803 Full

Physiol Rev
83: 803 834, 2003; 10.1152/physrev.00002.2003.
The Amygdaloid Complex: Anatomy and Physiology

P. SAH, E. S. L. FABER, M. LOPEZ DE ARMENTIA, AND J. POWER
Division of Neuroscience, John Curtin School of Medical Research, Australian National University,
Canberra, Australian Capital Territory, Australia
I. Introduction 804
II. The Amygdaloid Complex 804
A. Basolateral nuclei 806
B. Cortical-like nuclei 806
C. Centromedial nuclei 806
D. Other amygdaloid nuclei 806
E. Extended amygdala 806
III. Afferent and Efferent Connections 807
A. Sensory inputs 807
B. Polymodal inputs 808
C. Efferent connections 809
IV. Intra-amygdaloid connections 810
V. Morphology and Physiology: Basolateral Complex 811
A. Morphology 812
B. Physiology 813
C. Synaptic properties 816
VI. Morphology and Physiology: Central Nucleus 819
A. Morphology 819
B. Physiological properties 819
C. Synaptic properties 820
VII. Morphology and Physiology: Other Nuclei 821
A. Intercalated cell masses 821
B. Medial nucleus 822
VIII. Role of the Amygdaloid Complex 822
IX. The Amygdala and Fear Conditioning 822
A. Basolateral complex 823
B. Central nucleus 824
C. Where is the memory stored? 824
X. Synaptic Plasticity and Fear Conditioning 825
XI. Conclusions 827
Sah, P., E. S. L. Faber, M. Lopez de Armentia, and J. Power. The Amygdaloid Complex: Anatomy and
Physiology. Physiol Rev 83: 803 834, 2003; 10.1152/physrev.00002.2003.A converging body of literature over the
last 50 years has implicated the amygdala in assigning emotional significance or value to sensory information. In
particular, the amygdala has been shown to be an essential component of the circuitry underlying fear-related
responses. Disorders in the processing of fear-related information are likely to be the underlying cause of some
anxiety disorders in humans such as posttraumatic stress. The amygdaloid complex is a group of more than 10 nuclei
that are located in the midtemporal lobe. These nuclei can be distinguished both on cytoarchitectonic and
connectional grounds. Anatomical tract tracing studies have shown that these nuclei have extensive intranuclear and
internuclear connections. The afferent and efferent connections of the amygdala have also been mapped in detail,
showing that the amygdaloid complex has extensive connections with cortical and subcortical regions. Analysis of
fear conditioning in rats has suggested that long-term synaptic plasticity of inputs to the amygdala underlies the
acquisition and perhaps storage of the fear memory. In agreement with this proposal, synaptic plasticity has been
demonstrated at synapses in the amygdala in both in vitro and in vivo studies. In this review, we examine the
anatomical and physiological substrates proposed to underlie amygdala function.
www.prv.org 0031-9333/03 $15.00 Copyright 2003 the American Physiological Society 803
804 SAH ET AL.
I. INTRODUCTION of medial temporal lobe lesions in monkeys. These ani-

mals showed a range of effects including marked changes
The amygdala is an almond-shaped structure deep in emotional behavior that were described by them as
within the temporal lobe and was first identified by Bur- psychic blindness and has come to be known as the
dach in the early 19th century. Burdach originally de- Klu ver Bucy syndrome. However, these lesions were quite
scribed a group of cells that are now known as the baso- large and included the amygdala, hippocampus, and sur-
lateral complex. Subsequently, a large number of struc- rounding cortical areas. Subsequently, Weiskrantz (288)
tures that surround the basolateral complex have been showed that more restricted amygdala lesions could rep-
identified in many species and constitute what is now licate the results of Klu ver and Bucy, cementing the fun-
known as the amygdaloid complex. This structure has damental role of the amygdala in emotional processing.
attracted continued interest because of its central role in These studies made it clear that the amygdala is an es-
emotional processing. The word emotion is a difficult sential component of the circuitry that assigns emotional
concept that describes subjective experiences and feel- significance and produces appropriate behavioral re-
ings such as pain, fear, desire, and hope as well as aspects sponses to salient external stimuli (69, 96, 116, 231).
of the behavior of individuals both public and private. In While initial studies on the role of the amygdaloid
the past, emotions had traditionally been viewed as ex- complex used avoidance conditioning and instrumental
clusively human and distinct from other aspects of brain learning (245), the study of emotions reached a new level
function such as cognition and sensory perception. This of analysis with the development of the study of fear
separation between cognition and emotion persisted de- conditioning. Fear conditioning is a simple Pavlovian con-
spite there being little doubt that emotions can have a ditioning task in which a neutral stimulus, such as a tone
major impact on various aspects of mental function. In or a light, is paired with an aversive stimulus, typically a
the biological study of emotions, perhaps the single most footshock. Following a relatively small number of such
influential contribution was that of Charles Darwin. In his pairings, the neutral stimulus subsequently elicits a be-
seminal Expression of Emotions in Man and Animals havioral state similar to that evoked by the aversive stim-
published in 1872, Darwin (40) suggested that there are ulus alone. The fear response consists of freezing (a ces-
some fundamental aspects of emotion that find similar sation of movement), sweating, and changes in heart rate
expression in the behavior of both man and animals. This and blood pressure. In humans, there are also cognitive
was the first indication that it may be possible to make effects such as feelings of dread and despair associated
inferences about human emotion by examining animal with these autonomic effects. This learned behavior is
behavior. Around the same time William James (87) and rapidly acquired and long lasting. The simple nature of
C. G. Lange (111) independently suggested that emotions this learning task, and the readily measured physiological
are the cognitive responses that accompany our physio- changes that accompany it, have made the study of fear
logical responses to external stimuli. This idea came to be conditioning a very attractive model for the study of
called the James-Lange theory of emotion. Thus, in James learning and memory consolidation. Furthermore, be-
words, when we see a bear we dont run because we are cause of the physiological similarities between animal
afraid but are afraid because we run (88). Together with and human fear, fear conditioning is seen as relevant to
Darwins proposal, these ideas suggested that it is possi- the genesis of anxiety disorders in humans (41, 43, 237),
ble to study emotions by examining the physiological thus providing an additional incentive to study fear. These
responses to stimuli. advances have led to a rapid increase in the number of
The first neurophysiological theories of emotion studies examining the role of the amygdala in fear, learn-
emerged from the work of Cannon and Bard in the 1920s. ing, and memory in general. Although some controversy
Cannon and Bard were critical of the James-Lange theory persists in the precise role of the structures involved (114,
and instead suggested that the hypothalamus and its pro- 198), these studies have provided definitive evidence im-
jections to the cortex and brain stem were the central plicating the amygdala and its afferent and efferent pro-
element that both evaluated and initiated emotional re- jections in fear processing in mammals (1, 43, 96, 115).
sponses (25). Subsequently, Papez (197), reviewing the In this review, we discuss the anatomy and physiol-
anatomical and clinical data in 1937, added more medial ogy of the amygdala and the mechanisms proposed to
temporal structures to the circuitry involved in emotional underlie its involvement in fear conditioning.
expression. These ideas were further expanded by Paul
McLean (172), who named these forebrain circuits the
visceral brain and introduced the concept of the limbic II. THE AMYGDALOID COMPLEX
system. While McLean included the amygdala in the lim-
bic system, the involvement of the amygdala in emotional The amygdaloid complex, located in the medial tem-
processing arose from the now classic studies of Klu ver poral lobe, is structurally diverse and comprises 13
and Bucy (101, 102) who examined the behavioral effects nuclei. These are further divided into subdivisions that
Physiol Rev VOL 83 JULY 2003 www.prv.org

THE AMYGDALA 805
have extensive internuclear and intranuclear connections. mostly concentrate on results obtained in this species. We
These nuclei and subnuclei are distinguished on the basis use the nomenclature introduced by Price et. al (213) with
of cytoarchitectonics, histochemistry, and the connec- some modifications (159) (see below). In this classifica-
tions they make (105, 208). The architectonic organization tion, amygdala nuclei are divided into three groups (Fig.
and connectivity of the amygdala have been extensively 1): 1) the deep or basolateral group, which includes the
reviewed (4, 48, 159, 208). Although these reviews have lateral nucleus, the basal nucleus, and accessory basal
mostly concentrated on the rat amygdala, it has also been nucleus; 2) the superficial or cortical-like group, which
studied in the monkey (8) and cat (213). These studies includes the cortical nuclei and nucleus of the lateral
reveal that while there are many similarities between olfactory tract; and 3) the centromedial group composed
species, there are also clear differences in the organiza- of the medial and central nuclei. Finally, there is a separate
tion and the relative sizes of the different amygdaloid set of nuclei that do not easily fall into any of these three
nuclei. Because functional studies of the amygdala have groups and are listed separately. These include the interca-
largely been carried out in the rat, in this review we lated cell masses and the amygdalohippocampal area.
FIG. 1. Nuclei of the rat amygdaloid

complex. Coronal sections are drawn from
rostral (A) to caudal (D). The different nu-
clei are divided into three groups as de-
scribed in text. Areas in blue form part of
the basolateral group, areas in yellow are
the cortical group, and areas in green form
the centromedial group. ABmc, accessory
basal magnocellular subdivision; ABpc, ac-
cessory basal parvicellular subdivision; Bpc,
basal nucleus magnocellular subdivision;
e.c., external capsule; Ladl, lateral amygdala
medial subdivision; Lam, lateral amygdala
medial subdivision; Lavl, lateral amygdala
ventrolateral subdivision; Mcd, medial
amygdala dorsal subdivision; Mcv, medial
amygdala ventral subdivision; Mr, medial
amygdala rostral subdivision; Pir, piriform
cortex; s.t., stria terminalis. See text for
other definitions.

806 SAH ET AL.
A. Basolateral Nuclei (159), we will also separate this group from cortical nu-
clei. The CeA is located dorsomedially in the rostral part
The basolateral or deep nuclei comprise the lateral of the amygdala, bordered laterally by the basolateral
nucleus (LA), the basal nucleus (B), which is sometimes complex, dorsally by the globus pallidus, and medially by
called the basolateral nucleus (BLA), and the accessory the stria terminalis. The CeA has four divisions: the cap-
basal nucleus (AB), which is also known as the basome- sular subdivision (CeC), lateral subdivision (CeL), inter-
dial nucleus (Fig. 1). Often, these three nuclei are collec- mediate subdivision (CeI), and medial subdivision (CeM)
tively referred to as the basolateral complex (91). The LA (92, 149). The medial nucleus is found near the surface
is located dorsally in the amygdala where it abuts the bounded medially by the optic tract. It begins at the level
basal nucleus ventrally. It is bordered laterally by the of the NLOT and extends caudally. It has four subdivi-
external capsule and medially by the central nucleus. It sions: rostral, central (dorsal and ventral), and caudal.
has three subdivisions: the smaller celled dorsolateral
subdivision, the larger celled ventrolateral subdivision,
D. Other Amygdaloid Nuclei
and the medial subdivision. The basal nucleus is located
ventral to the LA and is subdivided into the rostral mag-
The final group of nuclei comprising the remaining
nocellular subdivision and the more caudal intermediate
amygdala areas are the anterior amygdala area (AAA), the
and parvicellular subdivisions. The AB is found ventral to
amygdalo-hippocampal area (AHA), and the intercalated
the basal nucleus and lies adjacent to the amygdalohip-
nuclei (I) (8, 210, 213). The AHA is the most caudal of the
pocampal area (AHA). It is comprised of the magnocellu-
amygdaloid nuclei and is comprised of the medial and
lar subdivision, the intermediate subdivision, and the par-
lateral subdivisions. The intercalated nuclei are small
vicellular subdivision (208, 210).
groups of neurons found in clusters within the fiber bun-
dles that separate the different amygdaloid nuclei (173).
B. Cortical-like Nuclei
E. Extended Amygdala
The second group is the superficial or corticomedial
nuclei (150, 213) (Fig. 1). Although these superficial struc-
Although the above classification has been adopted
tures are called nuclei, many have cortical characteristics
by many, several authors have suggested that a different
since they are located at the surface of the brain and have
classification is more appropriate. Initially, based on the
a layered structure (213). They comprise the nucleus of
known connections of the amygdala, Alheid and Heimer
the lateral olfactory tract (NLOT), bed nucleus of the
and co-workers (4, 5) argued that the centromedial amyg-
accessory olfactory tract (BAOT), the anterior and poste-
dala should be extended rostrally and medially. They
rior cortical nucleus (CoA and CoP, respectively), and the
pointed out that the amygdala innervates the BNST and
periamygdaloid cortex (PAC). The BAOT is at the very
the caudodorsal regions of the substantia inominata (ven-
rostral part of the amygdala where it is bordered laterally
tral pallidum). Furthermore, these two regions have sim-
by the CoA. The CoA is a layered structure located lateral
ilar efferent connections to the descending projections of
to the NLOT rostrally and the medial nucleus caudally.
the amygdala. Thus they argued that these regions are
The CoP is also three layered and is located in the most
part of the amygdaloid complex. By including these re-
caudal parts of the amygdala where it borders the AHA
gions they suggested that the centromedial complex
dorsally and the PAC laterally. The PAC is found ventral
should be termed the extended amygdala. More re-
to the basal nucleus and is subdivided into three subdivi-
cently, Swanson and Petrovich (267, 269) have argued
sions: the periamygdaloid cortex, the medial division, and
that the nuclei of the amygdaloid complex are a structur-
the sulcal division (208, 213).
ally and functionally heterogeneous group that have been
arbitrarily grouped. They suggest that these nuclei should
C. Centromedial Nuclei be divided into four functional systems. These systems
would be the frontotemporal, autonomic, main olfactory,
The centromedial nuclear group is found in the dor- and accessory olfactory systems. In this classification, the
somedial portion of the amygdaloid complex and consists basolateral nuclei, which embryologically are cortical-like
of the central (CeA), medial (M), and the amygdaloid part nuclei, receive afferents from similar sources and contain
of the bed nucleus of stria terminalis (BNST; Fig. 1). cells resembling cortical neurons (see below) from part of
Traditionally these nuclei were pooled with the cortical the frontotemporal (cortical-like) system. The central nu-
nuclei. However, it has recently been suggested that the clei, embryologically striatal in origin (214), contain cells
central, medial, and BNST have histochemical and devel- morphologically similar to those in the striatum (214, 269)
opmental characteristics that are distinct from the corti- and make many connections with regions involved in
cal nuclei (see below). Thus, as initiated by McDonald autonomic control comprise the autonomic system. Fi-

THE AMYGDALA 807
nally, the cortical nuclei, and the medial nucleus, which A. Sensory Inputs
are the major target of olfactory projections (see below),
are part of the main and accessory olfactory systems. The amygdala receives inputs from all modalities:
This grouping of the nuclei naturally divides the olfactory, somatosensory, gustatory and visceral, audi-
amygdaloid complex into distinct functional systems and tory, and visual.
fits well with the development of the structures. The Olfactory projections arise from the main and acces-
classification that we have described above is largely in sory olfactory bulbs as well as the primary olfactory
agreement with this proposal, with the basolateral nuclei cortex. The main olfactory bulb projects mainly to the
constituting the frontotemporal group, the centromedial nucleus of the lateral olfactory tract, anterior cortical
nuclei forming the autonomic group, and the cortical-like nucleus, and the periamygdaloid cortex, whereas the ac-
nuclei constituting the two olfactory groups. cessory olfactory bulb projects to the bed nucleus of the
accessory olfactory tract, the medial nucleus, and poste-
rior cortical amygdala (250). The piriform cortex and
III. AFFERENT AND EFFERENT CONNECTIONS anterior olfactory nucleus have projections to the lateral
amygdala, basal, and accessory basal nuclei (131). The
Data on afferent and efferent connections to the dorsal endopiriform nucleus additionally projects to all
amygdaloid complex come from studies in which antero- cortical nuclei of the amygdala as well as the nucleus of
grade or retrograde tracers have been injected into vari- the lateral olfactory tract, the periamydaloid cortex, and
ous amygdaloid, cortical, and subcortical regions. These medial amygdala (12). Thus all regions of the olfactory
studies reveal that each amygdaloid nucleus receives in- stream have projections to the amygdaloid complex.
puts from multiple yet distinct brain regions (159, 208, For somatosensory inputs, few projections arise di-
213). Efferent projections from the amygdala are also rectly from primary somatosensory areas. Most afferents
widespread and include both cortical and subcortical re- reach the amygdala via the dysgranular parietal insular
gions (208). There is a vast and complex literature on cortex in the parietal lobe (256). These projections target
the lateral, basal, and central nucleus (162, 256, 255). For
connections involving the amygdaloid complex, and there
the lateral amygdala, strong labeling is seen in the dorso-
have recently been two detailed reviews (159, 208). Here
lateral subdivision while in the basal nucleus these inputs
we briefly summarize the main afferent and efferent con-
are not segregated (256). Somatosensory information also
nections to the amygdala. Studies carried out in rats, cats,
reaches the amygdala by projections from the pontine
and monkeys show that in most cases there are extensive
parabrachial nucleus and thalamic nuclei, the medial por-
similarities in the organization of inputs and outputs in
tion of the medial geniculate and the posterior internu-
the three species. We have chosen to concentrate on the
clear nucleus (PIN), which have been suggested to be
connections in the rat, since most of the physiology has
involved in the transmission of nociceptive information
been performed in this species.
(16, 19, 121). Inputs arising in the PIN target all subdivi-
Based largely on the information content of the af- sions of the LA, but also innervate the accessory basal
ferents, inputs to the amygdala can be separated into nucleus and the medial subdivision of the central nucleus
those arising in cortical and thalamic structures and those (15, 129).
arising in the hypothalamus or brain stem. Cortical and Gustatory and visceral primary areas in the anterior
thalamic inputs supply information from sensory areas and posterior insular cortices provide strong projections
and structures related with memory systems. Hypotha- to the dorsal subdivision of LA, posterior basal nucleus,
lamic and brain stem inputs arise from regions involved in and central nucleus (255). Gustatory and visceral infor-
behavior and autonomic systems. The major source of mation also arrive from subcortical structures and, as
sensory information to the amygdala is the cerebral cor- with somatosensory projections, both cortical and sub-
tex (159). These projections are glutamatergic, predomi- cortical inputs converge in the amygdaloid complex (159).
nantly arising from layer V pyramidal neurons (7, 194). Inputs from the posteromedial ventral thalamic nucleus
The majority are ipsilateral and enter the amygdala via the (the thalamic gustatory nucleus) terminate in the LA, B,
external capsule (145). Most cortical projections originate and CeL (185, 274), and those from the parabrachial nu-
in association areas and transmit processed information cleus, which receives projections from the nucleus of the
by a series of cortico-cortical connections originating in solitary tract, target the CeL (15, 193).
the primary sensory cortex. These inputs can be divided Auditory and visual information also reach the amyg-
into those that relay modality-specific sensory informa- dala from association areas rather than primary cortex.
tion, those that are polymodal, and those arising in the These pathways are thought to be particularly relevant
medial temporal lobe memory system. The different in- during fear conditioning (see below). For auditory infor-
puts and their distributions in the amygdala are summa- mation, area Te1, the primary auditory cortex in rat, has
rized in Figure 2. no direct projections to the amygdala (145, 254). Injec-

808 SAH ET AL.
FIG. 2. Summary of the inputs to the amygdaloid nuclei. Neuromodulatory inputs (e.g., acetylcholine, serotonin)
have been omitted for clarity. See Fig. 1 and text for definitions.
tions of anterograde tracers in Te3 show fibers in the LA, dalities converges in the prefrontal cortical areas (224),
with the dorsolateral subdivision being the most common many of which are involved in behavior and reward cir-
target (119, 254). Retrograde tracing studies have shown cuitry in rats (231). In all species, a dense and topograph-
that these projections arise from cortical layers II and IV ically organized projection from the frontal cortex has
(119). Subcortical acoustic inputs arise from the thalamic been described (159). The basal nucleus is the main target
medial geniculate nucleus and target the same areas of of afferents from the prefrontal cortex, although projec-
the LA (118, 119, 274). As with acoustic inputs, visual tions to the LA as well as accessory basal, central, and
cortical projections to the amygdala also originate both medial nuclei have also been described (165).
from thalamic and high-order visual areas (253). Cortical Areas related to the long-term declarative memory
projections from these areas (Oc2) follow a cascade to system include the perirhinal cortex, the entorhinal cor-
the amygdala in large part via Te2 (159, 253). These fibers tex, the parahippocampal cortex, and the hippocampus
terminate in the dorsal subdivision of the LA, the CeL, and (175). Projections between the amygdala and these struc-
some in the magnocellular basal nucleus. tures are reciprocal and strong (159, 208). The medial
division of the LA receives the heaviest projection from
B. Polymodal Inputs the perirhinal cortex, but projections to basal and cortical
nuclei have also been described (257). The entorhinal
There are several sources of polymodal sensory in- cortex in comparison appears to project to most amyg-
formation to the amygdala. These include prefrontal cor- dalar nuclei (164). Inputs from hippocampus to the amyg-
tex, perirhinal cortex, and hippocampus. The prefrontal dala mainly originate in the subicular region, and although
cortex is a major source of cortical projections to the the basal nucleus is the main target, most other nuclei are
amygdaloid complex. Information from all sensory mo- also more sparsely innervated (26).

THE AMYGDALA 809
In summary, the amygdala receives sensory information has a major projection to prefrontal cortex, nucleus
tion from all modalities. These inputs target structures in accumbens, and the thalamus. Efferents from the baso-
the amygdaloid complex at all levels, from the tradition- lateral complex arise from pyramidal-like neurons and are
ally considered input side of the complex (basolateral thought to be glutamatergic (204).
complex and cortical nuclei) to the output side (centro- As mentioned above, the amygdala is involved in
medial nuclei) (208). Thus there are extensive levels of emotional responses, especially in fear and fear condi-
convergence between different sensory modalities. In tioning. These responses are characterized by freezing,
combination with access to information from the medial potentiated startle, release of stress hormones, and
temporal memory systems, the amygdala is in a good changes in blood pressure and heart rate which are elic-
position to form associations between current sensory ited by activation of the autonomic and hormonal systems
inputs and past experience. (42, 115). Activation of the central nucleus induces this
In addition to sensory information, the central, lat- autonomic response by stimulating groups of neurons in
eral, and medial nuclei receive substantial inputs from the the brain stem that control the autonomic system, or
hypothalamus while the other amygdalar areas receive alternatively by stimulating hypothalamic nuclei that
very meager projections. For brain stem inputs, the cen- modulate these centers (97, 120). In agreement with these
tral nucleus is a major target for a variety of inputs from behavioral responses, the medial subdivision of the cen-
the midbrain, pons, and medulla, while the other nuclei tral nucleus has substantial projections to the hypothala-
receive few or no inputs from these areas (208). mus, bed nucleus of the stria terminalis (49), and several
nuclei in the midbrain, pons, and medulla (279). Projec-
C. Efferent Connections tions to the brain stem are to three main areas: the peri-
aqueductal gray, which leads to vocalization, startle, an-
The amygdaloid nuclei have widespread projections algesia and cardiovascular changes (13, 226); the parabra-
to cortical, hypothalamic, and brain stem regions (Fig. 3). chial nucleus, which is involved in pain pathways (67,
In general, projections from the amygdala to cortical sen- 178); and the nucleus of the solitary tract (NTS), which is
sory areas are light and originate in cortical and basolat- connected with the vagal system (275).
eral areas of the amygdala. The perirhinal area, along with The hypothalamus contains a group of nuclei that
other areas in the frontal cortex that project to the amyg- have a major influence in the coordination of ingestive,
dala, receive reciprocal connections from the LA, B, AB, reproductive, and defensive behaviors (267). The medial
M, and periamygdaloid cortex (208). The cortical nuclei and capsular subdivisions of the central nucleus innervate
that receive olfactory projections all send substantial re- mostly the dorsolateral and caudolateral regions of the
ciprocal projections back to the olfactory cortex. hypothalamus (205). These areas of the hypothalamus
The basolateral complex (LA, B, AB) has a substan- project to autonomic cell groups in the brain stem and
tial projection to the medial temporal lobe memory sys- spinal cord (268). Efferents from the lateral subdivision of
tem with afferents to hippocampus and perirhinal cortex the central nucleus and from nuclei related with the ol-
(205, 208). A large projection is also found to the nucleus factory system in the amygdala also project to these ar-
accumbens (154). Similar to the LA, the basal nucleus also eas. Other hypothalamic nuclei innervated by the amyg-
has substantial projections to hippocampus, but in addi- dala are the medial nuclei of the behavior control column
FIG. 3. Summary of the main outputs from the

amygdaloid nuclei.

810 SAH ET AL.
(205). The ventromedial nucleus, which is involved in suggested to be inhibitory (209, 244). Functionally, acti-
reproductive behavior, is also innervated by nuclei related vation of CeA neurons in the rat results in rises in blood
to the olfactory system in the amygdala, particularly the pressure and heart rate. A GABAergic projection from the
medial nucleus, posterior basal nucleus, and posterolat- CeA suggests these fibers are likely to innervate local
eral cortical nucleus. The medial nucleus also sends pro- inhibitory cells in brain stem nuclei. However, no direct
jections to the hypothalamic neuroendocrine zone, mainly evidence for this is available.
to the anterior paraventricular nucleus (45, 205).
In addition to these direct projections to the hypo-
thalamus, the CeA has a strong projection to the BNST, IV. INTRA-AMYGDALOID CONNECTIONS
which also innervates hypothalamic nuclei. Furthermore,
both CeA and BNST have strong projections to ascending Tract tracing studies have revealed that amygdala
monoaminergic and cholinergic neuron groups. These in- nuclei have extensive intranuclear and internuclear con-
clude the noradrenergic locus coeruleus, the dopaminer- nectivity (105, 208). These studies indicate that sensory
gic substantia nigra and ventral tegmental area, the sero- information enters the amygdala through the basolateral
tonergic raphae, and the cholinergic nucleus basalis (8, nuclei, is processed locally, and then follows a predomi-
43, 213). These systems innervate large regions of the nantly lateral to medial progression to the centromedial
forebrain and temporal lobe memory systems as well as nuclei which act as an output station (210). However,
providing inputs to the amygdaloid complex. Rather than little is known about the synaptic physiology of these
the fast, point-to-point excitation mediated by most glu- circuits in the amygdala and how these networks inte-
tamatergic afferents, these ascending systems provide grate incoming information. The connections between the
modulatory inputs that affect information processing over different nuclei in the amygdala have been described in
large cell assemblies. great detail (208, 210). Here we summarize results (Fig. 4)
Large numbers of neurons in the medial subdivision obtained in the rat, although data are also available in cat
of the central nucleus and medial nucleus are GABAergic, (105, 204, 260, 261) and monkey (8). We restrict the
and these projections from the central nucleus have been discussion to connections involving the basolateral com-
FIG. 4. Intra-amygdaloid connections as de-

scribed by anatomical tract tracing studies. Most of
the connections between nuclei within the amygdala
are glutamatergic. See Fig. 1 and text for definitions.

THE AMYGDALA 811
plex and centromedial nuclei since these nuclei are the nucleus are largely restricted to the medial and capsular
best understood functionally. subdivisions. Within the CeA there are extensive intradi-
Within the LA, extensive rostrocaudal as well as in- visional and interdivisional connections (92) with each of
terdivisional connections have been described (211). The the four subdivisions making extensive intranuclear con-
dorsolateral subdivision projects to the medial subdivi- nections. The capsular and lateral subdivision make sig-
sion and to lateral aspects of the lateral subdivision. As nificant projections to the medial and capsular subdivi-
described above, unimodal sensory inputs enter the LA sions with a light projection to the intermediate subdivi-
laterally while the polymodal afferents and projections sion. The medial division largely sends projections out of
from declarative memory systems are largely confined to the amygdala, but also has a moderate projection to the
the medial subdivision (208). The presence of the lateral capsular subdivision (92). It is notable that the lateral
to medial intranuclear connections within the LA suggest subdivision, which forms the largest projections to the
that the medial subdivision might be a site for integration other central subdivisions, receives few reciprocal con-
of sensory information with assessments of past experi- nections. Interestingly, the lateral subdivision receives
ence. The LA sends extensive projections to the basal and extra-amygdaloid inputs from both cortical and subcorti-
accessory basal nuclei and the capsular part of the central cal sources (208), suggesting that this might also be a site
nucleus (211, 260). Of these, the heaviest projection is to for integration of inputs to the amygdaloid complex.
the accessory basal nucleus. Finally, the lateral nucleus In summary, there are extensive connections within
also sends projections to the periamygdaloid cortex. De- and between the different nuclei of the amygdaloid com-
spite early studies suggesting the contrary (8, 213), all plex. These connections indicate that there is extensive
these regions, except the central nucleus (92), send recip- local processing of information entering the amygdala
rocal connections back to the LA (248, 249). It is notable before it leads to the appropriate behavioral outcomes.
that most reciprocal projections terminate in the medial These intranuclear and internuclear connections have
and ventrolateral subdivisions in the lateral amygdala, mostly been studied using anatomical tract tracing tech-
while the dorsolateral subdivision is largely spared. Fur- niques, coupled in some cases with electron microscopic
thermore, these reciprocal connections are modest com- examination of the synaptic specializations. However,
pared with the lateromedial intranuclear connections. physiological studies indicate that amygdala nuclei con-
Most projections to and from the LA make asymmetrical tain many types of cells that cannot be readily distin-
synapses, indicating they are excitatory (249, 260). How- guished on anatomical grounds alone (see below). Fur-
ever, some of the reciprocal connections from the basal thermore, reconstructed neurons in the lateral and basal
nuclei make symmetrical synapses, suggesting that they nuclei show large dendritic trees, and neurons that have
are inhibitory (249). cell bodies in a particular nuclear subdivision (e.g., the
The basal and accessory basal nuclei, which receive dorsolateral subdivision of the lateral nucleus) may well
strong cortical inputs, have extensive internuclear as well have dendrites that extend into the next division (e.g., the
as intranuclear connections. Within the basal nucleus, all medial subdivision of the lateral amygdala) (56, 200, 219).
subdivisions have extensive rostrocaudal connections. This implies that inputs that anatomically terminate in a
The parvicellular subdivision has extensive projections to particular subdivision of these nuclei may well innervate
the magnocellular and intermediate subdivisions (246). neurons whose cell bodies are in a different subdivision.
The largest projection from basal nuclei is to the medial Thus the physiological impact of these local connections
subdivision of the central nucleus (204, 246, 247). These and their implications for information processing remain
afferents form asymmetric synapses with spines and den- elusive.
drites in the central nucleus and are therefore thought to
be glutamatergic (204). Because the hypothalamic and
brain stem projections from the amygdala responsible for V. MORPHOLOGY AND PHYSIOLOGY:
the autonomic responses of amygdala function largely BASOLATERAL COMPLEX
originate from the medial subdivision of the central nu-
cleus, these projections from the basal nuclei to the cen- The cell types present in the amygdala were, as in
tral nucleus have a key role in controlling the output of most other brain regions, initially described using Golgi
processed information from the amygdaloid complex. The techniques. More recently, single-cell recordings have
accessory basal nucleus has extensive rostrocaudal con- been made in both in vivo and in vitro preparations, the
nections and sends afferents to the LA, CeA, and medial cells filled with dyes, and their morphology reconstructed
divisions (247). after physiological recording. These studies have allowed
The central nucleus, which forms a major output of a correlation of morphological and physiological proper-
the amygdala, receives inputs from all the other amygda- ties of neurons in several nuclei. Although Golgi studies
loid nuclei but sends very meager projections back to have been carried out in most regions of the amygdala,
these nuclei (92). The amygdaloid inputs to the central investigations of the electrophysiological characteristics

812 SAH ET AL.
of neurons have centered mainly on the basolateral com- basolateral complex has slightly smaller somata (10 15
plex in vitro in the rat (56, 61, 219, 283) and in vivo in the m) and resembles nonspiny stellate cells of the cortex.
cat (107109, 196, 201). These were termed S, for spiny cells by Hall (73) and
stellate or class II cells by Millhouse and De Olmos
(174). These cells have two to six primary dendrites that
A. Morphology lack spines and form a relatively spherical dendritic field
(109, 150). There is no apparent apical dendrite and, as
Initially, two main types of neuron were described with the pyramidal like neurons, they form a heteroge-
based on Golgi studies. The first type comprises 70% of neous population that has been subdivided into multipo-
the cell population and has been described as pyramidal lar, bitufted, and bipolar cells according to their dendritic
(73, 174, 283), spiny, or class I cells (153, 150). Many have trees by McDonald (150). These neurons are clearly
pyramidal-like somata with three to seven dendrites em- GABAergic (160) and are local circuit interneurons. Their
anating from the soma. The secondary and tertiary den- axons originate from the soma or from the proximal
drites of these cells are spiny. One of the dendrites is portion of a primary dendrite (150). Consistent with local
usually more prominent than the others and thus has been circuit interneurons, the axons branch several times and
likened to the apical dendrite of cortical neurons (56, 73). thus have a cloud of axonal collaterals and terminals
Some neurons appear to have two apical dendrites and near the cell body (174). Some of these interneurons form
are more like the spiny stellate cells of the cortex (155). a pericellular basket or axonal cartridge around the
Unlike pyramidal neurons in the cortex or hippocampus, perikarya and initial segment of pyramidal cells, respec-
these cells are not arranged with parallel apical dendrites tively, allowing a tight inhibitory control over the output of
but are randomly organized, particularly close to the nu- the cell (28, 109, 161, 261).
clear borders (56, 155, 201, 219, 283). Thus, while this cell Like interneurons in other cortical areas, these cells
type has been described as pyramidal, these neurons express several calcium binding proteins (98, 163). About
differ from cortical pyramidal neurons in several ways.
one-half of the cells express parvalbumin, whereas the
The primary dendrite of the apical and basal dendrites is
other half express calbindin and/or calretinin in their
of equivalent length, the dendrites taper rapidly, the distal
cytosol (98, 158), suggesting that there are different
dendrites do not have an elaborate terminal ramification,
classes of interneurons in the basolateral complex. How-
and as mentioned above, there is no rigid orientation of
ever, there is significant overlap between these three
the pyramids in one plane (56, 112). Because of these
markers. While the calretinin and parvalbumin positive
clear differences, it may be more appropriate to call these
neurons form separate populations, a large proportion of
cells pyramidal-like or projection neurons (56, 201). The
the parvalbumin positive cells also express calbindin (98,
axons of these cells originate either from the soma or
from the initial portion of the primary dendrite (56, 150). 163). The functional relevance, if any, of these different
They give off several collaterals within the vicinity of the populations of interneurons is currently not known.
cell before projecting into the efferent bundles of the In addition, although uncommon, several other types
amygdala, showing that they are projection neurons (150, of cells have also been described in the basolateral com-
260). For neurons within the basolateral complex, cells plex on the basis of distinctive axonal or dendritic pat-
described as pyramidal comprise a morphological contin- terns. These have been termed extended neurons, cone
uum ranging from pyramidal to semi-pyramidal to stellate cells, chandelier cells, and neurogliaform cells (61, 95,
(56, 156, 201, 219, 283). However, it should be noted that 153, 150, 174). Extended cells are large cells with long
when reconstructed in coronal sections, cells can some- thick dendrites with few branches and few spines and are
times appear stellate because they have a largely rostro- found in the rostral parts of the basal nucleus. Cone cells,
caudal orientation (56, 174, 204, 283). In general, neurons which have only been described in the rat, have large cell
in the B are somewhat larger than in the LA with an bodies (20 30 m) and cone-shaped dendritic trees that
average soma diameter of 1520 m compared with are nonspiny and are found in the dorsal angle of the
10 15 m in the LA (150, 174). No clear morphological lateral nucleus (174). Chandelier cells resemble cortical
distinctions have been found between neurons in the chandelier cells and have clustered axon varicosities that
different subdivisions of the lateral or basal nuclei. As form synapses with the initial segment of pyramidal like
mentioned above, the large dendritic arbors of pyramidal- neurons (150). Finally, neurogliaform cells are another
like neurons indicate that the dendritic trees of these cells type of small nonspiny stellate neuron found in the baso-
would cover the boundary between subdivisions (56, lateral complex (95, 153, 150). These cells are small (10
200). These considerations call into question the func- m) with a restricted spherical dendritic tree and branch-
tional parcellation of neurons in the basolateral complex ing axons that travel little further than the confines of
into different subdivisions. their dendritic trees. They form numerous synaptic con-
The second main group of cells found within the nections along the dendrites of pyramidal-like neurons

THE AMYGDALA 813
and therefore probably represent inhibitory local circuit afterhyperpolarization (AHP), which is largely responsi-
neurons (150). ble for the spike frequency adaptation (57). The second
population fires short-duration action potentials (half-
width 0.7 ms) and shows little spike frequency adapta-
B. Physiology
tion in response to a prolonged depolarizing current in-
Electrophysiological studies of neurons in the baso- jection (109, 135, 201) (Fig. 6). Due to the similarities with
lateral complex have been made in vivo from cats and in cortical and hippocampal neuron firing properties (109,
vitro in acute brain slices, largely from the rat. These 135), the first type was classified as pyramidal or projec-
neurons have been divided according to whether they are tion neurons and the second as interneurons. This elec-
located in the lateral or basal nucleus. However, no at- trophysiological distinction between projection neurons
tempt has been made to separate neurons located in and local circuit interneurons is similar to that seen in
different subdivisions. In our experience, this is largely other brain regions (37, 148). A detailed analysis of repet-
because internuclear boundaries that can be delineated in itive firing patterns of pyramidal neurons in the lateral
Nissl-stained sections are not readily apparent in acutely nucleus has recently been carried out using whole cell
prepared coronal brain slices when viewed under the light patch-clamp recordings from coronal rat brain sections
microscope. However, the lateral and basal nucleus can (56). These characteristics were then correlated with
be readily distinguished (Fig. 5). morphological properties by filling cells with neurobiotin.
Recordings both in vivo and in vitro from neurons in In this study, cells were classified according to the degree
the LA show extremely low levels of spontaneous activity of spike frequency adaptation that they displayed in re-
(135, 200, 201). Based on their firing properties in response to a prolonged current injection. It was found that
sponse to current injections, neurons in the LA have been pyramidal-like neurons formed a continuum of firing
broadly divided into two types (Fig. 6) (135, 201). The first properties (Fig. 7). At one end of the spectrum cells fire
type, comprising 95% of total cells, fires broad action two to three spikes only and show marked spike fre-
potentials (half-width 1.2 ms measured at 28 30C) and quency accommodation, whereas at the other end of the
shows varying degrees of spike frequency adaptation in spectrum cells fire repetitively throughout the current
response to a prolonged depolarizing current injection. injection with little accommodation (Fig. 7A) (56, 65). In
Action potential trains are followed by a prolonged (15 s) between were cells that fire several times but show clear
FIG. 5. An example of the amygdaloid region as it appears in acutely prepared coronal sections. Left: a Nissl-stained
hemisection of a rat brain around bregma-3. The areas shown in the outlined region are shown in an acutely prepared
coronal brain slice as it appears under brightfield illumination (middle). Shown is the region containing the basolateral
complex and central nucleus. Right: approximate regions of the lateral (LA), basal (B), accessory basal (AB) and central
nucleus have been outlined. In the central nucleus, the approximate locations of the lateral (CeL) and medial (CeM)
subdivisions have also been shown. [Adapted from Paxinos G. and Watson C. The Rat Brain in Stereotaxic Coordinates
(2nd ed.). Sydney, Australia: Academic, 1986.]

814 SAH ET AL.
FIG. 6. Pyramidal-like neurons and

interneurons can be distinguished on elec-
trophysiological grounds. Traces show re-
cordings from typical pyramidal-like neu-
ron and interneuron in the basolateral
complex. Traces on the left are from a
typical pyramidal-like neuron, and those
on the right are from an interneuron. A:
injection of a 400-ms depolarizing current
injection in pyramidal neurons evokes ac-
tion potentials that show spike frequency
adaptation, while similar current injec-
tions into interneurons evoke a high-fre-
quency train of action potentials that do
not adapt. B: action potentials in interneu-
rons have a shorter duration than in pyra-
midal cells.
spike frequency adaptation. The majority of the cells lay (33, 56, 61, 297). In this cell type only a single action
at the end of the spectrum that fired fewer spikes and potential is evoked in response to a prolonged current
showed marked accommodation. These neurons did not injection; the excitability could not be enhanced by giving
show any difference in resting membrane properties. larger current injections or by depolarizing the cell. De-
Quantitative analysis of the morphological properties of spite the marked accommodation that it showed, no pro-
neurons at each end of the electrophysiological spectrum longed AHP followed the action potential (243). These
revealed no significant differences between cells (56, 65). cells appear to express a dendrotoxin-sensitive voltage-
Thus it was concluded that these neurons have differen- gated potassium current that is responsible for their
tial distributions of voltage-gated and calcium-activated marked spike frequency adaptation (E. Faber and P. Sah,
potassium channels that determine their repetitive firing unpublished observations). Thus this cell was considered
properties (56, 57). In accordance with this, cells that to be in a discrete class from the above neurons. Faulkner
show spike frequency adaptation were shown to have and Brown (61) recovered one of these cells for morpho-
larger AHPs than those that fire repetitively (Fig. 7B) logical analysis and found that it was pyramidal-like but
(56). This wide distribution of firing properties is consis- with few spines. In contrast, Yajeya et al. (296) who
tent with the distribution of morphological features that recovered two of these neurons reported them to have a
have been described for projection neurons in the baso- round soma from which four or five spiny dendrites em-
lateral complex (see above). However, no correlation was anated in a spherical fashion. Yajeya et al. (297) have
found between the cells firing properties and their mor- proposed the single firing neurons to represent the neu-
phology (56). rogliaform (type III) cells described by McDonald (150).
Finally, one other cell type, termed a single firer, that Intracellular recordings from LA neurons made in
stands out from the above classification has also been vitro in the cat and guinea pig and in vivo in cats have
described in the LA and comprises 3% of recorded cells found a large proportion of LA projection neurons to

THE AMYGDALA 815
FIG. 7. Electrophysiological and synaptic

properties of different types of pyramidal neu-
ron in the basolateral complex. The traces on
the left show recordings from neurons which
show marked spike frequency adaptation,
while the traces on the right are recordings
from a repetitively firing neuron. A: response
to 600-ms depolarizing current injection (400
pA). The neuron on the left fires a single ac-
tion potential, whereas the cell on the right
fires repetitively throughout the current injec-
tion. Spike frequency adaptation is due to
activation of a slow afterhyperpolarization. B:
the slow afterhyperpolarization (AHP) that
follows a train of action potential. The record-
ing on the left shows the AHP that follows two
action potentials in a cell which shows com-
plete spike frequency adaptation. Inset: re-
sponse to a 100-ms, 400-pA current injection
which evoked two action potentials. Traces
on the right show the AHP evoked in a cell
which fires repetitively during a prolonged
current injections. Note that the AHP that is
evoked is much smaller despite twice the
number of action potentials in response to the
current injection. C: late firing neuron from
the basal nucleus. Traces show the response
to a just threshold current injection and one
that is suprathreshold. Note the long delay to
action potential initiation during a threshold
current injection (left). A suprathreshold cur-
rent injection removes the long delay (right).
D: stimulation of the external capsule evokes
a depolarizing excitatory postsynaptic poten-
tial (inset on left) followed by a hyperpolar-
izing inhibitory postsynaptic potential (IPSP).
The IPSP has two components: a fast compo-
nent mediated by activation of GABAA recep-
tors and a slow component mediated by acti-
vation of GABAB receptors. The amplitudes of
the two components are highly variable be-
tween cells, and two examples are shown.
display intrinsic voltage-dependent oscillations. These in- modate. Action potentials are followed by an AHP. A
creased in frequency in a voltage-dependent manner until small number of nonoscillating bursting neurons were
repetitive spiking was evoked with larger depolarizing also described that fired a burst of two to three action
current injections (196, 201). As first reported in entorhi- potentials before firing in a sustained fashion and showing
nal cortex (6), these oscillations have been suggested to no accommodation. As with the results described in vitro
be due to the activity of subthreshold tetrodotoxin-sensi- (56, 65), no morphological differences were found be-
tive sodium channels (196). Upon depolarization, these tween neurons with different firing properties (201).
cells fire a burst of action potentials followed by a slow Thus, although there are some similarities, there are
rhythmic firing of single spikes, but do not fully accom- also clear differences in the description of pyramidal

816 SAH ET AL.
neurons recorded in vivo in cat and those recorded in age- or calcium-dependent potassium currents leading to
vitro in the rat. First, the in vitro recordings from neurons changes in spike frequency adaptation (57, 285) (E. Faber
in rat slices using whole cell patch-clamp techniques have and P. Sah, unpublished observations).
not shown the membrane oscillations described in vivo. Neurons in the basal nucleus have been examined in
Second, while the firing patterns were not described in vivo in the cat (200, 201). In contrast to LA neurons, which
detail for recordings in vivo, these studies did not de- are virtually silent, basal nucleus neurons were reported
scribe any fully accommodating projection neurons in the to fire in bursts at rest. The bursts of spikes were followed
LA, the major cell type described in vitro. The reason for by a nonadapting train of spikes, due to activation of a
this disparity is not clear. However, the recordings in vivo slow afterdepolarization. These constituted 80% of basal
were made with sharp intracellular microelectrodes, nucleus neurons recorded from. The remaining 20% of
which leads to a lower input resistance due to the mem- neurons in the basal nucleus were nonbursting cells that
brane leak around the electrode. This would have atten- accommodated and showed oscillations similar to those
uated the impact of the AHP in the neurons, and thus all recorded in the LA. After reconstruction of these cells,
cells would appear as repetitively firing neurons. Another they were all described as modified pyramids, and no
possibility is wash out of chemical mediators of the os- consistent differences in morphology were noted. In re-
cillations in whole cell recordings. In fact, in the few cells cordings made in in vitro brain slices, Rainnie et al. (219)
that showed full accommodation in the cat and guinea pig described the bursting cells to be spiny stellate, whereas
LA, the oscillations were absent, suggesting that adapta- Washburn and Moises (283) reported them to be spiny
tion may reflect an inactivation of the oscillations when pyramidal. In contrast, a third study (297) reported the
recorded in the whole cell mode (196). Finally, these repetitively firing neurons to be stellate. The simplest
differences may be species dependent, since microelec- explanation for these discrepancies is likely to be the
trode recordings in the LA in vivo have revealed that large variation in the orientation of the apical dendrite
accommodating neurons can also be found in rats (35). which makes clear classification of cell morphology dif-
Two studies have examined the electrophysiological ficult (201). In summary, in contrast to the discrepancies
properties of neurons in the basal nucleus using intracel- in recordings from LA pyramidal cells, there is more
lular recordings in rat brain slices and correlated them consensus in the properties of neurons in the basal nu-
with their morphological properties (219, 283). As in the cleus.
LA, these cells have been divided into pyramidal or pro- Recordings from interneurons in the basolateral
jection neurons, comprising 95% of the neuronal cell complex have been made both in vitro and in vivo in both
mass, and local circuit interneurons, which comprise the the lateral and basal nuclei. These neurons show a similar
remaining 5%. Pyramidal neurons have been further clas- pattern of physiological properties. In all cases, they gen-
sified into two electrophysiological groups based on their erate narrow action potentials (half width 0.7 ms) and in
firing patterns, burst firing, and repetitive firing. Burst response to a depolarizing current injection fire nonadapt-
firing cells fired one or two spikes before ceasing firing, ing trains of action potentials (109, 135, 201, 284). In
whereas repetitively firing cells fired throughout the cur- contrast to pyramidal neurons, interneurons fire sponta-
rent injection but showed little accommodation (219, 283, neously in vivo at high frequencies (10 15 Hz) (109,
297). A number of intermediate neurons have also been 201). As described above, interneurons can be divided
described (219, 283). Thus as with LA neurons, basal into at least two classes based on their content of calcium
neurons form a continuum of firing patterns. The repeti- binding proteins. However, no differences in physiologi-
tive firing neurons described by Washburn and Moises cal properties between interneurons have been reported.
(283) differ from those in the LA because they show a
delay in firing when depolarized from more negative mem-
brane potentials and have therefore been termed late C. Synaptic Properties
firing neurons (Fig. 7C). This effect has been shown to be
due to the presence of a low-threshold, slowly inactivat- Pyramidal-like neurons in the basolateral complex
ing potassium current (ID) in these neurons (283). Similar show high levels of immunoreactivity for glutamate and
to LA projection neurons, action potentials in basal pro- aspartate (260) but not glutamic acid decarboxylase (28).
jection neurons are broad, and accommodating neurons Thus these cells are presumed to be glutamatergic and
have a significantly larger AHP than repetitively firing form the output cells of this structure. These neurons
neurons. The difference in AHP is most likely the basis of receive both cortical and thalamic inputs which form
the difference in two extremes of firing pattern (56, 243). asymmetrical synapses (58). Consistent with their mor-
The electrophysiological properties of pyramidal neurons phology, these inputs are glutamatergic and form syn-
in the LA and B are subject to modulation by ascending apses containing both -amino-3-hydroxy-5-methyl-4-
and local transmitter systems. Acetylcholine, norepineph- isoxazolepropionic acid (AMPA) and N-methyl-D-aspar-
rine, glutamate, serotonin, and opioids all modulate volt- tate (NMDA) receptors (59, 60). Three types of ionotropic

THE AMYGDALA 817
glutamate receptors, AMPA, NMDA, and kainate recep- glutamate receptor have been suggested to underlie dif-
tors, are recognized in the mammalian central nervous ferent forms of synaptic plasticity in the amygdala (see
system (83). The presence of AMPA and NMDA receptors below).
at excitatory synapses within the central nervous has Glutamate also activates metabotropic receptors that
been known for many years. Electrophysiological studies are coupled via G proteins to phospholipase C or adenylyl
(Figs. 7D and 8) have confirmed that cortical and thalamic cyclase (207). These receptors are found both presynap-
afferents to pyramidal neurons form dual-component glu- tically and postsynaptically in many regions of the central
tamatergic synapses (32, 136, 220, 290). Analysis of spon- nervous system. However, only a few effects resulting
taneous, miniature synaptic currents has shown that from synaptically released glutamate have been described
AMPA and NMDA receptors are present at individual (206). Activation of metabotropic receptors by applica-
synapses in these neurons (136). The properties of these tion of exogenous agonists in basal amygdala neurons has
two inputs are similar with regard to the type of AMPA both presynaptic and postsynaptic actions (221, 222).
receptors that they express. However, it has been sug- However, effects of metabotropic glutamate receptors by
gested that NMDA receptors present at thalamic inputs synaptically released glutamate have only been described
might be different from those present at cortical inputs during the induction of synaptic plasticity (see below).
(291). At most synapses, NMDA receptors are not active Neurons in the LA have also been suggested to have
at resting membrane potentials due to their voltage-de- a fast excitatory inputs mediated by 5-hydroxytryptamine
pendent block by extracellular Mg2 (146, 191). The sug- (5-HT) receptors (264). However, since this initial report,
gestion is that NMDA receptors present at thalamic inputs subsequent experiments have been unable to reproduce
have lower levels of Mg2 block such that they are active these results as all inputs to these neurons can be blocked
at resting membrane potentials (127, 291, but see Ref. with a combination of glutamatergic and GABAergic an-
136). This finding has not been further studied but has tagonists (136, 259, 290). Instead, 5-HT3 receptors in this
major implications for the interpretation of experiments nucleus have been proposed to be present presynaptically
in which NMDA receptors are blocked by specific antag- on interneuron terminals (103, 104).
onists (see below). Interestingly, although heterogeneity in firing proper-
NMDA receptors are hetero-oligomers assembled ties has been described in pyramidal neurons, there have
from two types of subunits, NR1 and NR2. The NR1 been no reports of differences in synaptic properties be-
subunit is a single gene product, whereas the NR2 subunit tween cells, suggesting that the properties of exogenous
is encoded by four different genes: NR2ANR2D (147). inputs to all pyramidal neurons are similar. As discussed
Native NMDA receptors are thought to be heteromulti- above, the axons of pyramidal neurons have substantial
mers containing four or five subunits consisting of two local collaterals (150, 260). Many of the local targets of
NR1 subunits and two or three NR2 subunits (38). At most these collateral are interneurons (262), but they are also
synapses throughout the central nervous system, NMDA likely to contact nearby pyramidal neurons (150). The
receptors are composed of NR1 subunits in combination properties of any of these local connections are not
with either NR2A or NR2B subunits. NR2A and NR2B known.
subunits are ubiquitously distributed through the central Interneurons in the basolateral complex receive ex-
nervous system and have been shown to undergo a de- citatory inputs from local, cortical, and thalamic sources
velopmental switch in hippocampal and cortical neurons (109, 135, 270). In addition, these neurons are connected
(179). At birth NMDA receptors are composed of NR1/ in local networks such that interneurons have synaptic
NR2B subunits, and there is a switch from NR2B to NR2A connections with each other (109). In contrast to pyrami-
subunits around P7. However, in the LA, a recent study dal-like neurons, glutamatergic inputs to interneurons ac-
has shown that application of the NR2B-selective antag- tivate synapses that express few or no NMDA receptors in
onist ifenprodil blocks the induction of fear conditioning, the postsynaptic membrane (135). Furthermore, the
suggesting that receptors containing NR2B subunits are AMPA receptors present at these inputs show marked
present at synapses in the adult lateral amygdala where inward rectification and appear to be calcium permeable
they are involved in initiating synaptic plasticity (227). (135). AMPA receptors are heteromultimers assembled
While both NR2A and NR2B subunits are present in the from four genes, GluR1GluR4. Receptors that lack GluR2
lateral amygdala, the subunit composition of these recep- subunits have strong inward rectification (Fig. 8) and a
tors at synapses in the amygdala has not been determined. high calcium permeability (93, 286). Consistent with the
Recently, the presence of kainate receptors at syn- marked inward rectification reported at these synapses
apses has also been demonstrated (31 , 280). It has been (135), GABAergic cells in the basolateral complex have
suggested that kainate receptors are also present at some been shown to express low levels of GluR2 subunits
glutamatergic inputs to pyramidal neurons in the basal (157). As described above, interneurons in the basolateral
nucleus, where they are proposed to be involved in basal complex are a heterogeneous population of neurons that
synaptic transmission (126). All three types of ionotropic can be separated on morphological grounds and their

818 SAH ET AL.
FIG. 8. Pyramidal cell and interneu-

rons in the basolateral complex have dif-
ferent types of excitatory inputs. Record-
ings are shown from a pyramidal-like neu-
ron (left panels) and interneuron (right
right) in the lateral amygdala. A: excita-
tory synaptic currents recorded at the in-
dicated holding potentials following stim-
ulation of the external capsule. In pyrami-
dal cell, a slower current can be seen with
membrane depolarization, which is absent
in interneurons. Inhibitory synaptic cur-
rents have been blocked with picrotoxin
(100 M). B: peak current (I)-voltage (Vm)
relationships are shown for the fast in-
ward current in pyramidal neurons and
interneurons. Note that the I-V is linear in
pyramidal cells but shows marked out-
ward rectification in interneurons, indicat-
ing the presence of AMPA receptors
which lack GluR2 subunits. C: excitatory
synaptic currents in a pyramidal cell re-
corded at 80 mV and 40 mV (left traces).
Application of the NMDA receptor antago-
nist D-APV (30 M) blocks the slow compo-
nent. Records on the right are synaptic cur-
rents recorded from an interneuron at 80
mV before and after application of CNQX
(10 M), showing that the current is medi-
ated by AMPA receptors.
content of calcium binding proteins. Physiological studies neuron/pyramidal cell recordings, is currently lacking.
of these neurons have not thus far reported differences in However, stimulation of different afferents indicates that
synaptic properties between different cells. However, re- these different interneurons cannot be independently
cent results from our laboratory suggest that some inter- stimulated by extrinsic inputs in vivo (107). In contrast to
neurons in the basolateral complex do express synaptic most other cells, the slow component of the inhibitory
NMDA receptors (A. Woodruff and P. Sah, unpublished synaptic potential in LA pyramidal neurons is in part
observations). generated by a calcium-activated potassium conductance
As in most other parts of the central nervous system, that is activated by calcium influx via NMDA receptors
the fast spiking cells are GABAergic and constitute local (39, 108). This observation raises the possibility that, as
circuit interneurons (160, 189, 202, 209). Activation of recently described in the olfactory bulb (85), NMDA re-
these cells in the basolateral complex generates inhibi- ceptors in lateral amygdala neurons might be coupled to
tory synaptic potentials that have fast and slow compo- calcium-activated potassium channels.
nents (Fig. 7D) (108, 218, 284). As originally described in Interneurons can mediate both feed-forward or feed-
the hippocampus (52), the fast component is mediated by back inhibition (3). In the basolateral complex, whether
GABAA receptors while the slow component is mediated interneurons mediate feed-forward inhibition or feedback
by activation of GABAB receptors (135, 218, 284). Mea- inhibition (or both) has not been fully determined. Elec-
surements of spontaneously occurring miniature inhibi- trophysiological studies in acute slices in the rat have
tory synaptic currents have suggested that different inter- shown that these cells receive both cortical and thalamic
neurons in the basolateral complex are responsible for excitatory inputs, consistent with a role of these cells in
generating the GABAA and GABAB receptor-mediated feed-forward inhibition (109, 135, 270). However, it is
component of inhibitory synaptic current (263). Direct possible that the excitatory inputs to interneurons are due
evidence for this proposal, for example, by paired inter- to activation of axon collaterals of pyramidal cells, which

THE AMYGDALA 819
indicates a feedback role for interneurons. Tract tracing tributed homogeneously throughout the CeA. Immunohis-
studies in the cat and monkey found that cortical affer- tochemical studies have demonstrated the presence of a
ents form few if any synapses with parvalbumin positive wide variety of peptides in cells in the CeA as well as in
neurons, while local afferents do make synapses onto the afferents innervating these neurons (29, 30, 177). One
them (262). In contrast, a similar study in the rat has study has shown that the peptides enkephalin, neuroten-
described thalamic inputs to interneurons in the LA (295). sin and corticotropin releasing hormone (CRH) is found
These findings are consistent with the proposal that dif- in GABAergic neurons. There appear to be two popula-
ferent populations of interneurons can have a feed-for- tions of these cells: one contains enkephalin and the other
ward and/or a feedback role in the basolateral complex. CRH (44). Both populations have a partial overlap with
In summary, pyramidal neurons and local circuit in- neurotensin containing neurons (44, 258). Interestingly,
terneurons in the basolateral nuclei can be separated on intraperitoneal administration of the cytokine interleu-
electrophysiological grounds. As in the cortex, pyramidal kin-1 preferentially activated GABAergic neurons con-
neurons have a range of firing properties. However, unlike taining enkephalin (44), suggesting that neurons with dif-
in the cortex, these different repetitive firing properties ferent peptide content have different functional roles.
are not accompanied by clear morphological differences. Thus neurons in the central nucleus are morphologically
Among the interneurons, several classes of cell can be very different from those found in the basolateral com-
identified based on the presence of different calcium bind- plex with basolateral neurons having similar morphology
ing proteins. The roles of these cells with different firing to cortical structures and the central nucleus being more
properties are not currently understood. However, it striatal-like (73, 149). This finding is consistent with the
seems likely that neurons with differing electrophysiolog- different embryological origins of the two nuclei (214,
ical properties will involve different local circuits and 269). Finally, as with the majority of cells in the striatum,
have distinct afferent/efferent connections. projections from the central nuclei are predominantly
GABAergic while the basolateral nuclei have glutamater-
gic projections (42, 244, 269).
VI. MORPHOLOGY AND PHYSIOLOGY:
CENTRAL NUCLEUS
B. Physiological Properties
A. Morphology
Only a small number of studies have examined the
The morphology of neurons in the central nucleus electrophysiological properties of cells in the central nu-
has been studied using Golgi techniques as well as recon- cleus (144, 190, 251, 252). These recordings have been
struction after recording physiological properties in acute performed in vitro in coronal slices from rat and guinea
brain slices (144, 252). As with the basolateral complex, pig using either whole cell or microelectrode recordings.
the different subdivisions of the central nucleus (30, 105, As discussed above, cells have only been described in the
149, 208) cannot be easily identified in acute slices main- lateral and medial subdivisions as the boundaries for the
tained in vitro (Fig. 5). Thus, while Golgi studies have intermediate and capsular divisions are not apparent in
described neurons in the different subdivisions, results acute slices. At least three types of cells have been de-
from cell fills in slices have either not discussed subdivi- scribed that can be separated by their firing properties
sions (252) or divided cells into those in the lateral and (Fig. 9). Using intracellular recordings with sharp micro-
medial sectors (144, 251). Here we will therefore concen- electrodes, Schiess and co-workers (251, 252) described
trate on cells in the lateral and medial sectors of the two types of cells that they called type A and B cells. Type
central nucleus. There is general agreement that in both A (75%) cells fired throughout a prolonged current in-
subdivisions there is one predominant cell type that has jection, showing little spike frequency adaptation, and
been called medium spiny neurons in the CeL by com- action potentials were followed by a medium-duration
parison with neurons in the nearby striatum (73, 149). AHP (243) in response to short depolarizing current in-
These cells have an ovoid or fusiform soma and three to jections. Type B cells (25%) accommodated and exhib-
five nonspiny primary dendrites from which moderately ited both a medium and slow AHP. The two cell types had
spiny, sparsely branching secondary and tertiary den- similar passive membrane properties other than the rest-
drites arise (144, 149, 252). Axons give off several local ing membrane potential, which was more depolarized in
collaterals before leaving the nucleus. A second type of type B cells. Using whole cell recordings in guinea pig
neuron has also been described that has a somewhat slices in vitro, Martina et al. (144) divided cells into three
larger soma and a thick primary aspiny dendrite that types. The most common type was described as late
tapers into sparsely spiny secondary dendrites (29, 149, firing (95% in CeM and 56% in CeL; Fig. 9C). These cells
252). In addition, a small number of aspiny neurons have displayed a pronounced outward rectification in the de-
also been described (29). These three cell types are dis- polarizing direction (144). In addition to the late-firing

820 SAH ET AL.
neurons, the CeL also contained neurons that fired spikes

repetitively in response to a prolonged current injection
and were termed regular spiking cells (40%, Fig. 9B).
The regular firing cells of Martina et al. (144) likely cor-
respond to type A cells described in the rat (251).
Two further cell types were described in the guinea
pig that were called fast-spiking cells and burst-firing
cells (144). Fast-spiking cells were typical of interneurons
and fired fast action potentials at high frequency, showing
no accommodation. In contrast, burst-firing cells fired
repetitively, showing some accommodation, in response
to a prolonged current injection and fired rebound bursts
of action potentials, riding on a depolarizing potential
(Fig. 9A). These cells therefore are similar to the type B
cells described in the rat (251). However, after a hyper-
polarizing current step, low-threshold bursting neurons in
the guinea pig show a clear rebound depolarization, sim-
ilar to that reported by Scheiss et al. (251) in type A cells.
It should be noted that the whole cell recordings of Mar-
tina et al. (144) were made with potassium gluconate-
containing internal solutions, which makes comparison
with microelectrode recordings difficult since the slow
AHP is very sensitive to the anion present in the internal
solution (299). Thus some of the discrepancies in the
findings of Schiess and co-workers (251, 252) and Martina
et al. (144) may be due to the different recording tech-
niques used. In addition, it is notable that these studies
were done in different species, and it has recently been
suggested that there are differences in the distribution of
cells with distinct properties between rat, cat, and guinea
pig (50). Thus, for example, while late-firing neurons con-
stituted 90% of the cell population in the CeM in guinea
pigs, they accounted for only 2 and 6% of neurons in the
rat and cat CeM, respectively (50).
After recovery of physiologically identified neurons,
cells in the CeL were found to have generally smaller cell
bodies than cells in the CeM (144). However, while in the
rat Scheiss et al. (251) suggested that the two cells types
they found had different cell morphologies, Martina et al.
(144) did not find any systematic correlation between cell
firing properties and their morphology. In both studies,
the major physiological cell type recovered corresponded
to the medium spiny neurons described in Golgi studies.
C. Synaptic Properties
Consistent with tract tracing studies, experiments in

acute slices from rats and guinea pig have shown that
FIG. 9. Physiological properties of three types of neuron in the
central nucleus. Recordings are from the rat central nucleus showing the neurons in both the CeM and CeL receive glutamatergic
three most prominent types of neurons. A: low-threshold spiking neu- inputs from the lateral and basal nucleus which activate
ron. B: regular spiking neuron. C: late firing neuron. In each case, whole both AMPA and NMDA receptors (130, 239). In the guinea
cell recordings were made from neurons in the central nucleus and
depolarizing and hyperpolarizing currents injected as shown. pig, CeL neurons largely receive inputs from the lateral
nucleus while neurons in the medial subdivision receive
inputs from the basal nucleus (239). It should be noted,

THE AMYGDALA 821
however, that when recording from CeL neurons in acute amygdaloid complex has long been known to have a high
brain slices, cells are difficult to separate from those in density of benzodiazepine binding sites (188), and the
the capsular subdivision unless they are filled, recovered, actions of these agents may reflect their action at sites in
and divisional boundaries visualized histologically. the amygdala. Benzodiazepines are thought to produce
Excitatory inputs to CeL neurons also express pre- their anxiolytic actions by enhancing the activity of
synaptic metabotropic glutamate receptors (186). As in GABAA receptor-mediated inhibitory synaptic potentials
many other regions of the central nervous system (132, (192). The presence of a GABA receptor in the central
207), activation of these receptors leads to depression of amygdala with a novel benzodiazepine pharmacology sug-
the synaptic input. The activity of these receptors is mod- gests an alternative mechanism for the anxiolytic actions
ulated following kindling and chronic cocaine treatment of benzodiazepines.
(186). The central amygdala is particularly sensitive to a The subdivisions of the central nucleus have exten-
kindling stimulus (123a) and has been implicated in the sive intradivisional connections (see above) (92). Many of
reinforcing ability of repeated cocaine exposure (170). the neurons in the central nuclei are thought to be
These findings suggest that alterations in synaptic efficacy GABAergic (160, 189, 209). Both morphological (265) and
at inputs to the CeA may be involved in the changes seen electrophysiological (190) studies have indicated the
in epilepsy and behavioral sensitization to cocaine. presence of abundant local GABAergic connections
In the lateral subdivision of the central nucleus, two within the central nucleus. However, direct functional
distinct types of ionotropic GABA receptors have been evidence for this is not currently available.
demonstrated. One type is similar to typical GABAA re-
ceptors and is blocked by low concentrations of bicucul-
VII. MORPHOLOGY AND PHYSIOLOGY:
line and positively modulated by benzodiazepines and
OTHER NUCLEI
barbiturates (293). In addition, these cells express a sec-
ond type of GABA receptor that is markedly less sensitive
In comparison with the large number of studies that
to bicuculline. This bicuculline-resistant component was
have examined properties of cells in the basolateral com-
initially suggested to be due to activation of nicotinic
plex and central nucleus, there are very few detailed
acetylcholine receptors (190). However, recent experi-
studies of cells in the remaining amygdaloid nuclei.
ments have shown that it is blocked by the specific
GABAC receptor antagonist 1,2,5,6-tetrahydropyridine-(4-
yl)methylphosphinic acid (TPMPA) (46). These receptors A. Intercalated Cell Masses
have been called GABAC-like receptors due to their phar-
macological similarity to retinal GABAC receptors (89, The GABAergic intercalated cells (202) that lie in the
217). Bicuculline-insensitive receptors have also been fiber bundles between the basolateral complex (173) and
shown to be present in the medial subdivision (239), the central nucleus act as feed-forward interneurons to
indicating these receptors are present throughout the cen- cells in the CeA, leading to the generation of a disynaptic
tral nucleus. Although bicuculline-insensitive ionotropic inhibitory synaptic potential in these neurons following
GABA receptors have been described in other regions of stimulation in the basolateral complex (47, 203, 239).
the central nervous system (90), the presence of these There are two main types of neuron found in the interca-
receptors at synapses has not been shown outside the lated cell masses. The first, which accounts for the vast
central amygdala. The two GABA receptor types appear majority of cells, has medium (10 15 m) ovoid cell
to be localized to different GABAergic inputs onto CeL bodies with spiny, largely bipolar dendritic trees and ax-
neurons. Thus inputs from the intercalated neurons that ons that send collaterals into the lateral, basal, and central
form synapses onto the dendrites of CeL neurons express nuclei (173). The other type are very large cells (50 m)
both GABAA- and GABAC-like receptors. In contrast, a with very long thick spiny or aspiny dendrites that travel
different input that enters the central nucleus from a in parallel to the borders of the basal, lateral, and central
dorsomedial source activates synapses located on the nuclei (173, 203). These two cell types are very similar to
soma. These somatic synapses express only GABAA re- striatal neurons. Although detailed electrophysiological
ceptors (47). The initial segment of CeL neuron axons is studies of these neurons have not been reported, trains of
spiny (149). It is therefore tempting to speculate that if the action potentials in these neurons are followed by an
somatic GABAergic synapses were made onto these afterdepolarization (ADP) lasting several seconds (240).
spines, their activity would constitute a powerful means Activation of this ADP imparts a heightened excitability to
to inhibit the output of CeL neurons. These results sug- these cells. As the intercalated cells inhibit neurons in the
gest that the different GABA receptors may play distinct CeA, modulation of the activity of these neurons will have
roles in the local circuitry of the central amygdala (47). a significant impact on the output of the CeA. In addition,
Interestingly, the GABAC-like receptor is negatively mod- intercalated neurons are connected in local networks ori-
ulated by benzodiazepines such as diazepam (47). The ented in the lateral to medial direction such that activa-

822 SAH ET AL.
tion of intercalated cells preferentially inhibits neurons in drome, it is associated with some emotional deficits (2)
the medial direction (241). This organization within the including loss of the recognition of fear in others (1).
intercalated cell masses leads to a very specific control of Our understanding of the amygdala and its role in
inhibition in the central nucleus as information passing emotion is hampered by the abstract nature of emotion
through the basolateral complex activates different pop- itself. In humans, bilateral damage restricted to the amyg-
ulations of intercalated cells (241). dala is extremely rare. Animal studies are limited by their
inability to tell us how they feel. Thus much of our
understanding of the role of the amygdala in emotion
B. Medial Nucleus comes from the animal studies on fear (115). Fear, con-
ditioned and unconditioned, elicits a constellation of au-
The medial nucleus contains just one cell type that tonomic and hormonal responses that include cardiac
resembles the main neurons located in the CeM. They are effects (increased blood pressure, changes in heart rate),
small to medium-sized ovoid cells with two to four mod- hormonal effects (release of stress hormones, adrena-
erately spiny primary dendrites (156). There do not ap- line), defecation, vocalization, freezing, and a potentiated
pear to be any local circuit neurons in this structure. startle response (20, 115, 116, 141). These fear response
Neurons in the bed nucleus of the stria terminalis, which patterns are similar in animals and humans (110). Electri-
are similar to the main cell types found in the medial and cal stimulation of the amygdala elicits fear and anxiety
intermediate subdivisions of the central nucleus, have responses in both humans (34, 70) and animals (94), and
medium-sized somata and multipolar spiny dendrites lesions of the amygdala block the expression of certain,
(151). The anterior amygdaloid area contains cells that but not all, types of unconditioned fear. For example, rats
have ovoid somata and three to four primary dendrites with amygdala lesions show reduced freezing in response
that branch sparingly and have few spines (73). Thus they to cats (17) or cat hair (276), attenuated analgesia and
resemble the second class of neuron found in the CeM. heart rate responses to a loud noise (14, 298), and have
The cell types observed in the nucleus of the lateral reduced taste neophobia (182). However, amygdala le-
olfactory tract, the amygdalohippocampal area, and the sions do not affect other measures of fear such as open
cortical nuclei are similar to those in the basolateral arm avoidance in an elevated plus maze (271, 272) or
complex. The majority of the cells are pyramidal-like with analgesia to shock (287).
smaller stellate cells (which resemble the local circuit The amygdala is also necessary for many types of
neurons), spiny stellate cells, and neurogliaform cells also fear-motivated learning. Amygdala lesions disrupt the ac-
present to lesser degrees (156, 152). Orientation of neu- quisition, but not the retention, of both active avoidance
rons in the olfactory areas is more cortical-like with apical (escape from fear) (212) and passive avoidance (124, 235,
dendrites oriented parallel to each other. To our knowl- 271) conditioned responses. Moreover, emotional pro-
edge there have been no detailed studies of the electro- cessing in the amygdala is not limited to fear and aversive
physiological properties of neurons in these other nuclei. stimuli. The amygdala is also involved in conditioning
using appetitive stimuli such as food, sex, and drugs.
Amygdalar lesions disrupt appetitive Pavlovian condition-
VIII. ROLE OF THE AMYGDALOID COMPLEX ing (66, 266), conditioned place preference (54, 55, 166),
and conditioned taste aversion (180, 182, 230, 292) and
It has been known for over a century that the tem- reduce gustatory neophobia (51, 266). Finally, in addition
poral lobe, including the amygdala, is involved in emotion. to the direct role of the amygdala in learning and memory,
In 1888, Brown and Schafer (21) described taming in activation of the amygdala also has a modulatory effect on
monkeys affect associated with temporal lobe retraction. the acquisition and consolidation of memories that evoke
Klu ver and Bucy (102) elaborated on this finding by char- an emotional response (168, 169, 195). Although it is well
acterizing a collection of emotional disturbances caused recognized that the amygdala is involved in a myriad of
by temporal lobe damage, which became known as Klu - memory- and learning-related tasks, the best studied is its
ver-Bucy syndrome. Monkeys with temporal lobe lesions role in Pavlovian fear conditioning and fear-motivated
exhibited an absence of anger and fear, increased explo- operant conditioning. The remainder of this review will
ration, visual agnosia, hyperorality, hypersexualtity, and therefore focus on the neural substrates of classical fear
loss of social interactions. Subsequent work has shown conditioning.
that lesions restricted to the amygdala produced many of
these effects including a loss of fear and anger, increased IX. THE AMYGDALA AND
exploration, and hyperorality (288, 300). The reduced fear FEAR CONDITIONING
and anger, taming effect, of amygdalar lesions is seen in
many animal species (71). While amygdala damage in The neural circuitry of fear conditioning has been
humans rarely results in full-blown Klu ver-Bucy syn- extensively studied and recently reviewed (63, 116, 141).

THE AMYGDALA 823
In Pavlovian fear conditioning, a neutral conditioned stim- the basolateral complex and CeA play key roles in the
ulus (CS) is paired with an aversive unconditioned stim- acquisition and expression of fear-related behaviors. One
ulus (US), such as a mild foot-shock. CS-US pairings prevailing view is that during fear conditioning sensory
result in an association of the CS and US, whereby pre- input reaches the basolateral complex via the LA, which is
sentation of the CS subsequently elicits a conditioned fear the site for CS-US association. The basolateral complex
response (CR), such as freezing. The CS can be a discrete then controls the outputs of the CeA to evoke the behav-
presentation of auditory, visual, olfactory, or tactile stim- ioral and autonomic responses. The pathways involved in
uli or the CS can be contextual, a collection of numerous this model of fear conditioning are outlined in Figure 10.
environmental features. Foot-shocks or loud noises are
typically used as unconditioned stimuli. In rats, lesions to
the amygdala disrupt Pavlovian fear conditioning regard- A. Basolateral Complex
less of stimulus modality or response measures (17). Like-
wise, amygdalar lesions disrupt fear conditioning in non- A converging body of literature has implicated the
human primates and humans (11, 106). Studies over many basolateral complex in assigning affective value to stimuli
years have clearly established that within the amygdala, (27, 43, 82, 116). Anatomically, the basolateral complex is
FIG. 10. Functional model of the basic pathways

and synaptic plasticity proposed to underlie fear condi-
tioning. Sensory stimuli about the conditioned stimulus
(CS) and the unconditioned stimulus (US) converge on
the lateral nucleus of the amygdala. During fear condi-
tioning, convergence of inputs to single neurons results
in enhancement (long-term potentiation) of excitatory
postsynaptic potentials (EPSP) evoked by the CS (B).
This synaptic plasticity enhances the response of LA
projection neurons in response to the CS. CS-evoked
information reaches the CeA directly and via the B and
AB nucleus. Projections from the central nucleus control
the physiological responses, which include behavioral,
autonomic nervous system, and hypothalamic-pituitary
axis responses. As shown in B, the convergence of the
conditioned and unconditioned input to neurons in the
LA has been suggested to result in a larger output in
response to the CS. All projections, except those from
the CeA, are thought to be glutamatergic.

824 SAH ET AL.
well positioned for associative learning. Afferents convey- C. Where Is the Memory Stored?
ing conditioned and unconditioned stimulus information
from the neocortex, thalamus, and hippocampus con- Although there is general agreement about the cir-
verge on the basolateral complex (232). Lesions of the cuitry that participates in fear and fear conditioning, the
basolateral complex block acquisition and expression of exact role played by the different amygdala regions has
fear conditioning (24, 117, 142). Functional inactivation of been questioned. One issue that has been at the forefront
the basolateral complex by infusion of muscimol, a of discussion is whether the amygdala is involved in the
GABAA agonist, into the basolateral complex disrupts fear acquisition and/or storage of fear-related memories or is
conditioning when applied immediately before condition- its role largely in the expression of fear responses. There
ing or during testing, but not when applied immediately is a large amount of literature showing that the amygdala
after conditioning (77, 181, 294). plays a role in memory formation in other neural systems
Different CS modalities are mediated by different (168); however, it has been suggested that the same is also
amygdala afferents. Thalamic medial geniculate and audi- true for fear learning. Thus it has been argued that al-
tory cortical afferents are essential for conditioning to an though the amygdala has a key role in the analysis of
auditory CS (24, 122, 233), while projections from the emotional content, it largely modulates plasticity in other
perirhinal cortex are essential for conditioning to a visual brain regions that are the substrates for memory storage
CS (24, 236, 253). Foot-shock unconditioned information (23, 167). It has been shown that fear conditioning can
is conveyed to the basolateral complex by projections cause synaptic changes in regions outside the amygdala.
from the posterior parietal insula (IC) and the posterior Furthermore, rats can be fear conditioned even following
intralaminar nuclei of the thalamus (PoT/PIL) (253, 257). complete lesions of the basolateral complex (22, 277).
Combined, but not separate, lesions of the IC and PoT/PIL There is abundant evidence that synaptic plasticity
disrupt fear conditioning (234, 253, 257). occurs in the basolateral complex during fear condition-
ing (see below). However, the issue of whether these
changes are necessary and sufficient for fear conditioning
B. Central Nucleus remains to be resolved. The basolateral complex, rather
than simply controlling the CeA, has extensive projec-
As described above, the basolateral complex receives tions to the striatum and prefrontal cortex (27, 53, 54),
both direct and highly processed sensory information. allowing it to influence complex behaviors. Studies dem-
This information is processed locally and then transmit- onstrating that inactivation of the basolateral complex
ted to the central nucleus which projects to hypothalamic (by infusion of lidocaine or tetrodotoxin) many hours
and brain stem areas that mediate the autonomic and after the conditioning can interfere with consolidation of
behavioral signs of fear. Lesions of the CeA block the the fear memory (242, 278) is consistent with the proposal
expression of fear conditioned response using visual or that while plasticity within the amygdala is associated
auditory CS (24, 72, 76, 78, 79 , 99, 298). Furthermore, with fear conditioning, changes in other brain regions,
stimulation of the CeA produces the constellation of con- which require amygdala activation, are also involved.
ditioned fear responses even in the absence of prior fear Apart from fear conditioning, the amygdala is in-
conditioning (86, 97). These findings indicate that the volved in a range of memory tasks. It is well known that
complex behavioral pattern of the fear response is prob- the amygdala is necessary for fear-motivated operant con-
ably hard wired. In fear conditioning, it is only necessary ditioning. Unlike Pavlovian fear conditioning, in the in-
for the conditioned stimulus to activate the CeA; the strumental-avoidance task animals are able to avoid the
CS-US association occurs in or before the CeA. While the aversive stimulus by making the appropriate behavioral
CeA is often considered essential only for expression of response. Amygdala lesions disrupt the acquisition, but
the conditioned fear response (116), this view of the CeA not the retention, of both active avoidance (escape from
is probably an oversimplification. There are no studies fear) (64, 212) and passive avoidance (124, 235, 271)
involving selective reversible inactivation of the CeA dur- conditioned responses. These different learning tasks re-
ing conditioning. There is considerable evidence showing lated to fear can be double dissociated within the amyg-
that CeA is not simply an output pathway of the basolat- dala (99). In this study, one measure was a classical fear
eral complex. Data from other learning paradigms impli- response measured as a reduction in ongoing behavior
cate the CeA in the modulation of attention, arousal, and during CS presentation and was dependent on the CeA.
vigilance during conditioning (43, 82). These effects are This result is consistent with previous findings with CeA
mediated by the CeA via striatal and basal forebrain con- lesions and agrees with the idea that CeA outputs mediate
nections (27, 74, 80, 81). Finally, CeA activation of the behavioral responses. The other behavior, measured con-
cholinergic system modulates neuronal processing in sen- currently with the first, involved an operant choice depen-
sory and learning systems including the basolateral com- dent on the predictive value of the CS. This behavior was
plex, which is highly enriched in cholinergic receptors. independent of the CeA but required the basolateral com-

THE AMYGDALA 825
plex (LA and B). It was notable that the classical fear paired response could be blocked by hyperpolarizing the
responses were unaffected by basolateral lesions, a find- postsynaptic cell during pairing, indicating that the plas-
ing that suggests that information about the CS need not ticity required the postsynaptic neuron.
require activity within the basolateral complex. Although Together, these studies have led to the suggestion
the exact interpretation of these conclusions has been that NMDA receptor-mediated synaptic plasticity (long-
debated (183), it seems clear that while the outputs of the term potentiation, LTP) within the basolateral complex
CeA are involved in one type of emotional learning, the underlies the acquisition and storage of memory related
outputs of the basolateral complex need to be considered to fear conditioning (140). Moreover, it has been sug-
in other types of learning that involve the amygdala (168). gested that these changes occur in projection neurons in
Despite these caveats, simple Pavlovian fear conditioning the LA (116, 117). In agreement with this suggestion, in
remains the single most tractable model in which to ad- vivo recordings have shown that tetanic stimulation of
dress the cellular substrates that might underlie these thalamic inputs to the LA enhance both electrically
learned behaviors. evoked as well as auditory evoked responses in the LA
(228). Unfortunately, whether the changes in synaptically
evoked responses were sensitive to NMDA receptor
X. SYNAPTIC PLASTICITY AND blockade was not tested. In a different set of experiments,
FEAR CONDITIONING McKernan et al. (171) have shown that the amplitude of
AMPA evoked currents at thalamic inputs to neurons in
How does the amygdala mediate memory storage in the LA were larger in animals that had undergone auditory
fear conditioning? As described above, there is consider- fear conditioning. With the use of paired pulse facilitation
able evidence suggesting that the CS-US association oc- as an index of release probability (301), it has been sug-
curs in the basolateral complex. Inputs to the basolateral gested that the site of plasticity during fear conditioning is
complex use glutamate as the transmitter and activate presynaptic (171). Finally, if fear conditioning results
synapses expressing both AMPA and NMDA receptors. from LTP at inputs to neurons in the basolateral complex,
Results from several laboratories have shown that infu- one might expect that the induction of LTP at synapses
sion of the NMDA receptor antagonist D,L-2-amino-5-phos- that have participated in fear conditioning would be oc-
phonovaleric acid (AP5) into the basolateral complex cluded. This has recently been demonstrated at inputs to
blocks the acquisition of amygdala-dependent condition- pyramidal neurons in the LA (273). In summary, there is
ing (24, 68, 100, 176). Within the basolateral complex, convincing evidence that the response of neurons in the
selective lesions of the lateral amygdala disrupt fear con- LA to CS stimulation is enhanced after fear conditioning.
ditioning (9, 72, 117, 184, 281). Recordings of single units The basic properties of LTP, input specificity, coop-
with the LA in vivo have shown that these neurons re- erativity, and associativity, have made it an attractive
spond to both auditory (tone; CS) and somatosensory cellular model of associative learning (18). While LTP has
(shock; US) stimulation (199, 232). Auditory fear condi- been studied at many other synapses, the most exten-
tioning enhances short latency CS firing in LA neurons sively studied form occurs at excitatory synapses be-
(216). The enhanced neuronal firing to the CS observed in tween Schaffer collaterals and CA1 pyramidal neurons in
the LA precedes behavioral expression of the conditioned the hippocampus (133, 137). These studies have shown
response (225). Conditioning-associated increases in fir- that the induction of LTP requires a rise in postsynaptic
ing to the CS are also observed in the auditory cortex, but calcium (138). In the classical model of LTP, depolariza-
this change is subsequent to the firing rate changes in the tion of the postsynaptic neuron relieves the Mg2 block-
LA (215). Furthermore, animals can undergo auditory fear ade of NMDA receptors (191), allowing Ca2 to enter via
conditioning following lesions to either the thalamo- NMDA receptor channels to trigger signal transduction
amygdala or cortico-amygdala pathways, but not both (24, cascades that initiate the molecular changes that underlie
233). Finally, the auditory evoked potentials recorded in LTP (18, 139). The depolarization required to activate
the LA are enhanced in conditioned rats, but not in NMDA receptors is provided either by AMPA receptors
pseudoconditioned controls that receive unpaired CS and (during tetanically induced LTP) or by backpropagating
US presentations (229). action potentials (134). While the postsynaptic induction
Most recently, a study by Rosenkrantz and Grace of this LTP is certain, the final locus of the change that
(238) provides compelling evidence suggesting that the underlies LTP, whether it is presynaptic or postsynaptic,
LA is indeed a site of plasticity during fear conditioning. has been much debated (139). NMDA receptor-indepen-
Using intracellular recordings in vivo, these authors show dent forms of LTP have also been described. In NMDA-
that pairing an odor (CS) with a foot-shock (US) en- independent induction of LTP, postsynaptic calcium
hanced the amplitude of the response to the paired odor. comes from voltage-gated calcium channels (VGCCs)
Simultaneously they also showed that the response to an (296) or calcium-permeable AMPA receptors (135). Fi-
unpaired odor is unaffected. The potentiation of the nally, at some synapses, a presynaptic form of LTP, which

826 SAH ET AL.
does not require NMDA receptors, has also been de- changes in field potentials been correlated with changes
scribed (187). in synaptic potentials (84).
Both cortical and thalamic afferents to the LA are Recent in vitro studies have also demonstrated sev-
capable of long-term plasticity after tetanic stimulation. eral other forms of plasticity in the basolateral amygdala.
Tetanic stimulation of afferents to the basolateral com- First, glutamatergic inputs to interneurons in the basolat-
plex has been shown to result in LTP both in vivo (36, 143, eral complex activate synapses that do not express NMDA
228) and in vitro (10, 32, 33, 84, 289). All studies are in receptors. LTP at these inputs is triggered by a rise in
agreement that a rise in postsynaptic calcium is required calcium by influx via AMPA receptors (135). Because
for the induction of LTP at both cortical and thalamic interneurons are the only source of inhibitory potentials
inputs to basolateral neurons. The site of plasticity has in the basolateral complex, potentiation of excitatory af-
been proposed to be presynaptic with an increase in the ferents to interneurons leads to an increase in the ampli-
probability of transmitter release (84, 143, 273). However, tude of the disynaptic inhibitory potentials (135). Second,
whether induction of LTP at synapses to LA neurons as in the hippocampus, low-frequency stimulation of in-
requires activation of NMDA receptors is not clear. Al- puts to the basal nucleus leads to long-term depression
though some groups have shown NMDA-dependent LTP (LTD) (75, 282). This LTD is input specific and requires
in vitro (84), others have found that tetanically induced activation of metabotropic receptors and a rise in
LTP does not require activation of NMDA receptors (32). postsynaptic calcium levels (75, 128, 223). Lastly, low-
Furthermore, a recent study found that tetanic stimula- frequency stimulation of inputs to the basal nucleus
tion alone was ineffective in inducing LTP, but pairing evokes a slow onset of facilitation of these inputs (125).
action potentials with thalamic excitatory postsynaptic Its induction was shown to result from activation of kai-
potetials did lead to a form of LTP that required activation nate receptors present in the basal nucleus (125). This is
of L-type VGCCs for induction (273, 289). As mentioned a novel form of synaptic plasticity that has not been
previously described, and its relationship to LTD de-
above, NMDA receptors have also been suggested to con-
scribed by others, is not clear. The possible roles of these
tribute to basal transmission at thalamic inputs to the
phenomena in amygdala-mediated learning are not known
lateral amygdala (127, 291). In agreement with this find-
at present. In summary, although the role of the basolat-
ing, behavioral evidence has suggested that AP5 may
eral complex in the induction of fear conditioning is clear,
block the acquisition of fear conditioning by disrupting
there is considerable debate as to the nature of the
normal synaptic transmission in the amygdala (62, 123).
changes that occur during the conditioning.
These results challenge the notion that NMDA receptor-
If the underlying mechanism for the acquisition and
mediated plasticity within the amygdala is the mechanism
storage of fear-related memories is LTP within the amyg-
underlying the acquisition and storage of fear-related
dala, it is useful to consider how this might operate mech-
memories (23). anistically. As originally formulated, LTP results from the
One potential problem has been that all the in vivo coincident activation of two different inputs to single
studies (and some in vitro studies) have relied on field cells. One of these is considered the weak input and
potential measurements of inputs to the amygdaloid com- represents the CS. The other input is a strong input and
plex. However, unlike in clearly layered structures like represents the US. The strong input is capable of acti-
the hippocampus or cerebellum, there is little organiza- vating postsynaptic cells causing a strong depolarization
tion to the architecture of the neurons and where they and/or evoking repetitive action potential firing. The weak
receive their synaptic inputs. Thus, because there are no input, however, is not as effective as the strong input in
clear sources and sinks when inputs are stimulated, it is driving the postsynaptic cell. During conjunctive stimula-
difficult to separate field potentials associated with syn- tion, these two inputs undergo a pairing in which the
aptic currents and those associated with action potentials weak input is potentiated, that is, undergoes LTP and
(143). Thus changes in field potentials are difficult to becomes capable of driving the postsynaptic cell. This is
interpret with regard to the locus of change that is being the principle of associativity. Neurons in the basolateral
measured. Neurons in the lateral amygdala have extensive complex have been shown to respond to both the CS and
local connections (153). Following fear conditioning, an US. While the response to the CS is perhaps not as strong
increase in the correlated firing of neurons in the lateral as to that of the US (215), it is currently unclear how
amygdala have been reported (199, 215), suggesting that CS-US association during fear conditioning would lead to
there might also be changes in the local connections or LTP of the CS inputs. Interestingly, in a recent study,
cell properties following conditioning. Thus it is possible intracellular recordings maintained during the CS-US
that the changes in field potential measurements follow- pairing revealed that plasticity of the CS occurs despite
ing fear conditioning may result from changes in the the lack of significant postsynaptic depolarization during
strength of connections between neurons rather than of the conjunctive stimulation (238). This result suggests
excitatory inputs to these cells. Only in a few studies have that alternate mechanisms to classical LTP may be oper-

THE AMYGDALA 827
ating in the LA during fear conditioning. Conditioning to The major challenge for the future is in the study of
different modalities has been shown to require activation the local physiology and types and properties of receptors
of basolateral neurons. Fear conditioning to one modality and ion channels present in the amygdala. Fear condition-
preserves the response to a different modality showing ing is a relatively simple learning paradigm, and the in-
that the associated plasticity is restricted to the input that volvement of the amygdala in its induction is clear. An
has undergone conditioning (24). In vitro recordings in analysis of the physiological properties of the circuits
acute brain slices prepared from fear-conditioned animals underlying fear conditioning will provide a deeper under-
show enhanced responses when inputs to LA neurons are standing of the neural mechanisms underlying the acqui-
stimulated (171). These results suggest that different mo- sition and storage of emotionally related memories. These
dalities must converge on all neurons and does not re- studies will set the foundations for understanding the
solve how input specificity of conditioning is generated. neurological basis of fear and nervous disorders related to
These considerations show that a simple model involving fear.
plasticity of synapses made by a given input in one struc-
We thank Dr. Denis Pare and Alexander McDonald for
ture cannot entirely account for the complexity of the helpful comments on the manuscript and Dr. Pare for providing
phenomenon of learning. the data for Figure 9.
M. Lopez de Armentia is supported by a grant from the
Secretaria de Estado de Educacion y Universidades (Spain).
XI. CONCLUSIONS Address for reprint requests and other correspondence: P.
Sah, School of Biomedical Sciences, The University of Queens-
Studies performed over the last 50 years have clearly land, Brisbane, Queensland 4072, Australia. (E-mail: pankaj.sah
established the role of the amygdala in a range of related @anu.edu.au).
learning and memory tasks. The anatomy of the amygda-
loid complex, its local connectivity, and afferent and ef- REFERENCES
ferent systems that interact with the amygdaloid complex 1. ADOLPHS R, TRANEL D, AND DAMASIO AR. The human amygdala in
are now known in great detail. However, there is a paucity social judgment. Nature 393: 470 474, 1998.
of knowledge of information transfer through this struc- 2. AGGLETON JP. The contribution of the amygdala to normal and
abnormal emotional states. Trends Neurosci 16: 328 333, 1993.
ture. In addition, although much is known anatomically 3. ALGER BE AND NICOLL RA. Feed-forward dendritic inhibition of rat
about cell types present in most areas of the amygdala, hippocampal pyramidal cells studies in vitro. J Physiol 328: 105
the physiology of cell types has been examined in detail in 123, 1982.
4. ALHEID GF, DE OLMOS J, AND BELTRAMINO CA. Amygdala and ex-
only a few nuclei. Furthermore, an analysis of the physi- tended amygdala. In: The Rat Nervous System, edited by Paxinos
ology of local connections within the amygdaloid nuclei G. Orlando, FL: Academic, 1995, p. 495578.
has largely been ignored. An analysis of the physiology of 5. ALHEID GF AND HEIMER L. New perspectives in basal forebrain
organization of special relevance for neuropsychiatric disorders:
the local circuitry is essential to elucidate the nature of the striatopallidal, amygdaloid, and corticopetal components of
information processing within the amygdala. substantia innominata. Neuroscience 27: 139, 1988.
Studies of fear conditioning have provided the most 6. ALONSO A AND LLINAS RR. Subthreshold Na-dependent theta-like
rhythmicity in stellate cells of entorhinal cortex layer II. Nature
detailed understanding of the role of the amygdala and its 342: 175177, 1989.
afferent and efferent connections in this simple learning 7. AMARAL DG AND INSAUSTI R. Retrograde transport of D-[3H]-aspartate
paradigm. Long-term synaptic plasticity, involving activa- injected into the monkey amygdaloid complex. Exp Brain Res 88:
375388, 1992.
tion of glutamate receptors within the LA, has been pro- 8. AMARAL DG, PRICE JL, PITKA NEN A, AND CARMICHAEL ST. Anatomical
posed as the mechanism underlying the acquisition and organisation of the primate amygdaloid complex. In: The Amyg-
storage of fear conditioning. However, the types of plas- dala: Neurobiological Aspects of Emotion, Memory and Mental
Dysfunction, edited by Aggleton JP. New York: Wiley-Liss, 1992, p.
ticity that are present and their underlying mechanism are 1 66.
just beginning to be studied and in general are not always 9. AMORAPANTH P, LEDOUX JE, AND NADER K. Different lateral amygdala
in good agreement with the behavioral data. The modu- outputs mediate reactions and actions elicited by a fear-arousing
stimulus. Nat Neurosci 3: 74 79, 2000.
lation of processing within the amygdala by neuromodu- 10. ARONIADOU-ANDERJASKA V, POST RM, ROGAWSKI MA, AND LI H. Input-
lators and hormones that are known to play key roles in specific LTP and depotentiation in the basolateral amygdala. Neu-
memory formation (168, 169) has been largely ignored. A roreport 12: 635 640, 2001.
11. BECHARA A, TRANEL D, DAMASIO H, ADOLPHS R, ROCKLAND C, AND
recent report examining the changes in synaptic input to DAMASIO AR. Double dissociation of conditioning and declarative
neurons in the LA following odor-induced fear condition- knowledge relative to the amygdala and hippocampus in humans.
ing has found that that the plasticity induced during con- Science 269: 11151118, 1995.
12. BEHAN M AND HABERLY LB. Intrinsic and efferent connections of the
ditioning required activation of dopaminergic systems in endopiriform nucleus in rat. J Comp Neurol 408: 532548, 1999.
vivo (238). It is therefore tempting to speculate that some 13. BEHBEHANI MM. Functional characteristics of the midbrain periaq-
of the differences between results obtained in vivo and ueductal gray. Prog Neurobiol 46: 575 605, 1995.
14. BELLGOWAN PS AND HELMSTETTER FJ. Neural systems for the expres-
those seen in vitro might result from changes in the sion of hypoalgesia during nonassociative fear. Behav Neurosci
availability of some of these modulatory influences. 110: 727736, 1996.

828 SAH ET AL.
15. BERNARD JF, ALDEN M, AND BESSON JM. The organization of the 37. CONNORS BW AND GUTNICK MJ. Intrinsic firing patterns of diverse
efferent projections from the pontine parabrachial area to the neocortical neurons. Trends Neurosci 13: 99 103, 1990.
amygdaloid complex: a Phaseolus vulgaris leucoagglutinin 38. CULL-CANDY S, BRICKLEY S, AND FARRANT M. NMDA receptor sub-
(PHA-L) study in the rat. J Comp Neurol 329: 201229, 1993. units: diversity, development and disease. Curr Opin Neurobiol 11:
16. BERNARD JF, PESCHANSKI M, AND BESSON JM. A possible spino (tri- 327335, 2001.
gemino)-ponto-amygdaloid pathway for pain. Neurosci Lett 100: 39. DANOBER L AND PAPE HC. Mechanisms and functional significance of
83 88, 1989. a slow inhibitory potential in neurons of the lateral amygdala. Eur
17. BLANCHARD DC AND BLANCHARD RJ. Innate and conditioned reactions J Neurosci 10: 853 867, 1998.
to threat in rats with amygdaloid lesions. J Comp Physiol Psychol 40. DARWIN C. The Expression of the Emotions in Man and Animals.
81: 281290, 1972. Chicago, IL: Univ. of Chicago Press, 1872.
18. BLISS TVP AND COLLINGRIDGE GL. A synaptic model of memory: long 41. DAVIS M. The role of the amygdala in fear and anxiety. Annu Rev
term potentiation in the hippocampus. Nature 361: 3139, 1993. Neurosci 15: 353375, 1992.
19. BORDI F AND LEDOUX JE. Response properties of single units in 42. DAVIS M, RAINNIE D, AND CASSELL M. Neurotransmission in the rat
areas of rat auditory thalamus that project to the amygdala. II. Cells amygdala related to fear and anxiety. Trends Neurosci 17: 208 214,
receiving convergent auditory and somatosensory inputs and cells 1994.
antidromically activated by amygdala stimulation. Exp Brain Res 43. DAVIS M AND WHALEN PJ. The amygdala: vigilance and emotion. Mol
98: 275286, 1994. Psychiatry 6: 1334, 2001.
20. BROWN JS, KALISH HI, AND FARBER IE. Conditioned fear as revealed 44. DAY HE, CURRAN EJ, WATSON SJ JR, AND AKIL H. Distinct neurochem-
by magnitude of startle response to an auditory stimulus. J Exp ical populations in the rat central nucleus of the amygdala and bed
Psychol 41: 317328, 1951. nucleus of the stria terminalis: evidence for their selective activa-
21. BROWN S AND SCHAFER A. An investigation into the functions of the tion by interleukin-1beta. J Comp Neurol 413: 113128, 1999.
occipital and temporal lobes of the monkeys brain. Philos Trans R 45. DAYAS CV, BULLER KM, AND DAY TA. Neuroendocrine responses to
Soc Lond B Biol Sci 179: 303327, 1888. an emotional stressor: evidence for involvement of the medial but
22. CAHILL L, VAZDARJANOVA A, AND SETLOW B. The basolateral amygdala not the central amygdala. Eur J Neurosci 11: 23122322, 1999.
complex is involved with, but is not necessary for, rapid acquisition 46. DELANEY AJ AND SAH P. GABA receptors inhibited by benzodiaz-
of Pavlovian fear conditioning. Eur J Neurosci 12: 3044 3050, epines mediate fast inhibitory transmission in the central amyg-
2000. dala. J Neurosci 19: 9698 9704, 1999.
23. CAHILL L, WEINBERGER NM, ROOZENDAAL B, AND MCGAUGH JL. Is the 47. DELANEY AJ AND SAH P. Pathway-specific targeting of GABA(A)
amygdala a locus of conditioned fear? Some questions and cave- receptor subtypes to somatic and dendritic synapses in the central
ats. Neuron 23: 227228, 1999. amygdala. J Neurophysiol 86: 717723, 2001.
24. CAMPEAU S AND DAVIS M. Involvement of subcortical and cortical 48. DE OLMOS J, HARDY H, AND HEIMER L. Amygdala. In: The Rat Nervous
afferents to the lateral nucleus of the amygdala in fear conditioning System, edited by Paxinos G. Sydney, Australia: Academic, 1985, p.
317223.
measured with fear-potentiated startle in rats trained concurrently
49. DONG HW, PETROVICH GD, AND SWANSON LW. Topography of projec-
with auditory and visual conditioned stimuli. J Neurosci 15: 2312
tions from amygdala to bed nuclei of the stria terminalis. Brain Res
2327, 1995.
38: 192246, 2001.
25. CANNON WB. The James-Lange theory of emotion: a critical exam-
50. DUMONT EC, MARTINA M, SAMSON RD, DROLET G, AND PARE D. Phys-
ination and an alternative theory. Am J Phychol 39: 106 124, 1927.
iological properties of central amygdala neurons: species differ-
26. CANTERAS NS AND SWANSON LW. Projections of the ventral subicu-
ences. Eur J Neurosci 15: 544 552, 2002.
lum to the amygdala, septum, and hypothalamus: a PHAL antero-
51. DUNN LT AND EVERITT BJ. Double dissociations of the effects of
grade tract-tracing study in the rat. J Comp Neurol 324: 180 194,
amygdala and insular cortex lesions on conditioned taste aversion,
1992.
passive avoidance, and neophobia in the rat using the excitotoxin
27. CARDINAL RN, PARKINSON JA, HALL J, AND EVERITT BJ. Emotion and ibotenic acid. Behav Neurosci 102: 323, 1988.
motivation: the role of the amygdala, ventral striatum, and prefron- 52. DUTAR P AND NICOLL RA. A physiological role for GABAB receptors
tal cortex. Neurosci Biobehav Rev 26: 321352, 2002. in the CNS. Nature 332: 156 158, 1988.
28. CARLSEN J. Immunocytochemical localization of glutamate decar- 53. EVERITT BJ, CARDINAL RN, HALL J, PARKINSON JA, AND ROBBINS TW.
boxylase in the rat basolateral amygdaloid nucleus, with special Differential involvement of amygdala subsystems in appetitive con-
reference to GABAergic innervation of amygdalostriatal projection ditioning and drug addiction. In: The Amygdala: A Functional
neurons. J Comp Neurol 273: 513526, 1988. Analysis (2nd ed.), edited by Aggleton JP. New York: Oxford Univ.
29. CASSELL MD AND GRAY TS. Morphology of peptide-immunoreactive Press, 2000, p. 353390.
neurons in the rat central nucleus of the amygdala. J Comp Neurol 54. EVERITT BJ, MORRIS KA, OBRIEN A, AND ROBBINS TW. The basolateral
281: 320 333, 1989. amygdala-ventral striatal system and conditioned place preference:
30. CASSELL MD, GRAY TS, AND KISS JZ. Neuronal architecture in the rat further evidence of limbic-striatal interactions underlying reward-
central nucleus of the amygdala: a cytological, hodological, and related processes. Neuroscience 42: 118, 1991.
immunocytochemical study. J Comp Neurol 246: 478 499, 1986. 55. EVERITT BJ AND ROBBINS TW. The Amygdala: Neurobiological As-
31. CASTILLO PE, MALENKA RC, AND NICOLL RA. Kainate receptors medi- pects of Emotion, Memory and Memory Dysfunction, edited by J.
ate a slow postsynaptic current in hippocampal CA3 neurons. Aggleton. New York: Wiley-Liss, 1992.
Nature 388: 182186, 1997. 56. FABER ESL, CALLISTER RJ, AND SAH P. Morphological and electro-
32. CHAPMAN PF AND BELLAVANCE LL. Induction of long-term potentia- physiological properties of principal neurons in the rat lateral
tion in the basolateral amygdala does not depend on NMDA recep- amygdala in vitro. J Neurophysiol 85: 714 723, 2001.
tor activation. Synapse 11: 310 318, 1992. 57. FABER ESL AND SAH P. Physiological role of calcium-activated po-
33. CHAPMAN PF, KAIRISS EW, KEENAN CL, AND BROWN TH. Long-term tassium currents in the rat lateral amygdala. J Neurosci 22: 1618
synaptic potentiation in the amygdala. Synapse 6: 271278, 1990. 1628, 2002.
34. CHAPMAN WP, SCHROEDER HR, GUYER G, BRAZIER MAB, FAGER C, AND 58. FARB CR, AOKI C, AND LEDOUX JE. Differential localization of NMDA
POPPEN JL. Physiological evidence concerning the importance of and AMPA receptor subunits in the lateral and basal nuclei of the
the amygdaloid nuclear region in the integration of circulating amygdala: a light and electron microscopic study. J Comp Neurol
function and emotion in man. Science 129: 949 950, 1954. 362: 86 108, 1995.
35. CHEN J AND LANG EJ. Inhibitory role of K(Ca) channels in the lateral 59. FARB CR AND LEDOUX JE. NMDA and AMPA receptors in the lateral
amygdaloid (LAT) nucleus. Soc Neurosci Abstr 26: 1726, 2000. nucleus of the amygdala are postsynaptic to auditory thalamic
36. CLUGNET MC AND LEDOUX JE. Synaptic plasticity in fear condition- afferents. Synapse 27: 106 121, 1997.
ing circuits: induction of LTP in the lateral nucleus of the amygdala 60. FARB CR AND LEDOUX JE. Afferents from rat temporal cortex syn-
by stimulation of the medial geniculate body. J Neurosci 10: 2818 apse on lateral amygdala neurons that express NMDA and AMPA
2824, 1990. receptors. Synapse 33: 218 229, 1999.

THE AMYGDALA 829
61. FAULKNER B AND BROWN TH. Morphology and physiology of neurons awake but not anesthetized rats resemble conditioned emotional
in the rat perirhinal-lateral amydala area. J Comp Neurol 411: responses. Brain Res 418: 183188, 1987.
613 642, 1999. 87. JAMES W. What is an emotion? Mind 9: 188 205, 1884.
62. FENDT M. Injections of the NMDA receptor antagonist aminophos- 88. JAMES W. Principles of Psychology. New York: Holt, 1890.
phonopentanoic acid into the lateral nucleus of the amygdala block 89. JOHNSTON GAR. GABAA receptor pharmacology. Pharmacol Ther
the expression of fear-potentiated startle and freezing. J Neurosci 69: 173198, 1996.
21: 4111 4115, 2001. 90. JOHNSTON GAR. GABAC receptors: relatively simple transmitter-
63. FENDT M AND FANSELOW MS. The neuroanatomical and neurochem- gated ion channels? Trends Pharmacol Sci 17: 319 323, 1996.
ical basis of conditioned fear. Neurosci Biobehav Rev 23: 743760, 91. JOHNSTON JB. Further contributions to the study of the evolution of
1999. the forebrain. J Comp Neurol 35: 337 481, 1923.
64. FONBERG E. Effect of partial destruction of the amygdaloid complex 92. JOLKKONEN E AND PITKA NEN A. Intrinsic connections of the rat amyg-
on the emotional-defensive behaviour of dogs. Bull Acad Pol Sci daloid complex: projections originating in the central nucleus.
Biol 13: 429 432, 1965. J Comp Neurol 395: 5372, 1998.
65. FUNAHASHI M, MATSUO R, AND STEWART M. Propagation of synchro- 93. JONAS P, RACCA C, SAKMANN B, SEEBURG PH, AND MONYER H. Differ-
nous burst discharges from entorhinal cortex to morphologically ences in Ca2 permeability of AMPA-type glutamate receptor chan-
and electrophysiologically identified neurons of rat lateral amyg- nels in neocortical neurons caused by differential GluR-B subunit
dala. Brain Res 884: 104 115, 2000. expression. Neuron 12: 12811289, 1994.
66. GALLAGHER M, GRAHAM PW, AND HOLLAND PC. The amygdala central 94. KAADA BR. Somato-motor, autonomic, and electroencephalo-
nucleus and appetitive Pavlovian conditioning: lesions impair one graphic responses to electrical stimulation of rhinencephalic and
class of conditioned behavior. J Neurosci 10: 1906 1911, 1990. other structures in primates, cat, and dog. A study of responses
67. GAURIAU C AND BERNARD JF. Pain pathways and parabrachial cir- from the limbic, subcallodal, orbito-insular, piriform and temporal
cuits in the rat. Exp Physiol 87: 251258, 2002. cortex, hippocampus-fornix, and amygdala. Acta Physiol Scand 23:
68. GEWIRTZ JC AND DAVIS M. Second-order fear conditioning prevented 1285, 1951.
by blocking NMDA receptors in amygdala. Nature 388: 471 474, 95. KAMAL AM AND TO MBO L T. Golgi studies on the amygdaloid nuclei of
1997. the cat. J Hirnforsch 16: 175201, 1975.
69. GLOOR P. Amygdala. In: Handbook of Physiology. Neurophysiology. 96. KAPP BS. Amygdaloid contributions to conditioned arousal and
Washington, DC: Am Physiol Soc, 1960, sect. 1, vol. II, p. 13951420. sensory information processing. In: The Amygdala: Neurobiologi-
70. GLOOR P. Experiential phenomena of temporal lobe epilepsy. Facts cal Aspects of Emotion, Memory and Mental Dysfunction, edited
and hypotheses. Brain 113: 16731694, 1990. by Aggleton JP. New York: Wiley-Liss, 1992, p. 229 254.
71. GODDARD G. Functions of the amygdala. Psychol Bull 62: 89 109, 97. KAPP BS, GALLAGHER M, UNDERWOOD MD, MCNALL CL, AND WHITE-
1964. HORN D. Cardiovascular responses elicited by electrical stimulation
72. GOOSENS KA AND MAREN S. Contextual and auditory fear condition- of the amygdala central nucleus in the rabbit. Brain Res 234:
ing are mediated by the lateral, basal, and central amygdaloid 251262, 1982.
nuclei in rats. Learn Mem 8: 148 155, 2001. 98. KEMPPAINEN S AND PITKANEN A. Distribution of parvalbumin, calreti-
73. HALL E. The amygdala of the cat: a Golgi study. Z Zellforsch nin, and calbindin-D(28k) immunoreactivity in the rat amygdaloid
Mikrosk Anat 134: 439 458, 1972. complex and colocalization with gamma-aminobutyric acid.
74. HAN JS, HOLLAND PC, AND GALLAGHER M. Disconnection of the amyg- J Comp Neurol 426: 441 467, 2000.
dala central nucleus and substantia innominata/nucleus basalis 99. KILLCROSS S, ROBBINS TW, AND EVERITT BJ. Different types of fear-
disrupts increments in conditioned stimulus processing in rats. conditioned behaviour mediated by separate nuclei within amyg-
Behav Neurosci 113: 143151, 1999. dala. Nature 388: 377380, 1997.
75. HEINBOCKEL T AND PAPE HC. Input-specific long-term depression in 100. KIM JJ, DECOLA JP, LANDEIRA-FERNANDEZ J, AND FANSELOW MS. N-
the lateral amygdala evoked by theta frequency stimulation. J Neu- methyl-D-aspartate receptor antagonist APV blocks acquisition but
rosci 20: RC68, 2000. not expression of fear conditioning. Behav Neurosci 105: 126 133,
76. HELMSTETTER FJ. The amygdala is essential for the expression of 1991.
conditional hypoalgesia. Behav Neurosci 106: 518 528, 1992. 101. KLU VER H AND BUCY PC. Preliminary analysis of the temporal lobes
77. HELMSTETTER FJ AND BELLGOWAN PS. Effects of muscimol applied to in monkeys. Arch Neurol Psychiatry 42: 979 100, 1939.
the basolateral amygdala on acquisition and expression of contex- 102. KLU VER H AND BUCY PC. Psychic blindness and other symptoms
tual fear conditioning in rats. Behav Neurosci 108: 10051009, 1994. following bilateral temporal lobectomy in rhesus monkeys. Am J
78. HITCHCOCK J AND DAVIS M. Lesions of the amygdala, but not of the Physiol 119: 352353, 1937.
cerebellum or red nucleus, block conditioned fear as measured 103. KOYAMA S, KUBO C, RHEE JS, AND AKAIKE N. Presynaptic serotonergic
with the potentiated startle paradigm. Behav Neurosci 100: 1122, inhibition of GABAergic synaptic transmission in mechanically
1986. dissociated rat basolateral amygdala neurons. J Physiol 518: 525
79. HITCHCOCK JM AND DAVIS M. Efferent pathway of the amygdala 538, 1999.
involved in conditioned fear as measured with the fear-potentiated 104. KOYAMA S, MATSUMOTO N, KUBO C, AND AKAIKE N. Presynaptic 5-HT3
startle paradigm. Behav Neurosci 105: 826 842, 1991. receptor-mediated modulation of synaptic GABA release in the
80. HOLLAND PC AND GALLAGHER M. Amygdala central nucleus lesions mechanically dissociated rat amygdala neurons. J Physiol 529:
disrupt increments, but not decrements, in conditioned stimulus 373383, 2000.
processing. Behav Neurosci 107: 246 253, 1993. 105. KRETTEK JE AND PRICE JL. A description of the amygdaloid complex
81. HOLLAND PC AND GALLAGHER M. Effects of amygdala central nucleus in the rat and cat with observations on intra-amygdaloid axonal
lesions on blocking and unblocking. Behav Neurosci 107: 235245, connections. J Comp Neurol 178: 255280, 1978.
1993. 106. LABAR KS, LEDOUX JE, SPENCER DD, AND PHELPS EA. Impaired fear
82. HOLLAND PC AND GALLAGHER M. Amygdala circuitry in attentional conditioning following unilateral temporal lobectomy in humans.
and representational processes. Trends Cogn Sci 3: 6573, 1999. J Neurosci 15: 6846 6855, 1995.
83. HOLLMANN M AND HEINEMANN S. Cloned glutamate receptors. Annu 107. LANG EJ AND PARE D. Similar inhibitory processes dominate the
Rev Neurosci 17: 31108, 1994. responses of cat lateral amygdaloid projection neurons to their
84. HUANG YY AND KANDEL ER. Postsynaptic induction and PKA-depen- various afferents. J Neurophysiol 77: 341352, 1997.
dent expression of LTP in the lateral amygdala. Neuron 21: 169 108. LANG EJ AND PARE D. Synaptic and synaptically activated intrinsic
178, 1998. conductances underlie inhibitory potentials in cat lateral amygda-
85. ISAACSON JS AND MURPHY GJ. Glutamate-mediated extrasynaptic loid projection neurons in vivo. J Neurophysiol 77: 353363, 1997.
inhibition: direct coupling of NMDA receptors to Ca2-activated K 109. LANG EJ AND PARE D. Synaptic responsiveness of interneurons of
channels. Neuron 31: 10271034, 2001. the cat lateral amygdaloid nucleus. Neuroscience 83: 877 889,
86. IWATA J, CHIDA K, AND LEDOUX JE. Cardiovascular responses elicited 1998.
by stimulation of neurons in the central amygdaloid nucleus in 110. LANG PJ, DAVIS M, AND OHMAN A. Fear and anxiety: animal models

830 SAH ET AL.
and human cognitive psychophysiology. J Affect Disord 61: 137 signal for Hebbian plasticity in hippocampal neurons. Science 275:
159, 2000. 209 213, 1997.
111. LANGE CG. The mechanisms of emotion. In: The Classical Psychol- 135. MAHANTY NK AND SAH P. Calcium-permeable AMPA receptors me-
ogists, edited by Rand B. Boston: Houghton Mifflin, 1887, p. 672 diate long-term potentiation in interneurons in the amygdala. Na-
684. ture 394: 683 687, 1998.
112. LARKMAN A AND MASON A. Correlations between morphology and 136. MAHANTY NK AND SAH P. Excitatory synaptic inputs to pyramidal
electrophysiology of pyramidal neurons in slices of rat visual cor- neurons of the lateral amygdala. Eur J Neurosci 11: 12171222,
tex. I. Establishment of cell classes. J Neurosci 10: 14071414, 1990. 1999.
114. LEDOUX J. Fear and the brain: where have we been, and where are 137. MALENKA RC, KAUER JA, PERKEL DJ, MAUK MD, KELLY PT, NICOLL RA,
we going? Biol Psychiatry 44: 1229 1238, 1998. AND WAXHAM MN. An essential role for postsynaptic calmodulin and
115. LEDOUX JE. The Emotional Brain. New York: Simon and Schuster, protein kinase activity in long-term potentiation. Nature 340: 554
1996. 557, 1989.
116. LEDOUX JE. Emotion circuits in the brain. Annu Rev Neurosci 23: 138. MALENKA RC, KAUER JA, PERKEL DJ, AND NICOLL RA. The impact of
155184, 2000. postsynaptic calcium on synaptic transmission: its role in long term
117. LEDOUX JE, CICCHETTI P, XAGORARIS A, AND ROMANSKI LM. The lateral potentiation. Trends Neurosci 12: 444 450, 1989.
amygdaloid nucleus: sensory interface of the amygdala in fear 139. MALENKA RC AND NICOLL RA. Long-term potentiation: a decade of
conditioning. J Neurosci 10: 10621069, 1990. progress? Science 285: 1870 1874, 1999.
118. LEDOUX JE, FARB C, AND RUGGIERO DA. Topographic organization of 140. MAREN S. Long-term potentiation in the amygdala: a mechanism for
neurons in the acoustic thalamus that project to the amygdala. emotional learning and memory. Trends Neurosci 22: 561567,
J Neurosci 10: 10431054, 1990. 1999.
119. LEDOUX JE, FARB CR, AND ROMANSKI LM. Overlapping projections to 141. MAREN S. Neurobiology of Pavlovian fear conditioning. Annu Rev
the amygdala and striatum from auditory processing areas of the Neurosci 24: 897931, 2001.
thalamus and cortex. Neurosci Lett 134: 139 144, 1991. 142. MAREN S, AHARONOV G, AND FANSELOW MS. Retrograde abolition of
120. LEDOUX JE, IWATA J, CICCHETTI P, AND REIS DJ. Different projections conditional fear after excitotoxic lesions in the basolateral amyg-
of the central amygdaloid nucleus mediate autonomic and behav- dala of rats: absence of a temporal gradient. Behav Neurosci 110:
ioural correlates of conditioned fear. J Neurosci 8: 25172529, 1988. 718 726, 1996.
121. LEDOUX JE, RUGGIERO DA, FOREST R, STORNETTA R, AND REIS DJ. 143. MAREN S AND FANSELOW MS. Synaptic plasticity in the basolateral
Topographic organization of convergent projections to the thala- amygdala induced by hippocampal formation stimulation in vivo.
mus from the inferior colliculus and spinal cord in the rat. J Comp J Neurosci 15: 7548 7564, 1995.
Neurol 264: 123146, 1987. 144. MARTINA M, ROYER S, AND PARE D. Physiological properties of cen-
122. LEDOUX JE, SAKAGUCHI A, IWATA J, AND REIS DJ. Interruption of tral medial and central lateral amygdala neurons. J Neurophysiol
projections from the medial geniculate body to an archi-neostriatal 82: 18431854, 1999.
field disrupts the classical conditioning of emotional responses to
145. MASCAGNI F, MCDONALD AJ, AND COLEMAN JR. Corticoamygdaloid
acoustic stimuli. Neuroscience 17: 615 627, 1986.
and corticocortical projections of the rat temporal cortex: a
123. LEE HJ, CHOI JS, BROWN TH, AND KIM JJ. Amygdalar NMDA receptors
Phaseolus vulgaris leucoagglutinin study. Neuroscience 57: 697
are critical for the expression of multiple conditioned fear re-
715, 1993.
sponses. J Neurosci 21: 4116 4124, 2001.
146. MAYER ML, WESTBROOK GL, AND GUTHRIE PB. Voltage-dependent
123a.LE GAL LA SALLE G. Amygdaloid kindling in the rat: regional difffer-
block by Mg2 of NMDA responses in spinal cord neurones. Nature
ences and general properties. In: Kindling 2, edited by Wada JA.
309: 263, 1984.
New York: Raven, 1981, p. 31 44.
147. MCBAIN CJ AND MAYER ML. NMDA receptor structure and function.
124. LEHMANN H, TREIT D, AND PARENT MB. Amygdala lesions do not
Physiol Rev 74: 723760, 1994.
impair shock-probe avoidance retention performance. Behav Neu-
rosci 114: 107116, 2000. 148. MCCORMICK DA, CONNORS BW, LIGHTHALL JW, AND PRINCE DA. Com-
125. LI H, CHEN A, XING G, WEI ML, AND ROGAWSKI MA. Kainate receptor- parative electrophysiology of pyramidal and sparsely spiny stellate
mediated heterosynaptic facilitation in the amygdala. Nat Neurosci neurons of the neocortex. J Neurophysiol 54: 782 806, 1985.
4: 612 620, 2001. 149. MCDONALD AJ. Cytoarchitecture of the central amygdaloid nucleus
126. LI H AND ROGAWSKI MA. GluR5 kainate receptor mediated synaptic of the rat. J Comp Neurol 208: 401 418, 1982.
transmission in rat basolateral amygdala in vitro. Neuropharma- 150. MCDONALD AJ. Neurons of the lateral and basolateral amygdaloid
cology 37: 1279 1286, 1998. nuclei: a Golgi study in the rat. J Comp Neurol 212: 293312, 1982.
127. LI XF, PHILLIPS R, AND LEDOUX JE. NMDA and non-NMDA receptors 151. MCDONALD AJ. Neurons of the bed nucleus of the stria terminalis: a
contribute to synaptic transmission between the medial geniculate Golgi study in the rat. Brain Res Bull 10: 111120, 1983.
body and the lateral nucleus of the amygdala. Exp Brain Res 105: 152. MCDONALD AJ. Cytoarchitecture of the nucleus of the lateral olfac-
87100, 1995. tory tract: a Golgi study in the rat. Brain Res Bull 10: 497503, 1983.
128. LIN HC, WANG SJ, LUO MZ, AND GEAN PW. Activation of group II 153. MCDONALD AJ. Neuronal organisation of the lateral and basolateral
metabotropic glutamate receptors induces long-term depression of amygdaloid nuclei of the rat. J Comp Neurol 222: 589 606, 1984.
synaptic transmission in the rat amygdala. J Neurosci 20: 9017 154. MCDONALD AJ. Topographical organization of amygdaloid projec-
9024, 2000. tions to the caudatoputamen, nucleus accumbens, and related stri-
129. LINKE R, BRAUNE G, AND SCHWEGLER H. Differential projection of the atal-like areas of the rat brain. Neuroscience 44: 1533, 1991.
posterior paralaminar thalamic nuclei to the amygdaloid complex 155. MCDONALD AJ. Projection neurons of the basolateral amygdala: a
in the rat. Exp Brain Res 134: 520 532, 2000. correlative Golgi and retrograde tract tracing study. Brain Res Bull
130. LOPEZ DE ARMENTIA M AND SAH P. Excitatory projections to the 28: 179 185, 1992.
central lateral nucleus of the amygdala in the rat. Soc Neurosci 156. MCDONALD AJ. Cell types and intrinsic connections of the amygdala.
Abstr 713: 2, 2001. In: The Amygdala: Neurobiological Aspects of Emotion, Memory
131. LUSKIN MB AND PRICE JL. The topographic organization of associa- and Mental Dysfunction, edited by Aggleton JP. New York: Wiley-
tional fibers of the olfactory system in the rat, including centrifugal Liss, 1992, p. 6796.
fibers to the olfactory bulb. J Comp Neurol 216: 264 291, 1983. 157. MCDONALD AJ. Localization of AMPA glutamate receptor subunits
132. MACDERMOTT AB, ROLE LW, AND SIEGELBAUM SA. Presynaptic iono- in subpopulations of non-pyramidal neurons in the rat basolateral
tropic receptors and the control of transmitter release. Annu Rev amygdala. Neurosci Lett 208: 175178, 1996.
Neurosci 22: 443 485, 1999. 158. MCDONALD AJ. Calbindin-D28k immunoreactivity in the rat amyg-
133. MADISON DV, MALENKA RC, AND NICOLL RA. Mechanisms underlying dala. J Comp Neurol 383: 231244, 1997.
long-term potentiation of synaptic transmission. Annu Rev Neuro- 159. MCDONALD AJ. Cortical pathways to the mammalian amygdala. Prog
sci 14: 379 397, 1991. Brain Res 55: 257332, 1998.
134. MAGEE JC AND JOHNSTON D. A synaptically controlled, associative 160. MCDONALD AJ AND AUGUSTINE JR. Localization of GABA-like immu-

THE AMYGDALA 831
noreactivity in the monkey amygdala. Neuroscience 52: 281294, neophobia, learned taste aversions, and sodium appetite in rats.
1993. J Comp Physiol Psychol 87: 622 643, 1974.
161. MCDONALD AJ AND BETETTE RL. Parvalbumin-containing neurons in 183. NADER K AND LE DOUX JE. Is it time to invoke multiple fear learning
the rat basolateral amygdala: morphology and co-localization of systems in the amygdala? Trends Cogn Sci 1: 241244, 1997.
Calbindin-D(28k). Neuroscience 102: 413 425, 2001. 184. NADER K, MAJIDISHAD P, AMORAPANTH P, AND LEDOUX JE. Damage to
162. MCDONALD AJ AND JACKSON TR. Amygdaloid connections with pos- the lateral and central, but not other, amygdaloid nuclei prevents
terior insular and temporal cortical areas in the rat. J Comp Neurol the acquisition of auditory fear conditioning. Learn Mem 8: 156
262: 59 77, 1987. 163, 2001.
163. MCDONALD AJ AND MASCAGNI F. Colocalization of calcium-binding 185. NAKASHIMA M, UEMURA M, YASUI K, OZAKI HS, TABATA S, AND TAEN A.
proteins and GABA in neurons of the rat basolateral amygdala. An anterograde and retrograde tract-tracing study on the projec-
Neuroscience 105: 681 693, 2001. tions from the thalamic gustatory area in the rat: distribution of
164. MCDONALD AJ AND MASCAGNI F. Projections of the lateral entorhinal neurons projecting to the insular cortex and amygdaloid complex.
cortex to the amygdala: a Phaseolus vulgaris leucoagglutinin study Neurosci Res 36: 297309, 2000.
in the rat. Neuroscience 77: 445 459, 1997. 186. NEUGEBAUER V, ZINEBI F, RUSSELL R, GALLAGHER JP, AND SHINNICK-
165. MCDONALD AJ, MASCAGNI F, AND GUO L. Projections of the medial GALLAGHER P. Cocaine and kindling alter the sensitivity of group II
and lateral prefrontal cortices to the amygdala: a Phaseolus vul- and III metabotropic glutamate receptors in the central amygdala.
garis leucoagglutinin study in the rat. Neuroscience 71: 5575, J Neurophysiol 84: 759 770, 2000.
1996. 187. NICOLL RA AND MALENKA RC. Contrasting properties of two forms of
166. MCDONALD RJ AND WHITE NM. A triple dissociation of memory long-term potentiation in the hippocampus. Nature 377: 115118,
systems: hippocampus, amygdala, and dorsal striatum. Behav Neu- 1995.
rosci 107: 322, 1993. 188. NIEHOFF DL AND KUHAR MJ. Benzodiazepine receptors: localization
167. MCGAUGH J. Memory consolidation and the amygdala: a systems in rat amygdala. J Neurosci 3: 20912097, 1983.
perspective. Trends Neurosci 25: 456, 2002. 189. NITECKA L AND BEN-ARI Y. Distribution of GABA-like immunoreac-
168. MCGAUGH JL. Amygdala: role in modulation of memory storage. In: tivity in the rat amygdaloid complex. J Comp Neurol 266: 4555,
The Amygdala: A Functional Analysis, edited by Aggleton JP. New 1987.
York: Oxford Univ. Press, 2000, p. 391 424. 190. NOSE I, HIGASHI H, INOKUCHI H, AND NISHI S. Synaptic responses of
169. MCGAUGH JL, CAHILL L, AND ROOZENDAAL B. Involvement of the guinea pig and rat central amygdala neurons in vitro. J Neuro-
amygdala in memory storage: interaction with other brain systems. physiol 65: 12271241, 1991.
Proc Natl Acad Sci USA 93: 13508 13514, 1996. 191. NOWAK L, BREGESTOVSKI P, ASCHER P, HERBET A, AND PROCHIANTZ A.
170. MCGREGOR A AND ROBERTS DC. Dopaminergic antagonism within Magnesium gates glutamate-activated channels in mouse central
the nucleus accumbens or the amygdala produces differential ef- neurones. Nature 307: 462 465, 1984.
fects on intravenous cocaine self-administration under fixed and 192. NUTT DJ AND MALIZIA AL. New insights into the role of the GABA(A)-
progressive ratio schedules of reinforcement. Brain Res 624: 245 benzodiazepine receptor in psychiatric disorder. Br J Psychiatry
252, 1993. 179: 390 396, 2001.
171. MCKERNAN MG AND SHINNICK-GALLAGHER P. Fear conditioning in- 193. OTTERSEN OP. Afferent connections to the amygdaloid complex of
duces a lasting potentiation of synaptic currents in vitro. Nature the rat with some observations in the cat. III. Afferents from the
390: 607 611, 1997. lower brain stem. J Comp Neurol 202: 335356, 1981.
172. MCLEAN PD. Psychosomatic disease an the visceral brain. Psy- 194. OTTERSEN OP, FISCHER BO, RINVIK E, AND STORM-MATHISEN J. Putative
chosom Med 11: 338 353, 1949. amino acid transmitters in the amygdala. Adv Exp Med Biol 203:
173. MILLHOUSE OE. The intercalated cells of the amygdala. J Comp 53 66, 1986.
Neurol 247: 246 271, 1986. 195. PACKARD MG AND CAHILL L. Affective modulation of multiple mem-
174. MILLHOUSE OE AND DEOLMOS J. Neuronal configurations in lateral ory systems. Curr Opin Neurobiol 11: 752756, 2001.
and basolateral amygdala. Neuroscience 10: 1269 1300, 1983. 196. PAPE HC, PARE D, AND DRIESANG RB. Two types of intrinsic oscilla-
175. MILNER B, SQUIRE LR, AND KANDEL ER. Cognitive neuroscience and tions in neurons of the lateral and basolateral nuclei of the amyg-
the study of memory. Neuron 20: 445 468, 1998. dala. J Neurophysiol 79: 205216, 1998.
176. MISERENDINO MJD, SANANES CB, MELIA KR, AND DAVIS M. Blocking of 197. PAPEZ JW. A proposed mechanis of emotion. Arch Neurol Psych 38:
acquisition but not expression of conditioned fear-potentiated star- 725743, 1937.
tle by NMDA antagonists in the amygdala. Nature 345: 716 718, 198. PARE D. Mechanisms of Pavlovian fear conditioning: has the en-
1990. gram been located? Trends Neurosci 25: 436, 2002.
177. MOGA MM AND GRAY TS. Evidence for corticotropin-releasing factor, 199. PARE D AND COLLINS DR. Neuronal correlates of fear in the lateral
neurotensin, and somatostatin in the neural pathway from the amygdala: multiple extracellular recordings in conscious cats.
central nucleus of the amygdala to the parabrachial nucleus. J Neurosci 20: 27012710, 2000.
J Comp Neurol 241: 275284, 1985. 200. PARE D AND GAUDREAU H. Projection cells and interneurons of the
178. MOGA MM, HERBERT H, HURLEY KM, YASUI Y, GRAY TS, AND SAPER CB. lateral and basolateral amygdala: distinct firing patterns and differ-
Organization of cortical, basal forebrain, and hypothalamic affer- ential relation to theta and delta rhythms in conscious cats. J Neu-
ents to the parabrachial nucleus in the rat. J Comp Neurol 295: rosci 16: 3334 3350, 1996.
624 661, 1990. 201. PARE D, PAPE HC, AND DONG J. Bursting and oscillating neurons of
179. MONYER H, BURNASHEV N, LAURIE DJ, SAKMANN B, AND SEEBURG PH. the cat basolateral amygdaloid complex in vivo: electrophysiolog-
Developmental and regional expression in the rat brain and func- ical properties and morphological features. J Neurophysiol 74:
tional properties of four NMDA receptors. Neuron 12: 529 540, 1179 1191, 1995.
1994. 202. PARE D AND SMITH Y. Distribution of GABA immunoreactivity in the
180. MORRIS R, FREY S, KASAMBIRA T, AND PETRIDES M. Ibotenic acid amygdaloid complex of the cat. Neuroscience 57: 10611076, 1993.
lesions of the basolateral, but not the central, amygdala interfere 203. PARE D AND SMITH Y. The intercalated cell masses project to the
with conditioned taste aversion: evidence from a combined behav- central and medial nuclei of the amygdala in cats. Neuroscience 57:
ioral and anatomical tract-tracing investigation. Behav Neurosci 10771090, 1993.
113: 291302, 1999. 204. PARE D, SMITH Y, AND PARE JF. Intra-amygdaloid projections of the
181. MULLER J, CORODIMAS KP, FRIDEL Z, AND LEDOUX JE. Functional basolateral and basomedial nuclei in the cat: Phaseolus vulgaris-
inactivation of the lateral and basal nuclei of the amygdala by leucoagglutinin anterograde tracing at the light and electron micro-
muscimol infusion prevents fear conditioning to an explicit condi- scopic level. Neuroscience 69: 567583, 1995.
tioned stimulus and to contextual stimuli. Behav Neurosci 111: 205. PETROVICH GD, CANTERAS NS, AND SWANSON LW. Combinatorial
683 691, 1997. amygdalar inputs to hippocampal domains and hypothalamic be-
182. NACHMAN M AND ASHE JH. Effects of basolateral amygdala lesions on havior systems. Brain Res 38: 247289, 2001.

832 SAH ET AL.
206. PIN JP. Synaptic transmission. The two faces of glutamate. Nature sition but not the expression of fear conditioning. J Neurosci 21:
394: 19 21, 1998. 6889 6896, 2001.
207. PIN JP AND DUVOISIN R. The metabotropic glutamate receptors: 228. ROGAN MT AND LEDOUX JE. LTP is accompanied by commensurate
structure and functions. Neuropharmacology 34: 126, 1995. enhancement of auditory-evoked responses in a fear conditioning
208. PITKA NEN A. Connectivity of the rat amygdaloid complex. In: The circuit. Neuron 15: 127136, 1995.
Amygdala: A Functional Analysis, edited by Aggleton JP. Oxford, 229. ROGAN MT, STAUBLI UV, AND LEDOUX JE. Fear conditioning induces
UK: Oxford Univ. Press, 2000, p. 31115. associative long-term potentiation in the amygdala. Nature 390:
209. PITKA NEN A AND AMARAL DG. The distribution of GABAergic cells, 604 607, 1997.
fibers, and terminals in the monkey amygdaloid complex: an im- 230. ROLLINS BL, STINES SG, MCGUIRE HB, AND KING BM. Effects of
munohistochemical and in situ hybridization study. J Neurosci 14: amygdala lesions on body weight, conditioned taste aversion, and
2200 2224, 1994. neophobia. Physiol Behav 72: 735742, 2001.
210. PITKA NEN A, SAVANDER V, AND LEDOUX JE. Organization of intra- 231. ROLLS ET. The Brain and Emotion. New York: Oxford Univ. Press,
amygdaloid circuitries in the rat: an emerging framework for un- 1999.
derstanding functions of the amygdala. Trends Neurosci 20: 517 232. ROMANSKI LM, CLUGNET MC, BORDI F, AND LEDOUX JE. Somatosen-
523, 1997. sory and auditory convergence in the lateral nucleus of the amyg-
211. PITKA NEN A, STEFANACCI L, FARB CR, GO GG, LEDOUX JE, AND AMARAL dala. Behav Neurosci 107: 444 450, 1993.
DG. Intrinsic connections of the rat amygdaloid complex: projec- 233. ROMANSKI LM AND LEDOUX JE. Equipotentiality of thalamo-amygdala
tions originating in the lateral nucleus. J Comp Neurol 356: 288 and thalamo-cortico-amygdala circuits in auditory fear condition-
310, 1995. ing. J Neurosci 12: 4501 4509, 1992.
212. POREMBA A AND GABRIEL M. Amygdala neurons mediate acquisition 234. ROMANSKI LM AND LEDOUX JE. Information cascade from primary
but not maintenance of instrumental avoidance behavior in rabbits. auditory cortex to the amygdala: corticocortical and corticoamyg-
J Neurosci 19: 96359641, 1999. daloid projections of temporal cortex in the rat. Cereb Cortex 3:
213. PRICE JL, RUSSCHEN FT, AND AMARAL DG. The Limbic Region. II: The 515532, 1993.
Amygdaloid Complex. New York: Elsevier Science, 1987. 235. ROOZENDAAL B, KOOLHAAS JM, AND BOHUS B. The central amygdala is
214. PUELLES L. Thoughts on the development, structure and evolution involved in conditioning but not in retention of active and passive
of the mammalian and avian telencephalic pallium. Philos Trans R shock avoidance in male rats. Behav Neural Biol 59: 143149, 1993.
Soc Lond B Biol Sci 356: 15831598, 2001. 236. ROSEN JB, HITCHCOCK JM, MISERENDINO MJ, FALLS WA, CAMPEAU S,
AND DAVIS M. Lesions of the perirhinal cortex but not of the frontal,
215. QUIRK GJ, ARMONY JL, AND LEDOUX JE. Fear conditioning enhances
different temporal components of tone-evoked spike trains in au- medial prefrontal, visual, or insular cortex block fear-potentiated
ditory cortex and lateral amygdala. Neuron 19: 613 624, 1997. startle using a visual conditioned stimulus. J Neurosci 12: 4624
216. QUIRK GJ, REPA C, AND LEDOUX JE. Fear conditioning enhances 4633, 1992.
short-latency auditory responses of lateral amygdaloid neurons: 237. ROSEN JB AND SCHULKIN J. From normal fear to pathological anxiety.
Psychol Rev 105: 325350, 1998.
parallel recordings in the freely behaving rat. Neuron 15: 1029
238. ROSENKRANZ JA AND GRACE AA. Dopamine-mediated modulation of
1039, 1995.
odour-evoked amygdala potentials during pavlovian conditioning.
217. RAGOZZINO D, WOODWARD RM, MURATA Y, EUSEBI F, OVERMAN LE, AND
Nature 417: 282287, 2002.
MILEDI R. Design and in vitro pharmacology of a selective -ami-
239. ROYER S, MARTINA M, AND PARE D. An inhibitory interface gates
nobutyric acidc receptor antagonist. Mol Pharmacol 50: 1024 1030,
impulse traffic between the input and output stations of the amyg-
1996.
dala. J Neurosci 19: 1057510583, 1999.
218. RAINNIE DG, ASPRODINI EK, AND SHINNICK-GALLAGHER P. Inhibitory
240. ROYER S, MARTINA M, AND PARE D. Bistable behavior of inhibitory
transmission in the basolateral amygdala. J Neurophysiol 66: 999
neurons controlling impulse traffic through the amygdala: role of a
1009, 1991.
slowly deinactivating K current. J Neurosci 20: 9034 9039, 2000.
219. RAINNIE DG, ASPRODINI EK, AND SHINNICK-GALLAGHER P. Intracellular 241. ROYER S, MARTINA M, AND PARE D. Polarized synaptic interactions
recordings from morphologically indentified neurons of the baso- between intercalated neurons of the amygdala. J Neurophysiol 83:
lateral amygdala. J Neurophysiol 69: 1350 1362, 1993. 3509 3518, 2000.
220. RAINNIE DG, EFTIHIA K, ASPRODINI K, AND SHINNICK-GALLAGHER P. 242. SACCHETTI B, LORENZINI CA, BALDI E, TASSONI G, AND BUCHERELLI C.
Excitatory transmission in the basolateral amygdala. J Neuro- Auditory thalamus, dorsal hippocampus, basolateral amygdala, and
physiol 66: 986 997, 1991. perirhinal cortex role in the consolidation of conditioned freezing
221. RAINNIE DG, HOLMES KH, AND SHINNICK-GALLAGHER P. Activation of to context and to acoustic conditioned stimulus in the rat. J Neu-
postsynaptic metabotropic glutamate receptors by trans-ACPD hy- rosci 19: 9570 9578, 1999.
perpolarizes neurons of the basolateral amygdala. J Neurosci 14: 243. SAH P. Ca2-activated K currents in neurones: types, physiological
7208 7220, 1994. roles and modulation. Trends Neurosci 19: 150 154, 1996.
222. RAINNIE DG AND SHINNICK-GALLAGHER P. Trans-ACPD and L-APB 244. SAHA S, BATTEN TF, AND HENDERSON Z. A GABAergic projection from
presynaptically inhibit excitatory glutamatergic transmission in the the central nucleus of the amygdala to the nucleus of the solitary
basolateral amygdala (BLA). Neurosci Lett 139: 8791, 1992. tract: a combined anterograde tracing and electron microscopic
223. RAMMES G, EDER M, DODT HU, KOCHS E, AND ZIEGLGANSBERGER W. immunohistochemical study. Neuroscience 99: 613 626, 2000.
Long-term depression in the basolateral amygdala of the mouse 245. SARTER M AND MARKOWITSCH HJ. Involvement of the amygdala in
involves the activation of interneurons. Neuroscience 107: 8597, learning and memory: a critical review, with emphasis on anatom-
2001. ical relations. Behav Neurosci 99: 342380, 1985.
224. RAY JP AND PRICE JL. The organization of the thalamocortical con- 246. SAVANDER V, GO CG, LEDOUX JE, AND PITKA NEN A. Intrinsic connec-
nections of the mediodorsal thalamic nucleus in the rat, related to tions of the rat amygdaloid complex: projections originating in the
the ventral forebrain-prefrontal cortex topography. J Comp Neurol basal nucleus. J Comp Neurol 361: 345368, 1995.
323: 167197, 1992. 247. SAVANDER V, GO CG, LEDOUX JE, AND PITKANEN A. Intrinsic connec-
225. REPA JC, MULLER J, APERGIS J, DESROCHERS TM, ZHOU Y, AND LEDOUX tions of the rat amygdaloid complex: projections originating in the
JE. Two different lateral amygdala cell populations contribute to accessory basal nucleus. J Comp Neurol 374: 291313, 1996.
the initiation and storage of memory. Nat Neurosci 4: 724 731, 248. SAVANDER V, LEDOUX JE, AND PITKA NEN A. Topographic projections
2001. from the periamygdaloid cortex to select subregions of the lateral
226. RIZVI TA, ENNIS M, BEHBEHANI MM, AND SHIPLEY MT. Connections nucleus of the amygdala in the rat. Neurosci Lett 211: 167170,
between the central nucleus of the amygdala and the midbrain 1996.
periaqueductal gray: topography and reciprocity. J Comp Neurol 249. SAVANDER V, MIETTINEN R, LEDOUX JE, AND PITKA NEN A. Lateral
303: 121131, 1991. nucleus of the rat amygdala is reciprocally connected with basal
227. RODRIGUES SM, SCHAFE GE, AND LEDOUX JE. Intra-amygdala block- and accessory basal nuclei: a light and electron microscopic study.
ade of the NR2B subunit of the NMDA receptor disrupts the acqui- Neuroscience 77: 767781, 1997.

THE AMYGDALA 833
250. SCALIA F AND WINANS SS. The differential projections of the olfac- VY. Fear conditioning occludes LTP-induced presynaptic enhance-
tory bulb and accessory olfactory bulb in mammals. J Comp Neurol ment of synaptic transmission in the cortical pathway to the lateral
161: 3155, 1975. amygdala. Neuron 34: 289 300, 2002.
251. SCHIESS MC, ASPRODINI EK, RAINNIE DG, AND SHINNICK-GALLAGHER P. 274. TURNER BH AND HERKENHAM M. Thalamoamygdaloid projections in
The central nucleus of the rat amygdala: in vitro intracellular the rat: a test of the amygdalas role in sensory processing. J Comp
recordings. Brain Res 604: 283297, 1993. Neurol 313: 295325, 1991.
252. SCHIESS MC, CALLAHAN PM, AND ZHENG H. Characterization of the 275. VAN DER KOOY D, KODA LY, MCGINTY JF, GERFEN CR, AND BLOOM FE.
electrophysiological and morphological properties of rat central The organization of projections from the cortex, amygdala, and
amygdala neurons in vitro. J Neurosci Res 58: 663 673, 1999. hypothalamus to the nucleus of the solitary tract in rat. J Comp
253. SHI C AND DAVIS M. Visual pathways involved in fear conditioning Neurol 224: 124, 1984.
measured with fear-potentiated startle: behavioral and anatomic 276. VAZDARJANOVA A, CAHILL L, AND MCGAUGH JL. Disrupting basolateral
studies. J Neurosci 21: 9844 9855, 2001. amygdala function impairs unconditioned freezing and avoidance
254. SHI CJ AND CASSELL MD. Cortical, thalamic, and amygdaloid projec- in rats. Eur J Neurosci 14: 709 718, 2001.
tions of rat temporal cortex. J Comp Neurol 382: 153175, 1997. 277. VAZDARJANOVA A AND MCGAUGH JL. Basolateral amygdala is not
255. SHI CJ AND CASSELL MD. Cortical, thalamic, and amygdaloid con- critical for cognitive memory of contextual fear conditioning. Proc
nections of the anterior and posterior insular cortices. J Comp Natl Acad Sci USA 95: 1500315007, 1998.
Neurol 399: 440 468, 1998. 278. VAZDARJANOVA A AND MCGAUGH JL. Basolateral amygdala is involved
256. SHI CJ AND CASSELL MD. Cascade projections from somatosensory in modulating consolidation of memory for classical fear condition-
cortex to the rat basolateral amygdala via the parietal insular ing. J Neurosci 19: 6615 6622, 1999.
cortex. J Comp Neurol 399: 469 491, 1998. 279. VEENING JG, SWANSON LW, AND SAWCHENKO PE. The organization of
257. SHI CJ AND CASSELL MD. Perirhinal cortex projections to the amyg- projections from the central nucleus of the amygdala to brainstem
daloid complex and hippocampal formation in the rat. J Comp sites involved in central autonomic regulation: a combined retro-
Neurol 406: 299 328, 1999. grade transport-immunohistochemical study. Brain Res 303: 337
258. SHIMADA S, INAGAKI S, KUBOTA Y, OGAWA N, SHIBASAKI T, AND TAKAGI 357, 1984.
H. Coexistence of peptides (corticotropin releasing factor/neuro- 280. VIGNES M AND COLLINGRIDGE GL. The synaptic activation of kainate
tensin and substance P/somatostatin) in the bed nucleus of the stria receptors. Nature 388: 179 182, 1997.
terminalis and central amygdaloid nucleus of the rat. Neuroscience 281. WALLACE KJ AND ROSEN JB. Neurotoxic lesions of the lateral nucleus
30: 377383, 1989. of the amygdala decrease conditioned fear but not unconditioned
259. SMITH BN AND DUDEK FE. Amino acid-mediated regulation of spon- fear of a predator odor: comparison with electrolytic lesions.
taneous synaptic activity patterns in the rat basolateral amygdala. J Neurosci 21: 3619 3627, 2001.
J Neurophysiol 76: 1958 1967, 1996. 282. WANG SJ AND GEAN PW. Long-term depression of excitatory synap-
260. SMITH Y AND PARE D. Intra-amygdaloid projections of the lateral tic transmission in the rat amygdala. J Neurosci 19: 10656 10663,
nucleus in the cat: PHA-L anterograde labeling combined with 1999.
postembedding GABA and glutamate immunocytochemistry. 283. WASHBURN MS AND MOISES HC. Electrophysiological and morpho-
J Comp Neurol 342: 232248, 1994. logical properties of rat basolateral amygdaloid neurons in vitro.
261. SMITH Y, PARE JF, AND PARE D. Cat intra-amygdaloid inhibitory J Neurosci 12: 4066 4079, 1992.
network: ultrastructural organization of parvalbumin-immunoreac- 284. WASHBURN MS AND MOISES HC. Inhibitory responses of rat basolat-
tive elements. J Comp Neurol 391: 164 179, 1998. eral amygdaloid neurons recorded in vitro. Neuroscience 50: 811
262. SMITH Y, PARE JF, AND PARE D. Differential innervation of parvalbu- 830, 1992.
min-immunoreactive interneurons of the basolateral amygdaloid 285. WASHBURN MS AND MOISES HC. Muscarinic responses of rat basolat-
complex by cortical and intrinsic inputs. J Comp Neurol 416: eral amygadaloid neurons recorded in vitro. J Physiol 449: 121154,
496 508, 2000. 1992.
263. SUGITA S, JOHNSON SW, AND NORTH RA. Synaptic inputs to GABAA 286. WASHBURN MS, NUMBERGER M, ZHANG S, AND DINGLEDINE R. Differen-
and GABAB receptors originate from discrete afferent neurons. tial dependence on GluR2 expression of three characteristic fea-
Neurosci Lett 134: 207211, 1992. tures of AMPA receptors. J Neurosci 17: 93939406, 1997.
264. SUGITA S, SHEN KZ, AND NORTH RA. 5-Hydroxytryptamine is a fast 287. WATKINS LR, WIERTELAK EP, AND MAIER SF. The amygdala is neces-
excitatory transmitter at 5-HT3 receptors in rat amygdala. Neuron sary for the expression of conditioned but not unconditioned an-
8: 199 203, 1992. algesia. Behav Neurosci 107: 402 405, 1993.
265. SUN N AND CASSELL MD. Intrinsic GABAergic neurons in the rat 288. WEISKRANTZ L. Behavioural changes associated with ablation of the
central extended amygdala. J Comp Neurol 330: 381 404, 1993. amygdaloid complex in monkeys. J Comp Physiol Pharmacol 49:
266. SUTHERLAND RJ AND MCDONALD RJ. Hippocampus, amygdala, and 129 158, 1956.
memory deficits in rats. Behav Brain Res 37: 5779, 1990. 289. WEISSKOPF MG, BAUER EP, AND LEDOUX JE. L-type voltage-gated
267. SWANSON LW. Cerebral hemisphere regulation of motivated behav- calcium channels mediate NMDA-independent associative long-
ior (1). Brain Res 886: 113164, 2000. term potentiation at thalamic input synapses to the amygdala.
268. SWANSON LW AND KUYPERS HG. The paraventricular nucleus of the J Neurosci 19: 1051210519, 1999.
hypothalamus: cytoarchitectonic subdivisions and organization of 290. WEISSKOPF MG AND LEDOUX JE. In vitro analysis of synaptic input to
projections to the pituitary, dorsal vagal complex, and spinal cord the dorsal subdivision of the lateral amygdala. Soc Neurosci Abstr
as demonstrated by retrograde fluorescence double-labeling meth- 23: 786, 1997.
ods. J Comp Neurol 194: 555570, 1980. 291. WEISSKOPF MG AND LEDOUX JE. Distinct populations of NMDA
269. SWANSON LW AND PETROVICH GD. What is the amygdala? Trends receptors at subcortical and cortical inputs to principal cells of the
Neurosci 21: 323331, 1998. lateral amygdala. J Neurophysiol 81: 930 934, 1999.
270. SZINYEI C, HEINBOCKEL T, MONTAGNE J, AND PAPE HC. Putative corti- 292. WELZL H, DADAMO P, AND LIPP HP. Conditioned taste aversion as a
cal and thalamic inputs elicit convergent excitation in a population learning and memory paradigm. Behav Brain Res 125: 205213,
of GABAergic interneurons of the lateral amygdala. J Neurosci 20: 2001.
8909 8915, 2000. 293. WHITING PJ, MCKERNAN RM, AND WAFFORD KA. Structure and phar-
271. TREIT D AND MENARD J. Dissociations among the anxiolytic effects macology of vertebrate GABAA receptor subtypes. Int Rev Neuro-
of septal, hippocampal, and amygdaloid lesions. Behav Neurosci biol 38: 95138, 1995.
111: 653 658, 1997. 294. WILENSKY AE, SCHAFE GE, AND LEDOUX JE. Functional inactivation
272. TREIT D, PESOLD C, AND ROTZINGER S. Dissociating the anti-fear of the amygdala before but not after auditory fear conditioning
effects of septal and amygdaloid lesions using two pharmacologi- prevents memory formation. J Neurosci 19: RC48, 1999.
cally validated models of rat anxiety. Behav Neurosci 107: 770 785, 295. WOODSON W, FARB CR, AND LEDOUX JE. Afferents from the auditory
1993. thalamus synapse on inhibitory interneurons in the lateral nucleus
273. TSVETKOV E, CARLEZON WA, BENES FM, KANDEL ER, AND BOLSHAKOV of the amygdala. Synapse 38: 124 137, 2000.

834 SAH ET AL.
296. WYLLIE DJ, MANABE T, AND NICOLL RA. A rise in postsynaptic Ca2 299. ZHANG L, WEINER JL, VALIANTE TA, VELUMIAN AA, WATSON PL, JAHROMI
potentiates miniature excitatory postsynaptic currents and AMPA SS, SCHERTZER S, PENNEFATHER P, AND CARLEN PL. Whole cell record-
responses in hippocampal neurons. Neuron 12: 127138, 1994. ing of the Ca2-dependent slow afterhyperpolarization in hip-
297. YAJEYA J, DE LA FUENTE JUAN A, MERCHAN MA, RIOLOBOS AS, pocampal neurons: effects of internally applied anions. Pflugers
HEREDIA M, AND CRIADO JM. Cholinergic responses of morpho- Arch 426: 247253, 1994.
logically and electrophysiologically characterized neurons of the 300. ZOLA-MORGAN S, SQUIRE LR, ALVAREZ-ROYO P, AND CLOWER RP. Inde-
basolateral complex in rat amygdala slices. Neuroscience 78: pendence of memory functions and emotional behavior: separate
731743, 1997. contributions of the hippocampal formation and the amygdala.
298. YOUNG BJ AND LEATON RN. Amygdala central nucleus lesions atten- Hippocampus 1: 207220, 1991.
uate acoustic startle stimulus-evoked heart rate changes in rats. 301. ZUCKER RS. Short term synaptic plasticity. Annu Rev Neurosci 12:
Behav Neurosci 110: 228 237, 1996. 1331, 1989.

803 Full

Uploaded by

Copyright:

Available Formats

803 Full

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

803 Full

Uploaded by

Copyright:

Available Formats

Physiol Rev

83: 803 834, 2003; 10.1152/physrev.00002.2003.

The Amygdaloid Complex: Anatomy and Physiology

I. INTRODUCTION of medial temporal lobe lesions in monkeys. These ani-

Physiol Rev VOL 83 JULY 2003 www.prv.org

FIG. 1. Nuclei of the rat amygdaloid

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

FIG. 3. Summary of the main outputs from the

Physiol Rev VOL 83 JULY 2003 www.prv.org

FIG. 4. Intra-amygdaloid connections as de-

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

FIG. 6. Pyramidal-like neurons and

Physiol Rev VOL 83 JULY 2003 www.prv.org

FIG. 7. Electrophysiological and synaptic

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

FIG. 8. Pyramidal cell and interneu-

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

neurons, the CeL also contained neurons that fired spikes

Consistent with tract tracing studies, experiments in

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

FIG. 10. Functional model of the basic pathways

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

Physiol Rev VOL 83 JULY 2003 www.prv.org

You might also like