Hearing Research 274 (2011) 40e47

Contents lists available at ScienceDirect

Hearing Research
journal homepage: www.elsevier.com/locate/heares

Triadic synaptic interactions of large corticothalamic terminals in non-lemniscal

thalamic nuclei of the cat auditory system
Hisayuki Ojima a, b, *, Kunio Murakami b
a
Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo, Tokyo 113-8749, Japan
b
Department of Anatomy, Toho University School of Medicine, 5-21-16 Ohmori-Nishi, Ohta, Tokyo 143-8540, Japan

a r t i c l e i n f o a b s t r a c t

Article history: Large corticothalamic (CT) terminals, presumed to originate from cortical layer 5 pyramidal cells, are
Received 2 February 2010 distributed predominantly in non-specific thalamic nuclei in mammals. In the auditory system, little is
Received in revised form known about whether these CT projections participate in the synaptic aggregation referred to as the
6 May 2010
triad. We studied synaptic interactions of these terminals with neuronal elements in one of the auditory
Accepted 18 May 2010
Available online 26 May 2010
non-lemniscal thalamic nuclei, the dorsal nucleus of the medial geniculate complex (MGC), in cats. After
injections of an anterograde tracer in the primary auditory cortex, areas containing labeled large
terminals were examined using an electron microscope. It was revealed that a fraction of large CT
terminals participated in complicated synaptic arrangements: labeled terminals making synaptic
contacts with vesicle-free dendrites, probably of thalamic principal neurons, and/or vesicle-filled
neuronal profiles, probably of presynaptic dendrites (PSDs) of interneurons. In reconstructions or even in
single sections, we found that these synaptic connections participated in triadic arrangements. Thus,
PSDs postsynaptic to the labeled CT terminals were in turn presynaptic to the vesicle-free dendrites.
Ó 2010 Elsevier B.V. All rights reserved.

1. Introduction
The classical view of the role of the thalamus is currently
changing (see Jones, 2007 for review). It is becoming increasingly
important to consider that the thalamus is not only a relay station
transferring sensory information from the periphery to the cortex,
but also receives information already processed in cortical areas
and transmits it to adjacent cortical areas that are higher in the
hierarchical progression (Guillery, 1995). This transthalamic
pathway is mediated by descending axons of cortical layer 5
pyramidal neurons (Ojima, 1994; Bourassa et al., 1995; Bourassa
and Deschênes, 1995). It is different from the traditional cortico-
thalamic (CT) projection that originates from layer 6 and terminates
chiefly on distal parts of thalamic relay neurons (Guillery, 1969;
Abbreviations: AI, primary auditory cortex; CD, central dendrite; CG, colliculo-
geniculate; CT, corticothalamic; GABA, g-aminobutyric acid; IC, inferior colliculus;
MGC, medial geniculate complex; dMGC, dorsal nucleus of MGC; PHA-L, phaseolus
vulgaris leucoagglutinin; PB, phosphate buffer; PSD, presynaptic dendrite.
* Corresponding author. Graduate School of Medical and Dental Sciences, Tokyo
Medical and Dental University, 1-5-45 Yushima, Bunkyo, Tokyo, 113-8749 Japan.
Tel.: þ81 3 5803 5445; fax: þ81 3 5803 0186.
E-mail addresses: yojima.cnb@tmd.ac.jpè(H. Ojima), kunim@med.toho-u.ac.
jpè(K. Murakami).
Jones and Powell, 1969b). Thus, the cortex has two components
in terms of projection to the thalamus. In a simplified scheme, with
primary sensory cortex as an example of an area containing cells of
origin, small domains in each primary sensory cortex project back
to the corresponding principal (or specific) thalamic nucleus, with
axons originating in layer 6, and to non-specific thalamic nuclei
with axons originating in layer 5. In the former, reciprocal projec-
tions between cortex and thalamus are well preserved, but this is
not the case in the latter (Van Horn and Sherman, 2004; Llano and
Sherman, 2008; but see Huppé-Gourgues et al., 2006). Conse-
quently, as a whole, the entire CT projection originating from
a single cortical locus ends in thalamic nuclei in a divergent manner
(Guillery et al., 2001; Winer et al., 2001).
Target thalamic nuclei of large CT terminals include almost all
non-principal nuclei (see Rouiller and Welker, 2000 for review),
such as the pulvinar and lateral-posterior nucleus in the visual
system, the posterior nucleus in the somatosensory system, the
dorsal nucleus of the medial geniculate complex (dMGC) in the
auditory system, and others. The morphological characteristics of
the two distinct types of descending CT projections have been
characterized by terminal size, aggregation pattern of multiple
terminals, and the extent of their distribution in major sensory
systems of various animal species (Robson and Hall, 1977; Ogren
and Hendrickson, 1979; Hoogland et al., 1987; Rouiller and de

0378-5955/$ e see front matter Ó 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.heares.2010.05.009
 H. Ojima, K. Murakami / Hearing Research 274 (2011) 40e47
Ribaupierre, 1990; Rouiller and Welker, 1991; Rouiller et al., 1991;
Kuroda et al., 1993; Ojima, 1994; Rockland, 1994; Bourassa and
Deschênes, 1995; Bourassa et al., 1995; Bajo et al., 1995;
Vidnyánszky et al., 1996; Ojima et al., 1996; Winer et al., 1999;
Guillery et al., 2001; Li et al., 2003b; Rouiller and Durif, 2004;
Hazama et al., 2004; Kimura et al., 2005; Zikopoulos and Barbas,
2006), and also for the motor system (Rouiller et al., 1998; Kakei
et al., 2001).
Findings that layer 5-derived CT terminals are larger than layer
6-derived CT terminals have been confirmed in most cortical areas
in many animal species. Large CT terminals occasionally end with
the clustering of multiple boutons, giving an impression of a “bunch
of grapes” in cats (see Fig. 5 of Ojima, 1994 as example), or they end
in giant swellings in rodents (Hoogland et al., 1987, 1991; Bartlett
et al., 2000; Li et al., 2003b) and monkey (Rouiller et al., 1998),
occasionally forming aggregations connected by a short axon.
The ultrastructure of the large CT terminals has been investi-
gated with special reference to their resemblance to sensory
ascending afferents and the complexity of synaptic formations they
made. Sensory afferents projecting to thalamus from the periphery,
for example, the retina and trigeminal nucleus (Szentágothai, 1963;
Guillery, 1969; Jones and Powell, 1969a) end with terminal boutons
that participate in distinct synaptic interactions. These involve, in
cat and monkey, three elements: presynaptic dendrites (PSDs) in
interneurons, an afferent terminal, and dendrite(s) of thalamic
principal neurons. The synaptic arrangement formed by these
elements is traditionally referred to as a triadic arrangement,
synaptic aggregation, or glomerulus if they are encapsulated
together by glial lamellae (see Jones, 2007 for review). It is
proposed that large CT terminals derived from layer 5, despite
descending in the direction of the projection, form synaptic
arrangements similar to those found for ascending sensory affer-
ents in primary cortices (Hoogland et al., 1991; Schwartz et al.,
1991; Vidnyánszky et al., 1996; Feig and Harting, 1998: Van Horn
and Sherman, 2004).
In the auditory system, ultrastructural studies have been limited
to rats (Bartlett et al., 2000). Reconstructions from serial sections
have not been done. Since the rodent is characterized by very rare
occurrences of GABAergic interneurons in the sensory thalamus
(with the exception of the dorsal lateral geniculate nucleus)
(Thompson et al., 1985; Winer and Larue, 1988, 1996; Arcelli et al.,
1997), triadic synaptic arrangements would not be expected in the
auditory thalamus. Indeed, it was shown that large CT terminals do
not form triadelike synaptic interactions in the rat. Their (exclu-
sive) targets in the rat dMGC are the dendrites of principal neurons,
although these are occasionally enwrapped partially by glial
sheaths. Thus, it is still not known if large CT terminals are involved
in triadic synaptic arrangements in the auditory thalamus of non-
rodent species. This has been shown in other sensory systems. In
this report we examined the synaptic organization of non-
lemniscal large CT terminals grouped in small clusters that were
not contaminated by small, drumstick-like shaped terminals
derived from, presumably, layer 6 CT axons. Some of the labeled
large CT terminals were subjected to reconstruction from serial
sections.
2. Materials and methods
2.1. Animals
Two cats (male; 3.4 and 4.0 kg; Saitama Experimental Animal)
were used for electron microscopic examination. In addition, light
microscopy drawings (see Fig. 1a) were obtained from serial
sections used previously to reconstruct 3 dimensional distributions
of CT terminals across the cat MGC (Takayanagi and Ojima, 2006).
41
Fig. 1. a; Drawing of the distribution of small (small black dots) and large CT terminals
(large grey dots) in the cat thalamic nuclei after anterograde tracer, PHA-L, injections at
2 loci (dots in Fig. 1c) in the cat AI (8 kHz and 15 kHz for best frequency). b; Light
micrograph of an EPON embedded section to be cut serially that was obtained in
another animal at the site comparable to that indicated by the arrow in Fig. 1a. Note
that it contains several isolated large terminals as well as a cluster of large terminals
that are connected by a common preterminal axon. c; Drawing of sulcal pattern of the
cat left hemisphere, with injection sites indicated by two dots. Ds, superficial dorsal
nucleus of the MGC; Dd, deep dorsal nucleus of the MGC; M, medial nucleus of the
MGC; Vl, lateral nucleus of the MGC; Vo, ovoidal nucleus of the MGC; VL, ventrolateral
nucleus of the MGC, Sg, suprageniculate nucleus; BIC, brachium of inferior colliculus;
aes and pes, anterior and posterior ectosylvian sulci. Calibration bars in a, 1.0 mm and
in b, 20 mm.
Animals were subjected to surgery and care under the approval of
the Animal Care Committee of Toho University, in accordance with
the National Institutes of Health Guide for the Care and Use of
Laboratory Animals (NIH Publications No. 80e23).
2.2. Surgery
After being deeply anesthetized with Nembutal (sodium
pentobarbital, Dainippon Pharmaceutical, Osaka, Japan; 40 mg/kg),
animals were subjected to mapping and tracer injections under
aseptic conditions. Heart rate and body temperature were moni-
tored throughout. Atropine (Tanabe Pharmaceutical, Osaka, Japan;
0.2 mg/kg, s.c.) and dexamethasone (Banyu Pharmaceutical, Tokyo,
Japan; 0.25 mg/kg, i.m.) were applied during the experiment when
necessary. An antibiotic (Cefmetazon, Sankyo Pharmaceutical,
Tokyo, Japan; 50 mg i.m.) and an analgesic (Menamin, Chugai
Pharmaceutical, Tokyo, Japan; 0.5e1.0 mg/kg, i.m.) were adminis-
tered immediately after completion of the injection of the tracer,
and also for the following days.
2.3. Electrophysiology
Recording procedures were identical to those previously
described (Ojima and Murakami, 2002). Briefly, animals were set
into a stereotaxic apparatus (Narishige, Tokyo, Japan) following
Nembutal anesthesia, and a small opening on the left skull above
the middle ectosylvian gyrus was made. The right ear bar was
then replaced by a specially designed clamp bar that supported
the zygomatic arch, thus making the ear canal accessible to an
earphone. Recording of unit activities at 3e4 points were made
with a tungsten electrode (AM systems, USA; 1 MU impedance
at 1 kHz) for defined frequency response curves using
42 H. Ojima, K. Murakami / Hearing Research 274 (2011) 40e47

Fig. 2. Electron micrographs of glomeruli and glomerulus-like structures observed in the dMGC in the cat. In Fig. 2a showing a typical glomerulus, a central dendrite (CD),
presumably derived from a thalamic neuron, is surrounded by 9 profiles of neuronal elements, which are altogether encapsulated by glial sheets (g). It is obvious that one of the
profiles (p1) is filled with densely packed synaptic vesicles. This profile makes asymmetric synaptic contacts with the CD (black arrowhead, enlarged in 2c). The same profile also
makes asymmetric synaptic contacts (grey arrowhead, enlarged in 2c) upon one of the adjoining profiles (p2) containing less densely packed synaptic vesicles. Such profiles,
presumably derived from thalamic interneurons, make symmetric synaptic contacts with the CD (grey arrows, enlarged in Fig. 2d). Many profiles have puncta adherentia (open
arrowhead, enlarged in 2e) between different profiles. In Fig. 2b, a large terminal (CT), which contains a small clump of reaction product (asterisk), thus indicating cortical origin,
makes an asymmetric synaptic contact (black arrowhead) with a dendrite (CD) and an adjoining vesicle-containing profile (grey arrowhead, p2). This partial labeling reveals how
densely synaptic vesicles are packed in large CT terminals. The CD is also postsynaptic (grey arrow) to a terminal profile (p1) that is filled with less densely packed synaptic vesicles,
some of which are oval or pleomorphic in shape. Calibration bars in a and b, 500 nm.

a combination of sound intensities and frequencies, and a partial
frequency axis across the exposed primary auditory cortex (AI)
was constructed. An anterograde tracer, phaseolus vulgaris leu-
coagglutinin, PHA-L (Vector Laboratory, USA; 2.5% in phosphate
buffer at pH 7.2), was loaded into a glass micropipette (outer tip
diameter, 30 mm), then iontophoretically injected at a depth of
1400 mm from the cortical surface with 6 mA positive current
every 7 s for a total of 20 min.
Tone stimuli (5-ms rise and fall times, 50 ms duration, 1.1 s
interval) were generated by a Macintosh platform-based software
(MALlab, Kaiser Instrument, USA), and delivered to a dynamic
headphone driver (DT-47 Beyerdynamics, Heilbronn, Germany)
enclosed in a specially designed earphone. Best frequencies were
defined at a sound intensity 10e20 dB above the threshold.
2.4. Histology and electron microscopy
Histological and electron microscopy procedures were
described previously (Ohyama and Ojima, 1997; Takayanagi and
Ojima, 2006). Two weeks after injection, cats, deeply anes-
thetized with Nembutal (50 mg/kg, i.p.), were perfused with
a mixture of 4% paraformaldehyde and 0.1% glutaraldehyde in 0.1 M
phosphate buffer (PB; pH 7.2). For confirmation of injection site
using light microscopy, brains were permeated with 20% sucrose in
PB, and brain blocks including auditory cortex were cut at a thick-
ness of 50 mm on a freezing microtome. Sections were exposed to
a biotinylated antibody against PHA-L (Vector Labs, USA) in PB
containing Triton X-100 (Sigma, USA) for 24 h, followed by an ABC
reaction (Vector Labs, USA) in the same buffer for 3 h, and finally
reacted in DAB solution (0.035% diaminobenzidine and 0.015% H2O2
in PB). Then, sections were mounted in Permount (Fisher, USA). For
electron microscopy, brain blocks including MGC were cut serially
on a vibrating microslicer (Dosaka, Japan) at 40 mm. Sections were
thereafter subjected to freezeethawing in PB containing 30%
sucrose with liquid nitrogen. Sections were then stained for PHA-L
in the same way as used for light microscopy, except for the
omission of Triton X-100. Sections were then immersed in 1% OsO4
(TAAB, England) in PB at 4  C for 2 h, dehydrated through an
ethanol series and a mixture of ethanol and propylene oxide, and
finally immersed in EPON 812 (TAAB, England). Each section was
flat embedded between a silicon-coated slide and cover glasses and
polymerized in an oven at 40  C for 30 min and then at 60  C for
48 h. Areas of interest were sketched or photographed under a light
microscope, and trimmed for serial sectioning. Ultrathin sections
were cut serially with a diamond knife and collected on single-slit
meshes. Following staining in uranyl acetate and lead citrate,
sections were observed on a JEM 1010 electron microscope (JEOL,
MA, USA) at an acceleration voltage of 80 kV.
We first aimed to find glomerular structures in the non-
lemniscal nucleus, the dMGC, where AI layer 5 originating large
terminals predominantly project. Next, EPON blocks containing the
dMGC, that had been confirmed to contain labeled large terminals
but not small traditional CT terminals using light microscopy, were
trimmed out for cutting. Serial sections were obtained from these
 H. Ojima, K. Murakami / Hearing Research 274 (2011) 40e47
Fig. 3. Electron micrograph of a labeled corticothalamic terminal making synaptic
contacts with 2 profiles (black arrowheads). The postsynaptic profiles show ultra-
structure characterized by pleomorphic synaptic vesicles and asymmetric membrane
specializations. One of the postsynaptic profiles makes a second contact (grey arrow)
with a dendrite devoid of synaptic vesicles, probably a dendrite of a thalamic neuron
(CD), though the membrane specialization cannot be identified here because of the cut
angle. Calibration bar, 500 nm.
EPON blocks, and synaptic interactions of the tracer-labeled
terminals with target thalamic structures were examined. Recon-
structions of CT terminals labeled for synaptic interactions (see
Fig. 6) were made from 4 blocks. Contours of neuronal elements
were drawn on each of the thin sections using an image editing
software (Photoshop Elements, ADOBE, USA). If elements
continued across sections, sets of these contours were super-
imposed to give the maximum outline of the elements. The
maximum outlines of labeled elements and their postsynaptic
elements were extracted and schematically redrawn.
3. Results
3.1. Observation of labeling using light microscopy
Distribution and light microscopy images of small and large CT
terminals revealed by labeling cells of origin in the cat AI (Fig. 1)
were consistent with those described previously for the auditory
cortex (Rouiller and Welker, 1991; Bajo et al., 1995; Llano and
Sherman, 2008), and were also similar to those described for
other sensory and motor cortices. The differential projections of
small and large CT terminals, known to be derived from layers 6 and
5 pyramidal neurons, respectively, were confirmed, but the spatial
segregation of these two types of CT terminals was not complete. As
described previously, almost all labeled terminals in the ventral
nucleus of the MGC were small in size, while most labeled terminals
in the dMGC were large in size. Labeled terminals found in the
medial nucleus of the MGC were mostly small in size, with a few
large terminals dispersed throughout. In the dMGC, isolated large
terminals or their aggregations were dominant. They distributed at
multiple sites across the nucleus, occasionally intermingling with
loosely distributing small terminals. At some loci, such as those
indicated by an arrow in Fig. 1a, aggregates of large CT terminals
were fully segregated from small terminals. For electron microscopy,
43
areas corresponding to these loci were trimmed out from flat
sections embedded in EPON (the cluster shown in Fig. 1b was
examined in the electron microscope as shown in Fig. 4).
3.2. Synaptic arrangement of labeled corticothalamic terminals in
the dMGC
UItrathin sections were obtained from 29 blocks containing only
large terminals in the dMGC. From each of 25 blocks, 4e29 serial
sections were obtained and, from the remaining blocks, single
sections or a maximum of 3 serial sections were cut.
Using our fixation procedure, in the dMGC there were synaptic
structures typical of glomeruli, including a vesicle-free dendrite,
probably one derived from a thalamic projection neuron; medium-
to-large terminals containing highly packed synaptic vesicles, all of
which were round in shape; and large terminals, generally round or
oval in shape, containing relatively loosely packed synaptic vesicles,
most of which are round in shape, but of which some are pleo-
morphic (Fig. 2a). These profiles were encapsulated by glial sheaths.
In contrast to the rare occurrence of the typical glomeruli,
glomerulus-like structures encapsulated in part by glial sheaths
(Fig. 2b), or unencapsulated, were frequently found. The synapses
made by terminals with highly packed vesicles onto the centrally
located vesicle-free dendrites were asymmetric in membrane
specialization (Fig. 2c), while those with loosely packed vesicles
onto the central dendrites were symmetric in membrane speciali-
zation (Fig. 2d). Terminals with highly packed vesicles also made
asymmetric synaptic contacts with those with loosely packed
vesicles (Fig. 2c). Small terminals containing preferentially flat
vesicles, presumably axon terminals projecting either from
thalamic reticular neurons, interneurons (Montero and Scott, 1981;
Montero, 1983; Cucchiaro et al., 1991; Liu et al., 1995; Wanaverbecq
et al., 2008) or IC (Winer et al., 1996), were not identified in our
limited number of samples.
Most labeled terminals were fully filled with reaction products,
making characterization of the cytoplasmic ultrastructure of the
terminals difficult. However, some were only partially filled,
making it possible to identify their ultrastructure more clearly and
definitely. Such an example of labeled CT terminals, identified to be
large in shape using a light microscope, was shown in Fig. 2b. The
cytoplasm was filled with highly packed synaptic vesicles together
with a few mitochondria sparsely distributed among them. The
density of the vesicles was highest among other terminals found
within the same section. When the images of Fig. 2a,b were
compared, a non-labeled terminal, designated by p1 in Fig. 2a, may
represent a large CT terminal.
Observation of single sections revealed that labeled terminals
made synaptic contacts with other profiles of neuronal elements in
5 different ways. First, a single labeled CT terminal profile made
synaptic contact with one vesicle-filled profile (probably PSD, not
shown) or with multiple profiles with the same cytoplasmic
features (Fig. 3). Second, a single labeled CT profile made single or
multiple synaptic contacts with one vesicle-free profile (CT1 in
Fig. 4), probably a dendrite of a thalamic projection neuron. Third,
a single CT profile made synaptic contact with multiple profiles with
different cytoplasmic features (Fig. 2b). Fourth, a single CT terminal
profile made synaptic contact with one or two vesicle-filled profile
(s) that are in turn presynaptic to the vesicle-free profile (CT3-PSD2
or CT4-PSD1 in Fig. 4). Finally, a single labeled CT profile (CT1 in
Fig. 5) made synaptic contact with two profiles of different cyto-
plasmic features, one of which was free of vesicles and was also
postsynaptic to the other, that contained pleomorphic vesicles (see
PSD in Fig. 5). Altogether, these synaptic connections indicate that
large CT terminals participate in the formation of triadic synaptic
arrangements. Direct evidence for this arrangement is further
44 H. Ojima, K. Murakami / Hearing Research 274 (2011) 40e47

Fig. 4. Electron micrograph showing an aggregate of CT terminals (CT1-CT4) within a glomerulus-like structure. Each CT terminal makes single or multiple asymmetric synaptic
contacts with two different profiles, one having loosely packed synaptic vesicles (black arrowheads, CT3 and CT4), probably presynaptic dendrite (PSD), and the other having
virtually no vesicle (grey arrowhead, CT1), probably the central dendrite (CD) of a thalamic principal neuron. The PSD profiles are presynaptic to the latter (CD), with symmetric
membrane specializations as indicated by black arrows. Light microscope observation and reconstruction of serial sections reveal that these 4 large CT terminals (CT1-CT4) are
connected to a single preterminal axon (see Fig. 1b). Calibration bar, 500 nm.

illustrated in Fig. 5, where the three synaptic contacts that should,
by definition, be involved in this arrangement are present on
a single section (black arrow, black arrowhead and white arrow-
head). Additional support for this conclusion comes from the light
and electron microscopic correlation of clustered terminals (see
below).
3.3. Reconstruction
Reconstructions of serial sections from 4 blocks are schemati-
cally illustrated in Fig. 6. In these reconstructions, two labeled CT
terminals (Fig. 6b,c) were demonstrated to participate in the triadic
formation. As shown in Fig. 6b, for example, a set of 4 labeled
boutons, which appeared like a beaded cluster under the light
microscope (see Fig. 1b), were confirmed to be connected by thin
intervening axons in the reconstruction, and as a whole formed the
triad. Thus, though not demonstrable in single electron micro-
graphs, many CT terminals are possibly involved in a triadic
synaptic formation. In our samples, none of the labeled large CT
terminals participated in typical glomerular structures within the
dMGC.
4. Discussion
Our study has two major strengths that convince us of the val-
idity of our interpretation of the synaptic organization of large CT
terminals projecting to the non-lemniscal auditory thalamic
nucleus. First, samples used for observation using an electron
microscope contained only large terminals or clusters of large
terminals. This was made possible by trimming out small areas of
the dMGC that contained exclusively large terminals and had no
contamination of small, drumstick-like terminals originating from
fine axons (i.e., layer 6-derived CT terminals). Thus, interpretation
of the results was strictly confined to synaptic interactions of one
type of terminal. Second, the involvement of large CT terminals in
synaptic arrangements known as the triad was demonstrated
directly on single electron micrographs.
The present data clearly show that large CT terminals,
presumably derived from layer 5 pyramidal neurons, participate in
the formation of triadic synaptic arrangements. These triadic
arrangements are only partially encapsulated by glial sheathes,
indicating that the descending CT system, irrespective of layer 6
origin or layer 5 origin, are unlikely to participate in the formation
of glomeruli at the non-lemniscal thalamic nuclei, despite
glomeruli being typical of those found in the specific thalamic
nuclei (Jones and Powell, 1969a) (Fig. 2a).
The synaptic triad was classically found in the ascending
projection system and included a sensory afferent terminal from
periphery as extrinsic input. Recent studies in the cat visual and
somatosensory systems have demonstrated that the layer
5-derived CT descending projection forms the synaptic arrange-
ments almost identical to those found in the ascending system.
The large CT terminals participate in the formation of the triad in
the non-specific thalamic nucleus as extrinsic input. In the
 H. Ojima, K. Murakami / Hearing Research 274 (2011) 40e47 45

auditory system, however, there is a unique inhibitory (GABAer-

Fig. 5. Electron micrograph showing two large CT terminals (CT1 and CT2). One of
them (CT1) makes synaptic contact with two profiles of different cytoplasmic features
(black and white arrowheads), one of which is free of synaptic vesicles (white
arrowhead) and also postsynaptic to the other (arrow), containing pleomorphic
synaptic vesicles (PSD). Thus, the triadic synaptic formation is demonstrated on
a single thin section. Note that serial sectioning reveals that the symmetric speciali-
zation extends across at least 3 sections, (lower inset). Characteristics of the synapse
made onto the CD from CT1 (white arrowhead) are more convincingly recognized in an
adjacent section of a set of serial sections (upper inset). Additional synapses made on
this CD by another CT terminal (CT2) are indicated by a grey arrowhead (asymmetric
type) and arrow (symmetric type). Calibration bar, 500 nm.
gic) monosynaptic projection ascending from the IC to MGC in
cats (Winer et al., 1996; Saint Marie et al., 1997) and rats (Peruzzi
et al., 1997; Bartlett and Smith, 1999; Smith et al., 2007). This
includes all the IC nuclei (the central and lateral nuclei and the
dorsal cortex) as sources and almost all the MGC nuclei (the
ventral, dorsal and medial ones) as targets. We thus need to clarify
one important question in future; that is, whether the colliculo-
geniculate (CG) projection that is GABAergic participates in the
triadic synaptic arrangements or not. Specifically, it is possible
that the inhibitory terminals from the IC would participate in the
triad as inhibitory elements, some of which might have been
described incorrectly as presynaptic dendirites of thalamic
interneurons.
In rats, a significant number of thalamic neurons show inhibi-
tory postsynaptic potentials in response to stimulation of IC
neurons, and, for some neurons, inhibitory postsynaptic potentials
precede that of excitatory postsynaptic potentials, thereby gener-
ating feed forward inhibition. This inhibitory CG projection
apparently counterbalances the absence or scarcity of interneurons
in the rat MGC. However, considering the substantial similarity
between the CG inhibitory projections in the auditory system of
rats and cats (Winer et al., 1996; Saint Marie et al., 1997), the
involvement of large CT terminals in the triadic formation found
only in cats, but not in rats, is probably independent from the
inhibitory input from the IC. Thus, CT projections originating from
cortical layer 5 may simply follow the common pattern of synaptic
interaction among different sensory modalities (Vidnyánszky et al.,
1996; Feig and Harting, 1998). Consequently, the uniqueness of the
CG inhibitory projection found in the auditory system could be
considered in the context of modes of ascending activation, for
example timing in signal flow or spatial acuity in sound localiza-
tion. Further evaluation can be needed.

Fig. 6. Schematic drawings of synaptic arrangement reconstructed from serial sections obtained from 4 blocks of 2 cats, showing typical synaptic formation made by large CT
terminals in the cat dMGC. In each figure, CT terminals that are derived from a single preterminal axon are depicted. CT terminals are illustrated as hatched profiles, PSD as white
profiles with round and pleomorphic synaptic vesicles, and CD as a white polygon without vesicles, usually surrounded by two other structures. Reconstructions b and c show that
large CT terminals are involved in the formation of the triadic synaptic arrangement. Panel c is reconstructed from serial sections (one of which is Fig. 4) cut through the cluster of
large CT terminals shown in Fig. 1b. asy, asymmetric; sym, symmetric; p.a., puncta adherentia.
46 H. Ojima, K. Murakami / Hearing Research 274 (2011) 40e47
In relation to the dual CT systems, several questions will be
raised. First, are axons ending with large terminals in thalamic
nuclei collaterals of the corticocollicular projections? How are
response properties of layer 5 and layer 6 CT neurons within the
same cortical column different? Are the layer 5 and layer 6 CT
pyramidal neurons connected mutually within the same cortical
column, despite the fact that their cortical axon collateralization
and dendritic arborization are totally different (Ojima et al., 1992)?
We have some relevant answers to the first question. So far most
studies show that axons with large CT terminals are collaterals of
the main descending corticofugal axons (Ojima, 1994; Deschênes
et al., 1994; Veinante et al., 2000; Kakei et al., 2001). Major
targets of layer 5 corticocollicular neurons are collicular nuclei of
non- or weak-tonotopy, such as the dorsomedial nuclei, dorsal
cortex, lateral nucleus and the dorsal margin of the central nucleus
(Andersen et al., 1980; Kelly and Wong, 1981; Winer et al., 1998;
Bajo et al., 2007). If this is the case, it is likely that neurons in the
auditory thalamus and IC receive a common influence exerted by
the same population of cortical pyramidal neurons (Chomiak et al.,
2008). Our electron microscopic results imply that large CT termi-
nals may function as a strong driver (Guillery, 1995; Sherman and
Guillery, 2006), namely by exciting non-lemniscal thalamic
neurons through triadic synaptic interactions involved (Li et al.,
2003a), similar to the sensory ascending afferent. It is possible
that, in the thalamocollicular coordination, thalamic non-lemniscal
neurons are activated almost synchronously with the non-
lemniscal IC neurons, when the cortex is excited. This would
facilitate the ascending transfer of activation from these IC neurons
to the thalamic target neurons through the non-lemniscal pathway
(Kudo and Niimi, 1980; Hu et al., 1994), because the thalamic
nucleus is already in an activated state. Temporal timing of
ascending information transfer from colliculus to thalamus seems
to be more strictly regulated by the state of cerebral cortex. It is
hypothesized that responsiveness of layer 5 neurons could be
influenced by the internal state of the brain, such as expectation or
vigilance (Ojima et al., 2010). This raises the possibility that the
thalamus as well as brain stem nuclei, through the corticofugal
projections, could also be under the influence of a cortical state that
is altering over time with a changing environment.
Ackowledgements
Through our continuous conversation exchanged during the
Society for Neuroscience and ARO meetings, Jeff and I came to agree
that the large terminals projecting back from the auditory cortex to
the non-lemniscal MGC deserved more careful study. It is common
in anatomy that it is within the visual system that new, crucial
connections are found. However, the direct visualization of this
peculiar CT projection ending with large terminals, frequently
forming clusters, in non-specific thalamic nuclei was made first, in
the somatosensory system, and soon after in the auditory system,
using a newly introduced anatomical technique that enabled
anatomists to visualize axons and axonal terminals with the
defined topography of cells of origin. The labels appeared as if they
were stained by the Golgi method.
More recently, in 2008 when I visited his laboratory, Jeff and I
concluded that these terminals did not simply provide sensory
information back to the thalamus from the cortex but rather might
regulate the thalamic excitability via the brain internal state that
was in turn regulated by the thalamus and other brain stem struc-
tures. Jeff, as a keen anatomist, suggested that very few had shown
straightforward images of the contribution of the large CT terminals
to the triad synaptic organizations, which were believed to be
structural units that transfer information from sensory afferents to
thalamic principal neurons. Here, we present data that we promised
him to do at that time. It should be said that this work would not
have been accomplished without his suggestion and warm
encouragement made 2 years ago, just before he passed away.
Authors also thank Ms. M. Hiyomori for typing the manuscript.
References
Andersen, R.A., Snyder, R.L., Merzenich, M.M., 1980. The topographic organization of
corticocollicular projections from physiologically identified loci in the AI, AII,
and anterior auditory cortical fields of the cat. J. Comp. Neurol. 191, 479e494.
Arcelli, P., Frassoni, C., Regondi, M.C., De Biasi, S., Spreafico, R., 1997. GABAergic
neurons in mammalian thalamus: a marker of thalamic complexity? Brain Res.
Bull. 42, 27e37.
Bajo, V.M., Nodal, F.R., Bizley, J.K., Moore, D.R., King, A.J., 2007. The ferret auditory
cortex: descending projections to the inferior colliculus. Cereb. Cortex 17,
475e491.
Bajo, V.M., Rouiller, E.M., Welker, E., Clarke, S., Villa, A.E., de Ribaupierre, Y., de
Ribaupierre, F., 1995. Morphology and spatial distribution of corticothalamic
terminals originating from the cat auditory cortex. Hear. Res. 83, 161e174.
Bartlett, E.L., Stark, J.M., Guillery, R.W., Smith, P.H., 2000. Comparison of the fine
structure of cortical and collicular terminals in the rat medial geniculate body.
Neuroscience 100, 811e828.
Bartlett, E.L., Smith, P.H., 1999. Anatomic, intrinsic, and synaptic properties of dorsal
and ventral division neurons in rat medial geniculate body. J. Neurophysiol. 81,
1999e2016.
Bourassa, J., Deschênes, M., 1995. Corticothalamic projections from the primary
visual cortex in rats: a single fiber study using biocytin as an anterograde tracer.
Neuroscience 66, 253e263.
Bourassa, J., Pinault, D., Deschênes, M., 1995. Corticothalamic projections from the
cortical barrel field to the somatosensory thalamus in rats: a single-fibre study
using biocytin as an anterograde tracer. Eur. J. Neurosci. 7, 19e30.
Chomiak, T., Peters, S., Hu, B., 2008. Functional architecture and spike timing
properties of corticofugal projections from rat ventral temporal cortex. J. Neu-
rophysiol. 100, 327e335.
Cucchiaro, J.B., Uhlrich, D.J., Sherman, S.M., 1991. Electron-microscopic analysis of
synaptic input from the perigeniculate nucleus to the A-laminae of the lateral
geniculate nucleus in cats. J. Comp. Neurol. 310, 316e336.
Deschênes, M., Bourassa, J., Pinault, D., 1994. Corticothalamic projections from layer
V cells in rat are collaterals of long-range corticofugal axons. Brain Res. 664,
215e219.
Feig, S., Harting, J.K., 1998. Corticocortical communication via the thalamus: ultra-
structural studies of corticothalamic projections from area 17 to the lateral
posterior nucleus of the cat and inferior pulvinar nucleus of the owl monkey.
J. Comp. Neurol. 395, 281e295.
Guillery, R.W., 1969. The organization of synaptic interconnections in the laminae of
the dorsal lateral geniculate nucleus of the cat. Z. Zellforsch Mikrosk Anat. 96,
1e38.
Guillery, R.W., 1995. Anatomical evidence concerning the role of the thalamus in
corticocortical communication: a brief review. J. Anat. 187, 583e592.
Guillery, R.W., Feig, S.L., Van Lieshout, D.P., 2001. Connections of higher order visual
relays in the thalamus: a study of corticothalamic pathways in cats. J. Comp.
Neurol. 438, 66e85.
Hazama, M., Kimura, A., Donishi, T., Sakoda, T., Tamai, Y., 2004. Topography of
corticothalamic projections from the auditory cortex of the rat. Neuroscience
12, 655e667.
Hoogland, P.V., Welker, E., Van der Loos, H., 1987. Organization of the projections
from barrel cortex to thalamus in mice studied with Phaseolus vulgaris-leu-
coagglutinin and HRP. Exp. Brain Res. 68, 73e87.
Hoogland, P.V., Wouterlood, F.G., Welker, E., Van der Loos, H., 1991. Ultrastructure of
giant and small thalamic terminals of cortical origin: a study of the projections
from the barrel cortex in mice using Phaseolus vulgaris leuco-agglutinin (PHA-
L). Exp. Brain Res. 87, 159e172.
Hu, B., Senatorov, V., Mooney, D., 1994. Lemniscal and non-lemniscal synaptic
transmission in rat auditory thalamus. J. Physiol. 479, 217e231.
Huppé-Gourgues, F., Bickford, M.E., Boire, D., Ptito, M., Casanova, C., 2006. Distri-
bution, morphology, and synaptic targets of corticothalamic terminals in the cat
lateral posterior-pulvinar complex that originate from the posteromedial lateral
suprasylvian cortex. J. Comp. Neurol. 497, 847e863.
Jones, E.G., 2007. The Thalamus, second ed. Cambridge University Press, Cambridge.
Jones, E.G., Powell, T.P., 1969a. Electron microscopy of synaptic glomeruli in the
thalamic relay nuclei of the cat. Proc. R. Soc. Lond. B Biol. Sci. 172, 153e171.
Jones, E.G., Powell, T.P., 1969b. An electron microscopic study of the mode of
termination of cortico-thalamic fibres within the sensory relay nuclei of the
thalamus. Proc. R. Soc. Lond. B Biol. Sci. 172, 173e185.
Kakei, S., Na, J., Shinoda, Y., 2001. Thalamic terminal morphology and distribution of
single corticothalamic axons originating from layers 5 and 6 of the cat motor
cortex. J. Comp. Neurol. 437, 170e185.
Kelly, J.P., Wong, D., 1981. Laminar connections of the cat’s auditory cortex. Brain
Res. 212, 1e15.
Kimura, A., Donishi, T., Okamoto, K., Tamai, Y., 2005. Topography of projections
from the primary and non-primary auditory cortical areas to the medial
geniculate body and thalamic reticular nucleus in the rat. Neuroscience 135,
1325e1342.
 H. Ojima, K. Murakami / Hearing Research 274 (2011) 40e47
Kudo, M., Niimi, K., 1980. Ascending projections of the inferior colliculus in the cat:
an autoradiographic study. J. Comp. Neurol. 191, 545e556.
Kuroda, M., Sugiura, T., Shinkai, M., Murakami, K., Oda, S., Kishi, K., 1993. Synaptic
organization and prefrontal corticothalamic termination in the mediodorsal
thalamic nucleus of the cat. J. Hirnforsch. 34, 417e430.
Li, J., Guido, W., Bickford, M.E., 2003a. Two distinct types of corticothalamic EPSPs
and their contribution to short-term synaptic plasticity. J. Neurophysiol. 290,
3429e3440.
Li, J., Wang, S., Bickford, M.E., 2003b. Comparison of the ultrastructure of cortical
and retinal terminals in the rat dorsal lateral geniculate and lateral posterior
nuclei. J. Comp. Neurol. 460, 394e409.
Liu, X.B., Warren, R.A., Jones, E.G., 1995. Synaptic distribution of afferents from
reticular nucleus in ventroposterior nucleus of cat thalamus. J. Comp. Neurol.
352, 187e202.
Llano, D.A., Sherman, S.M., 2008. Evidence for nonreciprocal organization of the
mouse auditory thalamocortical-corticothalamic projection systems. J. Comp.
Neurol. 507, 1209e1227.
Montero, V.M., Scott, G.L., 1981. Synaptic terminals in the dorsal lateral geniculate
nucleus from neurons of the thalamic reticular nucleus: a light and electron
microscope autoradiographic study. Neuroscience 6, 2561e2577.
Montero, V.M., 1983. Ultrastructural identification of axon terminals from the
thalamic reticular nucleus in the medial geniculate nucleus in the rat: an EM
autoradiographic study. Exp. Brain Res. 51, 338e342.
Ogren, M.P., Hendrickson, A.E., 1979. The morphology and distribution of striate
cortex terminals in the inferior and lateral subdivisions of the Macaca monkey
pulvinar. J. Comp. Neurol. 188, 179e199.
Ohyama, J., Ojima, H., 1997. Labeling of pyramidal and nonpyramidal neurons with
lectin Vicia villosa during postnatal development of the guinea pig. J. Comp.
Neurol. 389, 453e468.
Ojima, H., 1994. Terminal morphology and distribution of corticothalamic fibers
originating from layers 5 and 6 of cat primary auditory cortex. Cereb. Cortex 4,
646e663.
Ojima, H., Honda, C.N., Jones, E.G., 1992. Characteristics of intracellularly injected
infragranular pyramidal neurons in cat primary auditory cortex. Cereb. Cortex 2,
197e216.
Ojima, H., Murakami, K., 2002. Intracellular characterization of suppressive
responses in supragranular pyramidal neurons of cat primary auditory cortex in
vivo. Cereb. Cortex 12, 1079e1091.
Ojima, H., Murakami, K., Kishi, K., 1996. Dual termination modes of corticothalamic
fibers originating from pyramids of layers 5 and 6 in cat visual cortical area 17.
Neurosci. Lett. 208, 57e60.
Ojima, H., Taoka, M., Iriki, A., 2010. Adaptive changes in firing of primary auditory
cortical neurons following illumination shift from light to dark in freely moving
guinea pigs. Cereb. Cortex 20, 339e351.
Peruzzi, D., Bartlett, E., Smith, P.H., Oliver, D.L., 1997. A monosynaptic GABAergic
input from the inferior colliculus to the medial geniculate body in rat. J. Neu-
rosci. 17, 3766e3777.
Robson, J.A., Hall, W.C., 1977. The organization of the pulvinar in the grey squirrel
(Sciurus carolinensis). II. Synaptic organization and comparisons with the
dorsal lateral geniculate nucleus. J. Comp. Neurol. 173, 389e416.
Rockland, K.S., 1994. Further evidence for two types of corticopulvinar neurons.
Neuroreport 5, 1865e1868.
Rouiller, E.M., Durif, C., 2004. The dual pattern of corticothalamic projection of the
primary auditory cortex in macaque monkey. Neurosci. Lett. 358, 49e52.
Rouiller, E.M., Liang, F.Y., Moret, V., Wiesendanger, M., 1991. Patterns of cortico-
thalamic terminations following injection of Phaseolus vulgaris leucoagglutinin
(PHA-L) in the sensorimotor cortex of the rat. Neurosci. Lett. 125, 93e97.
47
Rouiller, E.M., de Ribaupierre, F., 1990. Arborization of corticothalamic axons in
the auditory thalamus of the cat: a PHA-L tracing study. Neurosci. Lett. 108,
29e35.
Rouiller, E.M., Tanné, J., Moret, V., Kermadi, I., Boussaoud, D., Welker, E., 1998. Dual
morphology and topography of the corticothalamic terminals originating from
the primary, supplementary motor, and dorsal premotor cortical areas in
macaque monkeys. J. Comp. Neurol. 396, 169e185.
Rouiller, E.M., Welker, E., 1991. Morphology of corticothalamic terminals arising
from the auditory cortex of the rat: a Phaseolus vulgaris-leucoagglutinin (PHA-L)
tracing study. Hear. Res. 56, 179e190.
Rouiller, E.M., Welker, E., 2000. A comparative analysis of the morphology of cor-
ticothalamic projections in mammals. Brain Res. Bull. 53, 727e741.
Saint Marie, R.L., Stanforth, D.A., Jubelier, E.M., 1997. Substrate for rapid feedforward
inhibition of the auditory forebrain. Brain Res. 765, 173e176.
Schwartz, M.L., Dekker, J.J., Goldman-Rakic, P.S., 1991. Dual mode of corticothalamic
synaptic termination in the mediodorsal nucleus of the rhesus monkey. J. Comp.
Neurol. 309, 289e304.
Sherman, S.M., Guillery, R.W., 2006. Exploring the Thalamus. Academic Press, New
York.
Smith, P.H., Bartlett, E.L., Kowalkowski, A., 2007. Cortical and collicular inputs to
cells in the rat paralaminar thalamic nuclei adjacent to the medial geniculate
body. J. Neurophysiol. 98, 681e695.
Szentágothai, J., 1963. The structure of the synapse in the lateral geniculate body.
Acta Anat. (Basel) 55, 166e185.
Takayanagi, M., Ojima, H., 2006. Microtopography of the dual corticothalamic
projections originating from domains along the frequency axis of the cat
primary auditory cortex. Neuroscience 142, 769e780.
Thompson, G.C., Cortez, A.M., Lam, D.M., 1985. Localization of GABA immunoreac-
tivity in the auditory brainstem of guinea pigs. Brain Res. 339, 119e122.
Van Horn, S.C., Sherman, S.M., 2004. Differences in projection patterns between
large and small corticothalamic terminals. J. Comp. Neurol. 475, 406e415.
Vidnyánszky, Z., Borostyánkõi, Z., Görcs, T.J., Hámori, J., 1996. Light and electron
microscopic analysis of synaptic input from cortical area 17 to the lateral
posterior nucleus in cats. Exp. Brain Res. 109, 63e70.
Veinante, P., Lavallée, P., Deschênes, M., 2000. Corticothalamic projections from
layer 5 of the vibrissal barrel cortex in the rat. J. Comp. Neurol. 424, 197e204.
Wanaverbecq, N., Bodor, A.L., Bokor, H., Slézia, A., Lüthi, A., Acsády, L., 2008. Con-
trasting the functional properties of GABAergic axon terminals with single and
multiple synapses in the thalamus. J. Neurosci. 28, 11848e11861.
Winer, J.A., Diehl, J.J., Larue, D.T., 2001. Projections of auditory cortex to the medial
geniculate body of the cat. J. Comp. Neurol. 430, 27e55.
Winer, J.A., Larue, D.T., 1988. Anatomy of glutamic acid decarboxylase immunore-
active neurons and axons in the rat medial geniculate body. J. Comp. Neurol.
278, 47e68.
Winer, J.A., Larue, D.T., 1996. Evolution of GABAergic circuitry in the mammalian
medial geniculate body. Proc. Natl. Acad. Sci. U S A 93, 3083e3087.
Winer, J.A., Larue, D.T., Diehl, J.J., Hefti, B.J., 1998. Auditory cortical projections to the
cat inferior colliculus. J. Comp. Neurol. 400, 147e174.
Winer, J.A., Larue, D.T., Huang, C., 1999. Two systems of giant axon terminals in the
cat medial geniculate body: convergence of cortical and GABAergic inputs. J.
Comp. Neurol. 413, 181e197.
Winer, J.A., Saint Marie, R.L., Larue, D.T., Oliver, D.L., 1996. GABAergic feedforward
projections from the inferior colliculus to the medial geniculate body. Proc. Natl.
Acad. Sci. U S A 93, 8005e8010.
Zikopoulos, B., Barbas, H., 2006. Prefrontal projections to the thalamic reticular
nucleus form a unique circuit for attentional mechanisms. J. Neurosci. 26,
7348e7361.