Volume 21
Number 3
Reports
in response to alternating gratings can be useful to
distinguish peripheral from central diseases of the
visual neural pathway. Our findings supply a
pathophysiological interpretation for recently reported clinical findings on impairment of pattern
reversal ERG in patients with retinal degeneration, optic neuritis, and amblyopia.7"11
From Istituto di Neurofisiologia del C.N.R., Pisa,
Italy and Divisione Oculistica Spedali Riuniti, Livorno,
Italy (V. P.). Supported in part by C.N.R. Special Project on Biomedical Engineering. M. P. has a fellowship
from Scuola Normale Superiore Pisa. Submitted for
publication Feb. 9, 1981. Reprint requests: Dr. Lamberto Maffei, Istituto di Neurofisiologia del C.N.R., Via
S. Zeno, 51, 56100-Pisa, Italy.
Key words: ERG, flash stimulus, pattern stimulus, ganglion cells, occlusion, retinal artery, retrobulbar optic
neuritis
REFERENCES
1. Granit R: Sensory Mechanisms of the Retina. New
York, 1963, Hafner Publishing Co.
2. Granit R: The visual pathway. In The Eye, Davson
M, editor. New York, 1962, Academic Press, Inc.,
vol. 2.
3. Armington JC: The Electroretinogram. New York,
1974, Academic Press, Inc.
4. Johnson EP, Riggs LA, and Schick AM: Photopic
retinal potentials evoked by phase alternation of a
barred pattern. In Clinical Electroretinography,
Burian EM and Jacobson JH, editors. Oxford, 1966,
Pergamon Press, Ltd.
5. Maffei L and Fiorentini A: Electroretinographic responses to alternating gratings before and after section of the optic nerve. Science 211:953, 1981.
6. Maffei L and Campbell FW: Neurophysiological localization of the vertical and horizontal visual coordinates in man. Science 176:386, 1970.
7. Lawwill T: The bar-pattern electroretinogram for
clinical evaluation of the central retina. Am J Ophthalmol 78:121, 1974.
8. Sokol S and Nadler D: Simultaneous electroretinograms and visually evoked potentials from adult
amblyopes in response to a pattern stimulus. INVEST
OPHTHALMOL VIS SCI 18:848, 1979.
9. Groneberg A and Teping C: Topodiagnostik von
Sehstorungen durch Ableitung retinaler und kortikaler Antworten auf Umkehr-Kontrastmuster. Ber
Zusammenkunft Dtsch Ophthalmol Ges 77:409,
1980.
10. Arden GB, Carter RM, Hogg CR, Powell DJ, and
Vaegan: Reduced pattern electroretinogram suggest
a preganglionic basis for non-treatable human amblyopia. J Physiol 308:82, 1980.
11. Arden GB, Vaegan, Hogg CR, Powell DJ, and Carter RM: Pattern ERGs are abnormal in many
amblyopes. Trans Ophthalmol Soc UK (in press).
493
Development of binocular depth perception
in kittens. BRIAN TIMNEY.
By means of the jumping stand technique, binocular and
monocular depth thresholds were measured in kittens 4
weeks to 4 months old. Binocularly, performance improved very rapidly during the fifth and sixth weeks,
coinciding with the maturation of cortical disparitytuned neurons. Discriminations made monocularly took
longer to learn and thresholds were consistently poor.
The results demonstrate that kittens can use binocular
cues for depth at a very early age.
Under most circumstances, judgments about
the relative distance of nearby objects are far more
accurate if a person can use both eyes rather than
one.' The primary cue contributing to this binocular superiority is retinal disparity, and a great deal
is known about stereoscopic depth perception in
adults. However, it is only recently that data about
its origin and development have begun to accumulate.2' 3 I have been studying this problem in kittens using a behavioral technique4 and find evidence for binocular superiority at least by the end
of the fifth week of age. These data coincide with
physiological evidence showing that binocular
neurons develop their disparity specificity over
the first few weeks of life.5 The results suggest that
these disparity-specific cortical neurons, which are
thought to underlie stereopsis,6 are functional
from a very early age.
Most previous studies of depth perception in
young animals have used the visual cliff technique
developed by Walk and his colleagues,7' 8 who
have shown that kittens begin to discriminate the
deep from the shallow side reliably by the end of
the first month. The visual cliff technique relies on
the natural tendency of animals to avoid stepping
over a sharp drop, and it has proven invaluable in
demonstrating that very young animals are able to
discriminate large differences in depth. However,
the procedure seems inappropriate for obtaining
accurate threshold measurements, perhaps because the threshold for caution with respect to a
cliff is greater than that for recognizing that a slight
difference in height is present.
The present study capitalized on the kitten's
tendency to descend to the closer appearing of two
surfaces, rewarding such responses with a small
amount of food and punishing attempts to go to an
apparently more distant surface with a loud (>95
dB) high-frequency (4.5 kHz) tone. The procedure
is a modification of the jumping stand technique
used by Mitchell et al.9 for measuring visual
0146-0404/81/090493+04$00.40/0 © 1981 Assoc. for Res. in Vis. and Ophthal., Inc.
Downloaded From: http://jov.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933330/ on 06/17/2017
494
Invest. Ophthahnol. Vis. Sci.
September 1981
Reports
Fig. 1. Jumping stand apparatus for measuring
depth perception contains a seesaw arrangement
(A) so that either dotted pattern may be brought
closer to a Plexiglass landing surface (B) that
covers the whole top of the box. The separation
between the arms of the seesaw is varied by moving the whole unit up and down. An opaque mask
(C) placed directly beneath the landing surface
and restricting view of all but a small dotted region
may be removed to provide additional cues during
training. A centerboard (D) forces the kitten to
descend to the left or right side. Illumination is
from below through a translucent panel (E). Small
speakers (F) provide audible indication of an incorrect choice. The platform from which the kittens descend is not shown.
acuity. The testing procedure used relies on the
assumption that superior binocular performance
under otherwise identical viewing conditions occurs because the kitten is using retinal disparity
as a depth cue when viewing binocularly but is
forced to use monocular cues when one eye is
closed. This assumption has formed the basis of
other tests of stereopsis in cats10 and it seems to be a
reasonable one. Mitchell et al.4 have demonstrated
that binocular depth thresholds in normal cats are
similar to those obtained by others and are far
superior to monocular thresholds. Unpublished
experiments in this laboratory have shown that cats
which one may assume to be stereoblind, such as
Siamese or those made strabismic, do not show
such monocular/binocular differences. Similar results have been obtained for cats raised with alternating monocular occlusion.l0
The apparatus has been described in detail
elsewhere4 and is illustrated in Fig. 1. The kittens
descended onto the apparatus from a jumping platform of variable height (not shown). When training
first began, this platform was situated almost flush
with the landing surface so that the kittens could
simply step down to either side of the centerboard. As the animals became older and more
agile, the platform was raised until the jumping
height was between 70 and 80 cm. It might be
noted here that by the age of 30 days kittens are
becoming quite skilled at getting around and any
limitations in performance almost certainly are not
because of their inability to perform the task.
Initially, the opaque mask of the apparatus was
removed to provide the maximum number of
cues. The stimuli placed on each arm of the seesaw
consisted of a set of dots, varying in size from 8 to
20 mm, which were distributed randomly over the
surface. The number of dots per unit area was
approximately 25%. The kittens were trained to go
toward the closer arm, which was placed 20 cm
higher than the other arm. They were prevented
from falling by the presence of the transparent
Plexiglass cover. The side of the closer arm was
switched randomly over trials. Once a kitten had
mastered this problem to a criterion of 27 correct
out of 30 consecutive trials, the mask was replaced
to reduce the number of possible depth cues. The
kittens were retrained if necessary, and the threshold measurements were begun. Thresholds were
obtained with a modified staircase procedure, in
which the separation was decreased if the kitten
achieved criterion (4 out of 5 correct) and increased if the kitten failed to meet criterion.
Threshold was taken as the smallest separation at
which a kitten performed consistently with a minimum of 70% correct when trials were summed
over a session. Testing was conducted both binocularly and monocularly, the latter conducted
when the kittens wore an opaque occluder over
one eye. In the present study, five kittens were
tested, although they were not all tested through
the full age range.
With the jumping stand it is possible to begin
training kittens younger than 4 weeks, and we
were able to obtain thresholds on normal kittens
by about 30 to 35 days. Prior to this age the kittens
typically were very distressed and paid little attention to the task. Once they had learned the
discrimination they performed willingly when
tested binocularly. However, when one eye was
occluded with an opaque contact lens, performance
deteriorated markedly. Most kittens no longer
seemed to know what was required of them. Sev-
Downloaded From: http://jov.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933330/ on 06/17/2017
Volume 21
Number 3
eral additional sessions were often necessary before
it was possible to obtain a monocular depth threshold. It is unlikely that the poorer monocular performance was a result simply of the kittens' reactions to the lens; they paid little attention to it after
it was inserted. Furthermore, other kittens that
had been monocularly deprived by lid suture or
made strabismic at the time of natural eye opening
generally took much longer to learn the discrimination with the mask present than normal cats,
suggesting that level of performance is controlled
by the ability to use binocular cues.
The developmental functions for monocular and
binocular thresholds also differed markedly (Fig.
2). To make a meaningful comparison between
monocular and binocular performance, all thresholds were expressed in units of retinal disparity.
In the binocular case, disparity was calculated in
the conventional manner, taking into account interocular separation, viewing distance, and the
separation of the targets. When the kittens were
tested monocularly, threshold separations were
converted to the angular retinal disparity that
would have been present if both eyes had been
open. No functional significance should be attached to these units, they simply provide a common measure. Binocularly, thresholds improved
very rapidly up to the age of about 6 weeks. At this
age the best kittens were able to discriminate a
separation of 3 cm from a viewing distance of 40
cm, which is equivalent to about 20 min of retinal
disparity. After the sixth week there was a gradual
improvement toward adult levels. The normal
adult cat is able to resolve a separation of 2 cm or
less from a viewing distance of 70 cm, which represents a disparity of 5 to 10 min. Although there
was some variability among animals, the best of
them reached asymptotic values by about 80 days.
When tested monocularly, the kittens performed consistently worse than when they could
use both eyes. As mentioned above, they took
longer to learn the task initially, and even when
they had learned it they required much greater
separations to be able to discriminate between the
two landing surfaces. Asymptotic threshold values
were achieved also by about 80 days. At this age
most kittens were able to resolve a separation of 10
cm from a distance of 50 cm, which is equivalent to
about 30 min of retinal disparity. It may be seen
from Fig. 2 that final monocular thresholds are
worse by a factor of 5 or 6 compared with binocular
thresholds.
It is possible that the very rapid improvement in
binocular performance is because of the animals
learning better the task required of them. How-
Reports
W
•
c
E
_ •
#*
Oo
(^
o
cp
•
•
50 -
•
495
o
•
100
-
•
A
.tr
150 •
A
200 -
-
A
A
A
D
250 -
>300 -
•
1
AS 0 V
1
V
Age
1
1
1
10
12
14
(wks)
Fig. 2. Depth thresholds plotted as a function of
age. Closed symbols, Binocular thresholds; open
symbols, monocular thresholds. Different symbols
represent data from different kittens. To facilitate
comparison between performance under the two
viewing conditions and to accommodate the fact
that the kittens jumped from greater heights as
they got older, thresholds are expressed as angular
retinal disparity. For monocular testing, thresholds were interpreted in terms of the disparity
that would have been present had both eyes been
open.
ever, the fact that they performed well with large
separations and poorly when the separations were
decreased suggests that they were aware of what
to do but had reached the limits of their ability to
discriminate depth.
The most striking feature of the present data is
the very early age at which kittens are able to
utilize binocular cues in discriminating depth
compared with their use of monocular cues. It is
difficult to quantify the magnitude of this difference, since we were unable to obtain satisfactory
monocular thresholds when the kittens were very
young. However, their behavior under the two
viewing conditions was quite different. For the
Downloaded From: http://jov.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933330/ on 06/17/2017
496
Invest. Ophthalmol. Vis. Sci.
September 1981
Reports
most part, with both eyes open they descended to
the closer surface quite spontaneously; with one
eye covered the task seemed unfamiliar to them
and they had to be trained to go to the appropriate
side. Also, when the height of the jumping platform was changed, the first few jumps of the kittens typically were very awkward when they were
forced to use one eye.
The superiority of binocular thresholds, even in
animals as young as 34 days, suggests that stereoscopic vision is present in very young kittens.
Evidence for the mechanisms that would permit
such an ability has been provided by Petti grew. 5
He reported that after the fifth week, the proportion of binocular neurons tuned for retinal disparity begin to approach adult levels. His data correlate well with the present finding of a rapid improvement in binocular performance during the
fifth and sixth weeks. Taken together; the present
results and those of Pettigrew 5 provide strong
support for the view that disparity-specific neurons underlie stereoscopic depth perception and
that these neurons are able to mediate behavior as
soon as they are physiologically mature.
Recent studies by Held et al. 3 have suggested
that similar developmental changes may take place
in human infants. These authors found a very
rapid improvement in stereoacuity occurring over
a period of just a few weeks. Their data, as well as
those from the present study, suggest that the
mechanisms for stereoscopic depth perception develop independently of those responsible for spatial
resolution, which seem to develop with a much
different time course. 11 ' 12
I thank John Orphan for constructing the apparatus
and Betty Kay Williams, Carla Schneider, and Cathy
Fuller for help in testing the kittens.
From the Department of Psychology, University of
Western Ontario, London, Canada. This research was
supported by grant A7062 from NSERC Canada and MA
7125 from MRC Canada. Submitted for publication April
20, 1981. Reprint requests: Brian Timney, Department
of Psychology, University of Western Ontario, London,
Canada N6A 5C2.
Key words: stereopsis, kittens, visual development,
depth discrimination
REFERENCES
1. Howard HJ: A test for the judgment of distance. Am
J Ophthalmol 2:656, 1919.
2. Shea SL, Fox R, Aslin RN, and Dumais ST: Assessment of stereopsis in human infants. INVEST OPHTHALMOL VIS SCI 19:1400, 1980.
3. Held R, Birch E, and Gwiazda J: Stereoacuity of
human infants. Proc Natl Acad Sci USA 77:5572,
1980.
4. Mitchell DE, Kaye M, and Timney B: Assessment of
depth perception in cats. Perception 8:389, 1979.
5. Pettigrew JD: The effect of visual experience on the
development of stimulus specificity by kitten cortical neurons. J Physiol 237:49, 1974.
6. Barlow HB, Blakemore C, and Pettigrew JD: The
neural mechanism of binocular depth discrimination. J Physiol 193:327, 1967.
7. Walk RD and Gibson EJ: A comparative and analytical study of visual depth perception. Psychological Monographs: General and Applied 75:1, 1961.
8. Walk RD: The development of depth perception in
animals and human infants. Monogr Soc Res Child
Dev 31:82, 1966.
9. Mitchell DE, Giffin F, and Timney B: A behavioural
technique for the rapid assessment of the visual
capabilities of kittens. Perception 6:181, 1977.
10. Blake R and Hirsch HVB: Deficits in binocular
depth perception in cats after alternating monocular
deprivation. Science 190:1114, 1975.
11. Mitchell DE, Giffin F, Wilkinson FE, Anderson P,
and Smith ML: Visual resolution in young kittens.
Vision Res 16:363, 1976.
12. Gwiazda J, Brill S, Mohindra I, and Held R: Preferential looking acuity in infants from two to fifty-eight
weeks of age. Am J Optom 57:428, 1980.
Downloaded From: http://jov.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933330/ on 06/17/2017