Relation between Patterns of Intrinsic Lateral ...

Relation between Patterns of Intrinsic Lateral Connectivity, Ocular Dominance, and Cytochrome Oxidase-Reactive Regions in Macaque Monkey Striate Cortex

Takashi Yoshioka,1 Gary G. Blasdel,2 Jonathan B. Levitt,3 and Jennifer S. Lund3

To help understand the role of long-range, clustered lateral connections in the superficial layers of macaque striate cortex (area VI), we have examined the relationship of the patterns of intrinsic connections to cytochrome oxidase (CO) blobs, interblobs, and ocular dominance (OD) bands, using biocytin based neuroanatomical tracing, CO histochemistry, and optical imaging. Microinjections of biocytin in layer 3 resulted in an asymmetric field (average anisotropy of 1.8; maximum spread—3.7 mm ) of labeled axon terminal clusters in layers 1-3, with the longer axis of the label spread oriented orthogonal to the rows of blobs and imaged OD stripes, parallel to the V1/V2 border. These labeled terminal patches (n = 186) from either blob or interblob injections (a = 20) revealed a 71% (132 out of 186) commitment of patches to the same compartment as the injection site; 11% (20 out of 186) to the opposite compartment and 18% (34 out of 186) to borders of blob— interblob compartments, indicating that the connectivity pattern is not strictly blob to blob, or interblob to interblob (p < 0.005; \z). In injections placed within single OD domains (n = 11), 54% of the resulting labeled terminal patches (43 out of 79) fell into the same 0D territories as the injection sites, 28% (22 out of 79) into the opposite 0D regions, and 18% (14 out of 79) on borders, showing some connectional bias toward same-eye compartments (p < 0.02; AN0VA). Individual injection cases, however, varied in the degree (50-100% for CO patterns, 22-100% for 0D patterns) to which they showed same-compartment connectivity. These results reveal that while connectivity between similar compartments predominates (e.g., blob to blob, right eye column to right eye column), interactions do occur between functionally different regions.

noted that conspicuous interactions between different channels and modules occur in the visual cortex mediated by both excitatory and inhibitory neurons (Ts'o and Gilbert, 1988; Lund and Yoshioka, 1991; Kritzer et al., 1992; Malach et al., 1994). In this study, we further explore the commitment of anatomical connectivity patterns to two cortical modules—cytochrome oxidase (CO>rich and CO-poor compartments—as well as ocular dominance domains as visualized by functional imaging. Since these two compartments differ markedly in their geometry across the superficial layers, we were anxious to determine if one or other feature might take a predominant role in the distribution of connections, and thus expand upon the observations of Malach et al. (1993). The division of the superficial layer territory into CO-rich blobs and CO poor interblob regions is a prominent feature of its organization; the blobs lie in register with the OD bands of layer 4 (Horton and Hubel, 1981). The blobs receive direct input from a neurochemically distinct (CaM II kinase positive) population of the LGN cells found in the intercalated layers (Fitzpatrick et al., 1983; Hendry and Yoshioka, 1994; see Casagrande, 1994, for review) as well as relays from layers 4B and 4A (Yoshioka et al., 1994). Neurons in blobs and interblobs show certain physiological differences, cells in the blob territory are largely monocular, orientation-nonspecific, colorspecific, and show high contrast sensitivity, whereas cells in the interblob regions are mainly binocular, orientation specific, color-nonspecific, and possess high contrast sensitivity (Livingstone and Hubel, 1984a). Another robust functional characteristic of visual input to the cortex is the separation of information transmitted from each eye. Relays from the dorsal lateral geniculate nucleus (LGN) to the primary visual cortex establish topographically segregated terminal zones in layer 4C of area VI where left and right eye relays terminate in alternating stripe-like OD bands (Hubel and Wiesel, 1969; LeVay et al., 1975). Binocular interaction begins within layer 4C, particularly in 4C alpha, but becomes more prominent as visual information is relayed from layer 4C to the more superficial layers 4B and 3B where substantial binocular interactions take place (Wiesel, 1982). Despite considerable further binocular interaction in the superficial layers following relays from layers 4B and 3B to layers 1-3A, dominance of one or other eye is seen in the majority of recorded neurons in the superficial layers (Livingstone and Hubel, 1984a), and OD banding, in vertical alignment with the layout of thalamic inputs in layer 4C, can still be detected in the activity patterns of the neurons in layers 2 and 3 using 2-deoxyglucose (2-DG; Hendrickson and Wilson, 1979;Tootell et al., 1988) or optical imaging techniques (Biasdel and Salama, 1986; Bartfeld and Grinvald, 1992; Blasdel, 1992a,b; Malach et al., 1993). In New World monkeys such as owl and squirrel monkeys, however, ocular segregation is minimal in the superficial layers of VI as well as in V2, and the contribution of ocularity in forming organized connectivity therefore appears to be negligible in these species (Malach et

Cerebral Cortex Mar/Apr 1996;6:297-310; 1047-32ll/96/$4.00

Downloaded from cercor.oxfordjournals.org by guest on July 14, 2011

The superficial layers 2 and 3 of macaque VI are characterized by an extensive connection system of laterally spreading axon collaterals, arising from pyramidal neurons in the same layers and establishing repetitive patchy terminal fields (Rockland and Lund, 1983; Lund et al., 1993)- Previous studies, using cross-correlation analyses (Ts'o et al., 1986; Ts'o and Gilbert, 1988) and combined biocytin connectional anatomy and functional imaging (Malach et al., 1993), suggest a predominance of connections between cortical regions preferring similar stimulus orientation. Another feature of the superficial layers is the pattern of CO-rich blobs, which mark the termination sites of thalamic inputs from the intercalated layers of the dorsal lateral geniculate nucleus (LGN), and which possess the same repeat distance as previously described latticelike connections arising from pyramidal neurons (Rockland and Lund, 1983); it has also been shown that lateral connections preferentially link blobs together and interblobs together, less frequently linking blob to interblob regions (Livingstone and Hubel, 1984b; Lund et al., 1993; Malach et al., 1993). A different geometry characterizes eye-specific domains in the superficial layers, which form stripe-like patterns apparently in register with the pattern of ocular dominance columns derived from geniculocortical afferents in layer 4C (Hubel and Wiesel, 1969; Hendrickson and Wilson, 1979; Horton and Hubel, 1981; Blasdel and Salama, 1986). Although the clear segregation of functional and anatomical modules has been emphasized in these previous studies, it has also been

•Krieger Mind/Brain Institute, Johns Hopkins University, Baltimore, Maryland 21218, department of Neurobiology, Harvard Medical School, Boston, Massachusetts 02115, and 'Department of Visual Science, Institute of Ophthalmology, University of London, London EC IV 9EL, United Kingdom

1 cm Figure 1. Lateral view of macaque brain showing the left hemisphere. Opercular region of area VI is surrounded by lunate sulcus (iS) and inferior occipital sulcus (/OS), and the area outlined by a broken line represents the region where optical recordings and/ or biocytin injections were made.

Materials and Methods

Two macaque (JM. nemestrind) monkeys were used for both optical recordings and neuroanatomical tracing, and three additional monkeys (M. fasdcularis) were used solely for anatomical experiments in the current study. Opercular regions of area VI, representing predominantly the lower visual field, were used for both physiological and anatomical studies (Fig. 1). Biocytin Injections The methods employed in making extracellular biocytin injections to visualize patterns of intrinsic connectivity in macaque visual cortex have previously been described in detail (Yoshioka et al., 1992b). In brief, microinjections (n = 25; 120 (un average core diameter) of 4% biocytin (in 0.9% saline) were made iontophoretically through glass micropipettes (6-15 u,m tip inner diameter) by passing an anodal current of 3-6 uA for 3-15 min (7 sec ON/7 sec OFF) in layers

238 Lateral Connectivity of Macaque VI Cortex • Yoshioka et al.

Correlation of Biocytin Label with CORicb and -Poor Territories in Layers 2-3 We mapped the resulting labeled fibers, terminal arbors, retrogradely labeled neurons (Fig. 2) and CO-rich and -poor zones through a light microscope drawing tube at both low and high power. Tracings from tangential serial sections were aligned by matching blood vessels and tissue landmarks such as sulci. We defined the injection sites as a core zone where a dense biocytin deposit made recognition of cell or fiber structure impossible. There is no abrupt boundary to CO blob compartments; a transition zone of decreasing CO density merges into the surrounding CO-poor regions. The intensity and size of the CO-rich blobs also depends to some extent on the reaction time during histological processing. The position of the boundaries between CO-rich and -poor zones was mapped by two methods, both of which assigned a firm border. One method involved using a computer image analysis program (NIH Image) to draw boundaries based on the mean gray scale pixel value between blob and interblob territories; in the second, boundaries were drawn directly from the microscope image independently for the same tissue sections by at least two observers. These two methods were compared and found to give very similar results. Patches were identified by the presence of axon terminal branches that constituted a cluster of densely labeled processes at the same position in a number of consecutive tissue sections (Fig. 2). These terminal branches bore boutons, and often ran perpendicular to the axons emanating radially from the injection site. In mapping patch locations in relation to blob and interblob compartments, a patch was considered to fall inside a compartment when more than 67% of its area fell within the compartment; when the patch was identified as falling in neither blob nor interblob regions, it was defined as a patch on the blob/interblob edge. Most injections were spaced adequately to assure that labeled clusters were not overlapped. When clusters from two injections were situated closely, however, the origin of the axon terminals was discerned by examining the orientation of radiating fibers between a particular injection site and the clusters under light microscopy with medium- to high-power objective lenses. In general, the connectivity was examined by using more than three tissue sections per injection to help obtain precise origin-to-termination relationships. Cytochrome oxidase-rich structures were also traced from a few tissue sections in each injection, and were compared with biocytin patches traced from the sections most adjacent to the CO reacted sections. The use of multiple CO-reacted sections minimized possible alignment errors resulting from different blob outlines at different tissue depths. Optical Recording Procedures for Ocular Dominance Column Slapping The detailed methods of optical imaging have been previously described (Blasdel and Salama, 1986; Blasdel, 1992a,b). In short, the animals' anesthesia was induced with a mixture of ketamine and xylazine, and maintained by pentothal, supplemented by ventilation with a 2:1 mixture of nitrous oxide and oxygen through an intubation tube. During physiological recordings, animals were paralyzed with pancuronium or vecuronium bromide to stabilize the eyes. Following craniotomy and implantation of a recording chamber (25 mm diameter), the dura was opened and a 0.1% solution (in saline) of voltage-sensitive dye NK2367 (Nippon Kankoh-Shikiso Kenkyusha, Ltd., Okayama, Japan) was superfused to the cortical surface through

Downloaded from cercor.oxfordjournals.org by guest on July 14, 2011

al., 1994). In the cat, Lowel and Singer (1992) showed no correlation between superficial layer patchy lateral connections and single eye dominance; however, rearing kittens with artificially induced strabismus or with alternating monocular occlusion showed that these connections could be forced into purely monocular links during development (similar results are shown in monkey VI by rearing the animal with alternating monocular occlusion—see Burkhalter and Tychsen, 1993). In the present study, we were interested to determine whether the lattice of patchy lateral connections arising from pyramidal neurons are also formed without regard for ocularity in the normal adult monkey, and whether the connectivity is governed solely by differential characteristics of blobs and interblobs. The results of this study show the connections of CO blobs and interblobs to be less segregated than previously suggested (Livingstone and Hubel, 1984b), with approximately one-third of the connections linking blobs with interblob or blob edge regions. The elongated spread of the field of terminal clusters around single small injection sites appears to relate to the anisotropy of the cortical retinotopic map, and to link regions topographically related to the minimum response receptive field of neurons at the injected point, rather than extending far enough to include regions beyond the classical receptive field. The patterns of connectivity relative to the images of ocular dominance show that while connections provide ample substrates for interocular connectivity, they favor single ocular dominance domains. These results are in agreement with and expand upon observations in a recent report by Malach et al. (1993). A brief report of the present study has appeared previously (Yoshioka et al., 1992a).

2-3. Following a postinjection survival time of 2-24 hr, the animals were transcardially perfused with phosphate buffered 4% paraformaldehyde under deep anesthesia. The visual cortical regions containing biocytin injections were sectioned tangentially (parallel to the pia) on a freezing microtome at 40-50 u,m thickness. Prior to the sectioning, the tissues were flattened by pressure applied with a glass slide during the freezing process. As the bottom of the tissue was curved away in a manner that the bottom plane became parallel to the pial plane, the tissue distortion due to flattening process was minimal. Radial blood vessel patterns allowed verification of alignment of serial sections in cortical depth. Alternate sections were reacted for biocytin (Horikawa and Armstrong, 1988; King et al., 1989) or cytochrome oxidase (Wong-Riley, 1979). Occasionally, both biocytin and cytochrome oxidase (CO) reactions were performed on single sections to correlate the biocytin label with CO-rich and -poor zones directly in the same section (Lachica et al., 1991).

v

Downloaded from cercor.oxfordjournals.org by guest on July 14, 2011

Rgure 2. A Biocytin label following an iontophoretic injection into layer 3. Injection site (center) surrounded by orthogradely labeled terminal patches. Sections are cut parallel to the pial surface, two terminal patches are arrowed'{1A these are shown at higher power in B (patch 1) and C(patch 2). Arrow3 in A indicates a retrogradely labeled cell shown in D at higher magnification. Scale bar in ft 200 p-m for A 50 p-m for B and C, 25 \im for D.

Cerebral Cortex Mar/Apr 1996, V 6 N 2 299

the recording chamber. To visualize the ocular dominance of particular regions of the cortex, drifting grating patterns (1.5°/sec) were presented to the animals on a 20-inch TV monitor (Mitsubishi, CA) placed 1.5-2.0 m away, with alternate eye stimulation controlled by an electronically operated shutter. The average screen luminance and contrast were set to 3 cd/mm2 and 80%, respectively. The patterns of neuronal activity were captured by a video camera (COHU model 5300) as the light reflectance of the voltage-sensitive dye changed in the visually activated tissue regions. Images for each eye were subtracted with each other to generate differential images showing a response preference to the right eye (depicted by dark regions in Fig. 3D) or the left eye (light regions in Fig. 3D). The images are due to changes in the cortical tissue activity down to approximately 500 mm in depth (Blasdel and Salama, 1986).

Procedure for Aligning Injection Sites, Intrinsic Patches, and Ocular Dominance Maps

Statistical Analysis Chi-square analyses were performed to examine the relationship between patches and CO compartments against a hypothesis of strict blob-to-blob or interblob-to-interblob connections, with all blob or interblob injections pooled as a group. The relationship between patches and OD columns was analyzed by ANOVA (analysis of variance, one way). Yates's correction was applied to x2 test for cells with a sample number less than 10 (Bruning and Kintz, 1977). Results

A total of 25 microinjections of biocytin (see Table 1) were placed in the opercular region of area VI (outlined region in

300 lateral Connectivity of Macaque VI Cortex • Yoshioka et al.

Relationship between Patterns of Lateral Connectivity and CO Blob and Interblob Regions The relationship between biocytin-labeled patches and CO blob and interblob compartments was examined (Figs. 4-6). Injections placed in interblob regions yielded patches of terminal label largely in other interblob regions; one such interblob injection (Fig. 4; shown in details in Fig. 5A as a camera lucida tracing) labeled 17 distinguishable terminal clusters of which 12 patches (71%) were identified within interblob regions, 3 patches (17%) in blob borders, and 2 patches (12%) inside blobs. For the injections placed in blob territories, we saw a similar connectivity pattern; one such example is shown in Figure 6 with a blob injection yielding 12 identifiable patches (Fig. 6A). Out of 12 patches, 8 patches (67%) were identified to be inside blobs, 2 patches (17%) on blob borders, and 2 patches (17%) in interblob territories. Of the 25 injections made, 9 fell within blobs, 11 within interblobs, and 5 on edges between these compartments (Table 1). Although the majority of the patches of transported biocytin •were found in the same CO compartment as the injection site, confirming the observations previously made by Livingstone and Hubel (1984b), the percentage of connectivity between the same compartments varied from 100% to less than 50%, depending on the individual injections (Fig. 7, Table 1). In Figure 1A, all 9 blob injections and all 11 interblob injections are illustrated; this shows the variation in connectivity among individual injection sites made in the same CO compartment. While there may be some imprecision in determining the borders of blobs and patches, some of the terminal fields were clearly in opposite compartments to the injection site, and the locations of the population of patches around each injection generally included every compartment (Figs. 4-6). Overall, approximately one-third of the population

Downloaded from cercor.oxfordjournals.org by guest on July 14, 2011

Following the optical recordings, biocytin was iontophoretically injected within the optically imaged tissue regions (see above for the details of injections). The locations of the total 15 biocytin injections were superimposed on five video-imaged maps of ocular dominance (OD) by using blood vessel patterns on the tissue surface. The injection sites were first mapped onto a video). Retrogradely labeled cells were typically found in layers 2-3, with most in the lower part of layer 3, and were concentrated largely, but not exclusively, within the terminal patches. The most complete retrograde dendritic labeling occurred in cells with large somata as opposed to neurons of small somata where only the cell body and proximal dendrites were labeled. We restricted our analyses to anterograde labeling because biocytin travels more reliably in the anterograde direction, particularly in fine fibers. The total spread of label around single injection sites in the superficial layers of the opercular cortex measured up to 3.7 mm (range 1.6-3.7 mm, averaging 2.3 mm across the entire array of labeled patches) following our iontophoretic injections (Fig. 2/1). The number of labeled patches varied between 3 and 17 (average 10) for injections of core size between 60 and 210 mm (average 120 mm) (see Table 1). The size of patches varied slightly, depending on their distance from the injection site and the size of the injection. The patches were about 100-200 |j,m diameter in smallest injections or even in larger injections when the patch locations were further from the injection site, whereas they were 150-250 u,m diameter in larger injections or when they were closer to the injection site. Regardless of the size of the injection, the diameter of labeled terminal patches was consistently no greater than about 250 u.m (Fig. 2).

Downloaded from cercor.oxfordjournals.org by guest on July 14, 2011

I

•

"

.

.

-

'

"

•

:

Rgure 3. Illustrations of the procedures for locating injection sites in video-imaged ocular dominance maps. A Injection sites (asterisk^ that were marked on a blood vessel image taken during optical recordings; the positions of these injections can be identified in the surface section containing the vasculature pattern postmortem (6). Injection sites are more easily identified in the deeper section C (a/roivs). By superimposing the surface blood vessel pattern [rectangle in 6) from the histological section onto the imaged blood vessel pattern [A), it can be confirmed that injection sites found in the sections were at the same locations as those marked on the imaged vasculature pattern during the experiment The locations of these injection sites {asterisk^ and labeled patches traced by camera-lucida were transferred onto a video-imaged ocular dominance map (D). Scale bars: A and D, 1 mm; 6 and C, 2 mm.

Cerebral Cortex Mar/Apr 1996, V 6 N 2 301

Table 1 lontophoretic biocytin injections (n == 25) placed in layers 2-3 of area VI, some with optically imaged ocular dominance data In = 16)

Animal/ hemisphere DR DR DR DR DR DR DR DL CL

ei

1 2 3 4 5 6 8 2 1 2 3 4

5 8 9 10 17 18 19 21 1 2 3 5 6

Injection core diameter (urn)

Number of patches'

CO status of injection site

180 100 150 150 150 60 130 110 120 150 80 80 100 100 80 100 150 170 210 150 170 110 100 140 120

9(9) 6(6) 8(7) 5(5) 9(6) 3(1) 10(10) 6(6) 12(12) 1019) 9(9) 6(6) 12(121 10(7) 7(7) 9(6) 17 12 17 16 14 7 14 9 13

blob blob interblob interblob blob interblob blob blob blob interblob blob interblob blob interblob interblob interblob edge edge interblob interblob edge edge edge blob interblob

% Patches in same CO compartment*

Ocular dominance of injection site (right or left eye)

% Patches in same ocular dominance as injection site

67(6/9) 83(5/6) 75(6/8) 80(4/5) 78(7/9) 100(3/3) 50(5/10) 50(3/6) 67(8/12) 70(7/10) 56(5/9) 83(5/6) 75(9/12) 70(7/10) 86(6/7) 89(8/9) 29(5/17) 33(4/12) 71 (12/17) 56(9/161 21 (3/14) 43(3/7) 29(4/14) 89(8/9) 69(9/13)

R L L L R L R R R R/L L R R/L R R/L R/L no data no data no data no data no data no data no data no data no data

22 83 57 _< 67 100 60 . 33 50 0 56 33 17 86 14 17 no data no data no data no data no data no data no data no data no data

' Number in parentheses indicates the number of patches within the optically recorded regions. ' Calculations were made strictly between same CO compartments, and did not divide the possible contributions of CO edges into half blobs and half interblobs. 'This injection was not used for ocular dominance analysis due to the poor image quality in this tissue region.

of patches (Fig. IB) lay outside the injected compartment; of patches from blob injections, 68% (56 out of 82) were found in other blobs, 12% (10 out of 82) in interblobs, and 20% (16 out of 82) lay on blob/interblob borders. Similarly, 73% (76 out of 104) of patches from the interblob injections were found in other interblob territories, and 27% (28 out of 104) fell either within blobs or on blob/interblob edges. The patches resulting from injections placed on the borders of blobs with interblobs lay predominantly in interblob and edge positions (Fig. 1B\ 30%: 19 out of 64, 53%: 34 out of 64, respectively). The observed biases for injections in particular CO compartments to preferentially label patches in the same type of compartment (68% blob-blob or 73% interblob-interblob) were statistically different from a prediction of connections strictly between the same type of compartment (blob injections:^ < O.OO5, X2 = 8.8, df = 1; interblob injections:^) < 0.0O5, X2 = 8.6, df = 1). During this analysis, patches that were at least 67% within a compartment -were assigned 100% to that compartment. We compared this method with another approach where we measured exactly the area of overlap between patches and injection site and CO-rich and -poor compartments. For the interblob injection illustrated in Figure 5 using exact areal measures, 95% of the injection site fell within an interblob region, while the proportion of patch area that fell into interblob areas was 69% (i.e., 72% interblob-to-interblob connectivity from 95% interblob injection site), very close to our figure of 71% interblob-interblob projections for this injection using 67% overlap as criterion for 100% assignment to a compartment. Similarly, for the blob injection of Figure 6 (100% of whose injection site fell within the blob region), 64% of the area occupied by terminal patches fell within blobs when measured exactly. This is comparable to the 67% blob-blob connectivity originally determined for this injec-

302 Lateral Connectivity of Macaque VI Cortex • Yoshioka et al.

tion. Since there is a genuine imprecision in determining the borders of blob-interblob compartments, we have relied on our method of assigning patches of over 67% occupancy 100% to compartments, rather than measuring precise areal occupancy within each CO compartment, which we feel provides a false sense of precision without actually increasing accuracy. Relationship between Patterns of Lateral Connectivity and Ocular Dominance Images Fifteen injections were made within the regions where ocular dominance maps were obtained through optical recordings. Seven injections fell into right eye columns, four injections into left eye columns, and four injections •were on the borders of eye columns. Subsequent Figures 8 and 10 illustrate some of these injections, and quantitative analyses of all injections are listed in Table 1. Figure 9 summarizes the projection patterns of these injections. In relation to ocular dominance stripes, some injections showed an almost equal distribution of terminal patches in either right (darker regions) or left (lighter regions) eye columns, or on the border regions (e.g., injection 3 in Fig. 8, which is DR3 in Fig. 9.4), while other injections showed projection distributions skewed toward one eye (e.g., injection 2 in Fig. 8, which is DR2 in Fig. 9A). Figure 9A summarizes results of five injections placed in right eye columns and three injections into left eye columns. The right eye column injection CL8 in Figure 9A shows one example of an injection that exhibited a high percentage of connectivity between same eye columns; in this injection, six out of seven patches were found in the right eye columns (86%), and one patch was found on the border of eye columns (14%). In contrast, only 22% of the terminal projections of injection DR1 were to the same ocular dominance compartment as that injected.

Downloaded from cercor.oxfordjournals.org by guest on July 14, 2011

CL CL CL CL CL CL MK13R MK13R MK13R MK13R MK15L MK15L MK15L MK15L MK18L

Injection identification number

Downloaded from cercor.oxfordjournals.org by guest on July 14, 2011

*

Rgure 4. illustration of the relationship between the patterns of intrinsic connectivity and CO blobs and interblobs. A Tangential section with a biocytin injection placed in an interblob region of layer 3 showing transported label around the injection site. B, Adjacent section to A reacted for CO indicating the injection site in A (asterisk is in an interblob region (see also Fig. 5). Arrows indicate blood vessels used as fiduciary landmarks. Scale bar for A and B, 400 p.m.

Of the 11 injections that fell on either right or left eye columns, 4 injections showed greater than 67% connectivity between the same eye columns, 4 injections ranged from 34 to 66% same eye connectivity, and 3 injections had lower than 33% same-eye connectivity. The overall percentage of connectivity between same eye columns ranged from 22% to 100% (Table 1). However, for these 11 injections that were placed in either right or left eye columns, the average percentage of projection patches within the same eye territory as the injection locus was 54% (43 out of 79 patches), versus 28% (22 out of 79) of projection patches that fell into the opposite eye territories, and 18% (14 out of 79) on the borders. An analysis of variance shows this bias to same eye connectivity to be significant (p = 0.015). If we simply divide the contribution of border patches into 50% toward right eye columns and 50% toward left eye columns, the percentage of the sameeye connectivity becomes 63% and that of the opposite-eye connectivity becomes 37%. In either case, the percent connectivity between the same-eye columns is about twice that connectivity between opposite-eye columns. The overall distribution of patch locations to same-, opposite-, and mixed (i.e., borders of different eye columns>eye dominance regions is shown in Figure 9B, which illustrates the predominance of connectivity between the same-eye columns when injections

Rgure 5. A, Camera-lucida drawing of biocytin injection shown in Figure 4, and its labeled terminal patches [solid lined). B, Map of biocytin patches [filled regions) in relation to CO blobs (outlined by dashed lines) from the injection drawn in A Note that many patches (12 out of 17) are in interblob territories as is the injection site [hatched area in B), but some are either in blobs or on blob edges. Arrows indicate blood vessels used as fiduciary landmarks for alignment Scale bar for A and B, 200 tun.

were placed in either right or left ocular dominance columns. It should be noted that injections placed on the borders of the two ocular dominance stripes (Fig. 9B, Table 1; n = 4) showed a nearly equal distribution of patches into either right- or left-eye columns (47,41%, respectively) with a small percentage of patches (12%) falling onto the borders between columns. Furthermore, this proportion of projections to ocu-

Cerebral Cortex Mar/Apr 1996, V 6 N 2 303

Blob injections

Interblob injections

PARTM

100 -• id 9 0 -

S

8

8070605040302010-

1 1

1

O-rMt !

1 1 1

1

1

1

1

1

ut

INJECTION CASE (n=20)

100 r, 90 | 801 70 60 50 40 30 20 10 0 •

PATCH LOCATION

1

o

?'

*0I I o/

.rv

Figure 6. Another example of reconstruction of a biocytin injection. This injection fell in a CO blob territory and most patches (8 out of 12) were found in blob regions. Conventions same as in Figure 5. Scale bar, 200 ^m.

lar dominance borders differed little from injections located within either right or left eye stripes. This suggests that the border position does not constitute a distinct compartment.

•

Blob

B

|

y i1 Interblob

Blob-patch

H lnterblobpatch •

Edge-patch

—y

Edge

BIOCYTIN INJECTION SITE (n=25)

Figure 7. Summary histograms showing the relationship between patterns of intrinsic connectivity and CO blobs and interblobs. A, Histograms of all blob (n = 9) and interblob (n = 11) injections illustrating the range of distribution of patches that fell into the same CO compartment as the injection site. B, Histogram showing distribution of patches into blob, interblob, and edge territories for all injections pooled into categories of blob, interblob, and edge injections.

Discussion

Anisotropy of Biocytin Patch Distribution The overall field of labeled patches around each injection site (summarized in Table 2) was, in general, elongated orthogonal to the ocular dominance stripes (e.g., injections 1, 2, 3, and 5 in Fig. SB). An example of the relationship between the elongation of patch spread and ocular dominance stripes is shown in Figure 10 with a video-imaged ocular dominance map and a group of biocytin injections (see injections 1, 2, 3, 5, and 10). In addition to the orthogonal relationship between the axes of patch spread and ocular dominance stripes, where the injections were near the V1/V2 border, the spread ran parallel to the lunate sulcus (LS) and to the V1/V2 border (Fig. 10). Since the V1/V2 border represents the vertical meridian of the visual field, the patch spread also runs parallel to the vertical meridian at least for injections near the VI-V2 border. The elongation of spread of biocytin label was quantified for 25 injections by measuring the distance between the furthest patches through the injection site; the distance of the patches in the orthogonal direction was also measured to calculate the anisotropy ratio (= long axis/short axis). The average ratio for anisotropy of label spread was 1.78 (range 1-4.6, standard deviation 0.68), and the median anisotropy ratio was 1.6 (Fig. 11). Our injections were all placed well within the central 5° on the VI operculum, and 81% of the injections (13 out of 16 injections with ocular dominance maps) showed the anisotropy of projections orthogonal to ocular dominance stripes.

304 Lateral Connectivity of Macaque VI Cortex • Yoshioka et al.

The network of clustered lateral connectivity around each point on the cortex is a phenomenon seen most prominently in layers 1-3 of macaque VI as well as in layers 4B and 5 (Lund et al., 1993). Similar clustered patterns have been observed in many other cortical areas in our own studies (Yoshioka et al., 1992b; Levitt et al., 1993, 1994; Lund et al., 1993), and in studies by others (Callahan and Haberly, 1987; Gilbert and Wiesel, 1983,1989; Juliano et al., 1990; Huntley and Jones, 1991; Amir et al., 1993; Malach et al., 1994). Understanding the role of this ubiquitous cortical circuitry seems therefore crucial to our understanding of cortical information processing. The results of this study may thus be relevant to more general aspects of cortical function as well as to visual cortical function. It should be remembered that neurons in the compartments of blob-interblob or right- and left-eye ocular dominance stripes have dendrites that freely cross compartment borders (Malach, 1992; Lund et al., 1993). Therefore, the boundaries of the compartments as defined by afferent driving (as in the case of CO blobs or exclusively alternating monocular input in the case of OD domains) may not be a true reflection of the postsynaptic neuron's commitment to a compartment. Rather, the postsynaptic neurons may form a continuum of change in any property across the cortical surface. We would therefore recommend caution at this stage in interpreting these boundaries themselves as functionally meaningful compartments in the superficial layers.

Downloaded from cercor.oxfordjournals.org by guest on July 14, 2011

^..^

••

Figure 8. Biocytin injections (filled circled and transported terminal patches {open circled superimposed onto videoimaged ocular dominance maps (A without ocular dominance column outlines; 8, with ocular dominance column outlines indicated by white lined- Dark stripe regions in ocular dominance maps correspond to right eye columns and pale stripe regions correspond to left eye columns. Outlines separating the eye columns in B were derived from the result of a gradient operator analysis, which yielded vectors that reverse direction of the gray scale gradient as described in Blasdel (1992a). Numbers correspond to injections 1-6 of animal DR (see Table 1). Dashed lines in A separate adjacent injections and patches from each other. A, anterior, P, posterior; M, medial; L, lateral. Scale bar, 1 mm.

Downloaded from cercor.oxfordjournals.org by guest on July 14, 2011

Lateral Connectivity and Wob/Interblob Patterns Livingstone and Hubel (1984b) first described the specificity of connections between CO blob and interblob regions. The present study finds connectivity between like compartments is not absolute, since approximately one-third of the projections traced fall into compartments unlike that of the injection site (with at least 10% of projections to the opposite CO compartment). More importantly for the physiologist, projections from single points on the cortex cannot be predicted, as we have shown considerable variation between the pro-

jections arising from individual injections within particular compartments. The overall patch distributions are significantly different from a strict rule of 100% blob-to-blob, or interblob-to-interblob connectivity. Injections that were placed on the borders of blobs and interblobs vary in degree of connectivity with other border regions (21-43%, n = 5; see Table 1), suggesting that the border territory may not be an independent connectional unit in the neuronal networks in macaque VI. Although previous studies reported border-to-border connectivity of CO patterns in macaque VI (Livingstone

Cerebral Cortex Mar/Apr 1996, V 6 N 2 305

Right eye stripe injections

Left eye stripe injections

INJECTION CASE (n=ll) PATCH LOCATION

Right Eye Left Eye R/L border BIOCYTIN INJECTION SITE (n=15)

Rgure 9. Histograms illustrating relationship between intrinsic connectivity and ocular dominance (OD). A, Illustration of the range of distribution of patches in relation to the ocular dominance of injection sites for all individual injections placed in either right (n = 7) or left (n = 4) eye columns. Numbers above each histogram bin indicate the total number of patches with imaged OD maps that were used for the analyses. 6, Summary histogram of patch locations in relation to ocular dominance columns for all injections (15) including those on OD borders.

and Hubel, 1984a,b), no quantitative analyses were provided to enable a comparison with the present data. We have calculated the representation of interblob regions compared to blob territory as approximately 79% versus 21%. Given this greater areal representation of interblob territory over blob regions, it is clear that the 68% connectivity between blobs observed in the current study is very unlikely to have arisen by chance from a random distribution of patches of terminal label; it is more difficult to prove that the 73% figure of interblob to interblob connections is not the result of a random patch distribution. We have observed a similar phenomenon in area V2 (Levitt et al., 1994), where CO-rich regions are more likely to be interconnected than are COpoor regions. We suggest one factor that may be important in determining the development of these connections is the presence of common thalamic input to the CO-rich regions within each area. In macaque VI, direct geniculocortical afferents terminate at blob regions in the superficial layers (Fitzpatrick et al., 1983; Hendry and Yoshioka, 1994), whereas in macaque V2, pulvinar afferents terminate in both CO-rich thin and thick stripes (Livingstone and Hubel, 1982; Levitt et al., 1995). Because the intrinsic connectivity between CO-rich regions and between CO-poor regions is more prominent

306 Lateral Connectivity of Macaque VI Cortex • Yoshioka et al.

Lateral Connectivity and Patterns of Ocular Dominance Columns Our results demonstrate that clear connectivity between same-eye ocular dominance columns is present in about half (54%) of the intrinsic patches labeled in the superficial layers, whereas unambiguous links between different ocular dominance columns is present in 28% of the connectional patches. A statistical analysis indicates that the neurons in either-eye columns have a connectional bias toward same-eye ocular dominance columns (p = 0.015, ANOVA).This is in agreement with the recent report by Malach and his colleagues (Malach et al., 1993) that the majority of long-range lateral connections in monkey VI are targeted at territory dominated by the same eye. They, too, have examples of patches lying in columns of ocular dominance opposite to that of the injection site. Our data, containing a fairly large number of injection examples (n = 15) with projection patches superimposed on OD imaging maps, show that the projection patterns vary from injection to injection. Some injections show a majority of links between same-eye columns (e.g.,86% in CL8,83% in DR2, see Table 1), while other injections show only a minority of projections between same-eye columns (e.g., 22% in DR1, 33% in CL4, see Table 1). Malach et al. (1993) showed a border injection with projections predominantly to other borders, and concluded that the border zone represents a special binocular compartment. Since we observed no bias of border zones to project to other border zones, we believe it unlikely that the borders represent a special compartment. We suggest that with a bigger sample, the projections from border positions may be found to differ widely from case to case, as is the situation with monocular injections. We do suggest on the basis of results from our border injections, that the projections of borders •will equally favor right and left eye stripes, in contrast to the bias shown to same-eye compartments with injections placed confined to a single-eye stripe (see Table 1, Fig. 9JB). As discussed earlier, however, the overall projection bias toward same-eye columns suggests that ocular dominance information is one of the functional properties that influence the formation of clustered intrinsic connectivity in the superficial layers of macaque VI. This contribution of ocular information to the genesis of clustered connections observed in macaque VI does not necessarily apply to macaque extrastriate visual areas or striate cortex of nonmacaque species. For example, ocular dominance columns are not present in VI of New World monkeys, or in V2 of either macaque or New World monkeys, suggesting that ocular information is unlikely to play any role in determining the organized intrinsic connection pattern present in these visual areas and species. Our observations, and those of Malach et al. (1993), of the connectivity bias toward same-eye columns in macaque VI differ from findings in area 17 of the normal cat, where no bias to eye-specific connections has been observed (Lowel and Singer, 1992). Since ocular dominance columns can be clearly visualized in the superficial layers in both species using 2-DG autoradiography (cat: Fig. 2F of Lowel and Singer, 1992; monkey: Kennedy et al., 1976; Hendrickson and Wilson, 1979), the different findings in the two species is puzzling. A difference between the technical approaches used in the above studies is that the intrinsic connectivity was visualized by retrogradely labeled somata in the cat study (Lowel and Singer, 1992), while anterograde axon terminal labeling was

Downloaded from cercor.oxfordjournals.org by guest on July 14, 2011

B

than connections between different compartments, both in VI and V2 (Livingstone and Hubel, 1984b; Malach et al., 1994; Levitt et al., 1994), differential patterns of activity such as presence or absence of thalamocortical afferent drive may influence the intrinsic connectivity patterns.

RgureiO. A Tangential section showing several biocytin injection sites (indicated by numbers) near the lunate sulcus (LSI in animal CL Rectangle shows the frame video-imaged to obtain the ocular dominance map shown in B. B, Biocytin injections {asterisks] with labeled patches {open ovals) with arrows showing the major and minor axes of spread of label for each injection site. Dashed lines separate adjacent injections and patches from each other. A anterior; P, posterior, M, medial; L, lateral. Scale bar, 1 mm.

m-. • Downloaded from cercor.oxfordjournals.org by guest on July 14, 2011

LS

B used in the monkey studies (Malach et al., 1993; present study). Anterogradety transported biocytin terminal label occurs in more tightly focused clusters than is shown by retrogradely labeled somata (Fig. 2 in the present study; also Lund et al., 1993), which may be due partly to local connections not always being reciprocal (Boyd and Matsubara, 1991). It would be of interest to examine links between OD domains

in cat using orthograde transport of biocytin as employed in the present study. Functional Relationships of Intrinsic Connectivity Lattices Although the OD stripe territories are still preserved in the superficial layers, the majority of neurons in the superficial

Cerebral Cortex Mar/Apr 1996, V 6 N 2 307

Table 2 Spread of labeled patches and their anisotropy of distribution around 25 biocytin injections placed in layers 2-3 of area VI

Animal hemisphere

1 2 3 4 5 6 8 2 1 2 3 4 5 8 9 10 17 18 19 21 1 2 3

5 6

Long axis (mm)'

Short axis (mm)«

2.9 10 1.7 2.1 3.0 13 14 1.5 12 1.9 14 1.8 1.9 3.7 1.4 19 15 17 13 14 13 1.6 19 10 10

1.7 U 1.2 U 11 13 1.8 1.0 1.5 U 1.4 1.2 1.2 0.8 0.6 1.2 1.5 1.5 13 1.4 1.5 1.2 1.2 1.0 1.1

Long/ short

Orthogonal to 0D columns?

Along V1/V2 border?

1.7 1.5 1.4 1.6 1.4 1.8 13 1.5 1.5 1.5 1.7 1.5 1.6 4.6 13 14 1.7 1.8 1.0 1.7 1.5 13 14 10 1.8

Yes* Yes Yes Yes Yes No Yes Yes Yes Yes Yes No Yes Yes No Yes Yes No No No Yes Yes Yes Yes No

Yes' No Yes Yes Yes Yes No Yes Yes Yes Yes No Yes Yes No Yes Yes No No Yes Yes Yes Yes Yes No

•The long axis of the patch spread was the distance (mm) between the furthest labeled patches measured across the injection site; the short axis was the distance between furthest patches measured orthogonal to the long axis and passing through the injection site. ' Yes indicates that the orientation of patch spread was within ±30° of that of either ocular dominance columns or V1/V2 border. The orientation of ocular dominance columns was obtained by optical imaging in the first 16 injections (animals D and C), and by the orientation of blob rows in the remaining 9 injections (animals MK 13-18).

layers receive inputs from both eyes (Hubel and Wiesel, 1968; Wiesel 1982), in some cases, being tuned for binocular disparity (Barlow et al., 1967; Poggio and Fischer, 1977). The clustered long-range lateral connections observed in the present study arise almost exclusively from pyramidal cells (Rockland and Lund, 1982, 1983; Gilbert and Wiesel, 1983), and these are known to target dendritic spines on other pyramidal cells, as well as making a smaller proportion of contacts on local circuit GABAergic neurons (Rockland, 1985; Kisvarday et al., 1986; LeVay, 1988; McGuire et al., 1991). It is suspected that the ocular balance of cells in the superficial layers is created by activation of local inhibitory links since optically imaged ocular dominance columns disappear when inhibitory synapses are blocked by bicuculline, leaving a uniform binocular drive to superficial layer neurons (Haglund and Blasdel, 1991). Interaction between excitatory and inhibitory influences in the superficial layer lattices also seems to be an important feature of adult plasticity in which changes in receptive field size of cortical neurons takes place within minutes around a cortical scotoma created by bilateral retinal lesions (Gilbert and Wiesel, 1992). Spatial interaction between center and surround regions beyond the classical receptive field has been documented with unit recording techniques and optical imaging (Allman et al., 1990; Gilbert and Wiesel, 1990; Knierim and Van Essen, 1992; Grinvald et al., 1994). In these studies, the simultaneous presentation of stimuli placed outside the classical receptive fields have modified the cell's responses to the optimal stimulus placed within the receptive field. There is no consensus in these studies as to whether the modification of the response is likely to be inhibition or

308 Lateral Connectivity of Macaque VI Cortex • Yoshioka « al.

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 ANISOTROPY RATIO Figure 11. Distribution of anisotropy ratios among 25 injections. Anisotropy is the ratio between the measures of longer and shorter (orthogonal) axes of patch spread. Largest value was 4.6, smallest value was 1, and the median value was 1.6.

enhancement, and whether the same or orthogonal orientation between center and surround stimuli causes inhibitory effects. These long-range lateral interactions are seen in realtime imaging of cortical responses as a form of "cortical point spread function" with the spread "space constant" of 1.5 X 2.7 mm anisotropic fields (Grinvald et al., 1994). These observations suggest that lateral connectivity maintains a dynamic pattern of properties across the cortex that can change according to particular conditions of activation across the network as a whole. We have seen from the results of this study that blobs, which are the most monocular regions of the superficial layers as tested by unit recording (Livingstone and Hubel, 1984a), have variable connectivity and can project equally to regions of either-eye dominance or entirely to the same-eye territory (Table 1). A relatively weak connectivity bias toward same-eye columns in this study (p = 0.015, ANOVA) suggests that the connections made across the cortex are not structured to reinforce only monocular inputs. The commitment to the same-eye columns (54%) was lower than the 68-73% connectivity seen between the same CO territories (Fig. 7.8). This may suggest that the connectivity due to ocular dominance properties is emphasized in only a portion of lattice connections, and other functional features such as orientation, color, texture, or motion may also participate in determining the overall structure of cortical circuitry, and indeed, connectivity between neurons of similar orientation has been observed (Ts'o et al., 1986; Gilbert and Wiesel, 1989; Blasdel et al., 1992; Malach et al., 1993; Blasdel et al., unpublished observation). Another factor that could help determine the higher percentage of connectivity between same CO compartments is the different afferent drive of the compartments— direct thalamic input to blobs by a population of neurons in the LGN interlaminar layers (Fitzpatrick et al., 1983; Hendry and Yoshioka, 1994), and relays from 4C in the case of the interblob territory (Yoshioka et al., 1994; see also, Lachica et al., 1992). The physiological properties or high activity of the

Downloaded from cercor.oxfordjournals.org by guest on July 14, 2011

DR DR OR DR DR DR DR DL CL CL CL CL CL CL CL CL MK13R MK13R MK13R MK13R MK15L MK15L MK15L MK15L MK18L

Injection identification number

direct LGN pathway may exert a particularly strong influence on the formation of intrinsic circuitry.

Summary We conclude from these experiments that the intrinsic lattice connections of local clusters of neurons in superficial VI are influenced by multimodal properties, including retinotopic information, with a bias toward reinforcing properties of the local cluster. Notes

We are grateful for the support from Drs. Stewart Hendry and David Lewis during the course of this study. We also appreciate many valuable comments on the manuscript from Drs. Stewart Hendry and Vernon Mountcastle, and the assistance on statistical analyses from Dr. Michael Steinmetz. This work was supported by NEI Grants EY05282, EY10021, and EYO54O3; MRC Grant MRC9203679N, HFSPO Grant RG-98/94 B; and in part by EY08098 (Eye & Ear Institute of Pittsburgh), an ARVO/ALCON fellowship (T.Y.), and postdoctoral NRSA Award EY06275 (J.B.L.). Address correspondence to Dr. Jennifer S. Lund, Department of Visual Science, Institute of Ophthalmology, University of London, 1143 Bath Street, London EC1V 9EL, UK. References

AUman J, Miezin F, McGuinness E (1990) Effects of background motion on the responses of neurons in thefirstand second cortical visual areas. In: Signal and sense: local and global order in perceptual maps (Edelman GM, Gall WE, Cowan WM, eds), pp 131-142. New York: Wiley. Amir Y, Harel M, Malach R (1993) Cortical hierarchy reflected in the organization of intrinsic connections in macaque monkey visual cortex. J Comp Neurol 334:19-46.

Cerebral Cortex Mar/Apr 1996, V 6 N 2 309

Downloaded from cercor.oxfordjournals.org by guest on July 14, 2011

Anisotropic Projection of Pyramidal Neurons We have observed anisotropy in axonal projections from our injections predominantly in a direction parallel to the V1/V2 border and lunate sulcus, and in a direction orthogonal to ocular dominance stripes and CO blob rows (Tigs. 7, 9, 10). Anisotropy of visual field representation in macaque VI has been previously reported (Van Essen et al., 1984; Dow et al., 1985), and our median anisotropy ratio of 1.6 is somewhat reminiscent of the 1.5 physiologically obtained value in foveal regions (Dow et al., 1985), in which more cortical tissue is allocated to representation along the vertical meridian than along the horizontal meridian. Other recent studies using optical imaging also show a similar cortical anisotropy in macaque VI with the mean anisotropy 1.69 (+ 0.211 SD) in one report (Malach et al., 1993) and 1.8 (2.7 mm along the vertical meridian/1.5 mm along the perpendicular axis) in another report (Grinvald et al., 1994), which are comparable to the present observations (mean ± SD = 1.78 ± 0.68; median = 1.6). It was also noted by Hubel and Wiesel (1977) that visual field representation is duplicated across the ocular dominance bands—but not along them— and this creates an anisotropy of field representation across the ocular dominance bands. The elongation of the visual field representation along the vertical meridian in VI tissue and across ocular dominance bands resembles our observation of anisotropy of intrinsic connectivity in VI. The dimensions of the total field of patches around single injection sites closely matches the territory activated by single point stimuli as recently imaged by Grinvald et al. (1994) in macaque VI. The extent of these connections is such as to suggest links between neurons that share some part of their minimum response fields with neurons at the injection site; but the field does not appear extensive enough to allow much in the way of direct links between neurons at the injection site and regions outside their classical receptive field, though this could be accomplished via multisynaptic connections.

Barlow HB, Blakemore C, Pettigrew JD (1967) The neural mechanism of binocular depth discrimination. J Physiol (Lond) 193:327-342. Bartfeld E, Grinvald A (1992) Relationships between orientation-preference pinwheels, cytochrome oxidase blobs, and ocular dominance columns in primate striate cortex. Proc Natl Acad Sci USA 89:11905-11909. Blasdel GG (1992a) Differential imaging of ocular dominance and orientation selectivity in monkey striate cortex. J Neurosci 12: 3115-3138. Blasdel GG (1992b) Orientation selectivity, preference, and continuity in monkey striate cortex. J Neurosci 12:3139-3161. Blasdel GG, Salama G (1986) Voltage sensitive dyes reveal modular organization in monkey striate cortex. Nature 321:579-585. Blasdel GG, Yoshioka T, Levitt JB, Lund JS (1992) Correlation between patterns of lateral connectivity and patterns of orientation preferences in monkey striate cortex. Soc Neurosci Abstr 18:389BoydJ, MatsubaraJ (1991) Intrinsic connections in cat visual cortex: a combined anterograde and retrograde tracing study. Brain Res 560:207-215. Bruning JL, Kintz BL (1977) Computational handbook of statistics, 2d ed. Glenview, IL: Scott, Foresman. Burkhalter A, Tychsen L (1993) Alteration of intracortical connections in monkeys who lacked normal binocular experience in infancy. Invest Ophthalmol Vis Sci [Suppl] 34:1173. Callahan EC, Haberly LB (1987) An extensive intrinsic association fiber system within cat area 7 revealed by anterograde and retrograde axon tracing methods.J Comp Neurol 258:125-137. Casagrande VA (1994) A third parallel visual pathway to primate area VI. Trends Neurosci 17:305-310. Daniel PM, Whitteridge D (1961) The representation of the visual field on the cerebral cortex in monkeys. J Physiol (Lond) 159: 203-221. Dow BM, Vautin RG, Bauer R (1985) The mapping of visual space onto foveal striate cortex in the macaque monkey. J Neurosci 5: 890-902. Fitzpatrick D, Ito K, Diamond IT (1983) The laminar organization of the lateral geniculate body and the striate cortex in the squirrel monkey (Saimiri sciureus).) Neurosci 3:673-702. Gilbert CD, Wiesel TN (1983) Clustered intrinsic connections in cat visual cortex. J Neurosci 3:1116-1133. Gilbert CD, Wiesel TN (1989) Columnar specificity of intrinsic horizontal and corticocortical connections in cat visual cortex. J Neurosci 9:2432-2442. Gilbert CD, Wiesel TN (1990) The influence of contextual stimuli on the orientation selectivity of cells in primary visual cortex of the cat. Vision Res 30:1689-1701. Gilbert CD, Wiesel TN (1992) Receptive field dynamics in adult primary visual cortex. Nature 356:150-153. Grinvald A, Iieke EE, Frostig RD, Hildesheim R (1994) Cortical pointspread function and long-range lateral interactions revealed by real-time optical imaging of macaque monkey primary visual cortex. J Neurosci 14:2545-2568. Haglund MM, Blasdel GG (1991) Role of GABAergic activity in generation of upper layer ocular dominance columns in monkey striate cortex. Soc Neurosci Abstr 17:1088. Hendrickson AE, Wilson JR (1979) A difference in [14C] deoxyglucose autoradiographic patterns in striate cortex between Macaca and Saimiri monkeys following monocular stimulation. Brain Res 170:353-358. Hendry SHC, Yoshioka T (1994) A neurochemically distinct third channel in the macaque dorsal lateral geniculate nucleus. Science 264:575-577. Horikawa K, Armstrong WE (1988) A versatile means of intracellular labeling: injection of biocytin and its detection with avidin conjugates. Neurosci Methods 25:1-11Horton JC, Hubel DH (1981) Regular patchy distribution of cytochrome oxidase staining in primary visual cortex of macaque monkey. Nature 292:762-764. Hubel DH, wiesel TN (1968) Receptive fields and functional architecture of monkey striate cortex. J Physiol (Lond) 195:215-243. Hubel DH, Wiesel TN (1969) Anatomical demonstration of columns in the monkey striate cortex. Nature 221:747-750. Hubel DH, Wiesel TN (1977) Functional architecture of macaque monkey visual cortex. Proc R Soc Lond [Biol] 198:1-59. Huntley GW, Jones EG (1991) Relationship of intrinsic connections

310 lateral Connectivity of Macaque VI Cortex • Yoshioka et al.

Poggio GF, Fischer B (1977) Binocular interaction and depth sensitivity in striate and prestriate cortex of behaving rhesus monkey. J Neurophysiol 40:1392-1405. Rockland KS (1985) Intrinsically projecting pyramidal neurons of monkey striate cortex: an EM-HRP study. Soc Neurosci Abstr 11: 17. Rockland KS, Lund JS (1982) Widespread periodic intrinsic connections in tree shrew visual cortex (area 17). Science 215:15321534. Rockland KS, Lund JS (1983) Intrinsic laminar lattice connections in primate visual cortex.J Comp Neurol 216:303-318. TootellRBH,HamiltonSL,SilvermanMS,SwitkesE (1988) Functional anatomy of macaque striate cortex. I. Ocular dominance, baseline conditions, and binocular interactions. J Neurosci 8:1500-1530. Ts'o DY, Gilbert CD (1988) The organization of chromatic and spatial interactions in the primate striate cortex.J Neurosci 8:17121727. Ts'o DY, Gilbert CD, Wiesel TN (1986) Relationships between horizontal connections and functional architecture in cat striate cortex as revealed by cross-correlation analysis. J Neurosci 6:11601170. Van Essen DC, Newsome WT, Maunsell HR (1984) The visual field representation in striate cortex of the macaque monkey: asymmetries, anisotropies, and individual variability. Vision Res 24:429448. Wiesel TN (1982) Postnatal development of the visual cortex and the influence of environment. Nature 299:583-591. Wong-Riley MTT (1979) Changes in the visual system of monocularly sutured or enucleated cats demonstrable with cytochrome oxidase histochemistry. Brain Res 171:212-214. Yoshioka T, Blasdel GG, Levitt JB, Lund JS (1992a) Patterns of lateral connections in macaque visual area VI revealed by biocytin histochemistry and functional imaging. Soc Neurosci Abstr 18:299. Yoshioka T, Levitt JB, Lund JS (1992b) Intrinsic lattice connections of macaque monkey visual cortical area V4. J Neurosci 12:27852802. Yoshioka T, Levitt JB, Lund JS (1994) Independence and merger of thalamocortical channels within macaque monkey visual cortex: anatomy of interlaminar projections. Vis Neurosci 11:467-489.

Downloaded from cercor.oxfordjournals.org by guest on July 14, 2011

to forelimb movement representation in monkey motor cortex: a correlative anatomic and physiological study, j Neurophysiol 66: 390-413. Juliano SL, Friedman DP, Eslin DE (1990) Cortico-cortical connections predict patches of stimulus-evoked metabolic activity in monkey somatosensory cortex. J Comp Neurol 298:23-39. Kennedy C, Des Rosiers MH, Sakurada O, Shinohara M, Reivich M, Jehle HW, Sokoloff L (1976) Metabolic mapping of the primary visual system of the monkey by means of the autoradiographic [14C] deoxyglucose technique. Proc Natl Acad Sci USA 73:42304234. King MA, Lewis PM, Hunter BE, Walker OW (1989) Biocytin: a versatile anterograde neuroanatomical tract-tracing alternative. Brain Res 497:361-367. Kisvarday ZF, Martin KAC, Freund TF, Maglocsky ZS, Whitteridge D, Somogyi P (1986) Synaptic targets of HRP-filled layer HI pyramidal cells in the cat striate cortex. Exp Brain Res 64:541-552. Knierim JJ, Van Essen DC (1992) Neuronal responses to static texture patterns in area VI of the alert macaque monkey. J Neurophysiol 67:961-980. Kritzer MF, Cowey A, Somogyi P (1992) Patterns of inter- and intralaminar GABAergic connections distinguish striate and extrastriate (V2, V4) visual cortices and their functionally specialized subdivisions in the rhesus monkey. J Neurosci 12:4545-4564. Lachica EA, Mavity-Hudson EA, Casagrande VA (1991) Morphological details of primate axons and dendrites revealed by extracellular injection of biocytin: an economic and reliable alternative to PHAL. Brain Res 564:1-11. Lachica EA, Beck P, Casagrande VA (1992) Parallel pathways in macaque monkey striate cortex: anatomically denned columns in layer HI. Proc Natl Acad Sci USA 89:3566-3570. LeVay S (1988) Patchy intrinsic projections in visual cortex, area 18, of the cat: morphological and immunocytochemical evidence for an excitatory function. J Comp Neurol 269:265-274. LeVay S, Hubel DH, Wiesel TN (1975) The pattern of ocular dominance columns in Macaque visual cortex revealed by a reduced silver stain. J Comp Neurol 159:559-576. Levitt JB, Lewis DA, Yoshioka T, Lund JS (1993) Topography of pyramidal neuron intrinsic connections in macaque monkey prefrontal cortex (areas 9 and 46). J Comp Neurol 338:360-376. Levitt JB, Yoshioka T, Lund JS (1994) Intrinsic cortical connections in macaque visual area V2: evidence for interaction between different functional streams.J Comp Neurol 342:551-570. Levitt JB, Yoshioka T, Lund JS (1995) Connections between the pulvinar complex and cytochrome oxidase-defined compartments in visual area V2 of macaque monkey. Exp Brain Res 104:419-430. Livingstone MS, Hubel DH (1982) Thalamic inputs to cytochrome oxidase-rich regions in monkey visual cortex. Proc Natl Acad Sci USA 79:6098-6101. Livingstone MS, Hubel DH (1984a) Anatomy and physiology of a color system in the primate visual cortex.J Neurosci 4:309-356. Livingstone MS, Hubel DH (1984b) Specificity of intrinsic connections in primate primary visual cortex.J Neurosci 4:2830-2835. Lowel S, Singer W (1992) Selection of intrinsic horizontal connections in the visual cortex by correlated neuronal activity. Science 255:209-212. Lund JS, Yoshioka T (1991) Local circuit neurons of macaque monkey striate cortex: III. Neurons of laminae 4B, 4A, and 3B. J Comp Neurol 311:234-258. Lund JS, Yoshioka T, Levitt JB (1993) Comparison of intrinsic connectivity in different areas of macaque monkey cerebral cortex. Cereb Cortex 3:148-162. Malach R (1992) Dendritic sampling across processing streams in • monkey striate cortex.J Comp Neurol 315:303-312. Malach R, Amir Y, Harel M, Grinvald A (1993) Relationship between intrinsic connections and functional architecture revealed by optical imaging and in vivo targeted biocytin injections in primate striate cortex. Proc Natl Acad Sci USA 90:10469-10473. Malach R,Tootell RBH, Malonek D (1994) Relationship between orientation domains, cytochrome oxidase stripes, and intrinsic horizontal connections in squirrel monkey area V2. Cereb Cortex 4: 151-165. McGuire BA, Gilbert CD, Rivlin PK, Wiesel TN (1991) Targets of horizontal connections in macaque primary visual cortex.J Comp Neurol 305:370-392.