The Distribution of Retino-Collicular Axon Terminals in Rhesus Monkey

Transcription

1 The Distribution of Retino-Collicular Axon Terminals in Rhesus Monkey J. G. POLLACK 'School of Optometry/The Medical Center and Neuroscience Program, Uniuersity of Alabama in Birmingham, Birmingham, Alabama AND T. L. HICKEY ' ' ABSTRACT The retino-collicular projections in rhesus monkeys were studied using standard autoradiographic and degeneration techniques. A computer based technique was developed which provided a flattened visual display of the retinal projection onto the entire superior colliculus, quantified the area covered by such projections for different segments of the colliculus and showed how this morphological pattern varied with depth beneath the collicular surface. In the anterolateral third of the colliculus (i.e., the foveal representation) the retinal projection was light, confined to a narrow region of the superficial gray and contributed primarily by the contralateral eye. In the remaining binocular segment of the superior colliculus the retinal projections showed a marked degree of local patterning, that in many instances appeared as bands of label. By combining eye removal and eye injection procedures in a single animal and comparing adjacent sections processed for autoradiography and stained for degeneration, it was possible to assess the amount of overlap between retino-collicular projections from the two eyes. These experiments showed that total segregation of retinal afferents does occur in the monkey superior colliculus, but what occurs more often is a situation where the density of inputs from the two eyes varies reciprocally as one moves across the part of the colliculus that represents the binocular visual field. Recent autoradiographic studies (Graybiel, '75, '76; Hubel et al., '75) have demonstrated that the retino-recipient layers of the cat and monkey superior colliculus are organized partly on the basis of ocularity. In the cat superior colliculus ipsilateral retinal projections are discontinuous, forming a series of patches or "puffs" of label when viewed in single frontal sections. However, when serial sections are viewed in sequence, it is evident that the individual patches of label seen on single sections are, in fact, a part of longitudinal bands that traverse the surface of the colliculus. Graybiel ("76) has suggested that these bands form alternating left eye-right eye stripes, somewhat analogous to the ocular dominance columns seen in the visual cortex (Hubel and Wiesel, '68, '69, '72; Shatz et al., '77; LeVay et al., '78). In the monkey, the existence of a similar patterning of retino-collicular input has been suggested (Hubel et al., '751, but has not been demonstrated conclusively. Furthermore, even if retino-collicular projections from the J. COMP. NEUR. (1979) 185: two eyes do form bands in the monkey, the amount of overlap between the ipsilateral and contralateral projections is still unclear. In an attempt to answer these questions, we undertook an extensive analysis of the pattern of retino-collicular projections in seven rhesus monkeys. We utilized autoradiographic techniques, supplemented by computer reconstruction, to examine the distribution of retinal axon terminals in the superior colliculus as well as combined autoradiographic and degeneration techniques to determine the degree of overlap between the inputs from the two eyes. In general, our findings show that each colliculus receives a significant projection from both eyes. In all cases the contralatera1 eye is more strongly represented, at least in terms of the volume of collicular tissue innervated. In the anterolateral one-third of the Present address: Lt. J. Pollack, MSC. USNR, NAMI, NAS, Pensacola, Florida a Send reprint requests to: Lk. T. L. Hickey, School of Optometry, University of Alabama in Birmingham, Birmingham, Alabama

2 588 J. G. POLLACK AND T. L. HICKEY TABLE 1 Summary of animals used Monkey Right eye Survival Left eye Survival Plane of section 1 H3Proline (l.omc) 2 H3Proline (1.2 mc) 3 H3 Proline (500 pc) H3Leucine (250gC) 4 Removed H3Proline (250 pc) ladays H3Leucine (250uC) 5 Removed 5 days H3Leucine (500hC) 6 I H3Leucine (500 pc) H3Proline (500 pc) superior colliculus, the area representing approximately the central 10" of visual field (Cynader and Berman, '721, the retinal projection, as demonstrated by the autoradiographic technique, is light in some monkeys and totally absent in others. In the part of the colliculus representing more peripheral regions of the binocular visual field, the input from the two eyes is arranged in bands somewhat analogous to those seen in the cat superior colliculus (Graybiel, '75, '76). More peripheral regions of the visual field (i.e., the monocular crescent) form a fairly continuous projection onto the posteromedial surface of the contralateral superior colliculus. MATERIALS AND METHODS A summary of the experimental protocol followed for each of the seven adult rhesus monkeys used in the present study is shown in table 1. All animals received an intraocular injection of a tritiated amino acid (doses ranging between 500 and 1,200 pci). For five of the seven animals this was the only experimental intervention. The brains from these animals were embedded in paraffin, sectioned (14 p) in either the coronal, parasagittal or tangential plane, mounted on glass slides, dipped in photographic emulsion (Kodak NTB-2) and exposed for one month at 4 C. These sections were then developed and stained with cresyl violet. Camera lucida drawings, made from the coronal and parasagittal sections, were used in making the computer assisted reconstructions described below. Two animals had one eye removed 5 to 12 days before the other eye was injected. The brains from these animals were cut frozen (25 p) with every fourth and fifth coronal section processed for autoradiography and stained for degeneration (Wiitanan, '69), respectively. Since two adjacent sections then showed the 20 hours 36 hours 20 hours Coronal Coronal Parasagittal 20 hours Coronal 20 hours Coronal 20 hours Oblique 20 hours Tangential -- distribution of axon terminals from both eyes, it was possible to study the overlap of retinal input from the two eyes by aligning the sections on common landmarks. Computer assisted reconstructions In an attempt to reconstruct the morphological pattern of retino-collicular input, we developed a computer based technique which: provides a flattened visual display of the retinal projection onto the entire superior colliculus, quantifies the area covered by such projections for different segments of the col- A. 8. Fig. 1 Line drawing illustrating the method used to digitize and transform the anatomical data for computer plotting. A. The outlines of the collicular surface and radioactive labeled region are first drawn with the help of a camera lucida. This drawing is then digitized using a Graph-Pen digitizer. The important points and distances used in the transformation are shown. B. The collicular surface and retinal projection shown in A as it appears after the transformation. For additional information see text.

3 RETINO-COLLICULAR PROJECTIONS IN RHESUS MONKEYS 589 liculus and shows how this morphological pattern varies with depth beneath the collicular surface. Initially, every third section was drawn (63 x under darkfield illumination with the help of a camera lucida. Points of uncertainty were always checked using brightfield illumination and a higher magnification to insure that the silver grains were distributed around cell bodies. Each drawing consisted of the outline of the collicular surface, a common landmark (e.g., the midline) and enclosed areas corresponding to the regions covered by dense terminal label. These drawings were traced using a Graph-Pen digitizer and the data stored on magnetic tape for further processing. The data is stored in the form of X and Y coordinates where X is the distance from the origin (0) along the collicular surface and Y is depth below the collicular surface. In order to flatten the colliculus, the distance between successive digitized points on the collicular surface, such as Q in figure la, - is calculated from the origin (0). The distance OQ (X in fig. 1A) is then plotted as a straight line. This line represents the flattened surface of a segment of the colliculus with absolute distance preserved. We then reconstructed the perimeter points (P) of the labeled areas beneath the surface. This was accomplished by finding for each point P the point Q on the surface which was closest. The distance between P and Q corresponds to the Y coordinate for the flattened representation. RESULTS The computer based technique makes available two types of visual display. First, a flattened reconstruction of the entire colliculus can be obtained (for example, see fig. 2). This type of plot shows a flattened reconstruction of the retinal projection to each of the sections studied. To help in viewing the overall pattern of retino-collicular projection the individual sections have been lined up to show the colliculus as viewed from above. Starting at the top of the reconstruction shown in figure 2, one can seen how the pattern of the retino-collicular projection, as determined by the distribution of silver grains, changes as one moves from the rostra1 to the caudal pole of the colliculus. Although depth information is also contained in this reconstruction; i.e., in terms of the thickness of the blackened areas, this information is difficult to interpret. To assist in this interpretation, a second type of reconstruction was developed. This reconstruction shows the labeled regions as before, except that only one depth beneath the collicular surface is illustrated on any given plot (for example, see fig. 4). Since these reconstructions represent only one depth, they are drawn using lines rather than solid, blackened areas. Each line represents a scan at a given depth across one section with consecutive lines representing a series of sections through the rostral-caudal extent of the colliculus. Such plots can be obtained for any given depth below the collicular surface. Relative volume occupied by ipsilateral and contralateral retinal projections In the rhesus monkey, each visual hemifield is projected topographically onto the surface of the superior colliculus (Cynader and Berman, '72). The central visual field is represented anterolaterally and the lateral (monocular) field posteromedially. We assessed the distribution of label, in terms of relative labeled volume of collicular tissue, by dividing the colliculus into zones approximating the central 10" of the binocular visual field (CF) (Wilson and Toyne, '70; Cynader and Berman, '72; Hubel et al., '751, the remaining binocular visual field (BF) representation (as defined by the caudal boundaries of the ipsilateral retinal projection) and the monocular visual field (MF) representation. Since the borders of these zones have been defined quite arbitrarily, the relative volume of labeled tissue in any one zone is only approximate. Figure 3A shows the labeled volumes averaged over three animals. When the total volume of one superior colliculus receiving input from either eye is normalized to loo%, 4% of the labeled tissue is located in the CF, 82% in the BF and 14% in the MF. The relative inputs of the two eyes are shown in figure 3B. In the CF only one-fourth (1%) of the labeled tissue receives input from the ipsilateral eye. In the BF, where 82% of all labeled tissue is found, the relative inputs are again biased in favor of the contralateral eye by a factor of 2 (56%) to 1 (26%). The MF of the superior colliculus contains 14% of the total labeled volume, all of which is due to projections from the contralateral eye. The colliculus is therefore dominated by contralateral input in terms of the volume of collicular tissue innervated. It is important to note that for this analysis, and for the plots shown in figures 2 and 4, we

4 590 J. G. POLLACK AND T. L. HICKEY Fig. 2 Plot showing a series of transformed camera lucida drawings made from coronal sections taken from throughout the rostral-caudal extent of the right superior colliculus in a monkey that received a left eye injection. In this plot all three dimensions are represented; the thickness of the blackened areas corresponding to the depth dimension. A well-defined landmark, the optic disc representation (OD), can be seen as a large, oval break in the contralateral projection. The figure is aligned on the midline. M is medial, A is anterior. The overall shape of the colliculus is distorted since including the depth information required elongating the plot along the anterior-posterior axis.

5 RETINO-COLLICULAR PROJECTIONS IN RHESUS MONKEYS 59 1 Anterior A. Medial Lateral Posterior 8. I psi la t era1 Contralateral Posterior Fig. 3 The relative volume of collicular tissue receiving retinal projections. In each case the colliculi were first divided into segments approximating the central 10" of visual field (CF), the remaining binocular visual field (BF), and the monocular field (MF). Then the volume of tissue receiving a retinal projection was determined for each colliculus following an injection of radioactively labeled amino acid into one eye. The volumes of labeled tissue in the colliculi ipsilateral and contralateral to the injected eye were then summed, the resulting value being interpreted as the total volume of tissue occupied by the retinal projections onto one colliculus. In A, this volume was then set at 100% and the volume contained in each of the three segments calculated relative to this 100%. In B, the relative input from each eye is shown separately. Although the ipsilateral and contralateral projections are shown in two collicular outlines the values given still correspond to the relative volumes within one colliculus.

6 592 J. G. POLLACK AND T. L. HICKEY have been concerned with the dense patches of label that are quite obvious using darkfield illumination. In many instances light label could be seen extending between patches of dense label, especially contralateral to the injected eye. While this lighter label is only slightly above background level, it may correspond to a sparse retinal projection. If such areas had been included in the volume calculations, the relative amount of tissue innervated by the contralateral eye would have been even greater. Distribution of retinal projections The projections of the ipsilateral and the contralateral retinal fibers differ, both in their distribution across and in their depth below, the collicular surface (Graybiel, '75, '76; Hubel et al., '75). The projection of the central retina onto the anterolateral part of the superior colliculus has been the subject of some discussion in the literature. In this region, some studies using both degeneration and autoradiographic techniques have failed to find evidence of retinal input (e.g., Brouwer and Zeeman, '26; Bunt et al., '75; Wilson and Toyne, '701, while other studies have found a light, though definite projection (Hendrickson et al., '70; Hubel et al., '75). We have identified label of retinal origin in the anterolateral superior colliculus for some, but not all, of our monkeys. In those animals where a projection could be found (as in the case of fig. 4) the label was light, diminished rapidly as the anterolateral border of the colliculus was encountered, and was confined to the most superficial parts of the superior colliculus. These animals were our most heavily labeled monkeys. Consequently, the degree of labeling in the foveal representation may be, to some extent, a function of the amount of radioactive tracer injected. Where label was found, using the autoradiographic technique, the majority of label was contributed by the contralateral eye, that from the ipsilateral eye being less distinct or not present at all. In all cases the part of the lateral geniculate nucleus receiving input from central retina (Malpeli and Baker, '75) was well labeled, demonstrating that differences in the presence of label in the anterolateral superior colliculus are not due to uneven retinal uptake of radioactive material. In the two animals which also had one eye removed, we were not able to see any clear signs of degenerating axon terminals in this area. The majority of the retinal projection from both eyes, in terms of the volume of collicular tissue labeled, represents the region of visual field from approximately 10" off the vertical meridian to the outer boundaries of the binocular visual field (BF). Within this region, both eyes are represented by a patchy distribution of label on individual sections. When these patches are traced over serial sections, a banding pattern becomes evident that is similar to, but not as distinct as, that reported by Graybiel ('75, '76) for the cat superior colliculus. In addition, at least a partial laminar segregation of label becomes apparent as the ipsilateral retinal projection is most commonly displaced below that from the contralateral eye. Figure 4 represents a series of depth profiles for the ipsilateral and contralateral retinal inputs. The superior colliculus is viewed from above, and each consecutive computer plot shows the next level in depth - as if layers of an onion were being removed. The lines indicate the presence of dense label. Clearly, some label from each eye appears at each level examined, however, a partial laminar segregation does appear. The segregation is most clear in the region between the optic disc representation and the anterolateral edge of the colliculus. Here, ipsilateral label is found to be heavier from 70 p to about 150 p below the surface, whereas the contralateral input is absent below 110 p. Hence, the two projections are somewhat unbalanced at the most superficial level due to the predominance of the contralateral input, then combined at an intermediate level, and finally segregated at the deeper level due to the absence of a contralatera1 input. The monkey illustrated in figure 4 shows one region of label that, like the foveal projection, occurred in some of our more heavily labeled monkeys. While there is a tendency for the ipsilateral projection to be located superficially in lateral parts of the superior colliculus, we have also seen a very superficial tier of label in medial parts of the colliculus. Such superficial projections are always accompanied by a deeper ipsilateral projection. Examples of such tiering of ipsilateral input can best be seen in figure 7B (arrow), but is also apparent in figure 4. A very similar projection in the cat superior colliculus has previously been described by Graybiel('76) and was mentioned for the monkey superior colliculus by Hubel et al. ('75).

7 RETINO-COLLICULAR PROJECTIONS IN RHESUS MONKEYS 593 Near the medial and lateral borders of the superior colliculus differences in the distribution of ipsilateral and contralateral projections were even more pronounced. On the contralateral side, the medial and posterior borders of the superior colliculus were covered by an almost solid sheet of label forming a crescent some 1,200 to 1,300 p wide. This crescent, which is largely monocular, is more discontinuous superficially, giving a scalloped appearance in coronal sections. On the ipsilateral side, the collicular borders were, for the most part, devoid of label. However, three distinct patches of label consistently appeared here. Two of these have been previously described by Hubel et al. ( 75). At the lateral border of the ipsilateral superior colliculus and just anterior to the level of the optic disc representation is found an isolated patch of label 50 p wide and some 1,000 to 1,200 p long with a depth distribution from 50 p to almost 200 p (see arrow 1 in fig. 4). This relatively isolated projection has been found in all our animals. On the medial border, an anteriorly placed patch of label is present, which becomes continuous with other label posteriorly (see arrow 2 in fig. 4). This patch is also about 1,000 p in length and is present throughout the depth of the ipsilateral superficial gray. Finally, a third, not previously described, patch of label appears on the medial border at the level of the optic disc representation (see arrow 3 in fig. 4). This patch is connected to the main projection at its anterior end and does not seem to be as prevalent throughout the depth of the superficial gray. In the posteromedial segment of the superior colliculus the ipsilateral retinal projection is almost totally absent below 70 p. This region receives a strong projection from the contralateral eye and most certainly corresponds to the monocular segment of the superior colliculus (Cynader and Berman, 72). Finally, there is one other area where the ipsilateral and contralateral projections appear to be totally segregated. This area, which can reasonably be interpreted as the optic disc representation, can be seen quite clearly in figures 2 (OD) and 4 (arrow and outlined area). In figure 2, the optic disc representation appears as a break or hole in the contralateral projection in the caudal half of the colliculus. Although other smaller holes in the contralateral projection can be seen, they are never as consistently placed nor as large as the break we have interpreted to be the optic disc representation. Almost identical representations of the optic disc have been seen in every brain studied. In addition, when the ipsilateral retinal projection is studied, a corresponding solid projection (arrow in fig. 4) can be seen in this region. While this projection is greatest superficially, there is some ipsilatera1 input to this region throughout the depths shown in figure 4. Pattern of retinocotlicular projections As suggested by Hubel et al. ( 79, the retinocollicular projection in rhesus monkeys is organized in a manner similar to that seen in the cat (Graybiel, 75, 76). That is, the retinotectal projection exhibits a marked degree of local patterning that is similar to the banding seen in the cat superior colliculus; although such bands in monkeys are considerably more variable and much less distinct. While the overall orientation of the band-like patches of label in the monkey superior colliculus is longitudinal, the pattern is complex. Figure 5 shows an example of the complexity of the banding pattern for the ipsilateral retinal projection to a region just anterior and medial to the optic disc representation. This photomicrograph is a mosaic composed of several darkfield photomicrographs of serial sections cut tangential to the surface of the colliculus. Thus, the resulting photomicrograph represents a view of the top of the superior colliculus. The comparable region in the colliculus contralateral to the injected eye showed a similar, though much less striking, pattern. Overlap of retinal projections Given the rather complex, but definite, pattern of retino-collicular projection seen in figures 4 and 5, the question immediately arises as to what extent the inputs from the two eyes are segregated by having the contralateral label confined to the spaces on the ipsilateral side and vice versa. We have attempted to answer this question in two ways. First, in figure 6, we have redrawn one of the line plots (90 p) shown in figure 4. However, now the ipsilateral and contralateral projections have been superimposed using the optic disc representation for alignment. In addition, the isolated contralateral projections are shown in green and the isolated ipsilateral projections in red. Region of overlap are shown in black. This figure shows that the retino-collicular projections from both eyes overlap in some

8 DEPTH 50 U Fig. 4 The distribution of the retino-collicular projections, as seen on a series of coronal sections, is shown for several depths beneath the collicular surface. The ipsilateral projection is shown in the left column, the contralateral projection in the right column. The depth beneath the collicular surface is shown at the upper right hand corner of each depth pair. In all cases the outlines of the colliculus are approximate and are included only to assist in making comparisons between ipsilateral and contralateral projections. For each pair medial is to the center, anterior at the top. The optic disc

9 RETINO-COLLICULAR PROJECTIONS IN RHESUS MONKEYS 595 n DEPTH 170 U representations in the contralateral (outlined area) and ipsilateral (arrow) superior colliculi are best seen 90 fi below the collicular surface. A few examples of isolated ipsilateral input can be seen in areas that would normally be considered the monocular segment (for example see the line plot for 110 p). It is possible that these projections are aberrant. However, it is also possible that these few patches of silver grains do not correspond to retinal projections at all, but simply represent regions of high background label.

10 596 J. G. POLLACK AND T. L. HICKEY Fig. 5 Darkfield photomontage of tangentially cut sections (see drawing upper right) through the superior colliculus ipsilateral to the injected eye. The montage shows a region approximately 1 mm diameter located anterior and medial to the optic disc representation. areas and remain separate in other areas. Viewing the projections in this way also shows that both the overlapping and segregated projections are organized in a band-like pattern. While this pattern is complex, all of the bands appear to intersect the monocular segment at approximately right angles. This technique for comparing input to each colliculus is of limited use, however. Perfectly coronal sections, i.e., sections that cut through exactly the same regions of both hemispheres, are extremely difficult to obtain. Furthermore, the optic disc representation forms the only reliable landmark on which sections from the two colliculi can be aligned. Since the optic disc representation is seen clearly in only a few of the depth plots shown in figure 4, it is impossible to make accurate comparisons for all levels of retinal input. However, if comparisons between ipsilateral and contralateral projections can be made in a single colliculus, then the plane of section becomes relatively unimportant. To do this, we have made comparisons in animals who had one eye removed several days prior to having their remaining eye injected with a radioactively labeled amino acid. Figure 7A shows a darkfield photomicrograph of a coronal section through the contralateral colliculus at the level of the optic disc, and figure 7B the corresponding section on the ipsilateral side. Inspection of the two photomicrographs suggests that the retinal inputs are segregated in some regions. However, to

11 RETINO-COLLICULAR PROJECTIONS IN RHESUS MONKEYS 597

12 598 J. G. POLLACK AND T. L. HICKEY Fig. 7 Darkfield photomicrographs showing contralateral (A and C) and ipsilateral (B) projections onto coronal sections through the superior colliculus at the level of the optic disc (OD) representation. The region enclosed in the rectangle in A is shown in more detail in C. The regions labeled a-d are further described in figure 8.

13 RETINO-COLLICULAR PROJECTIONS IN RHESUS MONKEYS 599 Fig. 8 Brightfield photomicrographs of the section adjacent to the one shown in figures 7A and C. The regions labeled a-d in figure 7C are shown here on the adjacent section stained for degeneration.

14 600 J. G. POLLACK AND T. L. HICKEY verify this, it is necessary to examine the next section in the series stained for degeneration (fig. 8a-d). Figure 7C is a more highly magnified view of the outlined area in figure 7A. Each of the regions labeled a-d are also shown on adjacent sections stained to show degenerating axon terminals (fig. 8). In three of these regions (a, b, and d) the contralateral retinal projection, as measured by the distribution of silver grains, was markedly reduced. In every instance, examination of the adjacent section showed that degenerating axon terminals filled each of the regions under study. It is important to note, however, that while the degenerating axon terminals were most prevalent near the center of each of these regions, other examples of terminal degeneration could be seen extending into the adjacent areas, and thus, overlapping with a dense contralateral projection. In the remaining region (c) marked in figure 7C the contralateral retinal projection is only slightly reduced. However, figure 8c shows that the ipsilateral retinal projection also extends into this region, again overlapping with the contralateral input. These findings show that many areas of the superior colliculus receive input from both eyes, a situation that is also apparent in the line plot shown in figure 6. There are other regions in the superior colliculus where the dense label corresponding to the ipsilateral and contralateral retinal projections do not overlap. The two most obvious examples of this are the optic disc representation, where the ipsilateral projection fills a void in the contralateral projection, and the monocular segment, where the contralateral projection exists in isolation. Figure 6 shows that, at least for the projection to a region of the colliculus 90 p below the surface, the ipsilateral and contralateral inputs are segregated in many other areas. It must be emphasized, however, that these line plots show only the areas that contained the dense patches of terminal label. As mentioned before, there were other areas, especially contralateral to the injected eye, where more sparsely distributed label could be seen between the dense patches of label. If these regions of light label do, in fact, represent retinal projections, then there would be far more overlap of retinal projections than shown in figure 6. The findings in the animals having one eye removed and the other eye injected add further support to this hypothesis. In general, in areas where the silver grains were severely reduced or absent, the density of degenerating axon terminals tended to be greatest. As the number of silver grains increased in adjacent areas the number of degenerating axon terminals appeared to decrease. In a few instances, we were able to find small areas containing cell bodies that appeared not to receive retinal input from either eye. Such areas were not confined to any one part of the superior colliculus, sometimes even occurring in the monocular segment. It is possible that such cell areas receive input from other parts of the brain. However, it is also possible that cells in these areas have dendritic appendages that extend into nearby regions that do receive input. For the most part, our detailed analysis of the extent of overlap between retinal projections has been confined to the areas near the level of the optic disc representation. We were unable to examine the complimentarity of inputs to the anterolateral superior colliculus (foveal representation) since our staining techniques consistently failed to demonstrate any degenerating axon terminals in this region. DISCUSSION The findings reported here confirm and extend many of the results previously presented by Hubel et al. ('75). In addition, the present study shows that while the overall pattern of retino-collicular projections are qualitatively similar in monkey and cat (Graybiel, '75, '761, there are notable differences both in terms of the more prominent ipsilateral input to monkey superior colliculus (see also Hubel et al., '75) and the rather less distinct banding pattern seen in the colliculi of our monkeys. Hubel et al. ('75) and Graybiel ('75, '76) both report that the contralateral retino-collicular projection forms a rather continuous band throughout the more superficial level of the upper gray in both the monkey and the cat. Our findings differ from theirs slightly in that we also find the dense contralateral projection to be distributed in patches (figs. 2, 4,6, 7) at least somewhat similar to those seen ipsilatera1 to an eye injection. This finding is most evident for regions lying between the optic disc representation and the anterolateral border of the colliculus. Along the medial and posterolateral borders of the colliculus, and at deeper levels, the contralateral projection did appear more continuous. In addition, regardless of depth beneath the surface of the col-

15 RETINO-COLLICULAR PROJECTIONS IN RHESUS MONKEYS 601 liculus, the contralateral projection always becomes quite continuous near the posteromedial border of the superior colliculus; i.e., the monocular segment. In both the cat (Graybiel, '75, '76) and monkey (Hubel et al., '75) the ipsilateral projection is generally reported to occupy deeper levels of the superficial gray than the projection arising from the contralateral eye. With the exception of the superficial tier of the ipsilateral input seen anteromedially in some monkeys, our findings are in agreement with those previously reported. Such differences in depth are especially evident for the projections from parts of the binocular visual field beyond the central 10". Within the central visual field representation we find relatively little ipsilateral input, even though the contralateral projection is quite definite. The rather striking topographical pattern of retinocollicular projections in both cat and monkey has led to the suggestion that right and left eye inputs to the superior colliculus may be segregated in a fashion similar to that seen in layer IV of the monkey visual cortex (Hubel et al., '68, '69, '72; Wiesel et al., '74; LeVay et al., '75; and others). In the monkey superior colliculus, the projections from the two eyes are segregated in many areas, including the optic disc representation and the monocular segment. In addition, there are many parts of the colliculus that clearly receive a dense retinal input from both eyes. While both types of input; i.e., overlapping and segregated, are organized in a band-like pattern (we have refrained from calling these bands columns since the colliculus is not organized in a columner fashion equivalent to that seen in cortex), this pattern is by no means as distinct or as regular as the ocular dominance columns seen in visual cortex. In addition, our present findings suggest that there may be considerably more overlap between retinal projections in the colliculus, especially if one takes into account the fact that the areas containing sparse label are not illustrated in the reconstructions. If such label represents a less dense retinal projection (rather than label contained in fibers-of-passage), then one is faced with a more common situation where the density of inputs from the two eyes varies reciprocally as one moves across the part of the colliculus on which the binocular visual field is represented. One additional observation that would suggest this to be the case is that the optic disc representation contralat- era1 to the injected eye was always free of label, even when light label could be seen between dense patches of label in nearby parts of the colliculus. Since there is no reason to believe that fibers-of-passage would avoid the optic disc representation, it seems likely that any light label not contained in axon terminals would be as noticeable here as in other nearby parts of the colliculus. With this in mind, it is interesting to relate the pattern of retinocollicular input seen here in adult animals to the pattern of cortical ocular dominance columns seen in newborn monkeys (Hubel et al., '77). During the first few weeks of a monkey's postnatal life, there is a progressive segregation of ocular input to layer IV of the visual cortex. Initially, the input from the two eyes is largely overlapping. As the monkey's visual system matures more and more geniculo-cortical afferents become confined to their appropriate ocular dominance column until finally the inputs from the two eyes are totally segregated. Although it must be considered speculative at best, it is interesting to hypothesize that the segregation of input to the cortex and colliculus occurs in much the same way except that the segregation of retino-collicular afferents has not reached the level of development seen in the monkey visual cortex; actually being more similar to that seen in the cat visual cortex (Shatz et al., '77; LeVay et al., '78). The anatomical findings reported here coincide well with previous physiological recordings made in monkey superior colliculus (Cynader and Berman, '72; Schiller et al., '74). Given the distribution of retino-collicular projections described here it is quite easy to understand how most superior colliculus cells could be binocularly driven; although, given the pattern of overlap in retinal projections seen in our monkeys, one might expect to more often find areas where one eye is dominant over the other (Hubel et al., '75), especially in the parts of the colliculus near the optic disc representation. In the antero-lateral third of the superior colliculus the retinal projection, as demonstrated by the autoradiographic technique, is light. However, visually driven cells have been recorded in this region, even after the visual cortex has been cooled or removed (Schiller et al., '74). It is conceivable that the autoradiographic technique does not demonstrate all of the retinal projections to this region, possibly missing some of the fine fibered projections. Some support for this idea

16 602 J. G. POLLACK AND T. L. HICKEY comes from the fact that the animals showing a retinal projection to this region were consistently those that received the most radioactive amino acid. Given the balance of input to this region from the two eyes, one might expect to record more contralaterally dominated cells in a decorticate monkey. A slight tendency for this to occur is suggested in the histograms published by Schiller et al. ( 74). ACKNOWLEDGMENTS We thank David Simmons and Gloria Avery for their technical assistance and J. Gerard for much of the computer programming. This work was supported by N.I.H. Grants EY and EY and N.S.F. Grant BMS to T.L.H. Doctor Pollack was supported by N.I.H. Postdoctoral Fellowship F 32 EY LITERATURE CITED Brouwer, B., and W. P. C. Zeeman 1926 The projection of the retina in the primary optic neurons in monkeys. J. Physiol., 49: Bunt, A. H., A. E. Hendrickson, J. S. Lund, R. D. Lund and A. F. Fuchs 1975 Monkey retinal ganglion cells: Morphometric analysis and tracing of axonal projections, with a consideration of the peroxidase technique. J. Comp. Neur., 164: Cynader, M., and N. Berman 1972 Receptive-field organization of monkey superior colliculus. J. Neurophysiol., 35: Graybiel, A. M Anatomical organization of retinotectal afferents in the cat: An autoradiographic study. Brain Res., 96: Evidence for banding of the cat s ipsilatera1 retinotectal connection. Brain Res., 114: Hendrickson, A. E., M. E. Wilson and M. J. Toyne 1970 The distribution of optic nerve fibers in Macaca mulatta. Brain Res., 23: Hubel, D. H., S. LeVay and T. N. Wiesel 1975 Mode of termination of retinotectal fibers in macaque monkey: An autoradiographic study. Brain Res., 96: Hubel, D. H., and T. N. Wiesel 1968 Receptive fields and functional architecture of monkey striate cortex. J. Physiol. (London), 195: Anatomical demonstration of columns in the monkey striate cortex. Nature (London), 221: Laminar and columnar distribution of geniculo-cortical fibers in the macaque monkey. J. Comp. Neur., 146: Hubel, D. H., T. N. Wiesel and S. LeVay 1977 Plasticity of ocular dominance columns in monkey striate cortex. Phil. Trans. R. SOC. Lond. B., 278: LeVay, S., D. H. Hubel and T. N. Wiesel 1975 The pattern of ocular dominance columns in macaque visual cortex revealed by a reduced silver stain. J. Comp. Neur., 159: LeVay, S., M. P. Stryker and C. J. Shatz 1978 Ocular dominance columns and their development in layer IV of the cat s visual cortex: A quantitative study. J. amp. Neur., 179: Malpeli, J. G., and F. H. Baker 1975 The representation of the visual field in the lateral geniculate nucleus of Macaca mulatta. J. Comp. Neur., 161: Schiller, P. H., M. Stryker, M. Cynader and N. Berman 1974 Response characteristics of single cells in the monkey superior colliculus following ablation or cooling of visual cortex. J. Neurophysiol., 37: Shatz, C. J., S. H. Lindstrom and T. N. Wiesel 1977 The distribution of afferents representing the right and left eyes in the cat s visual cortex. Brain Res., 131: Wiesel, T. N., D. H. Hubel and D. K. Lam 1974 Autoradiographic demonstration of ocular-dominance columns in monkey striate cortex by means of transneuronal transport. Brain Res., 79: Wilson, M. E., and M. J. Toyne 1970 Retino-tectal and cortico-tectal projections in Macaca mulatta. Brain Res., 24: Wiitanen, J. T Selective silver impregnation of degenerating axons and axon terminals in the central nervous system of the monkey (Macaca mulattai. Brain Res., 14: