Organization of striatopallidal, striatonigral, and nigrostriatal projections in the macaque.

THE JOURNAL OF COMPARATIVE NEUROLOGY 304569-595 (1991)

Organization of Striatopallidal, Striatonigral, and Nigrostriatal Projections in the Macaque JOHN C. HEDREEN

AND MAHLON

R. DELONG

Departments of Pathology (J.C.H.), Neurology (M.R.D.),and Neuroscience (J.C.H.,M.R.D.), The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205

ABSTRACT The topographic organization of neostriatal connections was investigated by axonal transport of horseradish peroxidase, tritiated amino acids, or mixtures of both injected into the neostriatum of macaque monkeys. Striatal projections to pallidum and substantia nigra and the origin of projections to striatum from cerebral cortex and substantia nigra were examined. All striatal injections gave rise to projections to external and internal pallidum and to substantia nigra. Injections in caudate nucleus and in putamen both gave rise to substantial projections to pallidum and to substantia nigra, and the ratio of pallidal and nigral projections was generally similar. The striatopallidal projection showed prominent arborizations at right angles to the striatofugal pathway traversing the pallidum, forming in this manner terminal fields consisting of multiple bands or discs within a broad segment of the pallidum. Thus separate but neighboring regions of striatum appeared to have overlapping pallidal projection territories. In broad terms, rostral striatum projects to rostral pallidum, caudal striatum to caudal pallidum, and dorsal and ventral striatum, respectively, to dorsal and ventral pallidum. Inner (medial) and outer (lateral) putamen showed only subtle differences in pallidal projection patterns. The striatonigral projection from each injected area of striatum formed a longitudinal band extending over the entire length of the substantia nigra, with scattered, dense terminal fields occupying portions of pars compacta as well as pars reticularis. Rostral striatum projected to medial nigra and caudal striatum to lateral nigra. Terminal fields from ventral striatum were located somewhat more dorsally in the substantia nigra than those from dorsal striatum. Neighboring but separate regions of striatum appeared to have overlapping nigral projection territories, especially in caudal nigra. The nigrostriatal neurons projecting to an injected area of striatum generally were located in the same longitudinal band of the substantia nigra as the corresponding striatonigral projection. Labeled pars compacta neurons were often surrounded by a dense, labeled striatonigral terminal field, suggesting the existence of a striato-nigrostriatal loop. The rostromedial pars compacta contained labeled neuronal cell bodies in most cases, suggesting a widely divergent projection to striatum from this cell group. A slight tendency for preferential cell labeling rostrally in nigra with rostral striatal injection and caudally in nigra with caudal injections was noted. The preferred relationship of lateral nigra with caudal striatum and medial nigra with rostral striatum has implications for clinical expression of Parkinson's disease, which may vary with differential involvement of different nigral cell groups along the medial to lateral axis. In cases of mid-putamen injection, corticostriatal neuronal perikarya were labeled in a broad frontoparietal zone encompassing most motor and somatosensory areas, while a case of rostral caudate injection displayed labeled cells in four separate cortical zones: cingulate gyrus, prefrontal cortex medial to the sulcus principalis, superior temporal gyrus, and medial temporal cortex (including entorhinal cortex).

Accepted October 15,1990. Dr. DeLong is now at the Department of Neurology, Emory University School of Medicine, 401 Woodruff Memorial Building, P.O. Drawer V, Atlanta, GA 30322.

o 1991 WJLEY-LISS, INC.

J.C. HEDREEN AND M.R.DELONG

570

The projections to pallidum from injected areas of caudate nucleus and from injected areas of putamen showed little overlap. This finding is consistent with the concept of parallel, segregated pathways through the basal ganglia involving regions of neostriatum and pallidum controlled by association cortex and sensorimotor cortex. Neither the striatopallidal projection, nor the striatonigral projection, nor the origin of the nigrostriatal projection, nor the origin of the corticostriatal projection showed a simple point-to-point topographical organization, but rather ended in multiple terminal fields of complex organization or originated in multiple cell groups. This complex mode of connectional organization is shared by many other non-sensory neuronal circuits in the central nervous system and is likely to be a key feature in their function. Key words: neostriatum,cerebral cortex, substantia nigra, globus pallidus, putamen

The topographic organization of the connections of the neostriatum is of particular interest in relation to the functional subdivisions within this nucleus created by projections from different cortical regions (Alexander et al., 1986; DeLong and Georgopoulos, 1981; Divac, 1977; Selemon and Goldman-Rakic, 1985). It has been proposed that these functional subdivisions form part of segregated cortico-basal ganglia-thalamocortical circuits that are organized in a parallel manner (Alexander et al., 1986; DeLong and Georgopoulos, 1981) and which engage distinct regions of the neostriatum, globus pallidus, substantia nigra, and thalamus. Clarification of the divergent and convergent patterns of connections between neuronal populations of the neostriatum and the regions with which it maintains axonal connections is necessary for exploration of this parallel circuit hypothesis and is a prerequisite for understanding the functional organization of basal ganglia circuitry. In this regard, it is of considerable interest that many of these connections are not point-to-point connections, in the manner of those responsible for generating retinotopic maps in the lateral geniculate nucleus and primary visual cortex (Bowling and Michael, 1984; Gilbert and Wiesel, 1979; Mason and Robson, 1979; Perkel et al., 1986) but show considerable degrees of divergence and complex patterning in their axonal projections (Gerfen, 1985; Goldman and Nauta, 1977bj. The projections from caudate nucleus and putamen neurons to both segments of the globus pallidus and to the substantia nigra, and from nigral pars compacta neurons back to the striatum, were first established with certainty in the early 1960s (Anden et al., 1964; Szabo, 1962; Voneida, 1960). The topographic organization of these connections has always been a subject of interest but the techniques available to earlier investigators did not allow a complete demonstration of the striatofugal or nigrostriatal axons and terminal fields. The development of anterograde and retrograde axonal transport connection-tracing methods using tritiated amino acids, horseradish peroxidase (HRPj, and lectins provided the opportunity for numerous studies giving a more complete demonstration of the striatal efferent and afferent systems. Nevertheless, a number of questions concerning the topographic organization of these systems remain unresolved, especially in primates. For the striatopallidal projection, there is now broad agreement on the general organization of projections from different striatal regions along the dorsoventral and rostrocaudal axes. Dorsal, ventral, rostral, and caudal regions of neostriatum have been found to project to corresponding

regions of both external and internal segments of pallidum (Cowan and Powell, 1966; DeVito et al., 1980; Grofova et al., 1982; h o o k , 1965; Nauta and Mehler, 1966; Percheron et al., 1984; Smith and Parent, 1986; Szabo, 1962, 1967, 1970). However, the detailed divergent branching patterns of this projection have received little attention, except in reports of Golgi studies of the globus pallidus (Fox and Rafols, 1975; Percheron et al., 1984), which describe only individual axon collaterals. The topography of the striatopallidal projection originating from outer and inner striatal regions, i.e., regions close to cortical white matter vs. those close to the pallidum, is also not well understood. Cowan and Powell (1966) suggested the possibility that outer regions of caudate and putamen might terminate in the external pallidal segment and inner regions in the internal pallidal segment. A report by Kim et al. (1976) included an illustration of a case with an injection in outer putamen near subcortical white matter showing a projection to the lateral regions of both external and internal pallidal segments. This result suggests another possible basis for organization of these projections, with the corollary that inner striatal regions close to the pallidal external medullary lamina might project to medial parts of both pallidal segments. This issue has remained unresolved. There is less agreement on the topographic organization of the striatal projection to the substantia nigra than on that to the globus pallidus. A variety of studies in different species have produced somewhat conflicting findings and interpretations (Arbuthnott, 1978; Bunney and Aghajanian, 1976; Cowan and Powell, 1966; Desban et al., 1989; Domesick, 1977; Gerfen, 1985; Grofova et al., 1982;Johnson and Rosvold, 1971; Nauta and Mehler, 1966; Niimi et al., 1970; Parent et al., 1984; Royce and Laine, 1984; Smith and Parent, 1986; Szabo, 1962, 1967, 1970; Tulloch et al., 1978), and little information on the topographic organization of this projection in primates is available. The topographic organization of the nigrostriatal system in monkeys has also remained obscure. A number of reports (Beckstead et al., 1979; Deutch et al., 1986; Fallon and Moore, 1978; Francois et al., 1984a; Gerfen et al., 1987; Jimenez-Castellanos and Graybiel, 1989; Parent et al., 1984; Smith and Parent, 1986; Szabo, 1980) have described partial maps of the topographical distribution of nigrocaudate and nigroputaminal neurons in rats, macaques, and squirrel monkeys, but the findings and interpretations are not in complete agreement, even within single species. Most studies in rats conclude that medial striatum exchanges connections with medial nigra and lateral striatum with lateral nigra (Beckstead et al., 1979; Deutch et al., 1986;

PROJECTIONS IN THE MACAQUE Domesick, 1977; Fallon and Moore, 1978; Van der Kooy, 1979). However, some evidence also exists for a shift in axis, with the anterior striatum exchanging connections with medial nigra and the posterior striatum with lateral nigra (Faull et al., 1986; Van der Kooy and Kuypers, 1979). Studies in primates have not dealt clearly with this question, the emphasis being placed instead on contrasts in connectivity between caudate nucleus and putamen. The topography of the corticostriatal pathway was first studied in detail by Webster (1961) and Knook (1965) in rats, by Carman et al. (1963) in rabbits, by Webster (1965) in cats, and by Kemp and Powell (1970) in monkeys. These initial studies emphasized a point-to-point type of topography in this system. In contrast, use of the autoradiographic technique revealed that the projection from a particular cortical region terminates in a number of longitudinally oriented terminal fields in neostriatum (Donoghue and Herkenham, 1986; Goldman and Nauta, 1977b; GoldmanRakic, 1982; Jones et al., 1977; Kunzle, 1975, 1977a; Kunzle and Akert, 1977b; Kunzle, 1978; Malach and Graybiel, 1986; Selemon and Goldman-Rakic, 1985; Stanton et al., 1988; Yeterian and van Hoesen, 1978). Studies of this system using retrograde transport of HRP have concentrated on describing the laminar location of the corticostriatal neurons (Arikuni and Kubota, 1986; Kitai et al., 1976; Royce, 1982; Tanaka, Jr., 1987), and few studies of the regional topography of cortical afferents to striatum (McGeorge and Faull, 1989) have been made using this technique. Other unresolved questions concerning neostriatal connections in primates include the following: whether all regions of caudate nucleus and putamen show projections to both segments of pallidum and also to substantia nigra; whether tracer injections of equal size in caudate nucleus and putamen give rise to projections of equal size in the target zones (Smith and Parent, 1986); and whether projections from different functionally identifiable zones of striatum, such as that defined by projections from somatosensory, motor, and premotor cortical areas versus that defined by projections from association cortex (DeLong and Georgopoulos, 1981; Percheron et al., 1984), show distinctive features in their connectional patterns. This last question is of particular interest in relation to the issue of whether segregated parallel paths are maintained through the basal ganglia circuit by different functional systems (Alexander et al., 1986; DeLong and Georgopoulos, 1981). The present study addresses these issues using anterograde and retrograde axonal transport connection-tracing methods following injection of tritiated amino acids mixed with HRP or of HRP alone or tritiated amino acids alone into inner, outer, anterior, posterior, dorsal, or ventral regions of the putamen and four areas of the caudate nucleus in macaque monkeys.

METHODS Injections of tracer were placed in the putamen or caudate nucleus in eleven hemispheres of six Macaca fascicularis monkeys. Animals were anesthetized using sodium pentobarbital (35 mg/kg) and placed in a stereotaxic apparatus. The boundaries of cerebral cortex, subcortical white matter, putamen, and external pallidum were mapped in stereotaxic coordinates by means of multiple penetrations with recording electrodes using either a vertical approach or an oblique approach at an angle of 45" from vertical (Alexander, 1985b; Crutcher and DeLong, 1984a,b;

571

DeLong, 1972, 1973). The different classes of spontaneously active neurons in these structures were found to retain most of their characteristic firing patterns under barbiturate anesthesia. Stereotaxically placed injections of mixtures of tritiated amino acids and HRP, or HRP alone, or tritiated amino acids alone, were made with a 1 r~.1 Hamilton syringe, according to coordinates derived from the physiological mapping. In some cases a recording electrode was glued to the syringe needle, its tip extending 2.5 or 3.0 mm beyond the opening of the syringe needle. This electrode was used to guide the exact placement of the injection. The injected HRP solutions contained 8-20% HRP in buffered Hanks solution. Mixtures of tritiated proline, lysine, leucine, and phenylalanine (72-145 FCiI mM, 250-500 FCi total) (New England Nuclear, Boston, MA) were evaporated and dissolved in 5 pl of the HRP solution or in Hanks solution alone. Doses for each of the cases are given in the legend to Figure 1. Injections were made on both sides because the striatal efferent pathway remains unilateral (Nauta and Mehler, 1966; Voneida, 1960). The corticostriatal and nigrostriatal pathways include contralateral components; the contralateral cells of origin in these systems are reported to have the same topographical localization as the ipsilateral cells (Francois et al., 1984a; Jones et al., 1977). The possibility of labeling contralateral nigrostriatal and corticostriatal cells is taken into account in the description and interpretation of our retrograde transport results. A 2-day postoperative survival was selected to obtain optimal retrograde and anterograde labeling with the HRP technique (Hedreen and McGrath, 1977). After 2 days, the animals were anesthetized with pentobarbital and perfused with a cannula inserted through the left ventricle into the ascending aorta, following intravenous administration of heparin and sodium nitrite (Palay and Chan-Palay, 1974). A variety of perfusion protocols were used initially. In the protocol finally adopted, 1,000 ml of 0.12 M sodium phosphate buffer, pH 7.3, containing 2.5% sucrose and 1 ml of 1% sodium nitrite was perfused, followed by 5 L of 1.5% glutaraldehyde (Hedreen et al., 1976) in the same sucrosebuffer solution. The first half of the glutaraldehyde solution was perfused rapidly and the second half very slowly (total perfusion time 45 minutes) (Rosene and Mesulam, 1978). After perfusion the brain was blocked in the stereotaxic apparatus (a single coronal cut was made behind the substantia nigra), removed from the skull and kept in the fixative solution for 4-6 hours and then placed into 15% sucrose in 0.12 M phosphate, pH 7.3, for 24 hours (refrigerator) followed by 30% sucrose in this buffer for 3 days. It was then frozen on a sliding microtome stage and sectioned at 40 Fm. The sections were collected serially in Tris buffer (0.1 M, pH 7.6) in 24-compartment plastic boxes. Every sixth section was processed for HRP and the adjoining sections for autoradiography. Horseradish peroxidase histochemistry using 3,3',5,5'tetramethylbenzidine (TMB) was performed within 1week, following a procedure derived from that of DeOlmos et al. (1978). This horseradish peroxidase protocol was repeatedly compared with the TMB procedure of Mesulam (1978, 1982). No difference in sensitivity was noted, but the modified DeOlmos et al. procedure produced a less troublesome nonspecific crystalline deposit. The procedure is as follows: 1) Rinse in 0.01 M sodium acetate buffer pH 4.3. 2) Incubate in 100 p1 incubation solution without hydro-

572

J.C. HEDREEN AND M.R. DELONG

1L

3L

5L

2L

1R

3R

2R

4L

5R

Fig. 1. Injection sites in putamen and caudate nucleus (11 hemispheres in six monkeys). Case 1was given 0.2 pl of 8%HRP containing 10 pCi tritiated amino acids in each hemisphere. Case 2 was given 0.05 ~18% HRP and 2.5 pCi tritiated amino acids in each hemisphere. Case 3 received 0.2 pI 10%HRP and 10 pCi tritiated amino acids on the right

side and 0.1 p1 10% HRP and 5 pCi tritiated amino acids on the left. Case 4 received 0.15 p1 10% HRP. Case 5 received 0.1 ~ 1 2 0 % HRP in each hemisphere. Case 8 received 0.2 pl containing 10 pCi tritiated amino acids on each side.

gen peroxide for 10 minutes in a freezer at - 15 to -2O"C, agitating every 5 minutes. 3) Add 1.0 ml freshly diluted 1%hydrogen peroxidase and incubate for an additional 20-60 minutes at - 15 to -20°C

according to the results of an initial trial incubation, with agitation every 5 minutes. 4) Rinse in four changes of 0.01 M acetate buffer, pH 4.3, at 0-4"C, with 0.6 ml 1%hydrogen peroxide per 80 ml (rinses

PROJECTIONS IN THE MACAQUE of 1,2, 3, and 4 minutes with agitation), and allow to come to room temperature. 5) Rinse in fresh 0.02 M sodium periodate, pH 3.5, for 30 minutes (change if turns yellow). 6) Mount sections onto subbed slides from acetate buffer with hydrogen peroxide (step 4). Blot and dry for 60-90 minutes. 7) Rinse in acetate buffer with hydrogen peroxide (step 4) for 10-30 seconds and then counterstain lightly (10-30 seconds) with 0.5%cresyl violet in the same solution. 8) Rinse (30 seconds) in acetate buffer with hydrogen peroxide and then dehydrate and differentiate with agitation in 70% propanol in acetate buffer with hydrogen peroxide (15 seconds), 95% propanol in acetate buffer with hydrogen peroxide (15 seconds), 100%propanol(30 s), and three changes of 1-butanol (n-butyl alcohol) (30-90 seconds each). 9) Clear in three changes of xylene (1, 2, 3 minutes) with intermittent agitation. 10) Coverslip using DPX (or other nonreducing mounting medium).

573

varies between species and even between individuals of the same species and therefore does not form a useful basis for functional analysis of basal ganglia organization (Carman, 1966;Nauta, 1979).A more precise topographical nomenclature is preferable. The mutually perpendicular rostrocaudal, dorsoventral, and mediolateral axes are not necessarily optimal for use in studies of the topography of the corpus striatum; this structure may be better analyzed as a radial system (Cowan and Powell, 1966; Grofova, 1979; Nauta and Mehler, 1966; Percheron et al., 1984; Wilson, 1914). Thus, in the putamen, proximity to the subcortical white matter vs. proximity to the globus pallidus may be more relevant in the study of topographical organization of connections than position on the medio-lateral axis (i.e., distance from the midline).In this paper, we use the terms outer and inner to refer to areas of the lenticular nucleus according to their radial distance along axes radiating from the ventromedial edge of the globus pallidus to the subcortical white matter. In the substantia nigra, a longitudinal axis parallel to the course of the striatonigral fibers and of the underlying fibers of the cerebral peduncle should be understood when the terms The incubation solution contained 80 ml distilled water, anterior and posterior or rostral and caudal are used. The 10 ml 0.1 M acetate buffer, pH 4.3, 0.5 g gelatin, 1.0 ml terms medial and lateral and dorsal and ventral should be dimethyl sulfoxide, 6.4 ml 95% ethanol, 2.6 ml of 100% understood as referring to axes perpendicular to this longiethanol in which 7.0 mg of TMB had been dissolved by tudinal axis. sonication, and 67 mg sodium nitroferricyanide. Preliminary trials (e.g., two sections each at 20, 30, 45, and 60 RESULTS minutes, or two sections each for 45 minutes with 0.67,1.0, Injections were placed in the putamen in outer, inner, and 1.33 ml of 1%hydrogen peroxide) were run for each dorsal, ventral, rostral, and caudal positions, and in the case, using the same ratio of number of sections to incubation bath volume as in the final incubations (maximum 1.5 head of the caudate nucleus near the cortical white matter sections110 ml). An incubation time or hydrogen peroxide and internal capsule, at the edge of the internal capsule well concentration was selected such that nonspecific crystal away from the subcortical white matter, and in a far rostral deposits were just beginning to appear. Some of the above position. The stereotaxic procedure described above was procedures are based on our observations that the tetrame- highly accurate. The injection sites are shown in Figure 1. thylbenzidine reaction product is stable in acid and oxidiz- Both tritiated amino acids and HRP provided a clear ing solutions, but disappears in basic or reducing solutions, demonstration of the striatofugal pathway. The tritiated and is more stable in propanol and butanol than in ethanol. amino acid procedure labeled the axonal pathway more The coverslipped sectionswere mapped within a few months; strongly and the HRP procedure labeled the terminal fields subsequent fading of tetramethylbenzidine reaction prod- more strongly, at the survival time used. The distribution of the terminal fields in globus pallidus and substantia uct commonly occurred. Autoradiography was performed on every sixth section nigra was the same with both methods (see, e.g., Fig. 6). following standard methods for frozen sections with cresyl Neither the anterograde transport autoradiography method violet counterstaining (Cowan et al., 1972; Hendrickson et nor the HRP anterogradelretrograde transport method is al., 1972). Both types of sections were mapped freehand thought to show significant transport in intact fibers of under microscopic examination on outline drawings of passage (Cowan et al., 1972; Hedreen and McGrath, 1977; cytoarchitectural features and other landmarks made on a Oldfield and McLachlan, 1977). In some cases a portion of projection device (Bausch and Lomb Tri-Simplex micropro- the injected tracers was present in the external capsule, but jector modified for continuous vertical adjustment, with no significant transport from fibers of passage was detected (e.g., labeling of cortical cell bodies with HRP in layers Zeiss Luminar lenses). Anterograde projections to external and internal divisions of the globus pallidus and substantia nigra and retrogradely labeled neuronal cell bodies in substantia Figs. 2-5. Striatal projection in case lR, in which 0.2 I.LIof 8%HRP nigra and cerebral cortex were mapped. Retrogradely latritiated proline, leucine, and lysine was injected into the beled neurons in the thalamic intralaminar nuclei, dorsal containing outer edge of the midrostral putamen (A). The projections to external raphe nucleus, and other regions are not described in this pallidum (B-HI, internal pallidum (F-I), and substantia nigra (GU), paper. and the pathway of efferent axons from the injected site are illustrated. Although the putamen and caudate nucleus, divisions Note the band-like terminal fields in the external pallidurn, the dense created in the neostriatum by the internal capsule, corre- terminal field in the center of the internal pallidum, and the patchy spond roughly in macaque brain to the neostriatal regions terminal fields in the substantia nigra throughout its anteroposterior length. AC, anterior commissure; C , caudate nucleus; CP, cerebral receiving input from sensorimotor and from association peduncle; GPe, external pallidum; GPi, internal pallidum; IC, internal cortex, respectively, (DeLong and Georgopoulos, 1981; capsule; P, putamen; Po, pons; R, reticular nucleus of the thalamus; Percheron et al., 1984), this correspondence is imprecise. SNc, substantia nigra pars compacta; SNr, substantia nigra pars The division of the neostriatum by the internal capsule reticulata; STN, subthdamic nucleus.


574

2 Fig. 2 See preceding page.

3 Fig. 3. See precedingpage.

PROJECTIONS IN THE MACAQUE

575

4 Fig. 4. See page 573 for legend.

5 Fig. 5. See page 573 for legend


576

other than V or labeling of corticofugal axons in thalamus or pons). Each striatal injection gave rise to projections to both segments of the globus pallidus and to the substantia nigra. No other structures were identified which received striatal projections. In addition, each striatal injection resulted in retrograde labeling in both substantia nigra and cerebral cortex.

Striatal efferent pathway The striatofugal pathway and its pattern of terminations are shown in detail in a single case in Figures 2-5. The injection in this case (case 1R in Fig. 1)was at the outer and dorsal edge of the rostral putamen at a level just in front of the globus pallidus (Fig. 2, level A). Labeled axons are seen in bundles passing through the putamen (Fig. 2, levels B,C) to enter the rostral external globus pallidus midway up its lateral edge (Fig. 2C,D). Multiple vertical bands of terminal labeling are present in the external globus pallidus (Figs. 2B-D, 3E-H, 6A,B). The bands continue from one section to the next and thus are disc-shaped with an orientation parallel to the internal and external medullary laminae of the pallidum. In the internal pallidum a dense field of terminal labeling is seen in the midlateral portion (Fig. 3G,H). After passing through both segments of globus pallidus, descending axons traverse the internal capsule (Fig. 4I,J), and pass below the subthalamic nucleus to enter the lateral edge of the cross-sectionof the rostral substantia nigra (Fig. 4K). At a more caudal level, where the substantia nigra extends farther laterally, the labeled fibers come to be located in the ventral edge of the center of the substantia nigra (Figs. 4M, 5N). They maintain this position until the caudal end of the substantia nigra, giving rise to terminal fields throughout their rostrocaudal course (Figs. 4M-5U), except just after entering the nigra at the rostral-most end (Fig. 4L). In the other cases, the projections followed a similar course except for topographic shifts, as described below.

dorsal, and ventral to ventral. A topographic pattern in the radial dimension was much more difficult to discern. Comparison of projection patterns to pallidum after injections in inner and outer putamen suggested a difference in radial position of the multiple pallidal bands in these two situations. Thus, injections into inner putamen gave rise to terminal field bands at the outer and inner edges of external pallidum and the outer edge of internal pallidum as well as in the central region of each pallidal division (Fig. 7 , cases 1L, 3L). Outer putamen injections also gave rise to terminal fields in the central region of each pallidal division, but bands at the edges were less prominent (Figs. 2-3, case 1R; Fig. 8, case 2L). Injections in inner putamen appeared to give rise to terminal field bands in globus pallidus of approximately the same dorsoventral extent as injections of equal size in outer putamen (Figs. 2,3, and 7 , cases l R , 1L). Comparison of the different cases showed that the dorsoventral, rostrocaudal, and radial spread of terminal fields within the pallidum was great enough to indicate a considerable territorial overlap of the terminal fields deriving from neighboring portions of the striatum (Fig. 7, cases 1L vs. 3L; Fig. 10, cases 8L vs. 8R). Nevertheless, comparison of caudate projections with those from putamen (e.g., cases 8L, Fig. 10, vs. lR, Figs. 2,3) showed little or no evidence of overlap.

Substantia nigra

Injections of inner midlevel putamen (cases 1L and 3L) produced highly divergent band-like terminal fields in external pallidum and in the lateral and medial portions of internal pallidum (Fig. 7 ) .The terminal fields in the globus pallidus appeared less banded and more diffuse with injections of rostral caudate nucleus (cases 4L and 5R) and caudal putamen (cases 2R and 2L) (Fig. 8) than with injections of the midlevel putamen. It was difficult to discern whether this different appearance might be due to differences in the orientation of the disc-like labeling relative to the plane of section. The caudate nucleus injections (cases 4L, 5R, 8L, and 8R, Figs. 9 and 10) revealed projections to a rostral and dorsal region occupying approximately two fifths of the pallidum. In the rostrocaudal and dorsoventral orientations, the general organization of the striatal projection to globus pallidus was straightforward: rostral portions of striatum project to rostral pallidum, caudal to caudal, dorsal to

The striatal projections to the substantia nigra gave rise to terminal fields in pars compacta as well as in pars reticulata (Figs. 6C,D, ll),involving especially the fingerlike portions of pars compacta that extend downwards into the reticulata. In each case the striatonigral projection took the form of a longitudinally oriented band or strip containing labeled axons, giving rise to multiple dense patches of labeled terminals separated by unlabeled regions. As described earlier for case 1R (Figs. 2-5), labeled fiber bundles entering the rostrolateral substantia nigra appeared to travel a short distance within the nigra before beginning to arborize. With this exception, the longitudinal bands of terminal fields extended the full anteroposterior length of the substantia nigra. The bands were broad, occupyingfor a typical injection about one quarter to one third of the width of the substantia nigra. The position of these bands of terminal fields along the mediolateral axis of the substantia nigra depended on the position of the striatal injection in the rostrocaudal axis of the neostriatum. The most rostral striatal injections produced bands of termination in medial substantia nigra (Fig. 12, case 5R), and more and more caudal injections produced bands more and more laterally in substantia nigra (Figs. 12 and 13; compare also Fig. 11).Nevertheless, caudal striatal injections were sometimes noted to give rise to an apparent terminal field in the rostromedial pars compacta (Fig. 12, case 5L). Figure 14 shows a composite view of the substantia nigra from above, with the bands from four cases illustrated diagrammatically. These bands are not oriented in a straight anteroposterior direction but parallel instead the fibers of the underlying cerebral peduncle. The most

Fig. 6 . A,B: Band-like terminal field in external globus pallidus (case 1R) shown in closely spaced sections prepared by autoradiography (A) and for HRP (B). The distribution of apparent terminal fields seen with the two techniques was very similar. L, lateral; M, medial. Bar = 30 pm. C,D. Photomicrographs of striatonigral terminal field labeling

with HRP in cases 5L (C) with midlevel injection of putamen and 5R (D) with rostral injection of caudate nucleus. Note in both cases the presence of clusters of retrogradely labeled pars compacta cells within the labeled striatonigral terminal field. L, lateral; M, medial. Bar = 30 bm.

Globus pallidus


577


578

3

C

.-

Fig. 7. Band-like terminal fields in external and internal pallidum in two cases (3L, 1L) with inner putaminal injections at a middle anteroposterior level. For each case, the most rostral section is at upper

left and the most caudal at lower right. Dorsal, ventral, medial, and lateral directions are indicated. AC, anterior commissure; Gpe, external pallidum; GPi, internal pallidum; IC, internal capsule; P, putamen.


Fig. 8. Apparent absence of band-like outlines in striatopallidal terminal fields in internal pallidum in cases with caudal putamen injections (ZR,2L). For each case, the most rostra1 section is at upper

579

left and the most caudal at lower right. Dorsal, ventral, medial, and lateral directions are indicated. GPe, external pallidum; GPi, internal pallidum; IC, internal capsule; P, putamen.

580

Fig. 9. Apparent absence of band-like outlines in striatopallidal terminal fields in both external and internal pallidum in cases with rostral striatal injections (5R, 4L). For each case, the most rostral section is at upper left and the most caudal at lower right. Dorsal,


ventral, medial, and lateral directions are indicated. AC, anterior commissure; C , caudate nucleus; Gpe, external pallidum; GPi, internal pallidum; IC, internal capsule; P, putamen.


.. -.. .,., .,..

.

581

582

Fig. 11. Two maps (cases 5L, 5R) showing both terminal field labeling and retrogradely labeled neurons in substantia nigra. The 5R map is reversed for clearer comparison with the 5L map. The terminal


fields involve portions of the pars reticulata and some, but not all, clusters of labeled pars compacta neurons. SNc, substantia nigra pars compacta; SNr, substantia nigra pars reticulata.


583

lateral bands therefore are not present in cross-sections of the brainstem at the level of the rostralmost substantia nigra and after they enter the substantia nigra they proceed medially as they pass caudally. The result of this situation is that transverse sections cut symmetrically at right angles to the axis of the brainstem do not show projections of caudal striatum to substantia nigra at the most rostral nigral levels. If sections were instead cut at right angles to the striatonigral fiber bundles and the cerebral peduncle fibers, projections from caudal striatum to the lateral portions of the most rostral sections of nigra taken in this manner would then be observed. The bands from neighboring regions of striatum clearly overlap, including those from caudate (e.g., case 4L in Fig. 13) and sensorimotor zone of putamen (e.g., cases 1R in Figs. 2-5, 3L in Fig. 13), but whether synaptic overlap on postsynaptic neurons of nigral reticulata or compacta occurs is not certain, because, within the territory of the band, the terminals occur in patches leaving intervening regions free of termination. The territorial overlap is most extensive at the caudal end of the substantia nigra (Fig. 14). Note that caudate projections (first and second bands from medial edge on left in Fig. 14) show clear territorial overlap with putamen projections (second and third bands). Dorsalinjections (e.g., cases lL, lR, and 4L,Figs. 2-5,11, 13) gave rise to labeled striatonigral axons and terminal fields concentrated preferentially in the ventral aspect of the substantia nigra, while more ventral injections (e.g., case 3L, Fig. 13) gave rise to labeled axons and terminal fields preferentially located more dorsally in the substantia nigra. However, while this rule holds well for the striatonigral axon bundle within the substantia nigra, terminal fields do not always adhere to the rule; for example, some terminal fields in case 3L (Fig. 13) are located at the ventral edge of the pars reticulata. No clear variation in projection topography was discerned for the radial axis in the striatum (e.g., cases 1Rvs. 3L, Figs. 2-5, 13).

Q 5L

0

Nigrostriatal neuron labeling

"

,.

Fig. 12. Detailed mapping of striatonigral terminal fields showing their distributionthroughout the longitudinalextent of the substantia nigra in cases with far-rostral (5R) and middle-caudal (5L) striatal injections.The more rostral striatal injection gives rise to a more medial terminal field territory in the substantia nigra,whereas the more caudal injection gives rise to a terminal field occupying a more lateral territory in the nigra. The caudal injection nevertheless gives rise also to apparent terminals in the rostromedial pars compacta. Dorsal, ventral, medial, and lateral directionsare indicated.

Each case demonstrated numerous, widely distributed groups of retrogradely labeled neurons in the pars compacts. In particular, neuronal perikarya in those areas of pars compacta which contained labeled striatonigral terminal fields were often retrogradely labeled, although many other groups of retrogradely labeled cells had no accompanying fields of anterogradely labeled terminals (Figs. 6C,D, 11). Labeled pars compacta cells tended to occur in groups located mainly within or medial to the broad, rostrocaudally oriented bands of the striatonigral terminal fields (Fig. 11). Thus, there was a correspondence between the rostrocaudal position of the striatal injection and the mediolateral position of the main band of labeled neurons in the substantia nigra (Figs. 15, 16). Scattered labeled cells were often present medial to the main band, and, in particular, cells in the most rostral and medial pars compacta, often well outside the main band of labeling, were often labeled (Figs. 15, 16). Cases with smaller injections tended to produce a narrower band containing fewer groups of labelled cells, while still showing labeled cells in the anteromedial pars compacta. Cases with caudal striatal injections showed a somewhat larger number of labeled neurons at caudal levels of substantia nigra than cases with rostral striatal injections (case 5L vs. 5R, Fig. 15). Contralateral nigrostriatal neuron labeling was sought by comparing labeling patterns in cases with striatal injec-


584

D

+v Fig. 13. Striatonigral terminal field territories in cases with rostral (4L) and middle-rostra1 (3L)striatal injections. The terminal fields in Figures 12 and 13 occupy progressively more lateral territories in the

substantia nigra. Compare also case 1R (Fig. 2) with striatal injection rostral to that of 3L. SNc, substantia nigra pars compacta; SNr, substantia nigra pars reticdata.

tions at different anteroposterior levels in the two hemispheres, giving rise to bands of labeled cells in the substantia nigra on the two sides that were separated in the mediolateral axis. Assuming that contralateral nigrostriatal cells have the same topographic location as ipsilateral (Francois et al., 1984a), these cases showed little evidence of significant contralateral nigrostriatal neuron labeling. For example, in case 5R (Fig. 15), very few cells are labeled in the lateral substantia nigra on the right, where contralaterally projecting cells would be expected to be labeled resulting from the injection in the left caudal striatum (case 5L).

Corticostriatal neuron labeling Neurons in upper layer V of cerebral cortex were abundantly labeled in both hemispheres in each of the two animals with the largest HRP injections (cases 1and 51, but more sparsely labeled in the others. The labeling in each hemisphere in cases 1 and 5 covered broad portions of cerebral cortex. A continuous, uninterrupted population of labeled cells was present predominantly in layer V in the labeled cortical regions. After injection of putamen (Fig. 17, left hemisphere of case 51, retrogradely labeled cortical neurons were present


585

5L

0

o,,

Fig. 14. Semidiagrammatic outline of right substantia nigra seen from above, with striatonigral terminal field bands from four cases with progressively more caudal striatal injections (5R, 4L, lL, and 2L). Anterior and posterior directions and midline are shown on the left. The medial and lateral free boundaries of the cerebral peduncle (crus cerebri) are shown by dotted lines. The projections in the individual cases were mapped onto this outline using measurements in the mediolateral axis of the ratio of medial unlabeled substantia nigra, labeled substantia nigra, and lateral unlabeled substantia nigra. The elongated striatonigral projection fields tend to parallel the underlying cerebral peduncle fibers. Note the progressively more lateral territory corresponding to progressively more caudal striatal injections, the extensive territorial overlap corresponding to nonoverlapping striatal injections, and the increase in this overlap posterior in the substantia nigra.

in a broad region of frontal, parietal, and cingulate cortex, principally occupying the motor and premotor frontal regions and the somatosensory parietal regions. AfFerents to striatum are reported to arise from contralateral as well as ipsilateral motor and premotor areas (Fallon and Ziegler, 1979; Jones et al., 1977; Kunzle, 1975).Since the contralatera1 areas mirror the ipsilateral, interpretation of bilaterally injected cases becomes unclear in those cortical locations which are bilaterally labeled. Lack of cortical labeling in most motor and premotor areas on the right side of case 5 (Fig. 17, right), contralateral to the putamen injection, indicates that contralaterally projecting cells are poorly labeled by retrograde transport in this instance. The extent of ipsilateral labeling in the other two well-labeled hemispheres with putamen injections (cases lL, 1R)was similar to that in case 5L and was again predominantly confined to a broad region of caudal frontal and rostral parietal lobes and to the dorsal bank of the cingulate gyrus. Following the description of von Bonin and Bailey (1947), areas with labeled layer V neuronal perikarya in the left hemisphere of case 5 included FC, FB, FA (and FA-L), PA, PB, PC, PE, PF, and (with a small number of labeled cells) PG (Fig. 17, left). The ventral (vertical) component of cingulate gyrus has not been included in this list because it may have been labeled from the contralateral injection. Case 1, with bilateral injections of putamen, had an abundance of labeled cells in the dorsal (horizontal) portion of the cingulate gyrus, so that those found in this location in case 5L (Fig. 17, left) probably represent ipsilaterally projecting cells. In contrast to the broad zone of labeling seen after putamen injections, retrograde labeling from the rostral caudate nucleus was found in four separate cortical zones,

v

-

Q

n

v

Fig. 15. Detailed map of retrogradely labeled neurons in substantia nigra in cases 5L and 5R.Note that these cells are located principally within or medial to the zones previously defined for the nigrostriatal terminal fields. There is a tendency for cells labeled from caudal regions of striatum (case 5L) to be located in greater quantity in the caudal substantia nigra than cells labeled from rostral striatum (case 5R).


586

1R

U

Fig. 16. Map of retrogradely labeled neurons in the substantia nigra in cases 1L and 1R. Again, the labeled cells are located principally within or medial to the zones defined for the striatonigral terminal fields (compare case 1R in Figs. 2-5). In Figures 15 and 16, note that

each case includes labeled cells in the anteromedial pars compacta. SNc, substantia nigra pars compacta; SNr, substantia nigra pars reticulata; STN, subtbalamic nucleus.

in anterior and posterior cingulate gyrus (LA and LC), prefrontal granular cortex medial to the sulcus principalis (FD), superior temporal gyrus including the dorsal bank of the superior temporal sulcus (TA, TB), and medial temporal cortex (TF, TH), including entorhinal cortex at the level of amygdala and hippocampus (Fig. 17, right). The labeled cells in cerebral cortex in these cases were principally located within a single band centered in upper layer V (Fig. 18). Labeled cells extended into the region immediately above the boundary of layer V, where the latter is defined by a sudden increase in the number of very small neurons (Fig. 18),but otherwise were only sparsely distributed in more dorsal regions or in layer VI.

caudate nucleus; the apparent separation of the caudatopallidal projection territory from that deriving from putamen, in contrast to the overlap of the nigral projections from these two striatal regions; evidence in primate brain favoring a striatonigrostriatal loop via the pars compacta of the substantia nigra; the presence of a cell group in rostromedial pars compacta of the substantia nigra with projections to most of the neostriatum; and the different patterns of retrograde cortical labeling noted from putamen and from rostral caudate nucleus, including a projection to the latter from entorhinal cortex.

DISCUSSION A number of new findings emerge from these studies, including the absence of simple point-to-point connectional relationships between the striatum and its afferent and efferent partners; the connectional association of the rostrocaudal striatal axis with the mediolateral nigral axis; the projection to all three target areas (external and internal pallidum and substantia nigra) from all injected regions ofthe neostriatum; the approximately equal size of the nigral projections from equal volumes of putamen and of

Striatopallidal projection The widely accepted radial organization of the striatopallidal system was confirmed by our findings. But this projection is not of a point-to-point type. Rather, each injected region of neostriatum provided terminal fields to a broad, radial segment of both external and internal pallidum. The presumed terminal fields identified in the autoradiographic or horseradish peroxidase preparations consisted predominantly of multiple, apparently disc-shaped branchings oriented at right angles to the radially aligned main bundles of striatofugal axons, in both external and internal pallidum. This branching was thus in the plane of the main dendrites of the pallidal neurons (DiFiglia et al.,


587

Fig. 17. Retrograde cell labeling in cerebral cortex in cases 5L (injection in left putamen at level D) and 5R (injection in right rostralmost caudate nucleus). Note the broad labeling of dorsomedial

parietal and caudal dorsomedial frontal lobe on the left, in contrast to separate labeling of several distinct cortical fields on the right. For interpretation of bilaterally labeled cortical regions see text.

1982; Fox and Rafols, 1975; Francois et al., 1984b; Percheron et al., 1984; Yelnick et al., 1984). The pallidal projection field from a single injected region of striatum is large, encompassing a much greater proportion of the pallidum than the injection site does of the striatum. Thus, there is necessarily a high degree of overlap

in the pallidal projection territories of neighboring regions of striatum. However, the multiple discs comprising the projection field do not fill the territory they define, and projections from other striatal regions with overlapping pallidal territories may at least partly end on different sets of pallidal neurons, whose dendrites arborize in different

588

Fig. 18. Retrogradely labeled (filled) and unlabeled (outlines) neuronal perikarya in layer V and adjacent portions of layer IIIiIV and VI in motor cortex of case 1R. Note that labeled neurons are distributed


mainly in upper layer V, and in the immediately adjoining region just above the upper boundary of layer V. Upper and lower laminar boundaries of layer V are shown at each side of the drawing.


589

planes along the radial axis, rather than necessarily converg- positive and substance P-positive striatopallidal projections ing on the same neurons. Nevertheless, comparison of are demonstrated (e.g., Haber and Elde, 1981), a uniform different striatal injection cases suggests possible conver- innervation of the pallidum is seen, without prominent gence in certain regions, namely, upon neurons along the disc-like configurations. Thus, very large striatal injections external border of the external pallidal segment and possi- of anterograde tracer are likely to show more of a uniform bly along its inner border and along the external and terminal field with a loss of the disc-like images seen with ventral borders of the internal pallidal segment as well. The smaller injections (e.g., Smith and Parent, 1986). internal pallidal zones of possible convergence appear to correspond to those from which the pallidal projections to Striatonigral projection the habenula originate (Parent, 1979; Parent et al., 1981). The results indicate that populations of neurons located Questions of convergence and of differences between projections from outer versus inner striatum could better be within the approximately spherical regions of neostriatum addressed using double label anterograde techniques. receiving the injected tracer distribute their nigral terminal Some of the striatopallidal termination zones, including fields in patches within broad, longitudinally oriented those from the caudate nucleus, appeared to be more diffuse bands in the substantia nigra and that these terminal fields in nature. It is not clear whether this appearance is involve pars compacta in addition to pars reticulata. These indicative of their true three-dimensional shape or whether longitudinal bands traverse the entire length of the SN instead a disc-like organization is obscured because of the from the point of entry along the rostrolateral border to the plane of section. If these termination zones are truly diffuse caudal end of the substantia nigra. They are situated in the rather than disc-like, our findings are consistent with the medial to lateral axis of the substantia nigra according to possibility that the pallidal projections from sensorimotor the rostral to caudal position of the injection in the neostristriatum (as defined by corticostriatal projections) have a atum. This organization recalls that of the neocortical mode of termination with an emphasis on orientation projection in the cerebral peduncle underlying the substanperpendicular to the radial axis of the pallidum, whereas tia nigra, in which medial fibers derive from rostral and those from the association striatum (as defined by projec- lateral fibers from caudal cortex. This similarity in pathway tions from association cortex) have a more diffuse mode of topography between striatofugal and corticofugal projectermination, with fewer axons branching perpendicular to tions raises the question of whether there might be a the radial axis. However, this latter possibility seems relationship between the developmental mechanisms underunlikely, because the dendrites of the pallidal neurons and lying this topography for the two systems. It is unclear Golgi-stained striatofugal axon branches in the caudate whether individual striatal efferent neurons innervate the portion of the globus pallidus are reported to retain the entire length of the substantia nigra, or whether the orientation perpendicular to radial axes of the pallidum longitudinal strips are built up from much more restricted seen in the putaminal pallidum (Francois et al., 1984b; terminal fields of individual striatal neurons. The demonPercheron et al., 1984;Yelnick et al., 1984). It is of interest stration by Gerfen (1985) in rats of long, longitudinally that anterograde labeling of the subthalamic projection to oriented striatonigral terminal fields in phaseolus lectin pallidum (Carpenter et al., 1981; Nauta and Cole, 1978; tracing experiments involving only a few striatal axons Smith et al., 1990) produces disc-like images parallel to the favors the former alternative. The subthalamic nucleus pallidal laminae similar to those of the striatopallidal projection to the substantia nigra also gives rise to patchy projection. This mode of termination is also likely to be terminations along the full anteroposterior extent of the determined by the disc-like orientation of pallidal dendrites nucleus (Kita and Kitai, 1987; Nauta and Cole, 1978; Ricardo, 1980). (Percheron et al., 1984). The longitudinal, patchy nature of the terminal fields Each case shows separate discs (or multiple regions of termination of other apparent morphology)along the radial makes it clear that a simple point-to-point topographic axis of the pallidum. Thus, each region of striatum appears relationship does not exist between the neostriatum and to send projections to multiple regions of each of the two the substantia nigra. The functional interpretation of this segments of the pallidum. While the multiplicity of termi- patterned divergence is presently unclear. The retrograde nal zones within each pallidal segment could reflect the labeling results suggest that the neurons located in patches transport of label in multiple striatal efferent neurons, each throughout the longitudinal zone of nigra innervated by a with a specific pallidal target, it could also result from small region of striatum send axons back to that striatal multiple collaterals of single axons. The latter situation has region. Studies of nigral projections in rats (e.g., Beckstead been demonstrated following intracellular injection of HRP et al., 1979), however, demonstrate that small regions of into striatal efferent neurons in rats (Chang et al., 1981). substantia nigra project to large, longitudinally oriented Such multiple branching of a striatofugal axon is probably zones of striatum, so that a private circuit between a small restricted to a single segment of pallidum, however, for it region of striatum and a longitudinal zone of substantia now appears clear that striatal efferents to each pallidal nigra is apparently not maintained. The rostrocaudal to mediolateral90" shift in the striatosegment (and to substantia nigra) arise from distinct populations of striatal neurons (Feger and Crossman, 1984; nigral projection appears to differ from that reported by Parent et al., 1984; Beckstead and Cruz, 1986; Koliatsos et most investigators in rats, in which medial striatal injections are described as leading to anterograde labeling of al., 1988; Gimenez-Amayaand Graybiel, 1990). The size of the focal, sometimes disc-shaped striatopall- medial substantia nigra and lateral injections to labeling of idal terminal field images in these cases is determined by lateral nigra. However, a report by Faull et al. (1986) the size of the injection of transported marker. The disc-like confirms that at least to some extent a rostrocaudal striatal terminal fields are a composite image produced by the to mediolateral nigral relationship exists also in rats. The branching patterns of several hundred labeled striatal illustrations in Gerfen's study (1985) also provide some efferent axons. In contrast, when the entire enkephalin- support for this relationship in the rat striatonigral system.


590

A tendency for dorsoventral inversion of the striatonigral projection was noted, confirming reports in rats (Domesick, 1977; Gerfen, 1985). The functional significance of this partial dorsoventral organization remains uncertain, as the dendrites of nigral neurons extend long distances in the dorsoventral axis as well as in other directions (Francois et al., 1987; Yelnick et al., 1987), and it is therefore possible that a projection from dorsal striatum ends on nigral neurons that also receive input from more ventral striatal regions at the same anteroposterior level. No difference in termination pattern in the substantia nigra was detected for projections from inner versus outer striatum. As in the case of the pallidum, double label anterograde experiments would be needed to better address this question. A notable feature of the striatonigral projection is the patchy nature of the dense terminal fields in the substantia nigra (both reticulata and compacta). This mode of termination was found after very small injections of tracer by Gerfen (1985) in rats using the phaseolus lectin anterograde transport technique and was also demonstrated after very large injections by Smith and Parent (1986) in monkeys using HRP-labeled wheatgerm agglutinin. The nigral band in which the projection from a particular region of neostriatum terminates is broad, and these bands overlap to a considerable extent even for projections from relatively widely separated striatal regions, especially if these regions are at similar anteroposterior levels. This organization is suggestive of convergence. However, because of the patchy nature of termination within the nigral band corresponding to a particular striatal region, it is possible that projections from two regions of striatum, whose nigral terminal territories overlap, do not in fact converge onto the same nigral neurons. Comparison of projections from the zone of striatum that receives input from association cortex (e.g., case 4L in Fig. 13) with those from the zone of striatum receiving from sensorimotor cortex (e.g.,case 5L in Fig. 12) shows that the nigral bands from these regions overlap. Smith and Parent (1986) showed such territorial overlap in experiments in squirrel monkeys with multiple large injections of label in caudate nucleus on one side of the brain and putamen on the other. They interpreted the patchy terminal fields from the two sources as interdigitating, one projection filling the spaces left vacant by the other, rather than converging on equivalent groups of neurons. Their finding supports the hypothesis that separate striatonigral projections with nigral projection territories that overlap in a broad sense may in fact terminate on separate populations of nigral neurons. The terminal field labeling in the pars compacta is of an extent and density to make very unlikely the possibility that any significant proportion of this labeling results from retrograde transport of HRP in nigrostriatal axons followed by labeling of local collateral terminal fields of these axons. Furthermore, the tritiated amino acid labeling, which results from anterograde labeling of labeled protein and not from retrograde transport, shows the same distribution in the substantia nigra as the HRP terminal labeling.

Relative volume of striatopallidal vs. striatonigral projections Smith and Parent (1986) described in squirrel monkeys a distinctly greater projection of putamen to globus pallidus and of caudate t o substantia nigra. Our studies show that injections of tracer of approximately equal size in caudate and in putamen provide an approximately equivalent vol-

ume of projections to substantia nigra from each region (see Fig. 12, cases 5L and 5R). In contrast, the striatopallidal projection from an injected region of putamen appears to cover a greater territory in the globus pallidus than the projection labeled by an injection of equivalent size in the caudate nucleus (Figs. 7-10). However, this discrepancy need not indicate the presence of a greater number of synaptic terminals for individual putaminopallidal than for caudatopallidal neurons; terminals may instead be less densely dispersed over a wider territory and thus be more heavily intermixed with terminals of other putaminopallidal neurons. It has been estimated that the relative volumes of caudate nucleus and putamen are 43157 (Carman et al., 1965) in macaques (with a range in six animals of 38:62 to 49:51). The ratio of pallidum occupied by terminal fields from caudate nucleus versus putamen would appear to be roughly similar. Our observations, although limited to 11 cases each showing projections from a small fraction of the total striatal mass, are in accord with the concept that each region of striatum (of the size defined by our injections) contains neuronal populations which project to external pallidum, internal pallidum, and substantia nigra, and that the relative proportion of such projections to these three targets is about the same for each striatal region. Retrograde labeling studies (Beckstead and Cruz, 1986; Feger and Crossman, 1984; Koliatsos et al., 1988; GimenezAmaya and Graybiel, 1990), although also restricted in the fraction of the total striatal volume studied, support this conclusion.

Nigrostriatal projection and striato-nigro-striatal loop In many instances the labeled nigrostriatal neurons were surrounded by a dense cloud of terminal field label (presumed labeling of terminal boutons). This finding suggests the possible existence of a striato-nigro-striatal “lo~p,’’ although the synaptic contact cannot be proved from the present material. Strong evidence for striatonigral synapses on dopaminergic nigrostriatal pars compacta neurons has been provided by several ultrastructural studies in rats (Chang, 1988; Kawai et al., 1987; Somogyi et al., 1981, 1982; Van den Pol et al., 1985; Wassef et al., 1981). Gerfen (1984, 1985, 1989; Gerfen et al., 1987) has provided strong evidence for separate systems of connections involving striatal patch and matrix regions in rats, and Graybiel and colleagues (Jimenez-Castellanosand Graybiel, 1987; Langer and Graybiel, 1989) have described certain parallel findings in cats and squirrel monkeys. According to these authors, the more ventral dopaminergic neurons of the substantia nigra project to neostriatal patches. It is these more ventral pars compacta neurons, rather than more dorsally situated groups, which appear likely to receive input from the regions of striatum injected in the present study. The nigrostriatal neuronal cell bodies labeled by retrograde transport of HRP were broadly distributed, although the majority were located within the same broad, anteroposteriorly oriented bands as the labeled terminal fields generated by the same striatal injection. The nigral cells labeled following an injection of marker into a restricted region of striatum occurred in clusters, with intervening unlabeled cells. These findings rule out a simple, point-to-point topography for the nigrostriatal connection. Numerous


591

labeled nigrostriatal neuronal perikarya without any surSmith and Parent (1986) and Parent (1983) described the rounding terminal label were seen. This finding is compati- distribution of nigrocaudate and nigroputaminal neurons ble with the possibility that certain compacta neurons, in the squirrel monkey substantia nigra following multiple, labeled by retrograde transport from the striatal HRP large injections of two different retrograde markers into injection but showing no surrounding labeled terminals, caudate and putamen. They found that groups of labeled may receive such terminals from another region of stria- nigrostriatal neurons related to caudate nucleus and to tum, whose longitudinal terminal field band overlaps with putamen were present in an interdigitated fashion, occupythat of the first. If this is the case, the detailed topographic ing a wide mediolateral position, and extending through the relations in the postulated striato-nigro-striatal “loops” entire anteroposterior extent of the nigra. This result is become complex, with the striatal region at one end of the entirely consistent with our findings in macaque. Considloop communicating with a different striatal region at the ered in the light of the findings of Parent and colleagues, other end. This argument assumes the existence of striaton- our results suggest that nigrostriatal neurons projecting to igral synapses on pars compacta neurons, as discussed regions of putamen and caudate nucleus located at the same anteroposterior position within the corpus striatum would above. Many of the striatal injections, regardless of anteroposte- occupy separate but closely juxtaposed zones located at the rior position within the striatum, gave rise to retrograde same mediolateral level of the substantia nigra. The funclabeling of neurons in the anteromedial pars compacta. tional consequences and advantages of this arrangement This cell group evidently has an especially broad, divergent presently remain unclear. Szabo (1980) made injections of axonal projection in the neostriatum, although individual HRP in different regions of caudate nucleus and putamen cells may project only to limited regions of striatum. As this in macaque and squirrel monkeys. His illustrations reveal rostromedial cell group receives input from rostral as well the presence of nigrostriatal neurons in a broad band as other striatal regions (Fig. 121, it may provide a pathway extending the full anteroposterior length of the nigra in each case, in agreement with our findings. Although his from limbic to motor portions of striatum. interpretation does not address this issue, it is clear from The distribution of retrogradely labeled nigrostriatal the figures that the mediolateral position of the band of neuronal perikarya was similar to that of anterogradely labeled cells in the nigra corresponds to the anteroposterior labeled striatonigral terminal fields. Both occupied a longiposition of the striatal injection, consistent with the present tudinal band traversing the rostrocaudal axis of the substan- findings. tia nigra, whose mediolateral position corresponded to the position of the injection of label along the rostrocaudal axis Corticostriatal projection of the neostriatum. Thus, in a cross section of substantia nigra, the midlateral and middle dopaminergic cell groups Each of the three cases with large putamen injections are strongly related to the functional area of the striatum, (lL, lR, and 5L) displayed a broad distribution of labeled primarily located in the putamen, which is supplied with cortical neurons within caudal frontal and rostral parietal corticostriatal input from the motor and somatosensory lobes, including various motor and premotor subdivisions regions of neocortex. The medial dopaminergic cell groups (Muakkassa and Strick, 1979; Matelli et al., 19861, somaare preferentially related to the rostral neostriatum, includ- tosensory cortex, and area 5 . Labeled cells were also present ing the head of the caudate and the rostral putamen, in the dorsal bank of the cingulate g y r u s ; it is not clear regions whose functional attributes are created by input whether this region is to be regarded as a motor area, but it primarily from frontal association cortex and limbic areas has been described as projecting to area 4 (Muakkassa and (Divac, 1977; Nauta, 1979; DeLong and Georgopoulos, Strick, 1979). These findings suggest that projections from 1981). Assuming that the topographic relationships be- broad cortical areas converge onto the putamen but do not tween striatum and nigra are the same in human as in necessarily indicate convergence onto single neostriatal macaque brain, the present findings give a means of neurons (Alexander et al., 1988; Selemon and Goldmancorrelating clinical findings in Parkinson’s disease patients Rakic, 1985; Yeterian and van Hoesen, 1978). The wide with differential nigral cell loss in the mediolateral axis or extent of the retrograde cell labeling in cerebral cortex differential striatal loss of dopaminergictransmission mark- following focal injections of putamen and the labeling of ers in the rostrocaudal axis at autopsy. Bernheimer et al. distinct, noncontiguous regions of cortex following an (1973) found the most severe nigral cell loss in the middle injection of the rostral caudate nucleus indicate that a (alonga mediolateral axis) cell groups of the pars compacta. simple,point-to-pointtopographicrelationship is not present Kish et al. (1988) found that the motor representation between cortex and striatum. regions of putamen showed more severe dopamine loss than The location of labeled cortical neurons after a rostral the rostral head of caudate in autopsy brain tissue from caudate injection confirmedreports of convergenceof projecParkinson’s disease patients. These findings are consistent tions from separate (usually interconnected) regions of with the present results on topographical relationships association cortex onto single regions of the caudate nubetween substantia nigra and neostriatum. Rinne et al. cleus (Selemon and Goldman-Rakic,1985;Yeterian and van (1989) found that middle and lateral nigral cell groups Hoesen, 1978). The present case is the first to include showed the greatest cell loss in Parkinson’s disease, while entorhinal cortex as a site of origin in such a multiple patients with an accompanying dementia showed in addi- corticocaudate system. A projection from entorhinal area to tion increased cell loss in the medial nigral cell groups. The the neostriatum has been described by Sorensen and Witter medial nigra, according to the present study, connects with (1983) in rat, guinea pig, and cat. Selemon and Goldmanrostral striatum, a region receiving input from association Rakic (1985) and Yeterian and Van Hoesen (1978) came to areas of frontal lobe. Loss of dopaminergic innervation of opposite conclusions about whether the converging corticothis region may therefore provide at least a partial explana- striatal projections showed overlap of their terminal fields or interdigitation. Because the present study used the tion for the dementia in these cases.

592


tion territory. The above-mentioned physiological studies confirm this predicted organization but do not indicate any significant convergence between upper limb and lower limb projections to pallidum, despite the predicted strong apparent overlap. The physiological findings suggest that, as discussed above, the apparent overlap in projections from separate striatal regions may be illusory, with functional segregation being maintained for individual pallidal neurons. These physiological studies thus appear to indicate that the orthogonal orientation of the dendrites of pallidal neurons does not reflect an integration of diverse functional inputs; rather, they appear to collect highly specific input from functionally homogeneous groups of neostriatal neurons. In the substantia nigra, no clear separation was found between caudate and putamen projections. These projections (as well as the location of the cell bodies giving rise to the reciprocal nigroputaminal and nigrocaudate projections) in fact show considerable overlap in their distribution, confirming the findings of Parent et al. (1983) and Smith and Parent (1986). It should be emphasized, however, that the reports of Parent and colleagues suggest that the putaminonigral and caudatonigral projections predominantly end on interdigitating groups of nigral neurons, suggesting that convergence onto individual nigral neurons may not occur. Whether segregation is maintained in basal ganglia circuits passing through the substantia nigra is not clear from the present data. The elongated, longitudinally oriented somatotopicrepresentations in the putamen of upper limb, lower limb, and face are approximately coextensive in their anteroposterior extent (Alexander and DeLong, 1985a,b; Crutcher and DeLong, 1984a,b). This implies that their nigral projections are likely to consist of longitudinally oriented bands in the substantia nigra that are approximately coextensive in their mediolateral extent and that, because of the long anteroposterior extent of the cell groups of origin in the putamen, these bands are probably rather wide in the mediolateral axis. Preservation of somatotopic topography could result from a separate distribution of the patchy Parallel, segregated pathways terminal fields from each somatotopic region within these Evidence for segregation of parallel, functionally distinct wide bands, and from the slight differentiation in the pathways through the cortico-basal ganglia-thalamocorti- dorsoventral axis of the preferred region of nigral terminacal pathway has been reviewed previously (DeLong and tion. Georgopoulos, 1981; Alexander et al., 1986). The present findings are compatible with the concept of a distinct Axonal arborization patterns boundary between the pallidal territory for the injected For both the striatopallidal and the striatonigral projecregions of the head of the caudate (related to association cortex) and that for the injected regions of the putamen tions, one of the most interesting findings of this study was (related to sensorimotor cortex). These results suggests the divergent three-dimensional form of their terminal that circuits of the corticostriatopallidal pathway, involving fields. It is clear that the post-synaptic neurons for neurons on the one hand the sensorimotor part of putamen and on located within small areas of striatum are spread over the other the regions of caudate nucleus dominated by relatively large regions of the target nuclei. Furthermore, the cells of origin in both the corticostriatal and nigrostriassociation cortex, remain segregated. The sensorimotor representation in the putamen con- atal pathways are known to be spread over quite large areas sists of longitudinally oriented zones for lower limb (dorso- of the cortex and substantia nigra. Thus, it is evident that neither the corticostriatal, nor the lateral), upper limb, and face (ventromedial) (Alexander and DeLong, 1985a,b; Crutcher and DeLong, 1984a,b; striatopallidal, nor the striatonigral, nor the nigrostriatal Jones et al., 1977; Kunzle, 1975, 1977a). Physiological projection is organized in a point-to-point fashion in the studies (DeLong, 1971; DeLong et al, 1985; Mitchell et al., manner of the primary sensory systems. Instead, each site 1987; Hamada et al.,1990 in press) show that a somato- in a given region sends axons to multiple sites in its target topic organization is maintained within the external and regions and each site in a target region receives input from internal pallidum. The present findings would suggest that multiple, scattered groups of cells in regions projecting to it the pallidal region for (e.g.1 the upper limb is probably (Beckstead and Cruz, 1986; Gerfen, 1985; Koliatsos et al., elongated in the anteroposterior direction and is probably 1988). This type of complex connectional organization is centered below that for the lower limb and above that for common to many nonsensory regions of the brain, includthe face, but with considerable apparent overlap of projec- ing corticocortical connections (e.g., Goldman and Nauta,

retrograde method, it does not offer any resolution of this question. Corticostriatal neurons in our cases were found in a single band, largely confined to layer V, especially layer V-A (confirmingJones et al., 1977), with the upper edge of the band of labeled cells lying just above the upper boundary of layer V, a few were also found higher in layers IV and I11 or in upper layer VI. Significant numbers of corticostriatal cells in layers other than V have been reported by several investigators (Arikuni and Kubota, 1986; Jinnai and Matsuda, 1979; Kitai et al., 1976; Kubozono et al., 1986; Oka, 1980; Royce, 1982, 1983; Tanaka, Jr., 1987; Wilson, 19871, especially in nonprimate species. These were not encountered in our cases except immediately above the upper border of layer V. It is of interest that the labeled corticostriatal neurons were not restricted to particular “columns” or local cortical regions but extended continuously in layer V, even across areal boundaries, as was found also in rats (Hedreen, 1977). At the edges of this band of labeled neurons, however, some patchy grouping of labeled cells was noted, and the cases with smaller injections showed similar patchy grouping of cells, raising the possibility that projections to small regions of striatum may arise from discontinuous patches of cortical cells. The corticostriatal neurons were more difficult to label by retrograde transport than other striatal afferent neuronal populations. This finding may be due in part to the fact that corticostriatal axons terminate along a long anteroposterior strip within the neostriatum, and only a fraction of the terminals are exposed to the injected label. In addition, it has been reported that many cortical neurons projecting to the ipsilateral striatum also give rise to collateral projections to the contralateral striatum and other regions (Donoghue and Kitai, 1981; Fisher et al., 1984, 1986; Royce, 1983), so that a considerable additional proportion of the full terminal fields of these neurons would be outside of the zone containing injected label.

PROJECTIONS IN THE MACAQUE 1977a; Schwartz and Goldman-Rakic, 1984; Selemon and Goldman-Rakic, 1988)and corticopontine connections (e.g., Brodal, 1978; May and Andersen, 1986), among others. Some insight into the detailed organization of these complex systems of connections between neuronal populations has been provided by double-label anterograde and retrograde studies (e.g., Alexander et al., 1988; Beckstead and Cruz, 1986; Feger and Crossman, 1984; Gerfen, 1985; Koliatsos et al., 1988; Parent, 1983; Parent et al., 1984; Selemon and Goldman-Rakic, 1988; Smith and Parent, 1986); these experiments tend to reveal an interdigitation of the multiple terminal fields from one source with those from another, or of the multiple clusters of cell bodies innervating one target with those innervating another. But the functional significance of this complex mode of neuronal circuit organization remains unclear. Important principles of brain function will likely be revealed by a better understanding of these patchy, interdigitating, divergent and convergent modes of connection.

ACKNOWLEDGMENTS The authors thank Michael Crutcher, Glenn Holm, and Linda Watermeier for assistance with animal surgery and recording and with tissue processing, Christine YoungMcKenna for many of the drawings (the more elegant ones), Magid Fotouhi, Frank Barksdale, and Eric Rottenberg for assistance with planning and printing of the illustrations, and Charles Gerfen, Garrett Alexander, Vassilis Koliatsos, Yoland Smith, Cheryl Kitt, and Michael Crutcher for critically reviewing the manuscript. These studies were supported by NIH grants NS13812, NS15417, NS16375, NS23160. and NS20471.

LITERATURE CITED Alexander, G.E., and M.R. DeLong (1985a) Microstimulation of the primate neostriatum. I. Physiologicalproperties of striatal microexcitable zones. J Neurophysiol. 53:1401-1416. Alexander, G.E., and M.R. DeLong (1985b) Microstimulation of the primate neostriatum. 11. Somatotopic organization of striatal microexcitable zones and their relation to neuronal response properties. J. Neurophysiol. 53:1417-1430. Alexander, G.E., M.R. DeLong, and P.L. Strick (1986) Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Annu. Rev. Neurosci. 93357-381. Alexander, G.E., V.E. Koliatsos, L.J. Martin, J. Hedreen, and I. Hamada (1988) Organization of primate basal ganglia “motor circuit”: I. Motor cortex (MC) and supplementary motor area (SMA)project to complementary regions within matrix compartment of putamen. SOC.Neurosci. Abstr. 143720. (Abstract) Anden, N.-E., A. Carlsson, A. Dahlstrom, K. Fuxe, N.-A. Hillarp, and K. Larsson (1964) Demonstrating and mapping out of nigroneostriatal dopamine neurons. Life Sci. 3.523-530. Arbuthnott, G.W. (1978) Studies on the interactions between the substantia nigra and the neostriatum. Adv. Biochem. Psychopharmacol. 19.235248. Arikuni, T., and K. Kubota (1986) The organization of prefrontocaudate projections and their laminar origin in the macaque monkey: A retrograde study using HRP-Gel. J. Comp. Neurol. 2443492-510. Beckstead, R.M., and C.J. Cruz (1986) Striatal axons to the globus pallidus, entopeduncular nucleus and substantia nigra come mainly from separate cell populations in cat. Neuroscience 19:147-158. Beckstead, R.M., V.B. Domesick, and W.J.H. Nauta (1979) Efferent connections of the substantia nigra and ventral tegmental area in the rat. Brain Res. 1753191-217. Bernheimer, H., W. Birkmayer, 0. Hornykiewicz, K. Jellinger, and F. Seitelberger (1973) Brain dopamine and the syndromes of Parkinson and Huntington: Clinical, morphological and neurochemical correlations. J. Neurol. Sci. 20:415455.

593 Bowling, D.B., and C.R. Michael (1984) Terminal patterns of single, physiologically characterized optic tract fibers in the cat’s lateral geniculate nucleus. J. Neurosci. 4r198-216. Brodal, P. (1978) The corticopontine projection in the rhesus monkey: Origin and principles of organization. Brain 1013251-283. Bunney, B.S., and G.K. Aghajanian (1976) The precise localization of nigral afferents in the rat as determined by a retrograde tracing technique. Brain Res. 11 73423435. Carman, J.B., W.M. Cowan, and T.P.S. Powell (1963) The organization of cortico-striateconnexions in the rabbit. Brain 86:525-562. Carman, J.B., T.P.S. Powell, and W.M. Cowan (1965) A quantitative study of the striatum in some primates. J. Anat. 993950. Carman, J.B. (1966) The volume and neurone content of the striatum in man. Proc. Univ. Otago Med. Sch. 44:49-51. Carpenter, M.B., S.C. Carleton, J.T. Keller, and P. Conte (1981)Connections of the subthalamic nucleus in the monkey. Brain Res. 224r1-29. Chang, H.T., C.J. Wilson, and S.T. Kitai (1981) Single neostriatal efferent axom in the globus pallidus: A light and electron microscopic study. Science213:915-9 18. Chang, H.T. (1988) Substance P-dopamine relationship in the rat substantia nigra: A light and electron microscopy study of double immunocytochemically labeled materials. Brain Res. 4483391-396, Cowan, W.M., D.I. Gottlieb, A.E. Hendrickson, J.L. Price, and T.A. Woolsey (1972)The autoradiographicdemonstration of axonal connections in the central nervous system. Brain Res. 37321-514. Cowan, W.M., and T.P.S. Powell (1966) Strio-pallidal projection in the monkey. J. Neurol. Neurosurg. Psychiat. 293426439, Crutcher, M.D., and M.R. DeLong (1984a) Single cell studies of the primate putamen. I. Functional organization. Exp. Brain Res. 533233-243. Crutcher, M.D., and M.R. DeLong (1984b) Single cell studies of the primate putamen. 11. Relations to direction of movement and pattern of muscular activity. Exp. Brain Res. 53:244-258. DeLong, M.R. (1971) Activity of pallidal neurons during movement. J. Neurophysiol. 34:414427. DeLong, M.R. (1972) Activity of basal ganglia neurons during movement. Brain Res. 403127-135. DeLong,M.R. (1973)Putamen: Activity of single units during slow and rapid eye movements. Science 17931240-1242. DeLong, M.R., M.D. Crutcher, and A.P. Georgopoulos (1985) Primate globus pallidus and subthalamic nucleus: functional organization. J. Neurophysiol. 533530-543. DeLong, M.R., and A.P. Georgopoulos (1981) Motor functions of the basal ganglia. In Handbook of Physiology. The Nervous System. Motor Control. Sec. 1, vol. 2, pt. 2. Bethesda: American Physiological Society, pp. 1017-1061. DeOlmos,J.,H. Hardy, and L. Heimer (1978)The afferent connections of the main and the accessory olfactory bulb formations in the rat: An experimental HRP-study. J. Comp. Neurol. 181.213-244. Desban, M., C. Gauchy, M.L. Kernel, M.J. Besson, and J. Glowinski (1989) Three-dimensional organization of the striosomal compartment and patchy distribution of striatonigral projections in the matrix of the cat caudate nucleus. Neurosci. 293551-566. Deutch, A.Y., M. Goldstein, and R.H. Roth (1986) The ascending projections of the dopaminergic neurons of the substantia nigra, zona reticulata: A combined retrograde tracer-immunohistochemical study. Neurosci. Lett. 713257463, DeVito, J.L., M.E. Anderson, and K.E. Walsh (1980) A horseradish peroxidase study of afferent connections of the globus pallidus in Macaca mulatta. Exp. Brain Res. 38365-73. Difiglia, M., P. Pasik, and T. Pasik (1982) A Golgi and ultrastructural study of the monkey globus pallidus. J. Comp. Neurol. 212:53-75. Divac, I. (1977)Does the neostriatum operate as a functional entity? In A.R. Cools, A.H.M. Lohman, and J.H.L. van den Bercken (eds):Psychobiology of the Striatum. Amsterdam: North-Holland Publishing Company, pp. 21-30. Domesick, V.B. (1977) The topographical organization of the striatonigral connection in the rat. Anat. Rec. 187567 (Abstract). Donoghue, J.P., and M. Herkenham (1986) Neostriatal projections from individual cortical fields conform to histochemically distinct striatal compartments in the rat. Brain Res. 365:397403. Donoghue, J.P., and S.T. Kitai (1981) A collateral pathway to the neostriatum from corticofugal neurons of the rat sensory-motor cortex: An intracellular HRP study. J. Comp. Neurol. 201tl-13. Fallon, J.H., and R.Y. Moore (1978) Catecholamine innervation of the basal forebrain. J. Comp. Neurol. 80:545-580.

594 Fallon, J.H., and B.T.S. Ziegler (1979) The crossed cortico-caudate projection in the rhesus monkey. Neurosci. Lett. 15:29-32. Faull, R.L.M., W.J.H. Nauta, and V.B. Domesick (1986) The visual corticostriato-nigral pathway in the rat. Neuroscience 19:1119-1132. Feger, J., and A.R. Crossman (1984) Identification of different subpopulations of neostriatal neurones projecting to globus pallidus or substantia nigra in the monkey: A retrograde fluorescence double-labelling study. Neurosci. Lett. 49r7-12. Fisher, R.S., C. Shiota, M.S. Levine, C.D. Hull, and N.A. Buchwald (1984) Interhemispheric organization of corticocaudateprojections in the cat: A retrograde double-labeling study. Neurosci. Lett. 48:369-373. Fisher, R.S., M.K. Boylan, C.D. Hull, N.A. Buchwald, andM.S. Levine (1986) Branched projections of cat sensorimotor cortex: multiple retrograde labeling via commissural corticocortical, decussated corticostriatal and undecussated corticostriatal axons. Brain Res. 384:395400. Fox, C.A., and J.A. Rafols (1975) The radial fibers in the globus pallidus. J. Comp. Neurol. 159:177-200. Francois, C., G. Percheron, and J. Yelnik (1984a) Localization of nigrostriatal, nigrothalamic and nigrotectal neurons in ventricular coordinates in macaques. Neurosci. 13551-67. Francois, C., G. Percheron, J. Yelnick, and S. Heyner (1984b) A Golgi analysis of the primate globus pallidus. I. Inconstant processes of large neurons, other neuronal types, and afferent axons. J. Comp. Neurol. 227: 182-199. Francois, C., J. Yelnick, and G. Percheron (1987) Golgi study of the primate substantia nigra. 11. Spatial organization of dendritic arborizations in relation to the cytoarchitectonic boundries and to the striatonigral bundle. J. Comp. Neurol. 265473493. Gerfen, C.R. (1984) The neostriatal mosaic: Compartmentalization of corticostriatal input and striatonigral output systems. Nature 311:461-464. Gerfen, C.R. (1985) The neostriatal mosaic. I. Compartmental organization of projections from the striatum to the substantia nigra in the rat. J. Comp. Neurol. 236r454476. Gerfen, C.R., M. Herkenham, and J. Thibault (1987) The neostriatal mosaic. 11. Patch- and matrix-directed mesostriatal dopaminergic and nondopaminergic systems. J. Neurosci. 7:3915-3934. Gerfen, C.R. (1989) The neostriatal mosaic: Striatal patch-matrix organization is related to cortical lamination. Science 246t385-388. Gilbert, C.D., and T.N. Wiesel (1979) Morphology and intracortical projections of functionally characterised neurones in the cat visual cortex. Nature 280: 120-125. Gimenez-Amaya, J.M., and A.M. Graybiel(1990) Compartmental origins of the striatopallidal projection in the primate. Neuroscience 34r111-126. Goldman, P.S., and W.J.H. Nauta (1977a) Columnar distribution of corticocortical fibers in the frontal association, limbic, and motor cortex of the developing rhesus monkey. Brain Res. 122r393-413. Goldman, P.S.,and W.J.H. Nauta (197713)An intricately patterned prefrontocaudate projection in the rhesus monkey. J. Comp. Neurol. 171r369-386. Goldman-Rakic, P.S. (1982) Cytoarchitectonic heterogeneity of the primate neostriatum. Subdivision into Island and matrix cellular components. J. Comp. Neurol. 205t398-413. Grofova, I. (1979) Extrinsic connections of the neostriatum. In I. Divac and R.G.E. Oberg (eds): The neostriatum. Oxford: Pergamon Press, pp. 37-51. Grofova, I., J.M. Deniau, and S.T. Kitai (1982)Morphology of the substantia nigra pars reticulata projection neurons intracellularly labelled with HRP. J. Comp. Neurol. 208:352-368. Haber, S.N., and R. Elde (1981) Correlation between met-enkephalin and substance P immunoreactivity in the primate globus pallidus. Neuroscience 6:1291-1297. Hedreen, J.C., S. McGrath, and C. Warner (1976) Survival time, aldehyde fixation, and axonal transport of horseradish peroxidase. SOC.Neurosci. Abstr. 238 (Abstract). Hedreen, J.C. (1977) Corticostriatal cells identified by the peroxidase method. Neurosci. Lett. 4:l-7. Hedreen, J.C., and S. McGrath (1977) Observations on labeling of neuronal cell bodies, axons, and terminals after injection of horseradish peroxidase into rat brain. J. Comp. Neurol. 176:225-246. Hendrickson, A.E., L. Moe, and B. Noble (1972) Staining for autoradiography of the central nervous system. Stain Technol. 47:283-290. Jimenez-Castellanos, J., and A.M. Graybiel (1987) Subdivisions of the dopamine-containing AB-Ag-AlO complex identified by their differential mesostriatal innervation of striosomes and extrastriosomal matrix. Neuroscience 23r223-242.

J.C. HEDREEN AND M.R. DELONG Jimenez-Castellanos,J., and A.M. Graybiel(1989) Evidence that histochemically distinct zones of the primate substantia nigra pars compacta are related to patterned distributions of nigrostriatal projection neurons and striatonigral fibers. Exp. Brain Res. 74:227-238. Jinnai, K., and Y. Matsuda (1979) Neurons of the motor cortex projecting commonly of the caudate nucleus and the lower brain stem in the cat. Neurosci. Lett. 13:121-126. Johnson, T.N., and H.E. Rosvold (1971) Topographic projections on the globus pallidus and the substantia nigra of selectively placed lesions in the precommisural caudate nucleus and putamen in the monkey. Exp. Neurol. 33r584-596. Jones, E.G., J.D. Coulter, H. Burton, and R. Porter (1977) Cells of origin and terminal distribution of corticostriatal fibers arising in the sensorymotor cortex of monkeys. J. Comp. Neurol. I73:53-80. Kawai, Y., H. Takagi, Y. Kumoi, S. Shiosaka, and M. Tohyama (1987) Nigrostriatal dopamine neurons receive substance P-ergic inputs in the substantia nigra: Application of the immunoelectron microscopic mirror technique to fluorescent double-staining for transmitter-specific projections. Brain Res. 401:371-376. Kemp, J.M., and T.P.S. Powell (1970) The cortico-striate projection in the monkey. Brain 93:525-546. Kim, R., K. Nakano, A. Jayaraman, and M.B. Carpenter (1976) Projections of the globus pallidus and adjacent structures: An autoradiographic study in the monkey. J. Comp. Neurol. 169r263-290. Kish, S.J., K. Shannak, and 0. Hornykiewicz (1988) Uneven pattern of dopamine loss in the striatum of patients with idiopathic Parkinson’s disease. N. Engl. J. Med. 3182376880. Kita, H., and S.T. Kitai (1987) Efferent projections of the subthalamic nucleus in the rat: light and electron microscopic analysis with the PHA-L method. J. Comp. Neurol. 260:435-452. Kitai, S.T., J.D. Kocsis, and J. Wood (1976) Origin and characteristics of the cortico-caudate d e r e n t s : An anatomical and electrophysiological study. Brain Res. 118:137-141. Knook, H.L. (1965) The Fibre-Connections of the Forebrain. Assen: Van Gorcum. Koliatsos, V.E., L.J. Martin, J. Hedreen, G.E. Alexander, I. Hamada, D.L. Price, and M.R. DeLong (1988) Organization of primate basal ganglia “motor circuit.” 11. Putaminal projections to internal (GPi) and external (GPe) globus pallidus originate in distinct neuronal populations within the matrix compartment. Soc. Neurosci. Abstr. 14:720 (Abstract). Kubozono, T., K. Nakano, A. Todushige, Y. Hasegawa, and M. Kohno (1986) Topographical organization of the projections from the cerebral cortex to the head of the caudate nucleus: A horseradish peroxidase study in the cat. Neuroscience 19:113-123. Kunzle, H. (1975) Bilateral projections from precentral motor cortex to the putamen and other parts of the basal ganglia: An autoradiographic study in macaca fascicularis. Brain Res. 88r195-209. Kunzle, H. (1977a) Projections from the primary somatosensory cortex to basal ganglia and thalamus in the monkey. Exp. Brain Res. 30:481-492. Kunzle, H., and K. Akert (1977b) Efferent connections of cortical area 8 (frontal eye field) in macaca fascicularis: A reinvestigation using the autoradiographic technique. J. Comp. Neurol. 173:147-163. Kunzle, H. (1978) An autoradiographic analysis of the efferent connections from premotor and adjacent prefrontal regions (areas 6 and 9) in macaca fascicularis. Brain Behav. Evol. 15:185-234. Langer, L.F., and A.M. Graybiel (1989) Distinct nigrostriatal projection systems innervate striosomes and matrix in the primate striatum. Brain Res. 498r344-350. Malach, R., and A.M. Graybiel (1986) Mosaic architecture of the somatic sensory-recipient sector of the cat’s striatum. J. Neurosci. 6:3436-3458. Mason, C.A., and J.A. Robson (1979) Morphology of retino-geniculate axons in the cat. Neuroscience 4:79-97. Matelli, M., R. Camarda, M. Glickstein, and G. Rizzolatti (1986) Afferent and efferent projections of the inferior area 6 in the macaque monkey. J. Comp. Neurol. 251:281-298. May, J.G., and R.A. Andersen (1986) Different patterns of corticopontine projections from separate cortical fields within the inferior parietal lobule and dorsal prelunate gyrus of the macaque. Exp. Brain Res. 63t265-278. McGeorge, A.J., and R.L.M. Faull (1989) The organization of the projection from the cerebral cortex to the striatum in the rat. Neuroscience 29503-537. Mesulam, M.-M. (1978) Tetramethyl benzidine for horseradish peroxidase neurohistochemistry: A non-carcinogenic blue reaction-product with

PROJECTIONS IN THE MACAQUE superior sensitivity for visualizing neural d e r e n t s and efferents. J. Histochem. Cytochem.26:106-117. Mesulam,M.-M. (1982)Principles of horseradishperoxidase neurohistochemistry and their applications for tracing neural pathways- axonal transport, enzyme histochemistry and light microscopic analysis. In M.-M. Mesulam (ed): Tracing Neural Connections with Horseradish Peroxidase. NewYork: JohnWileyand Sons, pp. 1-151. R.T. Richardson, F.H. Baker, and M.R. DeLong (1987) The Mitchell, S.J., primate globus pallidus; neuronal activity related to direction of movement. Exp. Brain Res. 68:491-505. Muakkassa, K.F., and P.L. Strick (1979) Frontal lobe inputs to primate motor cortex: Evidence for four somatotopically organized “premotor” areas. Brain Res. 177:176-182. Nauta, H.J.W. (1979) A proposed conceptual reorganization of the basal ganglia and telenchephalon. Neuroscience4: 1875-1881. Nauta, H.J.W., and M. Cole (1978) Efferent projections of the subthalamic nucleus: an autoradiographic study in monkey and cat. J. Comp. Neurol. 18O:l-16. Nauta, W.J.H. and W.R. Mehler (1966) Projections of the lentiform nucleus in the monkey. Brain Res. 1:3-42. Nimi, K., T. Ikeda, S. Kawamura, and H.Inoshita (1970) Efferent projections of the head of the caudate nucleus in the cat. Brain Res. 21:327-343. Oka, H. (1980) Organization of the cortico-caudate projections. Exp. Brain Res. 40.203-208. Oldfield, B.J., and E.M. McLachlan (1977) Uptake and retrograde transport of HRP by axons of intact and damaged peripheral nerve trunks. Neurosci. Lett. 6:135-141. Palay, S.L., and V. Chan-Palay (1974) Cerebellar Cortex, Cytology and Organization, New York Springer-Verlag. Parent, A. (1979) Identification of the pdidal and peripallidal cells projecting to the habenulain monkey. Neurosci. Lett. 15:159-164. Parent, A,, S. Gravel, and R. Boucher (1981) The origin of forebrain aITerents to the habenula in rat, cat and monkey. Brain Res. Bull. 6:23-38. Parent, A. (1983) The subcortical fierents to caudate nucleus and putamen in primate: A flourescence retrograde double labeling study. Neurosci. 20:1137-1150. Parent, A,, C. Bouchard, and Y. Smith (1984) The striatopallidal and striatonigral projections: Two distinct fiber systems in primate. Brain Res. 303:38%390. Percheron, G., J. Yelnick, and C. Francois (1984) A Golgi analysis of the primate globus pallidus. 111. Spatial organization of the striato-pallidal complex. J. Comp. Neurol. 227:214-227. Perkel, D.J., J. Bullier, and H. Kennedy (1986) Topography of the afferent connectivity of area 17 in the macaque monkey: A double labelling study. J.Comp. Neurol. 253:376402. Ricardo, J.A. (1980) Efferent connections of the subthalamic region in the rat. I. The subthalamic nucleus of Luys. Brain Res. 202:257-271. Rinne, J.O., J. Rummukainen, L. Paljarvi, and U.K. Rinne (1989) Dementia in Parkinson’s disease is related to neuronal loss in the medial substantia nigra. Ann. Neurol. 26:47-50. Rosene, D.L. and M.-M. Mesulam (1978) Fixation variables in horseradish peroxidase neurohistochemistry. I. The effects of fixation time and perfusion procedures upon enzyme activity. J. Histochem. Cytochem. 26:2&39. Royce, G.J. (1982) Laminar origin of cortical neurons which project upon the caudate nucleus: A horseradish peroxidase investigation in the cat. J. Comp. Neurol. 205:B-29. Royce, G.J. (1983) Cortical neurons with collateral projections to both the caudate nucleus and the centromedian-parafascicular thalarnic complex: a Rourescent retrograde double labeling study in the cat. Exp. Brain Res. 50:157-165. Royce, G.J., and E.J. Laine (1984) Efferent Connections of the caudate nucleus, including cortical projections of the striatum and other basal ganglia: An autoradiographic and horseradish peroxidase investigation in the cat. J. Comp. Neurol. 226:28-49. Schwartz, M.L., and P.S. Goldman-Rakic (1984) Callosal and intrahernispheric connectivity of the prefrontal association cortex in rhesus monkey: relation between intraparietal and principal sulcal cortex. J. Comp. Neurol. 226:403-420. Selernon, L.D., and P.S. Goldman-Rakic(1985)Longitudinal topography and interdigitation of cortico-striatal projections in the rhesus monkey. J. Neurosci. 5:776-794.

595 Selemon, L.D., and P.S. Goldman-Rakic (1988) Common cortical and subcortical targets of the dorsolateral prefrontal and posterior parietal cortices in the rhesus monkey: Evidence for a distributed neural network subserving spatially guided behavior. J. Neurosci. 8:4049-4068. Smith, Y., L.-N. Hazrati, and A. Parent (1990) Efferent projections of the subthalamic nucleus in the squirrel monkey as studied by the PHA-L anterograde tracing method. J. Comp. Neurol. 294:306-323. Smith, Y., and A. Parent (1986) Differential connections of caudate nucleus and putamen in the squirrel monkey (Saimiri Sciureus). Neurosci. 18:347-371. Somogyi, P., J.P. Bolam, S. Totterdell, and A.D. Smith (1981) Monosynaptic input from the nucleus accumbens- ventral striatum region to retrogradely labelled nigrostriatal neurones. Brain Res. 217345-263. Somogyi, P., J.V. Priestly, A.C. Cuello, A.D. Smith, and J.P. Bolam (1982) Synaptic connections of substance P-immunoreactive nerve terminals in the substantia nigra of the rat. Cell. Tiss. Res. 223:469-486. Sorensen, K.E., and M.P. Witter (1983) Entorhinal efferents reach the caudato-putamen. Neurosci. Lett. 35.259-264. Stanton, G.B., M.E. Goldberg, and C.J. Bruce (1988) Frontal eye field efferents in the macaque monkey. I. Subcortical pathways and topography of striatal and thalamic terminal fields. J. Comp. Neurol. 271:473492. Szabo, J. (1962) Topical distribution of the striatal efferents in the monkey. Exp. Neurol. 5:21-36. Szabo, J. (1967) The efferent projections of the putamen in the monkey. Exp. Neurol. 19:463-476. Szabo, J. (1970) Projections from the body of the caudate nucleus in the rhesus monkey. Exp. Neurol. 27:l-15. Szabo, J. (1980) Organization of the ascending striatal afferents in monkeys. J. Comp. Neurol. 189:307-321. Tanaka, Jr., D. (1987) Differential laminar distribution of corticostriatal neurons in the prefrontal and pericruciate gyri of the dog. J. Neurosci. 7:40954106. Tulloch, I.F., G.W. Arbuthnott, and A.K. Wright (1978) Topographical organization of the striatonigral pathway revealed by anterograde and retrograde neuroanatomical tracing techniques. J. Anat. 127:425441. Van der Kooy, D., and H.G.J.M. Kuypers (1979) Flourescent retrograde double labeling: Axonal branching in the ascending Raphe and nigra projections. Science 204:873-875. Van der Kooy, D. (1979) The organization of the thalamic, nigral, and raphe cells projecting to the medial vs. lateral caudate-putamen in the r a t A flourescent retrograde double labelling study. Brain Res. 169:381-387. Van den Pol, A.N., A.D. Smith, and J.F. Powell (1985) GABA axons in synaptic contact with dopamine neurons in the substantia nigra: Double immunocytochemistry with biotin-peroxidase and protein A-colloidal gold. Brain Res. 348:146-154. von Bonin, G. and P. Bailey (1947) The Neocortex of Macaca Mulatta. Urbana: University of Illinois Press. Voneida, T.J. (1960) An experimental study of the course and destination of fibers arising in the head of the caudate nucleus in the cat and monkey. J. Comp. Neurol. 115:75-87. Wassef, M., A. Berod, and C. Sotelo (1981) Dopaminergic dendrites in the pars reticulata of the rat substantia nigra and their striatal input, combined immunocytochemical localization of tyrosine hydroxylase and anterograde degeneration. Neuroscience 6:2125-2139. Webster, K.E. (1961) Cortico-striate interrelations in the albino rat. J. Anat. 95:532-544. Webster, K.E. (1965) The cortico-striatal projection in the cat. J. Anat. 99:329-337. Wilson, C.J. (1987) Morphology and synaptic connections of crossed corticostriatal neurons in the rat. J. Comp. Neurol. 263:567-580. Wilson, S.A.K. (1914) An experimental research into the anatomy and physiology of the corpus striatum. Brain 36:427-492. Yelnick, J., G. Percheron, and C. Francois (1984) A Golgi analysis of the primate globus pallidus. 11. Quantitative morphology and spatial orientation of dendritic aborizations. J. Comp. Neurol. 227:200-213. Yelnick, J., C. Francois, G. Percheron, and S. Heyner (1987) Golgi study of the primate substantia nigra I. Quantitative morphology and typology of nigral neurons. J. Comp. Neurol. 265:455-472. Yeterian, E.H., and G.W. van Hoesen (1978) Cortico-striate projections in the rhesus monkey: The organization of certain cortico-caudate connections. Brain Res. 139r43-63.

Neuroprotective potential of pleiotrophin overexpression in the striatonigral pathway compared with overexpression in both the striatonigral and nigrostriatal pathways.

D1 and D2 dopamine receptor-regulated gene expression of striatonigral and striatopallidal neurons.

Dual role for Islet-1 in promoting striatonigral and repressing striatopallidal genetic programs to specify striatonigral cell identity.

D1 and D2 dopamine receptors differentially regulate c-fos expression in striatonigral and striatopallidal neurons.

Thalamic projections to sensorimotor cortex in the newborn macaque.

Bihemispheric organization of amygdalo-cortical projections in the rat.

Topographical organization of the striatonigral pathway revealed by anterograde and retrograde neuroanatomical tracing techniques.

Funicular organization of avian brainstem-spinal projections.

Dopamine-Induced Changes in Gαolf Protein Levels in Striatonigral and Striatopallidal Medium Spiny Neurons Underlie the Genesis of l-DOPA-Induced Dyskinesia in Parkinsonian Mice.

The development of cortico-motoneuronal projections investigated using magnetic brain stimulation in the infant macaque.

The organization of feline entopenduncular nucleus projections: anatomical studies.

Complementary Patterns of Direct Amygdala and Hippocampal Projections to the Macaque Prefrontal Cortex.

Central projections of intrinsically photosensitive retinal ganglion cells in the macaque monkey.

An autoradiographic study of cortical projections from motor thalamic nuclei in the macaque monkey.

Thalamic projections to sensorimotor cortex in the macaque monkey: use of multiple retrograde fluorescent tracers.

Topographic segregation of corticostriatal projections from posterior parietal subdivisions in the macaque monkey.

Efferent connections of the striatopallidal and amygdaloid components of the substantia innominata in the cat: projections to the nucleus accumbens and caudate nucleus.

Organization of the zona incerta in the macaque: a Nissl and Golgi study.

Correction: 1-Methyl-4-Phenylpyridinium Stereotactic Infusion Completely and Specifically Ablated the Nigrostriatal Dopaminergic Pathway in Rhesus Macaque.

Organization of medullary adrenergic and noradrenergic projections to the periaqueductal gray matter in the rat.

1-Methyl-4-phenylpyridinium stereotactic infusion completely and specifically ablated the nigrostriatal dopaminergic pathway in rhesus macaque.

Organization of amygdaloid projections to the prefrontal cortex and associated striatum in the rat.

Topographic organization of the basal forebrain projections to the perirhinal, postrhinal, and entorhinal cortex in rats.

Areal and laminar organization of corticocortical projections in the rat somatosensory cortex.