The Radial Fibers in t h e Globus Pallidus1 c. A.

FOX A N D J. A. RAFOLS Department of A n a t o m y , W a y n e State University School of Medicine, .Detroit, Michigan 48201

ABSTRACT In our Golgi collection of adult monkey brains the striatal efferents, i.e., the radial fibers in the globus pallidus and the “comb” bundle fibers in the internal capsule and in the cerebral peduncle, are well impregnated in the horizontally sectioned brain and in a sagittal sectioned brain. Since collaterals emerging from radial fibers are seen only in the horizontal series and not in the sagittal series, the interpretation is that they proceed anteriorly and posteriorly only, following the curvature of the pallidal segments, and do not run superiorly or inferiorly as they emerge. Although radial fibers emitting collaterals in the lateral segment and in the medial segment of the globus pallidus have been observed, it has not been possible to observe the same radial fiber emitting collaterals in both pallidal segments and the prospects of ever doing so are not good. The radial fibers converging in the globus pallidus pursue many radii and there is little coincidence between the plane of section and the planes in which they travel. At most only severed radial fiber segments 100-150 microns in length can be found i n the horizontal sections needed to observe the collaterals. Moreover, sagittal sections show that radial fibers are deflected in their course, either dorsoventrally or ventrodorsally, as they pass through the internal medullary lamina to enter the medial segment of the globus pallidus. The radial fibers in the medial segment of the globus pallidus are continuous with the “comb” bundle fibers and appear to be thinner than the radial fibers in the lateral segment of the globus pallidus. It is not proved; nonetheless, the view expressed here is that the radial fibers are thinner in the medial segment of the globus pallidus because they T a y be the same fibers that gave off collaterals in the lateral segment of the globus pallidus. This is discussed in the light of the electrophysiological disclosure of Yoshida et al. (‘71, ’72) that caudatopallidal fibers are collaterals off caudatonigral fibers. The afferent plexuses of fine, “bouton en passage” fibers, which completely ensheath the long radiating dendrites in the globus pallidus (Fox et al., ’66) are well impregnated in the horizontal series. Obviously, they are formed by a number of ultimate branches converging from the collateral branches of a number of different radial fibers. The divergence, too, in this system must be considerable; however, its true extent can only be surmised from the severed radial fibers and radial fiber collaterals seen in the incompletely impregnated Golgi section. One severed segment of a radial fiber displays three collaterals and one of these collaterals has five branches, one of which can be traced to a point where it gives off an ultimate branch in an afferent, dendrite-ensheathing plexus.

The caudate nucleus and the putamen project on the globus pallidus and the substantia nigra. This has been well established by a series of experimental studies employing various methods: degeneration utilizing silver techniques (Voneida, ’60; Szabo, ’62, ’67, ’70; Nauta and Mehler, ’66; Niimi et al., ’70); histochemical techniques (Olivier et al., ’70); analyzing the J. COMP.

NEUR.,159: 177-200.

gliosis in the globus pallidus and the substantia nigra resulting from lesions in different parts of the caudate nucleus and the putamen (Cowan and Powell, ’66). Prior studies indicated the existence of these connections (Riese, ’24a, ’24b; Rundle and Papez, ’37; Papez, ’41; and others). An ex1

Supported by NIH Grant NS-06925-08 from NINDS.

177

178

C . A . FOX AND J . A. RAFOLS

tensive review of this literature can be found in the paper of Nauta and Mehler (’66). The efferents from the caudate and the putamen run in radially arranged fiber bundles (Wilson’s pencils) which converge in the globus pallidus “like spokes in a wheel” (Papez, ’41) and continue by way of the “comb system” of fibers into the substantia nigra. Wilson (’14), who first called attention to the fine caliber and delicate myelinated sheaths of the fibers coursing from the putamen to the globus pallidus, gave this description: “They are massed into bundles or pencils, arising by the approximation of individual fibers, not always very close to their cells of origin, and running mesially; the anterior pencils converge a s they pass in a posterior direction, while the posterior converge as they travel anteriorly; the most ventral run in a dorsal direction, the dorsal in a ventral direction, and in this fashion they all converge towards the lateral zone of the globus pallidus . . .” Szabo (’62) likened the whole of the striatum to a dome, superimposed on the pallidum, which sends its fibers into a core, the centrally placed pallidum. Obviously there is a high degree of organization in these projections. Discussing this, Cowan and Powell (’66) considered two possibilities and illustrated them with diagrams. One possibility (their fig. 12a) is that the projection is organized topographically so that each segment of the globus pallidus receives afferents from different parts of the striatum. The other possibility (their fig. 12b) is that the striatofugal fibers as they course through the lateral segment of the globus pallidus give off collaterals and continuing on give collaterals in the medial segment of the globus pallidus so that all parts of the striatum project in a n organized manner upon both the lateral and the medial segments of the globus pallidus. Cowan’s and Powell’s (’66) speculation that collaterals from the same striatopallidal fibers may contribute to both segments of the globus pallidus is of special relevance now in the light of the recent electrophysiological studies of Yoshida et al. (’71, ’72) and Yoshida and Precht (’71), indicating that caudatopallidal fibers are collaterals emerging from caudatonigral

axons. These investigators examined in neurons of the cat entopeduncular nucleus, i.e., the homologue of the primate medial segment of the globus pallidus (Fox et al., ’66), responses evoked by stimulation of the substantia nigra, by stimulation of the caudate nucleus and by stimulation of the diencephalic area that includes Forel’s fields H1 and H2 and the subthalamic nucleus. Diencephalic stimulation evoked short latency IPSPs in entopeduncular neurons; substantia nigra stimulation evoked long latency IPSPs in entopeduncular neurons. The latter IPSPs resembled those evoked by caudate stimulation. Also there was strong interaction between caudate and substantia nigra evoked IPSPs, suggesting that both may be generated, at least in part, by activating the same fibers. These investigators reasoned: either the caudate efferents send collaterals to the entopeduncular nucleus a s they course to the substantia nigra or the substantia nigra efferents end in the entopeduncular nucleus as well as the caudate nucleus. Testing these possibilities, they destroyed the caudate nucleus >5 weeks prior to acute experiments and observed: “In these cats, SN stimulation did not evoke positive field potential in ENT or late IPSPs in ENT cells, although short latency IPSPs could still be evoked by diencephalic stimulation. These findings support the existence of terminations in the ENT nucleus of axon collaterals from caudatonigral fibers . . .” (Yoshida et al., ’71). Since there has been, to our knowledge, no anatomical demonstration of collaterals arising from the striatofugal fibers, i.e., the radial fibers, a s they course through the globus pallidus, a search through our collection of Golgi impregnations of the adult monkey (Macaca mulatta) was undertaken seeking evidence of their existence. In some of our preparations the radial fiber bundles are heavily impregnated in the globus pallidus. Fortunate horizontal sections reveal collaterals emerging from the radial fibers and branches from these collaterals giving rise to the long “bouton en passage” fibers, which contribute to the rich afferent plexuses that completely ensheath the long radiating dendrites of the globus pallidus. These formations, which make “longitudinal axo-

THE RADIAL FIBERS IN THE GLOBUS PALLIDUS

dendritic synapses,” were previously demonstrated in the globus pallidus (Fox et al., ’66;Fox et al., ’74).

179

of the globus pallidus and faces the cerebral peduncle. There are clumps of nerve cells and some radial fibers stained in the putamen, a few nerve cells and more of the MATERIALS AND METHODS radial fibers stained in the lateral segment The observations from Golgi material re- of the globus pallidus, but in the medial ported here are based on a study of the ex- segment of the globus pallidus the radial tensive collection of Golgi preparations of fiber bundles (RB) are intensely stained, the adult monkey brains (Macaca mulat- blackening this cone-shaped profile. t a ) available in this laboratory and preIt is not shown here, but when this and pared by the method of Fox et al. (’51). adjacent sagittal sections are examined The material used for electron micros- more closely under the microscope it can copy was taken from the squirrel monkey be seen that the radial fibers, as they pass (Saimiri sciureus) brain and processed as from the lateral to the medial segment of follows: monkeys were anesthetized with the globus pallidus, are deflected in their Nembutal (sodium pentobarbital) and per- course by curving either ventrodorsally or fused through the left ventricle with 300- dorsoventrally in the internal medullary 400 cc of isotonic Ringer’s solution followed lamina. Actually they contribute to the inby a hypertonic solution (750 mOs) of glu- ternal medullary lamina and after entertaraldehyde and paraformaldehyde in sodi- ing the medial segment of the globus pallium cacodylate buffer (pH 7.2-7.6) for 25 dus they resume their radial direction. There is no staining of neuronal or glial minutes. The brains were removed 30 minutes after the perfusion and placed in iso- elements in the internal capsule and in tonic fixative overnight. Then small pieces the cerebral peduncle and consequently of the globus pallidus were post-fixed in the intensely stained transpeduncular fi1% osmium tetroxide. The small blocks bers, the “comb” bundle system (CB), stand were placed in a saturated uranyl acetate out sharply against a clear background. solution overnight and subsequently dehy- Studying this section and the sections imdrated in ethyl alcohol, embedded in Mara- mediately adjacent on either side, it is obglas and sectioned with a LKB or Reichert vious that there is direct continuity beultratome. Ultrathin sections (silver) were tween the radial fibers and the fibers of picked u p on copper grids, stained with the “comb” system, Edinger’s “Kammlead citrate and observed with a RCA system des Fusses.” Edinger (’11) so named these fiber bundles, which interdigitate EMU3G electron microscope. with the fibers of the cerebral peduncle, OBSERVATIONS because they are arrayed like teeth in a comb, “wie sie gleich den Zinken eines T h e radial fiber bundles and their Kammes uberall aus der Zwischenschicht continuity with the “comb” zwischen die Bundle der Fussfaserung bundles hineingreifen.” In one sagittal series the radial fiber Partial impregnation of the “comb” bunbundles (RB) are very heavily impreg- dle system (CB) can be seen in the low nated. The low power photomicrograph (fig. power photomicrograph (fig. 2) which shows 1) shows an area from one of these sections the continuity of these fibers with the radial that includes part of the putamen (Put), fiber bundles (RB) in a horizontal section the anterior commissure (AC), the lateral cutting through the medial segment of the segment of the globus pallidus (LGP), the globus pallidus (MGP), the optic tract (OT), medial segment of the globus pallidus the lateral geniculate body (LGB), the (MGP), the optic tract (OT), the internal cerebral peduncle (CP), the nucleus subcapsule (IC), the cerebral peduncle (CP) thalamicus (Sth) and the substantia nigra and the thalamus (Th). Here the anterior (SN). In the section immediately adjacent commissure abuts the putamen and lies in to this section, which must have been neara groove in the cephalic aspect of the lat- er the surface of the block during impregeral segment of the globus pallidus. The nation, the “comb system” appeared to be optic tract is beneath the medial segment completely stained but the other nearby

180

C . A . FOX AND J . A. RAFOLS

areas, particularly in the region of the substantia nigra and the nucleus subthalamicus, are too blackened to be of any use for illustrative purposes. In this section (fig. 2) there are nerve cells, nerve fibers and blood vessels stained in the nucleus subthalamicus; also there are nerve fibers, longitudinal axodendritic plexuses ensheathing dendrites and a few nerve cells impregnated in the substantia nigra (SN). By way of contrast a different type of silver chromate reaction is shown in the low power photomicrograph (fig. 3 ) , a frontal section through the lenticular nucleus. Here the impregnation consists almost exclusively of nerve cell bodies and their dendrites. The island cortex, the claustrum (CL), the putamen (Put), the lateral segment of the globus pallidus (LGP) and the medial segment of the globus pallidus (MGP) are intensely stained. There is no staining in the internal capsule (IC). In the thalamus (Th) there is mostly precipitate. In regions where the staining of neurons is nearly complete the locations of the radial fiber bundle stand out as negative images since they are unstained. Some of these locations are marked with arrows in the putamen and in the lateral segment of the globus pallidus (fig. 3). In the globus pallidus the neurons have long radiating dendrites and the cell bodies are rather far apart so that in heavy impregnations this structure has a reticular appearance. Rectangular areas of the medial segment and of the lateral segment of the globus pallidus, respectively, taken from the above section (fig. 3 ) , are shown at higher magnifications in figures 4 and 5. Here again arrows mark the positions of the unstained radial bundles. The radial bundles, converging in the pallidum, pursue a number of radii; consequently the spaces they occupy, for the most part, appear in the medial segment (fig. 4) as round or oval “windows.” However, it should be noted that in regions of the lateral part of the lateral segment (fig. 5) these spaces have a “tunnel-Like’’ appearance. The plexuses o f f z n e “bouton en passage” fzbers r u n n i n g longitudinally on and ensheathing dendrites One horizontally sectioned series was invaluable for this study. The plane of sec-

tion is ideal for following the radial fibers and their collaterals and the silver chromate reaction is intense and varied. This is the series in which the long, sleeve-like plexuses of fine, longitudinally running, afferent fibers that fit snugly about the long radiating dendrites of the globus pallidus were first seen (Fox et al., ’66). The impregnation here is rare and most fortunate. In none of our other preparations are these formations that characterize the globus pallidus so richly stained. The low power view (fig. 6),taken from one of these horizontal sections, cuts through the lateral (LGP) and the medial (MGP) segments of the globus pallidus. It is used here for orientation: anterior is to the left and lateral is at the top. Incidentally, in this and in adjacent horizontal sections there is little or no impregnation in the regions of the external and internal medullary laminae. The area (arrow, 8) outlined in the outer portion of the lateral segment of the globus pallidus is the area displayed in the photomicrograph (fig. s), while a n area below the tip of arrow (10) in the inner portion of the medial segment of the globus pallidus is the area displayed in figure 10. These two photomicrographs (figs. 8, lo), from the opposite extremes of the lateral and medial segments of the globus pallidus, respectively, are arranged so that the radial fibers (RF), as they pursue their course, run vertically from the top to the bottom of the figures. The plane in which the radial fibers travel and the plane of section do not coincide; consequently individual radial fibers, when followed under the microscope, can, at most, usually be traced for only 100 to 150 microns. Comparing the radial fibers, one gets the impression that their calibers are thinner in the medial segment (fig. 10) than they are in the lateral segment (fig. 8). Arching across these microscopic fields, mostly at right angles to the radial fibers, are some of the densely impregnated, finefiber plexuses that ensheath the long radiating dendrites of the globus pallidus. These formations (open block arrows, fig. 8,lO) consist of a large number of extremely fine fibers bearing “bouton en passage” endings which form longitudinal axodendritic synapses on the long dendrites. But at this magnification it is impossible to re-

THE RADIAL FIBERS IN THE GLOBUS PALLIDUS

solve the individual fibers that make up these complexes; the fibers are too thin and there are too many of them impregnated en masse. However, the constituents of these unique afferent complexes have been demonstrated elsewhere in a variety of Golgi and electron microscopic preparations (Fox et al., '66; Fox et al., '74). The montage (fig. 9, taken with the x 53 oil immersion lens, shows a single nerve fiber entering a dendrite ensheathing plexus and at the same time it gives a slight inkling of the extensive divergence and convergence in the radial fiber system. The block arrow points to a nerve fiber, out of focus, that undoubtedly is a branch of a collateral off a radial fiber. It bifurcates. The resulting branch that ascends in the photomicrograph also bifurcates. These latter branches, which are lost in the section, must be heading for afferent fiber complexes that encase dendrites. The branch that descends joins the afferent fiber complex (open block mow). It does not show in the photomicrograph, but when this fiber is in focus under the microscope it can be seen dividing and sending a branch in one direction and a branch in the opposite direction in the dendrite-ensheathing plexus. Here, as in most of these impregnations, the dendrite within the formation is not impregnated. A fortunate mid-longitudinal section through one of these ensheathing plexuses is shown in the photomicrograph (open block arrow, fig. 7). The dendrite that it encases is not impregnated. Radial fiber collaterals Collaterals of radial fibers are demonstrated in photomicrographs of the lateral (figs. 11, 12) and of the medial (figs. 1316) segments of the globus pallidus. Taken with the X 53 oil immersion lens in various regions of the section (fig. 6), they are reproduced at a greater magnification than the photomicrographs (figs. 7, 9) and they are mounted so that up is laterad, down is mesad, left is anteriad and right is posteriad. Arrows indicate the direction of the radial fibers (RF) and their collaterals (C). The collaterals are numbered in order to distinguish them when two or more are impregnated on the same radial fiber. In the montage (fig. l l ) , obtained from the lateral segment of the globus pallidus, two radial fibers, paralleling each other

181

closely, extend from the top to the bottom of the figure. The one on the right displays three collaterals ( C l , C2 and C3). The first (Cl) goes anteriorly 42.5 microns and is severed in section. At a point 16.5 microns from its origin this collateral emits an anteromedially directed branch which proceeds for 20.5 microns before it is lost in the section. The second collateral (C2) is more extensive. It arises directly opposite the first collateral and moves posteriorly. In its course it gives off two branches and finally trifurcates. The first of these branches comes off the collateral at a point 18 microns away from the collateral's origin. It is a relatively thick branch that runs medially for a few microns, then curves sharply and is cut off in section. At a point 22 microns from its origin the collateral emits a branch, barely visible in the photomicrograph, that can be followed for a short distance in a posterolateral direction. The trifurcation takes place at a point 45 microns from the collateral's origin. The most lateral of these resulting branches soon disappears from view. The intermediate branch joins up with, and contributes to, the ensheathing plexuses (open block arrow) which is obliquely situated and out of focus in the photomicrograph. However, under the microscope the structure of this formation is easily recognized. The third branch of the trifurcation, the most medial, recurves and runs anteriorly for 13 microns and is then cut off in section. The third collateral (C3) off the radial fiber (RF) arises 33 microns mesad from the region where collaterals C1 and C2 emerge. It can be followed for 31 microns in an anteromedial direction. The radial fiber continues in a medial direction. Another radial fiber (RF) running for some distance in the lateral segment of the globus pallidus can be seen on the right side of figure 12. It gives rise to a collateral (C) that courses anteriorly. At a point 17 microns from its origin it emits a branch (not visible in the photomicrograph, but marked by arrows) that passes medialwards for 17 microns before disappearing in the section. The collateral ( C ) then curves laterally and shortly thereafter branches Tshaped. The resulting branches proceed anteriorly and posteriorly for 13 and 28 microns, respectively, before they are severed in section. The posterior branch cross-

182

C. A . FOX AND J. A. RAFOLS

es in front of the radial fiber from which it arose. The two fibers labeled X (fig. 12) are collaterals but they are not collaterals of the radial fibers seen in this photomicrograph. In the medial segment of the globus pallidus (figs. 13-16) it has not been possible to capture, within the plane of focus of the photomicrographs taken with the X 53 oil immersion lens, segments of radial fibers (RF) displaying the lengthwise extents that they do in the lateral segment (figs. 11, 12). This severing of the radial fibers into shorter segments here is due undoubtedly to the plane of section in this horizontal series relative to the greater angle of inclination the radial fibers have as they funnel into the medial segment of the globus pallidus (fig. l). Several radial fibers (RF) can be seen (fig. 13) in the medial segment of the globus pallidus; one gives off a collateral (C) that can be followed in an anterolateral direction for 23 microns. In the photomicrograph (fig. 14) there is an anteriorly running, thin collateral (C) emerging from a short, swollen segment of a severed radial fiber (RF). In the photomicrograph (fig. 16) there are four radial fibers displaying collaterals: (1) in the upper left there is a collateral (C) that runs posteriorly; (2) in the upper right there are two collaterals (C1 and C2) proceeding in opposite directions from a short segment of a radial fiber (RF); (3) similarly, in the lower right there are two collaterals (C1 and C2) coming off the opposite sides of a short segment of a radial fiber and proceeding in opposite directions; (4) below this there is a short segment of a radial fiber (RF) sprouting a collateral (C) running posteriorly. The rectangular area (15) outlined in the photomicrograph (fig. 10) is the area shown at a greater magnification (fig. 15). It contains an impregnated region of neuropil (NP) located between bundles of radial fibers. Only a few of the radial fibers in these bundles are impregnated. That such a n arrangement exists in the globus pallidus is clear from the electron micrograph (fig. 17). In the upper right hand corner of the photomicrograph (fig. 15) there is a radial fiber (RF) that has a collateral (C) directed towards the neuropil (NP). Lower in the photomicrograph there is a radial fiber (RF) sending a collateral (Cl) into the neuropil

and another collateral (C2) that goes mesad along the edge of the neuropil. The closed block arrow marks the point where the radial fiber is cut off in the section. Radialfibers in a low power electron micrograph The low power electron micrograph (fig. 17), taken from the lateral portion of the medial segment of the squirrel monkey's globus pallidus, cuts obliquely through bundles of radial fibers (RF). It is not surprising that accumulations of these myelinated fascicles are frequently found in globus pallidus sections prepared for electron microscopy, for the pale cast of the globus pallidus in the fresh conditionhence, its name - is due to its rich myelin content. Squeezed in between the radial fiber bundles is a small island of neuropil (NP) - compare with figure 15 - within which can be recognized the profiles of an oligodendrocyte (0),nerve cell bodies (NC), some dendrites (D) and a few myelinated fibers. The synaptic endings and the masses of fine unmyelinated fibers in the neuropi1 are poorly resolved at this magnification. In our opinion the large myelinated fiber (L) in the upper right hand corner of the electron micrograph is a pallidofugal fiber. DISCUSSION

The radial fibers are difficult to impregnate. In our extensive Golgi collection of adult monkey brains they were seen only in one sagittal series and .in one horizontal series, and judging from their caliber and disposition there can be no doubt that they are the same myelinated fibers Wilson ('14), Verhaart ('50) and others observed. We know from our own experience myelinated fibers, albeit infrequently, can be impregnated: e.g., myelinated Purkinje cell axons and their collaterals (Fox et al., '67); the myelinated axons of the large aspiny neurons in the striatum (Fox et al., '71/72); and myelinated fibers in the striatum related to oligodendroglia (Fox and Rafols, '7 1/72). It is interesting that the collaterals emerging from the radial fibers are seen only in the horizontal series and not in the sagittal series. Our interpretation is that the latter plane is not favorable for their revelation; that collaterals which come off

THE RADIAL FIBERS I N THE GLOBUS PALLIDUS

the radial fibers, more or less at right angles, initially proceed anteriorly and posteriorly only, following the curvature of the pallidal segments, and do not run superiorly and inferiorly as they emerge. The different radial fibers pursue many different radii a s they funnel into the globus pallidus and their emerging collaterals appear to be situated in horizontal planes intersecting these radii. The collaterals branch several or more times (C2, fig. 11) and it is these branches, or branches of the collaterals, that eventually give rise to the long “bouton en passage” fibers which individually contribute to the dense ensheathing plexuses that parallel and completely encase the long radiating dendrites (open block arrows, figs. 7, 8, 9, 10). The synaptic endings on the “bouton en passage” fibers as seen in Golgi preparations are separated from each other by intervals of 5 to 20 microns, yet electron micrographs show the dendrites covered by a continuous mosaic of synaptic endings (Fox et al., ’66; Fox et al., ’74). With synaptic endings on individual afferent fibers separated by such intervals, i t is clear that everywhere in the mosaics of synaptic endings covering dendrites and cell bodies contiguous endings arise from different afferent fibers. Considering this and the large number of fine fibers in the ensheathing plexuses, it is obvious that there must be a considerable convergence of branches from different radial fibers on individual neurons in the globus pallidus. In a n electron micrograph (fig. 11, Fox et al., ’66) of a transverse section of one of these formations a count was made of 270 profiles of the fine, densely packed fibers immediately surrounding the single layer of synaptic endings covering a dendrite. In another instance 292 profiles of fine fibers in this same relationship were counted (fig. 17, Fox et al., ’74). These counts give some indication of the convergence here but they are not an accurate index since it has been shown in Golgi preparations that these fine fibers, which run longitudinally for long distances on the dendrites, branch. This branching is especially noticeable in regions where the dendrites branch (fig. 9, Fox et al., ’66) and, interestingly, it sometimes occurs at the synaptic varicosities (fig. 14, Fox et al., ’66).

183

Just as a number of fine branches coming from a number of radial fibers converge to form each dendrite ensheathing plexus, each radial fiber, in turn, distributes, by way of its collaterals and their branches, a number of ultimate branches to a number of dendrite-ensheathing plexuses. This divergence must be considerable. Its true extent, however, can only be surmised from the severed radial fibers and severed radial fiber collaterals seen in both segments of the globus pallidus in the incompletely impregnated Golgi sections. Consider, for example, the incomplete information supplied by the radial fiber segment (fig 11) emitting three collaterals in the lateral segment of the globus pallidus. Collateral C2 has five branches, arising at different points and heading in different directions. Only one of these can be traced to a point where it gives rise to an ultimate branch participating in an ensheathing plexus which, fortunately, is impregnated in this section. The terminal distribution from the four remaining branches they may generate are not seen in this section. Likewise the other two collaterals, coming from this segment of the radial fiber, do not reveal their terminal distribution: collateral C 1 shows one branch and collateral. C3 shows no branching. But considering their caliber and that immediately prior to the distribution of ultimate branches the branching consists of very thin caliber fibers (fig. g), there is, most likely, further branching of these collaterals. Moreover, the emergence of these three collaterals appears not to have diminished the caliber of the parent radial fiber. In its course medialward it certainly must give rise to further collaterals and it seems safe to say that the revelation in fig. 11 is only a partial indication of this axon’s total divergence. Radial fibers giving off collaterals in the lateral segment (figs. 11, 12) and in the medial segment (figs. 13-16) of the globus pallidus have been observed but it has not been possible to observe the same radial fiber giving off collaterals in both segments of the globus pallidus. The distance such a fiber would have to be followed is just too great, and demonstrating this at any time in any Golgi preparation seems to us impossible. In the monkey, for example, where the segments of the globus pallidus

184

THE RADIAL FIBERS IN THE GLOBUS PALLIDUS

are adjacent to each other, a radial fiber’s transit through both segments of the globus pallidus covers a span of 6 to 7 millimeters. To observe adequately a radial fiber in both segments of the globus pallidus it would perhaps have to be followed for several millimeters or more. Yet in our preparations the severed segments of radial fiber extend at most for 100 to 150 microns in the lateral segment of the globus pallidus and their extent is even less in the medial segment of the globus pallidus, where the coincidence of the plane of section and the plane in which the radial fibers travel is less. In the cat it might be more difficult to follow the radial fibers adequately, for here the lateral and the medial (i.e., entopeduncular nucleus) segments of the globus pallidus are separated by an interval of several millimeters. Fiber pathways do not follow the conventional frontal, sagittal and horizontal planes of brain sections. The radial fibers, a s already noted, pursue a number of radii as they converge in the globus pallidus. But even if the radial fibers and their collaterals were well impregnated and if perchance the plane of section were tilted at some angle to the horizontal plane so that a few sections were obtained in which the coincidence of the plane of section and the plane of the coursing radial fibers were more favorable than it is in our best sections, there is still another difficulty. Ranson et al. (’41) have called attention to the deflections in the course of the radial fibers: “In passing through the external and internal laminas they are deflected slightly from their course, so that they do not appear to run directly through, but they do not run for any considerable distances in these laminas.” In the sagittal Golgi series we have noted this deflection of the radial fibers in the internal lamina and it is obvious that this deflection would make it difficult to trace fibers from the lateral to the medial segment of the globus pallidus in the horizontal sections needed for disclosure of the radial fiber collaterals. From examining a large number of radial fibers we have the clear impression that their calibers are smaller in the medial segment of the globus pallidus (fig. 10) than they are in the lateral segment of the globus pallidus (fig. 8). Undoubtedly, thin-

ning of axis cylinders and reductions of myelin sheaths are factors coritributing to the heavier impregnation of the radial fibers in the medial segment of the globus pallidus and of the “comb” bundle fibers seen in the sagittal Golgi series (fig. 1). It is not proved in the present study; nonetheless, it is our view that the radial fibers giving off collaterals in the medial segment of the globus pallidus may be thinner because they may be the same fibers that gave off collaterals in the lateral segment of the globus pallidus. Recalling Cowan’s and Powell’s (’66) suggestion that all parts of the striatum may project in an organized manner upon both the lateral and the medial segments of the globus pallidus and the disclosure of Yoshida et al. (’71, ’72) that caudatopallidal fibers are collaterals off caudatonigral fibers, it is conceivable that a radial fiber such as the one (fig. 11) giving off three collaterals in the lateral segment of the globus pallidus may distribute eventually in the substantia nigra and enroute give collaterals to the medial segment of the globus pallidus. This is discussed further by Fox et al. (’75) in the contribution dealing with computer measurements of striatal efferent axis cylinders at two points in their course: (1) as radial fibers before entering the globus pallidus; (2) as “comb” bundle fibers before entering the substantia nigra. The long thin varicose fibers that are the ultimate branches from the collaterals of the radial fibers and that make up the dendrite-ensheathing plexuses form “longitudinal axodendritic” connections, a type of synapse that permits a single fiber to make many contacts on a single dendrite. Previously these fibers and the climbing fibers were compared in Golgi preparations (Fox et al., ’66) because Ramon y Cajal (’34, ’54) used the climbing fiber to illustrate the “longitudinal axodendritic” type of synapse in his classification of synaptic patterns. It was shown then that the climbing fibers have stems approximately 2 microns in diameter and varicose branches that are slightly thicker than the fine fibers in the globus pallidus. The latter are 0.2 of a micron or less in diameter and their synaptic varicosities are smaller and farther apart than the varicosities on the climbing fibers. Since the dendrites are longer in the globus pallidus and their af-

185

C. A. FOX AND J. A. RAFOLS

ferents thinner, it was concluded that impulses here must spread more slowly and over longer intervals than they do on the smooth branches of the Purkinje cells. Meanwhile, more detailed ultrastructural information concerning these two “longitudinal axodendritic” synapses has come to light and it now appears that the only similarity they have is their geometric pattern. Contrasting them and pointing out their differences may be useful since it is now widely recognized that “. . . . the climbing fiber is the most powerful and specific excitatory synapse yet discovered in the central nervous system” (Eccles et al., ’66), while the latest evidence indicates that activation of the striatofugal fibers by stimulating the caudate nucleus produces IPSPs monosynaptically in the globus pallidus and in the substantia nigra (Yoshida and Precht, ’71; Yoshida et al., ’71, ’72). Within the last few years it has been learned that the climbing fiber endings contact short stubby spines arising from the smooth branches of the Purkinje cells. These synapses, correctly identified first in electron micrographs by Larramendi and Victor (’67), who pursued the problem embryologically, have been verified in a number of species (Kornguth et al., ’68; Fox et al., ’68; Hillman, ’69; Kaiserman-Abramof and Palay, ’69; Larramendi, ’69; Llinas and Hillman, ’69; Mugnaini, ’69; Sotelo, ’69; Uchizono, ’69; Palay and Palay, ’70). The short stubby spines, the critical structures for identifying the climbing fiber synapses, had been completely overlooked in prior studies and the first photomicrographs of them in Golgi preparations were published only recently (frog: Sotelo, ’69; cat: Mugnaini, ’70). The synaptic endings of the climbing fibers are invaginated by clusters of these short spines. Palay and Palay (’70) have shown a s many as six spine profiles in a single profile of a synaptic ending and have given a detailed, interesting discussion of how “. . . the climbing fiber incorporates in a single fiber a large number of morphological devices that could enhance the effectiveness of a synapse.” The endings of the climbing fibers and the endings of the striatopallidal fibers are structurally different. The former are larger, terminate with asymmetrical contacts on the short stubby spines, and contain round synaptic vesicles. The latter are

smaller; terminate with symmetrical contacts on cell bodies and dendrites (Kemp, ’70; Kemp and Powell, ’71); have dense projections on their presynaptic membranes; contain elongated, egg-shaped synaptic vesicles and some ellipsoid synaptic vesicles (Fox et al., ’74). This same type of ending with the same type of vesicles is found in the substantia nigra and Rinvick and Grofova (’70) describe these vesicles as pleomorphic. In the cat globus pallidus endings have been found in some instances on dendritic spines (Adinolfi, ’69; Kemp and Powell, ’71). In the adult monkey we have found occasional spines on the dendrites of the large neurons in Golgi preparations but we have not encountered them in electron micrographs (Fox et al., ’74). LITERATURE CITED Adinolfi, A. M. 1969 The fine structure of neurons and synapses in the entopeduncular n u c l e u s of the cat. J. Comp. Neur., 136: 225-248. Cowan, W. M., and T. P. S . Powell 1966 Striopallidal projection in the monkey. J. Neurol. Neurosurg. Psychiat., 29: 426-439. Eccles, J. C., R. Llinas and K . Sasaki 1966 T h e excitatory synaptic action of climbing fibres on the Purkinje cells of the cerebellum. J . Physiol. (London), 182 : 208-296. Edinger, L. 1911 Vorlesungen uber den Bau der nervosen Zentralorgane. Eighth ed. Vol. 1. Vogel, Leipzig. Fox, C. A , , A. N. Andrade, I . J. LuQui and J . A. Rafols 1974 The primate globus pallidus: A Golgi and electron microscopic study. J. Hirnforsch., 15: 75-93. Fox, C. A., A . Andrade and R. C. Schwyn 1969 Climbing fiber branching in the granular layer. In: Neurobiology of Cerebellar Evolution and Development. R. Llinas, ed. AMA-ERF Institute for Biomedical Research, Chicago, pp. 605-611. Fox, C. A , , A. N. Andrade, R. C. Schwyn and J . A. Rafols 1971172 The aspiny neurons and the glia i n the primate striatum: A Golgi and electron microscopic study. J. Hirnforsch., 1 3 : 341362. Fox, C. A,, D. E. Hillman, K. A. Siegesmund and C. R. Dutta 1967 The primate cerebellar cortex: A Golgi and electron microscopic study. I n : The Cerebellum. Progess in Brain Research. C. A. Fox and R. S. Snider, eds. Elsevier, AmsterdamLondon-New York, 25: 174-225. Fox, C. A , , D. E. Hillman, K. A. Siegesmund and L. A. Sether 1966 The primate globus pallidus and its feline and avian homologues: A Golgi and electron microscopic study. I n : Evolution of the Forebrain, Phylogenesis and Ontogenesis of the Forebrain. R. Hassler and H. Stephan, eds. Georg Thieme Verlag, Stuttgart, pp. 237-248. Fox, C. A,, and J. A. Rafols 1971172 Observations on the oligodendroglia in the primate striatum. Are thev Ram6n v Caial’s “dwarf” or “neurogliaform” neurons? J. Hirnforsch., 1 3 : 331-340. .

I

186

C. A. FOX AND J. A. RAFOLS

Fox, C. A , , J. A. Rafols and W. M. Cowan 1975 Computer measurements of axis cylinder diameters of radial fibers and “comb” bundle fibers. J. Comp. Neur., 159: 201-224. Fox, C. A,, M. Ubeda-Purkiss, H. D. Ihrig and D. Biagioli 1951 Zinc chromate modification of the Golgi technique. Stain Tech., 26: 109-114. Hillman, D. E. 1969 Neuronal organization of the cerebellar cortex in amphibia and reptilia. In: Neurobiology of Cerebellar Evolution and Development. R. Llinas, ed. AMA-ERF Institute for Biomedical Research, Chicago, pp. 279-324. Kaiserman-Abramof, I. R., and S. L. Palay 1969 Fine structural studies of the cerebellar cortex in a mormyrid fish. In: Neurobiology of Cerebellar Evolution and Development. R. Llinas, ed. AMA-ERF Institute for Biomedical Research, Chicago, pp. 171-204. Kemp, J. M. 1970 The termination of strio-pallidal strio-nigral fibres. Brain Research, 17: 125128. Kemp, J. M., and T. P. S. Powell 1971 T h e site of termination of afferent fibres in the caudate nucleus. Phil. Trans. R . SOC. (London), B . 262: 413427. Kornguth, S. E., J. W. Anderson and G. Scott 1968 The development of synaptic contacts in the cerebellum of Mnccicn mulnttti. J. Comp. Neur., 132: 531-545. Larramendi, L. M. H., and T. Victor 1967 Synapses on the Purkinje cell spines in the mouse. An electronmicroscopic study. Brain Research, 5 : 15-30. Larramendi, L. M. H. 1969 Analysis of synaptogenesis in the cerebellum of the mouse. In: Neurobiology of Cerebellar Evolution and D e velopment. R. Llinas, ed. AMA-ERF Institute for Biomedical Research, Chicago, pp. 803-843. Llinas, R., and D. E. Hillman 1969 Physiological and morphological organization of the cerebellar circuits in various vertebrates. In: Neurobiology of Cerebellar Evolution and Development. R. Llinas, ed. AMA-ERF Institute for Biomedical Research, Chicago, pp. 43-73. Mugnaini, E. 1969 Ultrastructural studies on the cerebellar histogenesis. 11. Maturation of nerve cell populations and establishment of synaptic connections in the cerebellar cortex of the chick. In: Neurobiology of Cerebellar Evolution and Development. R. Llinas, ed. AMA-ERF Institute for Biomedical Research, Chicago, 749-782. 1970 Neurones as synaptic targets. I n : Excitatory Synaptic Mechanisms. P. Andersen and J. K. S. Jansen, Jr., eds. Universitets Forlaget, Oslo, pp. 149-169. Nauta, W. J. H., and W. R. Mehler 1966 Projections of the lentiform nucleus in the monkey. Brain Research, 1 : 3-42. Niimi, K., T. Ikeda, S . Kawamura and H. Inoshita 1970 Efferent projections of the head of caudate in the cat. Brain Research, 21 : 327-343. Olivier, A., A. Parent, H. Simard and L. J . Poirier 1970 Cholinesterasic striatopallidal and striatonigral efferents in t h e c a t and the monkey. Brain Research, 1 8 : 273-282. Palay, V. C., and S. L. Palay 1970 Interrelations of basket cell axons and climbing fibers in the cerebellar cortex of the rat. 2. Anat. Entwickl: Gesch., 132: 191-227. Papez, J. W. 1941 A summary of fiber connec-

tions of the basal ganglia with each other and with other portions of the brain. Res. Publ. Ass. nerv. ment. Dis., 21: 2 1 4 8 . Ramon y Cajal, S. 1934 Les preuves objectives d e l’unite anatomique des cellules nerveuses. Trab. lab. invest. biol. (Madrid), 29: 1-137. 1954 Neuron Theory or Reticular Theory? Objective Evidence of the Anatomical Unity of Nerve Cells. Trans. by M. Ubeda-Purkiss and C. A. Fox. Consejo Superior d e Investigaciones Cientificas. Instituto Ramon y Cajal, Madrid. Ranson, S. W., S. W. Ranson, Jr. and M. Ranson 1941 Fiber connections of the corpus striatum as seen in Marchi preparations. Arch. Neurol. Psychiat., 46: 230-249. Riese, W. 1924a Zur vergleichenden Anatomie der striofugalen Faserung. Anat. Anz., 57: 487494. 1924b Beitrage zur Faseranatomie der Stammganglien. J. Psychol. Neurol., 31 : 81-122. Rinvik, , and I. Grofova 1970 Observations on the fife structure of the substantia nigra in the cat. Exp. Brain Res., 1 1 : 229-248. Rundles, R. W., and J. W . Papez 1937 Connections between the striatum and the substantia nigra in a human brain. Arch. Neurol. Psychiat. (Chicago), 38: 550-563. Schwyn, R. C., and C. A . Fox 1974 The primate substantia nigra: A Golgi and electron microscopic study. J. Hirnforsch., 15: 95-126. Sotelo, C. 1969 Ultrastructural aspects of the cerebellar cortex of the frog. In: Neurobiology of Cerebellar Evolution and Development. R . Llinas, ed. AMA-ERF Institute for Biomedical Research, Chicago, pp. 327-367. Szabo, J . 1962 Topical distribution of the striatal efferents in the monkey. Exptl. Neurol., 5: 2136. 1967 The efferent projections of the putamen in the monkey. Exptl. Neurol., 1 9 : 463476. 1970 Projections from the body of the caudate nucleus in the rhesus monkey. Exptl. Neurol., 27: 1-15. Uchizono, K. 1969 Synaptic organization of the mammalian cerebellum. In: Neurobiology of Cerebellar Evolution and Development. R. Llinas, ed. AMA-REF Institute for Biomedical Research, Chicago, pp. 549-581. Verhaart, W. J. C. 1950 Fiber analysis of the basal ganglia. J. Comp. Neur., 93: 4 2 5 4 4 0 . 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. Neur., 115: 75-87. Wilson, S. A. K . 1914 An experimental research into the anatomy and physiology of the corpus striatum. Brain, 36: 4 2 7 4 9 2 . Yoshida, M., and W. Precht 1971 Monosynaptic inhibition of neurons of the substantia nigra by caudato-nigral fibers. Brain Research, 32 : 225228. Yoshida, M., A. Rabin and M. Anderson 1971 Two types of monosynaptic inhibition of pallidal neurons produced by stimulation of the diencephalon and substantia nigra. Brain Research, 30: 235-239. 1972 Monosynaptic inhibition of pallidal neurons by axon collaterals of caudato-nigral fibers. Exp. Brain Res., 1 5 : 333-347.

PLATES

Abhreuiirtioiis

AC, anterior commissure ANT, anterior BV, blood vessel C, C1, C2, C3, collaterals from radial fibers CB, “comb” bundle fibers CL, claustrum CP, cerebral peduncle D, dendrite IC, internal capsule L, large nerve fiber, pallidofugal fiber LAT. lateral LGB, lateral geniculate body LGP, lateral segment of globus pallidus MED, medial

MGP. medial segment of globus pallidus NC, nerve cell body NP, neuropil 0 , oligodendrocyte OT, optic tract POST, posterior Put, putamen RB, radial bundle fibers RF, radial fiber SN, substantia nigra Sth, subthalamic nucleus Th, thalamus X , collaterals from radial fibers not in the section

PLATE 1 E X P L A N A T I O N OF F I G U R E S

1

A sagittal section cutting through the putamen (Put), anterior commissure (AC), lateral (LGP) and medial (MGP) segments of the globus pallidus, optic tract (OT), cerebral peduncle (CP), “comb” bundle system of fibers (CB), internal capsule (IC) and thalamus (Th). The radial fiber bundles (RB) are intensely impregnated, especially in the medial segment of the globus pallidus and they are continuous with the “comb” bundle fibers. Golgi preparation: Mitccrcc~mulnttci.

2

A horizontal section cutting through the medial segment of the globus

pallidus (MGP), the cerebral peduncle (CP), optic tract (OT), lateral geniculate body (LGB), subthalamic nucleus (Sth) and the substantia nigra (SN). Note the continuity of the radial bundle fibers with the “comb” bundle fibers (CB). Golgi preparation: Mrtcnccr mitlattn.

188

T H E RADIAL FIBERS I N T H E GLOBUS PALLIDUS C . A. Fox and J . A. R a f o l s

PLATE 1

189

PLATE 2 EXPLANATION OF FIGURES

190

3

A frontal section through the claustrum (CL), putamen (Put), lateral (LGP) a n d medial (MGP) segments of t h e globus pallidus, internal capsule (IC) a n d the thalamus (Th). T h e impregnation of nerve cells and their dendrites is nearly complete in the globus pallidus a n d in parts of the putamen. T h e almost complete blackening in t h e latter regions of the putamen and the reticular appearance of the globus pallidus reflect the greater cell density i n the putamen. T h e loci of some of t h e unstained radial fiber bundles stand out clearly and a r e marked by arrows. Golgi preparation. Mnccicct m u l n t t n .

4

A rectangular area of a photomicrograph taken from the medial segment of the globus pallidus (MGP) in section (fig. 3 ) . T h e radial fiber bundles pursue m a n y different radii as they funnel into the medial segment of the globus pallidus; hence, the spaces they occupy (marked by arrows) appear as oval “windows.” Golgi preparation: Mncacn mulntta. ( x 5.33 Obj.)

5

A rectangular area of a photomicrograph taken from the section (fig. 3 ) in a region of the lateral segment of th,e globus pallidus immediately adjacent to the external lamina. Here the spaces (arrows) for t h e radial fiber bundles have a “tunnel-like” appearance. Golgi preparation: Mnccicci mzclattu. ( x 5.33 Obj.)

THE RADIAL FIBERS IN THE GLOBUS PALLIDUS C . A. F o x and J . A. Rafols

PLATE 2

191

PLATE 3 E X P L A N A T I O N OF F I G U R E S

192

6

A horizontal section of a heavy impregnation cutting through the medial (MGP) and lateral (LGP) segments of the globus pallidus and a portion of the putamen (Put). The area outlined (arrow 8) i s the area in figure 8. Area below the tip of arrow 10 i s area in figure 10. Golgi prepara tion : M nc nc a m ulntt a .

7

A photomicrograph of a mid-longitudinal section through a n afferent plexus of “bouton en passage,” fine fibers that ensheath a dendrite. The dendrite is not impregnated. Golgi preparation: Mncaca mulnttn. ( X 43.46 oil immersion Obj.)

8

Photomicrograph of the area outlined in the outer portion of the lateral segment of the globus pallidus in figure 6. Note the radial fibers (RF) and their collaterals (C). The open block arrows point out the remarkable ensheathing plexuses that completely encase the long radiating dendrites and form longitudinal axodendritic synapses. Golgi preparation: Macacn mzclattn. ( X 10.25 Obj.)

9

A photomicrograph demonstrating divergence. The closed block arrow points to a n out-of-focus branch of a radial fiber collateral giving rise to a branch that contributes to an ensheathing plexus (open block arrow) and to a branch that proceeds in the opposite direction and bifurcates. Golgi preparation: Mncaca mztlattn. ( X 43.46 oil immersion Obj.)

THE RADIAL FIBERS I N THE GLOBUS PALLIDUS C . A. F o x and J. A. R a f o l s

PLATE 3

193

PLATE 4 EXPLANATION OF FIGURES

194

10

A photomicrograph of the area below the tip of the arrow 10 in the inner portion of the medial segment of the globus pallidus in figure 6. Observe the radial fibers (RF) and their collaterals (C). The radial fibers appear thinner here than in figure 8. The open block arrows point to ensheathing plexuses. T h e area 15 is the area shown in figure 15. (NP) neuropil. Golgi preparation: Mncncn mulnttn. ( X 10.25 Obj.)

11

A photomicrograph of a radial fiber (RF) in the lateral segment of the globus pallidus displaying three collaterals (C1, C2 and C3). Arrows indicate the extent in the section and the direction of the radial fibers (RF) and the collaterals. A branch from collateral C2 contributes to the ensheathing plexus (open block arrow) which is out of focus. Golgi preparation: Macaca mulnttn. ( X 43.46 oil immersion Obj.) N.B. Figs. 11-16, showing radial fibers and their collaterals, are reproduced at the s a m e magnification.

T H E RADIAL FIBERS IN THE GLOBUS PALLIDUS C. A. Fox and J . A. Rafols

PLATE 4

195

PLATE 5 EXPLANATION OF FIGURES

196

12

A photomicrograph of a radial fiber ( R F) in the lateral segment of the globus pallidus displaying a collateral (C). The fibers (X) are collaterals from radial fibers not in this section. Golgi preparation: Mncctcn mulnttn. ( x 43.46 oil immersion Obj.)

13

A photomicrograph of a radial fiber (RF) in the medial segment of the globus pallidus displaying a collateral ( C ) . Golgi preparation: Moccicri mulattn. ( x 43.46 oil immersion Obj.)

14

A photomicrograph of a collateral ( C ) in the medial segment of the globus pallidus arising from a thickened portion of a radial fiber (RF). Golgi preparation: Mnccicn m u l n t t n . ( X 43.46 oil immersion Obj.)

15

A photomicrograph of the area in the medial segment of the globus pallidus outlined (15) in figure 10. It shows a radial fiber (RF) and its collateral (C) and a radial fiber with two collaterals (C1 and C 2 ) . C1 enters the neuropil ( N P ) and C2 extends along the side of the neuropil. The closed block arrow points to the severed end of the radial fiber. Golgi preparation: M n c n c n mulatto. ( X 43.46 oil immersion Obj.)

16

A photomicrograph from the medial segment of the globus pallidus showing two radial fibers (RF) with a collateral and two radial fibers with two collaterals (C1 and C2). Golgi preparation: Mncciccc mulattci. ( X 43.46 oil immersion Obj.)

T H E RADIAL FIBERS IN T H E GLOBUS PALLIDUS C . A . Fox and J. A. R a f o l s

PLATE 5

197

PLATE 6 EXPLANATION

17

198

OF F I G U R E

An electron micrograph from the outer portion of the medial segment of the globus pallidus of the squirrel monkey (Saimiri sciureus) showing an island of neuropil (NP) trapped between masses of radial fibers (RF). Compare this with the photomicrographs of the Golgi preparation: viz. that seen in figure 15 and the area outlined in figure 10. T h e neuropil contains: a few myelinated fibers; two nerve cell body profiles (NC) which a r e either from the same or different neurons; a n oligodendrocyte ( 0 ) ;a few dendrites (D). The synaptic endings that form a mosaic covering the dendrites and the masses of extremely fine afferent fibers a r e poorly resolved at this magnification. The large myelinated fiber (L) is interpreted here as a pallidofugal fiber.

THE RADIAL FIBERS IN T H E GLOBUS PALLIDUS C . A. F o x a n d J . A. Rafols

PLATE 6

199