Académique Documents
Professionnel Documents
Culture Documents
The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elseviers archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
a r t i c l e
i n f o
Article history:
Received 21 April 2009
Revised 8 May 2009
Accepted 11 May 2009
Available online 13 June 2009
Keywords:
Temporal lobe epilepsy
Memory
Emotion
a b s t r a c t
In healthy humans, memory for words with emotional valence is better than memory for neutral words.
At the same time, the word preceding the emotional word in a word list learning task is remembered less
often than other neutral words. Both effects, enhanced memory for emotional words and retrograde
amnesia for preceding words, are dependent on intactness of the amygdala. In this study we asked
whether patients with temporal lobe epilepsy (TLE), a disease that often involves the amygdala as well,
show altered memory for emotional words and the words presented in close temporal proximity.
Whereas we found enhanced memory performance for emotional and decreased recognition performance for the preceding and successive word in our 19 control subjects, both effects were strongly
reduced in our 21 patients. No group differences occurred in memory for perceptually deviant words.
The lack of emotion effects on memory in the patients cannot simply be attributed to altered perception
of emotions as the patients rated the emotionality of the words no different than control subjects. Hence,
we conclude that patients with TLE have a specic decit in the emotion-driven encoding enhancement
mediated by the amygdalahippocampus loop.
2009 Elsevier Inc. All rights reserved.
1. Introduction
It is an everyday experience that we remember outstanding,
emotionally arousing events better than routine situations. In the
psychology laboratory, it was demonstrated as early as 1933 [1]
that perceptually deviant words (also called oddballs) are remembered better than those presented in the standard format (von
Restorff effect). Oddballs conveying a strong, usually aversive emotional content lead to even better recollection performance [2]. At
the same time they impair memory for words that have been presented just before the emotional oddball. This has been taken as
evidence that emotions interfere with the process of encoding
and/or consolidation of preceding events. It could be demonstrated
that these effects are indeed emotion specic and are not simply a
more pronounced von Restorff effect: When Strange and colleagues blocked the aymgdalas cholinergic output by administering the b-adrenergic antagonist propranolol to their subjects, the
emotion-related effects (enhancement for the emotional word,
anterograde amnesia for the preceding word) were abolished,
whereas the von Restorff effect for the perceptually deviant word
* Corresponding author. Address: Clinic for Neurology, Otto-von-Guericke-University, Leipziger Strasse 44, Magdeburg 39120, Germany. Fax: +49 (0)391
6715233.
E-mail address: notger.mueller@med.ovgu.de (N.G. Mller).
1525-5050/$ - see front matter 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.yebeh.2009.05.006
remained unaffected [3]. Hence, only the emotion effect on memory depends on integrity of the amygdalae, which are known to
play a key role in emotion processing. This assumption is further
supported by the nding that a patient with bilateral amygdala lesions behaved like pharmacologically treated subjects, that is, did
not show any emotion-related memory alteration [36]. The
amygdala is intensively connected with other structures in the
mesiotemporal lobe, including the hippocampus, which is critical
for encoding and recollection within episodic memory [7]. Regarding enhanced memory for emotional content, it has been speculated that the amygdala releases noradrenergic neurotransmitters
in response to emotional input which, via connections to the hippocampus, enhances encoding of emotional stimuli into episodic
memory by the hippocampus [3,810].
Tulving was the rst to show that emotional arousal also impairs memory for preceding events [11] and he speculated that this
was due to a premature termination of the encoding process
induced by the emotional event. Later it was speculated that this
effect is driven by the amygdalas release of neurotransmitter in response to emotional stimuli which may disrupt either the encoding
or consolidation of preceding items into memory [3].
Temporal lobe epilepsy (TLE) usually involves neural changes in
the hippocampus and surrounding areas, including to a variable
degree the amygdala, and is often accompanied by hippocampal
sclerosis [6,12,13]. Memory decits in these patients are a common
368
Table 1
Characteristics of patients.
Patient
Sex
Age
Years of
education
Hippocampal
sclerosis
Interictal
focus
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
M
F
M
F
F
F
f
M
F
F
M
F
F
M
M
F
F
M
F
F
M
32
32
23
50
61
34
37
54
28
35
67
22
43
31
34
38
53
33
38
47
35
12
11.5
17
12.5
12
14
21
18.5
10
20
16.5
15
9
12
21
17.5
15
18
22
12
19.5
No
Yes
Yes
No
No
No
Yes
Yes
No
No
No
Yes
No
No
No
No
No
Yes
Yes
Yes
No
Bilat
Left
Right
Left
Left
Unknown
Right
Left
Unknown
Right
Right
Right
Unknown
Left
Left
Left
Left
Right
Left
Left
Left
369
Fig. 1. Schematic experimental design. Note that no colors were used in the actual experiment.
the next block the starting number was one less, that is, 89 and so
forth, to prevent the numbers from being memorized.
After the distraction task, a list of 16 nouns was presented to
test recognition performance. Subjects had to indicate by a mouse
button press (counterbalanced across subjects) whether they believed that the presented word had been part of the preceding
learning list or not. They were told that reaction speed also mattered. Each word remained on the screen until button press or after
4 s had passed without a button press. Of the 16 nouns, 8 were old
(E, P, E 1, E + 1, P 1, P + 1, C1, C2). The eight new words were
those remaining from the initial 24-word list. Sequence of stimuli
was random, and they were all presented in the same standard font
(Times New Roman 95).
About a month after the patients had completed the experiment
they were sent a letter in which they were asked to rate the emotionality of the words in the same manner as described before for
the controls. This was done to ensure that the patients did not differ in rating of emotional valence per se.
2.4. Analysis
3.4. Oddball effects in controls
The main question of the study was whether the oddball nouns
and their predecessors and successors (hence called nouns of interest) would be remembered more often than neutral control words.
To do so, a factor was calculated that indicated whether the word
of interest was remembered more or less often than control words,
where a value of 1 means that the specic noun (e.g., E-word) was
remembered as often as the control words, values >1 indicate
superior memory performance, and values <1 indicated inferior
memory performance for the noun of interest compared with control words. To calculate this value, for each word of interest the
absolute number of correctly remembered words was divided by
half of the number of remembered control words (as there were
a total of 20 control words, but only 10 in the other categories).
If, for example a subject remembered 8 of the 10 E-words and 14
of the 20 C-words, this would yield a factor of 8/(14/2) = 1.14, indicating that E-words were remembered 1.14 times more often than
control words.
With these factors a t test was calculated across all subjects of a
group that tested whether the calculated factor for a noun of interest differed from 1 within a 95% condence interval.
Group comparisons involved an ANOVA with the factors oddball
type (emotional, perceptual), position relative to the oddball ( 1,
0, +1), and group. Because of the small sample sizes, subgroup
analyses were performed nonparametrically with two-sample KolmogorovSmirnov tests. All statistical analyses were performed
with SPSS software.
Fig. 2 shows that control subjects recognized the E-words signicantly more often (factor of 1.16) than control words; the words
preceding and following the E-word, however, were remembered
less often (E 1: 0.88, E + 1: 0.85). All effects reached signicance
at P < 0.005. Although P-words were also remembered more often
(1.04) and surrounding words less often (P 1: 0.96, P + 1: 0.98),
none of these effects reached signicance (P > 0.2 for P 1,
P > 0.6 for P, P > 0.5 for P + 1).
A type (emotional, perceptual oddball) position ( 1, 0, +1)
ANOVA revealed a main effect for position (F[2, 36] = 18.61,
P < 0.005) and the type position interaction (F[2, 36] = 6.83,
P < 0.05), conrming that only emotional and nearby nouns were
remembered better or worse, respectively, compared with neutral
control nouns.
3.5. Oddball effects in patients
Patients did not remember E-words signicantly better than Cwords (factor 1.05, t = 1.34, P = 0.19) (see Fig. 3). Memory for the
3. Results
3.1. Emotional rating
The emotional words were not rated differently by the patients
relative to the normal controls who had participated in our pilot
study (t = 0.16, P > 0.87; note, however, that only 14 of the 21 patients had completed the questionnaire).
3.2. Memory for control nouns
We rst assessed memory for the neutral control nouns to reveal a possible general memory decit. As could be expected from
370
group, namely, enhanced memory for emotional words and retroand anterograde amnesia for preceding and subsequent words,
respectively. The lack of an emotion effect was independent of
the specic form of TLE.
Taken together, the results in our patient group are very similar
to those previously seen in healthy subjects who had received a
drug that interfered with the amygdala-driven noradrenergic response to emotional stimuli and those seen in a patient with bilateral sclerosis of the amygdalae [3]. As in the present study, in this
earlier work no effects on memory for perceptually deviant oddball
words were induced by amygdala dysfunction. Our patients memory for the perceptual deviant word was comparable to that of controls; so was their memory for control words. Also, they did not
rate the emotional valence of the words differently from the controls, speaking against a general decit in emotion processing. This
is in accordance with previous studies that have shown normal
skin conductance responses to emotional stimuli in patients with
amygdala damage [16]. As emotional memory critically relies on
amygdalahippocampus interaction [6,9,10], our results lead us
to speculate that this interaction is disturbed in patients with
TLE. In demonstrating that emotional memory is altered even in
unoperated patients, this nding extends earlier results on emotional memory in patients after unilateral temporal lobectomy
[17].
Patients with TLE suffer from structural damage within temporal regions that to a varying degree can also affect the amygdalae
[12,18]. To what extent the amygdalae were structurally affected
in our patient group could not be reliably assessed from their
MRI scans (often we had to rely on written reports) and no volumetric analyses were performed. Nevertheless, the behavioral deficit observed in emotion memory indicates that there is at least a
functional decit in the amygdalahippocampus interaction during memory formation for emotional information.
This study had been designed to closely match earlier studies on
emotion-related memory effects [3,10]. Although we replicated
many of the ndings of these earlier studies, some ndings differ
and came as a surprise: First, we did not observe a von Restorff effect for the perceptually deviant word. Apart from issues of power
we believe this was due to the rather subtle perceptual changes
introduced by the font changes and the limited number of deviant
fonts we used so that they had to be presented repeatedly. The latter may have reduced the novelty effect of the P-word, and it has
been shown that novelty is associated with memory improvement
as well [19]. Thus, one can speculate that if we had used stronger
and more frequent perceptual changes, we would have observed
the usual von Restorff effect. Most crucial, however, there was no
difference between patients and controls in memory for the Pwords.
Another somewhat unexpected nding was the reduced recognition performance in controls for words following the emotional
word in the sense of anterograde amnesia. Tulving originally reported a mere retrograde amnesia induced by emotional content
[11]. However, other studies have reported anterograde amnesia
as well, suggesting this effect to be susceptible to the specic
experimental design. Hurlemann and colleagues found that emotional arousal rather than valence correlates with anterograde
amnesia so that both stimulus types, those with negative and those
with positive content, can cause anterograde amnesia as long as
they induce high arousal [20]. We cannot test this hypothesis
based on our data as we neither assessed arousal nor had positive
words, yet it is easily conceivable that the negative words induced
more arousal than neutral words and hence interfered with encoding of the subsequent word.
In this study, we did not observe a signicant difference in
memory performance between patients with a left- and those with
a right-sided epileptic focus. A verbal memory decit is more often
371
Fig. 4. Reaction time data for the control (left) and patient (right) groups.
[6] Richardson MP, Strange BA, Dolan RJ. Encoding of emotional memories
depends on amygdala and hippocampus and their interactions. Nat Neurosci
2004;7:27885.
[7] Pitknen A, Pikkarainen M, Nurminen N, Ylinen A. Reciprocal connections
between the amygdala and the hippocampal formation, perirhinal cortex, and
postrhinal cortex in rat: a review. Ann NY Acad Sci 2000;911:36991.
[8] Cahill L, Prins B, Weber M, McGaugh JL. Beta-adrenergic activation and
memory for emotional events. Nature 1994;371:7024.
[9] Phelps EA. Human emotion and memory: interactions of the amygdala and
hippocampal complex. Curr Opin Neurobiol 2004;14:198202.
[10] Strange BA, Dolan RJ. Beta-adrenergic modulation of emotional memoryevoked human amygdala and hippocampal responses. Proc Natl Acad Sci USA
2004;101:114548.
[11] Tulving E. Retrograde amnesia in free recall. Science 1969;164:
8890.
[12] Bernasconi N, Bernasconi A, Caramanos Z, Antel SB, Andermann F, Arnold DL.
Mesial temporal damage in temporal lobe epilepsy: a volumetric MRI study of
the hippocampus, amygdala and parahippocampal region. Brain 2003;126(Pt.
2):4629.
[13] Bonilha L, Kobayashi E, Rorden C, Cendes F, Li LM. Medial temporal lobe
atrophy in patients with refractory temporal lobe epilepsy. J Neurol Neurosurg
Psychiatry 2003;74:162730.
[14] Hermann BP, Seidenberg M, Schoenfeld J, Davies K. Neuropsychological
characteristics of the syndrome of mesial temporal lobe epilepsy. Arch
Neurol 1997;54:36976.
[15] Benuzzi F, Meletti S, Zamboni G, et al. Impaired fear processing in right mesial
temporal sclerosis: a fMRI study. Brain Res Bull 2004;63:26981.
[16] Tranel D, Damasio H. Intact electrodermal skin conductance responses after
bilateral amygdala damage. Neuropsychologia 1989;27:38190.
[17] LaBar KS, LeDoux JE, Spencer DD, Phelps EA. Impaired fear conditioning
following unilateral temporal lobectomy in humans. J Neurosci
1995;15:684655.
[18] Bonilha L, Rorden C, Kobayashi E, et al. Voxel based morphometry study of
partial epilepsies. Arq Neuropsiquiatr 2003;61(Suppl. 1):937.
[19] Bunzeck N, Dzel E. Absolute coding of stimulus novelty in the human
substantia nigra/VTA. Neuron 2006;51:36979.
[20] Hurlemann R, Hawellek B, Matusch A, et al. Noradrenergic modulation of
emotion-induced forgetting and remembering. J Neurosci 2005;25:
63439.