Immediate postoperative retention of visual discriminations following selective cortical lesions in the cat

BehaviouralBrainResearch,

145

17 (1985) 145-162

Elsevier BBR00475

IMMEDIATE POSTOPERATIVE RETENTION OF VISUAL DISCRIMINATIONS FOLLOWING SELECTIVE CORTICAL LESIONS IN THE CAT

JAMES M. S P R A G U E 1, GIOVANNI BERLUCCHI 2 and A N T O N E L L A A N T O N I N I 2

1Department of Anatomy, University of Pennsylvania Medical School, Philadelphia, PA 19104 (U.S.A.) and 2Istituto di Fisiologia Umana, Universit?t di Verona, Strada Le Grazie, 1-37134 Verona (Italy) (Received April 10th, 1985) (Revised version received June 25th, 1985) (Accepted July 1st, 1985)

Key words: visual cortex - striate area - extrastriate area - visual discrimination - cortical lesion - cat

Cats were trained preoperatively for brightness discrimination, and 7 pattern and form discriminations, and then retested for preoperative retention on each discrimination. Cortical lesions were then placed in areas 17 and 18 in one group (4 cats), in areas 17, 18 and 19 in another group (3 cats), and in suprasylvian cortex (areas 7, 21, and parts of 19, 5 and the lateral suprasylvian cortex) in a third group (4 cats). Results are also reported for a fourth group with extensive suprasylvian lesions, to which was added an unintended undercutting of areas 17 and 18 (4 cats). While during original preoperative learning the training continued until a fixed, stringent criterion of performance was attained, both preoperative and postoperative retention was tested in short sessions, involving a limited number of trials and a less stringent statistical criterion (significant run). After extensive removal of areas 17 and 18, all cats behaved as though following the cortical lesion they could immediately recognize the discriminative stimuli as efficiently as before, with no need for retraining. On the contrary, the group with areas 17, 18 and 19 lesions showed a substantial postoperative loss of all discriminations, and especially for the more difficult form discriminations, the reattainment of a significant level of performance was hard or impossible within the allotted number of trials. Also in the group with limited suprasylvian lesions, postoperative retention was generally impaired, but the reacquisition of efficient performance was superior to that of the previous group. Finally, large suprasylvian lesions encroaching on the white matter under areas 17, 18 and 19 proved disruptive for all discriminative capacities, both in retention and in relearning. The excellent retention of all discriminations following areas 17 and 18 lesions once again shows that these areas are by no means essential for complex vision in the cat. In addition, the results strongly indicate that the high-level visual capacities of destriate cats are not due to reorganization of readaptation processes occurring in extrastriate areas after a 17/18 removal. The clear-cut retention deficits which were present in cats with cortical lesions more extensive than areas 17 and 18 or outside of the latter areas prove the essential participation of extrastriate cortical areas in visual discrimination including form. However, the distribution of functions among the various visual cortical areas in visual discrimination remains poorly understood. Attempts at solving this problem will have to consider the variety and the multiplicity of inputs to the array of cortical visual areas as well as the highly interactive nature of the visual system.

INTRODUCTION

The cat's capacity for visual form discrimination is scarcely disrupted by removal of the primary visual cortical area, area 17 or striate a r e a 3'4"6'12'16"17"18"33'34. The effects of this lesion appear to be limited to a moderate reduction in

visual acuity 3,18 and to an impairment in the ability to extract figures from a masking noise (refs. 9, 17, but see ref. 16). By contrast, significant deficits in form vision follow cortical lesions which involve areas 18 and 19 in addition to a r e a 176"1°'~ 1,12,16,20,30,34,41,42, as well as the combined ablation of several cortical areas other than

Correspondence: G. Berlucchi, Istituto di Fisiologia Umana, Strada Le Grazie, 1-37134 Verona, Italy.

146 17 and 185'8'10-12'14"16"34. These behavioral findings must be interpreted in conjunction with the anatomical and physiological evidence on the organization of the neural pathways which carry visual information from the retina to the cortex. The retinal outputs include at least 3 populations of fibers, corresponding to the X-, Y- and Wganglion cell types. These cell types differ in size (and consequently in conduction velocity), as well as in response properties. Among the large-sized Y-cells and the medium-sized X-cells the former show a better temporal resolution of the visual stimulus while the latter show a better spatial resolution. The small-sized W-cells constitute a heterogeneous population with poor contrast sensitivity, poor spatial and temporal resolution, and, at least in some cases, good motion sensitivity (for reviews see refs. 25, 36). These retinal projections form several parallel pathways which convey visual information to avast expanse of the cerebral cortex, including the striate as well as extrastriate areas, via relays in the dorsal and ventral laminae and the medial interlaminar nucleus of the lateral geniculate complex, the superior colliculus, the pretectum and parts of the pulvinar complex (for reviews see refs. 25, 33, 36). An important property of these multiple pathways to the cortex is that the lateral geniculate cells which receive an X-retinal input project exclusively to area 17, whereas geniculate and extrageniculate projections conveying Y- and Winputs to the cortex terminate in virtually all visual cortical areas, both striate and extrastriate. Thus area 17 is presumably indispensable for passing information transmitted along the X-retinal channel to other visual cortical areas, whereas the Yand W-inputs can reach several visual cortical areas in addition to area 17 independent of the latter a r e a 25'33"36.

Taken together with the behavioral effects of selective cortical lesions this anatomical and physiological evidence suggests that in the cat areas 17 and 18 as well as the X-input from the retina are largely dispensable for form vision so long as the dimensions of the stimuli are well above acuity threshold, and therefore this function must rely on extrastriate cortical substrates coupled with the Y- and W-inputs. The

role of area 17 and the X-input in vision is still incompletely understood, but it seems clear that they are important chiefly for increasing the spatial resolution of the system in terms of both acuity and signal-to-noise sensitivity3,4,9,16-1s,25. However, before accepting the hypothesis that even in the normal cat the basic substrates for form vision lie outside area 17 and do not require the X-input, alternative possibilities qaust be ruled out. It is often suggested, for example, that cats with 17/18 lesions may solve form discrimination problems on the basis of visual cues other than form itself, such as luminance differences or local flux changes. This suggestion is hard to test directly, especially because such cues may normally contribute to form vision even in cats with an intact brain 41, and is unlikely for a number of reasons. The most important arguments against it are the demonstration by careful control of stimulus parameters that normal cats and cats without area 17 are guided by the same cues during form discrimination3; and the evidence that cats lacking area 17 exhibit complex capacities for perceptual grouping and figural synthesis which are considered fundamental to form vision ~6. Another possibility is that the capacity for form vision which survives removal of area 17 is due to a reorganization of the visual system, such that extrastriate areas may acquire functions which they do not normally possess in the intact brain. Spear and collaborators 2'27,32 have indeed contrasted the small deficits in visually guided behavior which follow the removal of the visual lateral suprasylvian areas in an otherwise intact brain with the prominent losses in visual functions which are caused by this same lesion when performed after an ablation of areas 17, 18 and 19. This difference is attributed to a major reorganization of the visual system, involving a radical change in the functional significance of the lateral suprasylvian areas, following a 17/18/19 removal 27.32" However, these same workers have been unable to find a direct physiological indication of a change in neuronal function in the lateral suprasylvian areas following ablation of areas 17, 18 and 19 which could account for the behavioral phenomena 28'29. It is only after neonatal removal

147 of areas 17, 18 and 19 that there is some electrophysiological sign of reorganization in the lateral suprasylvian areas 3s. On the other hand, in our previous work 16"34as well as in that of others ~8, there is circumstantial evidence that visual cortical areas beyond areas 17 and 18 can sustain visual discrimination of complex spatial stimuli in their own right, independent of a reorganization process. Cats submitted to a bilateral removal of areas 17 and 18 appear capable of remembering visual pattern discriminations learned before the operation without requiring any retraining, provided high acuity is not necessary for performing the task. To the extent that the reorganization hypothesis assumes that the functional changes in extrastriate areas following removal of the striate area are guided by visual experience and training, the t'mding of good immediate postoperative retention of form discrimination in cats lacking areas 17 and 18 appears to contradict it. In the present paper we report the result of a more systematic study of immediate postoperative retention of brightness and pattern and form discriminations in cats with different types of visual cortical lesions. Retention was assessed in brief tests shortly after the operation, such as to render unlikely the possibility of a reorganization of the visual system and/or the acquisition of new strategies for visual discrimination due to postoperative experience and training. We have compared the effects of a 17/18 lesion with those of more extensive lesions, involving areas 17, 18 and 19, the effects of lesions limited to the extrastriate areas in the middle suprasylvian gyri, and those in animals with suprasylvian lesions which inadvertently invaded the radiations innervating areas 17, 18 and 19. A partial summary of these results has appeared in two previous reviews 4"33. MATERIALS AND METHODS

Surgery and histology All surgical procedures were carried out under Nembutal anesthesia (35-40 mg/kg i.p.) and with aseptic precautions. A Zeiss stereomicroscope was used for the finer steps in the surgery. The original experimental plan contemplated 3 types of cortical lesions: (1) a bilateral lesion destroying

areas 17 and 18 in the lateral, postlateral and splenial gyri; (2)a similar bilateral lesion involving, in addition to the two above mentioned areas, area 19 in the lateral gyri and sulci and in the caudal portion of the posterior suprasylvian gyrus; and (3) a bilateral lesion of the crown of the anterior and middle suprasylvian gyrus (areas 7, 21 and parts of 19 and 5, largely sparing areas AMLS and PMLS in the suprasylvian sulcus; see ref. 40). The first two lesions were made by subpial aspiration of the grey matter under microscopic control. The latter lesion was made either by subpial aspiration or by stripping the pia off the cortical surface. While the fwst procedure is liable to produce a direct injury of the white matter under the site of aspiration, especially in the depth of the sulci, thereby at least partially interrupting the afferent and efferent projections of areas 17, 18 and possibly the lateral suprasylvian areas, the stripping procedure causes degeneration of the cortical grey matter deprived of its blood supply in the pia, and largely spares the connections of other areas lying outside the lesion. It is however, difficult to delimit sulcal lesions by this method. After sacrifice and reconstruction of the lesions the experimental animals were divided in 4 groups, based on the extent and the location of their cortical lesions, and taking into account any unintended involvement of the subcortical white matter, as well as the resulting thalamic retrograde degeneration. The 4 groups were: (1) 4 cats with a bilateral 17/18 lesion (cats 22, 23, 24, 28); (2) 3 cats with a bilateral 17/18/19 lesion (cats 1, 9, 27); (3)4 cats with a bilateral suprasylvian lesion (areas 7, 21 and part of 19), with variable degrees of involvement of AMLS and PMLS (cats 17, 18, 19, 25); and (4) 4 cats with a bilateral suprasylvian lesion complicated by various degrees of invasion of the white matter underlying areas 17, 18 and 19 plus AMLS and PMLS (cats 4, 7, 12, 29). After completion of the behavioral experiment all cats were deeply anesthetized with an overdose of Nembutal. Each animal was exsanguinated by perfusion through the heart with physiological saline followed by 10Yo formolsaline. The brain was exposed, hardened in situ in formol-saline and blocked stereotaxically. After

148 removal from the skull, it was drawn in several views, embedded in celloidin and cut in frontal sections 40/~m thick. The sections were alternately stained with cresyl violet and the Mahon or Woelcke method for cells and fibers respectively. Selected sections through the cortex and thalamus were drawn using an overhead projector, and the cortical lesions were reconstructed using cytoand myeloarchitectural criteria 19'23 to determine the borders of cortical areas, and by evaluating the extent of retrograde atrophy in dorsolateral geniculate nucleus (LGNd) and pulvinar-lateral posterior complex (for details of the method see refs. 16, 34). In analyzing the extent of the lesion and of the thalamic degeneration, we have attempted to relate the topography of the reconstructed lesion to the retinotopic organization of cortex and geniculate determined by physiological mapping (see refs. 22, 40) as indicated in Tables I, II and III.

Training apparatus and testing procedures The discrimination apparatus has been described in detail elsewhere34. Briefly, the cat, food deprived for 23 h, was trained to choose between two discriminanda placed on two side-by-side translucent panels illuminated from behind. The cat had to push open the door with the positive stimulus thus obtaining a food reward (a small piece of kidney). The panel with the negative stimulus was locked and colliding against it was the punishment for a wrong choice, although a correction procedure was used throughout. Cats were trained in daily sessions of 40 trials, on half

of which the positive stimulus was presented on the right door, while on the other half it was presented on the left door. The left/right position of the stimulus was alternated in a quasi-random order according to a modified Gellermann sequencel3.

All animals were trained binocularly as normals on a simple light-dark discrimination followed by 7 pattern discriminations presented in the order shown in Fig. 1. In the flux or brightness test, the dark stimulus (0.3 cd/m z) was chosen as positive (the mean luminance of the lighted panel was 39 cd/m2). On each discrimination the animal had to attain a learning criterion consisting of 90~o or more correct responses on two successive days. The figure-ground reversals used in the triangle and the cross-circle tasks were intended to control for use of local flux cues in making the discriminations between the two forms (see ref. 41). After learning the 8 discriminations, each cat underwent a preoperative retention test on all problems. These were presented in a random order which varied from cat to cat and was different from the sequence used during learning. Testing on each discrimination was continued until the cat performed a series of correct responses, allowing for at most one error, with a chance probability of occurrence, within the cumulative number of trials run on the retention test of that discrimination, equal to or lower than 0.01 ('significant run': see refs. 21, 37). After performing a significant run on one discrimination, the cat was immediately confronted with the next discrimination until the whole series of pr0blems was

IiI,•

m m

III --

m

Fig. 1. Pairs of stimuli used for discrimination. The 8 discriminations were learned successively from 1 to 8. In each pair, except the dark-light discrimination, the positive stimulus is on the left. In the dark-light discrimination the luminance of the dark stimulus was 0.3 cd/m 2, and that of the light stimulus 39 cd/m 2. In the pattern discrimination the size of the i~eal stimuli was about 10 times that in this figure.

149 completed. Since the earliest possible significant run with a 0.01 chance probability of occurrence is 7 correct responses in the first 7 trials, a cat making no errors in the retention test could complete all discriminations in 56 trials, and these were run in a single session. When errors occurred, so that more trials were required for performing a significant run, the test was broken into two or more daily sessions of about 50 trials. Following completion of the preoperative retention test, the cats were operated on and allowed a twoweek recovery period in their cage. The postoperative retention test followed the same paradigm as used for testing preoperative retention. The 8 discriminations were again presented in a random order, different from that used in the preoperative retention test, and the animal was required to perform a significant run on each discrimination within 200 trials before moving to the next discrimination. If a significant run could not be performed within 200 trials (the significant run for this number of trial is 17 correct responses out of 18 consecutive trials), testing on that discrimination was discontinued and the next discrimination was presented. This upper limit for postoperative retention testing was adopted because during initial learning and, afortiori, during the preoperative retention test, the first significant run was usually performed well within this limit. Non-parametric statistical analysis was performed on the total number of errors required for reaching the final criterion of learning during preoperative acquisition, as well as on the number of trials preceding the first significant run, both during original learning and during preoperative and postoperative retention tests. The best possible score was 0 (7 correct responses in the 7 initial trials or 10 correct responses in the first 11 trials), and the worst possible score was 200trials (failure to perform a significant run). The Kruskal-Wallis analysis of variance and the Walsh tests were used according to Siegel26. RESULTS

Anatomy Individual information on all cats in the group with the 17/18 lesion, in the group with a 17/18/19

lesion and in the group with a suprasylvian lesion is provided in Tables I, II and III respectively. These tables indicate the magnitude of damage to each of several cortical areas with the estimated relation to the visuotopic map, as well as the extent of retrograde degeneration in various thalamic nuclei. All the cats in the fourth group, not described in the tables, had an extensive removal of areas 19, 21, 5-7 and LSA, with an additional severe undercutting of areas 17 and 18 on both sides. They also showed widespread retrograde degeneration in all laminae in the L G N and in the pulvinar. The location and the extent of the lesions in one representative cat for each of the first 3 groups are reconstructed and plotted in surface view in Figs. 2-4. Selected coronal sections through the cortex of each of these 3 cats are shown in Figs. 5-7. Behavior Table IV shows the group means (with ranges) for errors committed before reaching the final criterion of preoperative learning on each discrimination. Although there was a considerable individual variability in learning rate, the groups did not differ significantly from one another on any problem (Kruskal-Wallis analysis of variance, P > 0.10 in all cases). The average learning

ED C

Pss ~

~

CAT 2 8

B

A

~

PLS

SSS

Fig. 2. Surface views of the lesion in the brain of one cat in the group with a lesion limited to areas 17 and 18 (cat 28). Letters refer to selected frontal sections shown in Fig. 5. Abbreviations: LS, lateral sulcus; PLS, posterolateral sulcus; PSSS, posterior suprasylvian sulcus; SS, splenial sulcus.

T24

T22

Cat

19

0

0

18

R e m o v a l subtotal (xx), V M to 1 ° to 2 ° on L, V M to 2 ° to 6 ° on R; on R only, U V F r e m o v e d + 5 ° to 40 °, 5 ° to 40 ° azi.

R e m o v a l subtotal (x) s p a r i n g LVF: - 1 ° to - 1 0 ° , 2 ° to 40 ° azi.; sparing U V F : H M to 60 ° , 5 ° to 70 ° azi.

17

R e m o v a l subtotal (xx), sparing LVF: - 30 ° to - 5 0 °, 20 ° to 90 ° azi. on L; sparing U V F : + 5 ° to + 6 0 ° , 2 0 ° to 60 ° azi.

R e m o v a l subtotal ( x - x x ) , sparing LVF: - 20 ° to - 5 0 °, 1 ° - 10 ° to 70 ° azi. on L; sparing U V F : - 3 ° t o + 6 0 ° , 15 ° to 60 ° azi:

nucleus, p a r s lateralis (see ref. 16).

0

0

20

0

0

21

0

0

5-7

0

0

LSA

LVF: A + A 1 = x x x , V M t o 10 ° - 3 0 ° C=xxxVM2 ° -5°, xx 2 ° - 5 ° to 4 0 ° ; C1 + C 2 = 0 ; N I M = 0 UVF: A+A1 =xxxVMto3 ° R, 20 ° L C = x x - x x x V M to 1 ° R, 20 ° L C1 + C 2 = 0 ; N I M = 0 c a u d a l pole intact = + 20: to + 45

LV F: A+A1 =xxx, VMto200 onL, VMto 45 ° on R; C = xxx, V M to 5 ° - 10°; C1 + C 2 = 0 ; N I M = 0 UVF: A+A1 =xxx, VMto25 ° C=xxx, VM to 2 5 ° ; C 1 + C 2 = 0 , NIM=0

LGNd

0

0

PuI-LP

Loss of tissue in various cortical a r e a s a n d retrograde a t r o p h y in dorsolateral geniculate nucleus ( L G N d ) a n d pulvinar-lateral posterior n u c l e a r c o m p l e x (Pulv.-LP) indicated by: mild (x), m o d e r a t e (xx), severe (xxx). L, left side of brain; R, right side; LVF, lower visual field; U V F , u p p e r visual field; Elev., elevation a b o v e or below horizontal meridian ( H M ) ; azi., eccentricity from vertical m e r i d i a n (VM); SS, splenial sulcus; L S A , lateral s u p r a s y l v i a n area, including b o t h b a n k s o f middle a n d posterior suprasylvian sulci (different subdivisions of L S A , i.e. P M L S , A M L S , etc., according to ref. 40). Different l a m i n a e o f L G N d (A,A,C,C1, N ! M see ref. 16). LP1, lateral posterior

Histological findings in cats with a 17/18 lesion

TABLE I

T28

T23

R e m o v a l e x t e n s i v e (xxx), s p a r i n g t i n y i s l a n d in L V F : - 15 ° t o 25 ° , 6 0 ° t o 90 ° azi. o n L (Fig. 5C), a n d in U V F : + 30 ° t o + 4 0 ° , 20 ° azi. o n R (Fig. 5 D )

V M to 60 ° azi. S p a r i n g UVF: +5°~to +60 ° , 10 ° t o 90 ° azi.

20 °, 20 ° t o 90 ° azi.; o n R o n l y - 30 ° t o 60 °,

-

Removal subtotal (x-xx), sparing LVF: - 5 ° to

R e m o v a l p a r t i a l (xx), V M to 2°-10 °

to 2 °-5 ° LVF and UVF, except caudal pole intact: +20 ° to +45 °

R e m o v a l s u b t o t a l (x), V M LVF:

1°-2°,

C1 + C 2 = x - x x o n L ,

0onR,

s p a r i n g + 20 ° t o + 60 ° C = xx-xxx VM to 2 ° NIM=0

A + A1 = xxx, c o m p l e t e R , p a r t i a l L, V M t o 50 ° A1 = x x x , V M t o 5 ° xx, 5 ° t o 4 0 ° R + L C = xxx, V M t o 5 ° , x, 5 ° t o 4 0 ° , R + L C1 + C 2 = x o n L , 0onR N I M = x - x x o n L, 0 o n R UVF: A = xxx complete R + L A1 = x x x , V M t o 20 °

LVF:

-2 ° R,

°

NIM=0

15 ° - 3 0

C=xxx, sameCl +C2=0, caudal pole intact: +20 ° to +45 °

A+A1 =xxx, VMto to3 ° -5 ° L

A+A1 =xxx, VMto C=xxx, VMto5 ° xx, 5 ° t o 4 0 ° CI +C2=0, NIM=0 UVF:

II

27

1

9

Cat

I n t a c t L V F (Fig. 6 A - D ) UVF: xxx on R,xx on L s p a r i n g 5 ° - 2 0 ° azi. o n L (Fig. 6 E - G )

A r e a 17 R e m o v a l e x t e n s i v e (xxx) s p a r i n g i s l a n d in L V F : 35 ° to - 60 °, 10 ° t o

(Fig. 6 F , G ) A r e a 18 R e m o v a l p a r t i a l (xx) sparing LVF: peripheral t o 1 0 ° - 1 5 ° azi. S p a r i n g UVF: +5 ° to +20 ° , 10 ° to 40 ° azi. o n L (Fig. 6 F , G )

45 ° azi. o n L (Fig. 6 A ) a n d o n U V F : + 15 ° to + 5 0 °, 40 ° t o 60 ° azi.

-

R e m o v a l e x t e n s i v e (xxx), sparing island approx. - 2 5 °, V M to 10 ° azi. r e m o v a l in U V F s u b t o t a l (x o n R ) s p a r i n g + 5 ° t o + 3 0 °, V M t o 10 ° azi.; (xx o n L ) s p a r i n g + 10 ° t o + 30 °, V M t o 10 ° azi.

- 35 ° t o 10 ° azi.; + 5 ° to 30 ° azi.

R e m o v a l e x t e n s i v e (xxx), s p a r i n g t i n y i s l a n d 17 U V F : + 2 0 ° to + 4 0 ° , 60 to 90 ° azi.

azi.

sparing LVF: -45 °,VMto sparing UVF: + 10 °, 10 ° t o

Removal extensive (xxx),

R e m o v a l e x t e n s i v e (xxx),

s p a r i n g i s l a n d 17 U V F : + 5 0 ° to + 9 0 ° , H M to 60 ° o n R ; s p a r i n g i s l a n d 18 U V F : + 20 ° to + 3 0 ° , 3 0 ° to 45 °

19

17-18

F o r a b b r e v i a t i o n s see T a b l e I.

Histological findings in cats with a 17/18/19 lesion

TABLE

0

0

0

20

21a = xx 21b = 0

0

0

21

0

0

0

7

0

0

0

LSA

C1 + C 2 = x x NIM = x-xx

C = xxx

=xxx

Pulv. R = x x x L=xx LPI: R = x x x L = xx LVF: A+A1 =xxx;C=xxx C1 + C 2 = xx, N I M = x - x x UVF: onR:A+A1

P + LP1 = x x x sparing caudal pole

P + LPI = xx-xxx

Pul.-LP

L V F: A + A1 = x x x ; C + C1 = x x x NIM = x-xx UVF: A + A1 = xxx, s p a r i n g + 5 ° t o + 20 °, 4 0 ° t o 70 ° azi. o n R ; C + C1 = x x - x x x , N I M = x - x x o n L, 0 o n R

NIM = xxx

L V F + U V F : all l a m i n a e = x x x ,

LGNd

tO

III

17

19

H M to - 2 0 °,

+ 10 ° ( U V F ) , V M t o

+ 2 0 ° to 40 ° azi.

A r e a 18 i n t a c t e x c e p t U V F

+ 2 0 ° t o + 5 0 ° azi.

A r e a 17 i n t a c t e x c e p t U V F 5 ° d a m a g e xxx o n

90 ° azi.

HMto

+20°,20

° to

r e m o v e d (xxx) e x c e p t

R,xx on L; U V F totally

-

L V F intact except H M to

ablatated bilaterally

in L V F f r o m + 5 ° t o

40 °

H M t o - 8 °, V M t o 3 ° (L). U V F = xxx t o t a l l y

A r e a 18 i n t a c t L , b u t o n R

- 25 ° V M to 10 ° ( R ) ;

c o m p l e t e = xxx, incl. A C

moval UVF

s p a r i n g - 5 ° t o 45 o. R e -

5 ° azi. f r o m H M t o 5 °,

R e m o v a l L V F (x) V M t o

90 ° azi. (Fig. 7 D )

+20 ° t o + 3 0 ° , 30 ° t o

7A,B). U V F , sparing

R e m o v a l L V F = xx, H M t o

0

25

-45°; 3~ - 5 ° to 50 ° azi. (Fig.

f i b e r loss in U V F (x)

0

18

L V F , sparing - 2 0 ° to

R e m o v a l s u b t o t a l (xx).

19

A r e a 17 i n t a c t b u t s o m e

17-18

Cat

F o r a b b r e v i a t i o n s see T a b l e I.

Histological findings in cats with suprasylvian lesions

TABLE

O

on R

onL,

2 0 a = xxx

0

0

0

20

on L

o n R, 0

2 1 b = xxx

2 1 a = xxx

21b = 0

2 1 a = xxx

21b = 0

2 1 a = xxx

21b = x

2 1 a = xxx

21

xxx

xxx

xxx

xxx

5-7

R = xx

L = xx

= 0 = x only

= 0 = 0

= 0 = 0 DLS + VLS = 0

PLLS = 0

PMLS

ALLS = 0

AMLS

DLS + VLS = 0

PLLS = 0

PMLS

ALLS = 0

AMLS

DLS + VLS = 0

PLLS = 0

periphery

PMLS

ALLS = 0

AMLS

VLS + DLS = 0

PLLS = 0

PMLS:

R = x

AMLS: L = 0

LSA

+ 2 0 ° to 40 ° NIM = 0

V M t o 10 °

UVF

A,A1,C on L

Laminae = 0 except

NIM = xx-xxx

and AC

d i f f u s e loss in L V F

L a m i n a e o n R = xx

Laminae on L = 0

NIM = xx-xxx

xx o n L

C,C1,C2 = x on R

A + A1 = 0

N I M = xxx

CI + C2 = xx-xxx

A + A I = 0, C = x

LGNd

LPI = xx-xxx

Pul = xx

U V F = xx

LPI:

P u l = xxx

U V F = xxx

LPI:

Pul = xxx

LP1 = xxx

Pul = xxx

Pul-LP

U~

.~.~

=

~

~-

65.2 (46-84)

3.5 (0-8)

3.7 (0-8)

12 (7-17)

47.3 (36-61)

49.7 (46-53)

7.5 (0-15)

2

47.25 (46-109)

1

=,--~

Suprasylvian plus 17/18 (n 4)

(n = 4)

=

17/18 (n : 4) 17/18/19 (n 3) Suprasylvian

Lesion group

Number of problem refers to Fig. 1.

3.7 (0-9)

4.7 (0-t9)

2 (0-16)

2.7 (0-6)

3

CO

__JJ

C7¢~

28.5 (0-99)

46.7 (0-83)

62 (0-167)

20.5 (8-28)

4

Problem

Number of errors to final criterion of learning for each lesion group (means and ranges)

TABLE IV

21.7 (11-37)

36.7 (13-51)

50.3(17-112)

25.2 (12-36)

5

°'~°-

b,~ ~~"

O" ~"

<.

29.2 (10-62)

54 (20-112)

9.6 (0-22)

23.5 (15-35)

6

27.7 (17-44)

74.6 (39-99)

46.6(12-50)

37.7 (20-73)

7

I

~

0

4.5 (0-11)

4.3(0-13)

7.7 (5-12)

8

4~

U,

22 23 24 28

1 9 27

17 18 19 25

4 7 12 29

17/18/19

Suprasylvian alone

Suprasylvian plus 17/18

Cats

17/18

Lesion

37 17 75 9 99 0 136 13

54 0 126 0 81 5 62 35

144 12 92 43 85 7

119 200 94 37

0 16 0 0

67 18 8

0 8 0 24

Pre Post

217 11 41 2 85 0 110 5

L

1

7 8 0 19

7 16 5 3

38 11 24

38 22 7 40

L

0 0 0 7

0 0 0 0

0 1 0

0 0 0 0

33 200 200 9

0 9 17 6

21 134 15

0 9 0 0

Pre Post

2

1 0 28 0

26 0 3 0

27 0 9

0 2 23 25

L

0 0 25 0

36 85 8

22 1 0 0

0 36 0 200 5 200 8* 73

8 19" 4 0

0 3* 0

30* 0 0 0

Pre Post

3

65 47 1 6

0 42* 174 0 99* 200 3 10" 200 126 0 4

77 4 0 87* 230 0 0 0

400 0 200 0 34* 200 49 5 61

1 0 27 1

Pre Post

41 3 70 6 45 15 42 3

L

4

72 19 26 47

127 39 25 93

5

0 0 0 0

0 0 0 0

0 0 0 0

81 200 200 51

0 17 25 8

0 200 0

0 0 0 0

Pre Post

130 0 37 0 36 0

74 48 43 54

L

Problem

0 0 0 0

0 0 0 7

200 200 200 74

12 54 123 142

0 200 3* 200 0 134

0 0 0 0

Pre Post

37 0 22 0 43 40 153 0

250 30 166 7

79 1 46

71 47 14 43

L

6

Number of problem refers to Fig. 1. * indicates cases in which retention was difficult to assess because of rapid original learning.

0 0 0 0

140 41 162 69

168 90 174 78

0 0 0 0

0 0 0 0

161 200 200 84

33 119 40 3

200 200 186

0 0 0 0

Pre Post

81 0 74 0 119 0

77 76 57 131

L

7

2 41 0 0

3 0 0 0

44 0 42

14 19 1 16

L

0 0 0 0

0 0 0 0

0 0 0

0 0 0 0

147 200 200 200

0 0 90 6

122 200 0

0 0 0 0

Pre Post

8

Number of trials necessary to perform the first significant run for each cat on each discrimination, during learning (L), the preoperative retention test (Pre) and the postoperative retention test (Post)

TABLE V

%rl

156 triangles. Generalization also varied considerably from cat to cat in the vertical-vs-horizontal set, so that some cats showed a complete transfer from one problem to the next, whereas other cats showed no transfer. However, cats showing transfer and cats showing no transfer were randomly distributed among the 4 experimental groups (see Table V). These groups were therefore roughly matched for preoperative learning and generalization capacities. On the preoperative retention test all cats showed a generally good retention, as indicated in Table V. For each cat the mean number of trials necessary to perform the first 'significant run' during the retention test was significantly smaller than that necessary to attain the same level of performance during original learning (Walsh test: P < 0.001 in all cases). In a few exceptional cases, i.e. on single problems in some cats, the score for preoperative retention was equal to or worse than that for original learning, owing chiefly to the very rapid initial acquisition of these problems. Thus the apparent lack of savings with respect to original learning in these cases was in fact a 'floor effect'. In a few cases, indicated by asterisks in Table V, the difference in favor of original learning was very strong probably because learning was facilitated by a generalization effect, while retention was not, presumably because of the different order of presentation. The degree of retention did not correlate with either learning rate or (except for the few cases mentioned above) order of presentation of the discriminations during the test. Generalization effects may in some cases have strengthened the performance on the problems which came later in the sequence, but usually there was no systematic difference between the retention of the problems tested rtrst and those tested subsequently in the series. Finally, there was no significant difference between the retention scores of the different experimental groups of animals on all problems (Kruskal-Wallis test, P > 0.10 in all cases). In the postoperative retention test, the performance was evaluated by comparison both with preoperative retention as well as with original learning. The comparison with preoperative retention allowed us to estimate the degree of post-

50°(17°1__..,,4°

I 19

'

18

~

, 18

":~' ~60c('15°)

~

19

~

yDLS~xN C

VM(*2°) ~ " ~

°) (V6°M)

L o "'"" 18 CAT 28

4001+20~)

E

Fig, 5. Selected frontal sections through the lesion of cat 28 (see Fig. 2). Abbreviations: DLS, dorsal lateral suprasylvian area; HM, horizontal meridian; PLLS, posterior lateral lateral suprasylvian area; PMLS, posterior medial lateral suprasylvian area; S S, splenial sulcus; VM, vertical meridian. Degrees outside of brackets refer to azimuth of visual field representation; degrees in brackets refer to elevation ( + , refers to upper visual field; - , to lower visual field).

operative retention deficit, whereas the comparison with the original learning provided a measure of savings compared to initial acquisition. Table V shows that in the group with lesion of areas 17 and 18 (cats 22, 23, 24, 28), the postoperative retention scores were not significantly different from the preoperative retention scores (Walsh test, P > 0.9 in all cats). It should be emphasized that the discrimination tasks are shown in Table V in order of presentation during preoperative learning; the order of presentation during both pre- and postoperative retention tests was different (see Methods). In fact all 4 cats showed a perfect re-

157 ._.... 10° (-35 °)

LS

50o (-I0°) PLS 2* (~) ~ "

PS$$

. :~;:~. . . .

~:J

60°

,t'i ".

-'~

DLS

::;.,:

,.~ss

15° + 450 (-40 °)

g.

g 5o (.I°1 • ..

~

,,~.-~

,

15°

?,

~'~50°('15°)

... ~,":;i;~

-:J'

~..

,c,.--" .18,~./5o* 1-30°) :':"

/[~

•' g

\

,=I.50" 1-3"1

ti:..

+ 1 0 0 (°10°)

/ I

G-__/

50°

7501-50°1

30 ° 0 ° { "8°)

)\--,'-, ,s, (..

,-,,,,.'.-L..

,

,

F

C /

~

1

5

O

1+15o)

, :.:,;.

;jr. '¢.; ...

•

VENT

~

~

, '.;,"

.--.,! :.::.., .

)

A

"-.

~

£AT 27 Fig. 6. Selected frontal sections through the lesion of cat 27 (see Fig. 3). Abbreviations: AC, area centralis; AMLS, anterior medial lateral suprasylvian area; MSSS, middle suprasylvian sulcus; VENT, ventricle. For other abbreviations and conventions see legends to Figs. 2 and 5.

tention of both triangle discriminations and both cross-vs-circle discriminations. On each of the 3 problems in the vertical-vs-horizontal set 3 cats showed perfect retention (cats 22, 24, 28 for the gratings; cats 23, 24 and 28 for the long bars; cats 22, 23, 28 for the short bars). On the light-dark discrimination two cats performed perfectly (cats 22, 24). Clear savings effects from original learning were present in the other cases. As in preoperative retention test, the order of presentation of the discriminations did not appear to affect the performance in the postoperative retention test. In conclusion, following the cortical lesion, cats with 17/18 lesion behaved as though they could recognize the discriminative stimuli as efficiently as before, with no need for retraining. In the group of cats with 17/18/19 lesions (cats 1, 9, 27), the postoperative picture was strikingly

different from that of the above group (Table V). In all cats the postoperative retention scores were significantly worse than the preoperative ones (Walsh test, P < 0.003 in all cats) and in fact cat 9 could not perform a significant run within 200 trials in 5 problems out of 8 (the short bars and both discriminations in the triangle and crosscircle problems). Cat 1 was similarly unsuccessful in 3 out of 8 problems (short bars, the black triangle and the black cross-vs-circle discriminations). On the remaining problems these two cats reattained a significant run within 200 trials, but with no savings relative to the original learning except on the light-dark problem and, for cat 1, the gratings problem also. Although the other cat (27) showed a postoperative retention that was significantly worse than preoperative retention, it performed a significant run within 200 trials on all

158

M~Ai'~~ ~'~ /

~-~o.~ /

//

t8 19

)

~ ,,~-,."-,2" N LS

MSSS.,~"PM '~

30~-5o 2 -.,-"

--~ PLS 19 i i 30°}

CA1 T8~ Fig. 7. Selected frontal sections through the lesion of cat 18 (see Fig. 4). For abbreviations and conventions see legends to Figs. 2 and 5.

problems. It is of interest that cat 27 had a less extensive cortical lesion (see Table II) than the other 3 cats in the group, in that areas 18 and 19 had more intact portions on one or both sides, and a more limited retrograde degeneration in the thalamus. In conclusion, this group showed a substantial postoperative loss of all discriminations, and relearning of the triangle and crosscircle discriminations was strongly impaired or impossible within the allotted number of trials. In the group with suprasylvian lesions (cats 17, 18, 19, 25) there was also a deficit in postoperative retention compared to preoperative retention (see Table V). In cat l9, postoperative retention scores were significantly worse than the preoperative ones (Walsh test, P < 0.06). A significant difference in the same direction was evident

in cat 25 after exclusion of the dark-light problem. In the other two cats (17, 18) there was a similar tendency toward a postoperative loss, although it did not reach significance. A parametric statistical test indicated that as a group these cats showed a postoperative retention deficit (P < 0.02 in a t-test for correlated scores considering the number of trials preceding the significant run in the preoperative and postoperative retention tests). On the other hand, with the exception of cat 17 the postoperative retention scores of these cats did not differ significantly from their performances during original learning. The exceptionally good postoperative retention exhibited by cat 17 can be attributed to the fact that it had a lesser involvement of area 19 and sparing of NIM in the lateral geniculate and more of P-LP (Table III). In conclusion, in this group with suprasylvian lesions, similar to the group with 17/18/19 lesions, postoperative retention was generally impaired. However, the deficit was smaller than in the latter group, and at least for the triangle and cross-circle discriminations, relearning was more efficient in the suprasylvian group than in the 17/18/19 group. In the group with suprasylvian lesions associated with various degrees of invasion of radiations underlying areas 17, 18 and 19 (cats 4, 7, 12, 29), there was a dramatic discrimination loss as well as an inability or a strongly reduced capacity to relearn all discriminations (see Table V). Cat 7 could not perform a significant run within 200 trials on any discrimination, and cat 12 was able to perform a significant run only on the light-dark discrimination, and within a number of trials very close to that required for original learning. Postoperatively, cats 4 and 29 could achieve a significant level of performance within 200 trials in 7 problems out of 8, but their postoperative retention was significantly worse than the preoperative one (Walsh test, P < 0.008 in both cases), and for cat 4 relearning took significantly longer than original learning (Walsh test, P = 0.008). Cats 7 and 12 had substantially greater undercutting of areas 17, 18 and 19 compared to cats 29 and 4 which had a larger number of intact neurons in NIM. In conclusion, a large cortical lesion involving suprasylvian areas plus

159 the undercutting of areas 17 and 18 was associated with a severe disruption of discriminative capacities both in retention and in relearning. It was more incapacitating than the other lesions, and the degree of impairment resulting from it was approached only by that seen after 17/18/19 lesions. It must, however, be stressed that these cats could use vision to avoid obstacles in their environments, and performed successfully in perimetry tests. Therefore they were not blind, although they must have been insensitive to the visual cues which normally guide performance in the discrimination apparatus, at least within the limits of our retraining procedure. DISCUSSION While we are fully convinced that attempts at defining the functional role of any given cortical area in behavior must include an analysis of the behavioral deficits which follow its selective removal4, we are also aware of the difficulties of the interpretation of the effects of brain damage. When examining the results of tests of retention of preoperatively learned visual discriminations following a cortical lesion, one must consider that failure on these tests may be due to a multiplicity of unspecific factors, such as changes in motivation, attention and emotion, rather than reflecting a specific loss of the discriminative capabilities and/or the memory required for performing the task. By contrast, it was underlined in the Introduction that a successful postoperative retention test after cortical damage is more revealing for neurological interpretation, especially when retraining is not required for good performance, since such a result must obviously mean that the ablated cortical area is not necessary for the discriminative ability under examination. It seems therefore very important to assess the presence or absence of postoperative retention with tests which are sensitive enough for distinguishing between a real inability to discriminate from other forms of inadequacy of performance. This problem has been tackled before by several experimenters and particularly by Thompson 37 who has illustrated the advantages of the 'significant run' analysis over traditional assessments of learning and memory which employ fixed stringent criteria.

The present data have been analyzed with this method and we feel that the results support Thompson's approach. The most important aspect of the results is the contrast between the behavior of the cats with a 17/18 lesion, which at the resumption of testing very shortly after the operation showed a perfect retention of all visual discriminations, and that of the other lesion groups, all of which exhibited various degrees of retention deficits. Thus a highlevel immediate postoperative retention of visual discriminations such as those used here requires the integrity of cortical substrates which are different from, or more extensive than, areas 17 and 18. These data are fully consistent with the idea that the cat's capacities for visual discrimination, including that of patterns and form, are subserved chiefly by extrastriate mechanisms 3,12.16,25,33,34. An immediate postoperative retention of visual discriminations by cats with a 17/18 lesion has been described in previous papers 16'18"31"34, but there have also been reports of a loss of visual discrimination after a striate lesion, followed by reacquisition upon retraining (see e.g. refs. 10-12). However, in the latter studies the cats with striate lesions were tested after a long postoperative period, which may have allowed the operation of normal forgetting and interference processes, and their performance was analyzed according to stringent criteria which may have assessed the ability for high-level performance rather than the capacity for discrimination. The immediate postoperative recognition of the discriminative stimuli by our cats with a 17/18 lesion can hardly be explained by reorganization or readaptation processes, on account of the short time separation between the operation and the retention test, as well as the absence of any kind of visual training during such brief recovery period. Nor can the successful retention be attributed to the postoperative employment of discriminative strategies based on the utilization of partial flux cues rather than total stimulus configuration, since these cats performed successfully on generalization and figure-ground reversal problems which can only be solved by overall shape recognition. It is also totally unlikely that

160 the successful performance in the retention tests might have been mediated by the parts of area 17 which had been spared by the lesion, because retention in one cat with a sparing of only a tiny portion of area 17 representing the extreme lower visual field was as good as in the other 3 cats with larger sparings. At any rate, in all cats with 17/18 lesions the sparing of area 17 was limited to portions representing the far periphery of the visual fields, and there was no evidence that in the retention tests cats with these lesions employed eye and head postures which could allow the utilization of the visual field regions projecting to the spared portions of area 17. The sparing of area 18 was more extensive, but the representation of the area centralis was removed in all cases, and the efficiency of retention was not related to the amount of cortex spared in this area. It must be emphasized that statements on the part of the visual field surviving the lesion are in the nature of an approximation. Even though the projections of the sections of the brain containing the lesion are carefully adjusted and compared with the figures of the maps of cortex and geniculate, there is no way to calculate the amount of variation of individual cats. It can be assumed that these estimates of sparing of upper and lower visual fields are subject to an error of about 10-15 ~ . Another point of importance in evaluating the functional significance of sparing in different cortical areas is that in the cat removal of all of area 17 and up to 50~o of area 18 (either unilaterally or bilaterally) does not lead to visual field deficit as measured by orienting in the food perimetry test 24"35. Other visual areas containing a representation of the visual field (see ref. 40) mediate the visual orienting seen in destriate cats. Similarly, extensive removals of other cortical areas receiving a geniculate input area 19, or 19 and 21a, or A M L S - P M L S , or 20, or 19, 21, 2 0 - do not lead to a scotoma in orienting behavior. Visual orienting is a basic function served by many areas of the visual cortex (perhaps all), as well as by the superior colliculus, but its presence is not necessarily related to an intact mechanism mediating pattern and form discrimination. Thus, although we have provided details of the retinotopic extent of sparing in lesions of

areas 17, 18 and 19 and of LGNd, the significance of this sparing for the perceptual discrimination process is by no means clear. Once again, the finding of very good visual capacities in destriate cats casts strong doubts on the rationality of many current attempts at interpreting the physiology and the pathology of the visually guided behavior of the cat solely in terms of the neuronal properties of areas 17 and 18 (for a critique of this approach see ref. 4). For example, the visual deficits exhibited by cats which suffered from early visual deprivation or experimental strabismus are not simply due to a disorganization of area 17, as repeatedly suggested in the past, since they are much more severe than those which follow a total removal of this area (see e.g. refs. 15,

18). The slight but significant retention deficits in the group with uncomplicated suprasylvian lesions, involving areas 7, 21 and 19, and to a smaller degree, area 5, and minimal damage to the visual lateral suprasylvian areas, are in agreement with similarly minor effects reported after comparable lesions in studies of both learning and retention 4'8'1°-12"14"16"34. It seems improbable that these effects were due primarily to damage to the lateral suprasylvian areas, whose relation to the visual modality is f'm'nly established on anatomical and physiological grounds 27,4°, since such damage was minimal if not completely absent. Further, Spear et al.32 have claimed that the selective ablation of the visual lateral suprasylvian areas has no affect on retention of various visual discriminations. However, the variability of their data, including the presence of massive retention deficits in some cats with suprasylvian lesions, does not allow det'mite conclusions. Devastating effects on retention were exerted by either the combined removal of areas 17, 18 and 19, or the extensive ablation of suprasylvian areas complicated by an undercutting of areas 17, 18 and parts of 19. In both cases, retention was very poor or absent on most problems, and some relearning was possible only for the brightness and/or gratings discrimination. Although previous work suggests that with a more prolonged retraining the cats with a 17/18/19 lesion might have regained some capacity for form

161 vision2,20,31,42, o u r present results and much other

evidence indicate that the discrimination of visual forms becomes impossible after massive cortical ablations, including the striate cortex as well as the extrastriate a r e a s 5"6"10-12'16"20. In this sense the definition of form vision as a cortical function seems justified, as opposed to the capacity to discriminate flux, brightness and perhaps pattern (gratings), which can be reacquired even after a complete destruction of all cortical substrates for vision'. While strongly supportive of the hypothesis that visual pathways terminating in cortical areas outside of 17/18 (i.e. the Y- and W-inputs) are sufficient by themselves for object vision, the results do not allow the identification of an area or a small group of areas which could be regarded as the basic cortical seat of visual form discrimination. Although it is true that significant deficits in visual discrimination result from lesion of different combinations of visual cortical areas, all of which include area 19 (see present results and refs. 5, 6, 16, 34), the removal of area 19 alone has generally been found ineffective on various visual performances16'17. An inescapable conclusion is that the visual system of the cat has a highly interactive organization in which each cortical area can hardly be considered a separable functional unit. Similarly, the various visual pathways ascending to the cortex through the lateral geniculate, the superior colliculus, the pretectum and the pulvinar are far from constituting parallel corticopetal channels because they converge at many cortical and subcortical stations. Given that very severe impairments in visual form discrimination are caused in the cat by removal of areas 17, 18 and 19, which are major recipients of the lateral geniculate projections, as well as by an extensive lesion of the superior colliculus and p r e t e c t u m 7,39, which are obligatory relays of the tectothalamocortical visual pathways, it does not seem too far-fetched to suggest that a necessary condition for form vision is the arrival at the cortex of a sufficiently strong input both through the lateral geniculate pathway, and through the tectothalamic pathway. A complete or nearly complete removal of any or both of these inputs would be incompatible with high-level processing of visual

information. Further analyses of the division of labor between the X-, Y- and W-inputs, between the striate and extrastriate cortical areas, and between the geniculocortical and tectothalamocortical pathways are necessary for accepting or rejecting this hypothesis. ACKNOWLEDGEMENTS

The experiments were performed at the Institute of Physiology of the University of Pisa and the Institute of Neurophysiology of the CNR, Pisa. The histological preparations and analyses were carried out in the Department of Anatomy of the University of Pennsylvania. The research was supported in part by the U.S. Public Health Service (Grant EY 00571), the Consiglio Nazionale delle Ricerche of Italy and a NATO Grant for Collaborative Research. J.M.S. was a recipient of a Josiah Macy Faculty Award.

REFERENCES 1 Alder, S.W. and Meikle, T.H., Visual discrimination of flux-equated figures by cats with brain lesions, Brain Res., 90 (1975) 23-42. 2 Baumann, T.P. and Spear, P.D., Role of the lateral suprasylvian visual area in behavioral recovery from effects of visual cortex damage in cats, Brain Res., 138 (1977) 445-468. 3 Berkley, M.A. and Sprague, J.M., Striate cortex and visual acuity functions in the cat, J. Comp. Neurol., 187 (1979) 679-702. 4 Berlucchi, G. and Sprague, J.M., The cerebral cortex in visual learning and memory, and in interhemispheric transfer in the cat. In F.O. Schmitt, F.G. Worden, G. Adelman and S.G. Dennis (Eds.), The Organization of the Cerebral Cortex, The MIT Press, Cambridge, MA, 1981, pp. 415-440. 5 Berlucchi, G., Sprague, J.M., Antonini, A. and Simoni, A., Learning and interhemispheric transfer of visual pattern discriminations following unilateral suprasylvian lesions in split-chiasm cats, Exp. Brain Res., 34 (1979) 551-574. 6 Berlucchi, G., Sprague, J.M., Lepore, F. and Mascetti G.G., Effects of lesion of areas 17, 18 and 19 on interocular transfer of pattern discriminations in split-chiasm cats, Exp. Brain Res., 31 (1978) 275-297. 7 Berlucchi, G., Sprague, J.M., Levy, J. and DiBerardino, A., Pretectum and superior colliculus in visually guided behavior and in flux and form discrimination in the cat, J. Comp. Physiol. Psychol., 78 (1972) 123-172.

162 8 Campbell, A. Jr., Deficits in visual learning produced by posterior temporal lesions in cats, J. Comp. Physiol. PsychoL, 92 (1978) 45-57. 9 Cornwell, P., Overman, W. and Ross, C., Lesions of visual cortex impair discrimination of hidden figures in cats, PhysioL Behav., 24 (1980) 533-540. 10 Cornwell, P. and Warren, J.M., Visual discrimination defects in cats with temporal or occipital decortications, J. Comp. Physiol. Psychol., 95 (1981) 603-614. 11 Cornwell, P., Warren, J.M. and Nonneman, J., Marginal and extramarginal cortical lesions and visual discriminations by cats, J. Comp. PhysioL PsychoL, 90 (1976) 986-995. 12 Doty, R.W., Survival of pattern vision after removal of the striate cortex in the adult cat, J. Comp. Neurol., 143 (1971) 341-369. 13 Gellermann, L.W., Chance order of alternating stimuli in visual discrimination experiments, Jr. Genet. Psychol., 42 (1933) 206-207. 14 Hara, K., Cornwell, P.R., Warren, J.M. and Webster, I.H., Posterior extramarginal cortex and visual learning by cats, J. Comp. PhysioL Psychol., 87 (1974) 884-904. 15 Holopigian, K. and Blake, R., Spatial vision in strabismic cats, J. Neurophysiol., 50 (1983) 287-296. 16 Hughes, H.C. and Sprague, J.M., Cortical mechanisms for local and global analysis of visual space in the cat, Exp. Brain Res., in press. 17 Kriager, K., Heitl~inder-Fansa, H., Dinse, H. and Berlucchi, G., Detection performance of cats lacking areas 17 and 18: a behavioral approach to the analysis of pattern-recognition deficits, Exp. Brain Res., in press. 18 Lehmkuhle, S., Kratz, K.E. and Sherman, S.M., Spatial and temporal sensitivity of normal and amblyopic cats, J. Neurophysiol., 48 (1982) 372-387. 19 Otsuka, R. and Hassler, R., l~ber Aufbau und Gliederung der corticalen Sehsphare bei der Katze, Arch. Psychiat. Z. ges. Neurol., 203 (1961) 212-234. 20 Ritchie, G.P., Meyer, P.M. and Meyer, D.R., Residual spatial vision of cats with lesion of the visual cortex, Exp. Neurol., 53 (1976) 227-253. 21 Runnels, L.K., Thompson, R. and Thompson, P., Nearperfect run as a learning criterion, J. Math. Psychol., 5 (1968) 362-368. 22 Sanderson, K.J., The projection of the visual field to the lateral geniculate and medial interlaminar nuclei in the cat, J. Comp. Neurol., 143 (1971) 101-118. 23 Sanides, K.J. and Hoffmann, J., Cyto- and myeloarchitecture of the visual cortex of the cat and the surrounding integration cortices, J. Hirnforsch., 11 (1969) 79-104. 24 Sherman, S.M., Visual field defects in monocularly and binocularly deprived cats, Brain Res., 49 (1973) 25-45. 25 Sherman, S.M., Functional organization of the w-, x-, and y-cell pathways in the cat: a review and hypothesis, Progr. Psychobiol. Physiol. Psychol., 11 (1984). 26 Siegel, S., Nonparametric Statistics for the Behavioral Sciences, McGraw-Hill, New York, 1956. 27 Spear, P.D., Behavioral and neurophysiological consequences of visual cortex damage: mechanisms of re-

covery, Progr. Psychobiol. Physiol. Psychol., 8 (1979) 45-90. 28 Spear, P.D. and Baumann, T.P., Neurophysiological mechanisms of recovery from visual cortex damage in cats: properties oflateral suprasylvian visual area neurons following behavioral recovery, Exp. Brain Res., 35 (1979) 177-192. 29 Spear, P.D. and Baumann, T.P., Effects of visual cortex removal on receptive field properties of neurons in the lateral suprasylvian visual area of the cat, J. NeurophysioL, 42 (1979) 31-56. 30 Spear, P.D. and Braun, J.J., Pattern discrimination following removal of visual neocortex in the cat, Exp. Neurol., 25 (1969) 331-348. 31 Spear, P.D. and Ganz, L., Effects of visual cortex lesion following recovery from monocular deprivation in the cat, Exp. Brain Res., 23 (1975) 181-201. 32 Spear, P.D., Miller, S. and Ohman, L., Effects of lateral suprasylvian visual cortex lesions on visual localization, discriminationand attention in cats, Behav. Brain Res., 10 (1983) 339-359. 33 Sprague, J.M., Hughes, H.C. and Berlucchi, G., Cortical mechanisms in form perception. In O. Pompeiano and C. Ajmone-Marsan (Eds.), Brain Mechanisms of Conscious Experience and Perceptual Awareness, Raven Press, New York, 1981, pp. 107-132. 34 Sprague, J.M., Levy, J., DiBerardino, A. and Berlucchi, G., Visual cortical areas mediating form discriminationin the cat, J. Comp. Neurol., 172 (1977) 441-488. 35 Sprague, J.M. and Meikle, T.H., The role of the superior colliculus in visually guided behavior, Exp. Neurol., 11 (1965) 115-146. 36 Stone, J., Dreher, B. and Leventahl, A., Hierarchical and parallel mechanisms in the organization of the visual cortex, Brain Res. Rev., 1 (1979) 345-394. 37 Thompson, R., Localization ofthe 'visual memory system' in the white rat, J. Comp. Physiol. Psychol. Monogr., 69 (1969) part 2. 38 Tong, L., Kalil, R.E. and Spear, P.D., Critical periods for functional and anatomical compensation in lateral suprasylvian area following removal of visual cortex in cats, J. Neurophysiol., 52 (1984) 941-960. 39 Tunkl, J.E. and Berkley, M.A., The role of the superior colliculus in vision: visual form discrimination in cats with superior colliculus ablations, J. Comp. NeuroL, 176 (1977) 575-587. 40 Tusa, R.J., Palmer, L.A. and Rosenquist, A.C., Multiple cortical visual areas; visual field topography in the cat. In C.N. Woolsey (Ed.), Cortical Sensory Organization, 1Iol. 2, Humana Press, New Jersey, 1981. 41 Winans, S.S., Visual cues used by normal and visual decorticate cats to discriminate figures of equal luminous flux, J. Comp. Physiol. Psychol., 74 (1971) 167-178. 42 Wood, C.C., Spear, P.D. and Braun, J.J., Effects of sequential lesions of suprasylvian gyri and visual cortex on pattern discrimination in the eat, Brain Res., 66 (1974) 443 -466.