Search

We have previously described the origins of neocortical inputs to the lateral nucleus of the macaque monkey amygdala based on retrograde tracing studies. Here we report results from studies that have attempted to confirm the projections from several candidate afferent regions using (3)H-amino acid autoradiography as an anterograde tracer. We have charted, based on the results of 33 separate injections, the topographic distribution of cortical projections throughout the amygdala. Areas TE and TEO of the inferotemporal cortex, portions of the superior temporal gyrus, and the granular region of the insula project primarily to the lateral nucleus, with little or no innervation of other amygdaloid nuclei. In contrast, orbitofrontal, medial prefrontal, and anterior cingulate regions project primarily to the basal and accessory basal nuclei and provide little innervation to the lateral nucleus. The orbitofrontal and medial prefrontal cortices, but not the anterior cingulate cortex, project to medially situated amygdaloid areas such as the cortical and medial nuclei and to the periamygdaloid cortex. The agranular and dysgranular insula, the parainsula, and rostral portions of the superior temporal gyrus project both to the lateral, basal, and accessory basal nuclei and to the medially situated nuclei. Projections to the central nucleus are particularly prominent from these regions. These data are discussed in relation to the hierarchical processing of sensory information that occurs within the amygdaloid complex.

We have used the anterograde tracer, Phaseolus vulgaris-leucoagglutinin (PHA-L) to study the intrinsic projections of the lateral nucleus of the Macaca fascicularis monkey amygdaloid complex. A reanalysis of the monkey lateral nucleus indicated that there are at least four distinct cytoarchitectonic divisions: dorsal, dorsal intermediate, ventral intermediate, and ventral. The major projections within the lateral nucleus originate in the dorsal, dorsal intermediate, and ventral intermediate divisions and terminate in the ventral division. The ventral division also projects to itself but does not project significantly to the other divisions of the lateral nucleus. Thus, the ventral division appears to be a site of convergence for information entering all other portions of the lateral nucleus. There are substantial regional and topographic differences in the projections from each of the lateral nucleus divisions to other amygdaloid nuclei. The dorsal division projects to all divisions of the basal and accessory basal nuclei, to the periamygdaloid cortex, the nucleus of the lateral olfactory tract, the dorsal division of the amygdalohippocampal area, and the lateral capsular nuclei. The dorsal intermediate division projects to the intermediate and parvicellular divisions of the basal nucleus, to the parvicellular division of the accessory basal nucleus, and to the periamygdaloid cortex. The ventral intermediate division projects to the magnocellular division of the accessory basal nucleus and to the parvicellular division of the basal nucleus. The major projections from the ventral division are directed to the parvicellular division of the basal nucleus, the parvicellular division of the accessory basal nucleus, the medial nucleus, and the periamygdaloid cortex. Projections from all portions of the lateral nucleus to the central nucleus are generally very light. It appears, therefore, that each division of the lateral nucleus originates topographically organized projections to the other amygdaloid areas that terminate in distinct portions of the target regions. The topographic organization of intrinsic amygdaloid projections raises the possibility that serial and parallel sensory processing may take place within the amygdaloid complex. J. Comp. Neurol. 398:431–458, 1998. © 1998 Wiley-Liss, Inc.

Previous studies have shown that the orbital and medial prefrontal cortex (OMPFC) is extensively connected with medial temporal and cingulate limbic structures. In this study, the organization of these projections was defined in relation to architectonic areas within the OMPFC. All of the limbic structures were substantially connected with the following posterior and medial orbital areas: the posteromedial, medial, intermediate, and lateral agranular insular areas (Iapm, Iam, Iai, and Ial, respectively) and areas 11m, 13a, 13b, 14c and 14r. In contrast, lateral orbital areas 12o, 12m, and 12l and medial wall areas 24a,b and 32 were primarily connected with the amygdala, the temporal pole, and the cingulate cortex. Data were not obtained on the posteroventral medial wall. Three distinct projections were recognized from the basal amygdaloid nucleus: 1) The dorsal part projected to area 12l; 2) the ventromedial part projected to most areas in the posterior and medial orbital cortex except for area Iai, 12o, 13a, and 14c; and 3) the ventrolateral part projected to orbital areas 12o, Iai, 13a, 14c, and to the medial wall areas. The accessory basal and lateral amygdaloid nuclei projected most strongly to areas in the posterior and medial orbital cortex. The medial, anterior cortical, and central amygdaloid nuclei and the periamygdaloid cortex were connected with the posterior orbital areas. The projection from the hippocampus originated from the rostral subiculum and terminated in the medial orbital areas. The same region was reciprocally connected with the anteromedial nucleus of the thalamus, which received input from the rostral subiculum. The parahippocampal cortical areas (including the temporal polar, entorhinal, perirhinal, and posterior parahippocampal cortices) were primarily connected with posterior and medial orbital areas, with some projections to the dorsal part of the medial wall. The rostral cingulate cortex sent fibers to the medial wall, to the medial orbital areas, and to lateral areas 12o, 12r, and Iai. The posterior cingulate gyrus, including the caudomedial lobule, was especially strongly connected with area 11m.

Previous anatomical studies of corticotropin-releasing factor (CRF)-like immunoreactivity in rat brain have reported prominent clustering of neuronal elements containing this peptide within the amygdala. The highest concentrations of both CRF-positive cells and fibers were evident in the central nucleus, an observation consistent with the putative role of this peptide in autonomic and endocrine regulation. In addition, lower densities of CRF-positive somata and processes have boon noted in other amygdaloid nuclei. However, the distribution of CRF-like immunoreactivity in the amygdala has not been described for any primate species. Such a description would be of interest since substantial differences in the distribution of CRF in rodent and primate have been reported for other brain regions. The present study uses immunohistochemical methods, with a polyclonal antiserum directed against the human form of CRF, to determine the distribution of this peptide in non-colchicine-treated monkeys (Saimiri sciureus). Within the amygdaloid complex, the most numerous and concentrated collections of CRF-positive neurons were seen in the basal and lateral nuclei. The highest densities of CRF-positive fibers and terminals were seen in the lateral and central amygdaloid nuclei. Moderately dense plexuses of CRF-positive fibers also were seen in layer Ia of the periamygdaloid cortex, nucleus of the lateral olfactory tract, anterior and posterior cortical nuclei, and the medial nucleus. Thus, the distribution of CRF-like immunoreactivity differs substantially in monkey and rat amygdala. Since CRF-positive perikarya in monkey are most prominent in nuclei with pronounced interconnections with neocortex, these differences may be an integral component of the increased cortical development that characterizes the primate brain. © 1992 Wiley-Liss, Inc.

The corticoamygdaloid and intraamygdaloid projections of the rat were studied by the use of retrograde transport of horseradish peroxidase (HRP). Observations based on anterograde transport of the enzyme were exploited to determine the course of the intrinsic connections. The HRP was injected stereotactically by means of iontophoresis. Most of the amygdaloid nuclei were selectively injected, and all but a few were reached by more than one approach. The vast majority of corticoamygdaloid fibers was found to originate in cortical areas defined as allocortical (Stephan, '75). From the medial frontal cortex the central amygdaloid nucleus (AC) receives a hitherto undescribed projection originating in the tenia tecta; and both the AC and the lateral amygdaloid nucleus (AL) receive fibers from the prelimbic and infralimbic areas. The anterior cingulate area entertains a weak connection with the basolateral amygdaloid nucleus (BL). As to the insular cortex, the posterior agranular insular area projects to all amygdaloid subdivisions; the BL, AC, and the anterior cortical nucleus (COa) receive, in addition, fibers from the ventral agranular area. The prepyriform cortex connects with the entire amygdala except the medial nucleus (Am). The amygdala receives afferents from a transitional area between the amygdala and the entorhinal area. The entorhinal area proper is related to the amygdala via projections from the ventral part of the lateral entorhinal area to the AL and from the dorsal part of the lateral entorhinal area to the BL. The former nucleus also receives fibers from the perirhinal region. Additional amygdalopetal connections from the hippocampal region include a previously undescribed projection from the temporal two-thirds of CA1 to the AL and BL and to the posterior cortical nucleus (COp) with the adjacent periamygdaloid cortex (PAC). The subiculum projections to the AL, and more modestly to other amygdaloid nuclei. There is an extensive network of intraamygdaloid connections, the Am and AC being the only nuclei not giving rise to the intrinsic fibers.

The association and commissural fiber systems arising in the olfactory cortical areas caudal to the olfactory peduncle (the piriform cortex, nucleus of the lateral olfactory tract, anterior cortical nucleus of the amygdala, periamygdaloid cortex and entorhinal cortex) have been studied utilizing horseradish peroxidase as both an anterograde and a retrograde axonal tracer. In the piriform cortex two sublaminae within layer II (IIa and IIb) and layer III have been found to give rise to distinctly different projections. Retrograde cell labeling experiments indicate that the association fiber projection from layer IIb is predominantly caudally directed, while the projection from layer III is predominantly rostrally directed. Cells in layer IIa project heavily to areas both caudal and rostral to the piriform cortex. The commissural fibers from the piriform cortex are largely restricted in their origin to layer IIb of the anterior part of the piriform cortex and in their termination on the contralateral side to the posterior part of the piriform cortex and adjacent olfactory cortical areas. A projection to the olfactory bulb has also been found to arise from cells in layers IIb and III of the ipsilateral piriform cortex, but not in layer IIa. In addition to those from the piriform cortex, association projections have also been found from other olfactory cortical areas. The nucleus of the lateral olfactory tract has a heavy bilateral projection to the medial part of the anterior piriform cortex and the lateral part of the olfactory tubercle (as well as a lighter projection to the olfactory bulb); both the anterior cortical nucleus of the amygdala and the periamygdaloid cortex project ipsilaterally to several olfactory cortical areas. The entorhinal cortex has been found to project to the medial parts of the olfactory tubercle and the olfactory peduncle. The olfactory tubercle is the only olfactory cortical area from which no association fiber systems (instrinsic or extrinsic) have been found to originate. A broad topographic organization exists in the distribution of the fibers from several of the olfactory areas. This is most obvious in the anterior part of the olfactory cortex, in which fibers from the more rostral areas (the anterior olfactory nucleus and the anterior piriform cortex) terminate in regions near the lateral olfactory tract, while those from more caudal areas (the posterior piriform cortex and the entorhinal cortex) terminate in areas further removed, both laterally and medially, from the tract. Projections to olfactory areas from the hypothalamus, thalamus, diagonal band, and biogenic amine cell groups have been briefly described.

Projections are described from the basolateral, lateral and anterior cortical nuclei of the amygdaloid complex, and from the prepiriform cortex, to several discrete areas of the cerebral cortex in the rat and cat and to the mediodorsal thalamic nucleus in the rat. These projections are very well-defined in their origin, and in their area of laminar pattern of termination. The basolateral amygdaloid nucleus can be divided into anterior and posterior divisions, based on cytoarchitectonic and connectional distinctions. In both the rat and cat the posterior division projects to the prelimbic area (area 32) and the infralimbic area (area 25) on the medical surface of the hemisphere. The anterior division projects more lightly to these areas, but also sends fibers to the dorsal and posterior agranular insular areas and the perirhinal area on the lateral surface. Furthermore, in the cat the perirhinal area is divided into two areas (areas 35 and 36) and the anterior division projects to both of these and also to a ventral part of the granular insular area; this last area is adjacent to, but separate from the auditory insular area and the second cortical taste area. In most of these areas, the fibers from the basolateral nucleus terminate predominantly in two bands: one in the deep part of layer I and layer II, and a heavier band in layer V (in the rat) or layers V and VI (in the cat). The lateral amygdaloid nucleus projects heavily to the perirhinal area, and also to the posterior agranular insular area. These fibers terminate predominantly in the middle layers of the cortex, although the cellular lamination in these two areas is relatively indistinct. The anterior cortical amygdaloid nucleus and the prepiriform cortex both project to the infralimbic area and the ventral agranular insular area, and the anterior cortical nucleus also projects to the posterior agranular area and the perirhinal area. In all of these areas, the fibers from these olfactory-related structures terminate in the middle of layer I. In the rat, the two divisions of the basolateral nucleus also project to the medial segment of the mediodorsal thalamic nucleus, with the anterior division projecting mainly to the posterior part of this segment and the posterior division to the anterior part. The endopiriform nucleus, deep to the prepiriform cortex, projects to the central segment of the mediodorsal nucleus; this may constitute the major olfactory input into the mediodorsal nucleus, since little or no projection could be demonstrated from the prepiriform cortex itself. Projections to the mediodorsal nucleus have not been found in the cat.

The connections of the amygdala with the insular and temporal cortices were examined by injecting wheat germ agglutinin conjugated to HRP (WGA-HRP) into the rat cortex. Following injections into the posterior agranular insular area (AIp) or perirhinal cortex (PR), bands of labeled neurons extending across nuclear boundaries were observed in the amygdala. These neuronal bands involved cells in the lateral, basolateral, and basomedial nuclei as well as the periamygdaloid cortex. Other nuclei of the corticomedial amygdala and the ventral endopiriform nucleus also exhibited retrogradely labeled cells. Anterograde label was observed in nuclei containing labeled neurons and in the central nucleus. Injections into gustatory, somatosensory, and auditory neocortical areas located dorsal to AIp and PR labeled small numbers of cells in the lateral and basolateral nuclei. Injections into AIp, PR, and, to a lesser extent, dorsally adjacent neocortical areas produced both retrograde and anterograde labeling in the contralateral amygdala. The main nuclei with contralateral insular and temporal projections are the basomedial nucleus, ventral endopiriform nucleus, and nucleus of the lateral olfactory tract. The contralateral central nucleus and to a lesser extent the lateral nucleus exhibited anterograde labeling. The pattern of retrograde labeling seen with injections at different rostrocaudal levels of the AIp-PR continuum indicates that amygdalocortical projections to these areas exhibit an overlapping topographical organization. Comparison of the results of this study with findings on amygdaloprefrontal cortical efferents suggests that amygdaloid projections to the entire fronto-insulo-temporal mesocortical field are topographically organized.

Previous studies have shown that the amygdala projects to both the mediodorsal thalamic nucleus (MD) and its cortical projection area, the prefrontal cortex (PFC). In this investigation rats received injections of different fluorescent retrograde tracers (true blue and diamidino yellow) into MD and either the lateral, polar, or medial PFC in order to examine the relationship of amygdaloid neurons with cortical and/or thalamic projections. PFC injections labeled neurons in the basolateral (BL), basomedial (BM), ventral endopiriform (Env), and rostral lateral nuclei as well as the periamygdaloid cortex (PAC) and the medial part of the amygdalohippocampal area (AHA). In BL, which contained the great majority of neurons projecting to PFC, most labeled cells were concentrated in particular parts of the nucleus and were topographically organized. The overwhelming majority of labeled neurons in BL were large pyramidal or piriform cells that correspond to class I neurons described in Golgi studies. Occasional small neurons with thin dendrites were also observed; these cells may be class II neurons. MD injections labeled numerous cells in the anterior division of the cortical nucleus, medial nucleus, and caudomedial part of the central nucleus. Moderate numbers of labeled cells were found in caudal portions of BM and PAC, whereas scattered cells were observed throughout the rest of the amygdala with the exception of the lateral nucleus. In BL and AHA many MD-projecting neurons were observed along nuclear boundaries and in the adjacent white matter. Neurons in BL, BM, and AHA usually had large elongated or irregular somata and two to four primary dendrites that branched sparingly. Other cells had smaller ovoid somata. The morphology and distribution of MD-projection cells in the basolateral amygdala indicate that they are primarily large class II neurons. Double-labeled amygdaloid neurons, labeled by both cortical and thalamic injections, were observed only in a small number of animals. Control experiments suggest that most of the double-labeled cells in these cases were artifacts caused by spread of the thalamic injectate into the third ventricle with subsequent uptake by fibers in the anterior commissure. Thus the findings of this study suggest that different neuronal populations in the amygdala project to the two poles of the MD-PFC system. In the basolateral amygdala class I neurons are the predominant cell type involved in PFC projections, whereas a subpopulation of class II neurons, hitherto thought to be primarily local-circuit neurons, project to MD.

Axonal projections are described from the lateral and basolateral nuclei of the amygdaloid complex, and from the overlying periamygdaloid and prepiriform cortices and the endopiriform nucleus, to the lateral entohinal area, the ventral part of the subiculum, and the parasubiculum in the cat and rat. All of these projections have well-defined laminar patterns of termination, which are complementary to those of other projections to the same structure. Based on these results, and on cytoarchitectonic distinctions, the lateral entohinal area has been divided into dorsal, ventral, and ventromedial subdivisions. The olfactory bulb and prepiriform cortex project to layers IA and IB, respectively, of all three subdivisions, but the lateral amygdaloid nucleus has a restricted projection to layer III of the ventral subdivision only. The periamygdaloid cortex projects to layer II of the ventromedial and adjoining parts of the ventral subdivisions. The ventral part of the subiculum receives fibers from the posterior division of the basolateral nucleus, which terminate in the cellular layer and the deep half to one-third of the plexiform layer. The periamygdaloid cortex and the endopiriform nucleus also project to the same part of the subiculum, but these fibers terminate in the outer part of the plexiform layer. None of these projections extend into the dorsal part of the subiculum. The posterior division of the basolateral nucleus also projects to the posterodorsal part of the parasubiculum ("parasubiculum a" of Blackstad, '56). These fibers end in the deeper part of the plexiform layer and the superficial part of the cellular layer.

The cytoarchitectonic structure and divisions of the amygdaloid complex are described in the rat and cat, with special reference to the axonal connections of each of the amygdaloid nuclei, and to interspecies variations and similarities. Several intra-amygdaloid connections are also described, based on autoradiographic experiments with small injections of 3H-amino acids into the individual nuclei. Although it has probably not been possible to determine all of the intra-amygdaloid projections from these experiments, the connections which have been shown are very specific. The lateral nucleus projects to the basomedial nucleus, the lateral part of the central nucleus, and the periamygdaloid cortex, while the basolateral nucleus projects only to itself, the medial part of the central nucleus, and the nucleus of the lateral olfactory tract. The basomedial nucleus projects to the cellular layer of the medial nucleus and the amygdalo-hippocampal area, while the molecular layer of these structures and of the posterior cortical nucleus, receives projections from the periamygdaloid cortex or from the endopiriform cortex. There are also interconnections between the medial and posterior cortical nuclei, and commissural connections between the posterior cortical nuclei of the two sides.

Cholecystokinin (CCK) binding sites were localized in the hippocampus, amygdala, and medial temporal cortices of macaque monkeys by using techniques of in vitro receptor autoradiography. Binding sites were labeled with 3H-CCK-8 and 125I-CCK-33, and nonspecific binding was assessed in the presence of 1 μM CCK-8. Comparison of autoradiograms with Nisslstained sections allowed precise correlation of autoradiographic grain distribution with cytoarchitecture. CCK binding in the amygdala varied among nuclear subdivisions. It was dense in the lateral, basomedial, endopiriform, and cortical nuclei, in the parvicellular portion of the accessory basal nucleus, the periamygdaloid cortex, the cortical transition area, and in the amygdalohippocampal area. Labeling was sparse in the central, medial, and basolateral nuclei as well as in the magnocellular accessory basal nucleus. In the hippocampal formation, a single dense band of CCK binding was observed over the granule cell layer and adjacent few millimeters of the molecular layer of the dentate gyrus, while in the polymorph and remaining portions of this layer binding was of very low density. Prominent label over the pyramidal layer in the presubiculum clearly distinguished this region from the adjacent subiculum in which binding just exceeded background levels. Moderate to light label was observed in the hilus and stratum pyramidale of CA3, CA2, and CA1, while other hippocampal layers showed minimal specific binding. Variation in CCK binding in the medial temporal cortex showed close correspondence to cytoarchitectonic subdivisions. In entorhinal cortex, for example, binding was concentrated in layers III-VI while label in area 35 was prominent in all laminae except layer IV. Area TH of von Bonin and Bailey ('47) was distinguished from other regions by evenly distributed binding across all layers, while in area TF a bilaminar pattern of label in layers II and IV was observed. The highly specific patterns of CCK binding in amygdala and transitional cortices of the medial temporal lobe can be related to terminal fields of neo- and allocortical afferents to these regions, while label in the hippocampal formation coincides with the terminals of intrinsic neurons which ramify among the somata of cells that are targets of neocortical afferents. Thus, in all structures of the medial temporal lobe the disposition of peptidergic binding sites suggests that CCKergic systems may be important in the modulation of cortical afferents.

The organization of the cholinergic innervation of the macaque monkey amygdaloid complex was investigated by means of immunohistochemical techniques and either a polyclonal antiserum or a monoclonal antibody directed against the specific synthetic enzyme choline acetyltransferase (ChAT). Adjacent series of sections were processed histochemically for the demonstration of the degradative enzyme acetylcholinesterase (AChE) or for cell bodies with thionin. The density of ChAT immunoreactivity differed substantially among the various nuclei and cortical regions of the amygdala. In general, the distribution of ChAT immunoreactivity paralleled the pattern of AChE staining. One notable exception was the presence of AChE containing cell bodies in addition to AChE positive fibers within nearly all of the nuclear and cortical regions. In contrast, ChAT immunoreactivity was associated only with fibers and terminals. The highest density of ChAT immunoreactive fibers and terminals was consistently observed in the magnocellular subdivision of the basal nucleus. Staining was substantially less dense in the more ventrally situated parvicellular subdivision. Medially, in the adjacent accessory basal nucleus, immunoreactive fibers and terminals were densest in the magnocellular and superficial subdivisions and least prominent in the parvicellular subdivision. Of the deep nuclei, the lateral nucleus generally contained the least ChAT immunoreactive terminals and processes. Only its more densely cellular ventrolateral portion contained appreciable fiber and terminal staining. One of the more distinctive patterns of ChAT immunoreactivity was seen in the nucleus of the lateral olfactory tract. Here, ChAT positive fibers formed pericellular basket plexuses around unstained cell bodies. This unique pattern of staining was used to delineate the boundaries of the nucleus and indicated that it is present for much of the rostrocaudal extent of the amygdala. Another region of conspicuous staining on the medial surface of the amygdala was the sulcal portion of the periamygdaloid cortex. This region, associated with the sulcus semiannularis and bordering the entorhinal cortex, consistently contained dense immunoreactivity. The central nucleus also presented a somewhat idiosyncratic pattern of ChAT staining. The lateral subdivision had a diffuse distribution of immunoreactivity in which focal patches of more densely stained terminals and occasional fine fibers were embedded. In contrast, the medial subdivision contained a larger number of thicker, stained fibers without diffuse background labeling. Of the remaining amygdaloid nuclei, the amygdalohippocampal area contained dense ChAT staining, whereas progressively lighter staining was seen in anterior amygdaloid, posterior cortical, anterior cortical, and medial nuclei.

The fields of origin of the amygdalofugal pathways in the albino rat were studied by using the Fink-Heimer technique to stain degenerating fibers and terminals in brains of animals sacrificed four and seven days after small unilateral lesions of the amygdale. A surprisingly restricted field of origin was found for the supracommissural stria terminalis — the component projecting to the ventromedial nucleus of the hypothalamus Only a lesion in the most posterior part of the cortical nucleus produced degeneration in this pathway. The postcommissural stria terminalis which apparently distributes to the bed nucleus of the stria terminalis and preoptic area originates in a more widespread area including parts of both the basolateral and corticomedial complexes. Although intense terminal degeneration could be seen within the amygdala following all lesions, long fibers projecting into the anterior amygdaloid area, lateral preoptic area and lateral hypothalamus were found only after lesions in which fibers of passage from periamygdaloid cortex could have been interrupted. Convincing evidence, therefore, of a long axon ventral amygdalofugal pathway of the rat was not found in this study.