NeuroscienceVol. 44, No. 1, pp. t-14, 1991 Printed in Great Britain

0306&22/91 33.GQ+o.o0 PcrgamonPmis plc 0 1991IBRO

ORGANIZATION OF AMYGDALOID PROJECTIONS TO THE PREFRONTAL CORTEX AND ASSOCIATED STRIATUM IN THE RAT A.J. MCDONALD Department

of Anatomy,

Cell Biology and Neuromiences, University of South Carolina School of Medicine, Cohnnbia, SC 29208, U.S.A.

Abstract-The organization of connections between the amygdala, prefrontal cortex and striatum was studied using anterograde and retrograde tract tracing techniques in the rat. The anterograde transport of Phaseok vrdgaris le~a~lutinin and wheat germ agglutinin conjugated to horseradish peroxidase was used to examine the striatal projections of the prefrontal cortex. These studies revealed that the p~~rnbic area of the medial prefrontal cortex projects mainly to the medial part of the striatum, whereas the dorsal agranular insular area of the lateral prefrontal cortex projects maiuly to the ventrolateml part of the striatum. The organization of amygdaloid projections to the prefrontal cortex and its associated portions of the striatum was investigated using the fluorescence retrograde tract tracing technique. Different color fluorescent dyes, True Blue and Diamidino Yellow, were injected into the prefrontal cortex and striatum. These studies demonstrated that medial portions of the basolateral nucleus, and adjacent portions of the lateral, basomedial and ~ygd~o-hip~pal nuclei, project to both the medial prefrontal cortex and its associated medial striatai region. The rostrai pole and lateral portions of the basolateral nucleus project to both the lateral prefrontal cortex and its associated lateral striatal region. Many neurons in the basolateral amygdaloid nucleus, and to a lesser extent other amygdaloid nuclei, were double-labeled in these experiments, indicating that these cells send collaterals to both the prefrontal cortex and striatum. These findings indicate that discrete areas of the amygdala, and in some cases individual amygdaloid neurons, can modulate information processing in the first two links of distinct cortico-striato-pallidal systems arising in the medial and lateral prefrontal cortex.

(limbic striatum) that receive projections from the The telencephalon is largely composed of cortical, striatal and pallidal structures.2*27Discrete portions of PFC.i6 Thus the basolateral amygdala can influence these three tiers of the forebrain are connected in the first two links of the prefront~o~ico-st~atopallidal system. Since the striatal projections3,z7 as series to form parallel co~co-s~ato-pallidal (CSP) systems that may subserve specific functions.‘~7*p*31 well as the functions* of the lateral and medial PFC differ, it is of interest to examine the topography of Although it is known that the basolateral amygdala has strong projections to the cortex and striatum, the the amygdaloid projections to the lateral and medial interrelationship of amygdaloid neurons projecting to PFC and their striatal targets. previous i~vesti~tions have demonstrated that each of these tiers has not been analysed. The prefrontal cortex (PFC) is one of the main both the amygdal~orti~ and amygdalost~a~l cortical targets of the basolateral amygdala.‘7*ut39 In projections in the rat are topographically organized.‘6’*~22~26 In the companion paper,% comparisons the rat these basolateral amygdaloid efferents terminate mainly in the dorsal agranular insular area of the amygdalocortical and amygdalostriatal topog (AId) of the lateral PFC and the prelimbic area (PL) raphy suggested that these projections were in of the medial PFC.” In addition to its projection to register; discrete portions of the amygdala appeared the PFC, the basolateral amygdala also has a strong to project to inte~onn~t~ portions of the cortex projection to the same portions of the striatum and striatum. In the present study the organization of amygdaloid neurons projecting to the medial and lateral PFC and associated striatal regions was Abbreviations: AId, dorsal agranular insular cortex; AIv, ventral agrannlar insular cortex; BL, basolateral directly investigated using a fluorescence retrograde amygdaloid nucleus; CSL neuron, single-labeled cortextract tracing technique. projecting neuron; CSP, ~rti~-st~to-paging DL neuron, doubl~la~l~ neuron; DY, Di~dino Yellow; NLOT, nucleus of the lateral olfactory tract; PFC, prefrontal cortex; PHA-L, Phaseolus vulgark leucoagglutinin; PL, prelimbic cortex; PLP, paraformaldehyde-lysine-periodate; SSL neuron, single-labeled striatum-projecting neuron; TB, True Blue; WGA-HRP, wheat germ a&utinin-conjugated horseradish peroxidase.

EXPERIMENTAL PROCEDURES Fluorescence retrograde tract tracing experiments

In this study, 47 Harlan Sprakme-Dawley rats (250-300 g) were anesthetized with sodium pentobarbital (5Omg/kg, i.p.) and received injections of different fluorescent retro-

2

A. J. MCDONALD

grade tracers into the medial and/or lateral prefrontal cortex (PFC) and the main striatal targets of the PFC. The fluorescent dyes True Blue (TB) and Diamidino Yellow hydrochloride (DY) were used in each animal. Cells retrogradely labeled by TB exhibited a blue fluorescent labeling of the perikarya and nucleolus. DY-labeled cells exhibited a bright yellow fluorescent labeling of the nucleus and a light labeling of the perikarya. These dyes have been used effectively in many different systems in the rat brain to identify individual neurons (double-labeled cells) projecting axonal collaterals to two different brain areas.‘S.L9 The striatal targets of the lateral and medial PFC were determined in experiments using anterograde transport of wheat germ agglutininconjugated horseradish peroxidase (WGA-HRP) and Phaseolus uulgaris leucoagglutinin (PHA-L) as tracers (see below). Injections of TB (5% in distilled water; 0.14.3~1) and DY (2% in 0.2 M phosphate buffer, pH 7.2; 0.3-0.5 ~1) were made stereotaxically using a 1.0~~1Hamilton microsyringe equipped with a 26gauge needle. Stereotaxic coordinates were obtained from the atlas of Paxinos and Watsou.33 Since the lateral and medial PFC and their terminal fields in the striatum extend for several millimeters in the rostrocaudal direction, some attempts to inject these regions involved two injections, one into the rostra1 half and the other into the caudal half of these regions. Four to seven days after the injections animals were anesthetized and perfused through the heart with h~rtonjc saline (1.5% NaCl) followed by 20% formalin in 0.1 M citrate buffer (pH 7.2). Brains were removed, postfixed for l-3 h, and sectioned at 4O/tm in the coronal or sag&al plane on a Vibratome. Alternate sections were mounted on gelatinized slides and stored in the dark at 4°C until examined with an OIympus BHA Auorescence microscope with a UC-1 excitation filter, DM400 dichroic mirror, and an L-420 barrier fitter. The remaining sections were mounted and stained with a Nissl stain (pyronin-Y”) to facilitate the localization of retrogradely labeled neurons observed in adjacent sections. A photographic technique was used to accurately chart the position of labeled neurons at six rostrocaudal levels of theamygdala (bregma -1.8, -2.3, -2.8, -3.3, -3.8, -4.3 of the atlas of Paxinos and Wa~on33). Photomon~~es assembled from black-and-white photbmicrographs -of fluorescent sections through the amygdala were made at each of the six levels. Labeled neurons and blood vessels were evident in these photomontages. Nuclear boundaries and blood vessels from adjacent Nissl-stained sections were traced on transparent plastic tilm with the aid of a drawing tube. These drawings were at the same ma~~tion as the photomontages. Each transparency was then placed on top of its corresponding photomontage using blood vessels as alignment landmarks. The actual section from which the photomontage was made was then examined at high magnification and the type of labeling exhibited by each neuron in the photomontage (double-labeled, TB singlelabeled, or DY single-labeled) was recorded on the transparency overlying the photo~aph~ cell. By this method the exact position and the type of labeling exhibited by each cell was accurately recorded. All animals in this study received injections of one fluorescent dye into the PFC and a different fluorescent dye into a portion of the striatal target of the injected cortical area. Injections into the medial PFC (centered in the prelimbic area) were paired with injections into either the rostromedial striatum (nucleus accumbens and rostra1 dorsomedial caudatoputamen; n = 12) or more caudal portions of the dorsomedial caudatoputamen (n = 14). Injections into the lateral PFC (centered in the dorsal agranular insular area) were paired with injections into the ventrolateral caudatoputamen (n = 12). In some rats (n = 9) injections of one dye into both the lateral and medial PFC were paired with injections of the other dye into both the

rostromedial striatum and the fundus striati. The pattern 01 retrograde labeling in the amygdala did not seem to depend on the particular dye injected, only on the exact placement of the injection site. In general, however, it appeared that DY produced larger numbers of labeled cells in any particular amygdaioid region than a similarly placed injection of TB. Thus, although the pattern of labeling was similar with each of the various combinations of dye placements, the numbers of cells labeled by each dye, including numbers of double-labeled cells, depended on the exact size and position of the injection sites as well as on the dye chosen for the striatal versus the cortical injections. The cases chosen to illustrate amygdaloid labeling (Figs 3- 7) were among those that had the largest numbers of double-labeled neurons. Because fibers of passage can take up and retrogradely transport fluorescent tracers,& it was important to determine if amygdalocorticai fibers might course through the striatum en route to the prefrontal cortex. Since the cells of origin of these fibers might be labeled by striatal injections (via uptake by fibers of passage) as well as by cortical injections (via axon terminal uptake), artifactual doublelabeling could result. To test this possibility l-mm-wide coronal knife cuts3’ were made just rostra1 to the striatal injections in some of the cases in this study, The finding that these knife cuts did not produce any noticeable decrease in double-laxly cells in the amygdala indicates that few. if any, ~ygd~o-p~front~ fibers pass through the striatum. Since adequate visualization of knife cuts required sectioning in the sagittal plane, which is not the customary plane for charting amygdaloid labeling, most cases with knife cuts were not charted. However, in cases where striatal injections were positioned near the rostra1 pole of the striatum, the tissue block containing the injection site and knife cut was sectioned in the sagittal plane while the block containing the amygdala was sectioned in the coronal plane. The results of one of these cases (R81) is illustrated (Figs 4 and 5).

Anterograde tract tracing experiments The projections of the lateral and medial PFC to the striatum were determined in experiments using anterograde transport of PHA-L or WGA-HRP. Four rats received injections of 2.5% PHA-L (in phosphate-buffered saline, pH 8.0) that were confined to the dorsal agranular insular area (AId) of the lateral prefrontal cortex (n = 2) or the prelimbic area (PL) of the medial prefrontal cortex (n = 2). Additional animals (n = 6) received PHA-L injections that involved areas located dorsal or ventral to Aid or PL. All injections were made iontophoreti~ally via glass micropipettes (2O~m outer tip diameter) using a Midgard high-voltage current source set at 5.0 nA (7 s on, 7 s off for 45 min). Following a 10 day survival, rats were perfused intracardially with 50 ml of ice-cold 0.1 M nhosuhate-buffered saline t&I 7.2) followed by 500 ml of &e-c&l paraformaldehyd~lysine$eriodate (PLP) fixative?s Brains were removed, postfixed in PLP overnight, and cut at 4Onm in the coronal plane with a Vibratome. Every third section was processed for PHA-L immunohistochemistry” using antibodies to PHA-L (diluted 1: 1000; Vector Laboratories) and Vectastain ABC kits (Vector Laboratories). Final visualization of PHA-L was accomplished using nickel-enhanced diaminobenzidine.” Alternate sections were stained with Cresyl Violet. Six rats received iontophoretic injections of 4% WGA-HRP into the lateral (n = 3) or medial (n = 3) PFC. One day later animals were perfused with 0.9% saline (50 ml) followed by 2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4 (750ml). Brains were removed and processed for horseradish peroxidase histochemistry using tetramethyl benzidine as a chromogen (see companion paper for technical details).% Sections were then counterstained with pyronin-Y.”

Amygdaloid projections to prefrontal cortex and striatum

3

Fig. 1. Drawings illustrating the results of experiments in which injections of WGA-HRP were made into the prelimbic area (PL) of the medial PFC (case H43, A-D) or the dorsal agranular insular area (AId) of the lateral PFC (case H50, A-D’). Note the topographic organization of the projections to the striatum [caudatoputamen (CP), nucleus accumbens @IA), and fundus striati (FS)]. The only striatal region which receives significant projections from both cortical areas is the rostra1 fimdus striati (B, B’). Additional abbreviations used in this and subsequent figures are as follows: AC, anterior commissure; AHA, amygdalobippocampal area; AI,, dorsal agranular insular area; BL,, anterior subdivision of the basolateral amygdaloid nucleus; BL,, posterior subdivision of the basolateral amygdaloid nucleus; BL,, ventral subdivision of the basolateral amygdaloid nucleus; BM, basomedial amygdaloid nucleus; BM,,

anterior subdivision of the basomedial amygdaloid nucleus; C, corpus cahosum; Ce, central amygdaloid nucleus; Co , posterior subdivision of the cortical amygdaloid nucleus; DB, nucleus of the diagonal band, F, fomix; & , frontal polar cortex; GP, globus pallidus; IC, internal capsule; Ld, dorsolateral subdivision of the lateral amygdaloid nucleus; L,, ventromedial subdivision of the lateral amygdaloid nucleus; M, medial amygdaloid nucleus; OT, olfactory tubercle; PAC, periamygdaloid cortex; S, septum. RESULTS

Anterograde tract tracing studies of prefrontostriatal projections PHA-L and WGA-HRP studies of the projections of the PFC to the striatum produced very similar results. Differences in anterograde labeling of the striatum were related to the exact placement of the injection site and were consistent between the two techniques. Since the WGA-HRP injections were

larger and involved a greater extent of the cortical area injected, cases with these injections will be. used to document the projections of the PFC to the striatum. Case H43 will be used to illustrate the striatal projections of the medial PFC (Fig. 1). This animal received an injection of WGA-HRP that involved the rostra1 PL and also impinged slightly on the ventrally adjacent medial orbital area

(Fig. 1A). At rostra1 levels of the striatum, anterograde label was observed in the dorsomedial caudatoputamen, rostra1 fundus striati, nucleus accumbens, and deep part of the olfactory tubercle (Fig. 1B). At a level just rostra1 to the crossing of the anterior commissure, dense label was seen in the dorsomedial caudatoputamen and sparse label was observed in the fundus striati (Fig. 1C). At caudal striatal levels, label was observed in the dorsomedial caudatoputamen adjacent to the stria terminalis (Fig. 1D). The two PHA-L cases with injections into PL each received injections into both the rostra1 and caudal portion of this cortical area; they exhibited anterograde labeling identical to that seen in H43. Injections of PHA-L confined to the infralimbic cortical area produced strong label in the medial olfactory tubercle and lighter label in the adjacent nucleus accumbens; almost no label was seen in the caudatoputamen.

A. J. MCDONALU

R4

;.

.

*00

a.

0

A-TB 0:DY 0 : TB-DY

Fig. 2. Drawing illustrating the results of case R4. Injections of TB were made into the PL of the medial PFC and the AI, of the lateral PFC (A). Injections of DY were made into the nucleus accumbens (NA) and fundus striati (FS) (B, C). The distribution of amygdeloid neurons that were double-labeled (closed circles), single-labeled by TB (triangles), or single-labeled by DY (open circles) are ilIustrated at the bregtna -2.8 leve13*of the rat brain (see D for nuclear subdivisions found at this level).

Fig. 3. Photomicrographs of case R4. (A) Photomicrograph of a pyronin-Y-stained section at the level of Fig. 2A. Arrow indicates the injection of TB into the lateral PFC. Box contains the injection of TB into the medial PFC. Compare with Fig 2A. Scale bar = 1 mm. (B) Fluorescence photomicrograph of the boxed field in A, photographed from an adjacent unstained section. (C) Photomicrogaph of a pyronin-Y-stained section at the level of Fig. 2B. Box contains the injection of DY into the nucleus accumbens. Compare with Fig. 2B. Same ~~fi~tion as in A. (D) FIuorescence photo~~o~aph of the boxed field in C, photo~ap~ from an adjacent unstained section. (E) Fluorescence photo~~ph of retrogradely labeled neurons in the anterior subdivision of the basolateral nucleus in R4. Open arrows, neurons single-labeled by the TB injections into the PFC; closed arrows, neurons single-labeled by the DY injections into the striatum. The remaining neurons were double-labeled (TB-DY). Scale bar = 25 Nrn. (F) Fluorescence photomicrograph of retrogradely labeled neurons in the lateral nucleus in M. Open arrows, double-labeled neurons (I’B-DY); dosed arrows, neurons single-lab&d by the DY injections into the striatum. The remaining neurons were single-labeled by the TB injections into the PFC. Same ma~~cation as in E.

~ygd~oid

projections to prefrontal cortex and striatum

Fig. 3.

5

A. J. MCDONALD

6

Case H50 will be used to illustrate the striatal projections of the lateral PFC (Fig. 1). This animal received an injection of WGA-HRP that was centered in the AId and also impinged on the medially adjacent ventral agranular insular area (Alv) and the dorsally adjacent sensorimotor cortex (Fig. IA’). At rostra1 striatal levels, anterograde label was observed primarily in the ventrolateral rostra1 fundus striati, lateral caudatoputamen, nucleus accumbens and adjacent olfactory tubercle (Fig. IB’, C’). At caudal striatal levels, label was seen in the ventral two-thirds of the caudatoputamen but avoided the dorsomedial third, which receives projections from the medial PFC (Fig. ID). Label in the ventral two-thirds of the caudatoputamen gradually attenuated at levels approaching the caudal pole of the caudatoputamen. PHA-L cases with more substantial involvement of AIv produced labeling in the ventrolateral strlatum that extended further medially and exhibited considerable overlap with that exhibited by medial PFC projections. A PHA-L injection that was centered in the dorsal part of AId and slightly involved the dorsally adjacent sensorimotor cortex, but not AIv, produced anterograde label in the ventrolateral caudatoputamen, but almost no labeling of the nucleus accumbens. These anterograde studies indicate that the PL of the medial PFC projects primarily to the dorsomedial caudatoputamen and adjacent nucleus accumbens, whereas the AId of the lateral PFC projects primarily to the ventrolateral striatum. Thus, the striatal projections of these two PFC regions are mostly complementary to each other, although some overlap does occur at rostra1 striatal levels. The striatal projection of the AIv of the lateral PFC, which receives projections primarily from the corticomedial rather than the basolateral amygdala,16 appears to overlap that of PL. The retrograde labeling in the cortex produced by the striatal dye injections described below is consistent with the findings of these anterograde tracing experiments. Thus, injections into the medial portions of the rostra1 striaturn retrogradely labeled cells in PL (and AIv), but not AId. Injections into the ventrolateral caudatoputamen retrogradely labeled cells in AId, but not in PL. Fluorescence retrograde

tract tracing studies

Cases in which injections into the lateral and medial prefrontal cortex were paired with striatal injections.

In nine animals, one fluorescent dye was injected into both the medial and lateral PFC and the other dye was injected into two locations in the rostra1 striatum that receive strong projections from the basolateral amygdala: the nucleus accumbens and fundus striati. The retrograde labeling in all successful cases was very similar; case R4 will be described in detail. True Blue (TB) was injected into both the medial and lateral PFC in R4 (Figs 2 and 3). The medial PFC

injection was confined to the PL, whereas the lateral PFC injection involved the AId and AIv. Diamidino Yellow (DY) was injected into both the nucleus accumbens and fundus striati (Figs 2 and 3). With both striatal injections the dorsally adjacent portions of the caudatoputamen were also involved. The accumbens injection was confined to a part of the striatum that receives projections from the medial PFC, whereas the fundus injection was in a striatal region that receives projections from both the medial and lateral PFC (see anterograde experiments above). Retrograde labeling was found throughout the rostrocaudal extent of the basolateral portion of the amygdala but was most dense in its rostra1 half and included many double-labeled neurons. The pattern of retrogradely labeled cells at the bregma - 2.8 level of the amygdala is illustrated in Fig. 2. As expected (see Discussion section of companion paper2’), involvement of both the medial and lateral PFC and portions of their striatal targets in the injection sites resulted in retrograde labeling (DY-labeled neurons, TB-labeled neurons, and variable numbers of doublelabeled neurons) in both the medial and lateral portions of the basolateral nucleus [and in other nuclei of the amygdala: lateral and basomedial nuclei, periamygdaloid cortex, and nucleus of the lateral olfactory tract (NLOT)]. With the exception of the periamygdaloid cortex, all amygdaloid nuclei contained more neurons labeled by striatal injections (DY cells in R4) than by the PFC injections (TB cells in R4). This was also seen in other cases in which the dyes were reversed. No retrograde labeling was observed in the central or medial amygdaloid nuclei in R4 or in any other of the animals examined in this study. The extensive double-labeling in the lateral and basal nuclei (Figs 2 and 3E, F; Table 1) indicates that many neurons in these regions have axons that branch to innervate both the PFC and the striatum. In all of these nuclei a greater percentage of PFC-projetting neurons than striatum-projecting neurons were double-labeled. In all cases the greatest percentage of double-labeled neurons was observed in the anterior subdivision of the basolateral nucleus (BLa; Table 1). The largest percentage of doublelabeled neurons in BLa was seen in case R9 (53.9%). In this animal, the dyes used to inject the PFC and striatum were opposite to that in R4 (i.e. DY into the PFC and TB into the striatum). In addition, the injection into the fundus striati in R9 was positioned more laterally than in R4 and therefore more fully involved the portion of the striatum innervated by the lateral PFC. In contrast to the widespread retrograde labeling seen in the amygdala in cases where medial and lateral PFC injections were paired with multiple striatal injections, a topographical organizational pattern was observed when medial PFC injections were paired with medial striatal injections and when lateral PFC injections were paired with lateral striatal injections (see below).

Amygdaloid projections to prefrontal cortex and striatum Table 1. Percentages of neurons labeled with True Blue, Diamidino Yellow, or both dyes (True Blue-Diamidino Yellow) in different amygdaloid nuclei in case R4 (actual neuronal counts am in parentheses) Nucleus Ld Lv BM BLv BLa

TB neurons 16.5% 26.7% 12.7% 12.0% 12.0%

(29) (66) (21) (16) (155)

DY neurons 70.4% 44.5% 57.2% 59.4% 45.7%

(124) (110) (95) (79) (590)

TRDY neurons 13.1% 28.8% 30.1% 28.6% 42.3%

(23) (71) (SO) (38) (547)

TB was injected into the prefrontal cortex. DY was injected into the striatum. Counts were made from several sections, Ld, lateral nucleus, dorsolateral subdivision; Lv, lateral nucleus, ventromedial subdivision; BM, basomedial nucleus; BLv, basolateral nucleus, ventral BLa, basolateral nucleus, anterior subdivision;

s&division.

Cases in which medial prefrontal cortex injections were paired with nucleus accumbens injections. Injections into the medial PFC were paired with injections centered in the nucleus accumbens in 12 rats. In all cases, l-mm-wide knife cuts were made just rostra1 to the accumbens injection site to prevent possible artifactual double-labeling (see Experimental Procedures section). The retrograde labeling was very

7

similar in all successful cases and R81 will be described in detail. The TB injection into the medial PFC in R81 involved mainly the PL and the ventrally adjacent medial orbital cortex (Fig. 4). The DY injection into the rostra1 striatum involved the medial and lateral portions of the nucleus accumbens as well as the dorsally adjacent dorsomedial caudatoputamen (Fig. 4). This injection was located just medial to the anterior limb of the anterior fussy in approximately the same location as the accumbens injection in R4 (see Fig. 2B). This striatal region receives a strong projection from the medial PFC, but little or no projection from the AId of the lateral PFC (see anterograde studies above). The distribution and relative density of amygdaloid neurons labeled by the TB injection into the medial PFC was almost identical to that labeled by the DY injection into the nucleus accumbens (Fig. 5). Neurons that were single-labeled by the cortical injection (CSL neurons), neurons single-labeled by the striatal injection (SSL neurons), and doublelabeled neurons (DL ne~ons) were observed in portions of the lateral, basolateral, and basomedial nuclei, as well as in the nucleus of the lateral olfactory tract (layer III, and to a lesser extent layer II; not

Fig. 4. Drawings of injaction sites in animals R81, R63, and RlO. Each animal received one or two injections of TR (cross hatch) into the prelimbic area of the medial PFC (R81 and R63) or AId of the lateral PFC (RlO), and one or two injections of DY (stipple) into the medial striatum (R81 and R63) or ventrolateral striatum (RlO). R63 and RlO are sectioned in the coronal plane. R8I is sectioned in the sagittai plane; note knife cut placed just rostra1 to the nucleus accumbens injection in R81.

A.

J. MC‘DQNALI)

Fig. 5. Coronal sections at OS-mm intervals through the amygdala from rostra1 (A, bregma - 1.8) to caudal (F; bregma -4.3) illustrating the retrograde neuronal labeling observed in IX31(TB injected into the medial PFC, DY injected into the nucleus accumbens; see Fig. 4). Only the boundaries of the nuclei of the basolateral portion of the amygdala are drawn at the four rostra1 levels (A-D). Squares, T&labeled neurons (each symbol represents six neurons}; triangles, DY-labeled neurons (each symbol represents six neurons); closed circles, double-labeled neurons (each sym&l represents three neurons).

illustrated) and amygdalohippocampal area (Fig. 5). The periamygdaloid cortex and posterior cortical nucleus contained moderate numbers of CSL neurons but almost no SSL or DL neurons. The anterior and posterior subdivisions of the basolateral nucleus contained the largest number, as well as the greatest density, of CSL, SSL, and DL neurons. The concentration of these cells was greatest in the medial two-thirds of the nucleus (Fig. 5). Numerous CSL and SSL neurons, and moderate numbers of DL neurons, were also seen in the ventral subdivision of the basolateral nucleus (Fig. SC-F). In the lateral nucleus, CSL and SSL neurons were located primarily in the ventromedial subdivision (Fig. 5ED). Numerous DL neurons were seen in the ventral portion of the nucleus where it borders on the basolateral nucleus. Contralaterai amygd~oid labeling in these cases was considerably lighter than ipsilateral labeling. Most of the labeled neurons in the contralateral amygdala were observed in the medial portions of BLa at the levels of Fig. 5C and D. Most of these cells were SSL neurons, but a few CSL and DL neurons were seen. Small numbers of SSL and CSL neurons

were observed contralaterally in the nucleus of lateral olfactory tract, lateral nucleus, and posterior and ventral subdivisions of the basolateral nucleus. Cases in which mediaf prefrontal cortex injections were paired with injection into the dor~o~ed~ai c~ud~to~ut~en. Injections into the medial PFC were paired with injections centered in the dorsomedial caudatoputamen in 14 rats. The density of retrograde amygdaloid labeling in these animals was proportional to the extent to which the injections involved the medial PFC and its terminal field in the dorsomedial caudatopu~men (see anterograde experiments above). The retrograde labeling in all successful cases was very similar and R63 will be described in detail. The TB injection into the medial PFC in R63 involved the prelimbic and dorsal cingulate cortical areas (Fig. 4). Injections of DY were made into the dorsomedial caudatoputamen at two different rostrocaudai levels (Fig. 4). This portion of the striatum receives projections from the medial PFC but not from the AId of the lateral PFC (see anterograde experiments). The more caudal striatal injection was situated quite dorsally in the striatum.

Amygdaloid projections to prefrontal cortex and striatum

R63

Fig. 6. Coronal sections at OS-nun intervals through the amygdala from rostral (A; bregma - I .8) to caudal (F; bregma -4.3) illustrating the retrograde neuronal labeling observed in R63 (TB injeckd into the medial PFC, DY injected into the dorsomedial caudatoputamen; see Fig. 4). Only the boundaries of the nuclei of the basolateral portion of the amygdala are drawn at the four rostra1levels (A-D). Squares, TB-labeled neurons (each symbol represents six neurons); triangles, DY-labeled neurons (each symbol represents six neurons); closed circles, double-labeled neurons (each symbol represents three neurons).

Since similar injections

in other animals produced only sparse retrograde labeling in the amygdala, the great majority of DY amygdaloid labeling in

R63 is probably due to the more rostra1 striatal injection. The TB retrograde labeling from the cortical injection in R63 is very similar to that seen in R81 except that there was less labeling in the basomedial and cortical nuclei, and periamygdaloid cortex in R63 (Fig. 6). This additional labeling in R81 may be due to the involvement of the medial orbital area by the cortical injection in that animal. The DY retrograde labeling from the striatal injections in R63 is far less extensive than that seen in R81 and is consistent with that seen with injections of WGA-HRP injections into the dorsomedial caudatoputamen (see companion pape?‘). Thus DY-labeled neurons were seen primarily in the medial part of BLa at the bregma -2.3 and -2.8 levels (Fig. 6B, C). There was also DY labeling in the BLv that extended to the caudal pole of the amygdala. Moderate numbers of DL neurons were observd in all amygdaloid domains populated by both TBand DY-labeled cells (Fig. 6). The number of DL

neurons in these areas was proportional to the density of TB- and DY-labeled neurons in the region. The greatest number and density of DL neurons, as well as CSL and SSL neurons, was seen in dorsal and medial portions of BLa (Fig. 6). Cont~ater~ly there were scattered CSL and SSL neurons in the lateral and basolateral nuclei. A few DL neurons were seen in the medial part of BLa. Several SSL neurons were observed in the contralatera1 NLOT, layer III in R63, in addition, some CSL neurons were sezn in other animals with similar injections. Ipsila~r~ly, a large number of SSL neurons and a few CSL and DL neurons were observed in layer III of NLOT. Cases in which lateral prefrontal cortex injections were paired with injections into the ventrolateral caudatoputamen. Injections into the lateral PFC were paired with injections into the ven~ola~r~ caudatop&amen in 12 rats. Although the exact rostrocaudal and dorsoventral coordinates of the ventrolateral striatal injections varied among cases, the pattern of retrograde labeling in the amygdala was very similar in all animals. Case RlO will be described in detail.

10

A. J. MCDONALD

0 =TBIPFCI n=DY .

ICPI

=TB-DY

Fig. 7. Coronal sections at OS-mm intervals through the amygdala from rostral (A; bregma - 1.8) to caudal (F; bregma -4.3) illwtrating the retrograde neuronal labeling observed in RIO (‘I73injected into the lateral PFC, DY injected into the wntrolateral caudatoputamen; see Fig. 4). Only the boundarks of the nuclei of the basolateral portion of the amygdala are drawn at the four rostra1kvels (A-D). !Squares, TBlabekd neurons (each symbol represents six neurons); t&r&s, DY-lahelcd neurons (each symbol represents six neurons); closed circles, double-labekd neurons (each symbol repxsents three neurons).

Two injections of TB were made into the lateral PFC in RIO (Fig. 4). Both injections were centered in the AId but also involved the dorsally adjacent sensorimotor area. A single DY injection was made into the rostra1 ventrolateral caudatoputamen. This striatal region receives a projection from the lateral PFC but is lateral to the striatal region innervated by the medial PFC (see anterograde studies). As illustrated in Fig. 4, the edge of the striatal injection extended as far lateralward as the external capsule. However, the density of DY was very low near the edge of the injection site and the dye did not appear to diffuse into the external capsule. The pattern of retrograde labeling in the amygdala also indicated that these fibers were not part of the effective injection site. Since the external capsule contains amygdalofugal fibers projecting to the AId, involvement of these fibers by the DY injection in RlO would produce retrograde amygdaloid labeling typical of AId injections (e.g. there would be numerous DY-labeled neurons in the lateral nucleus and the lateral part of the posterior subdivision of the basolateral nucleus). This labeling pattern was not observed. Instead the pattern of DY labeled cells was typical of that seen

with injections into the ventrolateral striatum (see companion papeti’). Moreover, other animals with dye injections into the ventrolateral striatum that did not come close to the external capsule produced amygdaloid labeling similar to that seen in RIO. Retrogradely labeled neurons in RIO were located in the lateral, basolateral, and basomedial nuclei (Fig. 7). The great majority of neurons labeled by the cortical and striatal injections, including doublelabeled neurons, were seen in the anterior and posterior subdivisions of the basolateral nucleus. Large numbers of cells labeled by the TB injections into the cortex were observed in the rostra1 pole of BLa (Fig. 7A, B). At more caudal levels, the area of the basolateral nucleus containing the greatest numbers of TB-labeled neurons gradually shifted into ventral and lateral portions of the nucleus (Fig. 7C-F). Most neurons labeled by the DY injection into the ventrolateral striatum were observed in the rostral half of the amygdala (Fig. 7A-D). At these levels the distribution of DY-labeled neurons in the basolateral nucleus was similar to that of TB-labeled neurons. Many DL neurons were seen in the rostra1 half of

Amygdaloidprojections to prefrontal cortex and striatum

the basolateral nucleus (Fig. 7A-C). In general, their distribution and relative density matched that of the CSL and SSL neurons at these levels. DL neurons, as well as SSL and CSL neurons, were also observed along the entire extent of BLv. Very few DL neurons were seen in the lateral nucleus. CSL and SSL neurons were observed in layer III of the nucleus of the lateral olfactory tract but only a few DL neurons were seen (not illustrated). Contralateral labeling was much lighter than that observed ipsilaterally. Moderate numbers of CSL and SSL neurons and a few DL neurons were seen in the rostra1 and caudomedial portions of BLa and the rostra1 part of BLv. The caudal tail of the contralatera1 BLv, which appeared to extend to the caudal pole of the amygdala, contained CSL and SSL neurons but no DL neurons. The contralateral NLOT in RIO had numerous CSL neurons in layer III, but no SSL or DL neurons. Other animals with injections similar to RlO had SSL neurons and an occasional DL neuron. The striatal injections in the latter cases were more dorsally situated than the striatal injection in RlO.

11

peroxidase retrograde tract tracing study of corticostriatal projections are also consistent with the present repOlt2

The pattern of retrograde labeling in the amygdala produced by fluorescent dye injections into the striatum in the present study was similar to that seen with striatal injections of WGA-HRP in the companion paper,” and was consistent with previous anterograde and retrograde tract tracing studies of ~yg~ost~a~l pro~tions.‘s,‘8*38 As ~rno~trat~ in the companion paper,= the main source of amygdalostriatal projections was the basolateral nucleus (BL). Neurons in BL projecting to the medial striatum (striatal target of the medial PFC) and ventrolateral striatum (striatal target of the lateral PFC) exhibited an overlapping to~~ap~cal organization. The projections to the medial striatum originated primarily from the medial two-thirds of BL, whereas the projections to the ventrolateral striatum arose primarily from the ventrolateral comer and rostra1 pole of BL. Injections into the medial or ventrolateral ~udatoputamen (cases R63 and RlO, respectively) labeled mainly the rostral part of BL. Injections that involved the medial part of the nucleus accumbens (e.g. case R81) labeled additional DISCUSSION caudal parts of BL. Portions of the lateral and Anatomical aspects of connections between the basomedial amygdaloid nuclei adjacent to labeled amygakia, prefrontal cortex, and striatum regions of BL usually contained retrogradely labeled The autoradio~ap~c studies of Krettek and neurons. The pattern of retrograde labeling in the amygPrice demonstrated that the main projections of dala produced by fluorescent dye injections into the basolateral amygdala to the PFC were directed at the PL of the medial PFC and the AId of the PL and AId in the present study closely resembled that seen in a previous study of these the lateral PFC.” The anterograde tract tracing experiments in the present study suggest that the projections performed in this laboratory.” The main striatal projections of these cortical areas are source of projections to these prefrontal areas was BL. The neurons in BL projecting to the PL to~grap~~~y organized and extend virtually the entire length of the striatum. The PL of the of the medial PFC and AId of the lateral PFC organizmedial PFC projects primarily to the medial exhibited an overlapping topographical ation. The projections to the medial PFC origicaudatoputamen and adjacent nucleus accumbens, whereas the AId of the lateral PFC projects pri- nated primarily from the medial two-thirds of marily to the ventrolateral ~udatopu~men and BL, whereas the projections to the lateral PFC arose mainly from the ventrolateral comer and the rostra1 fundus striati. There is significant overlap in the striatal projections of these corti- rostra1 pole of BL. Portions of the lateral and cal areas only in the rostra1 fundus striati basomedial amygdaloid nuclei adjacent to labeled retrogradely and adjacent lateral nucleus accumbens. On the regions of BL usually contained and other hand, the striatal projection of the AIv of the labeled neurons. Similarly, Sripanidkulchai lateral PFC, a cortical area which does not receive co-workers retrogradely labeled the medial part of BL with injections of ffuorescent dyes into the strong projections from the basolateral amygdala, appears to exhibit a greater amount of overlap medial PFC.” Although the latter investigators did with the prelimbic projection than that exhibited not make injections into AId, their injections into the caudally adjacent gustatory cortex produced by the AId. The striatal projections of the PL observed in retrograde labeling in the lateral part of BL which was very similar to that obtained in the present the present study closely resemble those obtained study with AId injections (see also Ref. 26). by Beckstead using autoradio~aphic techniques3 Although Beckstead did not report on the projections These results concerning amygdalocortical projecof the AId, his findings on the AIv appear to be very tions appear to contrast with the findings of Sarter similar to those obtained in the present investigation. and Markowitsch,39 who observed numerous cells in The striatal projections of AId and AIv in the the medial part of BL with fluorescent dye injections hamsteti’ resemble those obtained in the rat in the into the lateral PFC. One possible explanation for studv. this discrepancy may be the location of their injection Ioresent ~~~~~~~ ---~-I The results of a recent horseradish

12

A, J.

MCDONALD

sites in the lateral PFC. In the study by Sarter and Markowitsch,3g the injections into the lateral PFC appeared to involve primarily the AIv. These injections produced dense retrograde labeling in the central subdivision of the mediodorsal nucleus, which has a strong projection to the ventral, but not the dorsal, agranular insular area.36 With injections of tritiated amino acids into the basolateral nucleus, Krettek and Price” found silver grains over layer VI of AIv but could not determine whether this label represented axon terminals or fibers passing through AIv to reach AId. It is therefore of interest that injections of WGA-HRP into the deep part of AIv retrogradely label neurons in the medial, but not in the lateral, part of BL (unpublished observations). It seems likely therefore that the involvement of AIv in the lateral PFC injections of Sarter and Markowitsch resulted in large numbers of labeled cells in the medial part of BL; the majority of double-labeled cells seen with their injections into both the lateral and medial PFC may be neurons with axon collaterals projecting to the PL and deep part of the AIv. A novel finding of the present study was the remarkable similarity in retrograde amygdaloid labeling produced by injections into particular prefrontal areas and the striatal targets of these same cortical regions. This striking similarity in the position and relative density of retrogradely labeled neurons was observed not only in the basolateral nucleus, but also in the lateral nucleus, basomedial nucleus, and the amygdalohippocampal area. When injections into the medial PFC were paired with striatal injections that involved the entire rostra1 portion of the striatal target of the medial PFC (i.e. both the medial ~udatoputamen and the nucleus accumbens) the cortex-projecting and striatumprojecting amygdaloid neurons were congruent throughout the rostrocaudal extent of the amygdala (e.g. case R81). When injections into the medial or lateral PFC were paired with injections into the medial and ventrolateral ~udatoputamen, respectively, the cortex-projecting and striatum-projecting amygdaloid neurons were congruent only in the rostra1 half of the amygdala and in the caudal tail of the ventral subdivision of the basolateral nucleus (cases R63 and RlO). Another interesting finding of the present study was that many amygdaloid neurons were doublelabeled when injections of different color dyes were made into the PFC and the striatal target of that same PFC subfield. These results suggest that many amygdaloid neurons, particularly in the basolateral nucleus, send axonal coliaterals to distinct portions of the PFC and their striatal targets. However, fibers of passage can take up and retrogradely transport fluorescent dyes.” Therefore, if amygdalo-prefrontal fibers coursed through the striatum en route to the PFC, they could be labeled by striatal injections. The cells of origin of these fibers would be double-labeled

if injections of different color dyes were made mto the PFC and striatum. The finding of the present study that knife cuts placed just rostra1 to striatal injections did not produce any noticeable decrease in doublclabeled cells argues against the possibility that significant numbers of amygdalo-prefrontal fibers pass through the striatum. These results are also consistent with previous anterograde studies which indicate that amygdaloid projections to the cortex course medial, lateral, and ventral to the striatum rather than through it.” Since no individual animal in the present study received injections into all parts of the cortex and striatum, the exact percentage of amygdaloid neurons that send collaterals to both the cortex and striatum could not be determined. However, it appears that as many as three-quarters of PFC-projecting cells in the basolateral and basomedial nuclei may also send collaterals to the striatum, while as many as twothirds of the striatum-projecting cells in the basolatera1 nucleus, and one-third of the striatum-projecting cells in the basomedial nucleus, may also send collaterals to the PFC. The axonal collateralization demonstrated in the present study is consistent with previous Golgi studies which have demonstrated that axons of pyramidal (class I) projection neurons in the basolateral amygdala frequently give rise to major branches with different trajectories.4~2J.3” The present study indicates that layer III of the NLOT has bilateral projections to the prefrontal cortex, caudatoputamen, and nucleus accumbens. The latter two projections were also documented in the companion paper. 24 Since the fiber bundles containing NLOT efferents (anterior commissure and ~o~ssural bundle of the stria te~nalis) were not involved in the injections sites, these results cannot be due to the labeling of fibers of passage. At first glance these findings appear to contrast with the results of previous studies which only reported projections (bilateral) to more ventral portions of the striatum and cerebral cortex (i.e. mainly to the olfactory tube&e and piriform cortex, respectively13*M). However, Haberly and Price” retrogradely labeled cells primarily in layer II of NLOT with injections into the olfactory tubercle. Likewise, in the autoradiographic studies of NLOT by Luskin and Price,20 it appears that their injections into NLOT involved mainly layer II. Thus, apparently layer 11 of NLOT has bilateral projections to the olfactory tubercle and piriform cortex, whereas layer III projects bilaterally to the caudatoputamen, nucleus accumbens, and prefrontal cortex. In addition, the present study indicates that some of the layer III neurons have axons that branch to innervate both the striatum and prefrontal cortex (ipsilaterally). The failure to observe retrogradely labeled neurons in the central amygdaloid nucleus in the present study contradicts a previous report by Russchen and Price3’ that this nucleus projects to the ipsilateral and

Amygdaloid projections to prefrontal cortex and striatutn contralateral striatum. However, it appears that their PHA-L injection into the central nucleus also involved the basolateral nucleus (see Discussion section of companion paper). Functional

considerations

The telencephalon is largely composed of cortical, striatal and pallidal structures.2*3’ Discrete portions of these three tiers of the forebrain are connected in series to form parallel CSP systems that may subserve specific functions.‘*7*9s3’For example, studies in rats, cats, and primates have shown that lesions of specific portions of the PFC produce deficits in the same type of learning produced by lesions of the main striatal targets of these same prefrontal regions.7 The results of the present study demonstrate that distinct portions of the rat amygdala project to both specific PFC areas and their striatal targets. Moreover, many cells have axons that branch to innervate both an area of the PFC and its associated portion of the striatum. These findings indicate that individual areas of the amygdala, and in some cases individual amygdaloid neurons, can modulate information processing in the first two links of discrete CSP systems arising in the prefrontal cortex. The dense innervation of the striaturn and PFC by the basolateral amygdala suggests that the amygdala can powerfully activate these CSP systems via this dual innervation. Morphological, immunohistochemical, and ultrastructural studies have demonstrated that the basolateral amygdala shares many important features with the cortex and suggest that it can be considered a cortex-like structure.2,5*6~14~21~23,25.30 Since it has been suggested that putative amygdalo-striato-pallidal circuits are analogous to CSP systems,9 it is of interest to compare the organization of amygdalostriatal and corticostriatal connections.5~37 Studies in primates have demonstrated that interconnected cortical areas often project to overlapping regions in the striatum,43 although within the regions of overlap the terminal fields of these cortical areas may actually interdigitate.12 The results of the present investigation indicate

13

that specific basolateral amygdaloid domains, and the specific prefrontal areas with which they are interconnected, project to the same striatal regions. As discussed in the companion paper, a similar relationship may exist between discrete portions of the amygdala and the entorhinal, perirhinal, and subicular cortices. Thus, the triangular relationships between subregions of the amygdala, their associated cortical areas, and the striatum closely resemble the triangular relationships between interconnected cortical areas and the striatum.s*37 The extent of actual overlap versus interdigitation of associated amygdalostriatal and corticostriatal projections to the striatum remains to be determined. These finer details of the interaction of the amygdala and cortex in the striatum could not be observed with the methods utilized in the present study, but require the use of dual anterograde techniques. In addition, the putative triangular relationships of the amygdala and striatum with entorhinal, perirhinal, and subicular cortices must be verified using anterograde and retrograde double-labeling methods. Finally, the amygdala, cortex, and striatum all receive important projections from midline and intralaminar thalamic nuclei and from midbrain The organization of dopaminergic regions. 5*6,9~10,34 projections from the thalamus and midbrain may allow these regions to selectively modulate interconnected portions of the amygdala, cortex, and striatum. For example, as pointed out by Carlsen and Heimer, the interanteromedial thalamic nucleus projects to interconnected medial parts of the PFC, and basolateral amygdaloid nucleus, striatum.5T6 It remains to be determined whether other thalamic nuclei or midbrain dopaminergic neurons may have similar relations with amygdaloid domains and the cortical and striatal regions with which they have connections. Acknowledgements-This

work was supported by National

Institutes of Health grant NS19733. The author is grateful for the excellent technical assistance of Patricia Rnecht and Peggy Sullivan and the secretarial assistance of Melissa Avant and Carol Able.

REFERENCES

1. Alexander G. E., DeLong M. R. and Strick P. L. (1986) Parallel organization of functionally segregated circuits linking basal ganglia and cortex. A. Rev. Neurosci. 9, 357-381. 2.

Alheid G. F. and Heimer L. (1988) New perspectives in basal forebrain organization of special relevance for neuropsychiatric disorders: the striatopallidal, amygdaloid, and corticopetal component of substantia innominata.

Neuroscience 27, l-39. 3. Beckstead R. M. (1979) An autoradiographic examination of corticocortical and subcortical projections of the mediodorsal projection (prefrontal) cortex in the rat. J. camp. Neurol. 184, 43-62. 4. Braak H. and Braak E. (1983) Neuronal types in the basolateral amygdaloid nuclei of man. Brain Res. Bull. 11,349-365. 5. Carlsen J. (1989) New perspectives on the functional anatomical organization of the basolateral amygdala. Actu neural. scund. 79 (Suppl. 122), 4-29. 6. Carlsen J. and Heimer L. (1988) The basolateral amygdaloid complex as a cortical-like structure. Bruin Res. 441, 377-380. 7. Divac I. (1984) The neostriatum viewed orthogonally. In Functions of the Basal Ganglia (eds Evered D. and G’Conner M.), Ciba Foundation Symposium, Vol. 107, pp. 201-210. Pitman, London. 8. Eichenbaum H., Clegg R. A. and Feeley A. (1983) Re-examination of functional subdivisions of the rodent prefrontal cortex. Expl Neurol. 79, 434451. 9. Fallon J. H. and Loughlin S. E. (1987) Monoamine innervation of cerebral cortex and a theory of the role of

14

10. 11. 12.

13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 21. 28. 29. 30. 31. 32. 33. 34. 35.

A. J. MCDONALD monoamines in cerebral cortex and basal ganglia. In Cerebral Cortex (eds Jones E. G. and Peters A.). Vol. 6, pp. 41 I? 7 Plenum Press, New York. Fallon J. H. and Loughlin S. E. (1985) Substantia nigra. In The Rar Nerrous System (ed. Paxinos G.1, Vol. I. pp. 353-374. Academic Press, Florida. Gerfen C. R. and Sawchenko P. E. (1984) An anterograde ne~oanatomi~l tracing method that shows the detailed morphology of neurons, their axons and terminals: immunohistochemical localization of an axonally transported plant lectin, Phaseolus vuigaris leucoagglutinin (PHA-L). Brain Res. 290, 219-238. Goldman-Rakic P. S. and Selemon L. D. (1986) Topography of corticostriatal projections in nonhuman primates and implications for functional parcellation of the neostriatum. In Cerebral Cortex (eds Jones E. G. and Peters A.), Vol. 5, pp. 441465. Plenum Press, New York. Haberly L. B. and Price J. L. (1978) Association and commissural fiber systems of the olfactory cortex of the rat. J. camp. Neural. 178, 711-740. Hall E. (1972) Some aspects of the structural organization of the amygdala. In The ~~rob~oZQg,~ of the Amy&am (ed. Eleftheriou B. E.), pp. 94-121. Plenum Press, New York. Hiikfelt T. G., Skagerberg G., Skirboll L. and BjSrklund A. (1983) Combination of retrograde tracing and neurotransmitter histochemistry. In Handbook of Chemical Neuroanatomy (eds Bjorklund A. and HSkfelt T.), Vol. I, pp. 228-285. Elsevier, Amsterdam. Kelley A. E., Domesick V. B. and Nauta W. J. H. (1982) The amygdalostriatal projection in the rat-an anatomical study by anterograde and retrograde tracing methods. Neuroscieuce 7, 615-630. Krettek J. E. and Price J. L. (19771 Proiections from the amvadaloid corn&x to the cerebral cortex and thalamus in __ the rat and cat. J. camp. Neural. i72, &U-722. Krettek J. E. and Price J. L. (1978) Amygdaloid projections to subcortical structures within the basal forebrain and brainstem in the rat and cat. J. camp. Neurol. 178, 225-254. Kuypers J. G. J. M. and Huisman H. M. (1984) Fluorescent neuronal tracers. Ado. cell. Neurobiol. 5, 307-340. Luskin M. B. and Price J. L. (1983) The topographic organization of association fibers of the olfactory system in the rat, including centrifugal fibers to the olfactory bulb. J. camp. Neurol. 216, 264-291. McDonald A. J. (1982) Neurons of the lateral and basolateral amygdaloid nuclei: a Golgi study in the rat. J. romp. Neural. 212, 293-3 12. McDonald A. J. (1987) Organization of amygdaloid projections to mediodorsal thalamus and prefrontal cortex: a fluorescence retrograde transport study in the rat. J. camp. Neural. 262, 46-58. McDonald A. J. (1989) Coexistence of somatostatin with neuropeptide Y, but not with cholecystokinin or vasoactive intestinal peptide, in neurons of the rat amygdala. Brain Res. 500, 37-54. McDonald A. J. (199l) Topographical organization of amygdaloid projections to the caudatoputamen, nucleus accumbens, and related striatd-like areas of the rat brain. Neuroscience 44, 1.5-33. McDonald A. J. (1991) Cell types and intrinsic projections of the amygdala. In The Amygdah (ed. Aggleton 1. P.). John Wiley, New York (in press). McDonald A. J. and Jackson T. R. (1987) Amygdaloid connections with posterior insular and temporal cortical areas in the rat. J. camp. Neurol. 262, 59-77. McGeorge A. J. and Faull R. L. M. (1989) The organization of the projection from the cerebral cortex to the striatum in the rat. Neuroscience 29, 5033537. McLean I. W. and Nakane P. K. (1974) Periodat&lysine-paraformaldehyde fixative for immunoelectron microscopy. f. Histochem. Cytochem. 22, 1077-1083. Mesulam M. M., Hegarty E., Barbas H., Carson K. A., Gower E. C., Knapp A. G., Moss M. B. and Mu&on E. J. (1980) Additional factors influencing sensitivity in the tetramethyl benzidine method for horseradish peroxidase neurohistochemistry. J. Histochem. dytochem. 28, 1255-1259. Millhouse D. E. and DeGlmos J. (19831 Neuronal confinurations in lateral and basolateral amygdala. Neuroscience 10, _j 1269-1300. Nauta H. 1. W. (1979) A proposed conceptual reorganization of the basal ganglia and telencephalon. Neuroscience 4, 1875-1881. Palkovits M., Tapia-Ar~~bia L., Kordon C. and Epelbaum J. (1982) ~~tostatin connections between the h~o~lamus and the limbic system of the rat brain. SF& Res. 250, 223-228. Paxinos G. and Watson C. (1982) The Rat Brain in Stereotaxic Coordinates. Academic Press, New York. Price J. L., Russchen F. T. and Amaral D. G. (1987) The limbic region II: The amygdaloid complex. In Hdook ofChemical Neuroanatomy (eds Bjorklund A., Hijkfelt T. and Swanson L. W.), Vol. 5, pp. 279-388. Elsevier. Amsterdam. Reep R. L. and Winans S. S. (1982) Efferent connections of dorsal and ventral agranular insular cortex in the hamster, r

Mesocricetus aureatus. Neuroscience 7, 2609-2635.

36. Reep R. (1984) Relations~p

between prefrontal and hmbic cortex: a combative

25, 5-80.

anatomical review. Brain Behao. Eool.

31. Russchen F. T., Bakst I., Amaral D. 6. and Price J. L. (1985) The amygdalostriatal projections in the monkey. An anterograde tracing study. Brain Res. 329, 241-257. 38. Russchen F. T. and Price J. L. (1984) Amygdalostriatal projections in the rat. Topographical organization and fiber morphology shown using the lectin PHA-L as an anterograde tracer. Neurosci. Z.&t. 47, 15-22. 39. Sarter M. and Markowitsch H. J. (19841 Collateral irmervation of the medial and lateral prefrontal cortex by amygdaloid, thalamic and brain stem‘ neurons. J. camp. Newoh 244, 445-460. de&red pathways in the central nervous 40. Sawchenko P. E. and Swanson L. W. (198 1) A method for tracing ~~h~lly system using combined fluorescence retrograde transport and ~uno~st~he~~ techniques. Brain Res. 21% 3 l-5 1. 41. Sripanidkulchai K., Sripanidkulchai B. and Wyss J. M. (1984) The cortical projection of the basolateral amygdaloid nucleus in the rat: a retrograde fluore%%nt dye study. J. camp. Neurol. 229, 419-431. 42. Wouterlood F. G., Bol J. G. J. M. and Steinbusch H. W. M. (1987) Double-label hnmunocytochemistry: combination of anterograde neuroanatomical tracing with Phaseolus uulgurb leucoagglutinin and enzyme immunocytochemistry of target neurons. J. H&=tochem. Cytochem. 35, 817823. 43. Yeterian E. H. and Van Hoesen G. W. (1978) Cortico-striate projections in the rhesus monkey: the organization of certain cortico-caudate connections, Brain &s. 139, 4363. (Accepted

10

March

1991)