EXPERIMENTAL
The
NEUROLOGY
Effect
of
(1968)
22,638-660
Visual
Deprivation
on
Perceptual
Behavior LEO Department
of Psychology,
GANZ
AND
Uniewrsity Received
MARGARET of
California
June
FITCH at
l
Rielcvside, C&forma
92502
22, 1968
Monocularly reared cats manifest profound initial deficiencies in perceptualmotor coordination, in depth estimation on a visual cliff, and in the ability to discriminate visual forms differing in orientation when using only their deprived eyes. When the cats were subsequently exposed to a normal visual environment, not all these deficits proved permanent. Only the discrimination of visual form, the interocular transfer of form discrimination from the deprived to the experienced eye, and fine visually-guided spatial adjustments remained permanently absent. The more permanent visual deficiencies cannot be attributed to scotomata nor to losses in visual resolution, since measurements of visual acuity utilizing an optokinetic response elicited by fine black moving lines did not reveal large losses in resolution. The comparison of the nemophysiological and behavioral analysis on the same visually deprived cats leads us to conclude that visual cortex neurons with very selective receptive fields are particularly important in mediating visual form discrimination and fine spatial adjustments. Introduction
The preceding paper (3 j described a number of changes which monocular visual deprivation brought about in the activity of single cells of the cat’s visual cortex. The present paper describes the concomitant behavioral changes. Methods
Cats were visually deprived as described in the preceding paper (3). A larger number of subjects were tested behaviorally than were tested neurophysiologically. A monocular group consisting of 25 cats was reared with the eyelids of one eye sutured together. A second group consisted of five cats that had one eye sutured closed but in which openings of 24 mm developed spontaneously, usually within 2 weeks after suturing. After the period of deprivation (l-6 months), the animals were tested for visual 1 The support and assistance of Austin H. Riesen in many aspects of the work reported here are gratefully acknowledged. This work was supported by research grants NB-04802 and NB-04717 from the National Institute for Neurological Diseases and Blindness, U. S. Department of Health, Education, and Welfare. Dr. Ganz is now in the Department of Psychology, Stanford University. 638
VISUAL
DEPRIVATION
639
acuity, tested twice for depth perception, and some were tested for visual form discrimination. Thereafter, the cats were divided into two subgroups. Neurophysiological measurements were made on the first subgroup of ten (3). The second subgroup of 15 was used to determine to what extent the behavioral deficits found on initial testing were permanent or reversible. Following the initial deprivation period and behavioral testing, the deprived eye of subjects in this second subgroup was kept open for a period of time at least equal to the initial deprivation. In half of these subjects (binocular group), both eyes were left open during this development period. With the other half (reverse suture group), the previously open eye was now sutured closed. In the reverse suture group, the animal was forced to use its deprived eye if any use was to be made of visual information. Following this development period, behavioral tests were reapplied to both the binocular and the reverse suture groups to assess the effect of the delayed visual experience on open field behavior, depth perception, acuity, and visual form discrimination. Viszcal Acuit)~. When a number of parallel lines are moved in an animal’s visual field, an optokinetic response is often obtained, consisting of a movement of the eyes tracking the lines, followed by a rapid, ballistic movement in the opposite direction. If the animal is not otherwise restrained, tracking and return movements of the head and body are also obtained. We were concerned only with the eye movement response. The frequency of eye movement responses gave an estimate of the animal’s responsiveness to moving objects, and of its readiness to fixate and follow a moving array. We also obtained a measure of acuity by varying the thickness of the moving lines, noting the smallest width needed to obtain a threshold response. The procedure has been used to measure both the resolution and the detection acuity of the cat (17). In the present study, a measure of visual acuity was obtained comparable to the detection of a black wire on white surround (4). Four optokinetic drums were used, each containing a different width of line: 143, 9.3, 1.3, and 0 minutes of arc. The lines were strung vertically, at lo-cm intervals, and at a distance of 50 cm from the cat’s eye. The lines appeared against a white background of 12 ml luminance. The zero width stripe refers to a control condition which involves the presentation of a moving drum without lines. Since shields blocked the appearance of anything in the cat’s visual field except the lines, the cat could see nothing in motion during the zero condition. Therefore, the eye movement directions should not be affected by drum rotation in this latter case. The drum was rotated to and fro by hand, at a velocity of l-2 rpm. Every day a subject viewed the lines in each of the four drums. On a single day the subject spent 12 min in each drum. Order of line presentation was
640
GANZ
AND
FITCH
balanced over 8 days of testing. During the drum rotation, the cat was held in a restraining box. Testing was always monocular, one eye kept closed with black adhesive. Electrodes at the outer canthi of the two eyes were used to record changes in electrical potential associated with eye movements. A third electrode at ground potential was inserted in the skin above the eyes and halfway between them. Condenser coupled amplifiers were used. All responses greater than 250 pv were counted. Eye movement records were scored blindly. Depth Perception and Visual Orientation. The following tests were performed monocularly, first with the animal’s deprived eye open and the experienced eye shut, then with the eye patch switched to cover the deprived eye, leaving the experienced eye open. Visual Placing. The cat was held so that his forelimbs were free. It was lowered toward a table surface, preferably a surface with a strong bold pattern such as a checkerboard. A positive response consisted of extending the forelimbs, as if in anticipation of landing on the surface with the forelimbs. This is illustrated in Fig. 1. Visually Guided Placing Response. In order to test the cat’s capacity to guide its limbs with respect to smaller objects in visual space, we tested a few of the subjects on a recently developed modification of placing the response test (5). The cat was lowered toward a serrated edge. The prongs or extensions were 20 cm long and one cat paw wide. The spaces between prongs were twice the width of the prongs. The test consists of lowering the cat with one free forelimb toward the serrated edge. A normal cat will accurately place its paw onto one of the prongs. In the absence of visually guided placing, one expects a 50% proportion of landings onto the prong. Eight trials were administered to each eye separately. Visual Cliff, Spontaneous Responses. The subject was placed on a centerboard with a choice of a short apparent drop-off on one side or a long apparent drop-off on the other (18). The direction of descent was noted. If the animal did not leave the board after 30 set, a new trial was started. After five trials, the apparatus was rotated 180 deg in the testing room and another five trials were administered. The rotation counterbalanced for the possibility that the animal approached or avoided some object in the room other than the two apparent-depth surfaces. Visual Cliff, Forced Responses. Since the spontaneous descents on either side of the cliff were few when the deprived eye was tested, a second test of depth perception was given immediately after the spontaneous test, in which a descent from the centerboard was forced. While the cat was on the centerboard, the experimenter attempted to push it off, first onto the deep side and then onto the shallow side. The animal was held by the hind limbs, the forelimbs resting on the centerboard, and pushed until it descended onto
VISUAL
FIG. 1. Two tests response version of
of
depth
perception.
641
DEPRIVATION
In
the
two
top
photographs
the
forced
the visual cliff test is administered. The upper left photograph shows the cat, with its deprived eye open, resisting being forced onto the shallow side of the visual cliff. On the upper right we see that no resistance is offered to this maneuver when the experienced eye is open. The two bottom photographs were taken during the visual placing test. In the lower left, the deprived eye is open. As the experimenter brings the cat toward the table surface, no placing response is shown. The animal continues to lie flaccidly in the experimenter’s hands. On the lower right, the experienced eye is open. A strong extensor response toward the table edge is demonstrated. Subject 16E: Reared 3.5 months monocularly. The test seen here was administered shortly after the eyelids of the deprived eye were opened.
the glass. The cat’s resistance was estimated on a five-point scale: 0 = no resistance to being pushed off on a particular side of the centerboard; 1 = slight resistance; 2 = moderate resistance ; 3 = strong resistance ; 4 = vigorous resistance. The test was administered twice on each side. Open Field. The animal was placed on the floor with a variety of obstacles scattered about. Observations were then made about the extent and speed of locomotion, the use of visual cues in avoiding obstacles, the position of head and eyes during locomotion, and the presence of emotional behavior such as distress calls.
642
GANZ
AND
FITCH
Visual Discvinaination. Several of thl monocularly reared subjects were trained on a visual discrimination involving a vertical vs. a horizontal line. They were run monocularly: Only the deprived eye was open on the test day 1, only the experienced eye was open on day 2, etc. An eye patch was used to restrict stimulation to one or the other eye. The animals were trained in a Yerkes-Watson box (14). On each trial the vertical and horizontal lines were presented. The animal obtained a food reward if it pushed the door containing the positive stimulus. (The left or right position of the stimuli was sometimes changed after a correct trial.) Food was placed behind both doors, but the door holding the negative stimulus was locked. If the animal went to the incorrect side, it was replaced in the start box and that trial was repeated without changing the position of the stimuli. Food was restricted until the cats reach 80% of their body weight; they were then given 20 trials a day. The stimuli were a vertical and a horizontal black rectangle, 10 X 1.3 cm, against a white background. When the door of the start box opened, the animal was approximately 80 cm from the stimuli. Hence, the stimuli were at least 7 deg X 50 min in visual angle. At the choice point, they were approximately twice this angular size. In several cases, an even simpler task was administered wherein the stimuli consisted of a vertical and a horizontal grid. Each grid was composed of five black rectangles, 10 X 1.3 cm with intervening white spaces of the same dimensions. The subjects were run on visual discrimination training until 85% of their responses were correct or until 2,000 trials had elapsed. Results:
Immediately
following
Deprivation
Field Obserzmtions, Initial Testing. When first allowed to use their deprived eye in an open-field situation, the cats behaved in a manner most simply described as not mediated by visual cues. The more frightened animals did very little locomotion and when placed on a small platform 10 cm above the table, simply remained there for long periods, shivering and vocalizing. Usually they left the platform-sometimes after as long as 20-30 min-by slowly backing off, placing one hind leg back onto the table, then the second, etc. When such an animal was placed on the platform with his experienced eye open, he immediately stepped down, moving forward. Other animals showed less fear when their deprived eye was first opened and they were placed in an open-field situation. These animals bumped into objects. They had no visual appreciation of objects in space, since when they contacted such an object, even very gently, they typically reared back suddenly, displaying a strong startle response. Gross depth judgments were
trISUAL
DEPRIVATION
643
lacking : They were observed to step off steep ledges, fall from heights, etc. With their deprived eye the animals failed to blink or flinch to a rapidly approaching object until it actually touched their skin, hair, or vibrissae. When an object, such as a pencil, touched them, they often turned toward the direction of the object and attacked anything in that direction (box sponge, etc.) which happened to lie at hand. They did not follow objects, unless they were of very high contrast. such as a lighted bulb in an otherwise dark room. The monocularly reared animals failed to transfer their recognition behavior to the deprived eye. For esample, with only their normal eye open the cats typically ran toward the front of their cage when the experimenter opened their cage door and jumped onto his shoulder. With only their deprived eye open, they ran to the back of the cage and stayed there. .As another example, on the centerboard of the visual cliti, the cats, when using their experienced eye, usually positioned themselves on the end of the board closest to the experimenter, always minimizing the distance between the cat and the experimenter. They were never observed to do this when using their deprived eye. T,‘is~al Placing Rrspomr. \Vhen a cat with visuo-motor experience was brought toward a surface, it almost invariably extended its forelimbs. This responseis shown in the lower right quadrant of Fig. 1. The animal was being tested monocularly, using its experienced eve. When the same animal was tested with its deprived eye, for the first time (lower left quadrant of Fig. I), it lay relaxed. even limp, in the experimenter’s hands making no response until its vibrissae touched the edge of the table. The tactual stimulation then elicited a prompt and strong extensor response toward the table edge. The results for the initial tests were as follows. Of the 25 cats receiving monocular deprivation, 24 showed no visual placing response with their deprived eye. All showed a strong positive response when tested with their experiencd eye. None of the three darkreared kittens gave a visual placing response. The results of 19 cats that were tested at least twice are shown in Table 1. Visual Clifl, Spontaneous Response. In Table 2. the scoresrepresent the mean number of times the cats descendedspontaneously onto the visually “near” side, visually “far” side, and the number of times the animal did not step off the centerboard durin g the 3%set test. Turning again to the results of the first test administered 23 hours after the eyelids of the deprived eye were opened, we found that when the animals used their dc~rivctl ryr, they showed no such preference. They also gave more “neither” responses.Tn other words, they tended to stay on the centerboard of the visual cliff tnore frequently when only their deprived eye was open. This is probably related to the open-field observation that when they were
644
GANZ
AND
FIl’CH
TABLE VISUAL
PLACING
RESPONSE.
POSITIVE
Monocular depriv. 4 and
8 Weeks
Sample size
8
Partial closure
5
a Second testing during the interval.
was
ANALYSIS.
NUMBER
SHOWIKG
Deprived
2 weeks
Experienced eye
eye
First Second First Second First Second administered
A
RESPONSE
Test a
6
12 Weeks
1
RECOVERY
0 5 1 6 2 5 after
the
first,
6 6 8 8 5 5 both
eyes
being
open
placed on a slight elevation, with only their deprived eye open, many did not get off for a long time. Although the partial closure group did considerably better than the others in giving a visual placing response, they were about equally poor, on the visual cliff. Visual Cliff, Forced Response. In this test the cat was grasped by the hind limbs and pushed toward one or the other side of the cliff (Fig. 1, upper right). When the animal used its experienced eye, it could be pushed toward the visually “near” side of the visual cliff without offering any resistance. The animal resisted vigorously being pushed toward the visually “far” side. When viewing with its deprived eye, the animal often showed the same vigorous resistance to both the visually “near” side and visually “far” side (Fig. 1, upper left). Other animals viewing with their deprived eye showed no resistance to being pushed on either “near” or “far” side. The clearest difference between behavior when the experienced eye was open and when the deprived eye was open was therefore the diferencc in resistance to pushing the animals toward the “near” or “far” side. Table 3 illustrates that with the deprived eye, all groups showed just about the same resistance to being pushed toward the visually “near” side as toward being pushed toward the visually “far” side. All animals in all groups showed clear differences between “near” and “far” when viewing with their experienced eye. Optokinetic Response. When the eyes responded to a series of moving stripes, they typically showed a slow tracking movement in the same direction as the stripes succeededby a quick saccadein a direction counter to the moving stripes. The saccadeyields a larger response in a condensercoupled amplification system, such as the one used here, than the slow tracking movement. The results described below involve an analysis of the saccades.The optokinetic index we used was similar to the one applied
5
Partial closure
First Second
First Second
First Second
Test
was administered
8
12 Weeks
testing
6
4 and 8 Weeks
a Second
Sample size
Monocular depriv. period
u
2 weeks
RECOVERY
after
1.00 5.60
1.63 1.63
1.00 3.67
Near
both
1.20 1.00
2
open
-
during
> .lO < .0.5
> .lO > .lO
> .lO > .lO
Near vs. Far
t test
SPOKTANEOUS
eyes being
7.80 3.40
6.38 7.00
8.17 4.33
0.83 2.00 2.00 1.38
Neither
Responses
eye
CLIFF:
TABLE Deprived
VISUAL
Far
No.
the first,
Mean
ANALI’SIS.
Mean
the interval.
3.00 5.20
4.75 3.38
4.33 6.50
Near
RESPONSES
Far
0 0
1.25 1.00
eye
7.00 3.80
5.13 5.63
5.17 2.33
Neither
Responses
0.50 1.67
No.
Experienced
< .Ol < .Ol
< .Ol < .Ol
< .Ol < .Ol
t test Near JS. Far
E L? 2 > 2 2
5 z $
646
GANZ
AND
TABLE RECOVERY ANALYSIS.
Monocular denriv. 4 and 8 Weeks 12 Weeks Partial Partial closure
Sample size 6 8 5
VISUAL
CLIFF
FITCH
3
: FQKCED RESPONSES.
MEAN
RESISTANCE
Experienced
Deprived eye Test u
Near
Far
First Second First Second
2.67 0.67 1.13 2.00
2.67 2.33 1.13 3.63
First Second
0.80 0.80
1.20 3.60
a Second testing was administered during the interval.
INDEX
eye
t
Near
Far
p>.10 p < .025 P > JO p < .05
0 0.17 0.75 0.13
2.67 3.50 2.50 3.63
P< P< p< P<
P > JO p < .Ol
0 0
3.00 3.40
P < .a P < .Ol
t
.Ol .Ol .025 .01
2 weeks after the first, both eyes being open
by Reichardt (12) to analyze the visual system of the beetle chlorophanus: E-U/E+U ; E stands for the number of saccadesin the expected direction (expected relative to the direction of motion of the stripes) and U stands for the number of saccades in the opposite direction. The index has a maximum value of 1.0 when all tracking eye movements follow the lnoving lines. It ditninishes toward is made smaller and smaller.
zero as the width
of the rotating
lines
The optokinetic index for eleven monocularly reared cats is shown in Fig. 2. Looking first at the results with a drum without lines (zero minutes of arc condition), one seesthat the index is approximately zero in every case.This means that the shielding procedures-used to prevent the animal from seeing moving aspectsof the drum other than the lines themselveswere successful. The index diminishes as the lines are made thinner. When these indices were subjected to an analysis of variance the line width variable was found to account for a significant amount of the variance (Table 4). The monotonic relationship between line width and the optokinetic index is necessary if the index is to be used as a valid indicator of visual acuity. Turning to the effects of the experimental treatment, we see (Fig. 2) that for all line widths (excepting only the zero condition), the index was higher when the cat used his experienced eye than when he used his naive eye. For the three line widths and the 11 animals, the experienced eye gave the higher index in 32 of the 33 possible instances (p = 2-s2). Clearly there was some effect of visual deprivation which made the animals less prone to follow moving objects with their visually deprived eyes, whatever the width of the lines. The mean optokinetic indices are shoCn in Table 5. The difference between the mean indices was .I58 for the 143 minutes of arc line, and .161 for the 9.3 minutes of arc line. The 9.3
VISUAL
4_
O\\ \ 0.50o
647
DEPRIVATION
-2
‘1 ;,:r\;:
!50 . . . . . . . . . . . . ..:.\ . .o \[emailprotected]
!5.b
..__..
..o.. !&. @
5
0.50-
IO
IO
9
X
g -
o
0.50-
;s
.e.. --\ 13r”ib--. . . . . ..
a.0 -- l.;& ”
IO
14
y:.:.y.
..~. 2:
0, ‘--.
.
a O
24
0.50-, \
‘\
24 Eye
‘\ +--+
\ \
--QY,,
-
.8.. !4.d. . . . \ . . . ,..lJ.. 143
9.3
1.3
143 LINE
9.3
1.3
Deprive
* - -0 Experienced 143
9.3
1.3
WIDTH
2. Visual acuity in the monocularly-reared cat as measured by the optokinetic response. Each frame represents the results of a single animal. The number in the lower left of the frame identifies the subject. The number in the upper right refers to the period of monocular visual pattern deprivation in weeks. The abscissa depicts the various line widths employed (143, 9.3, 1.3, 0) in minutes of arc. The ordinate represents the optokinetic index : (E-u U) / (E -1-U). E Stands for the number of saccades in the expected direction (with respect to drum rotation). U Stands for the number of saccades counter to the expected direction. An optokinetic index of zero (horizontal dotted line) is expected when the moving lines cannot be resolved by the subject. Each plotted point represents data from sessions administerctl over 8 successive days. FIG.
648
GANZ
.4ND
FITCH
TABLE OPTOKINETIC
Source Experience Line width E X L Interact. Subtotal Within
groups
Total
RESPONSE
1
IKDES.
Sum of squares
AKALYSIS
OF VARIANCE
Meansquare
df
F Ratio
P
210,994
1
210,994
6.03
< .lO
1,429,487
3
476,496
13.623
< .Oj
104,931
3
34,977
4.77
< .Ol
1,745,412
7
586,589
80
2,332,001
87
7,332
minutes of arc line is clearly suprathreshold for the cats, even when viewing with their deprived eyes. But the 143 minutes of arc line, which was 15 times as wide as the suprathreshold line, still failed to evoke as vigorous a following response in the deprived as in the experienced eyes. This diminution in tracking is therefore independent of acuity. In other words, there appeared to be a deficiency in following moving objects which was independent of the ability to detect them as such. The second aspect of the results with the optokinetic drum relates to acuity. With the 1.3 minutes of arc line, 9 of 11 cats gave indices above zero when viewing with the experienced eye ; 9 out of 11 gave zero indices when using their deprived eye with that same line width. They responded as if they could not see the 1.3 minutes of arc with their deprived eye ; they apparently could see the 1.3 minutes of arc line with their experienced eye. In fact, the only two animals which gave any indications of responding TABLE
5
OPTOKINETIC STNPE
WIDTH
(IN
INDES MINUTES
OF AKC)
143
9.3
1.3
Deprived eye mean SD t test: line vs. 0
.259 .113 p < .OOl
.071 .062 p < .005
.005 .019 jJ > 0.100
.015 .036
Experienced eye mean SD t test: line vs. 0
.417 .158 p < .OOl
.084 .071 p < .OlO
.008 .017
p<
,233 .096 .OOl
Experienced-deprived mean SD t test: E vs. D
.158 .lOO p < ,001
.I61 ,071 p < ,031
.079 ,087 p < ,010
.003 .057 p > 0.500
VISUAL
649
DEPRIVATION
at all to the 1.3 minutes of arc line with their deprived eye were cats which had sustained the shortest deprivation period, i.e., 4 weeks. We obtained some measure of acuity by comparing for each animal and each eye, the difference between the optokinetic index obtained to a particular line width with the index obtained to the blank drum. A t test was applied to establish the statistical significance of these differences (Table 5). The test established that the kittens could see all the line widths used here with their experienced eye. With their deprived eye, in contrast, they could only seethe 143 and 9.3 line widths. /‘is& Discvi~&~~tion. In this test the animals were required to learn a visual discrimination between two large rectangular stimuli differing in orientation by 90 deg. The results are shown in Fig. 3. The animals were run monocularly with their experienced eye and their deprived eye on alternate days. Each symbol in Fig. 3 represents the proportion of correct responsesfor a group of 100 trials. Since 20 trials were administered during a session, each symbol represent data pooled from five sessions.The results show quite clearly that the cats, when using their experienced eye, learned the discrimination within a few hundred trials. At the same time, they failed to learn the discrimination with their deprived eye. The subjects were still discriminating at no better than chance levels after 2,000 trials had elapsed. Results:
Development
of Visual
Function
following
Deprivation
Dcalelopment of Open Field Behavior. The monocularly reared animals were usually tested a second time, 1-l days after the deprived eye was opened. Both eyes were left open during the interval between the two tests. Little improvement in open-field behavior was noticed on the second test. Some of the animals were observed in an open-field situation 3, 4, and 5 months after the opening of the deprived eye. Gradually, the more active subjects came to show open-field behavior with the deprived eye which resembledthat of a normal animal : They came to use visual cues to explore the open-field, they walked about in the presence of obstacleswithout bumping into them, they stepped off platforms quickly. Yet, though the animals showed considerable improvement, some deficiencies remained. When the cats were using only their deprived eye they would approach an object until 5-8 cm away, slow down, and sometimes jump back as if startled when their vibrissae actually made contact with the object. Secondly, they never came to show the signs of visual recognition toward the experimenter, described earlier, that they showed with their normal eye. In some of the subjects, the reverse suture group, the experienced eye was sutured closed. The more active of these subjects came to show precise
650
GANZ
AND
0 Control
FITCH
o Control
(L)
(RI
‘u0 0
4e
I t-
I
I
I
I
IO
I
ExPerienc
DeDrived
d
W a
0 0
a
0 0
I %o”
5
I
I
I
e 0
I3c
. 0
.
l.0 8 . 0,. .0.’ 0
l
*“”
f
o”
I
I
I
1000
2000
I 0
t
I
1000
2000
TRIALS FIG. 3. Visual form discrimination in the cat. Animals were trained to discriminate a horizontal from a vertical rectangle in a Yerkes-Watson apparatus. Each frame depicts the proportion of response to the positive stimulus as a function of the number of trials. The data symbols each represents 100 trials. Filled symbols: Deprived eye. Unfilled symbols: Experienced eye. Symbols above the short horizontal line represent discrimination behavior at criterion(85s correct). Top frames : Control subject, no visual deprivation. Middle and bottom frames: Subjects reared monocularly for 10 weeks, The acuity of the two monocularly-reared subjects is given in Fig. 2.
Vk3JAL
DEPRIVATION
651
perceptual-motor adjustments with their initially-deprived eye. For example, one cat was observed jumping from the centerboard of the visual cliff to avoid the “deep” side of the visual cliff. The jump was accurately adjusted in force so as to bring the animal to the far ledge. However, some animals in the reverse suture group remained inactive and appeared fearful, showing little inclination for exploration either in an open-field situation, or in their own cage. Even after many months, during which time the initially experienced eye was sutured closed, the fearful animals showed much less development of visuo-motor coordination. Viszull Placing. The ability of the deprived eye to elicit a visual placing response develops during the 14-day period between first and second testing (Table 1). All groups-those with short and those with longer periods of visual deprivation-came to show the placing response to an approaching edge. Seven animals (recovery group) that had been reared with complete monocular deprivation were retested after the deprived eye had obtained 1 month of visual experience. They all showed visual placing when using only the initially deprived eye at that time. They were tested again several months later and once more showed the visual placing response without exception. Development on V&al Clifl, Spontaneous Response. Performance on the visual cliff-where the relative aversion of the animal for the visually “deep” was ascertained-showed improvement after the deprived eye was given 14 days of visual exercise. The 4- and g-week deprivation group went from essentially chance preference (1.00/0.83) between “near” and “far” sides of the cliff in the first test to a preference for “near” (3.67/2.00) in the second test (Table 2). A t test of this preference failed to attain the .05 confidence level. This group also showed a rise in tendency to leave the centerboard (i.e., it showed a reduction in the average number of times it did not leave the centerboard) from 8.17 to 4.33 out of 10 opportunities. As was observed in the open-field situation, the tendency to stay on a small elevated platform was definitely associated with an absence of normal perceptual-motor adjustments to other spatial properties of the visual environment. In contrast, the 12-week group (combining subjects with longer durations of visual deficiences) did not show improvement in spontaneous performance on the visual cliff. The group that was only partially deprived of patterned vision through its deprived eye showed essentially complete development during the 14-&y period since on the second test the animals performed equally well with experienced or with deprived eye, Development tith the Forced Response. This version of the visual cliff test showed essentially the same results as the spontaneous response version (Table 3). We consider first the difference between the resistance
652
GANZ
AND
FITCH
to being pushed onto the “far” and “near” sides of the cliff. When tested with the deprived eye, this difference increased from first to second testing. The increased difference reflects both a lessened aversion to being pushed onto the “near” side and an increased aversion to being pushed onto the “far” side. Which-sometimes both-was affected during the 14 day period, varied markedly from subject to subject. A t test was applied to the pairs of resistance as paired comparisons. The differences in resistance are statistically significant in both the 4- and S-week deprivation group and in the 12-week deprived group (p < .025 and p < .05, respectively). Fourteen clays during which both deprived and experienced eyes were open was sufficient to develop a definite appreciation for the depth aspects of visual stimuli. In Table 6 the effects of more extensive visual experience on visual cliff behavior are shown. Again, when the deprived eye was first opened, the animals showed no greater resistance to being pushed onto the “deep” side than onto the “near” side. This contrasted markedly with the performance rCa. the experienced eye. After 13 days of binocular visual esperience, several of the subjects began to show more resistance to being pushed toward the “far” side than to the “near.” If the subjects were then given visual experience through the deprived eye equal to the original deprivation period, a.11came to show a resistance-difference of at least two points on our scale. This applied to the reverse suture group given the delayed experience to the deprived eye while the other eye was closed, and it applied equally to the binocular group with both eyes left open during the delayed experience period. Thus, ezter?! deprived animal in the delayed experience developed an appreciation of depth differences which included a greater relative aversion to a surface that appeared “deep.” The several indices of depth perception were not always correlated. For esample, one cat on later tests showed very strong resistance to being pushed on the “deep” side of the visual cliff and only moderate resistance to the “near” side. Yet this subject was so fearful, when its deprived eye was open, that it remained on a 15-cm high platform for over 10 min. By contrast, another, which appeared less emotional and explored more, did much better in both situations. Differences in performance with the deprived eye also were obvious in the two forms of the visual placing test. On the easier form of the test the animal was simply brought toward an edge of a horizontal surface. On this test all animals eventually showed a positive response when using only their deprived eye version of the test. Two of these animals had had a long period of reversed-suture monocular experience with their deprived eye. Their acuity had improved during this period of time until it was close to normal (Fig. 5). They both showed strong visual placing responses on the
22 4 14 12 4 10 12
Subjects
9E 15B 16E 18A 15A 16A 18C
9E 15B 16E 18-4 15‘4 16-4 18C
Initial deprivation (weeks)
EFFECTS
Group
VISUO-MOTOR
Reverse-suture “ “ I‘ I‘ “ “ Binocular “ ‘4
OF DELAYED
TABLE MEAN
(Initially 2 0 2 2 0 2 1
(Initially 4 4 0 3 4 0 3
6
experienced eye) 4 2 4 0 4 0 2 2 4 0 4 0 4 1
deprived eye) 4 3 4 0 0 0 3 4 4 1 0 0 3 1 4 4 4 4 4 4 4
4 0 0 4 4 4 4
Far
VISUAL
14 Days
INDEX.
Near
RESISTANCE
First Test Near Far
EXPERIENCE.
22 4 14 12 4 10 12
2 2 0 2 0 2 1
1 2 2 2 2 1 0
Near
FORCED RESPONSE
Delayed visuo-motor experience (weeks)
CLIFF:
Post-test
4 4 4 4 4 4 4
4 4 4 4 4 4 3
Far
g. 9.3 2 ?;8 Z:
-e s 2 r
654
GANZ
AND
FITCH
simple test. Both missed placing their paw on the prong on three and two out of eight trials, respectively, with their deprived eye. With their normal eye they performed perfectly. Thus, we conclude that gross adjustments to depth can be developed in the deprived eye of the monocularly-reared animals, but that finer adjustments to the position of visual objects remains deficient even after forced use of the deprived eye. Response. Two groups of cats were Development of the Optonzotor exposed to visual experience and then retested in the optokinetic drum a second time somewhat later. In the binocular group both eyes were left open during the interval between the first and second optokinetic tests. The results are shown in Fig. 4. The top frames represent the optokinetic indices obtained on initial testing, right after the deprived eye was opened. The bottom frames represent retests on the same animals after at least 4 months of binocular visual experience. The optokinetic index of the deprived eye was still inferior for all the animals on the retests. If anything, performance seemedto be somewhat worse on the second test when the 9.3 or 1.3 minutes of arc line was presented to the deprived eye, than on the first testing. One technical concern which had to be initially entertained was that the initial optokinetic testing of the deprived eye, coming as it did soon after the eyelids of the deprived eye had been opened surgically, might have prejudiced the results. Perhaps the lids were insufficiently open; perhaps healing was incomplete. The fact that the performance of the deprived eye did not improve after several months was some assurancethe postsurgery factors were not detrimental to the optokinetic testing of the deprived eye. In the reverse suture group, the deprived eye was kept open during the interval between the first and second optokinetic test and the experienced eye was closed by eyelid suturing. This had the effect of forcing the animal to use visual cues through his previously deprived eye if it was to orient visually at all. Figure 5 shows the results of the optokinetic test. To our surprise, the optokinetic index had become nearly equivalent for the two eyes. The initially experienced eye (given the late deprivation) did not necessarily become poorer. (This means that visual deprivation, if administered to cats older than 3.5 months, does not have anywhere near the effect it does if administered at 10 days of age.) Conversely, suturing the experienced eye led to a rise in the optokinetic index of the initially deprived eye. The data from one cat was of particular importance becauseit gave an optokinetic index to the 1.3 minutes of arc line, which was above zero, to a statistically significant extent. Thus, it appears that the resolution of fine lines can be accomplished after extensive periods of monocular deprivation. Development of I/ri~~4al Discriazinntinn. Two monocularly-reared animals
VISUAL
655
DEPRIVATION
FOLLOWIIJG
MONOCULAR
FOLLOWING
DEPRIVATION
BINOCULAR
EXPERIENCE
:;;I 143
93
13
13
143
93
LiNE
WIDTH
143
9.3
I3
b’m. 4. Three monocularly-reared cats are tested for visual acuity, then given a pcricd of binocular experience, and then retested. The deprived eye is consistently noor cr. Ordinate represents optokinetic index ; abscissa, line widths as in Fig. 2.
FOLLOWING
MONOCULAR
DEPRIVATION
0 50 -v,
$
o
;\;::_;6Q; .
0 F z E 2 %
FOLLOWING
REVERSED
SUTURE
0 50 -
;>No
-yc\
8
0 143
93
13 LINE
143
93
13
WIDTH
FIG. 5. Two monocularly-reared cats were tested for visual acuity, then the experienced eye was closed, the deprived eye given visual experience, and then the animals retested. Note how similar the acuity of the two eyes was on the second test. Ordinate and abscissa as in Fig. 2.
656
GANZ
AND
FITCH
in the reverse suture group were given 3.5 and 5 months of monocular visual experience to their deprived eye (with the experienced eye closed). When trained in the discrimination box with the vertical and horizontal line, performance was still clearly superior with the initially experienced eye. With this eye, these two achieved criterion in 300 and 500 trials, respectively (both were above chance levels after 150 trials). With their initially deprived eye they were still at chance levels after 700 trials. To see whether any form discrimination was possible the problem was made easier. We substituted a vertical VS. horizontal grid of lines and continued training. One subject still failed to get above chance levels after an additional 600 trials. The other subject may have achieved a weak discrimination after an additional 700 trials, being at 66% correct level on the average of the last 10 days of training and above 50% on 9 of those 10 days. Therefore, we concluded that the monocularly-reared subjects do have a small, residual capacity to establish a visual form discrimination through the deprived eye, but that stimuli presented to the experienced eye remain nzuch more effective in leading to form discrimination, even after the animal is forced to make use of his deprived eye for an extensive period of time. Discussion
The preceding paper (3) described the various changes in the behavior of single neurons in the visual system following deprivation. A group of permanent deficiencies were noted. Briefly, cells in the visual cortex of the monocularly-reared cat were much less likely to be binocularly activated, were more easily fatigued when stimulated through the deprived eye, responded to stimulation of the deprived eye in a less selective manner with respect to the direction of motion, and required much larger stimulus objects than when the experienced eye was stimulated. Confirming an earlier study (19, 20) it was clear that monocular visual deprivation had severely reduced the capacity of the deprived eye to control the activities of striate area neurons with small and highly selective receptive fields. In contrast, the deprived eye could activate neurons with large unselective receptive fields. By carefully examining which visual functions were permanently destroyed and which were intact, or only temporarily deficient, it should be possible to arrive at a preliminary formulation regarding the essential functions which the small, and highly selective receptive fields mediate. It is safe to exclude visual acuity of the black-line detection sort as an essential function of the selective cortical receptive fields. Acuity, as measured with an optokinetic drum, was relatively high when either the deprived or the experienced eye was used. However, 4 of the 11 animals
VISUAL
DEPRIVATION
657
were clearly able to detect a 9.3 minutes of arc line with their deprived eye. Two animals which were encouraged to use their deprived eye by having their experienced eye closed for several months showed even better acuity. The deprived eye was inferior to the experienced eye in the detection of very thin black lines. Such high visual resolution in the presence of fairly permanent marked losses in small, elongated cortical receptive fields is reasonable since it is known that a cat will continue to show high acuity, when measured with an optokinetic procedure, even after complete ablation of visual cortex (16). We confirmed these findings to our satisfaction by measuring the optokinetic response in two adult cats with visual cortex ablation. Thus, our optokinetic results imply that some residual discrete function at a subcortical level must exist. Many aspects of the subcortical visual system must be functioning essentially normally if an animal can detect and track smoothly a 1.3 minutes of arc line, using only his deprived eye. Although visual resolution, as measured by the optokinetic response, is clearly not mediated by neurons of the visual cortex, it remains quite possible that neurons with elongated receptive fields residing in the superior colliculus, a pretectal region, play some important role here (11). Such cells are known to exist in the gray ground squirrel (Michels. personal communication), but their existence has not yet been established in the cat. A second group of visual functions were clearly deficient when the deprived eye was first opened but gradually improved following visual experience to the deprived eye. Into this category we would put the following visual functions : Gross perceptual-motor adjustments in an open field, avoidance of objects while locomoting, the visual placing response to a simple edge, passive avoidance of “deep” looking surfaces, fixation and following of moving objects. Improvements in all of these visuallymediated activities were present after 10 days of binocular experience and continued to improve over a 3 to S-month period. These observations are similar to previous reports (1, 6, 15, IS). Since the deficiencies of the deprived eye in eliciting activity in the striate area neurons with selective receptive fields were found to be in large measure irreversible (3, Zl), it does not seem likely these neurons are essential parts of the mechanisms of the visual functions just listed. Finally, there was a series of visual functions which appeared to be irretrievably lost. First, certain aspects of depth perception never became totally normal. Although the animals learned to locomote rapidly in an open field, and step off small elevations quickly, with only their deprived eye open ( l), nevertheless they continued to show a startle response when their vibrissae touched an object. This startle response implied a lack of
658
GANZ
AND
FITCH
appreciation of small depth differences. The most dramatic and fairly permanent deficiencies in depth perception involved an inability to guide limbs accurately toward sn-aller visual objects (5 ) Another broad category that c!id not show improvement involves recognition abilities. For examp!e, th c monocularly-reared cats never learned to show signs of recognizing the experimenter with their deprived eye, when they quite obviouslv did with their experienced eye. These observations suggest the animals llave a specific deficit related to the visual recognition of an object. The almost total absence of visual form discrimination ability when these animals used their deprived eye implies the same deficit of visual recognition. This was true even though the animals could-in some operational sense of the wurtl--“see” the stimuli. That is to say, the discrimination involved stimuli clearly within the resolution capacities of the deprived eye, to judge from optokinetic tests made on the same animals. Moreover, the deficiency appeared rather permanent because many months of forced visual experience through the deprived eye failed to improve the deficiency in forming visual pattern discriminations. We conclude that our experiments support the hypothesis that the neurons of the visual cortex with highly selective receptive fields are essential for visual form discrimination, for the recognition of visual objects. Since formdeprived cats have no problem learning an intensity discrimination (13)) the deficit appears to be specific to the form aspects of the problem. A permanent deficiency in interocular transfer also results from monocular rearing. Even animals that had been given several months of binocular visual experience (13~ and 13d, in Fig. 3) continued to show essentially perfect performance through the experienced eye and chance-level performance through the initially-deprived eye on alternate training days. hbsence of interocular equivalence following monocular deprivation has been reported (2, 15). In our experiment, the same animals which showed absence of interocular equivalence also showed a much reduced proportion of binocular cells (137 0 in contrast to 85% in normal animals). Thus, our results support the hypothesis that the binocular cells play an essential role in integrating information coming into the two eyes. More specifically, an animal cannot tap the memory traces layed down via the feature extractors ( i.e., the striate cells with highly selective receptive fields) controlled by one eye using the feature extractors of the other eye atnless the very same feature extractors are excited because they are binocularly activated cells. In other words, the feature extractor is along the final common pathway to the memory trace. In one sense it is puzzling that the cats do not show interocular transfer of training. After all, rats can learn and retain a visual form discrimination with only l/60 of the striate area remaining (8) and cats can learn a visual
VISUAL
DEPRIVATION
659
iorm discrimination with the optic nerve deprived of all but 2% of its libcrs (10 ) . \\:hy then, can’t monocularly-reared cats show any interocular transfer at all when they still retain 1.370 of their binocular cells? It appears very likely that the binocularly activated cells are somehow sup1)ressed either during monocular acquisition of the discrimination through the experienced eye or during testing with the deprived eye. It is probably true that although the deprived eye elicits less activity in the striate cortex, the topographic projection of the deprived eye’s retina onto striate is in some measure retained. t\‘e concluded this from recordings of columns of cortical cells during vertical microelectrode descents. The columns contained cells of similar orientation and similar loci in the visual field. Both the cells stimulated by the deprived or experienced eyes were part of the same column (3 1. Hence, the topographic arrangements found at birth (7) are to some extent retained in the monocularly-reared cat. This must mean that when the monocularly-reared cat looks at a vertical line-whichever eye is opeii-a longitudinal ridge of cellular activity is aroused along an anterior-posterior axis on the lateral gyrus. When the cat looks at a horizontal line, a ridge of activity is established along a left-right asis. Ijut the cat cannot use this information to form a discrimination when he has only his deprived eye open. This must mean that the striate cells with elongated receptive fields evaluate the visual information given hy topographic projection. If these “feature extractors” are absent-as they are in effect when the cat views with his deprived eyethen the information is almost totally lost for the animal. For the adult cat, there does not appear to be any other decoding scheme of a field-type such as “interference waves” (9) by which the animal can use spatial information once the highly selective feature-extractors at the striate cortex are gone : Conversely, the nonselective, large receptive field neurons do not seem to play a major role in mediating visual form perception. Our main conclusion, then, is that animals that lack visual featureextractors have very great difficulty in learning form discrimination and fine perceptual-motor adjustments, although the peripheral visual apparatus has the capacity to resolve fine detail. to map the visual field topographically. to register gross depth, and to analyze movement. References B. L. 1966. Effect of visual deprivation during postnatal maturation on the electroencephalogram of the cat. Esptl. Nrzwol. 14: 224237. 2. C~row, KAO CHIANG, and H. 1%‘. NISSEK. 1955. Interocular transfer of learning in visually naive and experienced infant Chimpanzees. 1. Cowp. Ph~siol. I’sychol. 48 : 229-237. 3. GANZ, L., M. FITCH, and J. .I. SATTEWBEKG. 1968. The selective effect of visual deprivation on receptive field shape determined neurophysiologically. E.rptZ. Neural. 22 : 61&637. 1.
BAXTER,
660
GANZ
AND
FITCH
4. HECIIT, S., and E. V. MINTZ. 1939. The visibility of single lines at various illuminaticrrs and the retinal basis of visual resolution. J. Gcrt. I-‘hysio[. 22: 593612. 5. HEIN, A., and R. HELU. 1967. Dissociation of the visual placing response into elicited and guided components. Science 158 : [emailprotected] 6. HELD, R., and A. HEIN. 1963. Movement-produced stimulation in the development of visually guided behavior. J. COWIP. Physiol. Psychol. 56: 872-876. 7. HUBEL, D. H., and T. N. WIESEL. 1963. Receptive fields of cells in striate cortex of very young, visually inexperienced kittens. J. Neurophysiol. 26: 994-1002. 8. LASHLEY, K. S. 1939. The mechanism of vision: XVI. The functioning of small remnants of the visual cortex. J. COIMP. Ncnrol. 70 : 45-67. 9. LASHLEY, K. S. 1942. The problem of cerebral organization in vision. Biol. [emailprotected] 7 : 301322. 10. NORTON, T., G. FROMMER, and R. GALAMBOS. 1966. Effects of partial lesions of optic tract on visual discriminations in cats. Federation Proc. 25: 573. 11. PASIK, P., and T. PASIK. 1964. Oculomotor functions in monkeys with lesions of the cerebrum and superior colliculi, pp. 40-80. In “The Oculomotor System,” M. B. Bender [ed.] Harper (Hoeber), New York. 12. REICHARDT, W. 1961. Autocorrelation. A principle for the evaluation of sensory information by the central nervous system, pp. 00-00. In “Sensory Communication,” W. A. Rosenblith [ed.] Mass. Inst. Technol., Cambridge, Massachusetts and Wiley, New York. 13. RIESEN, A. H., and L. AARONS. 1959. Visual movement and intensity discrimination in cats after early deprivation of pattern vision. J. Camp. Physiol. Psychol. 52 : 142-149. 14. Riesen, A. H., M. E. KIJRKE, and J. C. MELLINGER. 1953. Interocular transfer of habits learned monocularly in visually naive and visually experienced cats. J. [emailprotected] Physiol. Psychol. 46 : 166-172. 15. RIESEN, A. H., and J. C. MELLINGER. 1956. Interocular transfer of habits in cats after alternating monocular visual experience. J. Camp. Physiol. Psychol. 46: 516520. 16. SMITH, K. V. 1938. Visual discrimination in the cat: VI. The relationship between pattern vision and visual acuity and the optic projection centers of the nervous system. J. &net. Psychol. 53 : 251-272. 17. SOUTH, K. U., W. E. KAPPAUF, and J. BOJAR. 1940. The functions of the visual cortex at different velocities of movement in the visual field. J. Gen. Psychol. 22 : 341357. 18. WALK, R. D., and E. J. GIBSON. 1961. A comparative and analytic study of visual depth perception. I’sychol. Mo~og. 75 : l-44. 19. WIESEL, T. N., and D. H. HUBEL. 1963. Single-cell responses in striate cortex of kittens deprived of vision in one eye. J. Neurophysiol. 26: 1003-1017. 20. WIESEL, T. N., and D. H. HUBEL. 1965. Comparison of the effects of unilateral and bilateral eye closure on cortical unit responses in kittens. J. Neurophysiol. 28 : 1029-1040. 21. WIESEL, T. N., and D. H. HUBEL. 1965. Extent of recovery from the effects of visual deprivation in kittens. J. Neurophysiol. 28: 1060-1072.