Orientation specificity in the mammalian visual pathway and retinal ganglion cell dendritic morphology.

The aim of the first investigation of this dissertation was to relate the orientation preferences of visual cortical neurons to the orientations of retinal ganglion cell dendritic fields in the cat. The preferred orientations of Simple cells in striate and extrastriate cortex and of Complex cells with narrow receptive fields in striate cortex but not those with wide receptive fields in striate and extrastriate cortex matched the orientations of the topographically corresponding ganglion cell dendritic fields. The columns representing the radial orientation are wider than those representing nonradial. This evidence indicates that the orientation specificity of cortical cells is related to the orientation sensitivity of retinal ganglion cells. The dendritic shape of ganglion cells in the macaque retina was also analyzed. Most of the dendritic fields were elongated, and there was a significant tendency for the dendritic fields to be oriented radially with respect to the fovea. The aim of the third investigation was to relate the variation of two attributes of dendritic morphology to the topography of the retina. The first attribute was the displacement of the center of the dendritic field from the center of the cell body. Ganglion cell dendritic fields tend to be displaced from the soma down the ganglion cell density gradient. The systematic displacement of the dendritic fields results in the dendritic field centers being arranged more regularly than are their cell bodies. The second attribute was the elongation and orientation of ganglion cell dendritic fields. The cat retina undergoes a process of maturation which begins at the area centralis prenatally and spreads over the retina in a horizontally elongated wave postnatally. The elongation and orientation of ganglion cell dendritic fields were correlated with the geometry of this wave of maturation. Dendritic competition can account for the systematic dendritic displacement but not for dendritic elongation and orientation which must, therefore, be accounted for by another mechanism which these results indicate is related to the wave of maturation.

ORIENTATION SPECIFICITY IN THE MAMMALIAN VISUAL PATHWAY AND RETINAL GANGLION CELL DENDRITIC MORPHOLOGY by Jeffrey O. Schall A dissertation submitted to the faculty of The University of Utah in partial fu1fil1ment of the requirements for the degree of Doctor of Phi losophy Department of Anatomy The University of Utah June 1906 THE UNIVERSITY OF UTAH GRADLATE SCHOOL SUPERVISORY CONLYIITTEE APPROVAL of a dissertation submitted by Jeffrey D. Schall This dissertation has been read by each member of the following supervisory committee and by majority vote has been found to be satisfactory. I C Keith A. Crutcher Kenneth W. Horch Audie G. Leventhal \J Hel ga Ko 1 b ~~ THE UNIVERSITY OF UTAH GRADUATE SCHOOL FINAL REi\DING APPROVAL To the Graduate Council of The University of Utah: I have read the dissertation of Jeffrey D. Schall in its final form and have found that (1) its format, citations. and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables, and charts are in place; and (3) the final manuscript is satisfactory to the Supervis­ory Committee and is ready for submission to the Graduate School. Dater f / Thomas N. Pa,rks Chairperson, Supervisory Committee Approved for the Major Department Marcus J~cobson Chairman I Dean Approved for the Graduate Council Clayton Dean of The Graduate School Copyright @ Jeffrey D. Schall 1986 All Rights Reserved ABSTRACT The aim of the first investigation of this dissertation was to relate the orientation preferences of visual cortical neurons to the orientations of retinal ganglion cell dendritic fields in the cat. The preferred orienta­tions of Simple cells in striate and extrastriate cortex and of Complex cells with narrow receptive fields in striate cortex but not those with wide receptive fields in striate and extrastriate cortex matched the orientations of the topographically corresponding gangl ion cell dendritic fields. The columns representing the radial orientation are wider than those representing nonradial. This evidence indicates that the orientation specificity of cortical cells is related to the orientation sensitivity of retinal gangl ion cells. The dendritic shape of gangl ion cells in the macaque retina was also analyzed. Most of the dendritic fields were elongated, and there was a significant tendency for the dendritic fields to be oriented radially with respect to the fovea. The aim of the third investigation was to relate the variation of two attributes of dendritic morphology to the topography of the retina. The first attribute was the displacement of the center of the dendritic field from the center of the cell body. Ganglion cell dendritic fields tend to be displaced from the soma down the ganglion cell density gradient. The systematic displacement of the dendritic fields results in the dendritic field oenters being arranged more regularly than are their oell bodies. The seoond attribute was the elongation and orientation of ganglion oell dendritic fields. The cat retina undergoes a process of maturation which begins at the area central is prenatally and spreads over the retina in a horizontally elongated wave postnatally. The elongation and orientation of ganglion cell dendritic fields were correlated with the geometry of this wave of maturation. Dendritic competition can account for the sys­tematic dendritic displacement but not for dendritic elongation and orientation which must, therefore, be accounted for by another mechanism whioh these results indioate is related to the wave of maturation. v TABLE OF CONTENTS ABSTRACT ............................................................................... i v LIST OF FIGURES .......................................................... viii ACKNOWLEDGMENTS ...................................................... xi CHAPTFR 1. ON THE ORIGIN OF ORIENTATION SPECIFICITY IN THE VISUAL SySTEM .............................................. 1 History of the Discovery of Orientation Specificity ............ 2 Attributes of Cortical Orientation Specificity ................... 7 Perceptual Implications ............................................... 10 Explanations of Cortical (lrientation Specificity ................ 14 Orientation Sensitivity in the Retina ............................... 24 Structural Basis of Ganglion Cell Orientation Sensitivity ...... 26 Orientation Sensitivity in the Dorsal Lateral Genicul ate Nucleus ................................... 2g Retinal Origin of Orientation Specificity in the Visual Pathway ........................................... 33 2. FACTORS AND EVENTS RESPONSIBLE FOR RETINAL GANGLION CELL DENDRITIC STRUCTURE ...................... 40 Structure and Function of Dendr ites ........................................ 41 Retinal Development .............................................................. 43 Ganglion Cell Differentiation .............................................. 47 Mechanisms of Dendritic Growth .................................. 48 Spatial Distribution of Ganglion Cell Dendrites ................... SO Conclusion ............................................................. 56 References ....................................................................... 59 3. RETINAL CONSTRAINTS ON ORIENTATION SPECIFICITY IN CAT VISUAL CORTEX .............................................. 82 Mater i a Is and Methods ................................................... 83 Results ................................................................ 86 Discussion ............................................................. 9 1 Conclusion ............................................................. 94 References ............................................................. 94 4. RETINAL GANGLION CELL DENDRITIC FIELDS IN OLD-WORLD MONKEYS ARE ORIENTED RADIALLy ...... 91 Materials and Methods ................................................................................. 98 Resul ts ..................................................... o,o, ••••••••••• 99 Discussion .......................................................................................................................... 100 References ........................................................................................................................ 102 5. RELATIONSHIPS BETWEEN GANGLION CELL DENDRITIC STRUCTURE AND RETINAL TOPOGRAPHY IN THE CAT ......... 104 Abstr act ................................................................... 1 D5 Introduction ..................................................................................................................... 106 Materials and Methods .......... """ .. "" .... " ...... " .. " .......................... " .... ,, .. 1 01 Results .................................. Ill ............................................................................................. 115 Discussion ....................................................................... 11 ...... ,. .......................................... 129 References ............................ " ............... " ..... " ............ 139 vii LIST OF FIGURES Figure Chapter 1 1. The first receptive field maps of frog and cat retinal ganglion cells .••....•..•.. " .•.•. -. ....• " ...• " ..... 3 2. Visual illusions illustrating that the perception of the orientation of a line is influenoed by the presence of other I jnes with different orientations ."" ....• " .•......•..•.... " ....• " •.......• " .. 12 3. Hubel and Wiesel's original model to aocount for Simple cell orientation specificity ......... II ••••••••••••• 16 4. Hubel and Wiesel's original model to aocount for Complex oell orientation specificity ............... II •••• II •• 18 5. Comparison of the sizes of recipient dendritic fields and afferent axonal fields in the retina and visual cortex ...................... " ....... -. ......•..... 30 6. A model to illustrate how lateral inhibition between neurons receiving orientation sensitive input can result in orientation selectivity"."" .... " ......... " ........................ :i7 Chapter 3 1. Computer-generated drawing of all the alpha cells in a patch of retina ...... " -. ........... "."." ........... II 85 2. Computer drawing of an alpha cell and its stick figure representation ... " ......................... " ....... 86 3. Stick fiQure representations of all of the alpha cells in three patches of retina .... """"" .... ""."""",, .. ,, .. 87 4. Distributions of the dendritic field orientations of all of the alpha cells in three patches and of alpha cells sampled randomly along each meridian .................. 88 5. Dendritic field orientations of alpha and beta ganglion cells and preferred orientations of area 11 S cells subservi ng central vis i on ......................................... 88 6. Orientations of beta cells and preferred orientations of area 11 S- and C-type cells subserving different meridians ..•....•........••...•.....•..•..............•. 89 7. Orientations of alpha cells and preferred orientations of area 18 Sand C cells ......................................... 89 8. Differences in orientation between neighboring alpha cells of I ike type and differences in preferred orientation between successively recorded cells in area 17 ..................................................... 90 9. Plot of preferred orientation versus distance in a penetration into striate cortex subserving the peripheral horizontal meridian ......•...•••....•........... 9 1 1 U. Plot of preferred orientation versus distance in a penetration into striate cortex subserving an ob I i que mer i dian ...•...........••.•............•...•.....•..... 9 1 11. Graph of preferred orientation versus distance in a penetration into striate cortex subserving the vertical meridian •............••.......................•.....• 92 12. Summary of the results and hypotheses of this study ........... 93 Chapter 4 1. Photomicrographs of an HAP-filled P-alpha and P-beta ganglion cell ......................................... 99 2. Computer representation and orientation analysis of a P-alpha and a P-beta cell ................................... 100 3. Orientation biases of all monkey ganglion cell analyzed ........ 101 4. Differences between dendritic field orientation and polar angle for oriented monkey ganglion cells ................. 101 ix Chapter 5 1. Computer representation and morphological analysis of an alpha cell .... "" .................................................... 112 2. Absolute and relative magnitude of dendritic displacement .......................................................... 116 3. Angle difference between the direction of displacement of the dendritic field from the cell body of alpha and beta ganglion oells more than 1 mm from the area centra I is ................................................. 11 B 4. Relation of dendritic displacement and the gangl ion cell density gradient .........•..................... 120 5~ Spatial arrangement of ganglion cell bodies and dendritic fields ............................................................. 122 6. An approximation of the geometry of the wave of maturation ......................................... 125 1. Relation of dendritic field orientation and the geometry of the wave of maturation ............................. 121 x ACKNOWLEDGMENTS I am first and foremost grateful to Dr.. Audie Leventhal.. He motivated me to do more than I thought I could; he taught me not only the art but the vocation of science .. Whatever success I might achieve is to his credit. I am also grateful to the chairman of my committee .. Dr. Tom Parks, who was always free with scientific and practical advice. Each of the members of my committee, Drs. Keith Crutcher, Ken Horch, Helga Kolb and Richard Mullen helped me expand and clarify my thinking. I am thankful for the opportunity to collaborate with Or. Hugh Perry who helped me formulate many of the ideas in this dissertation. Drs. Steve Ault and Dagmar Vitek have also contributed to my education. Wendy Wallace .. Marsha Phi II ips and Jean Holl is deserve much credit for their technical contributions. Each of the students; faculty and staff of the Anatomy Department has enriched my graduate training. I am grateful to the American Physiological Society for permission to reproduce the figures used in Chapter 1, to the Society for Neuroscience for permission to reproduce -Retinal Constraints on Orientation Speci­fic ty in Cat Visual Cortex· in its entirety as Chapter 3 and to Elsevier Science Publ ishers for permission to reproduoe -Retinal Gangl ion Cell Dendritic Fields in Old-World Monkeys are Oriented Radially- in its entirety as Chapter 4. The work presented in this dissertation was supported in part by a Graduate Research Fellowship from the University of Utah Research Committee to the author and by PHS grants EY03421 and EY04951 to Or .. Audie Leventhal .. finally, I would like to express my loving gratitude to my wife who put up with more than she should have had to. xi i CHAPTER 1 ON THE ORIGIN OF ORIENTATION SPECIFICITY IN THE VISUAL SYSTEM A false conclusion once arr ived at and widely accepted is not easi Iy dislodged .. and the less it is understood ... the more tenaciously it is held. George Cantor (J<nne ... 1980) 2 Hi$toLY.. of t!l.J_[!isco.Y~r.Y.!tl. Orientation Speoifioity The orientation specificity of neurons in the visual system was al­most not discovered.. Hartline (1938) was the first to record the activity of single neurons in the visual system, reoording from single ganglion cell axons in the frog retina. Hartline found that each ganglion cell responded to I ight fall ing wi thin a restr icted area of the visual field; he denoted this area the receptive field .. Some cells responded only at the beginning of il­lumination; others, only at the end of illumination, and others, to both light on and light off.. Hartline (1940) also mapped the spatial extent of gangl ion cell receptive fields using small (50 urn) flashing spots. The first published map of a ganglion cell receptive field is reproduced in Figure lA. Notioe that the reoeptive field is clearly elongated; Hartline makes no comment of this. Kuffler (1953) recorded from single ganglion cells in the cat retina. He also found that eaoh ganglion oell exhibited a oharaoteristio response to changes of illumination within its receptive field" Kuffler also discovered that each gangl ion cell responded to either I ight on or light off in the center of its receptive field and to the opposite stimulus contrast in a surrounding annulus. The component of a cell·s receptive field was denoted according to its response to light on or light off" For example .. a ganglion cell which responded to the onset of illumination in the center and offset A • tXPLORING )POT • 3 Imm IMM Figure 1. The first receptive field maps of frog and cat retinal gangl ion cells. In A is shown the first neuronal receptive field ever demonstrated (Hartl ine .. 1940). A 50 pm spot .. shown in the upper left .. of two intensities was flashed through the receptive field. The outer contour delimits the region responding to the brighter spot; the inner contour delimits the region responding to the dimmer spot. In B is shown the first published receptive field of a ganglion cell from cat retina O<uffler ... 1953). The receptive field was mapped with a 200 11m spot of unchanging brightness. The cell responded to the onset of the flash in the region denoted by the -+-. The cell responded to the offset of the flash in the outer region denoted by the ·0· and the diagonal hatching. The cel I responded to both the onset and offset of the flash in the intermediate zone. Notice that both receptive fields are clearly elongated .. but neither investigator remarked about this attribute. [Reproduced from the Journal of Neurophysiology with the kind permission of the American Physiological Society.] 4 in the surround was ON-center, OFF -surround. The response of a gang-lion oell increased as a spot covered more of the receptive field center. As the spot enlarged to cover the center and the surround, the response of the cell decreased. Within the center or surround there appeared summation; between center and surround there appeared mutual antagonisrn. Kuff ler also mapped the spatial extent of gangl ion cell receptive fields; an example is reproduced in Figure 1 B. Once again the investigator did not comment on the elongation of the receptive field. It was abstracted from these data that the reoeptive fields were circular; consequently .. the orientation of a stimulus was not a relevant variable. This abstraction was not challenged in subsequent investigations of the retina. Wiesel (1960) .. working in Kuffler1s laboratory .. recorded from cat ganglion cells over a wider area of retina; he noted that the receptive fields were -more or less circular.· The shape of a gangl ion cell1s receptive field was so unimportant that in the first study of gangl ion cell receptive fields in monkey retina by Hubel and Wiesel (1960) the shape of the receptive fields was not even investigated. Rodieck and Stone (1965) proposed the first model attempting to account for gangl ion cell receptive field properites. They noticed that the receptive fields were not always circular .. but this fact was not incorporated into their model. Meanwhile in Germany ... Jung ... Baumgartner and their group were the first to record the activity of single neurons in the visual cortex of the cat (reviewed by Jung ... 1 gSa, 1 g75). They found a variety of cells whose ac-tivi ty was either enhanced or attenuated by changes in overall illumination of the visual field. One-half of the neurons they recorded did not respond 5 to changes in general ilium ination but did respond to contours or edges of I ight. A meohanism was oonstruoted to oontrol the stimulus presentation. The machine, however, shone only vertically oriented edges, so they were unable to compare the responses of cortical neurons to edges of different orientations. Hubel (1958 .. 1959) was developing techniques to record trom single neurons in the visual cortex of unrestrained cats. He duplicated the results of Jung's group, finding many units responding to changes in diffuse illumination. One-half of the neurons Hubel reoorded, however, only re­sponded to small spots of I ight moving through their reoeptive fields. He also discovered that the response of many cells varied with the direction in which the spot moved. Sensitivity for the direction of stimulus movement had not been detected in retinal ganglion cells; hence, direction sensitivity was the first receptive field property which distinguished cortical neurons from retinal ganglion cells. Hubel (196D) also recorded from the dorsal lateral geniculate nucleus (LGNd) and was convinced that the receptive fields were not essentially different from the circular [sic], center-surround ganglion cell receptive fields originally described by Kuffler. In 1958 Hubel joined Wiesel in Kutfler's laboratory (see Huber, 1982. for the history). They commenced recording the activity of single neurons in the visual cortex of anesthetized ... paralyzed cats. Under these circumstances they were able to anal)'7e the receptive fields of neurons more carefully .. over longer periods. Hubel and Wiesel.. using flashing spots of light to stimulate cortical cell receptive fields .. proceeded slowly at 6 first .. They found that the arrangement of the receptive fields was differ-ent from the oenter-surround found in the retina and LGNd. For many neurons the ON and OFF regions lay side by side rather than surrounding each other as in the retina and LGNd; these were denoted Simple cells. For other cells the regions responding to light ON and light OFF over lapped; these were denoted Complex cells. The spots of I ight were produced with a slide projector. During one experiment, a crack in the slide cast a bar of I ight across a cell's receptive field, and the neuron responded more vigorously than had any before. Hubel and Wiesel soon found that the orientation of the bar was oritical. Cortical neurons, they found, only responded when the bar of I ight was aligned along a oertain angle whioh varied from oelt to oell.. If the bar deviated more than a few degrees to either side .. the cell did not respond. Orientation specificity had been discovered in the cortex. Hubel and Wiesel had begun the investigation with the aim of finding how neurons at different levels of the visual pathway operated on their input (1 959 .. 1 962 .. 1 965). This could be analyzed .. they believed .. by determining what stimuli were effective for cells at one level of the visual pathway and not at another level. These differences could reveal what each level was doing toward perception. Orientation specificity was considered a major success in this approach. The input to the visual cortex, it was bel ieved ... was not orientation sensitive ... and cortical cells were orientation selective. Hence .. this is what the cortex was doing - analyzing the orientation of edges in the world. Attributes of Cortioal Orientation Speoifioitv Orientation Preference and Selectivity The characteristics of orientation specific neurons in the visual cortex wi II be reviewed to provide a groundwork for subsequent discussion. The orientation specificity of a neuron can be defined by at least two at­tributes: 1) orientation preference and 2) degree of sensitivity or selectivity_ The two attributes may result from different mechanisms and are probably responsible for different aspects of visual perception. In this dissertation the phrase orientation specificity subsunnes both preference and selectivity_ Consider the response of a cortical neuron to elongated stimul i of different orientations. The cell fires the most action potentials when its receptive field is exposed to the stimulus in one orientation; this ori­entation is denoted the cel rs preferred orientation. Each neuron has just one preterred orientation. As the stinnutus rotates away trom the pre­ferred orientation, clockwise or counterclockwise, the cell responds less, until beyond some orientation it does not respond at all. Simple cells typically are more selective for the orientation of a stinnulus than are Complex cells. The degree of selectivity of a cortical cell varies with the nature of the stimulus. Both Simple and Complex cells exhibit less se­lectivity when short bars are used; the selectivity and the peak response improves with longer bars (Henry .. Dreher and Bishop .. 1974). Functional Architecture In his pioneering investigation of somatosensory cortex Mountcastle (1957) observed that all of the cells in a vertical penetration of cortex B responded to the same quality of stimulation (hair movement, skin pressure or pressure on deep tissue), and the quality of stimulation changed as the electrode passed horizontally through the cortex. He introduced the term column for this arrangement; neurons in the same vertical column of cortex shared optimal stirnulus properties. The visual cortex of higher mammals is organized in a simi lar fashion. Neurons in vertical slabs of cortex tend to prefer the same ori­entation, and neighboring slabs represent similar orientations in cat (Hubel and Wiesel, 1962, 1963b; Albus, 1975, 1979; Singer 1981; Albus and Sieber, 1984), tree shrew (Humphrey and Norton, 1980; Humphrey, Skeen and Norton, 1980) and monkey (Hubel and Wiesel, 1968, 1974; Hubel, Wiesel and Stryker, 1978). Orientation specific cells are not distributed so neatly in striate cortex of mouse (Drager, 1975), rat (Shaw, Yinon and Auerback... '1975; Wiesenfeld and Kornel, 1975), hamster (T iao and Blakemore.. 1 9-16b) or rabbit (Chow, Masland and Stewart, 1911; Van Sluyters and Stewart, 1974; Bousfield, 1977; Murphy and Berman, 1979). In the vertical dimension the representation of orientation is not as straightforward as originally described by Hubel and Wiesel. Recent work (Bauer .. 1982; Kruger and Bach .. 1982; Bauer .. Dow .. Snyder and Vautin .. 1983) has demonstrated that as the electrode advances perpendicular to the surface of the cortex through the supragranular and granular layers .. the cells exhibit the same orientation preference ... but as the electrode ad­vances into the infragranular layers the orientation preference often but not always shifts many degrees. Moreover .. if the activity of cells in layer 4 is blocked, neurons in the supragranular layers are sti II orientation 9 selective (Malpeli, 1983). If the the supragranular layers are inactivated, cells in the infr~gr~nul~r layers ~re lItso sti 11 or ientation selective (Schwark and Malpeli, 1984). Qriel!!~tion Anisotro~y The representation of orientation in the visual cortex is not iso­tropic. In the region of visual cortex subserving central vision most neurons prefer horizontally or vertically oriented stimul i (Pettigrew, Nikara and Bishop, 1968; Mansfield, 1974; Mansfield and Ronner, 1978; Leventhal and Hirsch, 1980; Orban and Kennedy, 1981). Most neurons in areas 17 (Leventhal, 1983) and 19 (leventhal, Schall and Wallace, 1984) of the cat subserving outside the area centralis prefer radially oriented stimuli. The overrepresentation of horizontal and vertical and of radial is more pronounced in Simple cells in area 17 and in cells with the narrowest receptive fields in area 19. The radial relation between orientation pre­ference and receptive field position is less evident in the Complex cells of area 17 and absent in the neurons with the widest receptive fields in area 19 .. Development of Cortical Orientation Specificity Hubel and Wiesel (1963a) presented evidence from seventeen neurons in four kittens that neurons in the striate oortex exhibit essentially adult-like reoeptive field properties. Subsequent work, however, found that neurons in the kitten visual cortex lack adult specificity and respons­iveness (Pettigrew, 1974; Blakemore and Van Sluyters, 1975; Buisseret and Imbert .. 1976; Fregnac and Imbert. .. 1978; Albus and Wolf, 1904). While Hubel and Wiesel's report indicated that all of the connecti~ity necessary 10 for cortical orientation specificity was diotated by the genome, more reoent work has demonstrated that the development of the orientation specificity of some neurons depends on experience.. Such dependence was first demonstrated by Hirsch and Spinelli (1970, 1971) and Blakemore and Cooper (1970).. These investigators raised kittens in environments which allowed the animal to see only contours of a single orientation. When the animals matured, the orientation preferences of many cells recorded from area 11 matched the orientation which the animal saw in early life. Neurons with Simple type receptive fields in layer 4 express their ori­entation speoifioity the earliest and independent of the rearing of the animal while the orientation specificity of neurons with Complex type receptive fields in the supra- and infragranular layers appears to be influenced by the early experience of the animal (reviewed by Hirsch and Leventhal ... 1978; Movshon and Van Sluyters ... 1981; Fregnac and Imbert ... 1904). Perceptual Implications The fact that orientation specificity can be demonstrated in the visual cortex does not necessarily indicate that orientation specificity is required for visual perception.. It Is ... therefore ... worth considering whether the orientation specificity is used and useful in visual perception. Is Orientation Specificity Used? Humans can deteot I ines and edges of all or ientat ions and can dis­criminate between two edges with different orientations.. Using central vision, humans detect and discriminate horizontally or vertically oriented 1 1 stimuli better than obliquely oriented stimuli; this has been termed the ·obl ique effeot· (reviewed by Appel Ie, 1972). The obi ique effeot appears to be innate since it is present in infants (Leehey, Moskowitz--Cook, Brill and Held, 1975). The obi ique effect dissipates as stimul i are presented to the peripheral visual field (Berkley, Kitterle and Watkins, 1975; Vandenbussche, Vogels and Orban" 1966). Using peripheral vision humans detect radially oriented stimuli better than nonradially (Rovamo, Virsu, Laurinen and Hyvarinen, 1982; Fahle and Braitenberg, 1983; Temme" Maino and Noell, 1984; Vandenbussche et aI., 1986). The meridional vari­ation in orientation detection and discrimination is correlated with the distribution of orientation preferences represented by the Simple cells in visual cortex (Pettigrew et aI., 1966; Mansfield, 1974; Mansfield and Ronner, 1976; Leventhal and Hirsch, 1960; Orban and Kennedy, 1961; Leventhal, 1963; Vandenbussche and Orban" 1983; Orban et aI., 1984). The presence of lines of one orientation influences the perception of the orientation of lines of another orientation (Andrews, 1961: Carpenter and Blakemore, 1913). Visual i lJusions such as illustrated in Figure 2 de­monstrate this phenomenon; the vertical lines appear bowed due to the presence of the obliquely oriented lines. These results have been inter-' preted as indicating inhibition between neurons with different preferred orientations in the visual system (Blakemore, Carpenter and Georgeson, 1910). Finally, ablation of striate cortex results in a loss not only of orientation discrimination abi lity but of high resolution form vision in 12 A B Figure 2. Visual illusions illustrating that the perception of the orientation of a line is influenced by the presence of other lines with different orientations. Ewald Hering publ ished the illusion shown in A in 1861. Wilhelm Wundt published the converse illusion shown in B in 1869. The vertical lines appear bowed due to the presence of the obliquely oriented lines. 13 general (Sohneider, 1969; Diamond and Ha", 1969; Weiskrantz, 1972; Berkley and Sprague, 1979). Is Orientation Specificity Useful? While these data imply that orientation specific cells are used in vision ... they do not demonstrate how orientation specificity is useful. To discuss whether orientation specificity is necessary for visual perception" it is beneficial to approach visual perception as an information processing task (Marr, 1982; Poggio, Torre and Koch, 1985). The input to the visual system is the I ight intensity detected by the photoreceptors, and that is all. From that a representation of the objects in the world must be de­rived. To demonstrate that orientation specificity is required for visual perception, its necessity at the level of the computational theory must be demonstrated. Edges contribute necessary information about objects. and a representation of an edge must include position and orientation. In this sense, then, it seems reasonable that orientation specific neurons are necessary in form vision. Why Radial? Certain regions of the brain appear to be required for visually guided movements; these areas include the superior coil iculus (Schneider, 1969) and the posterior parietal cortex (reviewed by Hyvirinen, 1982). Cells in the superior coli iculus are especially sensitive to movement .. and most cells tend to prefer stimul i moving away from the center of view (McIlwain and Buser ... 1968; Sprague ... Marchiafava and Rizzolatti ... 1968; Sterling and Wiokelgren, 1969; Strasohill and Hoffman, 1969; Berman and Cynader, 1972). A similar overrepresentation of neurons preferring 14 oentrifugal motion has been demonstrated in the posterolateral nuoleus of the thalamus (Rausoheoker, Friederiohs, yon GrOnau and Poul in, 1904) and the posteromedial lateral suprasylvian visual area in the cat (Poul in, von Grunau and Aauschecker, 1984). Neurons in the posterior parietal cortex of monkeys exhibit receptive fields which cover an extremely large part of the peripheral visual field. The cells are sensitive ... however, to the direction of movement of stimul i within the receptive field; most of the neurons prefer stimul i moving away from the center of view even though the absolute direotion varies in different parts of the reoeptive field (Motter and Mountoastle, 1981). Thus, in regions of the brain involved in providing information about the motion of an animal in the world most of the neurons subserve the peripheral visual field and respond best to radially moving stimuli. Form vision is accompl ished in the center of view41 whi Ie peripheral vision is suited for detecting moving objects (Trevarthen ... 1968; Schneider, 1969). Visual feedback to guide an animal's movement through the world is provided by the optioal flow field (Gibson", 1950; 1966). Forward movement results in optical flow which expands radially from the center of view" It seems reasonable.. therefore.. that the visual system has adapted with an overrepresentation of cells responding to radial stimuli. Explanations of Cortical Orientation Specificity Hubel and Wiesel's Model Hubel and Wiesel (1962) proposed the first model to account for the receptive field properties of cortical neurons. One group of neurons had receptive fields with discrete ON and OFF regions which lay side-by-side. 15 These neurons were enoountered most often in layer 4 of striate oortex wherein the LGNd afferents arrive. The reoeptive fields of these oells, Hubel and Wiesel argued, could be understood as ar ising from input to the cortical cell -from a row of LGNd cells as shown in Figure 3. These cells were denoted S;rn.l:tlLQ~lls. The arrangement of the LGNd oell reoeptlve fields was supposed to account for the organization of the Simple cell receptive fields and explain the cell's orientation specificity. A spot of light falling within the central ON region drove the cell to some degree" and a bar covering the entire central region drove the neuron more. If a spot of light covered both the central ON and flanking OFF regions. the ceJlls response was suppressed. Thus, within the ON or OFF regions there was spatial summation; between regions there was antagonism. In Hubel and Wiesel's view the interactions within and between ()N and OFF regions of cortical cells was just I ike that of retinal gangl ion cells and LGNd relay neurons; the only difference was that in the cortex the ON and OFF regions are arranged differently. Mu­tual antagonism between and length summation within the ON and OFF regions" Hubel and Wiesel hypothesized" is responsible for the orientation specificity of a Simple celt Hubel and Wiesel found that other neurons in striate cortex re­sponded to a stimulus flashing ON and OFF anywhere within their reoeptive field. The organization of these receptive fields did not clearly predict their orientation specificity" nor could the arrangement be simply under­stood as arising from exoitatory input from L~Nd oells. Aooordingly, such neurons were denoted Complex cells. Complex cells were essentially never 16 CD Simple cell Figure 3. Hubel and Wiesel's original model to account for Simple cell orientation specificity. Four lGNd relay neurons provide excitatory input to a Simple cell in the cortex. The center-surround receptive fields of the LGNd cells. represented by the concentric circles. are arranged in a row. The Simple cell receptive field. represented by the rectangles. has a central region with the same response as the centers of the LGNd receptive fields. and flanking regions derived from the surrounds of the LGNd cell receptive fip.lds. Hubel and Wiesel proposed that the preferrp.d orientation of a Sirnple cell was specified by convergence along the line of lGNd receptive fields and the selectivity resulted from antagonism between the center and the surround of the LGNd neuron receptive fields. 17 found in layer 4. Hubel and Wiesel proposed that each Complex cell received excitatory input from several Simple cells which all had the same preferred orientation and overlapping receptive fields as illustrated in figure 4. This hierarchy, they argued, was supported by the fact that Simple and Complex cells with the sarne preferred orientation were found in the same vertical column of cortex. Since they were in such proximity, they could be economically connected. Evaluation of Hubel and Wiesel's Model There is no evidence which conclusively supports and much which is counter to Hubel and Wiesel's model. Hubel and Wiesel's model predicted that the orientation specificity of a Simple cell was determined by the spatial arrangement of the ON and OFf regions in the receptive field. This prediction has been tested by Watkins and Berkley (1974) in the cat and by Schiller, Finlay and Volman (1976) in the monkey. They found that the degree of selectivity of a Simple cell was not correlated with the ar­rangement of the ON and OFF regions. Furthermore, Dreher and Sanderson (1973) showed that LGNd cells respond to bars of all orientations ex­tending over both center and surround. This indicates that antagonism between the oenter and surround of LGNd oells expressed in cortioal Simple cells cannot account for the lack of response to nonoptimal orientations as Hubel and Wiesel originally suggested. It is possible to selectively block the response of ON center gangl ion cells by injecting into the eye a glutamate analogue.. 2-amino- 4-phosphonobutyric acid (APB) (Slaughter and Miller. 1981). Following such treatment the activity of ON center cells in the LGNd ceases ... while 18 Complex cell Simple cells Figure 4. Hubel and Wiesel's original model to account for Complex cell orientation specificity. A group of Simple cells provide excitatory to a Complex cell. The receptive fields of the Simple cells, represented by the enclosed rectangles... are overlapping and have the same preferred orientation. The Complex cell receptive field", represented by the thick rectangle, does not have segregated ON and OFF regions. Hubel and Wiesel proposed that the preferred orientation of a Complex cell Is specified by the preferred orientations of the converging Simple cells", and the selectivity is derived from the selectivity of the Simple cells. 19 OFF oenter oells are unaffeoted in the oat (Horton and Sherk, 1984) and monkey (Sohiller, 1984). In the visual oortex of the oat (Sherk and Horton, 1904) and monkey (Schi lIer.. 1902) the APB treatment elim inates the response to ON flash and moving light edges. Even though the ON compo­nent of the receptive fields is absent .. however .. the orientation specificity of the neurons is unaffected. Hence .. mutual antagonism between the ON and OFF regions is not necessary or sufficient for orientation specificity_ In their original model Hubel and Wiesel assumed that each cortical oell ultimately reoeives input from many ganglion oells. Simultaneous recording of Simple type cortical cells and the ganglion cells with over-­lapping receptive fields indicates that each Simple cell receives excitatory input from one to a few gangl ion cells (Lee, Cleland and Creutzfeldt, -1977). Intracellular recording also indicates that the exci tatory input to cortical cells is from a few ganglion cells .. and the excitatory receptive field is not elongated (Creutzfeldt and Ito .. 1968; Creutzfeldt .. Kuhnt and Benevento, 1974). This evidence indioates that the exci tatory input to cortical cells is not sufficient to account for orientation specificity. Another premise of t-tubel and Wiesel's model is that the orientation specificity of Complex cells is conferred by SirTlple cells. Subsequent evi­dence, however, makes this proposal indefensible. Complex cells respond to stimul i which Simple cells do not (Hammond and Mackay, 1977). Furthermore, complex cells receive direct, excitatory input from the LGNd (Toyama, Maekowa and Takeda, 1977; Bullier and Henry, 1979). Finally, blookade of aotivity in the LGNd A layers and thus in layer 4 of 20 striate oortex does not el iminate stimulus speoifio responses of Complex cells in the supra- and infragranular layers (Malpel i, 1903). Lateral Inhibition Hubel and Wiesel initially underestimated the importance of lateral connections within the cortex since vertical connections were so much denser (Lorente de No .. 1943). Recently .. however .. both Hubel (Livingstone and Hubel, 1 904) and Wiesel (Gi Ibert and Wiesel, 1 983) as well as others (Fisken, Garey and Powell, 1975; Rockland, Lund and Humphrey, 1982; Rockland and Lund, 1983; Matsubara, Cynader, Swindale and Stryker, 1985) have sh~n that there is a considerable degree of organization in the lateral connectivity of striate cortex. Many of these lateral connections are inhibi tory (Fisken et aI., 1975; Hess, Negishi and Creutzfeldt, 1975). Hubel and Wiesel did not conceive of the possibi lity of intracortical inhibition playing a role in cortical orientation specificity_ The response of a cortical cell to an optimally oriented bar in the center of its receptive field, however, varies with the orientation of another stimulus presented in or around the receptive field (Blakemore and Tobin, 1972; Fries, Albus and Creutzfeldt, 1977; Nelson and Frost, 1978; Burr, Morrone and Maffei.. 1981; Morrone, Burr and Maffei.. 1982; Heggelund and Moors, 1983). When the second stimulus is parallel to the preferred orientation, the cells respond maximally. As the second stimulus is rotated away from the orientation preference .. the response of the cell to the optimally oriented bar is suppressed. These results are interpreted as indicating that the response of the recorded cell is inhibited by the activity of afferent neurons which are responding to the second stimulus. 21 Using two simultaneously presented stimul i, it is possible to de-monstrate suppressive and excitatory zones within a cortical cell receptive field (Heggelund, 1981a,b).. When the two stimuli are flashed in phase, suppressive flanks can be demonstrated around the central discharge zone of Sirnple cell receptive fields.. Such an arrangement can be accounted for if the Simple cell received direct, excitatory input from a single ON or OFF LGNd cell and indirect, inhibitory input from another LGNd cell whose receptive field ;s of the same sign and slightly displaced. The ar­rangement of exoitatory and suppressive zones in Complex cel I receptive fields oan be accounted for if the Complex cell receives direct, eXCitatory input from both ON and OFF LGNd cells and indirect, inhibitory input from both ON and OFF LGNd cells with slightly displaced receptive fields .. The orientation selectivity of cortical neurons can be predicted by the elongation and arrangement of the excitatory and suppressive regions (Heggelund and Moors, 1983). Intracortical inhibition mediated by gamma-aminobutyric acid (GABA) has been demonstrated directly. Iontophoretic application of the GABA antagonist, bicuculline, causes a reduction in the orientation se-· lectivity of Simple cells and an elimination of the orientation selectivity of some Complex cells in cat striate cortex (Sillito, 1975, 1979; Tsumoto, Eckart and Creutzfeldt, 1919). Using the more potent GABA antagonist, N-methyl-bicuculline (Si I lito .. Kemp, Milson and Berardi, 1980), the measurable orientation selectivity of Simple cells can be completely eliminated, and the remaining exoitatory reoeptive field is not elongated. 22 Intraoortioal inhibition and orientation seleotivity develop in parallel (Komatsu .. 1993; Albus and Wolf .. 1994). In the striate cortex of kittens from one to five weeks of age some neurons exhibit mature orientation selectivity; others respond to all orientations with one clear preference .. and others respond equally to all orientations. The orientation selectivity of the first group of cells was reduced or abolished upon application of bicuculline (Wolf. Albus and Hicks, 1903) or N-methyl­biououlline (Sato and Tsumoto. 1984). The orientation sensitive oells were less affeoted by the bicuoulline treatment. and the nonorientation sensitive cells were completed unaffected. Cortical Oendritic Field Elongation and Qrientati~~referenges Sutherland (1957) demonstrated that an octopus discriminates stimul i oriented horizontally or vertically better than those oriented ob­I iquely. Young (1960) showed that the dendritic fields of most neurons in the optio lobe of the ootopus are oriented horizontally or vertioally. and he suggested that the anisotropy may underlie the abi lity of the octopus to discriminate horizontally and vertically oriented stimuli better than obliquely" Motivated by YounQ·s hypothesis" Colonnier (1964) analyzed the shape of the dendritic fields of neurons in tangential sections of cat visual cortex. He found that most stellate cell dendritic fields in layer 4 and basal dendritic fields of pyramidal cells are elongated in the antero­posterior dimension, approximately parallel ~o the representation of the vertioal meridian of the visual field in the cortex. Colonf)ier suggested 23 that such elongation may provide for the preferential convergence of LGNd afferents. If this ~ere so, then the preferred orientation of a neuron should be parallel to the orientation of its dendritic field. Schiller et aL (1976) assumed that excitatory synapses were made near the cell body and inhibitory synapses .. on the distal regions of the dendritic field of cortical cells. They hypothesized that due to the elongation of the dendritic field into adjacent columns .. each cell was in­hibited by neurons with different preferred orientations. Consequently .. the preferred orientation of a neuron should be perpendicular to the tangential orientation of its dendritic field. The relation between cortical cell dendritic field orientation and its preferred orientation was tested indirectly by Tieman and Hirsch (1982). Cats were raised viewing only horizontal or vertical I ines. The tangential orientation of the dendritic fields of stellate cells in layer 4 was not different from normal. However .. the basal dendritic fields of layer 3 pyramidal cells were oriented orthogonal to the representation of the angle to which the cat was exposed .. Le ... in the cat reared in a vertical en­vironment the dendritic fields tended to be oriented perpendicular to the representation of the vertical meridian. Since the preferred orientation of the layer 3 pyramidal cells probably tend to correspond to what the animal saw in early J ife. this result indicates that the orientation pref­erences of a neuron may be orthogonal to the orientation of its dendritic field. Martin and Whitteridge (1984) directly tested the hypothesis that the preferred orientation of a cortical neuron is related to the tangential 24 orientation of the dendritic field by recording from single cells then intraoeltularly injeoting HRP. In a sample of just 24 oells of all types from all fayers there was no correlation between dendritic field orientation and orientation preference. Since the sample is so thin across cell types and laminae ... this finding must be considered tentative. Sti II ... the result indicates that the preferred orientation of a cortical neuron is not determined by convergence of LGNd afferents due to the elongation of the dendritic fields. In summary. there is no evidence to support Hubel and Wiesel's original model and much counterevidence. Furthermore, no other models have been proposed which can account for aft of the observed attributes of cortical orientation specificity. Hence, 25 years after its discovery the mechanisms underlying cortical orientation specificity are not understood. Recent investigations of the visual pathway peripheral to the visual cortex, however, may provide a solution. Orientation Sensitivity in the Retina In 1973 Rodieck reviewed the state of knowledge about the retina. Concerning ganglion cell receptive fields he wrote .. All the [ganglion cell] receptive fields were to some extent radially asymmetric. ... In most receptive fields the surround region was found to be asymmetrically situated about the center region... Since no order or pattern could be detected in these deviations from radial symmetry, they presumably reflect vagaries in neural wiring rather than some form of functional special ization. (p 566) The following year Hammond (1914) published the first analysis of the size and shape of gangl ion oell reoeptive fields in oat retina. It was accepted that the extent of a ganglion cells dendritic field defined the 25 oenter oomponent of its reoeptive field (Brown and Major, 1965; Boyoott and Dowling, 1969; Honrubia and Elliott, 1970; Boyoott and Wassle, 1974). Hammond recognized that ganglion cell dendritic fields had been described as irregularly shaped, more often elongated than oiroular (Leioester and Stone, 1961; Honrubia and Elliott, 1910; Boycott and Wassle, 1914). He found that most of the ganglion cell receptive fields were elongated; moreover, most of the receptive fields he sampled were oriented nearly horizontally. Order and pattern had been demonstrated. Subsequently, Leviok and Thibos (1982) demonstrated that the majority of ganglion oells in oat retina exhibit a signifioant degree of ori entation sensitivity ;n response to drifting sine wave gratings of high spatial frequenoies. Moreover, there is a signi"fioant tendency for ganglion cells outside the area centralis to prefer stimuli oriented approximately radially .. Le ... as the spokes of a wheel with the area oentralis at the hub. Thibos and Levick (1 9B~) considered the possible functional significance of gangl ion cell orientation sensitivity_ For the activity of a ganglion celt to discriminate the orientation of a stimulus, its response to nonoptimal orientations must be significantly different from the response at the optimal orientation. Thibos and Levick determined the standard deviation of the responses of ganglion cells to a stimulus at a Single orientation" If a ceilis response to a nonoptimal orientation was one standard deviation less than the response to the preferred orientation .. then the cell could be considered functionally orientation sensitive. According to this criterion, 73% of the cella they sampled diacriminated the orientation of drifting 26 gratings. Thus, the aotivity of retinal ganglion oells oan funotionally disoriminate the orientation of a stimulus. Structural Basis of Gangl ion Cell Orientation Sensitivity Retinal Gangl ion Cell Types Other than the distinction between ON--center and OFF ·--center ganglion cells described by Kuffler (1953), it once was thought that all retinal gangl ion cells were alike. Enroth-Cugell and Robson (1966), how­ever .. showed that ganglion cells in the cat retina are not all the same; some ganglion cells summed the illumination within their receptive field linearly while others did so nonlinearly. The ganglion cells which summed linearly were denoted X cells, and those which summed nonlinearly were denoted Y cells. Other parameters distinguish X and Y cells (Cleland and Levick, 1974) as well as other ganglion cell types (Stone and Hoffman, 1972). The differences between X, '( and other types of gangl ion cells have been reviewed (Aodieck, 1979; Stone, Dreher and Leventhal, 1979; Lennie, 1 9a 1; Stone, 1983; Orban 1984). Y cells have the largest receptive fields in each spot of retina; their axons conduct faster than any other gangl ion cells, and they comprise approximately 5% of the total population of gangl ion oells in the cat retina. X cells have the smallest receptive fields; their axons conduct slower than Y cells'.. and they comprise approximately 40% of the ganglion cells. There are other types of ganglion cells which have intermediate to large receptive fields; these axons conduct very 27 slowly, and such cells constitute 55% of the total population.. This last group of gang I ion cells is heterogeneous and has been denoted W cells .. In a Golgi study of gang I ion cells in the cat retina Boycott and Wassle (1974) recognized that in each spot of retina there are different morphological types of gangl ion cells. Some had large cell bodies and expansive dendritic fields with thick axons; these cells were denoted alpha cells. Other ganglion cells had medium sized somas with smal L densely branching dendritic fields and medium caliber axons; these were denoted beta cells. Other gangl ion cells with small and medium cell bodies with dendritic fields exhibiting a variety of patterns have also been characterized (Leventhal,. Keens and Tork, 1980; Kolb, Nelson and Mariani, 1981). Boycott and Wassle suggested that alpha cells correspond to Y cells; beta. to X cells. and other types to the W cells. Subsequent work (Cleland... Levick and Wassle.. 1975; Saito.. 1983) has confirmed this hypothesis. Dendritic Field Elongation and Orientation One of the fundamental enterprises of biology. relating structure and function ... has succeeded very well in the retina. It was of interest ... therefore, to ascertain if a structural basis was present for the orientation sensitivity of retinal ganglion cells. Recall that Hanlmond (1974) ori­ginally considered the elongation in ganglion cell receptive fields because the dendritic fields had been described as asymmetric. The first systematic investigation of the elongation and orientation of retinal ganglion cell dendritic fields was undertaken by Leventhal and Schall (1983).. The elongation and orientation of each dendritic field was 28 determined with a prooedure oomparable to that employed by Leviok and Thibos (1982) in their analys;s of the physiological orientation sensitivity of ganglion cells. Eighty--eight percent of the ganglion cells had sig-­nificantly oriented dendritic fields. The structural orientation biases for the sannpled dendritic fields nnatched the physiological orientation biases determined by Levick and Thibos ("1982). The elongation of the dendritic fields was the same as the elongation of the receptive fields determined by Hammond (1914). Moreover .. Peichl and Wassle (1983) have demonstrated directly that the contour of individual alpha cell receptive fields follows the shape of the cells' dendritic fields. Levick and Thibos (1982) found that there was a signi ficant tendency for elongated gangl ion cells outside the area centralis to respond best to radially oriented stimuli. Leventhal and Schall (1983) demonstrated that rnost elongated dendritic fields outSide the area centralis are oriented radially. Also, there are no dif­ferences in the elongation or tendency toward the radial orientation across gangl ion cell types. This evidence indicates that the elongation of a ganglion oell dend­ritic field confers upon the cell its physiological orientation sensitivity; the preferred orientation of the cell is given by the orientation of its dendritic field. The gangl ion cell dendritic fields in ferret (Vitek .. Schall and Leventhal.. 1985).. rnonkey (Schal L. Perry and Leventhal .. 1986b) and human (Rodieck. Binmoeller and Dineen, 1985) but not rat (Schall .. Perry and l.eventhal.. 1986a) are also elongated and oriented radially. The most likely explanation why elongated dendritic fields work in the retina and not in the cortex is found in the size of the excitatory 29 afferents relative to the size of the dendritio fields in retina and visual cortex (Tieman and Hirsoh, 1983; Martin and Whitteridge, 1984). The excitatory input to gangl ion cells is provided by bipolar cells (reviewed by Dowling, 1979). Because bipolar axonal fields are so much smaller than the dendritic fields of ganglion cell dendritic fields (Kolb et aI., 1981), the elongation of the dendritic field results in orientation selective convergence as shown in Figure SA. lGNd axons provide the excitatory input to stellate cells in layer 4 of striate cortex (reviewed by Gi Ibert, 1963; Martin, 1965). LGNd neuron axonal fields in cat extend from 500 to 2000 urn (Ferster and LeVay" 1978). Stellate cell dendritio fields in layer 4 span 200--300 urn (Lund" Henry" MacQueen and Harvey 1979). Since the geniculate afferents are so large" the elongation of the receipient dendritic fields cannot result in orientation selective oonvergence as shown in Figure 56. Orientation Sensitivity in the Dorsal La~!al_GenicuJ~!e Nucl~!JS Daniels, Norman and Pettigrew (1977) first demonstrated the ori­entation sensitivity of neurons in the LGNd of cat. They hypothesized that this orientation sensitivity was conferred upon the cells by the cortico­genioulate projection. To test this" they reared kittens viewing only horizontally or vertically oriented lines. The results of reoording from the LGNd were ambiguous. The normal distribution of preferred orienta-­tions was uniform; following either rearing condition the distribution of orientation preferences was peaked at horizontal and vertical. 30 Figure 5. Comparison of the sizes of recipient dendritic fields and afferent axonal fields in the retina and visual cortex. A. Ganglion cell dendritic fields superimposed on an array of bipolar cell axonal fields. An alpha cell dendritic field is represented on the left, and a beta eel L on the right. The dendritic fields receive excitatory input fronl the shaded bipo­lar cells. Because the gang I ion cell dendritic fields are so much larger than the afferent axonal fields .. the shape of the dendritic fields results in an orientation selective sampl ing of the afferents. B. A cortical stellate cell dendritic field in an array of LGNd relay cell axonal fields. The axonal fields which overlap the dendritic field are drawn with a heavier line. Since the axonal arbors are so much larger than the dendritic field .. the shape of the dendritic field does not result in an orientation selective sanlpi ing of the afferents. -------- ...... , .. ,' 31 32 Creutzfeldt and Nothdurft (1979) independently discovered orienta­tion sensitivity in the LGNd by accident. They were attempting to determine how individual neurons in the visual pathway -transferred­complicated visual stinlulL They rnoved a variety of stimuli, from pinwheels to pictures, systematically across each celtlts receptive field and noted the response of the cell to each incremental position of the stimulus. A two-dimensional plot of the cell's evoked activity revealed the pattern of its response to the entire stimulus. The response of some LGNd cells was better to one orientation in the pinwheel than to the others. They did not discuss the ramifications of this discovery .. Lee, Creutzfeldt and Elepfandt (1977) reported that many cells in the monkey LGNd were orientation sensitive. This finding, however, comprised a rninor part of this study and was not further commented upon. Vidyasagar and Urbas (1982) investigated more rigorously the orientation specificity of cells in the cat LGNd. They found that LGNd cells exhibit greater orientation sensitivity with long than short bars; this indicated that the surround of the receptive fields was involved in the orientation tuning of LGNd cells. The orientation tuning of LGNd cells to sine-wave gratings is sharper than that of retinal ganglion cells as measured by Levick and Thlbos (1962).. IontophoreSiS of blcuculline at­tenuates the LGNd neurons· or ientation sensitivity which indicates that GABAergic inhibition is responsible for enhancing the orientation sensitivity of LGNd celts (Vidyasagar .. 1984). The distribution of orienta­tion preferences of the sampled geniculate cells was peaked at horizontal, 33 but the stereotaxio ooordinates at whioh Vidyasagar and Urbas made their penetrations imply that they recorded primarily from the horizontal meridian (Sanderson, 1971). Overall, there was a significant radial correlation between the orientation preferences of the geniculate cells and the polar angle of the receptive fields. Vidyasagar and Urbas also demonstrated that the orientation sensitivity of LGNd cells remains after the corticogeniculate feedback is interrupted.. Orientation sensitivity in the LGNd develops prior to visual experience. In the second week after birth .. 30% of the cells in the kitten LGNd are orientation sensitive (Albus, Wolf and Beckman .. 1993). Retinal Origin of Orientation Specificity in the Visual Pathway Statement of the Problem This discussion wi II be a severe simpl if ication since it wi II consider orientation specificity in isolation from the other visual receptive field properties. A satisfactory model of cortical orientation specificity must account for the following attributes: 1) the anisotropic distribution of orientations represented in different regions of the visual cortex .. 2) the degree of orientation selectivity.. Le., the lack of response to non­optimally oriented stimuli.. 3) the functional architecture .. i.e ... smooth change of orientation with tangential distance and correspondance of orientation in granular and supragranular layers with shifted orientation in the infragranular layers ... and 4) the development in which orientation spe­cificity arises in some neurons independent of visual experience while the orientation specificity of other neurons requires experience. To date there 34 is no model whioh oan satisfaotorally aooount for all of these elements; in this seotion one oandidate is nominated. The recognition that intracortical inhibition plays a major role in determining the degree of orientation selectivity goes only part of the way toward a general theory of orientation specificity. Orientation selectivity can only be achieved by inhibition between cells which have different orientation preferences. The key element missing is how the preferred orientation of cortical cells is originally specified. Theoretical schemes (e.g . ., von der Malsburg., 1914; Bienenstook., Cooper and Munro., 1982) have been presented to deal with this bootstrapping problem through ·self­organizing networks. • These treatments., however., assume that the afferents to the visual cortex are not orientation sensitive and that visual experience wi th oriented contours is necessary for the systern to sort out. As reviewed above. however., some cortical cells as well as retinal gangl ion cells and LGNd eel Is develop orientation specificity independent of experience. The existence of orientation sensitivity in the peripheral visual pathway simplifies the problem of cortical orientation specificity; the incoming sensitivity may provide a foundation for the development of cortical orientation specificity. Specification of Orientation Preferences in the Visual Cortex All orientations are represented in each spot of retina .. but the distribution of the orientation preferences of ganglion cells is not iso­tropic in any spot of retina. It is., therefore" possible to test the hypo-thesis that preferred orientations of cortical cells are related to those of 35 gangl ion oells by oomparing the distribution of orientations represented by topographioa"y corresponding gangl ion cells and cortical neurons. As mentioned, there is an overrepresentation of radial orientations represented by Simple cells in area 17 (Leventhal, 1963) and by neurons with the narrowest receptive fields In area 19 (Leventhal et aI., 1904). Moreover", the distribution of the preferred orientations of Simple cells in areas 17, 1 Band 1 9 subserving any spot of the visual field is statistically indistinguishable from the preferred orientations of the ganglion cells subserving the same spot of the visual field (Schall, Vitek and Leventhal, 19860). This was true in the periphery where the radial orientation is overrepresented as well as in the center where horizontal and vertical orientations are overrepresented in both retina and cortex. The overrepresentation of radial orientations is weaker in the Complex cells in area 17 (Leventhal.,. 1983) and absent in the cells with the widest receptive fields in area 19 (l.eventhal et al.,. 1984). Complex cells in area 17 may be distinguished aocording to receptive field width; most Complex oells in the supragranular layers exhibit narrow reoeptive fields while Complex oells in the infragranular layers exhibit wide receptive fields (Gi Ibert, 1911; Leventhal and Hirsch, 1918). The distribution of the preferred orientations of the Complex celts with narrow receptive fields in area 17 matches the distribution of the orientations represented by topographically corresponding gangl ion cells, but the distribution of the preferred orientations of Complex cells with wide receptive fields in areas 17, 18 and 1 9 did not match the distribution of orientations represented by topographically corresponding gangl ion eel Is (Sohal I et al., 19860). 36 These data provide evidence for the following model, LGNd neurons reoeive exoitatory input from a singfe or few ganglion oells (Clefand et aI., 1971; Cleland and Lee, 1985), Since the distribution of orientations re-­presented in the LGNd matches that in the retina (Vidyasagar and Urbas, 1982), it is plausible that the orientation preference of an LGNd relay neuron is specified by the orientation of its excitatory ganglion cell input. The orientation sensitive LGNd relay neurons may then in turn specify the preferred orientations of the cortical cells to which they provide excitatory input. Cortical neurons rece;ve excitatory input from 10-30 LGNd cells (Tanaka, 1983), so it must be postulated that the preferred orientation of a cortical cel I is some sort of vector sum of the orientation preferences of the afferents. Mutual inhibition between the orientation sensitive cortical cells results in the receptive field organization and orientation selectivity. A diagram of this model is shown in Figure 6. This model provides a more parsimonious explanation for the functional architecture of orientation specific cortical neurons than do previous models, In other areas of the cortex the response properties represented by columns are defined in the peripheral receptors (reviewed in Schall et aI., 1906). A topographic map with local random scatter of orientation sensitive LGNd arbors will naturally result in a srnooth change in orientation with tangential distance in the cortex. In this scheme .. the width of orientation slabs wi II depend on the proportion of LGNd cells representing each orientation. In fact, in the regions of visual cortex subserving peripheral vision the slabs representing the radial orientation •• ••l ~ E ..'..- I-/-_- •.- .co. .-c• . .._ _. ...- '1/- '1/- 31 I 1111 , inhibitory-< '<, excitatory-4 '1/- '1/- Figure 6. A model to illustrate how lateral inhibition between neurons receiving orientation sensitive input can result in orientation selectivity. Four lower level cells which are orientation sensitive provide excitatory input to four upper level cells. The orientation tuning curves of the four afferent cells are drawn beneath each cell; each lower level neuron exhibits a different orientation preference. The four upper level cells mutually inhibit one another (either directly as shown or via an inhibitory interneuron). The orientation tuning curve of an upper level cell is drawn above it. The preferred orientation of the neuron is specified by the orientation preference of its excitatory input as represented by the shaded tuning curve. The cell is inhibited by cells with oblique and horizontal orientation preferences as represented by the dashed curves. The solid curve results when the inhibitory or ientation tuning curves are subtracted from the excitatory tuning curve. Orientation specific lateral inhibition confers upon the cell its orientation selectivity. 38 are wider than those representing nonradial orientations (Schall et aL .. 1986c). The independence of the orientation specificity of neurons in different layers of the cortex may also be accomodated by this rnodeL The supra- and infragranular layers as well as the granular layer receive direct LGNd input (Ferster and LeVay .. 1978; Leventhal.. 1979). Accordingly .. it is possible that the same enhancement of an afferent orientation sensitivity may operate in the different layers. Orientation sensitivity develops in the lGNd before it does in the visual cortex (Albus et aL ... 1983). The presence of orientation sensitivity in the afferents to the visual cortex provides for a clearer understanding of the deveioprnent of cortical orientation specificity and functional architecture. The initial ingrowth of LGNd terminals may be disorganized .. but actiVity-dependent developmental processes can result in ori­entation- specific segregation (Swindale ... 1982; Singer .. 1985). The first neurons in the cortex to exhibit adult orientation selectivity are the Simple cells in layer 4 which receive direct LGNd input; orientation specificity arises in these cells independent of visual experience. The cel Is which require visual experience for the expression of orientation specificity are those which are furthest removed fr'om the geniculate afferents and which represent orientations different from those present in the topographically corresponding area of retina. Certain tests of this hypothesis are possible. It is possible to record simultaneously from a cortical neuron and the ganglion cells (lee et al. .. 1977) or LGNd cells (Tanaka .. 1983) providing input to the neuron.. It is 39 necessary to demonstrate that cort ical neurons and their afferent gangl ion cells and LGNd neurons exhibit similar orientation preferences. It is also possible to record from a cortical cell and reduce the orientation semi­tivity of the geniculate input by injecting bicucull ine into the LGNd. Preliminary evidence indicates that the orientation selectivity of a cortical cell is actually reduced when the orientation sensitivity in the LGNd is reduced (Vidyasagar .. 1985). Finally .. it is possible to experi­mentally change the shape of retinal ganglion cell dendritic fields (Perry and Linden .. 1902; Eysel .. Peichl and Wassle .. 1905). It should be possible to record from the region of visual cortex representing the spot of retina with reoriented dendritic fields and ascertain if the distribution of orientation preferences is different from normal. CHAPTER 2 FACTORS AND EVENTS RESPONSIBLE FOR RETINAL GANGLION CELL DENDRITIC STRUCTURE The nervous system is the product of a superb architect and a sloppy workman, and in his plans the architect took into account the fact that the workman [is sloppy]. W.H. Huggins and J.C.R. Licklinder (1951) 41 Structure and Function of Dendrites The structure of a neuron's dendritic tree is critical to the cell's participation in the neuronal network in which it is embedded. The branching and cable properties of the dendrites dictate to a significant degree the behavior of the cell (Koch, Poggio and Torre, 1982; Rail, 1977). There are many examples in which the structure of a neuron's dendri tic field is clearly related to its behavior. For example .. the shape of visual interneurons in the crayfish (Kirk, Waldrup and Glantz, 1983), of cricket giant interneurons (Bacon and Murphey, 1984) and of cat retinal gangl ion cells (Peichl and Wassle, 1983) specifies the shape of the cell's receptive field. This relation, however, does not always hold; the size and shape of the dendritic fields of neurons in cat striate cortex are not correlated with their receptive field size or preferred orientation (Martin and Whltteridge, 1904). Dendritic tree structure has been described in several analytical investigations. The first such study was accomplished by Bok (1936) who measured the distances between branch pOints in the dendritic trees of cortical pyramidal cells. Sholl (1953, 1955) introduced a popular method of analyzing the coverage of a dendritic field. With the application of computers to the task (Glaser and Van der Loos, 1965) the number of analyses of dendritic structure multiplied (reviewed in Lindsay, 1977; Hillman, 1979). 42 The upshot of the numerous quantitative investigations is that a few fundamental parameters are neoessary and suffioient to desoribe a dendritic field. These parameters include: 1) the diameter of the dendritic trunks arising from the cell body, 2) the diameter of the terminal twigs, 3) the taper of the branch segments, 4) the length of the segments between bifurcations., 5) the number of branches formed at each bifurcation., 6) the ratio between the cross-sectional areas of the daughter branches, J) the orientation and elongation of the entire dendritic field and 8) the mag­nitude and direction of displacement of the center of the dendritic field from the ce II body. Some of these attributes do not vary significantly within a popu­lation of neurons, while other parameters are extremely variable. Those parameters which are invariant probably develop under more intrinsic genetiC control. The variabi lity of the remaining parameters probably results because those elements of dendritic form are influenced by ex­trinsic factors during development. Soma size, the sum of the cross­sectional area of the dendritic trunks, terminal twig diameter and the number of branches at each bifuroation are relatively invariant within a population of cells. The number of dendritic trunks arising from the soma, the length of the branch segments, the ratio of the cross-sectional areas of the daughter branches, the elongation and orientation of the dendritic field and the magnitude and direction of displacement of the center of the dendritic field from the cell body are extremely variable within a population of neurons. 43 Retinal Development An understanding of retinal development must precede an under­standing of the development of ganglion cell dendritic fields. This description will concentrate on the mammalian retina; while many aspects of its development are similar to nonmammalian retina, there are signi­f ioant differences (reviewed by Mann, 1928; Grun, 1982). The sequenoe of events appears to be essentially simi lar in rodents, carnivores and primates. The actual times which wi II be presented are those observed in the cat. The development of the neural retina is not significantly different from the development of other central nervous system structures. Mitosjs ceases, and the neuroblasts migrate to the appropriate level, forming the ganglion cell layer and the inner and outer nuclear layers; sub­sequent neurite outgrowth and synaptogenesis results in the inner and outer plexiform layers. There is a gradient of differentiation and maturation across layers in the vertical dimension of the retina; inner, vitreal elements differentiate before outer, scleral.. There is also a gradient of development within each layer, in the horizontal dimension of the retina; central retina develops before peripheral. Ver t i ca I Deve I opment The sequence of development will be described for the area central is of the retina; the same sequence occurs in peripheral retina with a delay. On embryonic day 21 (E21) of the cat's 65 day gestation the first beta ganglion cells undergo their final mitosis; on E25 alpha cells undergo their last mitosis .. and small ganglion cells begin to differentiate throughout this period (Walsh .. Polley .. Hiokey and Guillery .. 1983; Walsh and Polley .. 1905). Amaorine and photoreceptor cells begin to differentiate at ap- 44 proximately the same time as the ganglion cells; horizontal cells follow shortly, and bipolar cells are the last neurons in the retina to begin dif­ferentiation (Ramon y Cajal, 1960; Sidman, 1961; Morest, 1970). On EJ5-40 the IPL, clear of migrating neuroblasts, separates the ganglion cell layer from a pseudostratified neuroblast layer (Greiner and Weidman# 1980; Stone, Maslim and Rapaport, 1984). On E46 a layer of primitive horizontal cell nuclei are evident within the neuroblastic lamina (Greiner and Weidman# 1980). Around E50 the inner nuclear layer is separated from the outer nuclear layer at the level of the horizontal cell nuclei (Greiner and Weidman# 1980; Rapaport and Stone.. 1982). Cytogenesis slows (Rapaport and Stone .. 1983a); ganglion cell bodies develop distinct sizes (Rapaport and Stone .. 1983b) .. and the first conventional synapses between amacrine and gangl ion cells appear in the IPL (Stone et al .• 1984). Through the remainder of the prenatal period the OPl. becomes more distinct (Greiner and Weidman, 19aO; Rapaport and Stone, 19a2), more con­ventional synapses form in the IPL (Greiner and Weidman, 19aO; Stone et aL .. 1984t and the first ribbon synapses between horizontal cells and pho­toreceptors appear .. but only dyads are seen (Stone £It al.# 1984). At birth the area centralis of the kitten retina is still immature. Stili more conventional synapses appear In the IPL .. but there are no ribbon synapses (Greiner and Weidman, 1980; Stone et aI., 1984). More ribbon dyads form in the OPL (Greiner and Weidman~ 1900; Stone et al." 1 904). Photoreceptor outer segments are sti II matur ing (Greiner and Weidman~ 1980). Small and large ganglion cells can be distinguished at the area 45 centraJis, but they do not exhibit adult appearing NissJ substance (Oonovan, 1966; Tuoker, 1978). In the first postnatal week the OPL thickens as synaptogenesis pro'-' ceeds (VogeL 1978; Rapaport and Stone ... 1982) and the centraL bipolar member of the postsynaptic triad appears (Stone et al. .. 1984). Immature ribbon synapses between bipolar and ganglion cells are first seen in the IPL (Vogel.. 1978; Greiner and Weidman, 1980; Stone et aL, 1984). The gangf ion cell size distribution is not yet adult (Tucker .. 1978; Rapaport and Stone, 19S3b) .. but extensive ganglion cell maturation is characterized by an increase in cytoplasm volume, an increase in the number of organelles and an enlargement of the nuclei (Vogel.. 1978). In the second postnatal week rapid synaptogenesis ... both conventional and ribbon ... continues in the IPL (Vogel... 1918; Greiner and Weidman, 1980; Stone et aL... 1904). Gangl ion cell bodies are morphologically mature (Donovan ... 1966; Vogel... 1978) though smaller than adult (Rapaport and Stone ... 1983b). In the third postnatal week gangl ion cell bodies (Donovan ... 1966; Rapaport and Stone, 1983b) and beta cell dendritic fields (Rusoff and Dubin, 1918) are adult I ike. Final maturation of all layers is achieved by the fourth (Oonovan .. '1966) to sixth (VogeL. 1918) postnatal week. Horizontal Development The centrifugal gradient of retinal development has been appreciated for many years (e.g., Ramon y Cajal, 1960). The oytogenesis of af f of the oells types ocours in a centrifugal wave (Sidman, 1961). The ganglion oells in cat retina appear to pass through their final mitosis in a centrifugal spiral which is centered on the area centralis (Walsh and Polley, 1985). 46 Beta oells in the area oentralis undergo their final mitosis on E2l, and beta oells at the retinal margin undergo their final mitosis on E31; all the beta cells have been born by E35. Alpha cells are first born in the area central is on E25, and the wave of alpha cell birth reaches the retinal margin and conoludes on E3l. Throughout this period ganglion cells with srnall cell bodies are born across the retina. Following the wave of cytogenesis of ganglion cells, the other layers of the retina mature. In this period the ganglion cell density gradient appears as the number of cells in the gangl ion cell layer decreases dramatically (Stone .. Rapaport .. Wi II iams and Chalupa.. 1902). Gangl ion cells accumulate granular cytoplasm and grow into the distinct soma size groups (Rapaport and Stone, 1983b). Simultaneously the OPL appears (Rapaport and Stone, 1982), and cytogenesis ceases (Rapaport and Stone, 19B3a). This process begins at the area centralis around E50 and spreads over the retina in a horizontally elongated wave. The wave extends more into nasal than temporal retina; it reaches the nasal and temporal margins on postnatal day 5 and covers the whole retina by postnatal day 1 D. A similar process has been described in the rabbit retina (Stone, Egan and Rapaport, 1985). The wave begins in temporal retina, but the point of initiation is more variable, and the wave is significantly more elongated than in the cat. The rat retina undergoes a similar centrifugal develop­ment ... but the spatiotemporal pattern is poorly defined (Webster and Rowep 1985; Perry ... personal communication). The wave in human retina begins at the macula and is slightly horizontally elongated (Provis. Van Oriel, Bi IIson and Russel. 1985). 41 Gangl ion Cell Differentiation Ganglion cells are the first retinal neurons to differentiate; the process is not different from that which occurs in other areas of the ner­vous system. Ganglion cell development in mammals has been described by many investigators (e.g., Ramon y Cajal, 196U; Morest, 1910; Hinds and Hinds, 1974; Perry and Walker, 1980a). Ganglion cells undergo their final mitosis at the outer, scleral surface of the primordial retina; both daughter cells tend to be committed to become gangl ion cells. The cells assume a bipolar shape wi th one pro-­cess contacting the schleral surface and another reaching toward the inner, vitreal surface. The per ikarya migrates toward the inner surface. As it nears the inner surface. the vitreal process develops a gro~th cone and elongates into the optic nerve layer .. form ing the axon. Meanwhi Ie, the schleral process detaches and retracts; many thick, spinous processes emit from the perikarya. These processes elaborate with growth cones and fifopodia to contribute to the primitive inner plexiform layer (IPL). The gangl ion cell bodies distribute into a distinct layer. The dendri tic branches appear thicker ~ith more spines and appendages than the mature form. Following this period of exuberant growth .. the dendrites begin to thin; appendages are lost. and gro~th cones and filopodia become less abundant. During the period of remodeling the different morphological classes of gangl ion cells become distinguishable. Thereafter... the dendri tic trees expand to their adult dimensions (Ausoff and Dubin ... 1978). 48 Mechanisms of Dendritic Growth Dependence on Afferents Investigations of dendritio gro~th throughout the nervous systenn have indicated that the afferents induce and direct dendritic growth. Ob-­servations of normal dendritic development indicate that dendritic growth occurs when the afferents arrive (e.g., Ramon y Cajal, 1960; Morest, 1969). Furthermore .. the spatial distribution of a dendritic tree correlates with the arrangement of the afferents. For example .. in the malformed laminae of the neocortex of the neurological mutant .. reeler .. the dendritic fields distribute within the axon-rich strata (Pinto-Lord and Caviness .. 1979). The three dimensional dendritic field shape of the interneurons in the molecular layer of the cerebellum is correlated with the cells' birthdate and final position within the parallel fibers (Rakic .. 1972). If the afferents are removed, the dendritic trees consist of fewer, shorter branches. For example, elimination of the granule cells in the cerebellum results in a reduced distal arbor of Purkinje cell dendrites (Altman and Anderson, 1972; Altman, 1976; Rak;c and Sidman, 1973a,b; Sotelo and Changeux, 1974; Sotelo, 1975; Mariani, Crepel, Mikoshiba, Changeux and Sotelo, 1977; Bradley and Berry, 1976). Elimination of the 01 imbing fibers results in a decrement in the proximal dendrites of Purkinje cells (Sotelo and Arsenio-Nunes, 1976; Bradley and Berry, 1976). Elimination of the cochlear nerve input to brainstem auditory nuclei results in reduced dendritic growth (Parks, 1981; Oeitch and Rubel, 1904). Monocular enucleation results in a redistribution of the dendritic trees of neurons in the visual cortex (Valverde, 1968; Ruiz-Maroos and Valverde, 1970). 49 The relation between afferent and dendritic growth appears to depend on the activity of the afferents. The dendritic trees of neurons in the visual cortex are affected by the early visual experience of the animal (Volkman and Greenbough, 1972; Rutledge, Wright and Duncan, 1974; Parnavelas and Globus, 1976; Borges and Berry, 1976; Schapiro and Vukovich, "1970; Coleman, flood, Whitehead and Emerson, 198'1; Tieman and I"'lirsch, 1 983). Auditory deprivation leads to reduced dendritic growth in chick brainstem audi tory nuclei (Conlee and Parks, 1983). Synaptogenesis and Dendritic Growth Eleotron miorosoopio studies have demonstrated that there is a correlation between synaptogenesis and dendritic growth (Hayes and Roberts .. 1973; Kristt, 1978; Jacoby and Kimmel .. 1982). Immature syn-apses appear an the filopodia of dendritic growth cones, and mature synapses appear on the dendritic shafts (Skoff and Hamburger .. 1974; Vaughn .. Henrikson and Grieshaber .. 1974). This evidence prompted a filo­podial attachment hypothesis of dendritic growth; the filopodia of dend­ritic growth cones are stabilized by synapses with the afferents. and the dendrites extend where the filopodia attach. Topographic Dendritic Growth Mast studies have considered the three-dimensional dendritic shape across layers or within a nucleus. Consider. however .. the dendritic growth in the plane of topography... i.e .... the horizontal plane of the lamina. perpendicular to the ingrowing afferents. In layer 4 of the rodent so­matosensory cortex the oells are arranged in barrels which oorrespond to the distribution of facial vibrissae (Woolsey and Van der Loos .. 1970). The 50 dendritio trees of oells on the rim of the barrel extend into the oenter of the barre. (Woolsey, Dierker and Wann, 1975). If a row of vibrissae is removrd in the neonate .. the corresponding barrels coalesce .. and the dendritic fields change their distribution accordingly (Steffan and Van der Loos ... 1980; Harris and Woolsey ... 1981). Spatial Distribution of Gangl ion Cell Dendrites Retinal Coverage and Neuronal Mosaics A requirement of retinal development is to cover itself completely with all cell types so that no spot of the visual field goes unseen; hence .. the horizontal growth of dendrites is the major preoccupation of investigators of retinal cell structure and development. The cells in the retina are not distributed randomly in the horizontal dimension. Photoreceptors. hori­zontal cells and gangl ion cells are arranged in a mosaic (Wassle and Riemann. 1978). Alpha and beta ganglion cells which ramify in one sublamina of the IPL are arranged in a mosaic which is independent of the mosaio of the ganglion cells whioh ramify in the other sublamina (Wassle, Boycott and II ling, 1981a; Wassle, Peichl and Boycott, 1981c). Having retinal cells distributed precisely is necessary for vision. If the oells were distributed randomly, excessive dendritic growth would be required to completely cover the retinal surface, and interconnections between the laminae would be less efficient. Moreover, regular spacing of sampl ing pOints is required for high spatial resolution (French, Snyder and stavenga ... 1911). In fact ... the spacing of ON and OFF beta cells In the re­tina is correlated with the visual resolution of cats (Hughes .. 1981; Wassle et a I... 1981 a). The mosaics are not crystall ine.. but this measure of 51 sloppiness prevents the interferenoe effeots whioh occur when regular arrays are superimposed. Three possible mechanisms have been proposed to account for the observed regularity in retinal cell distribution (Wassle and Riemann, 1978): 1) contact inhibition between neurons of the sanle type, 2) com­petition for space/afferents between neurons of the same type and 3) the pattern of cell lineage. The mosaic is probably not due to the original layout at cel I birth. Mammal ian retinal development involves an overproduction of ganglion cells followed by a period of celJ death in the mouse (Young, 1904), rat (Jeffery and Perry, 1901; Perry, Henderson and Linden, 1983; Cunningham, Mohler and Giordano, 1982; Lam, Sefton and Bennett, 1982; Dreher, Potts and Bennett, 1983;), hamster (Sengelaub and Finlay, 1982), cat (Ng and Stone, 1962; Stone et aI., 1962), monkey (Rakic and Riley, 1983) and human (Provis et aI., 1985). Monocular enucleation attenuates the gangl ion cell death (Wi II iams, Bast ian i and Chalupa, 1983; Chalupa, Wi II iams and Henderson, 1984) and expands the gangl ion cells projecting to the ips; lateral side of the brain (Jeffery and Perry, 1982; Insausti, Blakemore and Cowan.. 1984). This indicates that binocular competition of the axon terminals is responsible for a measure of the cell death; furthermore, the process requires ganglion cell activity (Fawcett, O-Leary and Cowan, '1964). Following section of one optic tract the number of cells in the contralateral retina which project ipsilaterally is greater than normal in the rat (Linden and Perry., 1982) and the oat (Jaoobs., Perry and Hawken., 1984). These ipsi laterally prOjecting cells survive because the reduction of 52 neighboring ganglion cells in the nasal retina enables the ipsilaterally projeoting oells to oompete more suooessful Jy for afferents. Thus, only those ganglion cells with functionally valid connections of both dendrites and axon terminals survive. per,:.dritic Interaclio!l! The influence of neighboring gangl ion cells on dendritic structure has been investigated in the rat (Linden and Perry ... 1982; Perry and Linden .. 1982) and cat (Au I t, Schall and Leventhal, 1985; Eysel et aI., 1985). If the ganglion cells in a spot of developing retina are eliminated .. the dendritic trees of gangl ion cells on the border extend into the depleted area. This indicates that ganglion cell dendrites tend to grow where other dendrites are not. In the normal cat retina the center of the dendritic field of most ganglion cells in a spot of retina tend to be displaced down the ganglion cell density gradient in that spot of retina; this dendritic displacement results in the dendritic field centers being arranged in a more precise mosaic than are their cell bodies (Schall and Leventhal.. 1906). This evidence indicates that the dendritic interactions revealed by the experimental manipulations are operative during normal development of the cat retina. It is instructive to note that ganglion cell dendritic fields in the rat retina are not displaced down the shallow density gradient (Schal L. Perry and Leventhal, 1986a). Oendritic Field Elongation and Orientation Another attribute of retinal ganglion cell dendritic trees is their elongation and systematic orientation (Leventhal and Schall .. 1983)" The process responsible for the elongation and orientation has not been eluci- 53 dated, but several proposals have been advanced. Stretching of the dendritic fields due to retinal expansion might be responsible (Levick and Thibos, 1982). Mastronarde, Thibeault and Dubin (1984) demonstrated, however, that retinal expansion is insufficient to account for the elonga­tion and radial orientation of the ganglion cell dendritic trees. Dendritic growth within the ganglion cell mosaic (wassle, Peichl and Boycott, 1901b) may result in the elongation and orientation of the dendritic trees (Thibos and Levick~ 1985)" Other than a mutual radial tendency~ however" dendritio fields with simi lar orientations are not olustered in oat retina (Sohall et al,,~ 19860)" Direoted growth of the dendrites down the ganglion cell density gradient may be responsible for the elongation and orientation of the dendritic fields (Eysel et at..~ 1905) .. This cannot be, however, since the direction of gangl ion cell density gradient is not radial due to the visual streak. Furthermore.. neither the magnitude nor direction of dendritic displacement predict the elongation or orientation of ganglion cell dendritic fields. Therefore, an alternative explanation of ganglion cell dendritic field elongation and orientation is considered. Relation Between Retinal Development and Dendritic Field Elongation and Orientation From around E50 to P10 the cat retina undergoes a period of nlaturation as described above; this process is initiated at the area centra lis and spreads over the retina In a horizontally elongated wave (reviewed by Rapaport and Stone, 1984). The expansion of the mature area can be described with a vector. The angle of the vector represents the di­rection of expansion, and the length of the vector represents the relative 54 magnitude. The mean orientation of the ganglion cell dendritic fields in a spot of retina is correlated with the angle of the wave vector in that spot of retina. The elongation of the dendritic fields in a spot of retina is correlated with the relative magnitude of the wave of maturation in that spot of retina (Schall and Leventhal, 1986). The wave of maturation hypothesis can explain many elements of the elongation and orientation of the ganglion cell dendritic fields in different parts of the retina. Foremost .. because the wave of maturation begins at the area centralis .. it explains why the dendritic fields are oriented with respect to the area centra lis and not some other point on the retina. The horizontal elongation of the wave of maturation accounts quantitatively for the horizontal deviation from radial of the dendritic field ori­entations. The horizontal elongation of the wave of maturation also accounts for the finding that the dendritic fields along the horizontal meridian are more elongated than their counterparts on the oblique or vertical meridians .. The wave of matur at ion hypothesis can be tested in a number of ways. One test is an investigation of the relation between the geometry of the process of maturation and ganglion cell dendritic elongation and ori­entation in other species. Gangl ion cell dendritic fields in the macaque retina are elongated and oriented radially with respect to the fovea; the radial orientation and elongation is more pronounced on the horizontal meridian than on oblique and verticaL. and the degree of horizontal deviation from radial is less than in the cat (Schall et aI., 1986b). As mentioned, the wave of maturation in the human retina forms first at the 55 fovea and spreads over the retina in a slightly horizontally elongated wave. (Provis et al., 1985). The degree of elongation of the mature area, however, is not as extreme as that observed in the cat which correlates with the reduced horizontal deviation in the ganglion cell dendritic field orientations. In the rat retina the gangl ion cell dendritic fields are less elongated than those in primate or carnivore., and there is no systematic orientation (Schall et aL., 1 g86a). In correspondance., the wave of maturation in the rat retina is poorly defined spatiotemporally (Webster and Rowe .. 1985; Perry, personal communication). The wave of maturation in the rabbit retina begins in far temporal retina and is more horizontally elongated than in the cat (Stone et aI., 1985). The elongation and orientation of the ganglion cell dendritic fields throughout rabbit retjna can be predicted from the hypothesis. There should be an overall tendency to be oriented horizontally, and the only region to contain predominantly vertically oriented dendritic fields should be temporal retina outside of the visual streak.. If the elongation and orientation do not oorrespond to the predictions .. then the hypothesis wi II be falsified. Another test of the hypothesis is to examine the morphology of other retinal neurons since the wave of maturation passes through all the layers of the retina. In the rabbit retina horizontal cells are elongated. and there is a clear tendency for the horizontal cells along the visual streak to be oriented horizontally (Kolb and Normann, 1982; Bloomfield and Miller .. 1983). In the cat retina horizontal cells are elongated .. but there has been no systematic analysis of their orientation {Wassle., Peichl 56 and Boyoott, 1978). Amaorine oell dendritio trees in rat (Perry and Wolker, 1980b) and rabbit (Tauohi and Mas I and, 1984; Famiglietti .. 1985) retina are elongated; in the rabbit the arbors tend to be oriented horizontally. Another test of the hypothesis is to observe the tinle course of normal development of gangl ion cell dendritic trees in different parts of the retina at different ages. It wi II be of interest to compare the structure of the dendritic fields within the mature area to those outside the mature area. It may be that there is a temporal coincidence between the passage of the wave of maturation over a spot of retina and the imposition of the elongation and systematic orientation of the dendritic fields in that spot of retina .. but this is not a necessary condition of the hypothesis. Conclusion The two major insights from the original work disoussed in this overview are 1) the orientation specificity of cortical cells may depend on the orientation sensitivity of retinal ganglion cells which arises from the shape of the dendritic fields and 2) the elongation and orientation of gangl ion cell dendritic fields are related to the geometry of retinal matu­ration. These results may provide for a unified view of the visual pathways and visual behavior across species" Animals whioh do not rely on visually guided movements suoh as mouse (Drager and Olsen .. 1981) .. hamster (Tiao and Blakemore .. 1976a) and rat (Fukuda .. 1977) do not possess retinae with specialized regions of 57 elevated oell density. As described" ganglion celt dendritic fields in the rat are somewhat elongated and unoriented relative to any pOint on the retina (Schall et al., 1986a). Further, a low proportion of the neurons in the mouse (Drager, 1975), hamster (liao and Blakemore, 1976b) and rat (Shaw, Yinon and Auerback, 1975; Wiesenfeld and Kornel, 1975) visual cortex exhibit selective response properties,. The rabbit retina is characterized by a horizontal ridge of elevated cell density (Hughes, 1971; Provis, 1919) which provides for acute vision across the hor i20n. Some ganglion oells in the rabbit visual streak are orientation seleotive, and most prefer horizontally oriented stimuli (Levick, 1967). In the rabbit visual cortex approximately 70% of the neurons are orientation selective. and there is an overrepresentation of horizontal (Murphy and Berman. 1978). Visually guided movements of the eyes and limbs dominate the repetoire of cats and monkeys. The retinae of cat (Hughes .. 19-/5; Stone .. 1978) and monkey (Stone and Johnston .. 1981; Perry and Cowey .. 1986) exhibit special ized areas of elevated cell density. The spot of peak density provides for the most acute vision and subserves the center of view. Retinal ganglion cell dendritic fields in cat (Leventhal and Schal I.. 1983) and monkey (Schall et aL, 1986b) are significantly elongated and oriented relative to the point of peak acuity_ Essentially all 01 the neurons in the visual cortex 01 cat (Hubel and WieseL 1962) and monkey (Hubel and Wiesel. 1 968) exhibit selective response properties. Retinal topography is an adaptation to a species' I ifestyle in its particular habitat (reviewed by Hughes ... 1977).' As visually guided behavior 58 beoomes more important to a speoies, speoialized retinal topography is seleoted for. Retinal topography is a oonsequenoe of the pattern of retinal development, and the developmental process which results in retinal topography may also be responsible for retinal neuronal morphology. The structure of a neuron results in functional properties which provide a foundation for response seleotivlty in higher visual oenters. Finally .. a greater variety and specificity in the responses of visual neurons provides for more sophisticated visual behavior. SQ References Albus, K. (1975) A quantitative study of the projection area of the central and the paracentral visual field in area 17 of the cat. II. The spatial organization of the orientation domain. Exp. Brain Res. 24:181-202. Albus ... K. (1979) 14C-deoxyglucose mapping of orientation subunits in the cats visual cortical areas. Exp. Brain Res. 37:609-613. Albus, K. and BI Sieber (1984) On the spatial arranaement of ;so­or ientation bands in the eat's visual cortical areas 17 and 18: A 14C-deoxyglucose study. Exp. Brain Res. 56:384-388. Albus, K. and W. Wolf (19a.4) Early post-natal development of neuronal function in the kitten's visual cortex: A laminar analysis. J. Physiol. (Lond.) 348: 153-185. Albus" K,,, W. Wolf and R. Beckman (1983) Orientation bias in the response of kitten LGNd neurons to moving light bars. Oev. Brain Res. 6:308-313. Altman" J. (1976) Experirnental reorganization of the cerebellar cortex. VII. Effects of late X - irradiat ion schedules that interfere wi th cell acquisition after stellate cells are formed. J. Compo Neurol. 165:65-76. Altman" J. and W.J. Anderson (1972) Experimental reorganization of the cerebellar cortex. 1. Morphological effects of elimination of all microneurons with prolonged X-irradiation started at birth. J. Compo NeuroL 146:355-406. Andrews .. D.P. (1967) Perception of contour orientation in the central fovea. Vision Res. 7:975-1013. Appetle ... S. (1972) Perception and discrhnination as a function of stimulus orientation: The ·obl ique effect· in man and animals_ Psych. Bul L 78:266-278. Ault .. S.J .... J.O. Schall and A.G. Leventhal (1905) Experimental alteration of cat retinat ganglion cell dendritic field structure. Soc. NeuroscL Abstr. 11: 15. Bacon .. J.P. and R.K. Murphey (1904) Receptive fields of the cr icket giant interneurones are related to their dendritic structure. J. Physio!. (Lond.) 352:601-623. 611 Bauer, R. (1982) A high probability of an orientation shift between layers 4 and 5 in central parts of the cat striate cortex. Exp. Brain Res. 48:245-255. Bauer,. A .. ,. B .. M. Dow,. A .. Z. Snyder and R. Vautin (1983) Orientation shift between upper and lower layers in monkey visual cortex. Exp. Brain Res. 50: 133-145. Berkley, M.A. and J.M. Sprague (1919) Striate cortex and visual acuity functions in the cat .. J. Comp. Neurol. 187:679-702. Berkley,. M.A.,. F. Kitterle and D.W. Watkins (1915) Grating visibility as a function of orientation and retinal eooentrioity. Vision Res. 15:239-244. Berman, N. and M. Cynader (1972) Comparison of receptive-field organization of the super ior coli iculus in Siamese and norrYlal cats. J. Physiol. (Lond.) 224:363-389. Bienenstock, E.L., L.N. Cooper, and P.W. Munro (1982) Theory for the development of neuron selectivity: Orientation specificity and binocular interaction in visual cortex. J .. Neuroscience. 2:32-46. Blakemore, C. and G.F. Cooper (1970) Development of the brain depends on the visual environment. Nature. 228:477--478. Blakemore, C. and E.A. Tobin (1972) Lateral inhibition between ori­entation detectors in the cat's visual cortex. Exp. Brain Res. 15:439-440. Blakemore, C. and R.C. Van Sluyters (1915) Innate and environmental fac­tors in the development of the kitten's visual cortex. J. Physio). (Lond.) 248:663-716. Blakemore ... C .... A.H.S. Carpenter and M.A. Georgeson (1970) Lateral inhi­bition between orientations detectors in the human visual system. Nature 228:37-39. Bloomfield. S.A. and R.F. Mi Iler (1982) A physiological and morphological study of the horizontal cell types of the rabbit retina. J. Comp. NeuroL 208:288-303. Bok ... S.T. (1936) The branching of the dendrites in the cerebral cortex. Proc. Acad. ScL Arnst. 36:1209-1218. Borges .. S. and M. Berry (1976) Preferential orientation of stellate cell dendrites in the visual cortex of the dark reared rat. Brain Res. 112:141-147. 61 Bousfield, J.D. (1977) Columnar organization and the visual oortex of the rabbit. Brain Res. 136: 154-1 58. Boycott, B.B. and J.E. Dowling (1969) Organization of the primate retina: Light microscopy. Phil. Trans. R. Soc. Lond. B 255:109-184. Boycott, B.B. and H. Wassle (1974) The morphological types of ganglion oells of the domestic oat's retina. J. Physiol. (Lond.) 24:391-419. Bradley, P. and M. Berry (1976) The effects of reduced climbing and parallel fibre input on Purkinje cell dendritic growth. Brain Res. 109:133-151. Brown. J.E. and 0. Major (1965) Cat retinal ganglion cell dendritic fields, Exp. Neurol. 15:70--78. Buisseret.. P. and M. Imbert (1976) Visual cortical celts: Their developmental properties in normal and dark-reared kittens_ J. PhysioL (Lond.) 255:511-525. Sullier. J. and G.H. Henry (lg7Y) Ordinal position of neurons in cat striate curtex. J. Neur ophysiol. 42: 1251-1263. Burr ... D ... C. Morrone and L. Maffei (1981) Intra-cortical inhibition prevents simple cells from responding to textured visual patterns. Exp. Brain Res. 43:455-458. Carpenter.. R.H.S. and C. Blakemore (1973) Interaotions between orientations in human vision. Exp. Brain Res. 18:287-303. Chalupa.. L.M.... R.W. Williams and Z. Henderson (1984) Binocular interaction in the fetal cat regulates the size of the ganglion cell population. NeuroscL 12:1139-1146. Chow .. K.L ... A.H. Masland and D.C. Stewart (1971) The receptive field characteristics of striate cortical neurons in the rabbit. Brain Res. 33:337-352. Cleland .. B.G. and B.B. Lee (1985) A comparison of visual responses of cat lateral geniculate nucleus neurones with those of ganglion cells afferent to them. J. PhysioL (Lond.) 369:249-268. Cleland. B.G. and W.R. Levick (1974) Brisk and sluggish concentrically organized gangl ion cells in the eat's retina. J. Physiol. (Land.) 240:421-456. Cleland. B.G .... M.W. Dubin and W.R. Levick (1971) Simultaneous recording of input and output of lateral geniculate neurones. Nature New BioL 231: 191-192. 62 Cleland, B.G., W.R. Levick and H. Wassle (1975) Physiological identification of a morphological class of cat retinal gangl ion cells. J. Physiol. (Lond.) 248; 151--171. Coleman, P.O., D.G. Flood, M.C. Whitehead and R.C. Emerson (1981) Spatial sampling by dendritic trees in visual cortex. Brain Res. Z14:1-Z"1. Colonnier, M. (1964) The tangential organization of the visual cortex. J. Anat. (Lond.) 98:321-344. Conlee, J.W. and T.N. Parks (1983) Late appearance and deprivation­sensitive growth of permanent dendrites in the avian cochlear nucleus (Nuc. Magnocellularis). J. Compo NeuroL 211:216-226. Creutzfeldt, O. and M. Ito (1968) Functional synaptic organization of primary visual cortex neurones in the cat. Exp. Brain Res. 6:324-352. Creutzfeldt ... 0.0. and Nothdurft (1979) Representation of complex visual stimuli in the brain. Naturwissensch. 65:307-318. Creutzfeldt ... 0.0 .... U. Kuhnt ... and L.A. Benevento (1974) An intracellular analysis of visual cortical neurones to moving stimul i: Responses in a co·-operative neuronal network. Exp. Brain Res. 21:251-274. Cunningham... T .J.... LM. Mohler and D.L. Giordano (1982) Naturally ocourring neuron death in the ganglion cell layer of the neonatal rat: Morphology and evidence for regional correspondence with neuron death in superior call iculus. Dev. Brain Res. 2:203-215 Daniels, J.D., J.L. Norman and J.D. Pettigrew (1911) Biases for oriented moving bars in lateral genioulate neuoleus neurons of normal and str ipe-reared cats. Exp. Brain Res. 29: 155 -112. Deitch, J.S. and E.W. Rubel (1984) Afferent influences on brain stem auditory nuclei of the chicken: Time course and specificity of dendritic atrophy following deafferentation. J. Compo NeuroL 229:66-79. Diamond, I.T. and W.C. Hall. (1969) Evolution of neocortex. Science 164:251-262. Donovan ... A. (1966) The postnatal development of the cat retina. Exp. Eye Res. 5:249-254. Oowl ing ... J.E .. (1979) Processing by local circuits: The vertebrate retina as a model system. in The Neurosciences: Fourth Study Program. F.O. Schmitt and F.G. Worden (eds.) MIT Press. Cambridge. 63 Drager, U. (1975) Receptive fields of single cells and topography in mouse visual cortex. J. Comp. Neurol. 169:269-290. Drager., U. and J.F. Olsen (19B1) Ganglion cell distribution in the retina of the mouse. Inv. Ophthafmol Vis. Sci. 20:285-293. Dreher .. B. and K.J. Sanderson (1973) Receptive field analysis: Responses to moving visual contours by single lateral geniculate neurones in the cat. J. Physiol. (Lond.) 234:gS-118. Dreher .. B ... R.A. Potts and M.R. Bennett (1983) Evidence that the early postnatal reduction in the number of rat retinal ganglion cells is due to a wave of gangl ion cell death. NeuroscL Lett. 36:255-260. Dreher .. B ... A.J. Sefton .. S. Y .K. Ni and G. Nisbett (1985) The morphology .. number.. distribution and central projections of class I retinal ganglion oells in albino and hooded rats. Brain Behav. Evol. 26: 1 0- 48. Enroth-Cugell. C. and J.G. Robson (1966) The contrast sensitivity of retinal ganlgion cells of the cat. J. Physiol. (Lond.) 187:517-552. Eyset U.T .• L. Peichl and H. Wissle (1985) Dendritio plasticity in the early postnatal fel ine retina: Quantitative characteristics and sensitive period. J. Compo Neurol. 242:134-145. Fahle. M. and V. Braitenberg (1983) Curvature detection in the central and peripheral visual field of human subjects. Neurosci. Lett. 14:S 108. Famiglietti. E.V. (1985) Starburst amacrine cells; Morphological constancy and systematic variation in the anisotropic field of rabbit retinal neurons. J. Neurosci. 5:562-511. Fawcett. J.W ... D.D.M. O'Leary and W.M. Cowan (1984) Activity and the control of gangl ion cell death in the rat retina. Proc. Natl. Acad. Sci. 81 :5589-5593. Ferster. O. and S. LeVay (1918) The axonal arborizations of lateral geniculate neurons in the striate cortex of the cat. J. Compo Neurol. 182:923--944, Fisken" R.A ... L.J. Garey and T.P.S. Powell (1975) The intrinisic association and commissural connections of area 17 of the visual cortex. Phi I. Trans. A. Soc. Lond. B 212:487-536. Fregnac" Y. and M. 1m bert (1918) Early development of visual cortical cells in normal and dark-reared kittens: Relationship between orientation selectivity and ocular dominance. J. Physiol. 278:27-44. 64 Fregnao, Y. and M. Imbert (1 984) Development of neuronal seleotivity in primary visual cortex of cat. PhysioL Rev. 64:325-434. French .. A.S ... A.W. Snyder and D.G. Stavenga (19-'7) Image degradation by an irregular retinal mosaic. BioI.. Cybern. 27:229-233. Fries, W." K. Albus and 0.0. Creutzfeldt (1977) Effects of interacting visual patterns on single cell responses in eat's striate cortex. Vision Res. 17:1001-1008. Gibson, J.J. (1950) The Perception of the Visual World. Boston. Houghton Mifflin. Gibson,. J.J. (1966) The Senses Considered as Perceptual Systems. Boston. Houghton-Miffl in. Gi Ibert, C.D. (1983) Microcircuitry of the visual cortex. Ann. Rev. Neurosci.6:217-247. Gilbert, C.D. (1977) Laminar differences in receptive field properties of cells in cat primary visual cortex. J. Physiol. (Lond.) 268:391-421. Gilbert, C.D. and T.N. Wiesel (1983) Clustered intrinsic connections in cat visual cortex. J. Neurosci. 3: 1116-1133. Glaser,. E.M. and H. Van der Loos (1965) A semi-automatic computer microscope for the analysis of neuronal morphology. IEEE Tran. Biomed. Eng. 12:22-31. Greiner, J.V. and T.A. Weidman (1980) Histogenesis of the cat retina. Exp. Eye Res. 3U:439-453. Grun, G. (1982) The Development of the Vertebrate Retina: A Comparative Survey. New York. Springer-Verlag. Hammond .. P. (1974) Cat retinal ganQ.lion cells: Size and shape of receptive field centres. J. PhysioL (Lond.) 242:99---118. Hammond .. Po and D.M. MacKay (1977) Differential responsiveness of simple and complex cells in cat striate cortex to visual texture. Exp. Brain Res. 30:275-296. Harris .. A.M. and T.A. Woolsey (19B1) Dendritic plasticity in mouse barrel cortex following postnatal vibrissa foil icle damage. J. Compo NeuroL 196:357-376. Hartline .. H.K. (1938) The response of single optic nerve fibers of the vertebrate eye to illumination of the retina. Am. J. PhysioL 121:4(]0-415 65 Hartl ine, H.K. (1940) The receptive fields of optic nerve fibres. Am. J. Phys to I. -I J 0: 690 - 699. Hayes, B.P. and A. Roberts (19-13) Synaptic junction development in the spinal cord of an amphibian embryo: An electron microscope study. Z .. Zeit forsoh. 137:251-269 .. Heggelund .. P. (1981 a) Receptive field organization of simple cells in cat striate cortex. Exp" Brain Res. 42:89-98" Heggelund .. P. (1981b) Receptive field organization of complex cells in cat striate cortex. Expo. Brain Res .. 42:99-107. Heggelund .. P. and J. Moors (1983) Orientation selectivity and the spatial distribution of enhancement and suppression in receptive fields of cat striate cortex cells. Expo. Brain Res. 52:235-247. Henry, G.H ... B. Dreher and P.O. Bishop (1974) Orientation specificity of cells in cat striate cortex. J. Neurophysiol. 37:1394-1409. Hess, R. K. Negishj and O. Creutzfeldt (1975) The horizontal spread of intracortical inhibition in the visual cortex. Exp. Brain Res. 22:415-419. Hillman .. D.E. (1979) Neuronal shape parameters and substructures as a basis of neuronal form. in The Neurosciences: Fourth Study Program. F.D. Schmitt and f.G" Worden (eds.) MIT Press. Cambridge. Hinds ... J.W. and P.L. Hinds (1974) Early ganglion cell differentiation in the mouse retina: An electron microscopic analysis utilizing serial sections_ Dev. BioI. 37:381-416. Hirsch ... H.V.B. and A.G. Leventhal (1978) Functional modification of the developing visual system. in Handbook of Sensory Physiology. volume IX" M" Jacobson (ed.) Springer---Verlag. Berlin. Hirsch... H.V.B. and O.N. Spinelli (1970) Visual experience modifies distribution of horizontally and vertically oriented receptive fields in cats. Science. 168:869-87'1. Hirsch ... H.V.B. and O.N. Spinelli (1971) Modification of the distribution of receptive field or ientation in cats by selective visual exposure during development. Exp. Brain Res. 12:509-527. Honrubia ... F.M. and J.H. Elliott (1970) Dendritic fields of the retinal gangl ion cells in the cat. Arch. OphthoL 84:221-226. 66 Horton, J.e. and Sherk (1904) Receptive-field properties in the cat's lateral geniculate nucleus in the absence of retinal on-center input. J. Neurosci. 4:374-380. Hubel, D.H. (1958) Cortical unit responses to visual stimuli in non­anesthetized cats. Am. J. Ophthalmol. 46:110-122. Hubel, D.H. (1959) Single unit activity in striate cortex of unrestrained cats. J. Physiol. (Lond.) 147:226-238. Hubel, D.H. (1960) Single unit activity in lateral qeniculate body and optic tract of unrestrained cats. J. PhysioL (Lond.) 150:91-109 .. HubeL D.H .. (1982) Exploration of the primary visual cortex 1955-1978. Nature 299:515--524. Hubel, D.H. and T.N. Wiesel (1959) Rece~tive fields of single neurones in the eat's striate cortex. J. Physiol. (Lond .. ) 148:514-591. Hubel, D.H" and T.N. Wiesel (1960) Receptive fields of optic nerve fibres in the spider monkey. J. Physiol. (Lond.) 154:572-580. Hubel, D.H. and T .N. Wiesel (1962) Receptive fields, binocular interaction and functional architecture in the eat's visual cortex. J. Physiol. (Lond.) 160:106-154. Hubel, O .. H. and T.N. Wiesel (1963) Receptive fields of cells in striate cortex of very young, visually inexperienced kittens. Hubel, D.H. and T.N. Wiesel (1963) Shape and arrangement of columns in eat's striate cortex. J .. Physiol.. (Lond.) 165:559-568. 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 .. 28:229-289. Hubel, O"H. and T .N. Wiesel (1968) Receptive fields and functional architecture of monkey striate cortex. J. Physiuol. (Lond.) 195:215-243. Hubel, D.H. and T.N. Wiesel (1974) Sequence regularity and geometry of orientation columns in the monkey striate cortex. J. Comp. Neurol. 158:267-294 .. Hubel, D.H., T.N. Wiesel and M.P. Stryker (1978) Anatomical demonstration of orientation columns in macaque monkey. J. Conlp. Neurol. 177:361-380. 67 Huggins, W.H. and J.C.R. LickHnder (1951) Place mechanisms of auditory frequency analysis. J. Acoust. Soc. Am .. 23:290-299. Hughes .. A. (1971) lopographical relationships between the anatomy and physiology of the rabbit visual system. Documenta Ophthal. 30:33-159. Hughes .. A. (1915) A quantitative analysis of the cat retinal gangl ion cell topography. J. Compo Neurol. 163: 107 -128. Hughes" A. (1977) The topography of vision in mammals of contrasting life style: Comparative optics and retinal organization. in Handbook of Sensory Physiology. F. Crescitelli (ed.) New York: Springer-Verlag. pp 615-137. Hughes .. A. (1981) Cat retina and the sampling theorem; the relation of transient and sustained brisk-unit cut-off frequency to alpha and beta-mode oell density. Exp. Brain Res. 42:196-202. Humphrey, A.L. and T.T. Norton (1980) Topographic organization of the orientation column system in the striate cortex of the tree shrew (TuDaia Q.lli). 1. Microelectrode recording. J. Compo Neurol. 192:531-547. Humphrey.. A.L... L.C. Skeen and T.T. Norton (19aO) Topographic organization of the orientation column system in the striate cortex of the tree shrew (Tupaia !lll!>. II. Oeoxyglucose mapping. J. Compo Neurol. 192:549--566. Hyvarinen, J. (1982) The Parietal Cortex of Monkey and Man. Springer­Verlag. New York. Insausti, R ... C. Blakemore and W.M. Cowan (198'1) Ganglion oell death during development of ipsilateral retino-coll icular projection in golden hamster. Nature. 308:362-365. Jacobs .. D.S ... V.H. Perry and M.J. Hawken (1984) The postnatal reduction of the uncrossed projection from the nasal retina in the cat. J. Neurosci.4:2425-2433. Jacoby. J. and C.B. Kimmel (1982) Synaptogenesis and its relation to growth of the postsynaptic cell: A quantitative study of the developing Mauthner neuron of the Axolotl. J. Compo Neurol. 204:364-3'16. Jeffery. G. and V.H. Perry (1902) Evidence for ganglion cell death during development of the ipsi lateral projection in the rat. Dev. Brain Res. 2:176-180. ' 68 Jung, R" (1958) Excitation, inhibition and coordination of cortical neurones. Exp. Cell Res. Suppl. 5:262-271. Jung, R. (1975) Some European neuroscientists: A personal tribute. in The Neurosciences: Paths of Discovery. F .G. Worden, J.P Swazey and G. Adelman (eds.) Cambridge. MIT Press. Kirk ... M.D .... B. Waldrup and R.M.Glantz (1983) A quantitative correlation of contour sensitivity with dendritic density in an identified visual neuron. Brain Res. 274:231-237. Kline ... M. (1980) Mathematics: The Loss of Certainty. Oxford Univ. Press. New York. Koch. C .... T. Poggio and V. Torre (1982) Retinal ganglion cells: A functional interpretation of dendritic morphology. Phil. Trans. R. Soc. Lond. B 298:227-264. Kolb .. H. and R.A. Normann (1982) A-type horizontal cells of the superior edge of the I inear visual streak of the rabbit retina have oriented ... elongated dendritic fields. Vision Res. 22:905-916. Kolb ... H .... R. Nelson and A. Mariani (1981) Amacrine cells bipolar cells and ganglion cells of the cat retina: A Golgi study. Vision Res. 21 : 1081-1114. Komatsu ... Y. (1983) Oeveloprnent of cortical inhibition in kitten striate cortex investigated by a slice preparation. Dev. Brain Res. 8: 136-139. Kristt, O.A. (1 978) Neuronal differentiation in somatosensory cortex of rat. 1. Relationship to synaptogenesis in 1st postnatal week. Brain Res. 150:467-486. Kruger ... J. and M. Bach (1982) Independent systems of orientation columns in upper and lower layers of monkey visual cortex. Neurosci. Lett. 31 :225-230. Kuffler. S.W. (1953) Discharge patterns and functional organization of mammalian retina. J. Neurophysiol. 16:37-68. Lam .. K ... A.J. Sefton and M.R. Bennett (1902) Loss ofaxons from the optic nerve of the rat during early postnatal development. Dev. Brain Res. 3:~81-~91. Lee, B.B., B.G. Cleland and 0.0. Creutzfeldt (1977) The retinal input to cells in area 17 of the eat's cortex. Exp. Brain Res. 30:527-536. 69 Lee, B.B., 0.0" Creutzfeldt and A" Elepfandt (1979) The responses of magno- and parvocellular cells of the monkey's lateral geniculate body to moving stimul i. Exp. Brain Res. 35:547 - 557. Leehey, S.C., A. Moskowitz-Cook, S. Brill and A. Held (1975) Orientation anisotropy in infant vision. Science 190:900-903" Leicester, J. and J. Stone (1967) Ganglion, amacrine and horizontal cells of the cat's retina. Vision Res. 7:69S-7lJ5. Lennie, P. (1980) Parallel visual pathways: A review. Vision Res. 20:561-594. Leventhal., A.G. (1919) Evidence that the different classes of relay cells of the cat's lateral geniculate nucleus terminate in different layers of the striate cortex. Exp. Brain Res. 37:349-372. Leventhal" A.G. (1903) Rel'ationship between preferred orientation and receptive field position of neurons in cat striate cortex. J. Compo Neurol. 220:476'--483. Leventhal" A.G. and H.V.B. t-tirsch (1970) Receptive-field properties of neurons in different laminiB of visual cortex of the cat. J. Neurophysiol. 41 :948-962. Leventhal" A.G. and H.V.B. Hirsch (19811) Receptive-field properties of different classes of neurons in visual cortex of normal and dark-reared oats. J. Neurophysiol. 43: 1111-1132. Leventhal" A.G. and J.D. Schall (1903) Struotural basis of orientation sensitivity of cat retinal gangl ion cells. J. Compo Neurol. 220:465-415. Leventhal" A.G., J. Keens and I Tark (1 gao) The afferent ganglion cells and cortical projections of the retinal recipient zone of the cat's pulvinar complex. J. Compo NeuroL 1 g4:535-554. LeventhaL. A.G .... J.D. Schall and W. Wallace (1984) Relationship between preferred orientation and receptive field position of neurons in extrastriate cortex (area 19) in the cat. J. Compo NeuroL 222:445-451. Levick ... W.R. (1967) Receptive fields and trigger features of gangl ion cells in the visual streak of the rabbit's retina. J. Physiol. (Lond.) 188:285-307. Levick ... W.R. and L.N. Thibos (1902) Analysis of orientation bias in cat retina. J. Physiol. (Lond.) 329:243-261. 70 linden, R, and V, H, Perry (1902) Ganglion oell death within the developing retina: A regulatory role for retinal dendrites? Neurosci. 7:2813-2827 I lindsay, R.D. (ed) (1911) Computer Analysis of Neuronal Structures. New York. Plenum. livingstone, M.S. and D.H. Hubel (1904) Specificity of intrinsic connections in primate primary visual cortex. J. Neurosci. 4:2830-2835. Lorente de No, R. (1943) Cerebral cortex: Architecture, intracortical connections, motor projections. in Physiology of the Nervous System, J.F. Fulton (ed). New York. llxford University Press. lund, J.S., G.H. Henry, C.l. Macqueen and A.A. Harvey (1979) Anatomical organization of the primary visual cortex (area 11) of the cat. A oomparison with area 11 of the maoaque monkey. J. Compo Neurol. 184:599-618. Malpeli, J.G. (1963) Activity of cells in area 11 of the cat in absence of input from layer A of lateral geniculate nucleus. J,. Naurophysiol. 49:595-610. Mann, I.e. (1926) The regional differentiation of the vertebrate retina. Am. J. OphthalmoL 11 :515-526. Mansfield~ A.J.W. (1974) Neuronal basis of orientation perception in primate vision. Science '186: 1133-1135. Mansfield, A.J.W. and S.F. Ronner (1978) Orientation anisotropy in monkey visual cortex. Brain Res. 149:229-234. Mariani... J., F. Crepel, K. Mikoshiba, J. -P. Changeux and C. Sotelo (1977) Anatomical, physiological and biochemical studies of the cerebellum from reeler mutant mouse. Phil. Trans. R. Soc. Londo 281:1-28. Marr, D. (1982) Vision: A Computational Investigation into the Human Representation and Processing of Visual Information. W.H. Freeman. San Francisco. Martin ... K.A.C. (1984) Neuronal circui ts in cat striate cortex. in Cerebral Cortex Volume 2: Functional Properties of Cortical Cells. E.G. Jones and A. Peters (eds.) New York. Plenum. Martin, K.A.C. and O. Whitteridge (1984) The relationship of receptive field properties to the dendritic shape of neurones in the cat striate cortex. J. Physiol. 356: 291-302. 71 Mastronarde, D.M., M.A. Thibeault and M.W. Dubin (1984) Non-uniform postnatal growth of the cat retina. J. Compo Neurol. 228:598-608. Matsubara, J., M. Cynader, N.V. Swindale and M.P. Stryker (1985) Intrinsic projections with visual cortex: Evidence for orientation specific local oonneotions. Proc. Natl. Aoad. Sci. 82:935-939. McIlwain, J.T. and P. Buser (1968) Receptive fields of single cells in the cat·s superior colliculus. Exp. Brain Res. 5:314-325. Morest, O.K. (1 969) The growth of dendrites in the mammal ian brain. Z. Anat. Entwickl. -Gesch. 128:290 -317. Morest, O.K. (1 g711) The pattern of neurogenesis in the retina of the rat. Z. Anat. Entwickl. -Gesch. 131 :45-67. Morrison, J.D. (1982) Postnatal development of the area centralis of the kitten retina: An electron microscopic study. J. Anat. 135:255-271. Morrone, M.C .• D.C. (lurr and L. Maffei (1902) Functional implications of cross-orientation inhibition of cortical visual cells. 1. Neurophysiological evidence. Proc. R. Soc. Lond. B. 216:335-354. Motter, B.C. and V.B. Mountcastle (1981) The functional properties of light sensitive neurons in the posterior parietal cortex studied in waking monkeys: Foveal sparing and opponent vector orientation. J. Neurosci. 1 :3-26. Mountcastle, V.B. (1957) Modality and topographic properties of single neurons of cat's somatic sensory cortex. J. Neurophysiol. 20:408-434. Movshon, J.A. and A.C. Van Sluyters (1981) Visual neural development. Ann. Rev. Psychol. 32:477-522. Murphy, E.H. and N. Berman (1979) The rabbit and the cat: A comparison of some features of single cells in the prirnary visual cortex. J. Compo Neurol. 188:401-238. Nelson, J.I. and B.J. Frost (1978) Orientation-selective inhibition from beyond the classic visual receptive field. Brain Res. 139:359-365. Ng, A.Y.K. and J. Stone (1982) The optic nerve of the cat: Appearance and loss ofaxons during normal development. Dev. Brain Res. 5:263 --271. Orban, G.A. (1984) Neuronal Operations in the Visual Cortex. Springer-Verlag. New York. 72 Orban, G.A. and H. Kennedy (1981) The influence of eccentricity on receptive field types and orientation selectivity in areas 17 and 16 of the cat. Brain Res. 208:203-208. Orban, B.A., E. Vandenbussche and R. Vogels (1984) Human orientation discrimination tested with long stimuli. Vision Res. 24:121-128. P