The Projection from the Lateral Geniculate Nucleus onto the Visual Cortex in the Cat. A Quantitative Study with Horserad ish-Peroxidase HORSTMAR HOLLANDER ~ N DHORACIO VANEGAS I Mux-PlanLk-lnPtztutfur Pvychaatrzr, K r w p d z n s t r @ r 2, 8OfJO iMunrhvn 40,ARD

ABSTRACT Horseradish-peroxidase (HRP) was injected (9-18 p g in 0.03-0.06 p l ) into cortical areas 17, 18 or 19 of 11 adult cats. After survival times of 17 hours to 7 days, the thalamus was examined for retrogradely HRP labelled nerve cells in serial transverse sections. From these sections, the percentage of labelled cells occurring in each subdivision of the dorsal lateral geniculate nucleus (1,GNd) was calculated for each animal. One case each for injections in areas 17. 18 and 19 was then chosen for nerve cell size measurements irk each LGNd subdivision. Thc. perikaryal area of each labelled cell (N=689), and of representative samples of unlabelled cells (N=1137), was measured by planimetry. Size distribution histograms, mean values, standard deviations, and statistical significance levels were obtained by computer. It was found that area 17 receives a projection almost exclusively from laminae .4 and A l , and that the projecting cells belong to all cell size classes. Area 18 receives a projection mainly from laminar C and A l , and from the medial interlaminar nucleus (MIN). The projecting cells belong mainly to the large cell size classes. Area 19 receives a projection largely from MIN, and also from the (7-laminae and cxtrageniculate cell groups. The projecting cells helong to all cell size classes, with some emphasis on the large cells of lamina C . A significant projection was found to exist from the parvocellular laminae of LGNd onto area 19 and, to a lesser degrcbc. area 18. In conclusion, as one goes from area 17 to 18 and to 19 the projection source shifts from the A-laminae through the C-laminae on to VIN and extrageniculate cell groups. The cells which project to area 18 are on the whole larger than those which project to areas 17 and 19. ,I significant proportion of the contralateral visual input to area 18 is relayed via lamina C . These results provide a quantitative confirmation and extension of previous anatomical findings, and are in close relationship with physiological results regarding parallel channel processing in the visual system.

The results of investigations performed mostly during the last ten years have progressively disclosed increasing degrees of complexity in the projection from the dorsal lateral geniculate nucleus (LGNdj onto the visual cortex of the cat. One major aspect of such complexity is the fact that the various subdivisions of the LGNd, which in turn may receive separate projections from the retinae, project to different regions of the visual cortex (cf. Carey and Powell, '67; Wilson and Cragg, '67; Niimi J. W M P . NEIJR., 173. 519-536

and Sprague, '70; Burrows and Hayhow, '71; Rosencpist et al., '74; Maciewicz, '75; Gilbert and Kelly, '75;LeVay and Gilbert, '76). Another major aspect of this cotnplexity is borne out b y the finding that one and the same projection line may consist of parallel channels which differ in terms of cell size, conduction velocity and sensory information contents (Stone, '72; Hoffmaiin et Dozentenstiprndiat of the A . v . Humholdt-Stiftririg. On I r a v e of h e n c e from the Instituto \'cnr.zolanri ilv Invvstigacioiies Cientificas. Caracas, Venczriela.

519

520

IfOl{SThl4R IIO1.LANDEH A N D II0RAC:IO V.4NEGAS

al., '72; Stone and Dreher, '73).Among the or (2) poor labelling at the LGNd in terms different contributions which have heen of intensity of the HRP reaction or number made to the knowledge of geniculo-corti- of labelled nerve cells (see below); in some cal projections in the oat there has been cases such poor labelling seems to have substantial qualitative agreement, as well been due to wash out of the peroxidase by as some points of disagreement. In the bleeding arising from the needle track. present investigation we have undertaken The results to be described herein involve a quantitative study of the geniculo-corti- 11 cats, some of which were injected in cal projection b y means of small single both cerebral hemispheres (cf. table 1). injections of horseradish-peroxidase (HRP) into cortical areas 17, 18,or 19 and the sub- Surgery and injection procedure sequent analysis of retrogradely labelled Each animal was anesthetized with pennerve cells in the IXNd (cf. LaVail et al., tobarbital and placed in the stereotaxic ' 7 3 ) .The size and the relative number of frame. The cortical surface was then excells projecting to areas 17, 18 and 19 has posed, the dura mater was slit and reflectbeen evaluated here for each LGNd sub- ed, and the pia mater was carefully opened division (cf. Guillery, '70; Hickey and at the intended site of injection. A single Guillery, '74), including laminae A, '41, C injection of a 30% (9 mg in 30 pl) aqueous and C1 to C3, and the medial interlaminar solution of HRP (Roehringer, purity grade nucleus (MIN). It was found, that, in 1)was made in each hemisphere into areas general terms, area 17 receives its input 17,18 or 19, or the 17/18 border, at 0-8 mm almost exclusively from laminae ,4 and A l , rostra1 to the transverse interaural plane while the input to area 18 originates mainly (cf. table 1).The volume of each injection from laminae C and A1 and from MIN. was 0.03-0.06 kl (i.e., a total amount of 9Area 19 recieives its input largely from 18 p g of HRP), and was delivered in three MIK, and also from the C-laminae and from to eight minutes by means of a Hamilton extrageriiculate thalamic cell groups. microliter syringe hydraulically coupled to There is a significant projection from par- an automatic micropump (Schubert and vocellular laminae (C1 to C3) onto the Hollander, '75). The needle, 0.7 mm in divisual cortex, particularly to area 19. Final- ameter, was inserted 0.5-0.7 mm into the ly, the present investigation provides an cortex. After the end of each injection, evaluation of the differences among the several minutes were allowed to elapse LGNd projections to areas 17: 18 and 19. It before withdrawing the needle. Finally, will in fact be shown that the projections to the dural flaps were juxtaposed, the bone each of these cortical areas may arise from defect was closed with Gelfoam, and the a common set of LGPIjd subdivisions, but overlying muscle and skin were individthat, nevertheless, such projections are ually sutured. different from one another in terms of the Histological procedure und collection relative contribution made by each of of topographical data these subdivision and/or in terms of the size of the projecting nerve cells. After postoperative survival times of 17 hours to 7 days (cf. table 1), the animals MATERIAL AND METHODS were anesthetized with pentobarbital and This investigation was carried out in 25 transcardially perfused with 6% Macrodex adult cats of either sex. Several of these followed by a solution of 0.5% forwere used for pilot studies concerning pa- maldehyde and 2.5% glutaraldehyde in 0.1 rameters of cortical injection and postoper- M cacodylate buffer at pH 7.2 (cf. Karative survival time. Still other animals were novsky, '65 and Jacobson and Trojanowski, excluded from the data analysis because of '74). The brains were immediately ex(1) large cortical lesions unintentionally posed, blocked in transverse planes by made during the injection procedure, and/ means of a stereotaxieally guided knife,

LGNd PROJECTION TO CORTEX

and drawn in the frame by camera lucida after identification of the injection sites (cf. figs. ZA, 3A and 4A). The AP distance of the injection sites to the interaural transverse plane was then measured from the drawings. The blocks containing injection sites and thalamus bilaterally were stored overnight at 7°C in cacodylate buffer with 30% sucrose, and subsequently sectioned transversely, i.e., parallel to the stereotaxically cut block surfaces, at 40 p m on the freezing microtome. The sections were individually collected in cacodylate buffer, washed in tris buffer and incubated for at least 30 minutes in tris buffer containing 60 mgilOO ml 3,3'-diamimobenzidinetetrahyrochloride and 0.01% H z 0 2 (cf. Graham and Karnovsky, '66 and Jacobson and Trojanowski, '74). The sections were subsequently mounted, dehydrated without previous counterstaining, and enclosed in paraffin oil under cover slip. Each section was carefully scanned at 160 x to 400 x magnification, and the position of every retrogradely labelled nerve cell in the thalamus, as well as that of some landmarks such as blood vessels, was printed by means of an X-Y plotter coupled to the microscope and calibrated to a gain of 20 X . Every section or, in heavily labelled cases, every second section, was then counterstained with thionin, and the internal and external boundaries of the LGNd were carefully traced into the respective X-Y plot either by projecting the 40 pm section at 20 x onto the plot, or by superimposing the plot and the section by means of a camera lucida. In summary, each drawing contained a 20 X enlarged reproduction of the internal and external boundaries of the LGNd, as well as the position of every labelled nerve cell, in a 4 0 pm transverse section of the thalamus. In determining the internal divisions of the LGNd, the nomenclature and boundaries proposed by Guillery ('701 and Hickey and Guillery ('74) have been followed throughout this investigation. Some qualifications must be made in this respect. however. First, no distinctions were attempted between laminae C1, C2 and C3,

521

i.e., the parvocellular region located be-

tween lamina C and the optic tract was treated as a whole and referred to as C1-3. Second, some difficulty was generally experienced in determining the boundary between the posteroventral leaflet of the interlaminar plexus and lamina C, between lamina C and MIN, and, more ventrally, between MIN and C1-3. And third, the LGNd region which surrounds the medial edge of lamina A1 seemed to be a zone of confluence. In this region lamina C fuses with MIN, and even lamina A surrounding the medial edge of A1 may participate. Consequently in the cases where labelled cells were found in this region it was difficult to decide to which of these three subdivisions such cells belonged. Every cortical injection site was examined in serial sections conterstained with thionin. Selected sections were drawn in camera lucida, and the boundaries between areas 17, 18 and 19 (cf. Otsuka and Hassler, '62) were determined at higher magnification and introduced into the drawings (cf. figs. BB, 3B and 4B). Presentution of topographical data For each LGNd subdivision, the number of retrogradely labelled nerve cells was counted from the X-Y plots. In order to avoid counting some cells more than once, only the plot of every second section was used. Only cases with an overall total of 50 or more labelled cells are included in the present report (cf. table 1).For every thalamus ipsilateral to a given cortical HRP injection, the number of cells counted in each geniculate or extrageniculate subdivision was epxressed as percent of the total counted in that thalamus. An evaluation was thus obtained of the the relative contribution made by each LGNd subdivision or extrageniculate cell group to the projection onto the injected cortical area (cf. table 1). The mean values corresponding to each of these projection sources are shown by the column diagram of figure 1. In order to better visualize the topographic distribution of labelled nerve cells in each case, and to be able to compare it to

HORSl'hlAR HOI.I,AXDER

522 AREA

.ND HORACIO VANEGAS

5c

6%

17 + O/O

C

AREA

30

i 47%

18

+

54 Yo

I

1%

+ I

O/O

0

AREA

40

3

19

Yo

77 O/O

rfj !

20%

I I 1

Oh

I

0' Fig. I Percent of lahellcd nerve cells found in LGNd suhdivisions and in extragcniculatt: cttll groups (i.e., other), after HRP injection into the cortical areas indicated on the left. ilp indicates the interlamiiiar plexus located hetween the adjacently represented laminae. Each coliirnn represents the mean of the values obtained from the various experimental cases with a similarly placed injection (cf. table 1 ) . Values have heen approximated to the first integer, and therefore not always add up to exexactly IOU. The sums of' values bounded by dashed lines are given in figures. The ---t represent the values obtained in the cases chosen for nerve cell size measurement and illustrated in figures 2, 3 and 4.

the other cases, all brain sections containing a given LGNd were divided in ten equal groups arranged sequentially in rostrocaudal direction. The labelled nerve cells contained in all sections of any one

group were then all compiled upon the X-Y plot of the middle section of that group, as shown in figure 2C, 3C and 4C for three representative cases. In table 1, the figures in small print indicate in what rostrocaudal

1 24

6

2 1

131 239 30

110 141 250

17h

Sd

7d

Id

4d

4d

4.6

5.8

6.5

7.8

8

7.8

18

18

18

18

18

19

431r

408

406

407

"405

"456

25

426

Sd

6.4

17/18

4461

3

199

Jd

5.4

17118

4311

8

174

2d

3.7

17/18

430

9

162

2dh

0

17/18

4751

:34

1.53

0

24h

17/18

475r

:3 4

268

3

CUSBF

1-2

2-4

2

1-2

L

1-2

1

2

3 4

3-4

1 ~ 2

1 ~ 2

5d

6

I:

"446r

27

136

4d

4

17

454

26

91

5d

I

17

433

Summur!y of

1

2

2

2-3

2 3

2 4

2

1 2

13

9

2

28

3

6

2

3

2 3

9 4

2 3-4

1

6

3

2

3-4

12

4

2

2-3

2 - ,1

12

3 5

2 3

9

79

2

2-4

2-4

13

26

19

1

2-3

3-4

11

42

54

3-5

2-3

16

50

2-3

3-4

58

63

1

3

1-5

2-3

3

2-3

1

9

4-5

8

6

2

8

2-:i

3

4-5

18 6

?

4

3- 4

4

2

-

2

1

intrucorticall!i injected with IIRP

TABLE 1

4 4

3-4

3-4

2-3

3-5

2-5

.33

14

14

184 - r

,4

9 3-4

23

26

36

3-4

4-5

3-S

23

09

25

35

5

3-4

3 3

3-5

19

4-5

5

3

3-6

3-4

9

4

4-6

5

1-4

4

3

6

2

3

I1

2

3

3

6

2

3

1

40

44

46

3-4

3

3-6

4-ti

5-s

3-4

4-ti

5

25

5

4

5-6

10

3

4

80

4

1

4X

524

HORSTMAR HOLLhNDER AND HOHACIO VANEGAS

ARE A

A

17

11

Rr

1

- (I

446 r

10 rnrn

I 11111

2

3

4

A’

T L P 203 l a b 187

c1-3TLP

30i/I P i 05

40

20

65 lab 20

In

30

20 10 10

7 500

500

0 1000

0

500

1000

500

1000

CELL SIZE (p’)

Fig. 2 Case No. 446r. uith HRP injection into area 17. A. Camera lucida drawing of a dorsal view of the right cercbral hemisphere. showing the injection site (arrow), the stereotaxic houndaries of the hlock (dashed lines) and the location of the sections shown in B (Roman numerals). B. Serial sections at levels indicated in A showing the extent and approximate relative intensity of the peroxidase reaction. C . Diagram of serial sections in rostrocaudal sequence, each one representing the middle section of a group which comprises one-tenth of the LGNd (cf. hMTER1.U AND METHODS. table 1).Dots indicate location of labelled nerve cells found in each particular group of sections. D. The computer displays shown above give the relative position of all labelled (squares) arid unlabelled (points) nerve cells whose sizes were measured in order to obtain the values shown in table 2 as well as in the histograms below. The histograms indicate, for each LCNd subdivision, what percent of the total number of measured cells, included in any particular histogram, falls into each size class. The total numbcr of cells included in each histogram is shown in figures. Shaded histogram: labelled nerve cells. Thick-line histogram: “total local population” (TLP).The accompanying p-values apply in each case to the difference between the two distributions (not between the means). Mean values (cf. also table 2) are indicated by a black triangle (labelled nerve cells) and a white triangle (TLP). TLP involves only two of all 40 Frn sections where labelled nerve cells were found, and therefore does not include values obtained from all labelled cells.

525

LGNd PROJEC1'ION TO CORTEX

1 mm

lob

44

lob

20 -

20

10-

10

68

30 -

20-

2 UI

v

s OU 0

500

1000

CELL S I Z E ipm2)

Fig. 3 Case KO. 456 with HRP injection into area 18. A. Camera lucida drawing of a dorsal view of the right cerebral hemisphere, showing the injection site (arrow), the stereotaxic boundaries of the block (dashed lines) and the location of the sections shown in B (Roman numerals). The injection site and thalamus involved in this case were actually located in the left hemisphere, but, for the sake of convenience and uniformity. all original drawings for A, B and C have been reversed and arc shown here as if they were on the right hemisphere. R. Serial sections at levels indicated in A showing the extent and approximate relative intensity of the peroxidase reaction. C . Diagram of serial sections of LGNd in rostrocaudal sequence, each one representing the middle section of a group which comprises one-tenth of the LGNd (cf. blATERIALAND METHODS, table 1).Dots indicate h a tion of labelled nerve cells found in each particular group of scctions. D. Thc computer displays shown above give the relative position of all labelled (squares) and unlabelled (points) nerve cells whose sizes were measured in order to obtain the values shown in table 2 as well as in the histograms below. The higtograms indicate, for each LGNd subidivision, what percent of the total number of measured cells, included in any particular histogram, falls into each size class. The total number of cells included in each histogram is shown in fignres. Shaded histogram: labelled nerve cells. Thick-line histogram: "total local population" (TLP). The accompanying p values apply in each case to the difference between the two distributions (not between the means). Mean values (cf. also table 2) are indicated by a black triangle (labelled nervc: cells) and a white triangle (TLP).TLP involves only two of all 40 p m sections where labelled nerve cells were found, and therefore does not include values obtained from all labelled cells.

(RC) group the labelled nerve cells were located in all cases.

Andysis of nerce cell size One case each for HRP injections into areas 17, 18 and 19 was chosen for nieasurements of nerve cell size in the LGNd, and these are marked with asteriks in table 1. The choice was based on the concurrence of ( I ) good injection characteristics and, (2) the fact that the case’s percent distribution of labelled nerve cells was congruent with the central tendency of all cases with a similarly placed cortical injection. This latter point is depicted in figure 1, where the percent values of the cases chosen for cell size measurement are indicated. Cell size was measured only in those LGNd suhdivisions which showed a number of labelled nerve cells sufficiently large to warrant an adequate evaluation. Nerve cells sizes were measured by means of an apparatus which has been described elsewhere (Hollander et al., ’76). In short, the X-Y plots containing LGNd internal and external boundaries were placed on the plotter coupled to thc microscope so that every labelled nerve cell could again be found in the corresponding 40 p m section under oil immersion. By means of a beam splitter, a prism and a mirror, the image of the cell was projected onto a screen. Here the area of the perikaryon was measured planimetrically. This value, in pm2, was punched automatically on paper tape together with the x and y coordinates of the cell and a code indicating the LGNd subdivision involved plus whether the cell was labelled or unlabelled. All labelled nerve cells and representative samples of unlabelled nerve cells (see below), were measured in the LGNd of the three chosen cases provided that (1) the cell showed a clearly identifiable nucleolus, and (2) the cell was not located near or at a boundary between LGNd subdivisions. A total of 689 labelled and 1,137 unlabelled nerve cells met these two provisos and were therefore measured (cf. table 2, fig. 2D, 3L) and 4D), and the next paragraph refers only to such cells. All cell

size data were processed by means of an I130 IRM computer. In order to obtain representative samples of unlabelled nerve cells the following procedure was undertaken. For a given LGNd subdivision, the distance between the most rostral and the most caudal serial section containing labelled cells was divided by 3, and in the two sections located at the junction of the rostral and middle thirds, and of the middle and caudal thirds, respectively, all the unlabelled nerve cells were measured which were found within the cluster of labelled nerve cells in that particular LGNd subdivision (cf. figs. 2D, 3D and 4D). This population of unlabelled nerve cells may thus be used as a standard against which the size of labelled cells can be compared (cf. Gilbert and Kelly, ’75). Perhaps more appropriate, however, is to compare the labelled nerve cells against the total nerve cell population to which they belong. In order to obtain an indication of this population’s cell size distribution a pool was made of all values obtained from labelled and unlabelled nerve cells found in the two 40 p m sections defined above. This pool will be called “total local population” (TLP) for a given LGNd subdivision, and represents, again, the population of nerve cells some of whose members have been retrogradely labelled from the cortical HRP injection site (cf. table 2, figs. 2D, 3D and 4D). The differences between the distribution (not the mean) of labelled cell sizes vs. TLP cell sizes were statistically evaluated by means of the Kolmogorov-Smirnovtest, and the p values are shown with the distribution histograms in figs. 2D, 3D and 4D. RESU1,TS

General Retrograde HRP labelling found in this study in nerve cells of the LGNd and other thalamic cell groups was of the “granular” type described, among others, by LaVail and LaVail (’74) and Nauta et al. (’74) (fig. 5D-I). No “diffuse” type labelling was found in nerve cell bodies of subcortical structures. Abundant brown granules were

527

LCNd PROJECT109 TO CORTEX A

AREA 19 405

c ?

10 m m

I

11

n

h

/I\

I II I l l

U

I

1m m

D C

TLP 118 l o b 20

C1-3

TLP 154 l o b 41

. ...,~,...-.. ....... . . ....... . .. . . . _. . ..T‘ . . . . ....

M”TLP

131 lob 68

.. . . ........... . .. . %.

.;.%’:.**

/..

’

.

:

.’.

I .

I

.... . ...... . . . . :. ..... . . .;..*. . . . .*.. .. . .. .>,a

>,

:,!A:;.

. . . .. .. .. ... . ..

-. . ..

CELL S I Z E (pm’)

Camera lucida drawing of a dorsal view Fig. 4 Case No. 405, with HRP injection into area 19. -1. of the right cerebral hemisphere, showing the injection site (arrow), the stcreotaxic boundaries of the block (dashed lines) and the location of the sections shown in R (Roman numerals). The injection site and thalamus involved in this case were actually located in the left herriisphcre, but, for the sake of convenience and uniformity, all original drawings for A, B and C have been reversed and are shown here as if they were on the right hemisphere. B. Serial sections at levels indicated in A showing the extent and approximate relative intensity of the peroxidase reaction. C. Diagram of serial sections of LGNd in rostrocaudal sequence, each one representing tht, middle section of a group which comprises one-tenth of the LGNd (cf. M A ~ . E H I . ~AND L METHODS, table 1). Dots indicate location of labelled nerve cells found in each particular group of sections. D. The computer displays shown above givt: the relative position of all labelled (squares) and unlabelled (points) nerve cells whose sizes were measured in order to obtain the values shown in table 2 as well as in the histograms below. The histograms indicate, for each LGNd subdivision, what percent of the total number of measured cells, included in any particular histngram, falls into each size class. The total number of cells included in each histogram is shown in figures. Shaded histogram: labelled nerve cells. Thick-line histogram: “total local population” (TLP).The accompanying pvalues apply in each case to the difference between the two distributions (not between the means). Mean values (cf. also table 2) are indicated by a black triangle (labelled nerve cells) and a white triangle (TLP). TLP involves only two of all 40 pin sections where labelled nerve cells were found, and therefore does not include values obtained from all labelled cells.

528

17

4 A1

c Cl-3

18

456

A1

C C1-3

MIN 19

405

309.7

96

4461

C

C1-3

MIX

88 5

269.5

187

944

247.7

19

99.4

154.5

20

41 7

598.2

44

174.6

475.8

68

143.0

293.8

17

116.1

446i$0

109

519.7

20

177 2

261.y0

41

367.4

68

140.5

266.3

113

107.9

240.3

203

105.5

217.9

82

123.1

131.7

65

41.2

3 12.0

167

lfi4.2

242.9

111

137.3

154.6

79

48.6

317.1

98

148 7

11295ii5.3

231.1

154

794

13?19i%

also seen in some perivascular cells in the LGNd. These cells were always found within or adjacent to the clusters of labelled nerve cells.

237.5

1.2

1.3

239.0

1.1

1.1

219.9

1.1

1 .I

132.8

1.2

1.2

293.6

1.9

2.0

222.3

2.0

2.1

152.6

1.6

1.9

296.2

1.4

1.5

3H6i;8,1

1.3

1.4

232.!9,0

1.1

1.1

317.1

1.2

1.2

65

136 $3

61

156 97

75 97

106.5 112.6 129.2 41.4

146.7 120.5 46.4

144.4

110

149 118

147.0

used here as well as in another study involving similar HRP injections (Singer et al., '77), the spread of the brow7n reaction product in the cortex is considerable (cf. figs. 3B and 5 A ) . This product seems to be Postoperative sumivnl times subsequently cleared, and only remains -4wide range of survival times, i.e., 17 around the injection site (cf. figs. 2B, 4B hours to 7 days, was used in the present and 5B), which is presumably the only study in order to avoid inaccuracies due to place from which the peroxidase is taken putative different transport rates by nerve up by the nerve terminals, as judged by the cells of different size. In retrospect, such a limited and precise topographic location of wide range appears unnecessary, since no the labelled cells' clusters in the LCNd (cf. differences in the size distribution of la- figs. ZC, 3C, 4C and 5C).This problem, i.e., belled nerve cells were found. for example, the spread of intracortically injected HRP, between a case with 17 hours' survival and its uptake and transport by nervous strucanother with 4 days' survival, both with tures and its removal, will be considered in area 18 injection (see below). In agreement detail elsewhere (Vanegas et al., '76). In with other authors (Gilbert and Kelly, '75; connexion with the present study, howLaemle, '75; Maciewicz, '75) a postinjec- ever, it may be pointed out that at two tion survival of one or two days seems to be hours after injection the peroxidase is adequate and convenient for the present found in a large number of axons just below type of study. At the shorter survival times the cortex. Many of these axons show the

LGNd PROJECTION TO CORTEX

529

dichotomizing branchings (fig. 5J) which The cell size measurements indicate have been postulated (cf. Garey and (Case 446r, table 2 , fig. 2D) that the popuPowell, '67; Stone and Dreher, '73) for the lation of nerve cells labelled after area 17 LGNd projection to areas 17 and 18. The injection comprises all cell sizes in a relanumber of such branchings decreases at tively uniform fashion. four hours after injection, when the peroxArea 18 injection idase has descended from the cortex but cannot yet be found at the LGNd. Very few HRP injection into area 18 (table 1, figs. infracortical branchings can still be seen at 1,3) resulted in retrograde labelling of eight hours after injection, and at this time nerve cells located in all LGNd suhdivithe peroxidase begins to appear weakly sions. Compared to area 17 injections, the in nerve cell bodies at the LGNd. By 17 percentage of labelling in laminae A and hours after injection, the full picture of A 1 was smaller, and that in the C-laminae retrogradely labelled nerve cells in the was larger. In addition, a considerable amount of labelling was found in MIN. ParLGNd is obtained (figs. 3C and 5K). It was felt that the number of labelled ticularly noteworthy is the fact that after perivascular cells increased with survival area 18 injections, the mean percentage of labelling in lamina C was more than twice time, but this was in no way quantified. as large as that in lamina A. Some labelling Rostrocaudal topography of the geniculo- in extrageniculate nerve cells was found in cortical projection two cases (Nos. 4311- and 407). Of these, In agreement with great many studies one cell was located in the LP nucleus of (cf. Garey and Powell, '67; Bilge et a]., '67) the thalamus while seven cells were found it was here found that the more rostral the to be embedded in the sheet of radiation cortical injection site, the more rostral also fibers located between the LGNd and the was the cluster of labelled nerve cells in LP-pulvinar complex (cf. fig. 5L). N o lathe LGNd (i.e., lower RC value in table 1, belled cells were seen in the pulvinar, the small print). It is also clear from table 1 (cf. posterior nucleus or the ventral lateral also Gilbert and Kelly, '75; Maciewicz, '75) geniculate nucleus. The cell size measurements indicate that the clusters of cells labelled after any one cortical injection shifted within the (Case 456, table 2, fig. 3D) that the populageniculate from dorso-rostra1 ot ventrocau- tion of cells labelled after area 18 injection dal, thus involving the laminae sucessively involved almost exclusively the mediumfrom A to C1-3 in congruence with the ori- size and large nerve cells in any LGNd subentation of the "projection columns" division considered, including the par(Garey and Powell, '62; Bishop et al., '62; vocellular laminae C1-3. This is particularStone and Hansen, '66; Sanderson, '71), ly remarkable for laminae A l , C and C1-3, and, finally, the MIN (cf. figs. 2C, 3C and where the mean size of labelled nerve cells is twice as large as that of unlabelled ones. 4C). The labelled cells' size spectrum peaks Area 17 injection around 500 p m Zin lamina C and MIN, and HRP injection into area 17 (table 1, figs. around even higher values in lamina A l , 1, 2) resulted in intense retrograde label- where the largest nerve cells of the LGNd ling of LGNd nerve cells located almost ex- were found. In laminae C1-3 the labelled clusively in laminae A and A l . No labelled cells' size spectrum peaks around 250 pm2, cells were found in MIN, and a few were while the TLP peak is around 100 km2. It found in laminae C and C1-3. No labelled could be thought that such overall bias cells were found in thalamic structures towards the large cell end of the spectrum located outside the LGNd, including the might be due to the fact that, with the surposterior nucleus, the LP-pulvinar complex vival time of only 17 hours, smaller LGNd and the ventral lateral geniculate nucleus. cells would have had no time to be

530

HORSTMAR HOLL.4NL)EK AND HOHrZCIO VANECAS

retrogradely filled with HRP. However, cell size measurements done for lamina C and MIN in another case with a similar injection but with four days’ postoperative survival (No. 406, cf. table l),showed similar cell size distribution histograms and mean values. Indeed, the ratio of the mean size of labelled nerve cells vs. that of unlabelled ones was 2.0 for lamina C and 1.5 for MIN.

(table 2, fig. 4D) that the labelled nerve cells in laminae (21-3 and MIN belonged to all size classes, with only a slight emphasis on the medium size and large cells of MIN. Such an emphasis was more marked for lamina C, although b y no means so intensely as after area 18 injections. It must be pointed out that in this animal the nerve cells measured were overall bigger than in cases 446r and 456.

lnjection in the 1 7/18 boundary HRP injection into the boundary between areas 17 and 18 (table 1)resulted in a labelling pattern which combines elements of either area 17 or area 18 injections alone, including labelling in MIN alongside relatively large labelling in lamina A. Minimal labelling was found in the LP nucleus in two cases (No. 4461 and 475), but only in one of these reached such labelling 1%of the total (No. 4461, table 1). No labelling was found in the pulvinar, the posterior nucleus or the ventral lateral geniculate nucleus.

DISCUSSION

General The present investigation was based on injections of very small amounts of HRP (918 ~g in 0.03-0.06 pl). Such small quantities were used in an attempt to avoid HRP spread into cortical areas located adjacently to the area aimed at by the injection. It has been found (Vanegas et al., ’76) that such amounts of HHP remain highly restricted within the cortex throughout the first two hours following injection. Thereafter the enzyme begins to spread through a relatively large volume of cortex and this results in a picture similar to figures 3B and SA. Such spread, however, occurs after the geniculo-cortical axon terminals have been labelled by HRP (Vanegas et al., ’76) and, for yet unknown reasons, does not result in further retrograde labelling of LGNd

Area 19 injection An adequately placed HRP injection into area 19 was obtained only in one case !No. 405, table 1,fig. 4). Such paucity permits of course no evaluation of central tendency and variability, but it was nevertheless felt that the results obtained were sufficiently Fig. 5 A. Peroxidase reaction product in visual clear to warrant a tentative comparison cortex eight hours after injection, Nissl stained, 40 prn with those of injections in areas 17 and 18. transverse section. Calibration line, 1 mm. B. Similar to A, hut five days after injection. C. Cluster of HRP Such comparison indicates (fig. 1)a distinct labelled nerve cells in lamina A l , five days after area overall shift of the labelling pattern to- 17 injection. Transverse 40 p m section without counwards MIN, laminae C1-3 and extragenicu- terstains. Calibration line: 0.2 mm. D H . HRP labelled late cell groups. Labelling in lamina C was nerve cells in LGNd, Nissl stained, transverse 40 p m sections. Calibration line, 10 p. I. Nerve cell labelled less than after most of the injections in area after area 19 injection, located in thc sheet of fibers 18; labelling in lamina A1 was only 2 (cf. L, this figure) which separates LGNd from the LPpercent of the total; and there was no la- pulvinar complex. Nissl stained, transverse 40 pm. belling in lamina A. Of all labelled nerve Calibration line, 10 p m J. HRP labelled axonal cells, 2% were located in the posterior nu- branching in the white matter just beneath area 18, four hours after injection. l’ransverse 40 p m section cleus of the thalamus, and 18%were found without counterstain. Calibration line, 10 pm. K. HRP in the sheet of fibers interposed between labelled axon and cell in lamina A l , 17 hours after the LGNd and the LP-pulvinar complex area 18 injection. Transverse 40 p m section without (figs. 51,L). No labelled cells were seen in counterstain. Calibration line, 50 pm. L. Dashed lines delineate the sheet of radiation fibers between the the pulvinar or the ventral lateral genicu- LGNd and the LP-pulvinar complex. Nissl stained, late nucleus. transverse 15 p m celloidin section. Calibration line, The cell size measurements indicate I mm.

531

Figure 5

532

HOR5TM.IH HOLLANDER AND HORACIO 1 4 N Y C 13

nerve cells, as evidenced by the limited extension of the labelled cells' cluster (cf. figs. 2C, 3C, 4C and 32; cf. also Jones and Leavitt, '74). It can therefore be confidently assumed that in the present study the possibility of "falsely positive" results has been kept to a mininium. The possibility that the size of nerve cells was affected by storage of HRP could not be excluded.

Sources of the geniculo-cortical projection The sources of the geniculo-cortical projection in the cat have been earlier investigated by means of retrograde cell changes after axotomy (Garey and Powell, '67; Niinii and Sprague, '70; Burrows and Hayhow. T l ) , anterograde axonal degeneration methods (Wilson and Cragg, '67; Burrows and Hayhow, '71). anterograde axonal transport of radioactive tracers (Rosenquist et al., '74; LeVay and Gilbert, '76) and retrograde axonal transport of HRP (Maciewicz, '75; Gilbert and Kelly, '75; Laemle, '75). In many of these previous studies, two regions of the LGNd are considered as separate entities, namely, the "laminar" LGNd (sometimes involving only laminae A and A l ) , and the medial interlaminar nucleus. The results of the present investigation further provide an account of the individual projection arising from each of the LGNd subdivisions which could be distinguished in our material, as well as a quantitative appraisal of the degree to which these subdivisions contribute to the geniculo-cortical projection. The presence of a projection from laminae -4and A 1 to areas 17 and 18 is well documented (cf. Garey and Powell, '67; Niiini and Sprague, '70; Maciewicz, '75; Gilbert and Kelly, '75; LeVay and Gilbert, '76). We have now found that the percentage of cells projecting from lamina A to area 17 is quite high, and to area 18 comparatively low (table 1, fig. 1). In agreement with the authors just mentioned, no cells were found to project from lamina A to area 19. The projection from lamina A l to area 17 arises from a large number of cells, and that to area 18 arises from a not

so large, but nevertheless sizable, proportion of cells (table 1, fig. 1). Some cells were found to project from lamina A1 to area 19, but these were very few (table 1, fig. t ) , a fact which probably explains why this projection escaped observation in earlier studies (cf. Garey and Powell, '67; Wilson and Cragg, '67; Niimi and Sprague, '70; Burrows and Hayhow, '71; Rosenquist et al., '74; Maciewicz, '75; Gilbert and Kelly, '75; LeVay and Gilbert, '76). The projections from lamina C had not been previously studied in detail, although Garey and Powell's ('67) figures suggest a projection from the ventral leaflet of the nucleus interlaminaris centralis (which might have largely involved lamina C as defined later by Guillery, '70) onto area 17 and 18. Niimi and Sprague ('70) extended this finding to area 19, and Burrows and Hayhow ('71) report a projection from what appears to be lamina C onto area 18. By means of HRP retrograde labelling, Maciewicz ('75)showed (cf. his fig. 2) that "the C-laminae" project to areas 18 and 19, and LeVay and Gilbert ('76), by means of anterograde radioactive tracing, confirmed these findings and added that of a projection to area 17. From all of these results, however, it is not clear which of the C-laminae are responsible for the projection to which cortical areas. In the present investigation, lamina C was indeed found to project to area 17,18 and 19 (figs. 2C, 3C, 4C). The projection from lamina C onto area 17 (possibly also seen b Laemle, '75)involves relatively few cells table 1,fig. 11, and this may be the reason why it was not apparent to Maciewicz ('75). The projection from lamina C onto areas 18 and 19 involves a fairly large percentage of cells (table 1, fig. I), and this percentage is larger than that for lamina A. These results could be interpreted as evidence that lamina C is Inore important than lamina A as relay of contralateral visual input to area 18 and, certainly, to area 19. The information so far available on the projection arising from the parvocellular lawiinae C1-3 is rather vague. It can be inferred from the figures of Garey and

i:

Powell ('67) and Niimi and Sprague ('70) that laminae C1-3 project at least to area 17, although this was not clearly confirmed by similar experiments of Burrows and Hayhow ('71). Nevertheless, after HRP injection into area 17, Gilbert and Kelly ('75) report that the zone of retrograde labelling in the LGNd "stretched to the most ventral parts of the nucleus", and Laemle ('75) definitely saw retrograde HRP labelling in the parvocellular laminae. Finally, Maciewicz ('75) has shown retrograde labelling in the parvocellular laminae after HRP injection in areas 18 or 19, and LeVay and Gilbert ('76) found labelling in areas 17, 18 and 19 after radioactive proline injections that probably involved the parvocellular laminae. Our present results indicate that the parvocellular laminae C13 indeed project to areas 17, 18 and 19 (figs. 2C, 3C, 4C) and that the percentage of cells projecting to area 17 is minimal, to area 18 small, and to area 19 large (table 1, fig. 1). The projection from the medial interlaminar nucleus onto areas 18 and 19 is a well documented fact (cf. particularly Rosenquist et al., '74, and Maciewicz, '75, but also Niimi and Sprague, '70), although Garey and Powell ('67) admit that in their material, a projection from MIN onto area 18 could be only indirectly inferred, and Gilbert and Kelly ('75) have recently reported negative findings concerning the projection from MIN onto area 19. These latter authors also report retrograde labelling in MIN after HRP injections in area 17. Our results indicate that MIN projects to area 18 and, more considerably, to area 19 (table 1, figs. 1, 3C, 4C). A projection from MIN to area 17 was not apparent to us. Our findings regarding the projectiox of MIN are in agreement with those of Garey and Powell ('67),Wilson and Cragg ('67),Niimi and Sprague ('701, Burrows and Hayhow ('71), Rosenquist et al. ('74), Maciewicx ('75) and Laemle ('75). The presence of a projection from extrageniculate thalamic cell groups onto areas 18 and 19 has been reported by several authors. Although the issue is by no means

settled, the posterior nucleus (PN) and the nucleus lateralis posterior (LP) have been shown to project to area 19 and, less consistently, area 18 (Niirni and Sprague, '70; Burrows and Hayhow, '71; Graybiel, '72; Niimi et a]., '74; Rosenquist et al., '74; Maciewicz, '75j. We found a few lal>elled cells in LP after injections into area IS or the 17/18 border, and also a few in PN after injection in area 19, but by far most of the extrageniculate cells labelled after such injections (cf. fig. 51) were found to be located (cf. HF:SUL.TS) within the sheet of white matter which separates the LGNd from the LP-pulvinar complex (cf. fig. 5L). It is not clear whether this zone has been considered as part of PN or LP by the authors just mentioned. As regards the pulvinar, Maciewicz ('75) found labelled cells in this structure after HRP injections in area 18 or 19, but not after injections in area 17. Gilbert and Kelly ('75) likewise obtained labelling in pulvinar after HRP injections in area 18 or 19, and additionally after injections in area 17 or, even more intensely, in the 17/18 border, In our material, no labelled cells occured in the pulvinar after injections in area 17, 18 or 19, or in the 17/18 border (cf. HESLLTS), even though we purposely used, in one case with injection in the 17/18 border (No. 475, both hemispheres, cf. table l ) , the same injection placement (AP zero) and survival time (24 hours) as Gilbert and Kelly ('75) did. Our failure to find a projection from pulvinar onto area 17, 18 or 19 confirms equally negative results of Wilson and Cragg ('671, Cliiver and Campos-Ortega ('69), Graybiel ('72),Niimi et al. ('74) and Rosenquist et al. ('74).

Size of the projecting cells We have found that, for any one LGNd subdivision considered (ie., laminae A, A l , C, C1-31, the size spectrum of cells which project to area 17 is remarkably similar to the general cell size spectrum (TLP) of the particular LGNd subdivision (fig. 2D). The projection to area 18, however, arises from cells whose mean size is 1.4 to 2.0 times greater than the mean size of the general

cell population (TLP) of any LGNd subdivision considered, including laminae A l , C, C1-3 and the MIN.2 Gilbert and Kelly ('75) found that after HRP injections into area 17, the labelled cells in lamina A were on the whole smaller than the unlabelled ones. This is at variance with the conclusions of Garey and Powell ('67) that all cell sizes in LGNd project to area 17, as well as with our own data (fig. 2D) and those of Laemle ('75), who showed that, after HRP injection into area 17, the labelled cells in laminae A and A1 were of all sizes, i.e., 25% small, 50% medium size, and 25% large. Our data are in agreement with those of Gilbert and Kelly ('75) as regards the general size distribution in lamina A1 after area 17 injection, and fully support their findings as regards cell size measurements in lamina A1 after area 18 injections. The results obtained in the present study after HRP injections in area 17 or 18, together with those of Gilbert and Kelly ('75) after area 18 injections, provide a quantitative confirmation of the remarkable conclusion of Garey and Powell ('67) that area 17 receives afferents from the small, medium-size and large cells of the LGNd laminae, while only the large cells project to area 18. Although our data (fig. 3D) and those of Gilbert and Kelly ('75) indicate that a good number of medium size cells do project to area 18, these cell5 constitute only the lower tail of the distribution, and this in no way affects Garey and Powell's ('67) conclusion. Garey and Powell ('67) also indicated that the large LGNd cells project to areas 17 and 18 simultaneously by means of a branching axon. Such axons have not been anatomically identified, although Szentagothai ('72) shows axonal branchings just below the visual cortex (cf. his fig. 14).We have found large numbers of HRP labelled dichotomizing axons (fig. 5J) in the white matter immediately beneath the cortex up to four hours after injections in area 17 or 18 (j7anegas et al., '76). These dichotomizing axOnS have their branches pointing towards the cortex, and disappear as the

HRP retrogradely moves toward the LGNd (cf. RESULTS). The anatomical information revealed by the data of Garey and Powell ('67), Gilbert and Kelly ('7*5),and those presented herein, has it physiological counterpart in the experiments of Stone and Hoffmann ('711, Hoffmann et al. ('72) and Stone and Dreher ('73), among others (cf. also Stone, '72). These neurophysiological experiments have established that the Y-type (and presumably largest) relay cells in the LGNd laminae project to both areas 17 and 18 b y means of a branching axon, and that the X-type (and presumably not so large) relay cells project only to area 17. Furthermore, it has been shown (Mason, '75) that also MIN contains a large proportion of Ytype cells, and we have found that the cells which project from MIN to area 18 are roughly of the same size as those projecting from laminae A1 and C. If these large cells of MIN are the Y-cells found by Mason ('75), they would constitute an example of Y-cells without connection to area 17. There are no previous data regarding the size or physiological type of LGNd cells projecting onto area 19. Our data indicate (fig, 4D) that, as regards lamina C and MIN, the medium-size cells and - more considerably in the case of lamina C - the large cells, are the ones that project to area 19. These cells might presumably correspond to the physiologists' X- and Y-type cells. Wilson and Stone ('75) have found that neurons located "between lamina A1 and the optic tract" belong to the W-type (and presumably smallest) relay cells of LGNd and project to the visual cortex. On the other hand, we have shown (figs. 2D, 3D, 4D), that the parvocellular laminae C1-3 indeed project to the visual cortex, mostly to area 19 (fig. 1).It is noteworthy that, as with other LGNd subdivisions, the cells L)W

to the small number of labelled nerve cells in lamina

4 of the two cases w h e w cell sizc was measured after area 18 injections-i.e., Ns. 456 and 406, cf. table l-it was not possible t o obtain a balid qiiantification of the cell size distribution in lamina A aftrr area 18 injection.

LCNd PROJECTION TO CORTEX

535

which project from these Iaminae onto found in lamina A1 (ipsilateral visual areas 17 and 19 are of the same size as the input), lamina C: (contralateral visual input) general local population (TLY), but those and MIN. It would follow from the above that project to area 18 are larger. that the majority of the dichotomizing Finally, it may be worth to note that, axons observed beneath areas 17 and 18 after HRP injection in areas 17, 18 and 19, must come from cells located in lamina A 1. a large proportion of the smallest cells ACKNOWLEDGMENTS measured in any LGNd subdivision were unlabelled cells (figs. 2D, 3D, 4D). The The authors wish to thank Miss Ruth question whether these cells project to Muller, Miss Renate Scheil and Mrs. Briother cortical or subcortical regions, or are gitte Schwarz for excellent technical assisinstead short axon cells, or both, cannot be tance and secretarial help. The Dozenanswered at this time. tenstipendium from the A.V. HumboldtStiftung to H.V. is also gratefully acknowlFINAL CONCLUSIONS edged. 1.Most of the LGNd cells which project LITERATURE ClTED to area 17 are located in the A-laminae, Bilge, M.. A. Bingle, K. N. Scneviratne and D. Whitvery few in the C-laminae, and none in the teridge 1967 rl map of the visual cortex in the cat. MIN. Most of the cells which project to J. Physiol. (London), 151: 116P-118P. area 18 are located in lamina A l , lamina C Bishop, P. O., W. Kozak, W. R. Levick and G. J. Vakkiir 1962 The determination of the projection of and MIN. Most of the cells which project to the visual field onto the lateral gcniculate nucleus area 19 are located in the C-laminae, in of the cat. J. Physiol. (London). 163: 503-539. MIN and in parageniculate cell groups. Burrriws, G. R., and W. R. Hayhow 1971 The organization of the thalamo-cortical visual pathways in the That is, as one goes from area 17 to 18 and cat. An experimental degeneration study. Brain 19, the source of the projection shifts, withRehav. and Evol., 4: 220-272. in the LGNd, from the A-laminae through Cliiver, P. F. de V.;and J. .4.Campos-Ortega 1969 the C-laminae on to MIN. The source of the The cortical projection of the pulvinar in the cat. J. Comp. Neur., 137: 295-308. projection to area 18 cover? this whole range of LGNd subdivisions, but the pro- Carey, L. J., and T. P. S. Powcll 1967 The projection of the lateral geniculate nucleus upon the cortex in jecting cells are on the whole larger than the cat. Proc. Roy. Soc. (Biol.), 165: 107-126. those projecbting to area 17 or 19. There- Gilbert, C . D., and J. P. Kelly 1975 The projections of fore the projections from LGNd onto areas cells in different layers of the cat’s viual cortex. J. Comp. Neur.. 163: 81-106. 17, 18 and 19 can be said to be different from one another in terms of source, cell Graham, R. C., and M. J. Karnovsky 1966 Thc early stages of absorption of injected horseradish peroxsize compositon, or both. idase in the proximal tubules of mouse kidney: ul2. As far as the LGNd laminae are contrastructnral cytochemistry hy a new technique. J. Histochem. Cytochem., 14: 291-302. cerned, the input from the contralateral visual field into area 18 must be largely Graybiel, A. M. 1972 Some ascending connections of the pulvinar and nucleus lateralis posterior of the relayed via lamina C and less via lamina A. thalamus in the cat. Brain Research, 44: 99-125. This is even more valid for area 19, which Guillery, R. W. 1970 The laminar distribution of receives no input from lamina A. Both retinal fibers in the dorsal lateral geniculate nucleus of the cat: a new interpretation. J. Comp. Neur., areas 18 and 19 may of course receive con183: 339-368. tralateral, as well as ipsilateral, visual input Hickey, T. L., and R. W. Guillery 1974 An autoradivia MIN and, to a lesser degree, the parographic study of retinogeniculate pathways in the vocellular laminae. cat and the fox. J. Comp. Neur., 156: 239-254. 3. Since area 18 receives a relatively Hoffmann, K. -P., J. Stone and S. M. Sherman 1972 Relay of receptive-field properties in dorsal lateral small input from lamina A, it might be pergeniculatr nucleus of the cat. J. Neurophysiol.. 35: missible to speculate that the number of Y518-531. type cells in this lamina is small, and that Hollinder, H., M. Wickelmaier, W. und W. Pastor the large majority of Y-cells are to be 1976 Ein Koordinaten und Flachen registrierendes

536

HORSTMAR HOLL4NDER AND IiORAC.10 V4NEGc4S

Mikroskop zur topographischen Analyse von Tcilchengropenverteilnngen. Microscopica Acta, 78: 118-130. ]acobson, S., and J. Q. Trojanourski 1974 The cells of origin of the corpus callosurri in rat, cat and rhesus monkey. Brain Hesearch, 74: 149-135. Jones. E. G.. and A. Y. Leavitt 1974 Retrograde axonal transport and the denionstration of non-specific projections to the cerebral cortex and striatum from thalamic intralaminar nuclei in the rat, cat and monkey. J. Comp. Neur., 154: 349-378. Karnovsky, M. J. 1965 A fornialdehyde-glutaraldehyde fixative of high osmnlarity for use in electronmicroscopy. J. Cell Biol., 27: 137A. I,armle, L. K. 1975 Cell populations of the lateral geniculate nucleus of the cat as determined with horseradish pcroxidasc. Brain Research, 100: 650656. LaVail, J. H., and M.& LaVail I. 1974 ‘The retrograde intraaxonal transport of horseradish peroxidase in the visual system: a light and electron microscopic study. J. Comp. Neur., 157: 303-358. I,aVail, J. H., K. R. Winston and A. Tish 1973 A method based on retrograde intraaxonal transport of protein for identification of cell bodies of origin of axons terminating within the CNS. Brain Research, 58: 470477. LeVay, S., and C. D. Gilbert 1976 Laminar patterns of geniculocortical projection in the cat. Brain Research, 113: 1-19. MaciewicL, R. J. 1975 Thalamic affc:rents to arras 17, 18 and 19 of cat cortex traced with horseradish peroxidase. Brain Research, 84: 308-312. Mason, R. 1975 Cell properties in the medial interlaminar nucleus of the cat’s lateral geniculate complex in relation to thc transient/sustained classification. Exp. Brain Res., 22: 327-329. Nauta, H. J. W.,M. B. Pritz and R. J. Lasek 1974 Afferents to the cat caudoputamen studied with horseradish peroxidase. An evaluation of a retrograde neuroanatomical research method. Brain Research, 67: 219-238. Niimi, K.. M. Kadota arid Y.k t s u s h i t a 1974 Cortical projcctions of the pulvinar nuclear group of the thalamus in the cat. Brain Behav. and Evol., 9: 422457. Niinii, K., and J. M. Sprague 1970 Thalamo-cortical organization of the visual system in the cat. J. Comp. Neur., 138: 219-250.

Otsuka, R., and R. Hassler 1962 Uber Aufbau und Gliederung der corticalen Sehsphare bei der Katze. Arch. Psychiat. Nervkrankh., 20.3: 212-234. S. B. Edwards and I,. A. Palmer 1974 Rosenquist, 14,, An autoradiographic study of the projections of the dorsal lateral geniculate nucleus and the posterior nucleus in the cat. Brain Research, 80: 71-93. Sanderson, K . J. 1971 The projection of the visual ficld to the lateral geniculate and medial interlaminar nuclei in the cat. J. Comp. Neur., 143: 101118. Schubert, P., and H. Holliinder 1975 Methods for the delivery of tracers to the central nervous system. In: The Use of Axonal Transport for Studies of Neuronal Connectivity. W. M. Cowan and hl. CnCnod, eds. Elsevier Scientific Publishing Company, Amsterdam, pp. 115-125. Singer, W.. H. Holliindcr and H. Vanegas 1977 Decreased peroxidase labelling of lateral geniculate neurons following deafferentation. Brain Research, 120: 133-137. Stone, J. 1972 Morphology and physiology of the geniculo-cortical synapse in the cat: the question of parallel input to the striate cortex. Invest. Ophthalmol., 11: 338-346. Stone, J., and 8. Dreher 1973 Projection of X- and I’cells of the cat’s lateral geniculate nucleus to area 17 and 18 of visual cortex. J. Neurophysiol., 36: 551567. Stone, J., and A. M. Hansen 1966 The projection of the cat’s retina on the lateral geniculate nucleus. J. Comp. Neur., 124: 337-352. Stone, I., and K . -P.Hoffmann 1971 Conduction velocity as a parameter in the organization of the afferent relay in the cat’s lateral geniculate nucleus. Brain Research, 32: 454-459. Szentagothai, J. 1972 Synaptology of the visual cortex. In: Handbook of Sensory Physiology. Vol. VW3. R. Jung, ed. Springer, Berlin, pp. 269-324. Vanegas, H., €1. Hollander and H. Distel (1976, in preparation) The early stages of horseradish-peroxidase uptake by nervous structures in the visual cortex of cat. Wilson, M. E., and B. G. Cragg 1967 Projections from the lateral geniculate nucleus in the cat and monkey. J. Anat. (London), 101;677-692. Wilson, P. D., and J. Stone 1975 Evidence of W-cell input to the cat’s visual cortex via the C: laminae of the lateral geniculate. Brain Research, 92: 472-478.