Development of the Dorsal Lateral Geniculate Nucleus in the Cat RONALD KALIL Department of Anatomy, Uniuersity of Wisconsin, Madison, Wisconsin 53706

ABSTRACT The development of the lateral geniculate nucleus has been studied systematically in Nissl preparations from a series of cats t h a t ranged in age from newborn to adult. In addition, preliminary observations are reported a t two stages of fetal development. It was found t h a t laminae develop in the lateral geniculate nucleus near the time of birth and continue to differentiate during the first postnatal week. During development the major axis of the lateral geniculate rotates approximately 180" in the sagittal plane. The rotation begins prenatally and is not completed until after the twentieth postnatal week. The volume of the lateral geniculate was computed a t different ages and it was determined that during the first postnatal month the nucleus attains two-thirds of its adult size. However, the rate of growth declines markedly thereafter, and final volume, like final position, is not achieved until late in development. The cross-sectional areas of lateral geniculate neurons were measured a t four locations in the nucleus in each animal. The locations represented the following parts of the visual field: the paracentral and inferior peripheral fields in the binocular segment of lamina A; the monocular segment of lamina A; and the paracentral field in lamina A l . Neurons in each of these locations grow a t approximately the same rate and are essentially fully grown by 56 days. Cell size histograms show that more large cells are found in lamina A1 and more small cells in the monocular segment than elsewhere in the dorsal laminae. Unlike the retina, there appears not to be a gradient of development in the lateral geniculate nucleus from center to periphery, a t least in terms of cell body size a t the ages studied. On the contrary, that part of the lateral geniculate nucleus which represents the paracentral visual field is the last segment in the dorsal laminae to achieve a mature cell size distribution. Finally, a discrete class of small spindle-shaped neurons was observed in the lateral geniculate nucleus ventral and caudal to the C laminae during the first two postnatal weeks. These cells possess a leading and trailing cytoplasmic process and are distinctly different from cells in the main laminae. It is suggested that these spindle-shaped cells may be neurons that are still in the process of migration or differentiation in the postnatal animal. The anatomy of t h e dorsal lateral geniculate nucleus (LGN) of the adult cat has been studied intensively. The major afferent (Laties and Sprague, '66; Stone and Hansen, '66; Guillery, '67, '70; Garey, '68; Niimi et al., '71; Hickey and Guillery, '74; Updyke, '75) and efferent (Garey and Powell, '67; Niimi and Sprague, '70; Burrows and Hayhow, '71; Rosenquist et al., '74; Gilbert and Kelly, '75; J. COMP. NEUR. (1978) 182: 265-292.

Le Vay and Gilbert, '76; Hollander and Vanegas, '77; Shatz et al., '77) connections of the nucleus have been established. The cytoarchitecture of the LGN has been described in detail (for a review see Guillery, '70; also Hickey and Guillery, '741, and Golgi (O'Leary, '40; Guillery, '66; Famiglietti and Peters, '72) and horseradish-peroxidase labelling methods (Le Vay and Ferster, '77; Lin et al., '77) have

265

266

RONALD KALIL

been used to clarify neuronal morphology and provide a basis for distinguishing between relay cells and interneurons. Moreover, a considerable body of evidence is available regarding ultrastructural features and synaptic patterns in t h e LGN as seen with the electron microscope (Peters and Palay, '66; Szentagothai e t al., '66; Guillery, '69a,b; Guillery and Scott, '71; Famiglietti and Peters, '72). In recent years, anatomists and physiologists alike have demonstrated a vigorous interest in t h e development of t h e feline visual system. Curiously, however, despite t h e wealth of information about t h e mature LGN, relatively little attention has been paid to t h e development of t h e nucleus. Few normative anatomical (Garey et al., '73a; Cragg, '75; Elgeti et al., '76) or physiological (Adrien and Roffwarg, '74; Norman et al., '77) studies have been published, although t h e LGN has served as a focal point in numerous experiments in which visual deprivation has been employed as a tool to study development (Wiesel and Hubel, '63; Kupfer and Palmer, '64; Guillery and Stelzner, '70; Chow and Stewart, '72; Guillery, '72, '73; Garey et al., '73b, '76; Sherm a n e t al., '75; Cragg e t al., '76; Dursteler et al., '76; Wan and Cragg, '76; Garey and Blakemore, '77; Hickey et al., '77; Hoffman and Sireteanu, '77; Movshon and Dursteler, '77; Hoffman and Hollander, '78). I have also been interested in t h e effects of deprivation on t h e development of t h e visual system in t h e cat and have used t h e LGN to study t h e consequences of early enucleation, eyelid suture and dark rearing (Kalil, '75, '76, '77, '78). As these experiments progressed, i t became clear t h a t to interpret t h e effects of deprivation on the LGN, i t would first be necessary to understand t h e normal development of t h e nucleus. Nissl-stained material from a n age-graded series of normally reared cats was therefore collected, initially for t h e limited purpose of studying cell growth in t h e LGN during development. In t h e course of making cell area measurements to quantify neuronal growth, several additional observations and measurements were made concerning t h e general development of t h e LGN. This paper reports quantitative analyses of: (a) cell growth and cell size distribution in different segments of t h e LGN; and (b) changes in t h e volume and position of t h e nucleus during development. Qualitative descriptions of laminar organization and histology at different ages, including

preliminary observations on fetal kittens, a r e also presented. Details concerning t h e effects of deprivation on LGN development will be described in a subsequent communication. MATERIALS AND METHODS

Twenty-eight cats, ranging in age from newborn to adult, were used in this study. Most of t h e cats were born and raised in t h e laboratory colony, but a few animals were obtained from a commercial vendor who specialized in providing cats of known age. In all cases, t h e day of birth is considered day 1. However, since gestation in t h e cat may vary from 60 to 66 days, it is possible t h a t some of t h e youngest kittens may overlap in terms of their t r u e age postconception. A listing of all animals is provided in table 1. The cats were anesthetized with sodium pentobarbital and perfused through the left ventricle with 10% formol-saline. The brains were exposed, but not removed from the skull, and allowed to harden in t h e fixative for one month. Since t h e LGN undergoes a change in orientation during development (RESULTS), most of t h e brains were blocked sagittally to maintain a uniform plane of section in brains of different ages. Following dehydration in a graded series of ethanols, the brains were embedded at room temperature in low viscosity nitrocellulose (RS 0.5 second) and sectioned at 20 p. A cresyl violet series through t h e LGN of each animal was prepared by staining every third section in kittens 56 days and younger, and every fifth section in older cats. The volume of t h e LGN a t each age was estimated using t h e following procedure. First t h e sections in each series were arranged in order from t h e medial border of t h e nucleus to t h e lateral edge of t h e monocular segment of lamin a A. Then every other section in the series was projected at a linear magnification of 23 x and t h e outline of t h e LGN was drawn. The area of each tracing was measured with a n electronic planimeter, corrected for t h e magnification, and then t h e traced areas were summed. Multiplying this sum by t h e section thickness, in this case 20 F , yields a sample volume which can be converted to the t r u e volume of t h e LGN by determining t h e percent size of t h e sample. This value can be calculated by dividing t h e number of sections in t h e sample by t h e total number of sections through the LGN. Typically, t h e samples used in t h e volume computations represented 16% of the total for cats 56 days and younger, and

267

DEVELOPMENT O F THE LGN IN THE CAT TABLE 1

Postnatal growth of the lateral geniculate nucleus Mean cell size ( p 2 Cat

Age (days)

KIA K1B K1D

1 1 1

Binocular A

Monocular A

* S.E., n = 100) Peripheral A

Lamina A1

Volume mm'

74.32 2.6 91.92 3.0 61.92 2.6 Ave. 76.02 8.7 (75.5) 110.72 3.5 103.42 3.6 Ave. 107.12 3.6 (96.0)

66.5' 1.5 99.12 2.5 75.62 1.9 8 0 . 4 1 9.7 (79.5) 107.02 2.8 106.82 2.9 106.920.1 (102.7)

71.72 2.0 77.92 1.9 94.82 2.6 81.526.9 (79.5) 97.32 2.2 103.92 2.8 100.62 3.3 (93.5)

85.62 3.0 107.42 4.0 76.62 2.4 89.929.1 (82.8) 125.45 3.9 117.22 4.7 121.32 4.1 (117.5)

115.12 4.7 114.92 3.8 Ave. 115.02 0.1 (109.8)

103.02 3.5 101.02 3.3 102.02 1.0 (89.8)

99.52 3.1 104.42 3.5 101.92 2.5 (92.2)

136.42 6.0 125.42 5.0 130.92 5.5 (119.1)

9.9 10.3 10.12 0.2

122.22 3.5 137.72 3.9 129.92 7.8 (125.1)

136.12 6.9 165.9% 6.0 151.02 14.9 (134.7)

14.7 11.2 12.92 1.8 11.7 11.7 11.710.0

4.7 5.3 -

5.02 0.3 6.5 6.8 6 . 6 1 0.1

K3 K3A

3 3

K7 K7A

7 7

K10 KlOA

10 10

118.82 3.8 131.22 6.7 127.114.9 123.72 3.6 Ave. 127.42 3.8 (117.9) 122.954.1 (118.9)

K14 K14A

14 14

136.82 5.4 140.9% 5.0 Ave. 138.82 2.1 (129.4)

108.22 3.2 120.72 3.9 114.426.3 (110.1)

140.02 4.2 138.12 3.6 139.12 0.95 (133.5)

163.22 5.8 169.02 5.8 166.12 2.9 (157.8)

K21

21

147.82 5.4 (137.3)

150.32 4.9 (137.8)

169.52 4.1 (170.3)

176.72 7.5 (163.8)

-

K28 K28A

28 28

161.62 6.6 173.22 6.7 186.82 6.2 165.02 4.2 Ave. 174.22 12.6 (165.6) 169.1* 4.1 (159.7)

186.22 4.9 207.6% 5.6 196.92 10.7 (186.1)

187.32 7.5 208.95 7.1 198.11 10.8 (184.3)

27.0 29.3 28.121.2

K35

35

182.72 6.8 (172.1)

189.52 5.4 (186.2)

188.32 5.1 (185.1)

200.52 7.1 (191.7)

-

K42

42

189.6% 7.9 (180.9)

179.62 7.0 (167.8)

205.42 6.7 (192.4)

206.52 9.1 (190.5)

-

K49

49

196.92 8.6 (177.4)

203.72 5.6 (195.4)

195.6% 6.4 (180.0)

233.12 12.0 (206.1)

K56 K56A

56 56

204.82 6.8 215.62 6.2 Ave. 210.22 5.4 (215.9)

203.3k 7.0 220.0* 7.8 211.628.4 (213.8)

232.42 7.6 209.82 4.8 221.1211.3 (217.6)

224.62 8.8 245.42 9.1 235.02 10.4 (230.0)

31.7 37.3 34.52 2.8 -

K70

70

199.12 7.7 (179.5) 209.52 8.2 (205.5)

213.02 6.7 (211.8)

239.92 9.7 (220.1)

K84

84

218.92 8.0 (203.7)

209.026.8 (203.7)

226.12 6.2 (221.0)

2 5 2 2 2 11.0 (228.8)

-

140 140 140

269.42 10.5 233.52 8.6 216.15 7.9 Ave. 239.72 15.7 (215.4) 201.02 7.4 229.92 8.4 201.32 7.3 248.42 9.8 Ave. 220.12 11.6 (194.9)

230.52 7.0 225.2* 7.0 202.4f 7.3 219.4'8.6 (207.8) 226.926.7 190.62 5.3 186.03 6.4 204.1% 6.9 201.9'9.2 (191.1)

276.62 8.8 252.11 8.0 224.82 7.0 251.22 15.0 (238.4) 232.52 7.3 223.02 6.8 221.32 7.4 247.45 7.2 231.02 6.0 (222.9)

271.22 10.0 250.52 8.2 240.12 9.0 254.02 9.1 (233.5) 242.12 10.7 220.02 8.2 224.12 8.1 259.02 9.7 236.32 9.0 (215.7)

43.0 44.0 30.2 39.1 2 4.4 41.2 43.5 37.1 43.3 41.32 1.5

K140 K140A K140B CA2 CA3 CA4 CA5

Adult Adult Adult AduIt

Numbers in parentheses indicate medians.

8% in older animals. As a last step, the

volumes obtained were corrected for shrinkage. This was done a t each age by first dividing the area of the LGN in nitrocellulose sections by the area measured from matched frozen sectioned material. The square root of this value was then cubed to yield an estimate of volumetric shrinkage. Since there was relatively little variation in shrinkage with age, the average across ages, approximately 56%, was used to correct all volume measurements. The cross-sectional areas of LGN cells in lamina A and A1 were measured from a single

section a t each age near the medio-lateral midpoint of the nucleus. In the adult, the plane of the section used for these cell area measurements lies close to the middle of the binocular portion of the LGN, between the representations of the 5" and 10" isoazimuths (Sanderson, '711.' Cell areas were measured from two zones in lamina A. One zone was ' Although it is assumed throughout this paper that corresponding locations in the kitten and adult LGN represent similar parts of the visual field, the representation of the visual field has not been determined directly in the kitten LGN, and thus It 1s possible that the map which has been established for the adult is not valid at all ages.

268

RONALD KALIL

located at t h e rostro-caudal midpoint of t h e lamina near t h e projection of t h e 0” horizontal meridian. In referring to cells from this zone, the designation “paracentral A cells” will be used. The second zone lay at the anteroventral tip of lamina A between t h e - 40” and - 50” isoelevation projection lines. Cells from this zone will be called “peripheral A cells.” Cells measured in lamina A1 lay immediately subjacent to t h e paracentral A cell zone, and therefore represented essentially t h e same visual field projection line. The three zones a r e illustrated schematically in figure 1. In t h e sagittal plane, t h e monocular segment of lamina A is not as easy to locate as it is in frontal sections. In a closely spaced series, however, i t is possible to identify precisely t h e lateral margin of lamina A1 by t h e cell free interlaminar zone t h a t forms its border. Sections which lie lateral to t h e interlaminar zone pass through t h e monocular segment. Therefore, to measure t h e cross-sectional areas of monocular segment cells, hereafter referred to as “monocular A cells,” t h e lateral border of lamina A1 was determined and then a section midway between this point and t h e end of t h e series was selected. To

,/

B

A Fig. 1 Schematic drawing of the LGN in parasagittal section showing the zones in lamina A, paracentral (diagonal lines) and peripheral (shaded), and lamina A1 which were used for measuring the cross-sectional areas of geniculate cells. The dashed line AB, which indicates the major axis of the nucleus as described in text, was constructed by first extending the dorsal and ventral borders of the LGN with a straightedge to the point of their intersection. The included angle was then bisected with a protractor t o give AB.

check t h e validity of this procedure, t h e four adult brains were bisected sagittally and one hemisphere was sectioned in this plane, t h e o t h e r frontally. The mean cross-sectional areas of monocular A cells identified in t h e parasagittal and frontal sections were 202 p 2 and 208 p2 respectively. These values are in good agreement, indicating t h a t t h e method used to locate the monocular segment in parasagittal sections is reliable. In each animal t h e cross-sectional areas of 400 geniculate cells were measured; 100 from each of t h e four zones just described. All measurements were made at 1,000 x with a planapochromatic oil immersion objective. Perikaryal outlines were traced in t h e plane of t h e nucleolus with a Zeiss drawing tube and t h e area enclosed by each tracing was measured with a n electronic planimeter. These measurements were not corrected for shrinkage. A sample of 100 cells from lamina A1 and from t h e paracentral and peripheral parts of lamina A was collected by first positioning t h e microscope field (approximate area, 0.025 mm? at the dorsal border of each zone. All of t h e cells with visible nucleoli were drawn and then t h e field was advanced in steps directly across t h e lamina until t h e sample was complete or t h e inner border of t h e lamina was reached. If one sweep across t h e lamina did not yield 100 measurable cells, t h e microscope field was moved to a n adjacent area and t h e direction of t h e sweep was reversed. Generally, 1.5 sweeps were necessary t o collect 100 cells. While this procedure may have resulted in a slight bias in favor of cells which lie in t h e inner half of each lamina, i t offered t h e advantage of avoiding t h e systematic exclusion of cells located near laminar borders. It should be pointed out, however, t h a t care was taken to exclude t h e large cells which lie in t h e interlaminar zone between A and A l , and between A1 and C. Monocular A cell samples were gathered in a slightly different manner. When sectioned in t h e parasagittal plane, t h e monocular segment, near its medio-lateral midpoint, appears as a n oval-shaped mass of cells t h a t is flattened dorso-ventrally. To measure these cells t h e microscope field was first aligned with t h e center of t h e cell mass and then moved to its dorsal border. After tracing all cells with visible nucleoli, the field was swept ventrally until 100 cell outlines were drawn. Typically t h e sample was completed before t h e ventral border was reached.

DEVELOPMENT OF THE LGN IN THE CAT RESULTS

Laminar organization The laminar organization of the LGN in the three newborn kittens that were examined differed only slightly from animal to animal. In sections through the middle third of the nucleus, the dorsal laminae, A and Al, are prominent because of their high cell density, and lamina A shows a clear interruption which corresponds to the representation of the optic disc (Kaas et al., '73). In each brain A and A1 are separated by a relatively cell free interlaminar zone, but in only one of the kittens was there a similar cell sparse zone interposed between A1 and the C layers. Nevertheless, the presumptive C laminae can still be recognized in the newborn because the markedly reduced cell density in the posterior layers creates a sharp border contrast with lamina A1 (fig. 10A). Since the basic laminar organization of the LGN is present in the newborn kitten, lamination must develop prenatally. A systematic investigation of this problem was not undertaken, but the brains of two fetal kittens were available for study. Unfortunately, the exact ages of these fetuses were unknown, but reasonable estimates can be made based upon body measurements of crown-rump length, as well as previous work on the development of the brain in the prenatal cat (Marin-Padilla, '71). In his study of the prenatal development of the neocortex, Marin-Padilla ('71) examined a series of fetuses ranging in age from E l 7 (embryonic day 17) to E45. Figure 1of his paper illustrates the gross development of the brain a t different fetal ages, and measurements show that at E40 the brain is approximately 14.0 mm from the tip of the frontal pole to the rudiment of the cerebellum. By E45, the length has increased to 18.0 mm. Furthermore, the 45-day fetus can be distinguished from earlier stages by the presence of the corpus callosum which develops around E43. The brain of the younger fetus in the present study measured 15.0 mm in length and microscopic examination indicated that the corpus callosum had not yet formed. Crownrump length of this fetus was 8.0 cm which corresponds to a gestational age of 42 days (Farris, '50). Allowing for individual differences in development, t h e brain and body measurements place the age of this kitten at between E40 and E45.

269

Horizontal sections (fig. 8A) reveal that the LGN is compressed medio-laterally, and near its midpoint it resembles a flattened teardrop. In parasagittal section, the nucleus appears as a triangular mass of densely packed cells. The apex of the triangle is directed dorsally, and the base is marked by a shallow concavity (fig. 8B). In both planes it is clear that the LGN extends forward to the rostral wall of the diencephalon, occupying proportionally a large volume of the developing thalamus. At this stage of development the LGN is comprised of small neurons (about 4 p in diameter) which are uniform in size, circular in cross section, and very immature. Viewed with the light microscope, the cells lack cytoplasm, and contain two or more nucleoli surrounded by dense specks of chromatin. The LGN a t this age shows no sign of lamination. The brain of the older fetus measured 19.5 mm from the rostral tip of the neocortex to the cerebellum, and in this animal the corpus callosum was well developed. Crown-rump length was 12.0 cm (estimated gestational age 56 days, Boyd, '71). Since the brain of a newborn kitten measures approximately 23 mm, the older fetus may be assumed to be midway in age between the younger fetus and the newborn, probably near the middle of the eighth week of gestation. The LGN in this kitten is considerably more mature than that of the younger fetus. In horizontal sections (fig. 8C) one can see that the nucleus is no longer compressed along its medio-lateral axis, although i t is still contiguous with the rostral wall of the diencephalon. In the parasagittal plane the LGN is elongated dorsoventrally with a curvilinear profile characteristic of the postnatal animal (fig. 8D). This new conformation appears to result from a rotation of the LGN in the sagittal plane of approximately 45". Thus the hilum of t h e nucleus, from which fibers of the optic radiation emanate (Thuma, '28; Hayhow, '581, was directed ventrally in the younger fetus, but now faces rostrally. Geniculate cells in the older fetus also show little or no cytoplasm, but in contrast to the younger animal most contain only a single nucleolus, and all of them display evenly dispersed chromatin. More importantly, the cells vary in size and shape, from spherical to elliptical, suggesting an early differentiation that anticipates the diversity of cells evident in Nissl preparations of the mature LGN. Although a clear cell density gradient exists in the LGN of the older

270

RONALD KALIL

fetus and lends prominence to the presumptive dorsal layers, i t is impossible to distinguish a laminar organization. Based on these preliminary observations, one may conclude t h a t lamination of t h e LGN occurs very late in gestation, even though the rudimentary n u cleus is probably formed before E40. In comparing t h e structure of the LGN in the 3-day postnatal kitten with t h e newborn, the most obvious difference is t h e presence of a clear-cut interlaminar zone separating lamina A1 from t h e C layers (fig. 10B). During t h e next ten days, t h e interlaminar zones widen, and by two weeks postnatal they assume the relatively cell free character typical of t h e mature LGN (fig. 10D). As is well known, however, t h e interlaminar zones are not devoid of neurons, but, in fact, contain some of t h e largest cells in t h e adult geniculate (Guillery, '70). In the developing LGN, distinct interlaminar cells are not seen until seven days after birth. The abrupt appearance of these cells coincides with a general emergence of medium sized neurons (mean size 250 p 2 )scattered throughout t h e dorsal laminae (figs. 9C, 1OC). Although t h e number of these medium-sized cells is too small at this age to affect quantitative measurements dramatically (on average fewer than 6 medium-sized cells a r e seen in a single oil immersion field), their presence represents a major step in t h e postnatal development of the LGN. Thus in contrast to the newborn kitten in which geniculate cells are essentially uniform in size, t h e LGN in older cats is marked by a variety in cell size which first becomes apparent at seven days. One further qualitative observation t h a t bears noting is t h e presence of a discrete class of small geniculate neurons which is seen only during t h e first two postnatal weeks. These neurons are found ventral and caudal to the C laminae, adjacent to t h e optic tract where they appear to be aligned parallel to t h e long axis of t h e LGN (fig. 9A). The cells a r e spindle-shaped, with a n elongated nucleus, and have a prominent leading and trailing cytoplasmic process (fig. 9B). Since these neurons are very basophilic, they a r e easily distinguished from other neurons in t h e LGN, and on t h e basis of cell body size, nuclear appearance, and prominent cytoplasm they can be separated from glial cells.

Rotation of the LGN If one studies t h e LGN i n a n age-graded series of kittens, i t is immediately obvious that the nucleus undergoes a progressive rota-

TABLE 2

Sagittal rotation Cat

of

the LGN duringpostnatal deueloprnent

Age (days)

Angle (degrees)

K1A K1B K1D

1 1 1

63 74 78 Ave. 72

K3 K3A

3 3

K7 K7A

7 7

K10 KlOA

10 10

K14 K14A

14 14

K2 1 K28 K28A

21 28 28

K35 K56 K56A

35 56 56

51 37 Ave. 44 40 55 Ave. 48 36 32 Ave. 34 43 49 Ave. 46 21 30 27 Ave. 28 28 28 22 Ave. 25

K70 K84 K140 K140A

70 84 140 140

CA3 CA5

Adult Adult

I

20

15 27 7 Ave. 17 11 2 Ave. 6

' Acute angle with respect to the horizontal tional displacement during development (figs. 10A-HI. The rotation begins prenatally and occurs chiefly in t h e sagittal plane. As mentioned, t h e hilum of t h e nucleus initially faces ventrally, but rotates 90' to assume a rostrally directed position. A surprisingly long time is required to complete this process since t h e LGN probably does not achieve its final position until after t h e twentieth postnatal week. One is confronted by two problems in attempting to measure t h e angular position of t h e LGN at different ages. Firstly, in parasagittal sections, t h e major axis of t h e LGN is not linear, but instead resembles a mild S-shaped curve. Secondly, when the brain is cut in the sagittal plane, i t presents few obvious landmarks t h a t can be considered stable during development and therefore useful for constructing a horizontal or vertical coordinate system against which t h e rotation of t h e nucleus can

271

DEVELOPMENT OF THE LGN IN THE CAT

be measured. In a n effort to circumvent these problems, two assumptions were made. The first was to define t h e major axis of the LGN by drawing a line through t h e nucleus which approximately bisected its middle third (AB: fig. 1). The second was to define a frame of reference by taking account of t h e fact that in a parasagittal section through t h e middle third of t h e LGN, t h e zero horizontal plane passes approximately through the center of the medial geniculate body and t h e cerebellum (plate 55: Berman, '68). Since t h e relation of these structures to each other does not change significantly during postnatal development, a line drawn through their midpoints can be used as a horizontal referent. Therefore to determine t h e angular position of the LGN at each postnatal age, a section through t h e middle third of the nucleus was projected a t a linear magnification of 14 X , and a protractor was used to measure t h e acute angle between the major axis of t h e LGN and t h e horizontal. The measured values a r e presented in table 2, and illustrated schematically in figure 11. Although t h e measurements at some ages show considerable variation, which may be attributed, at least in part, to individual differences among animals as well as discrepancies in t h e plane of section, t h e averaged results demonstrate a clear trend. Thus in t h e newborn kitten, the major axis of t h e LGN is close to vertical, whereas in t h e adult i t is nearly horizontal. Volume of the LGN

45r

7 14

28

56 " AGE IN DAYS

140 ADULT

Fig. 2 Mean volume of the LGN in cats ranging in age from newborn to adult.

a1 area of all cells measured at that age. For clarity, standard error bars have been omitted, but numerical values are given in table 1. During t h e first postnatal month LGN neurons approximately double in cross-sectional area, and essentially achieve adult size by the end of t h e eighth week.2 Individual growth curves have not been analyzed mathematically, but simple inspection suggests t h a t their slopes are not markedly different. While the r a t e of growth is similar for cells in t h e four laminar areas t h a t were measured, cells vary somewhat in size depending upon their location. Thus anteroventral peripheral A cells tend to be larger t h a n paracentral A cells, which are in t u r n larger t h a n monocular A The volume of t h e LGN, corrected for cells, and lamina A1 cells are the largest of shrinkage, at various postnatal ages is plotted all. However, a consistent difference in size in figure 2. Each point represents a n average throughout development is seen only for lamiof at least two animals; values for individual n a A1 cells which are larger at each age than kittens are given in table 1. During t h e first cells in t h e other three groups. 28 days after birth LGN volume more t h a n In t h e adult cat, A1 cells are approximately quintuples (the average value in t h e newborn 7% larger t h a n binocular A cells, but this difis 5.0 mm3, in the four week kitten, 28.2 mm3). ference is not statistically significant (p > Such rapid growth does not extend into t h e 0.1, two-tailed t test for related samples, second postnatal month, although the LGN n = 4 ) . Similarly, paracentral A cells in t h e continues to increase in size. An additional adult a r e 9% larger than monocular A cells, 22%in volume is gained in this period, and by but again t h e difference is not statistically 56 days t h e LGN reaches approximately 85% significant (p > 0.1). Indeed, if cell size differof its adult size. Final volume appears to be ences between groups in t h e adult cat are achieved slowly, but additional measurements evaluated using conservative statistical proare needed between 56 and 140 days to specify cedures, e.g., a t test for related samples with precisely when LGN growth ceases. n = 4, none of them are significant. However,

Cell growth Figure 3 plots t h e growth of geniculate cells in each of t h e four zones that were sampled. Each point represents t h e mean cross-section-

A t 56 days the mean cross.sectiona1 area of LGN neurons IS not different statistically from adult values (two-tailedt test. 4 d.f.1. The peak seen at 140 days IS due chiefly t o one animal (K140) that had abnormally large cells. If K140 1s diacounted, sample means at 140 days are: paracentral A, 224.8 p ' , monocular A, 213.8 p d ;peripheral A, 238.4 pi.A l . 245.3

,L'.

272

RONALD KALIL

250230210N

=i 190-

a W

I1z

a

170-

_I

w

u z

a W

e--.Lamina A, Paracentral

2

*----o

Lamina A, Monocular

uLamina A, Peripheral @---a Lamina A 1

I

I

7

14

21

28

I

I

I

I

35 42 49 56 AGE I N DAYS

I

I

70

84

~ ’ ! ” , +

Fig. 3 Mean cross-sectional areas of LGN neurons in cats at various postnatal ages. Note that cells grow at essentially the same rate in each segment and are almost fully grown by 56 days.

since other tests can be applied which might lead to different statistical interpretations, complete mean cell area measurements for each of the animals in this study are presented in table 1. Histograms showing cell size distribution a t different postnatal ages are given in figures 4-7. Each histogram represents combined results from two cats with the following exceptions: three animals were used a t 1 day and 140 days, and the histograms for adults are based on four cats. Above each histogram the mean of the distribution is indicated by a filled circle on the right and the median by the circle on the left.3 It is clear from the histograms that the size distribution of geniculate cells a t any age is continuous and tends to be marked by a single modal value. Thus in terms of perikaryal size alone, Nissl preparations provide little correlative evidence to support LGN cell classifications that can be made in Golgi impregnated material (Guillery, ’66; Famiglietti and Peters, ’72) or by physiological recording (Hoffman et al., ’72; Cleland et al., ’76; Wilson

et al., ’76; Dubin and Cleland, ’77). Nevertheless, in discussing age related changes in the distribution of cell sizes it is convenient, if arbitrary, to group cells into classes on the basis of size. For this purpose, “small” cells are considered to be those with cross-sectional areas of 150p or less, “medium” cells those with areas between 150 p2 and 300 g2,and “large” cells those which are greater than 300 p z .The relative proportions of small, medium and large geniculate cells a t different postnatal ages are shown in table 3. In general, cell size distributions a t any one age are similar regardless of laminar origin or location. The changes which occur during development may be summarized as follows, although table 3 should be consulted for specific In certain cases. combining histograms from two or more animals may lead to a composite cell size distribution which is an artifact. This will surely result if individual distributions differ significantly from each other. In the present instance, however, it is clear from table 1 that mean cell sizes at any one age rarely vary by more than 25 p 2 (one-half of the bin width in the histograms), and the standard errors of the means are in close agreement. Median values are presented because this measure of central tendency is less affected by extreme values than is the mean, and thus in older animals it yields a more accurate estimate of the modal value of each histogram.

273

DEVELOPMENT OF THE LGN IN THE CAT

60

..

5.1

i DAY

28 DAYS

K

w

rn

fz

6ol

6ol 50

50

40t

60-

50

-..

..

140 DAYS

60r

ADULT

14 DAYS

100 200 300 400 500 600

CELL AREA

(pm2)

Fig. 4 Cell size distribution in the paracentral part of lamina A at different postnatal ages. Note that the adult complement of large cells does not emerge until after 56 days. The filled circle at the right indicates the mean of the distribution. the circle at the left the median.

274

RONALD KALlL TABLE 3

Percentage

of

small, medium and large cells in the LGNat different postnatal ages Monocular A.

Paracentral A Age (days)

1 3 7 14 28 56 140 Adult

s.

M.

97

3 10 14 25 57 74 67 66

90

85 74 37 19 10

18

L

0 0 1 1

6 7 23 16

S

M

99 93 92 88 43 26 15 23

1 7 8 12 53 62 73 67

Penpheral A.

L.

s.

M

0 0

99 95 91 69 18 15 7 12

5 9 31 76 73 72 73

0 0 4 12 12 10

1

A1

L.

0 0 0 0

6 12 21 15

S

M

92 84 77 45 25 22 6 19

8 15 21 50 65 56 67 60

L.

0

1 2 5 10 22 27 21

S small cells 150 +'and smaller. M, medium cells 151 c2 300 p' L large cells, greater than 300 p 2

details. In t h e newborn kitten, geniculate neurons a r e clustered tightly about t h e population mean, but a broadening of t h e distribution is evident as early as three days postnatal. Relatively little change occurs between three and seven days, however by 14 days there is a noticeable increase in the number of medium-sized neurons. During t h e next two weeks t h e percentage of small cells declines rapidly. The corresponding gain in mediumsized cells is dramatic, and is accompanied by t h e emergence o f a modest number of large cells. Although t h e small cell population continues to decrease between 28 and 56 days, i t is the large cell group t h a t undergoes the most striking change during this period. With the exception of cells in the paracentral A zone, t h e number of large cells approximately doubles during t h e second postnatal month, and t h e resulting proportions of small, medium and large cells in the LGN of the 8-week-old kitten are nearly identical to those in t h e adult. In comparing cells sampled at different locations in t h e LGN i t is clear that small but interesting variations exist in t h e distribution of cell sizes. Thus in lamina A1 (fig. 7) large cells develop earlier t h a n in lamina A, and also account for a greater percentage of t h e total population. In contrast, the fewest large cells are found in t h e monocular segment of lamina A (fig. 5 ) , but small cells are well represented. This bias in favor of small cells is seen throughout development since t h e upward shift in cell size takes place slowly in t h e monocular segment. A comparison o f paracentral (fig. 4) with peripheral A cells (fig. 6) reveals similar cell size proportionalities during the first two postnatal weeks, but between 14 and 28 days medium sized cells develop about 50% faster in the periphery. The tendency for

development in t h e paracentral region to be delayed can also be seen if one compares t h e relative proportions of large cells. In A1 and in t h e peripheral and monocular parts of lamina A t h e adult complement of large cells is acquired by the end of t h e second postnatal month, but not in t h e paracentral zone until sometime between 56 and 140 days. As noted earlier in discussing mean cell area measurements, t h e values for the 140day animals appear discrepant since they a r e larger than t h e adult. This divergence is also reflected in t h e cell size histograms (see also table 3) which show t h a t a t 140 days there are fewer small cells t h a n in t h e adult, or even at 56 days, and correspondingly more medium and large cells. Again, however, if t h e results from K140 are excluded, t h e percentages compiled from t h e two remaining 140-day cats show proportionally fewer large cells and more medium-sized cells. Although the relative number of small cells remains low, the percentages of medium and large cells a r e in line with adult values. DISCUSSION

At birth, t h e LGN in t h e cat is structurally immature. Lamination is not fully developed, t h e cross-sectional areas of nerve cells average only one-third of adult size, and t h e major axis of t h e geniculate is displaced nearly 70" in t h e sagittal plane with respect to its mat u r e orientation. This study shows t h a t during t h e first two postnatal months rapid changes occur in t h e morphological organization of t h e nucleus. Thus the basic trilaminar arrangement of t h e LGN is established during t h e first week, and is clearly delineated at 14 days. Between birth and 28 days t h e volume of t h e LGN increases almost sixfold. By t h e end of t h e eighth week cell growth is substantially

275

DEVELOPMENT OF THE LGN IN THE CAT

701 60

50t

70

50 140 OAYS

7 DAYS 401

6 70[

40

CELL AREA

ADULT

(p2)

Fig, 5 Cell size distribution in the monocular part of lamina A at different postnatal ages. More small cells are found in the monocular segment than elsewhere in the dorsal laminae. Mean and median as in figure 4.

276

RONALD KALIL

-

?Or

60 -

60-

50 -

50

40 -

40-

30 -

30

70

-

-

28 DAYS

LI

20 -

-

10-

7060-

..

P

50 40 -

3 DAYS

3020 10-

- , ,

J

'Or

50

40

t

140 DAYS

70r

501

40

ADULT

100 200 300 400 500 600

CELL AREA (rmz) Fig. 6 Cell size distribution in the peripheral part of lamina A at different postnatal ages. In comparison with paracentral A, cells in peripheral A achieve a mature size distribution earlier, but the final distributions in the two zones do not differ markedly in the adult animal. Mean and median as in figure 4.

277

DEVELOPMENT OF THE LGN IN THE CAT

60-

50

-

40 -

i DAY

20-

30

10-

-

56 DAYS

60 -

50 0.

1

7 DAYS

40

-

140 DAYS

30 -

0 .

50 40

-

14 DAYS

ADUL?

10

100 200 300 400 500 600

CELL AREA

100 200 300 400 500 600

(p2)

Fig. 7 Cell size distribution in lamina A1 at different postnatal ages. Excluding the newborn, lamina A1 contains the greatest proportion of large cells at all ages. Mean and median as in figure 4.

278

RONALD KALIL

completed, and the LGN has acquired an orientation that approximates the adult configuration. However, the time required for full development of the LGN extends beyond the first two months since the final position of the nucleus, adult volume, and mature cell size distribution in the paracentral part of lamina A are not achieved until later. Previous work on the morphological development of the feline visual system has dealt chiefly with the retina (Donovan, '66) and visual cortex (Cragg, '751, and relatively little attention has been paid to the LGN. Nevertheless, several important observations have been made concerning the development of the LGN which are relevant to the present report. Anker ('77) has studied the prenatal development of the retino-geniculo-striate system in the cat using light microscopic degeneration techniques. Afferents from the retina arrive a t the LGN a t approximately E37 and the presumptive geniculostriate pathway can be detected a t E48. Neither of these events, however, appear to play a critical role in t h e laminar development of the LGN since the present results indicate that lamination occurs very late in gestation, most likely within the week t h a t precedes birth. A similar situation obtains in the rhesus monkey in that axons from the retina are distributed throughout the LGN up to 30 days before the onset of lamination (Rakic, '77a). A satisfactory explanation of the delay in laminar formation is not available a t present, but preliminary evidence in the monkey suggests that retinogeniculate synaptogenesis may be important. Thus Hendrickson and Rakic ('77) have described a caudorostral gradient in the formation and differentiation of retinogeniculate synapses. This gradient can be correlated with the establishment of laminae in the LGN during normal development since lamination begins a t the caudal pole and then sweeps anteriorly. Furthermore, Rakic ('77b) has shown that monocular enucleation in the fetal monkey a t E63, before the onset of lamination and the major period of synaptogenesis, prevents the formation of laminae. However, if eye removal is performed a t E91, when synaptogenesis is increasing and laminae are beginning to develop, a permanent laminar organization results, but only a t the caudal pole of the nucleus. Again, a correlation can be drawn between synaptogenesis and lamination which is especially clear in the E91 experiment since laminae are formed only in a specific part of

the LGN. Presumably, layers do not develop anterior to the caudal pole of the nucleus because synapse formation by one set of retinogeniculate afferents is aborted by the enucleation. In view of the rapid development of laminae in the cat LGN, late in gestation, it should be relatively easy to determine if synaptogenesis is correlated with the initiation of lamination. In common with primates the LGN of the cat undergoes a displacement and rotation during development. As the dorsal thalamus expands during development, the primate LGN is first displaced caudally and then laterally and ventrally (Clark, '32; Chacko, '48; Rakic, ' 7 7 ~ ) In . association with the ventrolateral displacement, the LGN undergoes a rotation in both the transverse and sagittal planes. The details of this complex displacement-rotation are described clearly by Rakic ( ' 7 7 ~ and ) will not be repeated here. Suffice i t to say t h a t neurons which lie initially lateral in the LGN are destined to occupy a ventral position, while those which first migrate to dorsal locations eventually form the caudal pole of the nucleus. In the fetal cat the LGN occupies a major part of the anterior-posterior extent of the lateral margin of the diencephalon. Interestingly, this embryonic position in the cat is similar to that of the mature LGN in mammals such as the tree shrew (Clark, '32; Chacko, '48). Horizontal sections of the fetal brain show that as the LGN develops the anterior pole of the nucleus is displaced caudally. In parasagittal view it is clear that the displacement of the embryonic anterior pole, which becomes the caudal pole in the mature nucleus, results from a rotation of the entire LGN in the sagittal plane. Thus the anterior pole in the younger fetus (E40-E45) is directed somewhat ventrally, but it rotates almost 45" during the next 7 to 12 days to assume a position in the older fetus which is similar to t h a t seen in the newborn kitten (compare figs. 8B and 8D with 10A). Postnatally, the orientation of the LGN continues to change (see also Elgeti et al., '761, but i t appears that the global rotation of the nucleus seen in the fetal animal ceases soon after birth. Thus the angular position of the caudal pole remains essentially fixed after the first postnatal week, although the major axis of the LGN shifts progressively from vertical to horizontal. This axial shift can best be explained by assuming a differential pattern of growth that leads to an anterodorsal translo-

DEVELOPMENT OF THE LGN IN THE CAT

cation of the rostra1 two-thirds of the nucleus. It is clear from the measurements made in the present study that this process requires several months to complete, but the exact time remains to be determined. A second feature of LGN development which is protracted is the acquisition of final volume. As mentioned, LGN volume increases very rapidly during the first postnatal month, reaching approximately two-thirds of adult size. Thereafter, LGN volume continues to increase but the rate of growth is slower. Cell density measurements in the binocular part of lamina A show that during the first postnatal month there is a marked decrease in the total number of nerve cells per unit volume, but throughout development the absolute density of glial cells remains essentially constant (Elgeti et al., '76). These measurements suggest that no postnatal neurogenesis occurs in the LGN, however gliogenesis is significant. Elgeti et al. ('76) estimate that the major wave of glial cell proliferation takes place between the second and sixth postnatal weeks. The production of neuroglia undoubtedly plays an important role in the rapid increase in LGN volume that is seen during this period. Probable factors contributing to the subsequent enlargement of the LGN during the second postnatal month are the continued proliferation of glial cells, the completion of growth and differentiation (Morest, '69) of LGN neurons, and perhaps the acquisition of myelin by intrageniculate optic tract fibers and geniculocortical axons (Moore e t al., '76). In contrast to the relatively long time required for LGN to attain its final position and volume, geniculate neurons themselves reach mature size quickly. The rate of cell growth is very rapid in the first 28 days after birth, but declines during the second month. By the end of the eighth week, however, the mean crosssectional area of LGN neurons in laminae A and A1 is a t minimum 95% of adult size, and thus with minor exceptions the cells are almost fully grown. Considering the early morphological (Donovan, '66) and physiological (Hamasaki and Flynn, '77; Rusoff and Dubin, '77) development of the area centralis relative to more peripheral parts of the retina, it is surprising that cells in the LGN which receive input from the central retina do not develop more rapidly than those which subserve peripheral vision. A careful comparison of the growth curves for the paracentral, peripheral, and monocular segments of lamina A fails,

279

however, to reveal any significant differences among them.4 Indeed, if one takes the establishment of an adult-like cell size distribution as an indication of maturity, then paracentral A cells develop somewhat slower than those in the periphery because they are the last group to achieve a mature distribution. However, we know little about the factors which regulate cell body size, especially those which involve the effects of stimulation. For example, complete visual deprivation during development causes LGN neurons to atrophy, but this effect is only temporary since cells eventually reach full size, even in the absence of visual input (Kalil, '78). Since cell growth in the LGN bears no simple relation to stimulus conditions in this extreme case, it is perhaps unreasonable to expect that morphological and functional developmental gradients in the retina should be reflected by differential cell growth in the LGN. Possibly, a correlation between LGN development and retinal gradients can be established with the electron microscope, but it will be necessary to study systematically retinogeniculate synapse formation a t central and peripheral locations in the LGN. At present, the only available electron microscopic data are those which show that most of the synaptogenesis in the LGN takes place during the first 30 postnatal days (Cragg, '75), although an adult ratio of symmetric to asymmetric synapses is not found until 6 0 to 70 days (Winfield et al., ' 7 6 ) . Perhaps the most puzzling finding in the present report is the occurrence of small, dark neurons, immediately caudal to the C laminae of the LGN, which appear to be completing their migration during the first two postnatal weeks. Young neurons possessing a spindleshaped cell body with cytoplasmic extensions are regarded to be migrating cells (Shimada and Langman, '70; Sidman and Rakic, '73). Indeed, in the material available for this study, migrating neurons with one or both of these features occur commonly a t many sites in the fetal brains. In postnatal animals they are seen frequently in the fiber tracts underlying various neocortical areas, for example visual cortex, as late as three weeks. A second possibility worth considering is related to the fact that spindle-shaped cells can be found in the mature LGN, in the region be< Similar results have been reported by Garey et al. ('73a) for cell growth in the binocular and monocular segments of lamina A during the first postnatal month. However, the present experiments do not confirm their finding that cells in these segments grow at differential rates later in development.

280

RONALD KALIL

tween lamina C1 and the optic tract (Hickey and Guillery, '74). Therefore, the small immature neurons seen during the first two postnatal weeks may be simply precursors of the adult spindle-shaped cells which have finished their migration, but have not completed differentation. Unfortunately, the genesis of the LGN in the cat has not been studied with thymidine autoradiography as it has in the monkey (Rakic, '77c), and thus critical details such as the time of neuron origin and the pattern of cell migration are unknown. Nissl-stained preparations by themselves cannot provide this information, and thus until the appropriate autoradiographic experiments are conducted, the precise stage of development of the small dark neurons cannot be specified. Functional considerations Spontaneous unit activity can be recorded in the LGN a t two days postnatal and adult discharge patterns are seen a t the beginning of t h e third week (Adrien and Roffwarg, '74). Moreover, nerve impulses elicited by visual stimulation are transmitted through the LGN as early as the third postnatal day, but cortical neurons are not driven until day 4 (Huttenlocher, '67). At eight to nine days t h e geniculocortical system is sufficiently mature that some cortical neurons display reliable stimulus specific responses such as orientation selectivity (Hubel and Wiesel, '63; Blakemore and Van Sluyters, '75). Although little specific information has been published concerning receptive field organization of kitten LGN neurons, two brief reports have appeared which indicate that surround antagonism in young geniculate cells is either weak or nonexistent as late as the third postnatal week (Glendenning and Norton, '73; Norman et al., '75). More recently, Norman et al. ('77) have studied the development of X - and Y-cells in the LGN. They found that prior to the beginning of the third postnatal week LGN cells are difficult to drive reliably, even following electrical stimulation of the optic chiasm, and of the cells which they were able to record from only one-third could be classified as X or Y. Although the percentage of classifiable cells increases after the third week, response latencies to stimulation of the optic chiasm remain abnormally long in kittens as old as six weeks. In comparing t h e development of X- and Y-cells, using measures such as response latency, receptive field size

and center-surround antagonism, Norman et al. ('77) conclude t h a t some X-cells are mature by 34 days, but most Y-cells have yet to achieve adult properties. As suggested by Norman et al. ('771, the long response latencies of kitten LGN cells probably result from the fact that many axons in the optic tract are unmyelinated prior to the fourth postnatal week (Moore et al., '76). However, myelin formation by itself cannot account for the earlier development of mature X-cell latencies, since there is no evidence (Moore et al., '76) that the small diameter optic axons which are presumed to drive LGN X-cells (Hoffman et al., '72; Wilson et al., '76) acquire myelin before the larger fibers t h a t innervate Y-cells. It is possible that retinogeniculate axons in the X-system form mature synapses before axons of the Y-system, and the resulting stability in synaptic coupling accounts, a t least in part, for the relatively early development of adult response latencies. At present it is difficult to test this hypothesis since there are no known morphological features which can be used to discriminate between different functional classes of retinogeniculate synapses. Nevertheless, the notion is worth pursuing for two reasons. Firstly, tentative physiological evidence has been adduced which suggests t h a t optic tract synapses upon geniculate X-cells may indeed develop earlier than Y-cell synapses (Norman e t al., '77). Secondly, if a conclusive demonstration can be made that Y-cell retinogeniculate synapses mature a t a slower rate than those on X-cells, then a plausible morphological basis would be available to account for the greater vulnerability of Y-cells to visual deprivation (Sherman et al., '72, '75; Kratz et al., '77). It is clear from the present results that the growth of cell bodies in the LGN does not appear to be related intimately to major functional sequences in the development of the retino-geniculo-corticalsystem. In fact, the normal growth of geniculate cells is essentially a linear process, as is general synaptic development in the LGN (Cragg, '75; Winfield et al., '76), and does not reflect such events as opening of the eyelids, functional gradients in the maturation o f the retina, emergence of mature response properties, or the onset and completion of the critical period for cortical ocular dominance (Hubel and Wiesel, '70). With regard to the relatively late development of large cells in the paracentral part of

DEVELOPMENT OF THE LGN IN THE CAT

lamina A it is possible to draw a parallel between the development of the cat LGN and that of the human. In his study of cell growth in the human LGN, Hickey ('77) determined that cells in the parvocellular layers reach full size a t about six months after birth, but neurons in the magnocellular laminae require nearly two years to complete their growth. Hickey ('77) proposed that this period of protracted growth might be related to a critical period in the development of the human visual system. In view of current physiological knowledge, it seems unlikely that a similar situation exists in the cat, but the late emergence of large cells in paracentral A does indicate that full development requires a longer time than has been generally suspected. Physiological changes occurring relatively late in development have not received much attention, however, monocular deprivation experiments have demonstrated functional alterations in the striate cortex in cats as old as 17 to 18 weeks. Thus if the lids of the experienced eye are sutured and the deprived eye is allowed normal vision for three or four months, a small but significant increase occurs in the percentage of cortical neurons driven by the initially deprived eye (Smith et al., '78). Moreover, if the experienced eye is enucleated following four months of monocular deprivation, the ability of the deprived eye to influence cortical neurons is enhanced dramatically (Kratz et al., '76; Hoffman and Cynader, '77). In summary, these physiological studies show that the geniculocortical system of the cat retains a capacity for limited modifiability which persists for several months after birth. Relevant to this is the present morphological evidence which indicates that maturation of the LGN is completed slowly, raising the interesting possibility that protracted development and long term plasticity are related. ACKNOWLEDGMENTS

I thank R. W. Guillery and P. D. Spear for their comments on the manuscript, M. J. Meyer, M. Rhodes, and R. Schmitt for technical assistance, and D. Reierson and 0. Feusahrens for typing. Supported by g r a n t EYO-1331 from the National Eye Institute. LITERATURE CITED Adrien, J., and H. P. Roffwarg 1974 The development of unit activity in the lateral genlculate nucleus of the kitten EXD.Neurol.. 43: 261-275. Anker, R.'L. 1977 The prenatal development of some of the visual pathways in the cat. J. Comp. Neur., 173: 185-204.

28 1

Berman, A. L. 1968 The brainstem of the cat. University Wisconsin Press, Madison. Blakemore, C., and R. C. Van Sluyters 1975 Innate and environmental factors in the development of the kitten's visual cortex. J. Physiol. (London),248: 663-716. Boyd, J. S. 1971 The radiographic identification of the various stages of pregnancy in the domestic cat. J . Small Anim. Prac., 12: 501-506. Burrows, G. R., and W. R. Hayhow 1971 The organization of the thalamo-cortical visual pathways in the cat. An experimental degeneration study. Brain, Behav. and Evol., 4: 220-272. Chacko, L. W. 1948 The laminar pattern of the lateral geniculate body in t h e primates. J. Neurol. Neurosurg. Psychiatry, 11: 211-224. Chow, K. L., and D. L. Stewart 1972 Reversal of structural and functional effects of long-term visual deprivation in cats. Exp. Neurol., 34: 409-433. Clark, W. E. Le Gros 1932 A morphological study of the lateral geniculate body. Brit. J. Ophthal., 16: 264-284. Cleland, B. G., W. R. Levick, R. Morstyn and H. G. Wagner 1976 Lateral geniculate relay of slowly conducting retinal afferents to cat visual cortex. J. Physiol. (London), 255: 299-320. Cragg, B. G. 1975 The development of synapses in the visual system of the cat. J. Comp. Neur., 160: 147-166. Cragg, B., R . Anker and Y. K. Wan 1976 The effect of age on the reversibility of cellular atrophy in the LGN of the cat following monocular deprivation: a test of two hypotheses about cell growth. J. Comp. Neur., 268: 345-354. Donovan, A. 1966 The postnatal development of the cat retina. Exp. Eye Res., 5: 249-254. Dubin, M. W., and B. G. Cleland 1977 Organization of visual inputs to interneurons of lateral geniculate nucleus of the cat. J. Neurophysiol., 40: 410-427. Dursteler, M. R., L. J. Garey and J. A. Movshon 1976 Reversal of the morphological effects of monocular deprivation in the kitten's lateral geniculate nucleus. J. Physiol. (London),261: 189-210. Elgeti, H., R. Elgeti and K. Fleischhauer 1976 Postnatal growth of t h e dorsal lateral geniculate nucleus of the cat. Anat. Embryol., 249: 1-13. Famiglietti, E. V., and A. Peters 1972 The synaptic glomerulus and the intrinsic neuron in the dorsal lateral geniculate nucleus of t h e cat. J. Comp. Neur., 144: 285-334. Farris, E. J. 1950 The Care and Breeding of Laboratory Animals. John Wiley, New York. Garey, L. J. 1968 The projection of the retina in the cat. J. Anat. (London), 202: 189-222. Garey, L. J., and C. Blakemore 1977 The effects of monocular deprivation on different neuronal classes in the lateral geniculate nucleus of the cat. Exp. Brain Res.. 28: 259-278. Garey, L. J., and T. P. S. Powell 1967 The projection of the lateral geniculate nucleus upon the cortex in the cat. Proc. Roy. SOC.B., 169: 107-126. Garey, L. J., R. A. Fisken and T. P. S. Powell 1973a Observations on the growth of cells in the lateral geniculate nucleus of the cat. Brain Res., 52: 359-362. 1973b Effects of experimental deafferentation on cells in the lateral geniculate nucleus of the cat. Brain Res., 52: 363-369. 1976 Cellular changes in the lateral geniculate nucleus of the cat and monkey after section of the optic tract. J. Anat. (London), 221: 15-27. Gilbert, C. D., and J . P. Kelly 1975 The projections of cells in different layers of the cat's visual cortex. J. Comp. Neur., 263: 81-106. Glendenning, R. L., and T. T. Norton 1973 Receptive field properties of lateral geniculate neurons in kittens. Third

282

RONALD KALIL

Ann. Meeting, SOC. Neuroscience, San Diego (Abstract). Guillery, R. W 1966 A study of Golgi preparations from the dorsal lateral geniculate nucleus of the adult cat. J. Comp. Neur., 128: 21-50. 1967 Patterns of fiber degeneration in the dorsal lateral geniculate nucleus of the cat following lesions in t h e visual cortex. J. Comp. Neur., 130: 197-213. 1969a The organization of synaptic interconnections in the laminae of the dorsal lateral geniculate nucleus of the cat. Z. Zellforsch. Mikroskop. Anat., 96: 1-38. 1969b A quantitative study of synaptic interconnections in the dorsal lateral geniculate nucleus of the cat. 2. Zellforsch. Mikroskop. Anat., 96: 39-48. 1970 The laminar distribution of retinal fibers in t h e dorsal lateral geniculate nucleus of the cat: a new interpretation. J . Comp. Neur., 138: 339-368. 1972 Binocular competition in the control of geniculate cell growth. J. Comp. Neur., 144: 117-130. 1973 The effect of lid suture upon the growth of cells in the dorsal lateral geniculate nucleus of kittens. J. Comp. Neur., 124: 149-160. Guillery, R. W.. and G. Scott 1971 Observations on synaptic patterns in the dorsal lateral geniculate nucleus of the cat: the C laminae and perikaryal synapses. Exp. Brain Res., 12: 184-203. Guillery, R. W., and D. J. Stelzner 1970 The differential effects of unilateral lid closure upon the monocular and binocular segment of the dorsal lateral geniculate nucleus in the cat. J. a m p . Neur., 139: 413-422. Hamasaki, D. I . , and J . T. Flynn 1977 Physiological properties of retinal ganglion cells of 3-week-old kittens. Vision Res., 17: 275-284. Hayhow, W. R. 1958 The cytoarchitecture of the lateral geniculate body in the cat in relation t o the distribution of crossed and uncrossed optic fibers. J. Comp. Neur., 110: 1.64. Hendrickson, A., and P. Rakic 1977 Histogenesis and synaptogenesis in the dorsal lateral geniculate nucleus (LGd) of the fetal monkey brain. Anat. Rec., 187: 602 (Ahstract). Hickey, T. L. 1977 Postnatal development of the human lateral geniculate nucleus: relationship to a critical period for the visual system. Science, 198: 836-838. Hickey, T. L., and R. W. Guillery 1974 An autoradiographic study of retinogeniculate pathways in the cat and t h e fox. J. Comp. Neur., 156: 239-254. Hickey, T. L., P. D. Spear and K. E. Kratz 1977 Quantitative studies of cell size in the cat's dorsal lateral geniculate nucleus following visual deprivation. J . Comp. Neur., 172: 265-282. Hoffman, K.-P., and M. Cynader 1977 Functional aspects of plasticity in the visual system of adult cats after early monocular deprivation. Phil. Trans. Roy. Soc. Lond. B., 278: 411-424. Hoffman, K.-P.,and H. Hollander 1978 Physiological and morphological changes in cells of the lateral geniculate nucleus in monocularly-deprived and reverse-sutured cats. J. Comp. Neur.. 117: 145-157. Hoffman, K.-P., and R. Sireteanu 1977 Interlaminar differences in t h e effects of early and late monocular deprivation on the visual acuity of cells in the lateral geniculate nucleus of the cat. Neuroscience Letters, 5: 171-175. Hoffman, K.-P., J. Stone and S. M. Sherman 1972 Relay of receptive-field properties in the dorsal lateral geniculate nucleus of the cat. J. Neurophysiol., 35: 518-531. Hollander, H., and H. Vanegas 1977 The projection from t h e lateral geniculate nucleus onto t h e visual cortex in the cat. A quantitative study with horseradish-peroxidase. J. Comp. Neur., 173: 519-536.

Huhel, D. H., and T. N. Wiesel 1963 Receptive fields of cells in striate cortex of very young, usually inexperienced kittens. J. Neurophysiol., 26: 994-1002. 1970 The period of susceptibility to the physiological effects of unilateral eye closure in kittens. J. Physiol. (London), 206: 419-436. Huttenlocher, P. R. 1967 Development of cortical neuronal activity in the neonatal cat. Exp. Neurol., 17: 247-262. Kaas, J. H., R. W. Guillery and J. M. Allman 1973 Discontinuities in the dorsal lateral geniculate nucleus corresponding to the optic disc: a comparative study. J. Comp. Neur., 147: 163-180. Kalil, R. E. 1975 Monocular enucleation in young kittens: effects on cell growth in t h e dorsal lateral geniculate nucleus. Paper presented at the spring meeting, Assoc. for Res. in Vision and Ophthalmology, Sarasota, (Abstract). -1976 Transneuronal atrophy of cells in the lateral geniculate nucleus of kittens. Paper presented a t the spring meeting, Assoc. for Res. in Vision and Ophthalmology, Sarasota, (Abstract). 1977 Effects of monocular deprivation on the growth of lateral geniculate cells in the cat. Neuroscience Abstracts, 3: 664 (Abstract). 1978 Dark rearingin the cat: effects on visuomotor behavior and cell growth in the dorsal lateral geniculate nucleus. J. Comp. Neur., 178: 451-468. Kratz, K. E., R. E. Kalil and S . M. Sherman 1977 Effects of dark rearing on physiology and anatomy of the cat's lateral geniculate nucleus. Neuroscience Abstracts, 3: 566 (Abstract). Kratz, K. E., P. D. Spear and D. C. Smith 1976 Postcriticalperiod reversal of effects of monocular deprivation on striate cortex cells in the cat. J. Neurophysiol., 39: 501-511. Kupfer, C., and P. Palmer 1964 Lateral geniculate nucleus: histological and cytochemical changes following afferent denervation and visual deprivation. Exp. Neurol., 9: 400-409. Laties, A. M., and J. M. Sprague 1966 The projection of optic fibers to the visual centers in the cat. J. Comp. Neur., 127: 35-70. LeVay, S., and D. Ferster 1977 Relay cell classes in the lateral geniculate nucleus of the cat and the effects of visual deprivation. J. Comp. Neur., 172: 563-584. LeVay, S., and C. D. Gilbert 1976 Laminar patterns of geniculocortical projection in the cat. Brain Res., 113: 1-19. Lin, C. S., K. E. Kratz and S. M. Sherman 1977 Percentage of relay cells in the cat's lateral geniculate nucleus. Brain Res., 131: 167-173. Marin-Padilla, M. 1971 Early prenatal ontogenesis of the cerebral cortex (neocortex) of the cat (Felis Domestica). A Golgi study. I. The primordial neocortical organization. Z. Anat. Entwick1.-Gesch., 134: 117-145. Moore, C. L., R. E. Kalil and W. Richards 1976 Development of myelination in optic tract of the cat. J. Comp. Neur., 165: 125-136. Morest, D. K. 1969 The growth of dendrites in the mammalian brain. Z. Anat. Entwick1.-Gesch., 128: 290-317. Movshon, J. A,, and M. R. Dursteler 1977 Effects of brief periods of unilateral eye closure on t h e kitten's visual system. J. Neurophysiol., 40: 1255-1265. Niimi, K., S. Kawamura and S . Ishimaru 1971 Projcctions of the visual cortex to the lateral geniculate and posterior thalamic nuclei in the cat. J. Comp. Neur., 143: 279-312. Niimi, K., and J. M. Sprague 1970 Thalamo-cortical organization of t h e visual system in the cat. J. Comp. Neur., 138: 219-250. Norman, J. L., J. D. Daniels and J. D. Pettigrew 1975 Absence of surround antagonism in unit responses of the

DEVELOPMENT OF THE LGN IN THE CAT LGN in young kittens. Neuroscience Abstracts, 1: 88 (Ahstract). Norman, J. L., J. D. Pettigrew and J. D. Daniels 1977 Early development of X-cells in kitten lateral geniculate nucleus. Science, 198: 202-204. O’Leary, J. L. 1940 A structural analysis of the lateral geniculate nucleus of the cat. J. Comp. Neur., 73: 405-430. Peters, A., and S. L. Palay 1966 The morphology of laminae A and A1 of the dorsal nucleus of the lateral geniculate body of the cat. J. Anat. (London), 100: 451-486. Rakic, P. 1977a Prenatal development of the visual system in the rhesus monkey. Phil. Trans. Roy. SOC.Lond. B., 278: 245-260. 197% Effects of prenatal unilateral eye enucleations on the formation of layers and retinal connections in the dorsal lateral geniculate (LGd) of the rhesus monkey. Neuroscience Abstracts, 3: 573 (Abstract). 1977c Genesis of the dorsal lateral geniculate nucleus in the rhesus monkey: site and time of origin, kinetics of proliferation, routes of migration and pattern of distribution of neurons. J.Comp. Neur., 176: 23-52. Rosenquist, A,, S. B. Edwards and L. A. Palmer 1974 An autoradiographic study of the projections of the dorsal lateral geniculate nucleus and the posterior nucleus in the cat. Brain Res., 80: 71-93. Rusoff, A. C., and M. W. Dubin 1977 Development of receptive-field properties of retinal ganglion cells in kittens. J. Neurophysiol., 40: 1188-1198. Sanderson, K. J. 1971 The projection of the visual field to the lateral geniculate and medial interlaminar nuclei in the cat. J. Comp. Neur., 143: 101-118. Shatz, C. J., S. Lindstrom and T. N. Wiesel 1977 The distribution of afferents representing t h e right and left eyes in the cat’s visual cortex. Brain Res., 131: 103-116. Sherman, S. M., K:P. Hoffman and J.Stone 1972 Loss of a specific cell type from the dorsal lateral geniculate nucleus in visually deprived cats. J. Neurophysiol., 35: 532-541. Sherman, S. M., J. R. Wilson and R. W. Guillery 1975 Evidence that binocular competition affects the-postna-

283

tal development of Y-cells in the cat’s lateral geniculate nucleus. Brain Res., 100: 441-444. Shimada, M., and J. Langman 1970 Cell proliferation, migration and differentiation in the cerebral cortex of the golden hamster. J.Comp. Neur., 139: 227-244. Sidman, R. L., and P. Rakic 1973 Neuronal migration, with special reference to developing human brain: a review. Brain Res., 62: 1-35. Smith, D. C., P. D. Spear and K. E. Kratz 1978 Role of visual experience in postcritical period reversal of effects of monocular deprivation in cat striate cortex. J. Comp. Neur., 178: 313-328. Stone, J.,and S. M. Hansen 1966 The projection of the cat’s retina on t h e lateral geniculate nucleus. J. Comp. Neur., 126: 601-624. Szentagothai, J., J.Hamori and T. Tomhol 1966 Degeneration and electron microscopic analysis of the synaptic glomeruli in the lateral geniculate body. Exp. Brain Res., 2: 283-301. Thuma, B. D. 1928 Studies on the diencephalon of the cat. I. The cytoarchitecture of t h e corpus geniculatum laterale. J.Comp. Neur., 46: 173-199. Updyke, B. V. 1975 The patterns of projection of cortical areas 17,18, and 19 onto the laminae of the dorsal lateral geniculate nucleus in the cat. J. Comp. Neur., 163: 377-396. Wan, Y. K., and B. Cragg 1976 Cell growth in the lateral geniculate nucleus of kittens following the opening or closing of one eye. J. Comp. Neur., 166: 365-372. Wiesel, T. N., and D. H. Huhel 1963 Effects of visual deprivation on morphology and physiology of cells in the cat’s lateral geniculate body. J. Neurophysiol., 26: 978-993. Wilson, P. D., M. H. Rowe and J. Stone 1976 Properties of relay cells in cat’s lateral geniculate nucleus: a comparison of W-cells and with X- and Y-cells. J. Neurophysiol., 39: 1193-1209. Winfield, D . A., M. P. Headon and T. P. S. Powell 1976 Postnatal development of t h e synaptic organization of the lateral geniculate nucleus in t h e kitten with unilateral eyelid closure. Nature, 263: 591-594

The LGN in horizontal section in a fetal cat a t approximately the middle of the eighth week of gestation. The nucleus in this animal occupies a n intermediate position, having completed about one-half of its caudalward displacement. Other details same as figure 8A.

C

comparison with the younger fetus (fig. 8B),the LGN in the older animal has rotated about 45' in the sagittal plane. Although cells are densely packed in the presumptive dorsal layers, laminae are not present. Other details same as figure 8B.Scale is 0.5mm and applies to all figures.

D A parasagittal section through the middle third of t h e LGN in the same fetus as shown in figure 8C.In

A parasagittal section through the middle of the LGN in the same cat as shown in figure 8A. The hilum of the nucleus faces ventrally, and continuity between the LGN and the anterior wall of the diencephalon is evident. Rostral is to the left, dorsal is a t the top of the figure.

B

8A A horizontal section through the middle of t h e LGN in a fetal cat a t approximately E42. Note the mediolateral compression of the LGN, and its location along the anterolateral margin of the developing diencephalon. At this age, the presumptive caudal pole of t h e nucleus is directed rostrally. Rostral is to the left, lateral is a t the top of the figure.

EXPLANATION OF FIGURES

PLATE 1

PLATE 2 EXPLANATION OF FIGURES

286

9A

Small, spindle-shaped neurons in the LGN of a newborn kitten. The C laminae are on the left. The optic tract is on the right. In comparison with neurons in t h e C laminae, the spindle-shaped cells are very basophilic. They are much larger than glial cells which can be seen scattered throughout t h e optic tract. Collectively, t h e spindle-shaped cells are oriented dorso-ventrally in this micrograph. Rostra1 is to the left and dorsal is a t the top of the figure. Scale, 50 p.

B

At higher power it can be seen (arrows) t h a t each spindle-shaped cell has a leading and trailing cytoplasmic process. Scale, 25 p.

C

Medium-size cell8 in lamina A of t h e LGN a t seven days postnatal. These cells are not seen in younger kittens, but appear abruptly a t one week. The arrow indicates a cell in the interlaminar zone that separates A from A l . Scale, 25 p.

DEVELOPMENT OF THE I G N IN THE CAT Ronald Kalil

PLATE 2

PLATE 3 EXPLANATION OF FIGURES

10A-€1 P a r a s a e t t a l sections through the middle third of the LGN in cats ranging in age from newborn to adult as follows: A,, newborn; B., 3 days; C., 7 days; D., 14 days; E., 28 days; F., 56 days; G., 140 days, H., adult. In the newborn kitten (10A) an interlaminar zone is seen only between A and A l , but a t 3 days (10B) a ventral interlaminar zone appears between A1 and C. By 7 days (1OC) the interlaminar zones have widened, and are prominent a t 14 days (10D) and thereafter. The rapid increase in cross-sectional area of the LGN during the first postnatal month is striking. In all figures rostra1 is to the left, dorsal is a t the top. Scale is 1.0 mm and applies to all figures.

288

DEVELOPMENT OF THE LGN IN THE CAT Ronald Kalil

PLATE 3

t4

0

W

the LGN during development. R, rostral; C, caudal; D, dorsal; V, ventral.

11 Projection drawings of the LCN in the sagittal plane at different postnatal ages. Each drawing has been aligned with the zero horizontal plane as described in the text to show the rotation of the major axis of

EXPLANATION OF FIGURES

PLATE 4

q v)

>a n

0

w -

m

d

291