in, 338
Learning theory, 325331 memory; Episodic memory calcium role in, 336337
configural approach, 330 Limax, 281285 cerebellar, 170, 335, 336, 436
context of conditioning and associative learning in, 281285 437
extinction, 330331 central olfactory centers and, conversion to long-term
current status, 329331 281282 potentiation, 337
Guthries contribution, 219220, imaging of odor memories, 285 functional roles of, 338
326 model of odor memory gene knockout mouse models,
history of, 325328, 590591 formation, 283, 284 170
Hulls contribution, 235237, modulation of procerebral lobe hippocampal, 335, 337
326327 dynamics, 283 hippocampal metaplasticity and,
James parliamentary theory, odor-memory storage in 388, 389
298299 procerebral lobe, 282 induction and observation of,
Kamins blocking effect and, procerebral lobe inputs and 335
301 outputs, 283285 metabotropic glutamate
Lorenzs influence, 356 Limbic system receptors in, 337
mathematical, 359362 avoidance learning and, 451 in motor learning, 430
models of learning, 329, 329 453, 452
neocortical, 335
330 cholinergic input and output,
nitric oxide and cGMP in, 337
Pavlovian concepts and, 326 190
338
Rescorla-Wagner model, 329, connection to olfactory cortex,
protein kinases and
329330 210
phosphatases in, 338
Skinners contribution, 327 hippocampal connection to
protein synthesis in, 338
temporal model, 331 mammillary bodies, 204
receptors involved in, 336
Thorndikes contribution, 326, occipitotemporal pathway and,
signal transduction underlying,
648 543
335338
Tolmans contribution, 326 postsynaptic muscarinic and
subtypes, 335
Watsons contribution, 326 nicotinic receptors, 189
Leask, J., 131 in reinforcement, 569 Long-term facilitation, in Aplysia,
Least mean square (LMS) algorithm, Line coding, 500 367
1516 Linguistic development. See Long-term habituation. See
Leaton, R. N., 225 Language learning, in humans; Habituation
LeDoux, J., 347 Language learning, in nonhuman Long-term memory
Left hemisphere of brain primates See Knowledge
anomalies of, in learning Link mnemonics, 394 See Secondary memory (long-
disabilities, 323 Lipofuscin granules, in neurons, 205 term memory)
modality-specific agnosia and, Lip reading, 398, 399 Long-term potentiation (LTP), 339
307 Livingstone, M. S., 322 354
multiple knowledge systems in, lMAN (lateral part of magnocellular abnormal, in cortical regions of
307 nucleus of anterior neostriatum), memory storage, 368
Leg-position learning, orthoptera, in birdsong learning, 64, 66 adult visual cortex, 420
259 LMS (least mean square) algorithm, amygdala, 342345, 467
Lehrman, Daniel S., 356 1516 back-propagating action
Letter span, memory testing and, Local field potential oscillation, potentials in, 351352
380 Limax, 281, 283, 285 behavioral roles, 346348
Levels of processing, 8081 Locality principle, 433 conversion to long-term
dissociation of implicit and Localization of memory traces. See depression, 337
explicit memory, 244 Memory traces, localization of cooperativity and associativity,
incidental vs. intentional Loci, method of, 393, 395 340342
processing, 81 Lockhart, R. S., 80, 135, 136 defined, 339
materials to be remembered, 81 Locusts, 259 dendritic spine changes in, 350
698 GENERAL INDEX
Long-term potentiation (LTP) Magnetic resonance imaging (MRI), radial maze and working
(continued) 323 memory, 488490, 670
disruption of, hippocampal Magnetic resonance imaging, win-shift tasks, 415
learning and, 174175 functional (fMRI), 306, 370371, win-stay tasks, 415
expression of, 341 473, 474 McAdams, D. P., 53
glutamate and memory, 475 Magnitude, relative, 599 McDermott, Kathleen, 559
476 Magnocellular nucleus McEchron, M. D., 8
hippocampal metaplasticity and, anterior neostriatum, in McGaugh, James L., 546
388 birdsong learning, 64, 66, 67 McGeoch, John A., 362363, 363
hippocampus, 172175 lateral geniculate nucleus, discrediting of Thorndikes law
induction threshold modulation, dyslexia and, 322 of disuse, 269
340341 Mammals. See Animals; Primates; theory of forgetting, 269, 363
interneurons, 342 Vertebrates; specific kinds verbal learning research, 362
maintenance, 349351 Mammillary body, hippocampal Meador, D. M., 265
manipulation of, learning connection to, 204 Meaney, M., 123
enhanced by, 174 Mammillothalamic tract, 204, 452 Meaningful material and
memory effects of, 347348 Mandler, George, 12 relationships
motor learning, 430 experts memory performance,
Mansui, I., 354
as neuronal model of learning 141, 142
MAP kinase. See Mitogen-activated
and memory, 340, 341 in oral traditions, 496
protein (MAP) kinase
non-associative, 342 See also Mnemonic devices;
Markov chain, 361
overview, 340342 Schemata
Marquis, D. G., 2
place field development and, Measurement of memory, 364366
Marr, D.
531 explicit memory tests, 29
postsynaptic calcium increase in, on aperture problem, 442
cerebellar theory of, 197 in frontal lobe damage, 158161
352 implicit memory measures, 365
postynaptic enhancement of, on codon formation, 440
366
353354 on cortical processing speed,
442 indirect memory tests, in aging,
presynaptic expression of, 352 11
353 on neural computation in
infants, 7273
presynaptic mechanisms of hippocampus, 439
prospective memory measures,
maintenance, 349350 Mass action, Lashleys notion of, 318
365
protein kinases in, 352, 353 Massed repetition, distributed
recall measures, 158, 364, 377
protein phosphatase regulation practice effect vs., 115117, 126
379, 378
and, 353354 Massero, Dominic W., 608
recognition memory measures,
receptor property changes in, Matching-to-sample (MTS) concept
365, 373, 373
350 learning
relationship among measures,
signal transduction mechanisms bees, 262
366
and early events, 351354 pigeons, 92 working memory tests, and
stress and, 389390 Material-specific memory disorders, intelligence, 265266, 266
synaptic efficacy in conventional knowledge systems and, 306308 The Mechanisms of Perception (Piaget),
learning vs., 346347 Mathematical learning theory, 359 528
synaptogenesis and, 407 362 Medial temporal lobe. See Temporal
See also Synaptic plasticity connectionism, 362 lobe, medial
Lorenz, Konrad, 355, 355366 Hull and habit strength, 359 Medin, Douglas L., 362
Lotze, Hermann, 411 360 Medulla oblongata, 180, 181
LTD. See Long-term depression law of habit growth, 360 Melton, A. W., 115, 269, 363
(LTD) probability matching, 360361, Memantine (Akatinol), 521
LTP. See Long-term potentiation 361 Memory
(LTP) statistical learning theory, 361 aging and memory in animals,
Lubow, R. E., 123 362 710
Lures, in forced-choice recognition See also Algorithms, learning; aging and memory in humans,
tests, 365 Neural computation 1013
Luria, A. R., 357, 357358, 396 Mathews, A., 133 Aristotelian theory of, 4647
Lysosomes, in neurons, 205 Matire et mmoire (Bergson), 585 attention and, 4851
Matsuzawa, T., 315, 316 autobiographical, 5154
M Maze learning characteristics, 507508
Mackintosh, N. J., 524 double dissociation in, 415 classification, 118119, 306
Magnesium ions, 177, 337 Morris water maze, 9, 490, 492, collective, 8688
Magnetic fields, in migratory 545, 670 computer file metaphor, 507
orientation, 390391 plus-maze tasks, 533534 508
GENERAL INDEX 699
declarative, 105107 See also Forgetting; Memory rehabilitation of, 561563

dj vu phenomenon, 108109 headings below; Primary residual, after head injury, 228
development of, in children, memory (short-term in schizophrenia, 588589
7174 memory); Secondary memory See also Amnesia headings;
dissociation between implicit (long-term memory); specific Dementia; Forgetting
and explicit, 244245 aspects and types Memory drum, 412
Memory and Intelligence (Piaget et al), Memory-enhancing drugs. See
distributed repetitions and,
528 Cognitive enhancers; Drugs and
115117
Memory consolidation, 366372 memory
drugs and memory, 118120,
ACTH for, 476 Memory impairment
520522
in animals, 620 See Memory disorders/
Ebbinghauss studies, 125126 blocking of LTP and prevention impairment
echoic, 49, 397398 of, 347348 Memory loss. See Forgetting;
emotional, 133135, 466468 coding vs., 580581 Interference and forgetting
episodic, 135137 CREB and, 366367 Memory processing. See Coding
experts memories, 140143 Drosophila memory phases, 293 processes
genetic substrates, 165175 editing memory during recall, Memory recovery
hereditary, 606 368 hypnotic, 241242
elaboration in, 50 therapeutic, 24, 663
hormones and, 232234
glucocorticoids in, 232233 Memory retention interval. See
hypnosis and, 24, 240242
gradual, 371, 371372 Forgetting; Memory search;
implicit, 243246 Hermissenda, 278, 280 Memory span
individual differences, 251253 hierarchical storage, 634 Memory retrieval. See Memory
in infancy, 254257 hippocampal/neocortical search; Recall; Retrieval processes
intelligence and, 263267 interactions in, 367368 in memory
James on, 299 infant visual recognition Memory search, 373379
measurement of, 364366 memory, 254255 direct-access in item
modal model of, 49 interference and forgetting information retrieval, 374
theory, 268 375, 376
mood and, 133135
mechanisms of memory long-term memory retrieval,
morphologic basis of, 401408 formation and, 334 376377
Mller on, 412413 molecular and cellular order retrieval and short-term
multiple memory systems, 413 processes, 366368, 555556 memory recall, 377379, 378
416 Mller and Pilzecker on, 412 recognition time and short-term
natural settings, 417419 413 memory, 374, 374
organization of, 8283 neuroimaging of, 370371 retrieval in short- and long-term
parallel distributed processing in observational learning, 482, memory, 373374
models, 507511 483 search processes in recovery of
prefrontal cortex and, 535538 passive avoidance and, 513515 order information, 377
prospective, 13, 229, 365366 prolonged process of short-term memory in relation
reorganization, 369372 to other abilities, 379
protein synthesis in long-term
protein synthesis in, 366, 555 time course of short-term
memory, in vertebrates, 553
556 memory retrieval, 376, 377
556
retention latency, 512513 underlying retrieval
recognition memory, 399, 543,
sleep and, 618621 mechanisms, 374, 375
623
stress hormones in, 513, 514 See also Retrieval processes in
reconstructive, 558561 temporal lobe in, 369370 memory
reference, 489, 490 temporally graded retrograde Memory span, 379381
retrieval processes, 580584 amnesia and, 370 defined, 379
in savant syndrome, 587 temporal patterns, 515 digit span test, 267, 364, 380
schizophrenia and, 588590 Memory disorders/impairment factors in, 380381
self-monitoring of, 385387 Alzheimers disease, 1819, 521 individual differences, 252
semantic, 597604 causes of, 520 information-processing in, 381
cognitive enhancers for, 8486, item recognition in order recall,
sensory, 607608
120, 522 377378, 378
skilled, 397
dementia, 109112 phonemic devices, 381
social, 621624 drug treatment for, 119120, primary memory, 380381
stages of, 580581 520522 response time in relation to,
stress and, 641643 in electroconvulsive therapy, 377378, 378
tip-of-the-tongue phenomenon, 132133 temporary groupings and, 380
650651 mood-related, 134135 working memory, 380
700 GENERAL INDEX
Memory span (continued) control processes, 386387 Mind as Observable Object

working memory tests, 265, 266 defined, 385 (Singer), 218
Memory strategies episodic memory tasks, 137 The Mind at Work and Play (Bartlett),
childhood memory feeling-of-knowing judgments, 56
development, 7172 385, 386387 Mineralocorticoid receptor,
See also Mnemonic devices individual differences, 252 hippocampal, 389
Memory tests. See Measurement of learning strategies, 592 Minimal brain damage/dysfunction,
memory; Recall tests prospective monitoring, 385 320
Memory traces retrospective confidence Mirror-reading, in amnesia, 548
in orienting reflex habituation, judgments, 385386 Misinformation, in reconstructive
498 test of, in frontal lobe damage, memory, 559560
in reconstructive memory, 559 160 La mission de lide (Piaget), 527
repetition and, 578 Metaplasticity, 387390 Mitogen-activated protein (MAP)
Semons theories, 606 activity-dependent, 388389 kinase
Memory traces, localization of, 332 BCM rule, 388 in conditioned taste aversion,
334 defined, 387 348
definition of memory trace, 332 modulatory, 389 Drosophila learning, 294
for different memory types, stress and, 389390 long-term depression induction,
333334 in synaptic operation, 341, 387 338
eyeblink conditioning and, 333 388 long-term potentiation
history of, 332333 Method of constant stimuli, 411 induction, 352, 353
Lashleys studies, 318319, 332 Metyrapone, epinephrine blocked second messenger systems, 596
333 by, 234 Miyake, A., 264
memory formation mechanisms Mice, transgenic Die mneme (Semon), 605, 606
and, 334 Alzheimers disease models, Die mnemischen Empfindungen (Mnemic
Menard, M. T., 322 910, 2122 Psychology) (Semon), 605, 606
Meningeal membranes, 181 Mnemonic devices, 393395
eyeblink classical conditioning,
Menstrual cycle, learning and, 614, imagery as, 7779, 496
169170
615 in memory rehabilitation, 562
visual cortical plasticity models,
Mental capacity tests. See memory span and, 381
428
Intelligence tests
See also Rodents practical applications, 395
Mental Imagery in the Child: A Study of
Microfilaments, of neurons, 205 principles of, 393
the Development of Imaginal
Microtubules in school learning, 591
Representation (Piaget and
of neurons, 205 in self-paced learning, 386
Inhelder), 528
of synapses, 215 types of, 393395, 394t
Mental images. See Imagery
Midbrain (mesencephalon) Mnemonists, 395397
Mental retardation, 381384
classification of, 381382 gross anatomy, 180 Modality effects, 397400
cultural-familial, 382 neuronal groups in, 181 in absence of sound, 399
degree of impairment, 382 Migration, navigation, and homing, auditory, 397, 608
dementia vs., 110 390393 birdsong learning guided by,
developmental concerns in, 384 bird food storage and, 9091 63, 6567
eidetic imagery as marker of, bird foraging and, 150, 151 classic, 397, 398
131 childrens geometric navigation defined, 398
etiology-based interventions, strategies, 91 in free recall, 399
383384 homing phenomenon, 391 long-term, 399
operant conditioning for, 61 learning in, 391392 in serial recall, 398, 399
organic, 382384 navigation in, 391392, 631 stimulus timing and, 400
recall in, 264265 orientation in, 390391 Modality-specific aphasias, 307
savant syndrome in, 587588 Mild cognitive impairment, 19, 112 Modality-specific knowledge loss,
See also specific syndrome names Miles, H. L. W., 315 306307
Mesencephalon (midbrain), 180, 181 Miller, N. E., 236, 327 Modeling
Metabotropic glutamate receptors Miller, Stephan, 309 in observational learning, 483
(mGluR) Millers Analogies Test, 92 484
characterization of, 176, 178 Milner, B., 414 symbolic, 484
long-term depression induction, Milner, Peter, 485 Modulatory synapses, 215
337, 338 Mimicry Mollusks. See Aplysia; Hermissenda;
in metaplasticity, 388 cryptic prey, 150 Limax; Tritonia
Metacognition, 385387 evolutionary learning and, 139 Mongolism. See Down syndrome
comparative cognition studies, in parallel retrieval mechanisms, Monitored learning systems, 433
93 374 435
GENERAL INDEX 701
Monkeys and apes neocortical plasticity and, 429 Mutations

abstract concepts based on 430 Drosophila, 137138, 293
categories, learning of, 92 neural substrates, 430 familial Alzheimers disease
analogical reasoning, 92 Motor skill learning, 409410 genes and proteins, 20
number scale learning, 8990 cerebellar genetic substrates mutant mouse models in
same/different (S/D) concept and, 170171 classical eyeblink
learning, 92 ecological approach, 410 conditioning, 169170
sex differences in learning, 613 GFAP knockout mouse model, Myelin sheath, axons, 209
614 170171 Myers, C. S., 55
spatial learning, 631 motor map plasticity, 429430
visual attention, 543 neural substrates, 430, 548 N
visual perception and memory, Schmidt schema theory, 409 Nadel, L., 414
540543, 541 410 Naloxone
visual recognition memory, 543 somatosensory cortex in, 431 blocking of memory effects, 234
working memory, 670671 tests for, 547548 fear conditioning blocking, 303
Monocular deprivation, cortical See also Procedural learning Names
plasticity in, 426, 427 Motor system aging and forgetting of, 11
Mood and memory, 133135 avoidance learning and, 451 memorizing, levels of processing
memory impairment related to, 453, 452 and, 81
134135 birdsong learning, 6465, 66 retrieval by, 508509
mood-congruent memory cerebellum and, 193199 Narrative prose. See Prose retention
(MCM), 133134 Movshon, J. A., 442443, 443 Natural selection. See Evolution and
mood-dependent memory Mowrer, O. H., 2, 327 learning
(MDM), 134135 mRNA (messenger RNA), 5, 6, 167, Natural settings, memory in, 417
Morgan, C. L., 325326 167 419
Morphological basis of learning and Mller, Georg Elias, 411, 411413 cues triggering memories, 418
memory, in invertebrates, 401403 early life and career, 411 419
neurite outgrowth, 401402 interference theory, 268 flashbulb memories, 419
structure and function, causal main contributions, 411413 laboratory methods vs., 417
relationships between, 403 memory consolidation concept, linearity and, 418419
synaptic ultrastructure, 402403 369 long-term memories, 418
Morphological basis of learning and Multi-infarct dementia, 110 recall of life experiences, 418
memory, in vertebrates, 404408 Multiple-memory systems, 413416 The Nature of Human Conflicts (Luria),
changes in nonneural elements, 357
basal ganglia in stimulus-
407408 Nausea
response, 544545
enriched environment and, food aversion learning related
dual-memory theories, 414415
405406 to, 148
hippocampus and, 414415, 534
number of synapses, 404405 See also Taste aversion and
implicit memory, 245246
plastic neural change and preference learning in
infants, 257
synapse formation, 406407 animals
long-term, 602603
Navigation
synapse efficiency charges, 407 neuroanatomy, 415416 bird food storage and, 9091
Morris, R., 347 place vs. response learning, bird foraging, 150, 151
Morris water maze, 9, 490, 492, 545, 534535 childrens geometric navigation
670 temporal lobe amnesic strategies, 91
Morton, J. R. C., 122 syndrome and, 414 Clarks nutcrackers and
Morton, John, 398 Multiple personality disorder, 23 landmark navigation, 631
Mossy fibers See also Amnesia, functional in migration, navigation, and
cerebellar cortex, 169, 436437 Multiple synaptic boutons, 407 homing, 391392
cerebellar long-term depression, Multiple task performance See also Spatial learning
170 aging and, 13 Nectar, foraging for, 151
cerebellum, 169, 195, 197 attention limits and, 4849 Negative patterning, 330
nerve terminals, 206, 209 Multiple-trace theory, 578579 Negative reinforcer
Mosteller, F., 360361 Multiplexing, in coding, 500 in operant conditioning, 495,
Motivation Muscarinic receptors 564
incentive motivation, 327, 328 limbic and cortical areas, 189 side effects of, 565
learned helplessness changes, pirenzepine-sensitive m1 taste aversion, 646
319 subtype, 189 Neilson, D. R., 225
observational learning, 482, 483 Mushroom body structure, insect Neobehaviorism, 57
Motor cortex, 429430 brain, 293 Neocortex
mapping of, 429430 Music, in oral tradition, 496 anatomy, 425, 441442
702 GENERAL INDEX
Neocortex (continued) Neural networks Neuromigration, focal injury in,

basal ganglia connections, 191 defined, 433 learning disabilities related to,
evolution of, 441, 445446 Hebbs cell-assembly idea and, 321322
hippocampal interaction with, 231 Neuromodulators
during memory consolidation, memory consolidation model, long-term potentiation
367368, 371 371, 371 induction, 341
integration in, 444 structural organization, 434 memory consolidation, 513515
long-term depression in, 335 See also Neural circuits neural computation, 444
memory consolidation and, Neural plasticity stress-related, 642643
367368, 371, 372 defined, 422 See also Drugs and memory;
neural computation, 441445 synapse formation and, 406407 specific names
processing speed, 444 Neural substrates, 451468 Neurons, 204210
topographic maps, 442 avoidance learning, 451454 Alzheimers disease effects,
Neocortical plasticity, 419432 emotional memory, 466468 2122
adult visual cortex, 419, 420 habit learning, 549 anatomy, 204210
422 imprinting, 249250 Aplysia, 42, 100
auditory cortex, 422425 motor map plasticity, 430 auditory, song-selective, 67
monocular deprivation and, 420 procedural learning, 544546 axons, 206, 207, 208, 208209
motor cortex, 420, 429430 Neural substrates of classical cerebellar cortex, 436
somatosensory cortex, 419420, conditioning, 454465 circuits, 206, 207, 208, 209210
431432 cardiovascular response, 454 cytoarchitecture of brain, 180
visual system development, 419, 457 181
425429 discrete behavioral response, dopamine-containing, 192, 303
Neonatal handling, behavior and 458460 glial cells vs., 204
learning effects, 121123 fear conditioning, freezing, high vocal center (HVc) song-
Neppe, V. M., 108 461462 selective, 66
fear-potentiated startle, 463465 imprinting changes and, 250
Nerve cells. See Neurons
Neuritic plaques, 20, 110 nhibitory-bursting, Limax, 281
Nerve growth factor (NGF), 420421
Neurobiology, semantic memory 282
for memory enhancement, 521
and, 602604 local circuit, 207, 209
ocular dominance plasticity and,
Neurodegenerative dementia, 110 myelinated axons, 209
420, 427
Neuroendocrinology. See Hormones nucleus basalis of Meynert Ch4
susceptibility to Alzheimers
Neurofibrillary tangles complex, 187, 188, 188190,
disease and, 190
in Alzheimers disease, 19, 110 189
Networks, associative, organization
in Ch4 neurons of basal projection, 208, 209
of memory and, 82
forebrain, 190 soma, 205
Neural circuits
Neurofilaments, of neurons, 205 synchronization, 500504, 502,
analog, 433 Neurogenesis, 469473 503
cerebellar, 436, 437 adult, 469, 471472 terminal boutons, 206, 209
reinforcement and reward, 567, adult birds, 69, 469, 471 visual cortical, 426
574575 defined, 471 See also Dendrites; Synapses
Neural computation, 432449 hippocampal, 469, 470, 471 Neurons, postsynaptic. See
approaches to learning, 432 473 Postsynaptic neurons
435 learning and, 406 Neurons, presynaptic, modulation
cerebellum, 436438 new gatekeeper, 472473 of, in Aplysia, 4243
defined, 433 Neuroimaging, 473474 Neuropeptides, 182183
feedforward mechanisms, 446 implicit memory, 246 Neuropeptide Y, 234
447, 449 knowledge category studies, 306 Neuropharmacology. See Drugs and
hippocampus, 439441, 440 Limax procerebral lobe odor memory
neocortex, 441445 response, 289 The Neurophysiological Basis of Mind
nervous system organization, memory consolidation areas, (Eccles), 128, 129
434 370371 Neuropsychology
olfactory cortex, 445449 prefrontal cortex, 161, 162 Freuds contribution, 156
principles of, 433 schizophrenia, 589 Lurias contribution, 357
subclustering activity, 448, 449 semantic memory, 604 Neuropsychology of Memory (Luria),
taxonomy of learning systems, source monitoring, 630 357
433434 spatial memory, 634 Neurotransmitter receptors, drug
telencephalic processing model, working memory, 674 effects on, 119120
445449 See also specific techniques Neurotransmitters, 474477
See also Parallel distributed Neuromedin K peptide circuitry, activity-dependent regulation of
processing models of memory 183 synthesis of, 46
GENERAL INDEX 703
Limax procerebral lobe, 283 Nodes of Ranvier, 208 Nystagmus, 659, 660
release by neurons, 204 Nominal groups, recall testing in,
releasing sites, on neurons, 209 621622, 622
synaptic response to, 215 Nonassociative learning O
visual cortical plasticity, 428 Aplysia, 3741, 401 Obesity, in Prader-Willi syndrome,
See also specific neurotransmitter C. elegans (Caenorhabditis elegans) 383
New gatekeeper, 472473 learning, 288, 290 Object concept, development of,
New-learning impairment. See Nonassociative long-term 479481
Amnesia, anterograde potentiation, 342 looking vs. reaching, 480481
Nicotine, as cognitive enhancer, 85 Nondeclarative memory. See Implicit Piaget on, 479
Nicotinic receptors, limbic and memory Object recognition
cortical areas, 189 Non-NMDA receptors, 176177 collaboration and, 623
Nif (nucleus interfacialis), in Nonprescription drugs, as cognitive memory span testing and, 380
birdsong learning, 64, 66, 66 enhancers, 8586, 522 Observational learning, 482484
Nim (chimpanzee), 315 Nonpyramidal cells, cerebral cortex, abstract, 482483
Nissl bodies, 205 201, 201 subfunctions, 482, 483
Nitric oxide Nonsense syllables Obsessive-compulsive disorder,
activation of guanylyl cyclase, interference theory, 268 autobiographical memory loss in,
596597 invention by Ebbinghaus, 125, 54
long-term depression induction, 412 Occasion setting, in discrimination,
337338 memory consolidation studies, 75, 113, 114
synthesis of, Limax procerebral 369 Occipital lobe
lobe, 283 memory testing and, 380, 412 anatomy and functions of, 200,
Nitric oxide scavenger, blocking of 413 200
long-term depression, 338 serial learning of, 610 damage to, autobiographical
Nitric oxide synthase, bee learning Nonspatial memory, 491492 memory loss in, 54
and memory, 275, 276 Nonverbal memory Occipitotemporal pathway
NK3 receptor, 183 head injury and, 228 anatomy, 540541
NMDA receptor antagonists See also Spatial learning lesions, 542
blocking of hippocampal- Nootropics (drug class), 522 limbic system and, 543
dependent learning, 172173 Norepinephrine neuronal properties, 541542
long-term potentiation and, activity-dependent regulation of OConnor, W. J., 127
348, 388 synthesis of, 46 Ocular dominance
metaplasticity and, 388 epinephrine in release of, 232 competitive synaptic
prevention of second-order functions of, 182 mechanisms, 435
conditioning, 98 memory consolidation, 513 defined, 426
types of, 177 memory effects, 232 development of, 426
NMDA receptor-mediated EPSCs, 67 neural computation, 444 neocortical plasticity, 420, 426
NMDA receptors (NMDAR) visual cortical plasticity, 428 Odor learning/perception. See
Aplysia sensitization, 34 Norman, D. A., 49 Olfactory entries
back-propagating action Noxious events, avoidance learning, Oedipus complex, 156
potentials in opening of, 352 1 OKeefe, J., 414, 529
birdsong learning, 67, 69 Nucleus accumbens, in Olds, James, 484487, 485
characterization of, 177178 reinforcement, 568569 Olfactory bulb, 210, 211, 211
glycine site, 177 Nucleus basalis accessory (AOB), 210
hippocampal learning and, 173, Ch4 complex, 188190, 189 Olfactory cortex, 210213
174 cholinergic neurons, 187, 188, amygdalar connections, 185,
imprinting and, 250 188190, 189 186
on latent synapses, 353 cholinergic system, in learning, anatomy, 210213, 446, 447
localization of, 177 424, 424 basic cortical circuit, 211213
long-term depression signal Nucleus interfacialis (Nif), in complex computations, 447448
transduction, 336, 337 birdsong learning, 64, 66, 66 cortical areas, in humans, 211
long-term potentiation and, Number scale, learning by animals, 212
340342, 341, 344, 347348 8990 neural computation, 445449
olfactory cortex, 212 Nutrition and diet, 147149 overall structure, 210211, 211,
visual cortical plasticity, 428 See also Food aversion and 212
See also Glutamate receptors preference learning in piriform cortex, 210, 212, 212
N-methyl-d-aspartate (NMDA) humans; Food caching/storing simple computations, 446447
receptors. See NMDA receptors nXIIts (hypoglossal nucleus, spoken sounds and, 448449,
(NMDAR) tracheosynringeal portion of), in 449
N, Lorente de, 231 birdsong learning, 64, 65, 66 subclustering activity, 448, 449
704 GENERAL INDEX
Olfactory nucleus, anterior (AON), Orbital cortex. See Prefrontal cortex Paired-associate learning
210 and memory in primates defined, 655656
Olfactory pathway Orbitofrontal neurons, 577 diagnosis of amnesia, 106
bee learning and, 276 Order information imagery and word learning, 79
olfactory cortex in, 210 order retrieval and short-term Paired-comparison technique, infant
Olfactory sensory epithelium, 210 memory recall, 377379, 378 visual recognition memory, 254
Olfactory system search processes in recovery of Paired helical filaments (PHF),
anatomy, 446, 447 order information, 377 poorly soluble, in Alzheimers
Drosophila, 293295 Organic amnesia. See Amnesia, disease, 19
Limax, 281285 organic Paivio, A., 78, 131
neurogenesis, 469 The Organization of Behavior (Hebb), Paleocortex. See Olfactory cortex
Olfactory tract, lateral (LOT), 210 230, 332, 485 Pallidum. See Globus pallidus;
212, 212 Organization of memory, 8283 Striatum
Olfactory tubercle (OT), 210 associations and, 82 Palmer, J. C., 146
Oligodendrocytes, 181 associative networks and, 82 Papez circuit, 204
Oligodendroglial cell, 209 autobiographical memory and, Parahippocampal region
Oligodeoxynucleotides, antisense, in 27
cortical regions, 202
taste aversion learning, 168 coding and organization, 82
cortical system, 203204
Olton, David, 488492, 489 embodiment and representation
general features, 202
Omission, in instrumental in, 83
perforant pathway, 203, 214
conditioning, 103 experts memory, 143, 251
perirhinal cortex connections,
On Aggression (Lorenz), 356 localization of memory, 332
213
One-trial passive avoidance training, mnemonic devices in, 393
schemata in, 83, 551552 subcortical connections, 214
2 trisynaptic organization, 203
On Memory (Ebbinghaus), 412 semantic networks and semantic
features in, 8283 wiring, 202204
On Sensations of Color: Psychophysical
serial, 609612, 610 See also Hippocampus
Investigations (Mller), 412
Orientation Parallel distributed processing
Operandum. See Operants
dependence on, 633, 633634 models of memory, 507511
Operant behavior, 493495
magnetic fields and, 390391 characteristics of PDP systems,
aversive stimuli, 495
in migration, navigation, and 508
defined, 564
homing, 390391 connectionist models, 441
Forman paradigm, 259
visual experience and, 427 distributed model, 509510, 511
Lorenzs view, 356
Orienting reflex habituation, 497 Hebbs cell-assembly idea and,
rate of response, 494, 494
499 231
reinforcement and, 463465,
event-related potentials, 498 hybrid systems, 510511
493494
Skinners concept of, 58, 493 neuronal mechanisms, 498499, learning rules, 511
495, 564, 616617 499 localist model, 508509, 509,
stimulus control, 495 Orienting tasks, levels of processing 510
Operant conditioning. See and, 8081 Parallel retrieval mechanisms, 374,
Conditioning, instrumental Origin of Species (Darwin), 238 375
(operant) Orthogonalization, 440 mimicry of serial mechanisms,
Operants Orthoptera learning, 259 374
Oscar-Berman, M., 51 Parasubiculum
defined, 564
Oscillations, synchrony, and amygdalar connections, 214
inapplicability to studying
neuronal codes, 500504 entorhinal cortex connections,
associative learning, 103
Outlining, prose retention and, 552 213, 214
Skinners concept of, 58, 616
Output interference, 399 as region of parahippocampus,
617
Overload principle, cues, 582 202
Operation-span test, 265, 266
Overshadowing Parietal lobe, 200, 200
Opiate antagonists, memory effects,
of conditioned stimulus, 76 Parkinsons disease
233234
verbal, 146 basal ganglia involvement, 191
Opioid peptides
Over-the-counter drugs, as cognitive cholinergic malfunction in, 476
blocking of, by other hormones,
enhancers, 8586, 522 dorsal striatum abnormality in,
234
Oxytocin, memory effects, 233 192
memory effects, 233234
Oxytremorine, endorphins blocked double-dissociation in, 549
neurotransmitter role, 476
by, 234
Optimal foraging theory, 149150 Lewy bodies in, 110
Optokinetic eye-movement response memory impairment, 520
(OKR), 659661 P Parrots
neuronal circuits, 659660 PACAP (pituitary adenylyl cyclase concept learning, 92
Oral traditions, 417, 496497 activating polypeptide), 5, 6 song learning, 64
GENERAL INDEX 705
Pars compacta, of substantia nigra, Perceptron neural-network model, Phrenology, 318

192 231 Physical activity, adult neurogenesis
Pars reticulata, of substantia nigra, Perceptual identification, implicit and, 471
191192 memory and, 246 Physiology
Passeriformes. See Birdsong learning Perceptual learning Lashleys studies, 318319
Passive (inhibitory) avoidance, 512 adaptation and, 421 Olds studies, 484
516 in amnesia, 548 The Physiology of Nerve Cells (Eccles),
behavioral testing apparatus, in discrimination, 114 128
512, 513 implicit memory and, 245, 246 Physiology of the Brain and Comparative
defined, 451 Perforations, postsynaptic, 407 Psychology (Loeb), 616
fear conditioning vs., 515 Performance Physostigmine, 85, 234
fear learning, 512 of multiple tasks, 13, 4849 Piaget, Arthur, 526
memory consolidation, 513515 stress and, 641642 Piaget, Jean, 526, 526529
retention latency, 512513 Periaqueductal gray. See Reticular on cognitive development, 528
temporal patterns, 515 formation on egocentrism, 526527
See also Avoidance learning, Perirhinal cortex, 213214 experimental methods, 527528
active and passive cortical connections, 213 on genetic epistemology, 527
Past events, memory for. See as extension of ventral visual influence of, 526, 529
Secondary memory (long-term stream, 540541 on memory, 528529
memory) hippocampal connections, 214 on object concept, 479
Patches of food, foraging, 150151 lesions, 542 on stages of development, 527
Pate, J. L., 315 parahippocampal connections, Piaget, Valentine Chatenay, 526
Pattern analysis/discrimination, 213 Pick bodies, in dementia, 110
positive and negative patterning, subcortical connections, 214 Picks disease
113114 Perry, Ralph Barton, 652 memory impairment, 520
Pattern completion, 439, 510 Perseveration-consolidation semantic dementia in, 603
Pavlov, Ivan, 516520 hypothesis, 412413 Picrotoxin, as CNS stimulant, 84,
conditioned inhibition Personal facts. See Autobiographical 120
procedure, 75 memory; Episodic memory Pictorial images and memory. See
conditioned (orienting) reflex Personality and Psychotherapy (Dollard Eidetic imagery; Iconic memory;
discovery, 497, 516 and Miller), 236 Imagery
dog-salivary conditioning PET (positron emission Picture superiority effect, 78
studies, 518 tomography), 306, 497 Pigeons
experimental neurosis work, as Phantom pain, somatosensory cortex comparative cognition studies,
basis for desensitization, 60 in, 431 92, 93
influence of, 235, 327, 519520 Pharmacological treatment of navigational system, 392
political skill, 517 memory deficits, 520522 Pilzecker, Alfons, 268, 369, 412413
reinforcement methods, 563 See also Drugs and memory Piracetam, for memory
See also Conditioning, classical; Phillips, W. A., 50 enhancement, 522
Pavlovian conditioning Phobias Piriform cortex, 210, 212, 212
Pavlov, Petr, 517 behavioral fear-reducing Pituitary adenylyl cyclase activating
Pavlovian conditioning procedures, 5960, 525 polypeptide (PACAP), 5, 6
avoidance learning, 23 biological determinants, 525 PKA. See Protein kinases, cAMP-
in brainwashing, 518 causes, 523524 dependent (PKA)
conditioned reflex concept, 220 classification, 523 Place cells, 529532, 530
conditioned/unconditioned conditioning theory, 523524 ensembles, 532
stimuli, 518 neoconditioning concepts, 524 in spatial learning, 9, 531
fear conditioning, 166, 461, 525 Place fields
466467 verbal information causing, 525 development of, 531
Hermissenda, 277280, 279 Phormia regina (blowflies), 260 firing of, 529, 530, 531
influence of, 519 Phospholipase A2 (PLA2), 337 hippocampal circuitry, 532
repeated exposure to Phospholipase C (PLC), 337, 595 hippocampal pyramidal cell
conditioned stimulus, 123 Phosphorylation firing, 174
Rescorla-Wagner model of, 329, AMPA receptors, in LTD, 338 in spatial learning, 9
329330 AMPA receptors, in LTP, 354 Place vs. response learning, 533535
See also Conditioning, classical of tyrosine hydroxylase, activity- Planum temporale, dyslexia and,
Pavlovian Society, 519520 dependent, 5, 6 322323
Pearce, John M., 330 Photographic memory, eidetic Plaques, neuritic, 20, 110
Perception impairment, in visual imagery vs., 130, 131 Plasma membrane, of neurons, 204
object agnosia, 306 Photoreceptors, Hermissenda, 278, Plasticity
Perceptron learning algorithm, 16 279 metaplasticity, 387390
706 GENERAL INDEX
Plasticity (continued) functions, 200, 535 retrieval mechanisms, 373374,

neocortical, 419432 learned bradycardia in rabbits, 375
short-term, 389 455456 search processes in recovery of
spinal, 639641 localization of function, 161162 order information, 377
vestibulo-ocular reflex (VOR), in memory selection, 457 time course of short-term
659661 orbital and dorsolateral regions, memory retrieval, 376, 377
See also Synaptic plasticity 161 working memory vs., 672673
Plato, 251, 591 in working memory, 674 Primary model, serial learning, 612
Plesiomorphic traits, defined, 258 Prefrontal cortex and memory in Primates, 535543
Pleydell-Pearce, C. W., 52 primates, 535538 language learning in nonhuman
Plus-maze tasks lesions and developmental primates, 314317
in behaviorist/cognitivist delay, 536 nonhuman, visual perception
learning debate, 533534 short-term memory and, 535, and memory in, 540543
multiple memory systems, 534 536, 538 prefrontal cortex and memory,
535 specificity of areas, 536 535538
Poetics, in oral tradition, 496 Pregnenolone, as cognitive visual attention in, 539540
Pointing tests, in frontal lobes enhancer, 86
Priming
damage, 159 Premack, D., 315
amnesia patients, 30
Pollination, constancy and, 140 Prepyriform cortex. See Piriform
implicit memory tests, 243244,
Polyribosomal aggregates, 407 cortex
245
Pons, 180, 181 Presenilin-1 gene, 20, 21
indirect memory tests, 11
Popper, K., 128, 130 Presenilin-2 gene, 20, 21
Population dissociation, implicit and Presubiculum Primitive societies, eidetic imagery
explicit memory, 244 amygdalar connections, 214 study, 131
Positional association, 611, 612 entorhinal cortex connections, Principles of Behavior (Hull), 235, 236
Positron emission tomography (PET) 213 Principles of Behavior: An Introduction
knowledge category studies, 306 as region of parahippocampus, to Behavior Theory (Hull), 236
sensory stimulation, 497 202 Principles of Psychology (James), 125,
Postsynaptic conductances (PSCs), Presynaptic inhibition, 128 297
amygdalar, 343 Presynaptic neurons, modulation of, Principles of Psychotherapy (Janet), 218
Postsynaptic neurons in Aplysia, 4243 Proactive inhibition, 153, 159, 656,
calcium increase, LTP and, 352 Presynaptic terminals 670
density of, in imprinting, 250 expression of long-term Probabilistic classification task, 549
long-term potentiation and, potentiation, 352353 Probability matching theory, 360
340, 341, 353354 long-term potentiation 361, 361
modulation of, in Aplysia, 43 induction, 340, 341 The Problem of Noise (Bartlett), 56
Postsynaptic potential maintenance of long-term Proboscis extension reflex
amygdalar, 343 potentiation, 349350 bees, 261, 274
Hermissenda, 278, 279 Preverbal memory. See Infancy, flies, 260
See also Excitatory postsynaptic memory in Procedural knowledge (skills). See
potential (EPSP) Prey Skills
Posttetanic potentiation, amygdalar cryptic, 150 Procedural learning, 544549
long-term potentiation, 344 selection, 150 animals, 544546
Posttraumatic amnesia (PTA), after Primary memory (short-term
cerebellum in, 544
head injury, 228 memory)
cognitive skills and habits, 548
Potentiation, long term. See Long- aging effects, 1213
549
term potentiation (LTP) attention and, 4950
declarative learning vs., 544
Practice, memory rehabilitation, 562 C. elegans (Caenorhabditis elegans),
defined, 547
Prader-Willi syndrome, 383 288290, 289
Pragmatics, and language, 311 confusion over findings about, eyeblink classical conditioning,
Precategorical acoustic storage, 398, 1213 544
607608 memory span, 380381 humans, 547549
Predatory hunting behavior, 461 order retrieval and short-term memory trace localization and,
Preference learning. See Food memory recall, 377379, 378 333
aversion and preference learning prefrontal cortex in, 535, 536 perceptual skills, 548
in humans; Taste aversion and recognition tests following sleep and skill consolidation,
preference learning in animals short-term memory list, 619620, 620
Prefrontal cortex example, 373, 373 See also Motor skill learning
anatomy, 535536, 537 recognition time and, 373, 374 Procedural memory. See Implicit
connection with olfactory cortex, relation to other abilities, 379 memory
211 response time and, 374375 Procerebral lobe. See Limax
GENERAL INDEX 707
Processing of memory. See Levels of second messengers and, 594 Psychophysics, Mller on, 411412
processing; Memory consolidation; 595 Psychoses
Parallel distributed processing Protein kinases, autobiographical memory loss
models of memory Ca2+/calmodulin-dependent. See in, 54
Process mnemonics, 394395 Calcium-calmodulin-dependent See also Schizophrenia
Production, in observational protein kinase (CA/CAM kinase II) Public memory. See Collective
learning, 482, 483 Protein kinases, cAMP-dependent memory
Progenitor cells, in neurogenesis, (PKA) Punishment
471 Aplysia sensitization, 34, 4344 avoidance learning, 12
Programmed instruction, 617 bee learning and memory, 275, instrumental conditioning, 103
Project for a Scientific Psychology 276 reinforcement and, 564565
(Freud), 156 catecholamine synthesis, 5 Purines, as neurotransmitters, 6
Prolactin, 234 hippocampal learning, 173 Purkinje cells
Propranolol, blocking of naloxone long-term potentiation calcium influx, in cerebellum,
by, 234 induction, 352 197
Prosencephalon (forebrain), 180, regulation, 595 cerebellar, 195, 197, 198, 436
210 Protein kinases, mitogen-activated. 437
Prose retention, 550553 See Mitogen-activated protein eyeblink classical conditioning,
assessment of, 553 (MAP) kinase 78, 169170
Bartletts studies, 417 Protein phosphatase, regulation of long-term depression, 335, 336,
causal relations, 552 LTP induction, 354 338
encoding and reading processes, Protein phosphorylation, long-term mutant mouse models, 169170
550551 potentiation induction and, 354 vestibulo-ocular reflex plasticity,
factors influencing, 550, 551 Protein synthesis, 553556 661
intervening information, 552 amygdalar learning, 167168 Purpose, in Tolmans view of
553 correlative studies, 555 behaviorism, 57
organization and prose type, Hermissenda memory Purposive behavior. See Operant
551552 consolidation, 280 behavior
schemata and, 550551 interventive strategies, 554 Purposive Behavior in Animals and Men
time factors, 552553 long-term depression induction, (Tolman), 652
Prospective memory 338 Putamen
aging effects, 13 long-term memory in in forebrain, 180
head injury and, 229 vertebrates, 553556 neuron structure of, 181
measures of, 365366 memory consolidation and, 366, Parkinsons disease and
Prospective monitoring, 555556 Huntingtons disease
metacognition and, 385 See also Activity-dependent associated with, 192
Protein kinase C (PKC) regulation of neurotransmitter as striatal component, 192
synthesis Pyramidal cells
Aplysia sensitization, 44
Protein tyrosine kinase (PTK), long- amygdalar, 185
Hermissenda synaptic efficiency,
term depression induction, 338 apical dendrites of, 210
278
Pseudo-conditioned and sensitized basal dendritic region, 206, 209
long-term depression and
responses, 102 cerebral cortex, 201, 201
vestibulo-ocular reflex
excitatory feedback, within
adaptation, 170 PS-1 gene (presenilin-1), 20, 21
piriform cortex, 212
long-term depression induction, PS-2 gene (presenilin-2), 20, 21
eyeblink classical conditioning
338 Psychoanalysis, origins of, 155156
and, 89
long-term potentiation Psychogenic amnesia, 23
olfactory cortex, 212
induction, 352, 353 Psychogenic fugue, 23
processes of, 209
phosphorylation of AMPA Psychological Abstracts, 238
Pyriform cortex. See Piriform cortex
receptors, 354 Psychological Care of Infant and Child
second messenger activation, (Watson), 668
594596 Psychology, physiological, 318319
Protein kinases Psychology and Primitive Culture Q
cascades, in Aplysia sensitization, (Bartlett), 55 Quantitative Aspects of the Evolution of
4344 Psychology and the Soldier (Bartlett), Concepts (Hull), 234
in catecholamine synthesis, 46 55 Quinn, W. G., 293
inhibitors of, memory effects, Psychology for Neurologists Quinoxaline derivatives, as AMPA/
120 (Freud), 156 QA receptor antagonists, 176
long-term depression induction, The Psychology of Human Learning: An
338 Introduction (McGeoch), 363
long-term potentiation Psychopharmacology. See Drugs and R
induction, 352, 353 memory R. B. (amnesia patient), 29
708 GENERAL INDEX
RA (robust nucleus of archistriatum), ordered recall and short-term Reinforcement, 563566

in birdsong learning, 64, 65, 66, memory, 377379, 378 anatomical substrates, 566, 566
66, 67, 68 Recency effect, modality effect and, 570
Rabbits 398 computational, 565
eyeblink classical conditioning, Recent memory. See Working defined, 563, 570
78, 8 memory Hulls drive-reduction
nictitating membrane/eyelid Receptive fields hypothesis of, 328
conditioning, 458459 plasticity, 422, 424, 424425 methods, 493494
Radial maze spinal cord neurons, 432 Mowrers concept, 327
Olton, David, and, 488490 Receptors in operant conditioning, 493
reference-memory tasks, 491 long-term depression signal 495
use of, 492 transduction, 336338 operant nature of, 563564
working memory and, 488490, non-NMDA receptors, 176177 positive vs. negative, 564
670 See also NMDA receptors punishment and, 564565
Rain Man (movie), 588 (NMDAR); Synapses; Synaptic reward vs., 563
Rajan Srinavasen Mahadevan plasticity schedules for, 494495
(mnemonist), 397 Recherche (Piaget), 527 shaping of behavior, 493, 565
Ramn y Cajal, Santiago, 557558, Reciprocal inhibition, phobia
Skinners concept of, 58, 493
558 treatment, 60
494, 564, 617
law of dynamic polarity, 557 Recognition masking, 608
Thorndikes law of effect and,
on neurons, 215, 401, 404 Recognition memory
326
on protein synthesis in memory collaboration and, 623
Thorndikes studies on, 563564
formation, 553 modality effects on, 399
unobservable, in learning
trisynaptic circuit of visual, in monkeys, 543
theory, 326, 328
(para)hippocampal system, Recognition memory tests
See also Reinforcement or
203 following short-term memory
reward in learning
Ranck, James B., 529 list, example, 373, 373
Reinforcement learning algorithms,
Ranschburg effect, 580 forced-choice recognition, 365
1617, 565
Ranviers nodes, 208 single-item (yes/no) tests, 365
Reinforcement or reward in
Rapid eye movement (REM) sleep, Recognition time
learning, 566577
618621, 619 defined, 373
anatomic substrates, 566570,
Rate estimation theory (temporal 567
376
model of learning), 331 cerebellum, 570572
short-term memory retrieval,
Rats. See Rodents cortical contributions, 569
373, 374
Rayner, Rosalie, 667668 electrical self-stimulation of
Reconstructive memory, 558561
Reading, prose retention, 550551 brain, 567, 573575
evaluation and attribution, 560
Reading-span test, 265, 266, 266 factors influencing, 567
561
Recall historical perspectives, 567568
implanted memories, 560
collaborative inhibition, 622 nucleus accumbens, 568569
misinformation, 559560
cued vs. free, 581 studies on, 558559 striatum, 575577
editing memory during, 368 Redundancy removal, 439440 Reinforcement therapy, as clinical
by groups, 621623 Reference memory, 489, 490 application of operant
in infancy, assessment methods, Reference systems, spatial, 634, 635 conditioning, 60
73 Reflex Reinstatement, in hypnotic age
judgment of learning in habituation and sensitization of, regression, 241
predicting, 385 223226 Relational information theories,
memory span and, 380 Skinners conception of, 58 599600
in mental retardation, 264265 Reflexes of the Brain (Sechenov), 517 Relations between relations,
modality effect in, 398, 399 Rehabilitation of memory disorders, analogical reasoning, 92
neuroimaging, 474 561563 Remembering. See Cues, retrieval;
repeated testing of, 581 errorless learning, 563 Forgetting; Recall; Retrieval
with retrieval cues, 581582 external aids and environmental processes in memory; Working
working memory in, 380 supports, 562 memory
See also Retrieval processes in mnemonic strategies, 562 Remembering (Bartlett), 56, 417
memory practice and rehearsal Reminiscence bump
Recall tests techniques, 562 autobiographical memory, 52,
cued recall, 364 vanishing cues, 562 53, 418
free recall, 364 Rehearsal collective memory, 88
frontal lobe lesions, 158 in memory rehabilitation, 562 Reminyl (galanthamine), for
immediate serial recall, 364 as memory strategy, 71 Alzheimers disease, 521
GENERAL INDEX 709
Remote events, memory for, amnesia Retention of information. See Retroactive inhibition, 153, 363, 656
patients, 3031 Memory consolidation; Retrieval Retroactive interference, 268, 269,
REM sleep, 618621, 619 processes in memory; Secondary 671
Renan, Ernest, 585 memory (long-term memory) Retrograde amnesia. See Amnesia,
Repetition and learning, 578580 Reticular formation retrograde
distributed practice effects, 115 in fear behavior, 462 Retrospective confidence judgments,
117 mesencephalic, and long-term 385386
ineffectiveness in, 579580 habituation, 225 Retrosplenial cortex, 213
memory rehabilitation, 562 structure, 181 Revivification, in hypnotic age
memory traces in, 578579 Retina, lesions and neocortical regression, 241
Semons theories, 606 plasticity, 420421 Revue Philosophique, 585
subproblems, 578 Retrieval cues. See Cues, retrieval Reward, 575577
verbal materials, 81, 579580 Retrieval processes in memory, 580 behavioral functions, 576
Repression, infantile amnesia and, 584 classical conditioning, 101
24, 26, 241 aging differences, 9, 12 dopamine-containing neurons
Reproductive inhibition theory of by content, 509 and, 192
forgetting, 269 control in, 386387 instrumental conditioning, 103
Reproductive success, learning in, defined, 373, 580 pathway for, in bees, 276
138 direct-access in item reinforcement vs., 563
Reptiles and amphibians, 445446 information retrieval, 374 striatum in, 575577
375, 376 See also Reinforcement or
Rescorla, R. A., 2, 16, 99, 519, 524
drug effects, 119 reward in learning
Rescorla-Wagner model/rule
dual coding and retrieval, dj R - f(S, A) (Skinner equation), 616
of blocking effect, 302, 329330
vu and, 109 Rhombencephalon (hindbrain), 180,
eyeblink conditioning and, 302,
individual speed differences, 187
303
252253 Rhythm, in oral tradition, 496
Hulls principles and, 360
infant memory, 255 Ribot, Thodule, 585, 585
Konorskis implementation of,
interference and forgetting,
310 law of regression, 30, 585
271272
least mean square algorithm Risk sensitivity, bees, 262
and, 16 Rivastigmine (Exelon), for
376377
of Pavlovian conditioning, 329, Alzheimers disease, 521
by name, 508509
329330 RNA, messenger
order retrieval and short-term
Pearces configural approach vs., suppression of transcription,
memory recall, 377379, 378
330 amygdalar learning and, 167,
parallel distributed processing
as special case of connectionist 167
models, 508510
network, 362 tyrosine hydroxylase, 5, 6
prefrontal cortex function, 161
Resemblance, and representation, in Robust nucleus of archistriatum
162
Aritotelian theory, 4647 (RA), in birdsong learning, 64, 65,
recognition time and short-term
Response, 533535 66, 66, 67, 68
memory, 373, 374
decrement of, in habituation, Rodents
retrieval in short- and long-term
224 memory, 373374 avoidance learning, 23
place vs., in learning, 533535 search processes in recovery of early handling of, 121122
prevention of, in phobia order information, 377 eyeblink classical conditioning
treatment, 59 Semons ecphory theory, 606 in rats, 459
See also Stimulus-response (S-R) serial and parallel mechanisms, sex differences in learning, 613
learning 374, 375 sleep and brain activation, 621
Response competition, in source monitoring, 630 spatial learning and memory, 9,
interference and forgetting, 270 state-dependent retrieval, 119, 631
Response-shock (R-S) timer, 582583, 583t working memory, 670
avoidance learning, 1 tests for, 581582 See also Mice, transgenic
Response time, in retrieval time course of short-term Roediger, Henry, 559
defined, 374 memory retrieval, 376, 377 The Role of Speech in the Regulation of
linear increase, in relation to list tip-of-the-tongue phenomenon, Normal and Abnormal Behavior
length, 374 650651 (Luria), 357
memory span and, 377378, transfer-appropriate processing, Roman High Avoidance and Roman
378 583584 Low Avoidance rat strains, 23
recency and position in study underlying retrieval Romski, M. A., 315
list, 374375, 376 mechanisms, 374, 375 Rosen, G. D., 322
Retardation. See Mental retardation See also Direct-access retrieval; Rosen, J. B., 166
Retention latency, 512513 Recall Rough endoplasmic reticulum, 205
710 GENERAL INDEX
Routine and variation memory. See Secondary memory (long-term Self-talk strategies, 592
Working memory memory) Seligman, Martin, on phobias, 525
Rovee-Collier, C., 255 attention and, 50 Semantic memory, 597604
Rumbaugh, D. M., 315 C. elegans (Caenorhabditis elegans), aging effects, 11
Rumination, in anxiety, 134 290 Alzheimers disease and, 1819,
Russell, James E., 649 CREB transcriptional activators 111
Ryanodine receptors, 337 in induction of, 44 categorical knowledge, 598599
defined, 49 cognitive aspects, 598601
in dementia, 111 conceptual combinations, 600
S head injury and, 228229 601
S. See Shereshevskii (mnemonist)
Sadness, mood-congruent memory hippocampus in, 439441 declarative memory system and,
and, 133 hormone effects, 232 105
Salmon, navigation in, 391392 interference and forgetting, 269 dementia and, 111, 602603
Same/different (S/D) concept modality effects, 399 domains in, 604
learning, animals, 92 multiple long-term systems, episodic memory vs., 597, 598
Sarah (chimpanzee), 92 602603 false, 145
SAT (Scholastic Assessment Test), natural settings, 418 head injury and, 228229
266 protein synthesis in, 366, 553 judgments of relative
Satiety, sensory-specific, 148 556 magnitude, 599
Savage-Rumbaugh, E. S., 316, 317 retrieval mechanisms, 376 neurobiology, 602604
Savant syndrome, 587588 search processes in recovery of organic amnesia and, 30
Schacter, D. L., 135, 136 order information, 377 relational information theories,
Schemata taxonomy, 548 599600
autobiographical memory and, transcription and translation separate vs. unitary semantic
27 requirements, 173174 stores, 603
Bartletts beliefs, 56 See also Memory consolidation; Semantic networks, organization of
collective memory, 88 Retrieval processes in memory memory and, 8283
memory organization and, 83 Second messenger systems, 593597 Semon, Felix, 605
prose retention and, 551552 Aplysia sensitization, 4344 Semon, Richard, 605, 605606
recognition memory based on, degradation, 596 Senility. See Alzheimers disease;
29 function, 594595 Dementia; Drugs and memory
school learning, 591 Hermissenda, 278 Senior moments, 520
source monitoring, 629 MAP kinases, 596 Senses. See Sensory memory; specific
Schenk, D., 9 production of molecules, 594 sense
Schizophrenia, 588590 protein kinase cascades and, The Senses and Intellect (Bain), 325
autobiographical memory loss 4344 Sensitization, 223226
in, 54 protein kinases and, 594596 Aplysia studies, 34, 38, 3840,
cognitive deficits, 498 signal transduction pathway 4144, 401
frontal lobe abnormalities, 589 variations, 596597 C. elegans (Caenorhabditis elegans),
590 synapses, 215 288
memory deficits, 588589 Second-order conditioning cellular analogue, in Aplysia, 39
neuroimaging studies, 589 Aplysia, 98 defined, 102
Schleiden, Jacob, 557 bees, 274 mechanisms of, 224225
Schmidt, R. A., 409 higher-order classical in spinal cord, 225
Schnabel, I., 123 conditioning, 98, 99, 99100 startle reflex, 225226
Scholastic Assessment Test (SAT), -Secretase, 22 Tritonia, 291292
266 -Secretase, in familial Alzheimers vertebrates, 223226
School learning, 590593 disease, 20, 22 Sensitization, long-term
history of, 590591 Seizure, dj vu associated with, 108 in Aplysia, cellular correlates of,
knowledge acquisition, 592 Selective association, taste aversion, 42
research on, 591 645 in Aplysia, molecular basis of,
Schwann, Theodor, 557 Selective neuronal activity, in active 4144
Schwann cell, 181 avoidance learning, 451, 453 Sensory information, Hunters
Science and Human Behavior The Self and Its Brain (Eccles), 129 theories, 239
(Skinner), 617 130 Sensory memory, 607608
Scopolamine, memory effects, 120 Self-monitoring/control of memory. attention and, 49
Scoville, W. B., 414 See Metacognition auditory, 398, 607608, 670,
Scripts, defined, 83 Self-paced learning, control 671
Search rate, memory span and, 380 processes in, 386 defined, 49
Sea snail. See Aplysia Self-stimulation of brain. See Brain recognition masking, 608
Sechenov, Ivan, 517 stimulation reward (BSR) visual, 607, 670, 671
GENERAL INDEX 711
Sensory-specific satiety, 148 Shock treatment. See Sleep, 618621

Septal nucleus, medial, of basal Electroconvulsive therapy and memory consolidation and,
forebrain, 187188 memory loss 618621
Septohippocampal system, blocking Short-term memory. See Primary memory reactivation, 620621
and, 304 memory (short-term memory); REM sleep, 618621, 619
Serial organization, 609612, 610 Working memory stages, 618
definitions and distinction, 609 Shuttle box, avoidance learning, 2, 3 visual cortical plasticity and, 429
610 Sight. See Visual headings Slug. See Limax
distractor paradigm, 611, 612 Signal transduction Smell, sense of. See Olfactory cortex
frontal lobe damage and, 159 Hermissenda, 278 S memory system, imprinting, 250
method of anticipation, 610 hippocampal learning, 173 Smith, Emily Mary
611, 611 long-term depression, 335338 Smith, M. E., 51
theoretical models, 611612 long-term potentiation Smith, Stevenson, 218
Serial recall tasks mechanisms, 351354 Smooth endoplasmic reticulum, 205
modality effects in, 398, 398 second messenger systems, 593 Snail. See Aplysia
as recall measure, 364 594 Social factors and issues. See Culture
Serial retrieval mechanisms, 374, variation in pathways, 596597 and society; Schemata
375 Sign-gestalt-expectation, 652 Social Learning and Imitation (Miller
exhaustive serial search Silence, modality effects and, 399 and Dollard), 236
properties, 374 Silliman Lectures, Yale University, Social loafing, 622
mimicry by parallel 638, 639 Social memory processes, 621624
mechanisms, 374 Similarity, law of, 219 collaboration in recall, 621622,
recovery of order information, Simon, H. A., 140, 141 622
377 Simonidess method of loci, 77, 78,
collaborative vs. nominal
Serine phosphorylation, 5 393, 396
groups, 622
Serotonin Singer, Edgar, 223
Social phobias, 523
memory modulation, 476 Single-item (yes/no) recognition
Social Psychology (McDougall), 652
mimicry of tail-shock induced tests, 365
Society for Experimental Psychology,
inhibition of reflexes in Siphon-gill withdrawal reflex, in
411
Aplysia, 40 Aplysia, 33, 37, 42, 401
Socrates, 591
defensive, 389
modulation of Aplysia Sokolov, E. N., 224
second-order conditioning, 98
presynaptic and postsynaptic Solomon, R. L., 2
neurons, 42, 43 tail-shock induced inhibition of,
3940 Soma, anatomy, 205
Sevcik, R. A., 315 Somatosensory cortex
Skilled memory, theory of, 397
Sex differences in learning, 613615 Hebbian processes, 432
Skill learning
birdsong learning, 6769, 68 neocortical plasticity and, 431
preserved, in amnesia, 30, 106
gonadal hormones and, 615 432
working memory and, 264
neural structures, 614615 in phantom pain, 431
Skills
spatial learning, 635636 topographic maps, 442
cognitive, 548549
stress and, 642 defined, 547 Somatostatin, 234
Sex hormones. See Gonadal in memory and intelligence, Sometimes opponent process (SOP)
hormones 266267 model, 624628
Sexual behavior, aberrant, aversion perceptual, 548 acquisition and extinction, 626
therapy for, 61 preserved, in amnesia, 547 627, 627
Sexual imprinting, 248 procedural, 548 cue-competition effects, 627
Sexuality, infantile, Freuds theory See also Procedural learning 628
of, 156 Skin galvanic response (SGR), in learning rules, 625626
Seyle, Hans, on stress response, 642 orienting reflex habituation, 497 performance rules, 626
Shah, P., 264 498 principles, 624625, 625
Shand, Michael A., 399 Skinner, B. F., 616617, 617 stimulus, 625, 626
Shaping, reinforcement and, 493, behaviorism theory, 57, 58, Songbirds. See Birdsong learning
565, 616 493495 SOP model. See Sometimes
Shaughnessy, J. J., 116, 117 cognitive revolution in reaction opponent process (SOP) model
Shereshevskii (mnemonist), 395396 to his behaviorism, 236 Source amnesia, in episodic
Sherman (chimpanzee), 315 learning theory and, 328 memory, 136
Sherrington, C., 127, 215 occasion setting coined by, 113 Source monitoring, 628630
Shiffrin, R. M., 672673 reinforcement methods, 493 brain regions involved, 630
Shimamura, A. P., 160 494, 563, 564 factors affecting, 630
Shock-shock (S-S) timer, avoidance on response rate, 616 false memories and, 146
learning, 1 Skinner box, 493 historical context, 629630
712 GENERAL INDEX
Source of information Spinal cord, habituation and Hulls view, 236237, 327
confusion in, dj vu associated sensitization in, 225 interval timing, 89, 89, 90
with, 108 Spinal plasticity, 639641 mental processes between
forgetting, in aging, 12 Spinal reflex excitability, 640 stimulus and response, 8889
frontal lobe damage, 159 Spines, dendritic. See Dendritic modality effects, 400
Space-rate code, 444 spines Pavlovian influence on, 518519
Spacing effect, in distributed Spoken traditions. See Oral temporal factors, 518519
repetition, 116 traditions theoretical approaches to, 533,
Sparrows, and song learning, 62, 64, Spontaneous recovery, in 655656
65, 6869, 435 interference and forgetting, 269, Thorndikes view, 326
Spatial learning, 633636 270 Watsons view of, 57
aging animal model, 9 Squire, L. R., 160 Stimulus sampling theory, 361362
Alzheimers disease and, 9 S-R units. See Stimulus-response Stimulus-to-stimulus (S-S)
animal hippocampal size and, (S-R) learning conditioning paradigms, 101
139 Stanford-Binet test, mental Stone, C. P., 227
animals, 631632 retardation and, 381 Storage of memory. See Memory
bees, 273 Startle reflex, 463465 consolidation
blocking of long-term acoustic pathway, 463 Stories. See Prose retention
potentiation and, 347348 fear-potentiated, 463465 Story mnemonics, 393394
brain and, 634 glutamate receptors in, 464 Storytellers. See Oral traditions
frontal lobe damage, 159, 161 habituation and sensitization of, Strabismus, experimental, 427
hierarchical storage, 634 225226 Stream of consciousness, 299
hippocampus and, 9, 472, 489, neural pathways, 464, 465 Strength theory of memory, 578579
529, 531 State-dependent retrieval of Stress and memory, 641643
language and culture, 455 information, 119, 582583, 583t fight or flight, 641
in navigation, 392 Statistical learning algorithms, 15 learned helplessness, 642
nonspatial learning vs., 491492 Statistical learning theory, 361362 long-term potentiation, 389390
orientation dependence, 633, Stellate cells, 185 metaplasticity, 389390
633634 Stephens, D. L., 264 performance and, 641642
radial maze, 488489 Stephens, D. W., 150 Stress hormones, memory
reference systems, 634, 635 Stereotypes, modeling of, 484 consolidation and, 513, 514
route vs. survey knowledge, 633 Stern, William, 417 Striatum
temporally graded retrograde Sternberg, S., 374 amygdalar connections, 187
amnesia and, 370 Steroid hormones, birdsong learning anatomic subdivisions, 192
Speaking-span test, 265, 266 and, 6869 association with substantia nigra
Special education, 321 Stimulus and subthalamic nucleus,
Species, avoidance learning aversive, 495 191192, 192
differences, 2 configurations of, 330 basal ganglia connections, 191,
Specificity principle, encoding, 582 contextual approach to 191
Specific language impairment (SLI), understanding, 330 dopaminergic neurons and
322323 dimension of, 495 blocking, 303
Speech. See Language; Language discriminative, 495 neurophysiology, 576577
learning, in humans; Language preexposure, in classical nucleus accumbens and, 568
learning, in nonhuman primates; conditioning, 7576 reinforcement or reward in
Oral traditions sometimes opponent process, learning, 575577
Speech impairments 625, 626 ventral, 192, 193
prefrontal cortex in, 536 timing and modality, 400 Stroke, dementia associated with,
See also Aphasia See also Conditioned stimulus 110
Spellman, B. A., 271 (CS); Unconditioned stimulus Students. See School learning
Spence, Kenneth, 636639, 637 (US) Studies on Hysteria (Freud and
delivering Silliman Lectures, Stimulus control. See Discrimination Breuer), 155156
637, 638 and generalization A Study in the Psychology of
on discrimination learning in Stimulus generalization Testimony (McGeoch), 362
animals, 637, 638 defined, 113 Study-phase retrieval theory, in
Hulls theories and, 57, 236, Hulls principle of, 360 distributed practice effects, 116
327, 360, 637 Stimulus-instrumental response (CS- Subcortical dementia, 110111, 111t
on Pavlovian conditioned IR) paradigms, 101 Subiculum
response, 519 Stimulus-response (S-R) learning in parahippocampal trisynaptic
Spencer, Herbert, 585 amygdala in, 415416, 545546 circuit, 203204
Spencer, W. A., 223224, 225 basal ganglia in, 544545 projection to entorhinal cortex,
Sperling, George, 607 cognitive learning vs., 533 213
GENERAL INDEX 713
as subdivision of hippocampus, Eccless studies of, 128 Tegmentum, ventral

202 Hermissenda, 278, 279 blocking and, 303
Subjacency theory, of grammar, 313 imprinting changes and, 250 dopamine-containing neurons
Subjective persistence, 607 mechanisms of memory in, 192, 303
Substance P, 234, 515 formation and, 334 Telencephalon (endbrain)
Substantia innominata, 187 metaplasticity, 341 basal ganglia as component of,
Substantia nigra neural change and synapse 191
association with striatum and formation, 406407 neural computation models,
pallidum, 191192, 192 second-order conditioning and, 445449
dopaminergic neurons and 99, 99100 Television, infant memory of, 256
blocking, 303 short-term, 389 257
neuronal groups in, 181 See also Long-term depression Temporal difference algorithms, 17
parts reticulata and pars (LTD); Long-term Temporal discrimination, 114115
compacta, 191192 potentiation (LTP);
Temporal lobe
Subthalamic nucleus, 191192, 192 Metaplasticity
Alzheimers disease and, 18
Successive approximation, 493 Synaptic transmission, control
amnesic syndrome, 414
Suffix effect, 399, 608 mechanisms, 594
anatomy and functions, 200,
modality effect and, 398 Synaptic vesicles, 209, 215, 407
200
Sulcus(i), of cerebral cortex, 200, Synchronization, neuronal, 500504
binocular rivalry, 501, 502 memory consolidation and,
200 369371
Summation test, conditioned EEG patterns, 501
feature binding, 503504 schizophrenia and, 590
inhibition, 75
perceptual grouping, 503, 504 seizure in, dj vu associated
Superior memory performance. See
Synesthesia, 396 with, 108
Experts memories
Syracuse High Avoidance and Temporal lobe, medial
Supervised learning. See Error-
Syracuse Low Avoidance rat memory consolidation and,
correction learning
strains, 3 369371
Supple, W. F., 225
Syrinx, in birdsong learning, 65, 67 neuroimaging of, for memory
Symbolism, in mnemonic learning,
Szentgothai, Jnos, 129 consolidation, 371
393, 394
organic amnesia and, 29
Synapses, 215218
temporally graded retrograde
boutons and, 206, 209, 215, 407 T amnesia and, 370
characterization of, 215, 216 Tacrine, for memory impairment,
dendritic spines, 216, 217218 Temporally graded retrograde
111, 521 amnesia, 368, 369370
electrical-chemical transmission Taglialatela, J. P., 317
controversy, 128 Temporal model of learning (rate
Tail-shock induced inhibition of
excitatory, 215, 216 estimation theory), 331
siphon withdrawal reflex, Aplysia,
glial cell process and, 216, 217 Temporal order information, search
3940
Hebbian, 230, 231, 332, 340 processes in recovery of, 377
Tail-siphon withdrawal reflex,
imprinting changes in, 250 Temporal order of events. See Serial
Aplysia, 42, 401
inhibitory, 215, 216, 217 organization
Tang, A., 122
learning and efficiency in, 407 Terrace, H. S., 315
Tap withdrawal response, C. elegans
locations on neuron, 216, 217 Testosterone (androgen)
(Caenorhabditis elegans), 288
modulatory, 215 Task-specific processing, in working birdsong learning and, 6869
neurotransmitter functions, 215 memory, 265266 memory effects, 234
olfactory cortex, 212 Taste aversion and preference spatial learning and, 614
overview, 181 learning in animals, 645647 Tests of memory. See Measurement
postsynaptic density, 215, 216 amygdalar injection with of memory; Recall tests; Word lists
silent, activation of, 353354 antisense Tetracycline-controlled transactivator
vesicles contained by boutons, oligodeoxynucleotides, 167 system, and LTP regulation, 172,
209, 215 168 354
See also Synaptic plasticity behavioral characteristics, 645 Tetrahydroacridine, with tacrine, for
Synaptic connections, changes in, evaluative conditioning in, 148 Alzheimers disease, 521
404405 gustatory cortex in, 348 Textura del sistema nervioso del hombre
Synaptic plasticity reproductive success related to, y de los vertebrados (Ramn y Cajal),
Aplysia sensitization, 34, 389 138 558
cellular mechanisms of, 388 Teaching. See School learning Teyler, T. J., 344
conventional learning changes Tectum, of midbrain, 180 Thalamus
vs. long-term potentiation, Tegmentum amygdalar projections, 185, 186
346347 neuronal groups in, 181 cortical connections, 200
defined, 387 sparing of, in Parkinsons dorsolateral nucleus (DLM), in
Drosophila, 294 disease, 193 birdsong learning, 64, 66
714 GENERAL INDEX
Thalamus (continued) latent learning experiments, classical conditioning, 74, 101

dorsomedial nucleus, in 653 Unconditioned stimulus (US)
birdsong learning, 65 on learned behavior, 533 Aplysia siphon-gill withdrawal
in forebrain, 180 learning theory of, 327, 328 reflex, 33
mediodorsal nucleus of, Tomoyori, Hideaki (mnemonist), blocking effect and, 98, 100,
connection to olfactory cortex, 395 301302, 302t
210 Topicality effect, 598 classical conditioning, 74, 101,
recent memory and, 333 Tower of Hanoi puzzle, 160, 548 454
Thematic organization, in oral Trace conditioning, eyeblink classical excitatory classical conditioning,
traditions, 496 conditioning and, 7, 8 7475
Theoretical psychology, Spence and, Trace decay theory of forgetting, eyeblink classical conditioning,
637639 152 7, 169170
Theory of distributed associative Trace procedure, excitatory classical in reinforcement, 570572
memory (TODAM), 612 conditioning, 74 taste aversion, 646
Therapeutic modeling (behavior Trans-ACPD (trans-1-amino-1,3- unpaired control procedure,
therapy), as fear-reducing cyclopentanedicarboxylic acid), 102
procedure, 59, 60 176, 178 See also Interstimulus intervals
Thinking (Bartlett), 56 Transcription (ISI)
Thinking, Watsons view of, 667 in memory consolidation, 366 Underwood, Benton, 655657, 656
Thompson, Richard F. 367 associationist perspective, 656
on associative learning, 570 required for long-term memory, 657
on cerebellum in conditioned 173174 functionalist perspective, 655
eyeblink response, 544 Transfer-appropriate processing, 656
on eyeblink classical 246, 583584 McGeoch and, 363
conditioning, 7, 169 Transference, in psychoanalysis, on memory retention, 656, 657
on habituation, 223224, 225 156157 on verbal learning, 656657
Thomson, D. M., 153 Transfer of learning/training Universal grammar, 312314
Thorndike, Edward, 647, 647649 forgetting and, 268 Unlearning and spontaneous
associationism, 648649 Harlows studies, 227 recovery, in interference and
avoidance learning, 2 Transforming growth factor b, in forgetting, 269, 270
on brain evolution, 649 Aplysia sensitization, 43 Unpaired control procedure, 102
characterization of stimuli, 103 Transgenic mice. See Mice, Unsupervised learning, 433
law of disuse, 268, 269 transgenic
Unsupervised learning algorithms,
law of effect, 235, 326, 533, 564, Transient global amnesia, 3132
16
637, 648649 Translation, required for long-term
Urbache-Wiethe disease, 184
law of exercise, 648 memory, 173174
Uvaform (Uva) nucleus, in birdsong
learning theory of, 326, 648 Traumatic Aphasia (Luria), 357
learning, 64, 66
puzzle-box experiment, 648 Treisman, A., 48
on reinforcement, 563564 Triadic design, 319
William James and, 298299 Trisomy 21. See Down syndrome V
Tritonia, habituation and Vanishing cues, 562
Thorndikian conditioning. See
sensitization in, 291292 Vascular dementia, 110
Conditioning, instrumental
True/false tests (single-item Vasopressin
(operant)
recognition tests), 365 as CNS stimulant, 8485
3H-Thymidine, birthdating of
Truly random control, 102 in memory consolidation, 515
birdsong nuclei, 67
Tryon, Robert, 653 memory effects, 233, 234
Time sampling, freezing behavior,
Tubulin, of neurons, 205 memory modulation and, 476
461
Tully, T., 293, 295 as neuropeptide, 182
Tinklepaugh, O. L., 327
Tulving, E., 135, 136, 153, 597, 602 Ventral striatum, 192, 193
Tip-of-the-tongue phenomenon, 11,
Tuning curve, in receptive fields, Verbal-associative learning/memory
650651
422, 423 Alzheimers disease and, 18, 111
Tissue plasminogen activator (TPA),
Turner, M. L., 265 head injury and, 228
transgenic, 174
Tyrosine hydroxylase, 46, 182
TODAM (theory of distributed individual differences, 252
associative memory), 612 McGeochs research, 362363
Tokai High Avoider rat strain, 3 U Underwood on, 655, 656657
Token economy, 61 ber das Gedchtnis (Ebbinghaus), Verbal behavior, Skinners concept
Tolman, Edward C., 651654, 652 127 of, 58
goal-oriented behaviorism, 57 Ubiquitin C-terminal hydrolase, Verbal overshadowing, 146
58, 652 Aplysia (apUCH), 44 Vertebrates
Hull and, 236 Unconditioned response (UR) habituation and sensitization in,
influence on Spence, 639 blocking, 98 223226
GENERAL INDEX 715
higher-order classical nonhuman primates, 540543 on Pavlovian conditioned

conditioning, 98 occipitotemporal pathway response, 519
morphological basis of learning lesions and, 542 Waugh, N., 49
and memory, 404408 retrosplenial cortex in, 213 Weber-Fechner law, 411
protein synthesis in long-term Visual system, 425429 Wechsler Adult Intelligence Scale
memory, 553556 binocularity, 426, 427 (WAIS), 267, 381
Vesicles, synaptic, 209, 215, 407 cortical neurons, 426 Wechsler Memory Scale (WMS), 267
Vestibulo-ocular reflex (VOR) development of, 425429 Weinberg, J., 122
plasticity, 659661 disparity detectors, 427 Weiner, I., 123
cerebellar long-term depression neocortical plasticity, 426429 Weingartner, H., 134
and, 170 neurochemistry, 428 Wernicke-Korsakoff syndrome. See
neural computation and, 433 nonhuman primates, 540541, Korsakoffs syndrome
neuronal circuits, 659660 541 Wernickes area, 200
velocity storage, 660 ocular dominance, 421, 426 White matter
Video games, dreaming and, 621 orientation and, 427 of brain, electrical stimulation,
Vigotsky, Lev, 357, 396 sleep and cortical plasticity, 429 571572
Vinpocetine, as cognitive enhancer, strabismus, 427 of cerebral cortex, 200
85 transgenic mice models, 428 Whitman, Charles O., 356
Virchow, Rudolf, 557 visual experience and Widrow, B., 15
Visible persistence, 607 development, 426427 Williams, J. M. G., 133
Visual attention Vitamin E, for cognitive Williams syndrome, 383384
nonhuman primates, 543 improvement, 522 Witness testimony. See Eyewitness
primates, 539540 Vocal communication. See Aphasia; testimony
Visual cortex, adult Birdsong learning; Language; Woike, B., 53
anatomy, 540541 Language learning, in humans; Wolpe, J., 60
cortical lesions and, 421 Language learning, in nonhuman Wong, R., 122
long-term depression, 420 primates; Oral traditions; Verbal- Word chaining, 611
long-term potentiation, 420 associative learning/memory Word-fragment completion test, 11,
neocortical plasticity and, 420 Voeks, V. W., 220 365
422 Voles, spatial abilities, 139 Word lists
neural computation, 442443 Voluntary behavior. See Operant distributed practice effects, 115
neurons, 426, 541 behavior 116
retinal lesions and, 420421 VP (mnemonist), 396 false memory studies, 145146
visual memory storage, 663 Vygotsky, Lev. See Vigotsky, Lev interference and forgetting, 269,
Visual experience 271
neocortical development and, last item as cue for next item,
426, 428 W 377
orientation and, 427 Wagner, A., 16, 302 memory search and, 373379
synaptic connections and, 404 See also Rescorla-Wagner model/ memory span and, 380
405 rule organization of, for memorizing,
Visual field defects, cortical lesions Walden Two (Skinner), 58, 617 82
and, 421 Wallace, Alfred Russell, 645 part-list cuing, 271
Visual memory Warner, L. H., 2 recency and memory search
brightness and flux in, 663664 Warning signal (WS), avoidance processes, 374376, 376, 377
eidetic imagery, 130132 training, 1 repetition and learning, 579
infant visual recognition Warning signal (WS)-shock interval, unlearning and spontaneous
memory, 254255 1 recovery, 269
nonhuman primates, 540543 Washoe (chimpanzee), 315 Words
visual cortex and, 663 Watson, John B., 665668, 666 Alzheimers disease and recall
See also Imagery; Sensory advertising career, 668 of, 18, 111
memory; Visual memory, avoidance learning, 2 concrete, dual verbal and
brightness and flux in; Visual behaviorism, 666668 imaginal coding, 79
object agnosia bird studies, 666 lexicon acquisition, 313314
Visual memory, brightness and flux on child rearing, 668 Word-stem completion test, 11, 243
in, 663664 human learning research, 667 244, 366
Visual object agnosia, 306 668 Working memory, 669676
Visual perception, 540543 Hunter and, 238 aging and, 13, 74
adaptation aftereffects, 421 influence of, 533, 668 in animals, 669671
attention and, in primates, 539 influences on, 666667 Atkinson/Shiffrin model, 672
540 Lashley and, 317318 673
Mller on, 412 learning theory of, 326 attention and, 4950
716 GENERAL INDEX
Working memory (continued) Olton on, 488492 Y

Baddeley/Hitch model, 673, prefrontal cortex in, 674
Yerkes, Robert M., 637, 666
673674 processing system in, 264, 264
radial maze and, 488490, 670 Yes/no single-item recognition tests,
children, assessment of, 7374
recall in, 380, 444 365
defined, 13, 49, 490, 669
in dementia, 111 short-term memory vs., 672673 Yoked-control design, 103, 259
frontal lobes and working-with task-specific processing and
memory, 157158, 158 intelligence, 265266, 266
hippocampus and, 489491 tests of, 265266, 266
in humans, 672676 time factors, 670671 Z
visual and verbal components, Zebra finches, and song learning,
individual capacity differences,
79 67, 69
252
Wynne, L. C., 2 Zechmeister, E. B., 116, 117
intelligence and, 264, 264, 265
267, 266 Zif268 transcription factor, CREB
mechanisms, 669670, 674676, X regulation of, 174, 367
675 X area, in birdsong learning, 66, 67 Ziv-Harris, D., 123

Byrne - Learning & Memory PDF

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Byrne - Learning & Memory PDF

Transféré par

Droits d'auteur :

Formats disponibles

Learning

Henry L. Roediger, III

MACMILLAN REFERENCE USA

Jill Lectka, Director of Publishing Operations

John H. Byrne, Editor in Chief

LIBRARY OF CONGRESS CATALOG-IN-PUBLICATION DATA

Learning & Memory / edited by John H. Byrne. - - 2nd ed.

BF318 .E53 2002

Printed in the United States of America

Learning and Memory, Second Edition ..........................................1

IMAGING AND MULTIMEDIA

Active and Passive Avoidance DEVELOPMENT OF PROCESSES LEVELS OF PROCESSING

Activity-Dependent Regulation of MOLECULAR BASIS OF LONG-TERM Roger W. Schvaneveldt

Guthrie, Edwin R. (18861959) HERMISSENDA

Morris Moscovitch Head Injury William N. Frost

Knowledge Systems and Material- Mathematical Learning Theory Multiple-Memory Systems

MAINTENANCE and Memory Joseph E. Steinmetz

Neurogenesis Prefrontal Cortex and Memory in Savant Syndrome

Taste Aversion and Preference Underwood, Benton (19151994) Working Memory

Charles I. Abramson Shaowen Bao Diego E. Berman

Thomas H. Brown Genetic Substrates of McGeoch, John A.

W. K. Estes Bennett G. Galef, Jr. Ann M. Graybiel

Alice F. Healy Masao Ito Yuri Koutcherov

Fred D. Lorenzetti James L. McGaugh Neil W. Mulligan

Steven E. Petersen Mark R. Rosenzweig James H. Schwartz

Brian H. Smith Rodney A. Swain Almira Vazdarjanova

Gordon Winocur Brain: Hippocampus and Anthony A. Wright

ACTIVITY-DEPENDENT REGULATION equilibrium with choline acetyltransferase (CAT), the

Bibliography formance on the basis of its own experience. Also

XB(t) = W1(t)X1(t) + . . . + Wn(t)Xn(t). (1)

firing neuron B at a time t is often represented by the Figure 1

Wi(t + t) = Wi(t) + cXB(t)Xi(t), (2)

where t is a small time increment and c is a small pos-

This means that when the elements output, X(t),

puts and comparing the resulting evaluations. This is ALZHEIMERS DISEASE

Bibliography Amnesia in the Dissociative Disorders

Trauma, Repression, and Dissociation

Paradoxically, the idea that trauma plays a role in Bibliography

Figure 1 dependent protein kinase; the subsequent protein

Behavioral Studies Figure 2

Second Messengers and Protein Kinase

training or extended 5-HT treatment results in a pro- Immediate Early Genes

Although there are other sites of neural plasticity,

Attention and Long-Term Memory Attention and Implicit Memory

See also: LANGUAGE LEARNING: HUMANS; WORKING ATTENTION DEFICIT DISORDER

Autobiographical Memory and Personality erativity. In contrast, those participants without a

always be at the mercy of old habits; but with a schema

Effects of Isolation and Deafness Song Overproduction, Social Interactions,

then to nXIIts constitutes a motor pathway special-

A Motor Pathway for Song

Figure 1 that they are more likely to be able to provide a code

Cue Competition There are several ways to demonstrate timing of

lieved to play an important role in everyday human CODING PROCESSES

Figure 1 required for subjects to scan from one location to an-

Bibliography such as whether the word is a synonym for a specified

ORGANIZATION OF MEMORY at least one fundamental process of organization. As-

3. collective memory presupposes an unchanging James V. Wertsch

An information-processing model of temporal processing representative of scalar timing theory.

goal site in relationship to geometrical properties than configural-landmark navigation (Shettleworth,

CONDITIONING, CELLULAR AND generally activated by the US or whether qualitatively

production of a CR but also causes active inhibition Figure 1

Blocking See also: APLYSIA: CLASSICAL CONDITIONING AND

Instrumental Conditioning when the experimental subject performs the target