Regional Variation in the Representation of the Visual Field in the Visual Cortex of the Siamese Cat

Transcription

1 THE JOURNAL OF COMPARATIVE NEUROLOGY 193: (1980) Regional Variation in the Representation of the Visual Field in the Visual Cortex of the Siamese Cat MICHAEL LEE COOPER AND GARY G. BLASDEL Division of Biology, California Institute of Technology, Pasadena, California ABSTRACT In Siamese cats, many ganglion cell fibers from the temporal retina misproject to the contralateral hemisphere; as a result, each lateral geniculate nucleus contains an abnormally large representation of the ipsilateral visual field. The manner in which the visual cortex processes this aberrant visual information has been examined in several previous studies. In some Siamese cats, the region of the 17/18 border was found to contain an extensive, systematic map of the ipsilateral field, while in other animals no such map was found, and the 17/18 border appeared to represent the zero meridian of azimuth (as in normal cats). These results have led to the suggestion that there are two distinct types of Siamese cat ("Boston" and "Midwestern") which can be distinguished on the basis of cortical topography and the anatomical organization of the geniculocortical pathway. In the present study, we have recorded from four Siamese cats in order to examine the visual field map in the region of the 17/18 border; in each cat we recorded at anterior coronal levels corresponding to the representation of the lower visual field, and also at more posterior levels near the horizontal meridian representation. In all of the animals we found that the anterior penetrations (corresponding to mean receptive field elevations inferior to -7") yielded 15-20" of ipsilateral field representation at the 17/18 border; however, the posterior, horizontal meridian penetrations (with mean elevations from +lo to -4") showed excursions of only about 5" into the ipsilateral field. This large difference in the representation of azimuth was not due to rotation of the eyes during our recording sessions. The finding of appreciable differences in the amount of ipsilateral field represented at different anterior-posterior levels of the same animal might lead to the suggestion that there are not two distinct populations (or types) of Siamese cat with regard to the cortical map of the ipsilateral field. Rather, we raise the possibility that Siamese cats form one population in which there is a continuous variation in the extent of ipsilateral field represented in the cortex. In the Siamese cat, as in other albino mammals, a genetic defect results in the misrouting of axons arising from ganglion cells in the temporal retina (Guillery, '69; Guillery and Kaas, '71; Kalil et al., '71; Hubel and Wiesel, '71; Kirk, '76; Shatz, '77a; Stone et al., '78; Cooper and Pettigrew, '79c; and Mason et al., '79). Many of the temporal ganglion cells, which normally send their fibers ipsilaterally, instead misproject to the contralateral hemisphere; in this way, each dorsal lateral geniculate nucleus (LGN) comes to contain an abnormal representation of much of the ipsilateral visual field. The significance of this chiasmatic misrouting for the organization of higher order visual pathways has aroused considerable interest. What have been considered to be secondary re- arrangement of visual pathways in response to the primary chiasmatic defect have been described for the geniculocortical system (Hubel and Wiesel, '71; Kaas and Guillery, '73), corticogeniculate system (Montero and Guillery, '78; Shatz and LeVay, '79), callosal projections (Shatz, '77b,c), and cortical associational pathways (Shatz and LeVay, '79). Most attention has been focused on the geniculocortical pathway, where previous work has reported two seemingly distinct modes of processing the aberrant retinal input to the LGN (see Fig. 1). Dr. Cooper's present address: Section of Neuroanatomy, Yale University SehoolofMedicine, 333 CedarStreet,New Haven, Conn Dr. Blaedel's present address: Dept. of Ophthalmology, Storm Eye Institute, Medical University of South Carolina, 171 Ashley Ave., Charleston, South Carolina, ALAN R. LISS, INC.

2 238 M.L. COOPER AND G.G. BLASDEL On the basis of their cortical recordings, Hubel and Wiesel ( 71) described a Boston cortical modification, in which geniculate fibers carrying aberrant visual field information appear to be inserted as a block into their own region surrounding the border between cortical areas 17 and 18. In the normal cat, areas 17 and 18 both contain representations of the contralateral visual field, with the border between these areas representing the vertical midline (Otsuka and Hassler, 62; Hubel and Wiesel, 65; Bilge et al., 67; Tusa et al., 78, 79); however, in Boston Siamese cats there is also a well-ordered topographic map of much of the ipsilateral visual field, and the 17/18 border comes to represent the maximum ipsilateral excursion of - 15 to Thus, in penetrations traversing the border (entering area 18 at the crown of the lateral gyrus; see Fig. 21, Hubel and Wiesel ( 7 1) found that the receptive fields of cells from which they recorded moved from the contralateral visual field, across the zero azimuthal meridian, and into the ipsilateral hemifield as the track approached the 17/18 border. When the electrode reached the border itself, the receptive fields were located at their most ipsilateral travel (- 15 to - 20 ); and as the electrode moved beyond the border, the receptive fields moved back toward the zero mendian and then into the contralateral hemifield. A second geniculocortical pattern was reported by Kaas and Guillery ( 73). In their Midwestern cats, cortical topography appeared basically normal, since the 17/18 border represented the zero meridian; there was no systematic representation of the aberrant retinogeniculate input. Kaas and Guillery ( 73) showed anatomically that LGN fibers arising in the abnormal segment of lamina Al (or abal, the segment of lamina A1 receiving abnormal crossed input from the contralateral temporal retina) do indeed reach their normal, topographically correct regions of the cortex, in spite of the fact that they carry aberrant ipsilateral field information. In other words, these fibers project to the same cortical locations as they would if they were driven through the ipsilateral eye and conveyed normal information about the contralateral hemifield. Since the recordings of Kaas and Guillery ( 73) showed that only a relatively small number of cortical units respond to stimuli in the ipsilateral hemifield, it appears that the aberrant visual input from the LGN is somehow suppressed in the Midwestern cortex. Within the otherwise normal map of the contralateral hemifield, Kaas and Guillery did find a few receptive fields in the ipsilateral visual field; these receptive fields were placed mirrorsymmetrically across the zero meridian from the receptive fields of surrounding units. This result shows that the suppression of aberrant input is not complete and gives further evidence that aberrant fibers from abal end in their topographically normal cortical locations. The major difference between cortical recordings from normal and Midwestern Siamese cats is that in the Midwestern, there is relatively little representation of the ipsilateral eye, the great majority of units being driven only through the contralateral eye. The previously published reconstructions of electrode penetrations through the cortex of Boston Siamese cats have been from anterior coronal levels representing the inferior visual field (elevations below about -5 ; Hubel and Wiesel, 71; Shatz, 77a; Cooper and Pettigrew, 79c). We had been working on Siamese cats which showed a seemingly typical Boston modification at these cortical levels, and it became of interest to us to determine whether the amount of ipsilateral visual field represented in the cortex was similar at all coronal levels. Accordingly, in the work reported here we constructed detailed maps of the aberrant ipsilateral field representation in the region of the 17/18 border, both at anterior levels, representing lower visual field (below about -7 1, and at more posterior levels representing elevations near the horizontal meridian. To our surprise, we observed great differences in the amount of aberrant visual field represented below and at the horizontal meridian. In the anterior penetrations through the map of the inferior field, we found usual Boston topographies, with the 17/18 border corresponding to a point ipsilaterally; near the horizontal meridian, however, our maps of the 17/18 border yielded only about 5 of ipsilateral field, an ipsilateral representation which is much closer to that seen in normal cats (see Discussion) and Midwestern Siamese cats. Abstracts of this work have appeared elsewhere (Cooper et al., 78; Cooper and Blasdel, 79). METHODS The cats The cats used in this study had typical Siamese markings. One animal (BCC-6) was obtained from outside sources, while the other three (BCC-3, BCC-4, and BCC-5) were littermates from our colony (and were thus related to the previously studied cats LGNS-21 and LGNS-25 of Cooper and Pettigrew, 7912). When

3 VISUAL FIELD IN VISUAL CORTEX OF SIAMESE CAT 239 NORMAL MIDWESTERN -2p BOSTON tex.gn T contra eye ips# aye contra eye ipsi eye contra eye ipsi eye Fig. 1. Diagram summarizing the pathways from the eye to the visual cortex in normal and Siamese cats (modified from Shatz, 77a). In the normal cat, ganglion cells in the temporal retina (stippled) project ipsilaterally to laminae A1 and C1 of the dorsal lateral geniculate nucleus (LGN) and to the medial interlaminar nucleus (MIN). The MIN and laminae A, C, and C2 (not labeled) of the LGN receive input from the nasal retina of the contralateral eye. In the Siamese cat (labeled here as Midwestern and Boston ), many of the ganglion cells in the temporal retina misproject to the contralateral hemisphere. As a result of this chiasmatic defect, the geniculate laminae which normally receive ipsilateral eye input (A1 and C1) are, in the Siamese, broken into discrete islands receiving either normal input from the ipsilateral retina or abnormal projections from the contralateral retina. Lamina A1 tends to be split into at least three islands (Guillery, 69; Guillery and Kaas, 71j Shatz, Wa), two with normal input and one with aberrant input. The island of normal input near the medial border of the LGN, called the medial normal segment of A1 (or MNS, labeled M here, nomenclature of Guillery and Kaas, 711, has beenconsidered to receive fibers from the ipsilateral temporal retina close to the zero meridian, while the island farthest laterally, the lateral normal segment of Al (LNS), has been thought to be the target of more peripheral ganglion cells in the ipsilateral temporal retina. (The possibility has been raised that the MNS and/or LNS might contain compressed maps representing more ofthe ipsilateral temporal retina than would be expected solely on the basis of position in the LGN. See Lund, 75, Stone et al., 78, and Cooper and Pettigrew, 79~). Contralaterally misprojecting fibers from the temporal retina end in what has been called the abnormal segment of Al (abal), as well as in part of C1. The contralateral projections from the nasal retina appear to end normally in laminae A, C, and C2. Thegeniculocortical pathway from the A-laminae to area 17 is schematized in the upper part of the figure. In the normal cat corresponding points in laminae A and A1 project to the same point in the cortex to give a topographic map of the contralateral hemifield. The maps in areas 17 and 18 are roughly mirror images, with the 17/18 border (unmarked arrow) representing the zero azimuthal meridian (0 ). In what has been called the Midwestern Siamese modification (Kaas and Guillery, 731, geniculate fibers from abal (dashed line), which carry aberrant ipsilateral field information, project to their topographically normal locations in the cortex. However, this aberrant input is somehow suppressed, so that the overall topography ofthe visual field map is normal, with the 17/18 border again representing the zero meridian of azimuth. In what was called the Boston Siamese modification (Hubel and Wiesel 71), fibers from abal are routed to a new position in the cortex, so that a topographic map ofthe ipsilateral visual field is organized in the region around the 17/18 border (diagonal stripes), with the border itself coming to represent a point placed up to about 20 ipsilaterally (-20 ). As a result of the creation of this new ipsilateral field map, the zero meridian representation (0 ) is displaced away from the 17/18 border; in fact the zero meridian is represented twice, once deep in area 17 and again deep in area 18.

4 240 M.L. COOPER AND G.G. BLASDEL alert, none of these animals had an obvious squint; however, under paralysis, the areae centrales diverged by an average of nine degrees (loo, lo", I", and 15" in BCC-3, -4, -5, and -6, respectively). In the normal cat, paralysis leads to a divergence of about 5" (Bishop et al., '62). Surgical procedures Our recording procedures were similar to those described previously (Blasdel and Pettigrew, '78; Cooper and Pettigrew, '79a). These procedures have been devised so as to limit surgical intervention and minimize traumasteps which facilitate the later recovery of the animal. After inducing anesthesia with Halothane in a mixture of nitrous oxide and oxygen (2:1), we quickly intubated the animals with a plastic endotracheal tube that had been coated with lidocaine jelly (2%). We then inserted a 25 gauge butterfly needle into the cephalic vein and paralyzed the animals with 40 mg of Flaxedil (gallamine triethiodide), after which we transferred them to a stereotaxic headholder. The anesthetic gases were changed at this time to 75% nitrous oxide in carbon dioxideioxygen without Halothane. The level of expired carbon dioxide was monitored periodically and maintained at 3.5% by adjusting the amount of inspired carbon dioxide. In our experience, the removal of Halothane never led to a reduction in surgical analgesia as judged by changes in ECG or dilation of the pupils (prior to the application of atropine) following noxious stimuli. The animals were maintained in a healthy state throughout the recording sessions with a continuous infusion of essential amino acids and salts (Ambex, 1 m1150 ml infusion; 4.6 ml infusion/hr.) in 10% dextrose. We added Flaxedil to this infusion (10 mgikglhr) in order to eliminate eye movements. In addition, the animals received one large dose of dexamethasone (4 mg, iv) and antibiotics (penicillin and streptomycin, im) every 12 hours. In order to maximize the chances of obtaining satisfactory electrode penetrations (see below), each cat was studied during two separate recording sessions. In the first session, we recovered the animals after 24 hours of study (Blasdel and Pettigrew, '78). This usually allowed us enough time to make detailed examinations of the topography at two different locations in the cortex. At the end of this time, we discontinued the use of Flaxedil and in its place infused a solution of 1Wo dextrose. After closing the wound margins of the scalp with sutures, we removed the animals from the stereotaxic instrument. The anesthetic gases were then discontinued and replaced with oxygen-enriched air. When the animals showed signs of recovery from paralysis, we reversed the paralysis completely by administering atropine followed by prostigmine. The cats were then allowed to rest for several days before the second recording session. At the end of this second session, the animals were anesthetized with a near-fatal dose of sodium pentobarbitol and perfused through the heart with saline followed by formal-saline. We cut blocks of the visual cortex on the freezing microtome at 40 pm and stained the sections with cresyl violet or the Loyez myelin stain. The criteria of Otsuka and Hassler ('62) and Tusa et al. ('79) were used to identify the position of the 17/18 border on each track reconstruction. Recording procedures We recorded cortical activity with tungstenin-glass electrodes (Levick, '72) and plotted the receptive field positions for cortical units on a tangent screen placed 57cm in front of the animal. Conventional stimulating and recording techniques were used (Nikara et al., '68; Pettigrew, '74). As mentioned above, the object of these experiments was to compare the amount of ipsilateral visual field represented at different coronal levels in the visual cortex of our Siamese cats. Thus, in each animal we made two sets of penetrations in the same hemisphere, one set between HC + 1 and +6 and the other set between HC -3 and -5. The coronal level of the first group of tracks was chosen so as to explore the representation of inferior visual field (5-15" below the horizontal meridian), while the second, posterior tracks were intended to map the horizontal meridian representation itself (Bilge et al., '67; Tusa et al., '78, '79). In order to reliably estimate the amount of ipsilateral hemifield represented at each coronal level, we made as many tracks as were necessary to ensure that at least one penetration crossed the 17/18 border and contained a closely spaced sample of units around this border. In each track we attempted to record visual activity at least every 100 pm; although most of the data in this paper come from well-isolated single units, we sometimes plotted the position of multi-unit activity if the electrode had been advanced much beyond 100 pm without isolating a clear single unit. In this study we mapped receptive fields for 329 units. During the recording sessions, we inferred when the

5 A B lateral SUICUS supra- A a I, medial \\ lateral Fig. 2. (a) Showing the manner in which 17/18 border shifts laterally as one proceeds caudally in the visual cortex. In the center of the figure is a dorsal view of the right hemisphere; the dashed line represents the border between area 17 (located medially in the lateral gyrus) and area 18 (located laterally). Also shown on this hemisphere are the locations of two coronal planes which correspond approximately to the AP levels chosen for mapping in this study. In the more rostra1 section A (on the left; medial is to the left and lateral to the right in both sections), the 17/18 border lies medially, near the greatest convexity of the lateral gyrus. At the more caudal level B (on the right), corresponding to the approximate horizontal meridian representation, the 17/18 border is shifted laterally and lies on the lateral bank of the gyrus. (b) Photograph of a Loyez myelin-stained section corresponding approximately to coronal level A in Figure 2a. This section came from the region near the anterior electrode track in BCC4. The arrow shows the approximate location ofthe 17/18 border. electrode traversed the 17/18 border by noting the maximum excursion of receptive fields into the ipsilateral hemifield. Electrolytic lesions were made in each track in order to facilitate subsequent reconstruction. Monitoring of eye position and torsion The positions of the opticdisks were projected on the tangent screen with a fiber optic light pipe (Pettigrew et al., '79). In the Siamese cat the areae centrales are often not distinct under ophthalmoscopic examination (Hubel and Wiesel, '71; Shatz, '77a), so we took the zero meridian for each eye to lie 15.8" nasal to the projection of the optic disk (Nikara et al., '68). The positions of the ocular landmarks were checked frequently during each recording session in order to monitor residual movements of the eyes. The eyes appeared to move very little between consecutive measurements. We did find, however, that over the course of an entire recording session the two eyes drifted gradually downward by 1-2"Jday. In three cats (BCC-4, -5, -6) we estimated the amount of eye rotation due to paralysis (Hubel and Wiesel, '62; Nelson et al., '77; Cooper and Pettigrew, '78a). Before each recording session we photographed the cat's eyes in moderately bright light and determined the relative orientations of the slit-pupils (see Cooper and Pettigrew, '79a). After paralysis, but before the eyes were dilated, we again photographed the slit-pupils. Comparison of the two sets of photographs yielded the amount of rotation. Tracing of retinothalamic pathways using horseradish peroxidase In addition to making cortical recordings, in BCC-5 we also filled the left LGN (as well as the medial interlaminar nucleus and ventral lateral geniculate nucleus) with horseradish peroxidase (HRP). The injection and histological procedures used here, as well as the method of analysis of the retinal whole-mount data, were identical to those employed by Cooper and Pettigrew ('79b; also see this paper for the controls against contamination of the results by retinomescencephalic axons). The results of the HRP injection into BCC-5 were similar to those described previously (Cooper and Pettigrew, '7912). In particular, we found HRP-labeled ganglion cells throughout the entire ipsilateral temporal retina (including central retina) in BCC-5. In the contralateral retina we observed labeled ganglion cells extending at least 20" into the temporal periphery. However, there were relatively fewer contralateral HRP-filled cells temporal to 20" than in the cats studied by Cooper and Pettigrew ('79~). Fading of the blue HRP-reaction product in the

6 242 M.L. COOPER AND G.G. BLASDEL TABLE 1. Summary of data for tracks which were histologically confirmed to have crossed the 17/18 border Left or HC Mean Elevation Range of Maximum Ipsilateral Cat Track Right Cortex Level 2 Stand. Dev. Elevations Excursion BCC3 A1 right A2 right P1 right P2 right BCC4 A1 right PI right BCC5 A1 right P1 right P2 left BCC6 A1 left P1 left +3.0 f f f " 2 3.5" -7.8" i 1.8" -2.9" & 1.1" -3.8" 2 0.7" -8.4"? 1.2" +0.2" 2 1.1" -12.8" 2 1.3" +Lo" -c 0.9" -2.3" 2 1.0" -7.0"? 1.1" +LO" * 0.5" -6.3" to -17.8" -3.7" to -10.3" -1.0" to -6.1" -2.3" to -4.9" -5.2" to -10.1" -3.1" to +1.6" -10.2" to -15.5" -1.2" to +2.1" -3.9" to -0.5" -4.6" to -8.4" +1.8" to +0.3" -14.2" (unit 22) " -5.9" (unit 4) -5.2" (units 35, 36) " (unit 6) -4.6" (unit 33) -18.8" (unit 38) -5.2" (unit 58) -5.8" " (unit 33) -4.4" (unit 8) The mean receptive field elevation is given for each track, as is the range of receptive field elevations encountered in the track. In addition, the maximum excursion into the ipsilateral hemifield (in degrees of azimuth) is included here. contralateral retina prevented us from making detailed maps of HRP-filled cell density in this case. RESULTS The main part of these results will describe our attempt to obtain detailed maps of the cortical topography in the Siamese cat. Table 1 summarizes the data derived from each of our four experimental animals; this table lists the mean and standard deviation of the receptive field elevations for all of the units in each penetration, as well as the maximum amount of ipsilateral field represented in each track. In the table and text below, all receptive field coordinates refer to the positions of the receptive field centers. Except for the seven units noted below, all receptive fields were mapped through the eye contralateral to the recorded hemisphere. Extent of ipsilateral field representation Cat BCC-3 We placed our first penetration (track Al, Fig. 3a) anteriorly in the visual cortex of BCC-3. The electrode was positioned so that it would enter area 18 at Horsley-Clarke anterior 3 (HC +3), cross the 17/18 border, and then run vertically down the medial bank of the lateral gyrus. The first visually responsive unit in area 18 (unit 11) had a receptive field which was located 2" into the contralateral hemifield and 12" below the horizontal meridian. (We had the impression that the units in our penetrations in BCC-3 had receptive fields lying lower in the visual field than we would have expected at the same Horsley-Clarke coordinates in a normal animal.) As we advanced the electrode toward the 17/18 border, the receptive fields of subsequent units drifted gradually across the zero Fig. 3. Reconstructions of electrode tracks through the right cortex of Siamese cat BCC3. Figure 3a presents the reconstruction of an anterior track (HC +3) with mean receptive field elevation of ", while in 3b we show two penetrations (with mean receptive field evelations of -2.9" and -3.8") through more posterior cortex. The diagram at the lower right of Figure 3b shows a dorsal view of the brain on which are indicated the locations of the various electrode tracks in this animal. Coronal sections containing the penetrations appear on the right side of Figures 3a and 3b. The location of each electrolytic lesion in a penetration is denoted by L. Medial is to the left and lateral to the right in each section, and the 17/18 border is represented in each case as a heavy arrow. The degree values lying along the cortical surface indicate the approximate azimuth values for underlying receptive fields. On the left side of Figures 3a and 3b are shown receptive field plots for the units encountered in the various microelectrode penetrations through the cortex of BCCB. Each receptive field position is depicted relative to the area centralis, which is represented for each cell by a horizontal cross-mark. Azimuth is represented along the horizontal axis, with the contralateral hemifield to the left and the ipsilateral field to the right. These conventions also apply to the track reconstructions in Figures 4-6. The anterior track A1 (Fig. 3a) entered the cortex in area 18, and the first unit (11) encountered in this penetration had a receptive field centered near the zero azimuthal meridian. As the electrode was advanced towards the 17/18 border (arrow), the receptive fields moved into the ipsilateral hemifield, until the maximum excursion of -14.2" was reached at unit 22. The receptive fields moved back towards the zero meridian as the electrode proceeded deeper, into area 17. Track A2 (mean receptive field elevation = -7.8") is not reconstructed here; it yielded 17.7" of ipsilateral field representation. The crown of the lateral gyrus is distort+ in this anterior region due to edema occuring after the first recording session (see footnote 2). In the posterior cortex (Fig. 3b), the electrode penetrations entered area 17 and crossed the border to end in area 18. The maximum ipsilateral excursion noted in these tracks was only -5.9". Thus there was a substantial difference in the representation of aberrant homolateral field between the anterior map of the lower visual field and the more posteriorly located map of the horizontal meridian region.

7 a CONTRA VISUAL FIELD IN IPS TRACK P1 lpsl IPS1 CONTRA IPS1-100, -2'J' BCC-3 RIGHT CORTEX TRACKS P1 & PZ nc -4.5 medial 13 r TRACK P2 lpsl r IPS1-200 F f 41 b Figure 3

8 244 M.L. COOPER AND G.G. BLASDEL azimuthal meridian' into the ipsilateral hemifield and then shifted rather more quickly (units 18-21) until they were about 14" into the ipsilateral hemifield. In the region of the 17/18 border itself, we found unit 22 with a receptive field centered 14.6" ipsilaterally. The movement into this hemifield was accompanied at first by a drift downward in receptive field elevation, so that unit 18 was situated about 20" below the horizontal meridian, almost twice as inferior as the first units in penetration. This movement downward was not due to the angle of the electrode penetration, inasmuch as the receptive fields of following cells moved back upward until unit 23, located near the furthest ipsilateral excursion, had a receptive field which was only 6 degrees below the horizontal meridian. As the electrode crossed the 17/18 border and moved on into area 17, the receptive fields moved, as expected, back toward the zero meridian (unit 26); they also returned to a lower position in the visual field. This anterior penetration (with mean elevation of -11.3") thus contained near the 17/18 border a wellordered map of much of the ipsilateral visual field, and on this basis we initially identified BCC-3 as a typical "Boston" Siamese cat. A second anterior electrode penetration (A2, mean receptive field elevation = -7.8") at about Horsley-Clarke + 1 also seemed to identify this animal as "Boston," since the receptive fields moved almost 18" into the ipsilateral hemifield. A reconstruction is not shown for this track. As in the first penetration, the receptive fields moved upward as they progressed out into the ipsilateral visual field. Wishing to explore the representation of the horizontal meridian, we placed our third and fourth electrode tracks (P1 and P2, Fig. 3b) at posterior 4.5 (HC -4.5). These penetrations were similar to the anterior ones in that they each crossed the 17/18 border; however, since the border at this posterior level is situated on the lateral side of the postlateral gyrus (see Fig. 21, we found it necessary to angle the electrode laterally so that it crossed more tangentially from area 17 into area 18. The first unit which we encountered in track P1 lay some distance below the surface of the cortex. Its receptive field was still located 3.5" below the horizontal meridian. This receptive field was centered 4.7" into the homolateral hemifield and that of another unit (unit 41, which was located a few hundred micra farther along the track, lay only 5.9" into the ipsilateral field. Subsequent track reconstructions revealed that unit 4 lay very close to the 17/18 border (see Figure 3b). As we advanced the electrode into area 18, the receptive fields moved slowly back to the zero meridian (unit 19); we did not find any receptive fields centered more ipsilaterally than that for unit 4. Fearing that we might have missed a more substantial ipsilateral representation by insufficiently sampling the region of the 17/18 border, we placed a fourth track, P2, slightly medial to P1. There was very little change in receptive field position in the evenly spaced sample of units made as track P2 approached the 17/18 border; all of the receptive fields (except the first, unit 21) lay 3.5"-5" ipsilaterally and about 4" below the horizontal meridian. The greatest ipsilateral excursion was again found near the 17/18 border, with units 34 and 35 having fields centered at -5.2". The fields then drifted back toward the zero meridian (unit 40) as the track passed into area 18. Our subsequent reconstructions of tracks P1 and P2 showed that our sample of units in the vicinity of the 17/18 border was sufficiently closely spaced to give confidence that only the first five to six degrees of ipsilateral field were expressed near the horizontal meridian representation (mean elevation for P1 = -2.9", mean elevation for P2 = -3.8"). Hence, in contrast to the anterior tracks, these posterior penetrations showed very little representation of the ipsilatera1 field. The recordings from BCC-3 thus revealed substantial differences between the amount of aberrant ipsilateral hemifield represented at different cortical levels. Since all of the tracks were placed in the right cortex and all of the units were mapped through the left (contralateral) eye, asymmetries between the cortical hemispheres or errors in estimating the position of the zero meridian can not be invoked as explanations for these differences. Cats BCC-4, -5, and -6 The cortical penetrations in BCC-4 and BCC-5 gave results very similar to those obtained from their littermate, BCC-3. In BCC-4, the anterior track (Al, Fig. 4) crossed the 17/18 border at HC +3; the mean receptive field elevation for the track was -8.4", while the greatest ipsilateral excursion was -14.5". In contrast, in BCC-4's posterior track P1 (HC -4.21, with a mean elevation of +0.2", the receptive fields were never centered more than 4.6" 'For the reasons discussed elsewhere (Cooper and Pettigrew, '79a), we prefer the terms "zero meridian'' or "zero azimuthal meridian" to the more commonly employed "vertical meridian." The terms "homolateral" and "ipsilateral" will be employed interchangeably here.

9 TRACK P1 lpsl n -700 '200 I U 20 IPS1 IPS1-10" ' -20' TRACK A1 HC '3 15 Q11 '60 Fig. 4. Track reconstructions for BCC4. The results are very similar to those obtained in BCC3 (Fig. 31; 14.5" ofhomolateral field were represented in the anterior, lower field track (mean elevation = - 8.4") and only 4.6" in the posterior, horizontal meridian penetration (mean elevation = +0.2"). Units 13, 23, 37, and 38 were identified as LGN fibers. ipsilaterally (unit 33, on the 17/18 border). In cat BCC-5, the rostra1 penetration (Al, Fig. 5) was placed in the right cortex at HC +6; the mean receptive field elevation of the 24 units around the 17/18 border was -12.8", while the extreme of the ipsilateral representation, which was reached near the border at unit 38, was ". The caudally placed track P1 in this cat centered the right cortex at HC -4, very close to the horizontal meridian representation (mean elevation = + 1.0"); again, the receptive fields in this track never progressed more than 5.2" (unit 58) into the right, homolateral hemifield. A second posterior track, P2, was placed in the opposite (left) cortex of BCC-5. No reconstruction is shown for this track, although it did cross the 17/18 border (at HC -3.5). The mean elevation of the 17 units on this 3 mm-long track was -2.3"; no units were encountered which were centered more than 5.8" ipsilaterally. This track demonstrates that the representation of the homolateral field is fairly

10 246 M.L. COOPER AND G.G. BLASDEL ncc-5 RIGHT CORTEX TRACK A1 lpsl -10- IPS1 r\-= Fig. 5. Microelectrode penetrations in Siamese cat BCC5. Ninetten degrees of ipsilateral field were expressed in the anterior track (mean elevation = , while the fields in the posterior penetration P1 (mean elevation = + 1.9") were never found ipsilateral to-5.2". Track P2 was placed in the opposite hemisphere to P1; no reconstruction is shown for P2, which did cross the 17/18 border and yielded an almost identical ipsilateral excursion to track P1. symmetrical between the two hemispheres at these posterior levels. Thus, the recordings from the littermates BCC-3, -4, and -5 were all quite consistent in showing a large, 10-14" difference between the amount of aberrant ipsilateral field represenbd at the anterior and posterior cortical levels' In Order to ensure that this finding was not peculiar to animals from our Colony, we from siamese cat BCC-6, which was obtained from outside sources. The two tracks cortex in the figure is distorted due to edema which occured after the end of the first recording session.? The results from this animal conform to those obtained from our colony-bred "In two cats, BCC-3 (Fig. 3a) and BCC-6 (Fig. 61, the cortex in the region of the anterior tracks was distorted due to edema. However, the tracks were clearly visible and the tissue preservation was sufficient so that the 17/18 border was still identifiable using the Loyez myelin stain. In the reconstructions for these anterior tracks, our anatomical estimates of the position of the 17/18 border agreed well with the location ofthe maximum excursion into the ipsilateral hemifield found during recording. In the other two animals, BCC-4 and BCC-5. there which we made in the left cortex ofthis animal was no edema inthe anterior cortex, and in none of the cats did we find any sign of edema in the region of the posterior (horizontal meridian) are reconstructed in Figure 6. (The anterior tracks.

11 VISUAL FIELD IN VISUAL CORTEX OF SIAMESE CAT 247 TRACK PI BCC-6 LEFT CORTEX -209 IPS1 IPS1 IPS1 TRACK A1-2"" n Fig. 6. Electrode tracks through the left cortex of BCC6. The same conventions are used here as in Figures %5, except that the ipsilateral hemifield is to the left, and medial is to the right in the tracings of the coronal sections. BCCG was unrelated to the littermates BCC3,-4, and-5; however, the results from BCCG are almost identical to those from the other three animals. The anterior, lower field penetration A1 (mean elevation = -7.0") yielded receptive fields placed up to 14.8" ipsilaterally, while the posterior track (mean elevation = +1.0") gave a maximum excursion of only ahout 4.4" ipsilaterally. The coronal section for track A1 is somewhat distorted due to edema (see footnote 2). subjects. The anterior track at HC +4, yielding units with a mean elevation of -7.0", gave a maximum ipsilateral excursion of " (unit 331, while the posterior track again gave units very close to the horizontal meridian (mean elevation = + 1.0') and yielded a greatest travel of only 4.4" into the ipsilateral field (unit 8). Control for eye torsion When we obtained the electrophysiological results from BCC-3, it occurred to us that the observed differences in ipsilateral field representation might be due to some outward cyclotorsion of the eyes during paralysis; outward rolling of the contralateral eye would cause an ipsilateral shift of inferior field receptive fields with respect to fields near the horizontal meridian. In fact, this possibility seemed rather unlikely to us, since it is known that in normal cats paralysis induces an intorsion of 4-5" per eye (Hubel and Wiesel, '62; Nelson et al., '77; Cooper and Pettigrew, '79a).3 However, as a 3Eye-rolling would seem a rather improbable explanation in any case, given the large amount of extorsion needed to account for the results. For instance, the mean elevations of the two tracks in BCC-4 differ by 8.6", while the maximum ipsilateral excursions differ by 9.9". Inordertoaccount for thisdifferenceon the basisofocular torsion, each eye wouldbavetoextortbyarctan( ") = 47'. (Weuse 9.l"instead of 9.9" because in the normal alert cat each eye's zero meridian is extorted by about 5", which would produce a difference of about 0.8" over an elevation change of 8.6" [Cooper and Pettigrew, '79al.) Similar calculations for BCC-5 and BCC-6 would require extorsions of 42" and 50" per eye, respectively, in order to explain the observed greater representation of the ipsilateral field in the anterior track. Such substantial eye rolling is unlikely.

12 248 M.L. COOPER AND G.G. BLASDEL control for eye rolling, we determined the amount of eye torsion in BCC-4, -5 and -6 (see Methods). By measuring the angles between the slit-pupils before and after paralysis, we found that each eye rotated inward by 1-4". This inward rolling of the eyes is in the wrong direction for the observed differences in visual field representation to be explained by eye torsion. Summary of mapping results The main electrophysiological results of this study are summarized in Table I and in Figure 7, which is a scatter diagram of receptive field BCC-4 8=+0.2" 0-20'1 I -20" I 0 O0 O A1 B = -8.4' 4, O +5F P1 1 B=+l.O" A1 e = -?.OO O o? 00 '0 r0 A1 B = "

13 elevation (ordinate) versus azimuth (abscissa) for all of the cortical units encountered in the tracks which were histologically confirmed to have crossed the 17/18 border. With great consistency, these scatter plots emphasize the limited amount of ipsilateral field represented in the posterior penetrations; these penetrations correspond to mean elevations near the horizontal meridian (+l.o" to -3.8"), and the receptive fields in these tracks were never centered more than about 5" ipsilaterally. In contrast, the anterior penetrations, with mean elevations inferior to about -7", all included 15-20" ipsilateral field. It is interesting to note from Figure 7 that the greatest homolateral excursion of " for anterior track A2 of BCC-3 (upper left-hand scatter plot) corresponds to a receptive field only 3.7" below the horizontal meridian; such an elevation overlaps with the elevations of the posterior tracks in this animal. This may reflect the scatter of receptive field elevations at a given region of the cortex (Hubel and Wiesel, '74) and demonstrates the importance here of comparing not simply elevations of individual units on given tracks, but rather of using the mean receptive field elevations or the anterior-posterior levels of the tracks as the basis for comparisons of the ipsilateral field representation. In addition, this observation is one piece of evidence that the observed cortical differences are not due merely to the lack of contralaterally misprojecting ganglion cells near the horizontal meridian in the retina (see below). Receptive fields through the ipsilateral eye Although recent studies of the retinofugal pathways in the Siamese cat have revealed the presence of significant numbers of ipsilaterally projecting ganglion cells in the central temporal retina (Kirk, '76 Stone et al., '78; Cooper and Pettigrew, '79c), there is to date only slight evidence for a cortical representation of the central contralateral visual field through the ipsilateral eye (see Cooper and Pettigrew, '79c for further mention of this question). Hubel and Wiesel('71) and Kaas and Guillery ('73) found a total of only nine units influenced through the ipsilateral eye within the central approx. 12" (see also Chino and Shansky, '77). Although we did not search vigorously for ipsilateral eye responses, we did encounter seven units which could be driven through this eye and which had receptive fields within the central lo" of contralateral visual field. The receptive fields for these units are shown in Fig. 8; unit 9 in the figure was found in track P2 of BCC-5, and the VISUAL FIELD IN VISUAL CORTEX OF SIAMESE CAT 249 Tt5" 40 FIELDS THROUGH IPS1 EYE Fig. 8. Positions of receptive fields plotted through the eye ipsilateral to the recorded hemisphere in Siamese cats BCC4 and BCC5. Azimuth is represented along the horizontal axis and elevation along the vertical axis. All of these receptive fields were located in the contralateral visual field (positive azimuth values). Unit 36 was an LGN fiber; the other units were cortical neurons. Of these, units 9 and 23 could only be excited through the ipsilateral eye (Hubel and Wiesel 1'621 ocular dominance group 71, while the other units could be driven through either eye. other units were recorded in two tracks in BCC-4. One of the units (36) was classified as an LGN fiber. Of the cortical units, units 9 and 23 could be driven only through the ipsilateral eye (Hubel and Wiesel "621 ocular dominance group 7), while the others could be influenced by either eye, units 27 and 40 being group 2,35 being group 3, and 25 being group 5. These units may represent a geniculocortical input from the "medial normal segment" (see legend for Fig. 11, although (except for unit 36) such ipsilateral responses could also arise through callosal connections. DISCUSSION The ipsilateral field representation in the anterior and posterior penetrations The principal finding of this study is that there can be substantial differences in the amount of ipsilateral visual field represented at different anterior-posterior levels of the Siamese cortex. Penetrations through the inferior field representation reveal an extensive map of the ipsilateral hemifield, with the 17/18 border corresponding to a point 15-20" ipsilaterally; on the other hand, more posterior tracks near the horizontal meridian representation yield at most about 5" of ipsilateral excursion near the 17/18 border. One might be concerned that our result seflects an inadequate sampling along the posterior penetrations; in fact, it is quite unlikely

14 250 M.L. COOPER AND G.G. BLASDEL that we missed any substantial representation of the ipsilateral hemifield along the horizontal meridian. In each of the six reconstructions of the posterior tracks, we obtained a closely spaced sample of units in the vicinity of the 17/18 border. Given the relatively larger magnification factor in the representation of the horizontal meridian (i.e., more millimeters of cortex per degree of visual field), one would expect a given excursion into the ipsilateral hemifield to occur over a greater cortical distance at these posterior levels than anteriorly; yet in none of our six posterior penetrations did we see a receptive field centered more than 6" ipsilaterally. In addition, due to the position of the 17/18 border at about HC -4 and to the angle of our penetrations near this level, our posterior tracks were approximately parallel to the cortical surface as they crossed the region of the 17/18 border. (None of these tracks passed through the white matter anywhere near this border.) In contrast, the anterior penetrations traversed the 17/18 border near the greatest convexity of the lateral gyrus and usually passed through the deeper cortical layers (in BCC-3, -4 and -6); the rostra1 tracks thus travelled more rapidly across the radially fanning lines of iso-azimuth, making it more possible that we missed the very most ipsilaterally placed fields in the anterior penetrations. Therefore, it is more likely that we underestimated the extent of ipsilateral field representation in the anterior tracks than in the posterior tracks, so that, if anything, the difference in ipsilateral field representation may be greater than we report. Three of the four cats studied here were littermates from our breeding colony. However, since our fourth animal, BCC-6, gave the same results as the other three, our findings are not unique to animals from our colony. This point is supported by the work of Spear and Guillery (personal communication), who have obtained mapping results similar to ours. According to Shatz and LeVay (footnote #9: '79), however, there are Siamese cats in which this variation in the representation of azimuth does not exist. These authors mention finding a significant ipsilateral representation along the horizontal meridian as well as in the lower field. In addition, Kaas and Guillery ('73) and Shatz ('77a) have presented data suggesting that in some Siamese cats the cortical topography may be approximately normal at all coronal levels. From inspection of the scatter plots in Figure 7, it might appear that the difference in azimuth representation between our posterior and anterior penetrations occurs in an abrupt, step-like fashion, with a 10-15" difference in azimuth occuring for a difference in elevation of only about 3" (elevations from +lo to -4" showing only little ipsilateral field and elevations below about -7" showing the more extensive representation). The seemingly abrupt nature of this transition may be somewhat deceptive, however, since the distance on the cortex between the anterior and posterior sets of tracks was quite substantial (usually at least 7 mm). Owing to the way in which the visual field is mapped onto the cortical surface (the magnification factor), significant anterior-posterior movements of electrode position in the regions between our two sets of penetrations would yield only relatively small changes in mean receptive field elevation. Thus, it may be that further recordings would reveal a gradually decreasing representation of ipsilateral field as the penetrations were moved from anterior to posterior within the region intermediate between our two standard recording levels. Alternatively, it is conceivable that the variation in ipsilateral field representation at the 17/18 border is more irregular and may even shift in and out along the border. (We did not see strong evidence for this within the range of cortical levels chosen for our two sets of penetrations.) Further experiments will be needed to clarify these points. More work is also needed in order to explore the question of how the ipsilateral superior field is represented in the cortex. One might find in our animals that the ipsilateral field representation would increase again in the regions of cortex corresponding to the superior visual field. Cortical topography in Siamese cats: "Boston"l"Midwestern" dichotomy or continuum? Our anterior (lower field) penetrations yielded maps which are similar to those used to describe the "Boston" cortical pattern (Hubel and Wiesel, '71). However, it is not clear whether one should consider our posterior tracks, with their representation of only 4-6" of ipsilateral field, to indicate '' '5-degree' Boston" or "Midwestern" topographies. Even in the normal cat, receptive field centers at the 17/18 border typically have been reported to extend as far as 2" into the ipsilateral hemifield (Leicester, '68; Tusa et al., '78), and Harvey ('80) finds that this normal ipsilateral representation can reach -5", with occasional receptive fields centered as far as 15" into the ipsilateral hemifield. These findings, coupled with

15 VISUAL FIELD IN VISUAL CORTEX OF SIAMESE CAT 251 what is often a 1-2 uncertainty in placing the zero meridian, indicate that our posterior tracks approach the range of ipsilateral representation which has been found for normal animals. Our finding of only about 5 of ipsilateral field in the posterior tracks, when viewed in light of these uncertainties concerning the exact determination of the maximum ipsilateral excursion in the normal, shows that it can at times be difficult to decide on the basis of cortical recordings whether to identify a cortical topography as Boston or Midwestern. Such a result, taken in conjunction with Shatz s ( 77a) finding of two Siamese cats (WRl and 5R5) in which the receptive fields did not progress more than 10 ipsilaterally, may indicate that there are not actually two distinct Siamese populations with regard to the extent of the cortical representation of the ipsilateral field. In fact, the simple finding here of a varying representation of the ipsilateral field depending on elevation might in itself lead one to consider the possibility that there is a continuum in the amount of aberrant visual field that can be expressed in the cortex; there may be a continuous range of maximum ipsilateral excursions extending from the normal topography of what has been known as the Midwestern cortex up to the -25 representation in some of Shatz s ( 77a) animals. Ifthis is the case, it may not be appropriate to speak in terms of two distinct types of Siamese cat or cortical pattern (i.e., Midwestern and Boston ); rather, it might be that Siamese cats should be considered as a single population in which the representation of the ipsilateral field varies continuously, both between cats and maybe even within individual cats. Previous cortical recordings (Hubel and Wiesel, 71; Kaas and Guillery, 731, as well as anatomical studies of retrograde degeneration in the Siamese LGN following cortical lesions (Kaas and Guillery, 73; Guillery and Casagrande, 77; Shatz and LeVay, 79), strongly imply that there are two distinct mechanisms for the routing of individual geniculocortical fibers in the Siamese cat. Fibers carrying aberrant ipsilateral field information are directed to their topographically normal cortical locations by one mechanism, while they are routed to a new position by the other ( Boston ) mechanism. However, it may be that the total population of aberrant geniculocortical axons (i.e., those carrying aberrant ipsilateral field information) in a given Siamese cat contains fibers directed by both mechanisms. Thus, although individual aberrant fibers within a single cat may fall into two distinct populations on the basis of cortical termination, there could be a continuous variation between cats in the degree to which one or the other mechanism predominates within the total population of aberrant geniculocortical afferents. In this light, it might be mentioned here that there is some physiological evidence which suggests that both mechanisms for routing aberrant fibers can coexist in a single cat. In some of the animals used by Hubel and Wiesel( 71) and Shatz ( 77a) to identify the Boston mechanism, penetrations through the representation of the first approx. 20 of contralateral hemifield revealed patches of units with receptive fields placed mirror-symmetrically in the ipsilateral hemifield. Some single units even had two receptive fields, one in the contralateral hemifield and one in the mirror-symmetric location in the ipsilateral field. This finding might be explained on the basis of callosal connections, or a recently described abnormal corticocortical associational connection (Shatz and LeVay, 79). However, such an explanation does not appear likely because the mirror fields were found only in cortical layers IV through VI (Hubel and Wiesel, 71; Shatz, 77a), whereas there are strong callosal and associational inputs to layer 111, but none or few to layer IV (Shatz and LeVay, 79). It seems more probable that the presence of mirror fields reflects the routing of aberrant geniculocortical fibers to their topographically normal locations in the cortex. Additional evidence along these lines is provided by work of Shatz (personal communication), who found labeled cells in the MNS after HRP injections at the 17/18 border of cats she identified as Boston. This finding is further support for the suggestion that the segregation of normal and aberrant geniculocortical fibers is not always complete in what have been called Boston Siamese cats. Some results of HRP injections in hypo-pigmented mink (mink 895 of Guillery et al., 79) also might be interpreted as giving evidence of intermingling in the cortex of normal and aberrant geniculate fibers. From this discussion, it becomes clear that in order to determine in our cats which mechanism for routing aberrant geniculate fibers might predominate at posterior cortical levels, it would be necessary to trace the projections of individual geniculate cells by making lesions or HRP injections at the appropriate cortical locations. However, we might reemphasize here that this issue of the routing of single aberrant geniculocortical fibers is distinct from

16 252 M.L. COOPER AND G.G. BLASDEL that of the overall visual field topography found in the cortex. On some possible factors determining the extent of the cortical representation of the ipsilateral field It has been suggested (Shatz, 77a) that there is a systematic relationship between the size of the lateral normal segment of Al (taken as relating to the proportion of retinogeniculate fibers projecting ipsilaterally from the peripheral temporal retina; see Fig. 2) and the extent of ipsilateral field represented in the cortex. Shatz ( 77a) noted that since the size of the lateral normal segment (LNS) seems to vary inversely with that of abal, one might expect that a larger LNS would lead to a smaller extent of homolateral field in the cortex (see Fig. 11: Shatz, 77a). Shatz did report finding a loose inverse correlation between the size of the LNS and the extent of ipsilateral field represented at the 17/18 border (r = -0.72). However, this correlation should be viewed with caution, since Shatz took her data on LNS size from LGN levels close to the horizontal meridian representation, while her cortical data came from anteriorly placed tracks in the lower field representation. Thus, an extensive comparison between cortical and geniculate levels corresponding to the same elevation remains to be done. It is difficult to propose other factors which might determine the amount of aberrant ipsilateral field expressed in the cortex and which might explain the differences between our two sets of penetrations. Our cats do not lack aberrantly projecting ganglion cells along the horizontal meridian, since thalamic injections of HRP into BCC-5 and the cats of the previous study (two of which were related to the three colony-bred cats here; Cooper and Pettigrew, 79c) revealed substantial numbers of misprojecting ganglion cells throughout much of the temporal retina (including the horizontal meridian). It is interesting to note here that some of the previously published retinal isodensity maps of the percentages of retinothalamic ganglion cells projecting ipsilaterally in the Siamese show a nasally directed bulge near the horizontal meridian (Cooper and Pettigrew, 79c). However, it is not clear to what extent, if any, the differences in cortical topography observed in the present study reflect the quantitative aspects of the retinothalamic topography. In this respect, it would be useful to combine cortical recordings at several coronal levels with large HRP injections into the LGN in order to directly compare cortical and retinal topography in the same animal. It is possible that the differences between anterior and posterior tracks may be related in some way to the magnification factor (the amount of cortex devoted to a given amount of visual field) normally found at the different cortical levels. It is interesting to note that in the posterior tracks visual field position shifts much more slowly across the cortical surface than it does in the anterior tracks. Thus, the magnification factor of the aberrant ipsilateral field representation around the 17/18 border (in millimeters of cortex per degree of ipsilateral visual field) is much larger for the horizontal meridian than for the inferior field. This is similar to the situation at the 17/18 border in normal cats, where the magnification factor for the area centralis representation is much greater than that for more peripheral visual fields (Bilge et al., 67; Tusa et al., 78, 79). It may be that whatever factors produce expansion of the central representation with respect to the peripheral representation in normals also act in the Siamese to produce the observed greater magnification at the 17/18 border for the horizontal meridian map than for the more peripheral, lower field map. It is conceivable that the tendency toward lower magnification in the anterior cortex would allow this cortex to accomodate a larger representation of the ipsilateral field, thus yielding a greater excursion into the ipsilateral periphery. This discussion raises the possibility that consideration of the ipsilateral field representation in terms of its magnification factor (or evenly simply in terms of the absolute amount of cortex devoted to this representation) may be as instructive as considering only the maximum ipsilateral excursion itself. In the future, careful examination and comparison of cortical magnification factors in both normal and Siamese cats may yield data of great interest. CONCLUSION Although at present the developmental events which give rise to the differences observed here in ipsilateral field representation remain unclear, we feel that our results provide some simplification of the understanding of cortical topography in Siamese cats. We present here the suggestion that Siamese cats form a single population in which two mechanisms for routing aberrant geniculate fibers can be expressed to varying degrees, both between cats and perhaps even within individual cats, to produce a continuous variation (again, between cats and within individual cats) in the amount of ipsilateral hemifield found in the cortex. Further work will be needed to test this suggestion and to reveal the factors which de-

17 VISUAL FIELD IN VISUAL CORTEX OF SIAMESE CAT 253 termine which geniculocortical routing mechanism predominates at a given cortical level. ACKNOWLEDGMENTS We wish to express our deep thanks to Dr. J.D. Pettigrew for his constant support and encouragement. We are also grateful to Drs. Sheila and David Crewther, who bred three of the cats used in this study. In addition, we would like to express our gratitude to Drs. R. Guillery, J. Lund, R. Lund, C. Shatz, P. Spear, and J. Stone, who offered helpful comments on drafts of this paper. Sarah Kennedy provided expert technical assistance and Betty Thomas typed the manuscript. Lederle Labs kindly provided Flaxedil for these experiments. This work was supported by grants MH25852 and EY1909 from USPHS to J. Pettigrew. G.G.B. held a Weizmann Postdoctoral Fellowship. LITERATURE CITED Bilge, M., A. Bingle, K.N. Seneviratne, and D. Whitteridge (1967) A map of the visual cortex in the cat. J. Physiol. 191 : 11& 118. Bishop, P.O., W. Kozak, and G.J. Vakkur (1962) Some quantitative aspects of the cat s eye: Axis and plane of reference, visual field coordinates and optics. J. Physiol. 163:46& 502. Blasdel, G.G., and J.D. Pettigrew (1978) Effect ofprior visual experience on cortical recovery from the effects of unilateral eyelid suture in kittens. J. Physiol. 274: Chino, Y.M., and M.S. Shansky (1977) Relationship between interocular alignment and binocularity of striate cortical neurons in Siamese cats. Neurosci. Absts. 3:555. Cooper, M.L., and G.G. Blasdel(1979) Cortical topography in Siamese cats. Neurosci. Abstr. 5:780. Cooper, M.L., G.G. Blasdel, and J.D. Pettigrew (1978) Evidence in the Siamese cat for simultaneous expression of Boston and Midwestern patterns. Abstracts of Association for Research in Vision and Ophthalmology Annual Meeting, May, 1978, p Cooper, M.L., and J.D. Pettigrew (1979a) A neurophysiological determination of the vertical horopter in the cat and owl. J. Comp. Neurol. 184:l-26. Cooper, M.L., and J.D. Pettigrew (1979b) The decussation of the retinothalamic pathway in the cat, with a note on the major meridians of the cat s eye. J. Comp. Neurol. 187: Cooper, M.L., and J.D. Pettigrew (1979~) The retinothalamic pathways in Siamese cats. J. Comp. Neurol. 187: Guillery, R.W. (1969) An abnormal retinogeniculate projection in Siamese cats. Brain Res. 14:73%741. Guillery, R.W., and V.A. Casagrande (1977) Studies of the modifiability of the visual pathways in Midwestern Siamese cats. J. Comp. Neurol. 174: Guillery, R.W., and J.H. Kaas (1971) A study of normal and congenitally abnormal retinogeniculate projections in cats. J. Comp. Neurol. 143:7%100. Guillery, R.W., M.D. Oberdorfer, and E.H. Murphy (1979) Abnormal retinogeniculate and geniculo-cortical pathways in several genetically distinct color phases of the mink (Mustela vison). J. Comp. Neurol. 185: Harvey, A.R. (1980) A physiological analysis of subcortical and commissural projections of areas 17 and 18 of the cat. J. Physiol. In Press. Hubel, D.H., and T.N. Wiesel(1962) Receptive fields, binocu- lar interaction and functional architecture in the cat s visual cortex. J. Physiol. 160:1W154. Hubel, D.H., and T.N. Wiesel (1965) Receptive fields and functional architecture in two nonstriate visual areas (18 and 19) of the cat. J. Neurophysiol. 28:22%289. Hubel, D.H., and T.N. Wiesel (1971) Aberrant visual projections in the Siamese cat. J. Physiol. 218: Hubel, D.H., and T.N. Wiesel(1974) Uniformity of monkey striate cortex: A parallel relationship between field size, scatter and magnification factor. J. Comp. Neurol. 158:29&306. Kaas, J.H., and R.W. Guillery (1973) The transfer of abnormal visual field representations from the dorsal lateral geniculate nucleus to the visual cortex in Siamese cats. Brain Res. 59: Kalil, R.E., S.R. Jhaveri, and W. Richards (1971) Anomalous retinal pathways in Siamese cat: An inadequate substrate for normal binocular vision. Science 174: Kirk, D.L. (1976) Projections of the Visual Field by the Axons of Cat Retinal Ganglion cells. Section 2. Decussation of Optic Axons in Siamese Cats. Ph.D. Thesis, Australian National University. Leicester, J. (1968) Projection of the visual vertical meridian to cerebral cortex of the cat. J. Neurophysiol. 31: Levick, W.R. (1972) Another tungsten microelectrode. Med. Electron. Biol. Engng. 10: Lund, R.D. (1975) Variations in laterality of the central projections of retinal ganglion ~- cells. Exp. - Eye - Res. 21 : Mason, C.A., E.H. Polley, and R.W. Guillery (1979) Retinotopic organization of axons in the optic nerve and tract of normal and Siamese cats. Neurosci. Abstr. 5:795. Montero, V.M., and R.W. Guillery (1978) Abnormalities of the corticogeniculate pathway in Siamese cats. J. Comp. Neurol. 179:l-12. Nelson, J.I., H. Kato, and P.O. Bishop (1977) Discrimination of orientation and position disparities by binocularly activated neurons in cat striate cortex. J. Neurophysiol. 40:26G283. Nikara, T., P.O. Bishop, and J.D. Pettigrew (1968) Analysis of retinal correspondence by studying receptive fields of binocular singleunits in cat striatecortex. Exp. Brain Res. fi:35%372 ~ ~~~ Otsuka, R., and R. Hassler (1962) Uber Aufbau und Gliederung der Corticalen Sehesphare bei der Katze. Arch. Psychiat. Nervenkr. 203:21%234. Pettigrew, J.D. (1974) The effect of visual experience on the development of stimulus specifity by kitten cortical neurones. J. Physiol. 237: Pettigrew, J.D., M.L. Cooper, and G.G. Blasdel (1979) Improved use of tapetal reflection for eye-position monitoring. Invest. Ophth. Vis. Sci. 18t Shatz, C. (1977a) A comparison ofvisual pathways in Boston and Midwestern Siamese cats. J. Comp. Neurol. 171 : Shatz, C. (197713) Abnormal interhemispheric connections in the visual system of Boston Siamese cats: a physiological study. J. Comp. Neurol. 171 :22%246. Shatz, C. (1977~) Anatomy of interhemispheric connections in the visual system of Boston Siamese and ordinary cats. J. Comp. Neurol. 173: Shatz, C.J., and S. LeVay (1979) Siamese cat Altered connections of visual cortex. Science 204: Stone, J., J.E. Campion, and J. Leicester (1978) The nasotemporal division of retina in the Siamese cat. J. Comp. Neurol. 180:78%798. Tusa, R.J., L.A. Palmer, and A.C. Rosenquist (1978) Retinotopic organization of area 17 (striate cortex) in the cat. J. Comp. Neurol. 177:21%236. Tusa, R.J., A.C. Rosenquist, and L.A. Palmer (1979) Retinotopic organization of areas 18 and 19 in the cat. J. Comp. Neurol