- Email: [email protected]
The development of visual acuity in very young kittens. A study with forced-choice preferential looking
0042-6989’85
C?.WJ~ Res. Vol. 25. No. 6. pp. 781-789. 1985 Pnnred m Great Britain. All rights reserved
53.00 + 0.00
Copyright Q 1985Pergamon Press Ltd
THE DEVELOPMENT OF VISUAL ACUITY IN VERY YOUNG KITTENS. A STUDY WITH FORCED-CHOICE PREFERENTIAL LOOKING RUXANDRA SIRETEANU
Max-Planck-Institut
fiir Himforschung,
Deutschordenstrasse
46, D-6000 Frankfurt 71. F.R.G.
(Received 27 February 1984; in final reuised /arm I3 Februury 1985) Abstract-The
development of visual acuity was studied longitudinally in young kittens, using a modification of the forced-choice preferential looking method (FPL) devised by Teller er al. [Vision Res. 14. 1433-1439 (1974)] for human infants. Acuity, defined as the spatial frequency which yields 70% correct responses by a naive observer, shows a l6-fold increase between 2 and IO weeks of age. At comparable ages. acuity evaluated by this method falls short of the acuity values obtained with the mumping stand or with electrophysiological methods. FPL acuity estimated with a more lenient criterion (58%) comes close to the resolution of the highest-resolving single cells in the striate cortex. These results suggest that the preferential looking procedure provides a method that can be used in kittens over a wide age range, including ages at which it is impossible to use the jumping stand method. Visual acuity
Kittens
Preferential looking
INTRODUCIION
The kitten is an excellent model for the study of normal visual development. Due to its visual immaturity at birth, stages of development are available in the newborn kitten, which in other species occur prenatally. A wealth of data exists on the development of the physiological properties of single cells at different levels in the visual system of young kittens, and on their anatomical substrate (for a review, see Norton, 1981). Taken together, these studies suggest that the most rapid developmental changes occur during the first postnatal weeks. Unfortunately, there have been very few quantitative behavioural studies to cover this early developmental period. Due to the impossibility of training very young kittens, and to the tediousness of testing unreinforced behaviour, the earliest available data are from kittens 5-6 weeks old (Mitchell ef al. 1976; Timney, 198 I). The aim of the present work is to describe the development of visual acuity in kittens starting at 2 weeks of age. To evaluate the kittens acuity, a modification of the method of forced-choice preferential looking developed by Teller et al. (1974) for the study of human infants was used. The method is based on the observation that, given the choice between a square-wave high-contrast grating and uniform gray surface of matched average luminance, the infants prefer to stare at the grating (Fantz, 1958). Based on the infants behaviour, a naive observer is able to guess the side of location of the grating. This method is used extensively for the study of visual development in human infants (cf Gwiazda et al., 1978; for a review, see Dobson and Teller, 1978). and more recently in infant monkeys V.R.?J&D
(Teller er al., 1978; for a complete discussion of the technique see Teller, 1979). I felt encouraged to adapt this method for kittens by the recent observation that kittens as young as If14 days orient reliably toward visual stimuli presented not further than 45 deg in their peripheral visual field (Sireteanu and Maurer, 1982). METHODS
The subjects were 26 healthy kittens from 7 litters, aged from 2 to 10 weeks. Two kittens were tested longitudinally over most of this period. Twelve other kittens were tested extensively over periods of 2-3 consecutive weeks. The remaining 12 kittens were tested at only one age. Since there was no difference whether the kittens were tested longitudinally or cross-sectionally, the results shall be presented together. The kittens were tested daily, five times per week. During one experimental session, the kitten was held gently by an experimenter on a white table. A second experimenter drew the cat’s attention to the midline by noises or a centering visual stimulus. When she judged that the gaze of the kitten was properly centered, a white cardboard containing the stimuli was presented at 19cm in front of the kitten. The acuity cards were 40 cm wide (93 deg) and I5 cm high (40 deg). This covered most crthe effective visual field of the youngest kittens (Sireteanu and Maurer, 1982). The stimuli were square-wave vertical gratings of high contrast (close to I), paired with a gray surface of matched mean luminance (58cd/m’). Both the gratings and the gray surfaces were circular, 21 deg in diameter, and centered at 33.5 deg from the midline (see Fig. 1). 781
For all kittens . gratings of SIX different spatial frequencies were used: 0.10, 0.10. O.-RI.0.80, I .60 and 3.20~ deg. During each session. each spatial frequency was presented eight times. four times on the right side and four times on the left. The sequence of presentation of the gratings was randomized (method of constant stimuli). Each session contained 48 trials and lasted about 30 min. The observer was not aware of the side of presentation of the gratings. After each trial, she was forced to guess the location of the grating, basing her judgement on the kitten’s behaviour. In some instances, the card containing the stimuli was moved slowly from left to right. This maximized the chances of the stimuli to fail on the retinal area of highest spatial resolution, and gave the kittens a chance to follow the stimuli, thus providing an additional cue for the observer. Trial-by-trial feed-back was provided by the experimenter holding the kitten. For the youngest ages (2-4 weeks), IO-12 kittens were tested for each week of development; Ior older age groups, at least four kittens were tested for each week. Four additional kittens 14-20 days old (3rd week) were tested with the stimuli centered at 20 deg, instead of 33.5 deg. During this control experiment, the experimenter holding the kitten was prevented from seeing the stimuli by opaque goggles (doubleblind procedure). The responses were pooled over all kittens for each spatial frequency and each week of development. The descending part of the resulting psychometric functions was fitted by straight lines. The curves were fitted by eye. The interpolated spatial frequency yielding a correct response of 70% was taken as a measure of visual acuity. This way of processing the data was chosen to enable a direct comparison with a recently completed study on human infants (Sireteanu et al., 1984).
toire increased. and their Interest in other ~ispe~t< tit the experiment made them less suited for rhis !est. Even for kittens of the same age. the responses depended on their physical condition. The srrongrst kittens were also the best experimental subjects. Part of this variability might be due to the {presumable) different gestational ages of the different litters. HOLVever. clear ditrerences were seen even between iitermates of obviously diKerent physical strength. Pq&ametric
fk-lions
(a) Individuaf kittens. Figure 2 shows a few representative psychometric functions for individual kittens of different ages. For each kitten, the data were pooled over one week of development. It is obvious from this figure that kittens of increasing age respond to gratings of progressively higher spatial frequencies. The visual acuity of individual kittens can be estimated by the spatial frequency which yields 709); correct responses from the observer (shown by a dotted iine in Fig. 2). Figure 3 shows the acuity of a part of the tested kittens, plotted as a function of age. Each symbol indicates the acuity of a kitten, tested daily over one week. Kittens tested longitudjna~ly are indicated by the same symbols. Although there are some inconsistencies, it is clear that, on the average, visual acuity increases with age. This increase occurs very rapidly during the first weeks of testing, and more slowly after 6 weeks of age. Between 2 and 10 weeks of age, the averaged acuity develops by a factor 15, from 0.12 to 1.80c/deg (shown by star symbols in Fig. 6). (b) Cun~ufff~e~reszrlrs. To provide a more reliable estimate of the kittens visual acuity at various ages, 100
1
RESULTS
Behavioural observations Most kittens easily completed the 48 daily trials. However, their responses varied, according to their age and physical condition. During the third week of life, the kittens responsiveness to gratings was poor, although they clearly oriented toward auditory stimuli and sometimes toward high-contrast moving objects presented in their peripheral visual field. However, some kittens definitely oriented toward gratings with the lowest spatial frequencies (see Figs 1 and 3). With increasing age, the gratings became increasingly effective; at 7-8 weeks, the gratings were SO com~I~ng that some of the kittens litera&+ jumped on them, ignoring the pairing gray surface. fiowever, even at these ages, the responses rarely exceeded 85-90x, probably because they were not reinforced. For the oldest kittens, the overall responsiveness to gratings decreased slightly; their behavioural reper-
m 0 40.
30
0.1
0.2
SPATIAL
g.4
0.B
FREQUENCY
1.6 ( c ldag
3 3.2
)
Fig. 2.
Examples of psychometric functions for kittens of different ages. For each kitten, data were accumulated over one week of development. Number of trials for each data point: week IV (solid circles): 24; week V (open triangles): 40; week VI (solid triangles): 32; week VIII (open squares): 32; week X (solid squares): 32. The age of the kitten is shown
by numerals next to each curve.
Fig. 1. Experimental arrangement. The kitten shown in this figure was 3 w&ks otd at the time of the experiment.
783
The development of visual acuity in very young kittens
During the fourth week of life, the kittens oriented clearly toward gratings of the three lowest spatial frequencies, but practically randomly toward gratings of the highest spatial frequencies (Fig. 4, solid circles). In the following weeks, the effectiveness of the lower spatial frequencies increased progressively. Higher spatial frequencies also became increasingly effective. At IO weeks of age, the kittens exceeded the acuity threshold for ail but the highest spatial frequency used in this experiment (see Fig. 5, open squares).
0.Y
^
4
1
785
i
Visual acuity development AOC
trook01
Fig. 3. Visual acuity of individual kittens as a function of age. Each point indicates the spatial frequency which yietds 70% correct observer’s response, For each kitten, data were accumulated over one week of development. Kittens tested longitudinally are indicated by the same symbols.
in the following analysis the results are pooled over all kittens tested in each age group. Figures 4 and 5 show the percentage of correct observer’s responses for each spatial frequency and each week of deveiopment. During the third week of development, the pooled responses did not exceed the acuity threshold of 70% for any of the spatial frequencies used in this experiment. However, the kittens’ orienting toward the two lowest spatial frequencies was definitely better than chance. The extrapolated visual acuity of these kittens was about 0. IO c/deg (see Fig. 4, open circles). Similar results were obtained in four additional kittens tested in a control experiment with the stimuli centered at 20deg, instead of 33.5 deg. This shows that the low responsiveness of these very young kittens was not due to their restricted visual field. WEEK III 12 kmens
1.0
Visual acuity, taken as the interpolated spatial frequency which yielded a 70% correct observer’s from about response, increased monotoni~lly O.lOc/deg during the third week to 1.60 c/deg at 10 weeks of age (see Fig. 6, solid circles). These values come very close to the mean visual acuity obtained from individual psychometric functions (Fig. 6, star symbols). Using a less conservative criterion (58x), acuity of the cumulated functions developed from 0.75 to 6.4cfdeg (Fig. 6, open circles). DISCUSSION
Comparison with other behavioural studies A comparison of the developmental curve obtained here with the developmental curve described by Mitchell el al. (1976) using a jumping stand technique is shown in Fig. 7. Although a direct comparison is difficult, since the luminances and contrasts of the stimuli were not identical, and the psychophysical techniques were also different (constant-stimuli vs staircase}, it appears that the jumping stand tends to produce acuities about one octave higher than the
WEEK Ip 10 kittens
WEEK P 6 Iuttens
I
i
Jb
1
I
al
a2
u
SPATIAL
u
u
12
FREOUENCV
al.1uuuu22
OF
THE
L1uuoJu2.2
GRATING
STIMULUS
tc/deg)
Fig. 4. ~ychomet~c funetions of kittens aged 2-5 weeks. Each point was based on 207 trials (week III, open circles), 198 trials (week IV, sotid eireles) and 184 trials (week V, open triangles).
RUXANDRA SIRETEA~-L WEEK 900
!Zl
WEEK PI] 5 ktttens
4 kIttens
I
a1
cz
Cl
SPATIAL
*.I
i‘
11
a.,
FREOUEUCY
c*
I4
OF
WEEK
0..
I.#
THE
12
,,
a*
GRATING
Pm
a1
u
WEEK X 4 klt!ens
t.
1.2
STIMULUS
u
0.2
w
01
1.6
1.2
(c/des)
Fig. 5. Psychometric functions of kittens aged 610 weeks. Each point was based on 112 trials (week VI,
solid triangles). 96 trials (week VII. open diamonds), 120 trials (week VIII, solid diamonds) and 96 trials (week X, open squares).
method of forced-choice preferential looking (compare the solid symbols in Fig. 7). One reason for this difierence might be that the two methods tap acuity at different neural levels: the jumping stand might require the kittens to make a decision based on cortical activity, while the visual orienting reaction (on which preferential looking is based) might be mediated by the superior colhculus; visual acuity of single celis in the colliculus is much lower than the acuity of cortical cells (Bisti and Sireteanu, 1976). Another reason might be that on the jumping stand. the kittens fixate the patterns with their area centralis, while here their decision is probably based on a more peripheral position of the patterns. Finally, it might be that the jumping stand, with its relatively high penalty for failure (after a
Fig. 6. Development of preferential-looking acuity of kittens aged 2-10 weeks. Solid citi acuity d&red as the spatial frequency which yields too/, cormct observer’s responses. The bars indicate one standard error of the mean. Open circles: acuity deftned by the Sg% observer correct criterion. Data from Figs 3 and 4.
miss, the kittens were let to fail 4Ocm to the floor) is more likely to tax the kittens at their behavioural limit than the compkte freedom of the preferentiailooking situation. Note, however, that this difference holds only for the 70% correct observer’s criterion. For a 58% criterion, the acuity values obtained with the FPL procedure are actually higher than those obtained with the jumping stand (compare Fig. 2 in Mitchell et al., 1976, with Figs 4 and 5 in the present paper). Comparison with electrophysiologicul studies The development of the kittens’ visual acuity has been studied with several ekctrophysiological methods: using the cortical visually evoked potential in anaesthetized (Freeman and Marg, 1975) or awake kittens (Fiorentini et al., 1983), the responses of single cortical cells (Derrington, 1977) or of cells in the lateral geniculate nucleus (Ikeda and Tremain, 1978) and the retina (Ikeda, 1979). The results of these studies are shown in Fig. 7. Although all show a clear development in the first postnatal weeks, these results again cannot be compared directly, hecause the stimuli had different luminances and contrasts, and sometimes also different wave-forms, Freeman and Marg (1975) used square-wave gratings, while Derxington (1977), Ikeda and Tremain (1978), Ikeda (1979) and Fiorentini et al. (1983) used sinusoidal gratings. The acuity thresholds were also obtained differently: by extrapolation of the evoked potential (Freeman and Marg, 1975; Fiorentini er al., 1983), as the resolution of the highest-resolving cortical cell (Derrington, 1977), or as a mean of the resolution of the special class of “sustained” retinal and LGN celts from the area centralis (Ikeda and Tremain, 1978; Ikeda, 1979).
787
The development of visual acuity in very young kittens
Freeman and Marg. 75 Fiorontini et al.. 93 Derrington, 77 Ikeda Jnd fromain. 79 Ikeda. 79 Mitchell et al.. 79
0
Sir*teanu.
94
- 0 _ q I “EP 1 ;
,ingle
l l
* I
cells
I behauiwr
I 0
1
2
3
4
5
6
7
g
9
10
*12
AGE ( weeks ) Fig. 7. Visual acuity as a function of age. Solid circles: forced-choice preferential looking acuity, taken as the spatial frequency which yields a 70”/, correct observer’s responses (data replotted from Fig. 6). Small closed squares: acuity determined with the jumping stand (700/, correct); data from Mitchell er al.. 1976. Small open circles: acuity determined with the visually evoked potential in anaesthetixed kittens (data from Freeman and Marg, 1975). Small open squares: acuity estimated by extrapolation of the contrast sensitivity function estimated with steady-state visually evoked potentials in awake kittens (data from Fiorentini et al., 1983). x-symbols: optimal acuity of x ganglion cells in the area centralis of the retina (data from Ikeda, 1979). Crosses: mean spatial resolution of “sustained” cells in the LGN (data from Ikeda and Tremain, 1978). Star symbols: spatial resolution of the highest-resolving cortical cell (data from Derrington, 1977). The two sets of acuity values obtained
with visually
evoked potentials agree with each other for the younger age groups; for older kittens, VEP acuity is higher in awake kittens (open squares) than in anaesthetized kittens (open circles). The acuity obtained with the jumping stand (solid squares) is close to the VEP acuity of awake kittens (open squares). While the mean acuity of “sustained” retinal and the LGN ceils with inputs from the area centralis (X-symbols and crosses) agrees with the jumping stand acuity, the acuities determined for single best-resolving cortical cells (star symbols) are, understandably, much higher. Preferential-looking acuity, defined as the spatial frequency which yields a 70% correct observer response, is lower than the acuity determined will all the other methods. Again, if the 58% criterion is considered, the discrepancy decreases, and the acuity of the very young kittens comes close to the values reported for single cortical cells (compare Figs 6 and 7). It is remarkable that the different behavioural and electrophysiological methods tend to produce developmental curves which run parallel to each other; apparently, these methods probe different aspects of the same biological function. It is thus possible, by using a conversion factor, to pass from one developmental curve to the other. This is of importance for the very young age groups, for which acuity testing is particularly frustrating.
Possible neural substrate of acuity development
The results presented in the previous section suggest that the low visual acuity of the young kitten is determined at or before the level of the retinal ganglion cells. In the adult cat, acuity is inversely related to the size of the dendritic field of a particular class, of ganglion cells (the beta-cells described by Boycott and Wlssle, 1974); these cells have sustained, or X-type physiological properties (Levick, 1975). In the young kitten, the receptive fields of ganglion cells are difficult to classify, and their sizes are abnormally large (Russoff and Dubin, 1977). Surprisingly, however, the dendritic fields of identified beta-cells from the area centralis are equal or smaller than in the adult (Russoff and Dubin, 1978). It follows that the low resolution of these cells has to be determined at the preganglionar level. Factors such as poor optics or the smaller size of the eyeball could play a role. However, the kitten’s optic media are cloudy only until about four weeks after birth (Thorn ef al., 1976), and according to Bonds and Freeman (1978), poor optics is not a limiting factor of visual acuity. On the other hand, the combined growth of the retina and of the eyeball accounts for only a factor 1.5 increase in spatial resolution between 3 weeks and adulthood (Russoff, 1979). Therefore, as Russoff and Dubin (1978) point out, the abnormally large size of the receptive fields of retinal ganglion cells may be due to an abnormally extensive extrasynaptic excitatory input or to the lack
of inhibitory synapses onto these cells in the young kitten. The postnatal maturation of these mcchanisms might constitute the neural basis of the development of acuity described in this paper. COSCLUSION
The orienting response toward a patterned stimulus in preference over a uniform surface seems to be a universal property of young mammals: human babies, infant monkeys and kittens show this response. in ail three species, the visual acuity estimated on the basis of this behaviour develops rapidly during the early postnatal periad. The results of the present experiment show that the preferential looking procedure provides a method that can be used in kittens over a wide age range, including ages at which no other behaviourai methods are availabIe. am grateful to Professor Wolf Singer, whose generous and interested support made this research possible. Thanks are due to Angefika Sachse for help with testing tke kittens, to Ulrich P&me, Hedwig Thomas anti Lisa Kannenbrock for preparing the illustrations and to Gisela Knott for typing the manuscript. Acknowledgements-I
REIXRENCES Bisti S. and Sireteanu R. (1976) Sensitivity to spatial frequency and contrast of visual cells in the cat superior collicutus. Vtsion Res. 16, 247-251. Bonds A, B, and Freeman R. D. (1978) Development of optical quality in the kitten eye. Ytrion Res. t8, 391-398. Boycott B. B. and W%sle N, (1974) The morpholo&al types of @nglion c@ls of the domestic cat’s retina, J. Physiot. 240, 397-419.
Derrington A. M. (1977) Deveiopment of selectivity in kitten striate cortex. J. Phy&& Z76, 46p47P. Dobson V. and Teller D. Y. (1978) Visual acuity in human infants: a review and comparison of behavioural and electrophysiological studies. Vision Res. 18, 1469-1483. Fantz R. L. (1938) Pattern vision in young infants. Psychof. Rec. I&43-47. Fiorentini A., Pirchio M, and Spinelli D. (1983) Development of retinal and cortical responses to pattern reversal in infants: a selective review. B&M. Bruin Res. 1% 99-H&5
Freeman D. N. and Marg E. (1975) Visual acu~rk de;&opment coincides with the sensitive period in kittens ‘valure 254, 614-615. Gwiazda J.. Briii S., Xohindra f. and Held R. (1978) Infant visual acuity and its meridional variation. i&ion &.Y. t8,
1557-1564. lkeda ‘cf.and Tremain K. E. (1978) The development of spatial resolving power of lateral geniculatc neuroses in kittens. E.rpL Brain Res. 31, 193-206. ikeda H. (1979) Physiological basis of visual acuity and its development
in kittens.
Child
Care
&t&k
ftecl
S,
375-383.
Levick W. R. (1975) Form and function af cat retinal
ganglion cells. Narure 254, 659-662. Mitchell D. E., Gifin F., Wilkinson F., Anderson P. and Smith M. L. (1976) Visual resolution in young kittens. Vision Res. f6, 363-366.
Nortan 7. T. (1981) Development of the visual system and visually guided behaviour. In ~#efo~~e~r of Perception. Voi. 2. pp. i13-$56. Academic Press, New York. RussofT A. (1979) Development of ganglion cells in the retina of the cat. In Developmental Neurobiatogy of Vision (Edited by Freeman R. D.) Plenum Press, New York. Russoff A. C. and Dubin M. W. (1977) Development of receptive field properties of ret&al gang&on cells in kittens. 2. ~e~~o~k~s~ot. 40, 5188-t 198. Russoff A. C. and’ l&bin M. W. f1978) Kitten ~n~lj~n cells: dendritic field site at 3 weeks of age and correlation with receptive field size. Irtncest. Opkrhal. &suat Sci. 17, 819-821.
Sireteanu R., Kellerer G. and Bargen K.-P. (1984) The development of peripheral visual acuity in human infants. A preljminary study. Hunt, Netrrobiol. 3, 81-85.
Sireteanu R. and Maurer D. 0982) The develoament of the kitten’s visual field. Vis~on‘Res.~22, 1105-l if l. Teller D. Y. (1979) The forced-choice preferential looking procedure: a psychophysical technique for use with human infants. Infanr Behau. Dent. 2, 135-153.
Teller D. Y., Morse R., Borton R. and Regal D. (1974) Visual acuity for vertical and diagonal gr&ngs in human infants. Vision Res. 14, 143s$439.
Teller D. Y., Regal D. M., Videen T. 0. and pufos E. (fQ78) Development of visual acuity in infant monkeys (Macuca nemesrrina) during the early postnatal weeks. Vision&x 18, X1-566. Thorn F., Golfender M. and Erickson P. (1976) The develooment of the kitten’s visual ootic. @ion Res. 16.
lirls-1149. Timney B. (1981) Development of binocular depth percep rion in kittens. !nuesr. Uphihnl: rzswl Sci. 21, 493-496.
C?.WJ~ Res. Vol. 25. No. 6. pp. 781-789. 1985 Pnnred m Great Britain. All rights reserved
53.00 + 0.00
Copyright Q 1985Pergamon Press Ltd
THE DEVELOPMENT OF VISUAL ACUITY IN VERY YOUNG KITTENS. A STUDY WITH FORCED-CHOICE PREFERENTIAL LOOKING RUXANDRA SIRETEANU
Max-Planck-Institut
fiir Himforschung,
Deutschordenstrasse
46, D-6000 Frankfurt 71. F.R.G.
(Received 27 February 1984; in final reuised /arm I3 Februury 1985) Abstract-The
development of visual acuity was studied longitudinally in young kittens, using a modification of the forced-choice preferential looking method (FPL) devised by Teller er al. [Vision Res. 14. 1433-1439 (1974)] for human infants. Acuity, defined as the spatial frequency which yields 70% correct responses by a naive observer, shows a l6-fold increase between 2 and IO weeks of age. At comparable ages. acuity evaluated by this method falls short of the acuity values obtained with the mumping stand or with electrophysiological methods. FPL acuity estimated with a more lenient criterion (58%) comes close to the resolution of the highest-resolving single cells in the striate cortex. These results suggest that the preferential looking procedure provides a method that can be used in kittens over a wide age range, including ages at which it is impossible to use the jumping stand method. Visual acuity
Kittens
Preferential looking
INTRODUCIION
The kitten is an excellent model for the study of normal visual development. Due to its visual immaturity at birth, stages of development are available in the newborn kitten, which in other species occur prenatally. A wealth of data exists on the development of the physiological properties of single cells at different levels in the visual system of young kittens, and on their anatomical substrate (for a review, see Norton, 1981). Taken together, these studies suggest that the most rapid developmental changes occur during the first postnatal weeks. Unfortunately, there have been very few quantitative behavioural studies to cover this early developmental period. Due to the impossibility of training very young kittens, and to the tediousness of testing unreinforced behaviour, the earliest available data are from kittens 5-6 weeks old (Mitchell ef al. 1976; Timney, 198 I). The aim of the present work is to describe the development of visual acuity in kittens starting at 2 weeks of age. To evaluate the kittens acuity, a modification of the method of forced-choice preferential looking developed by Teller et al. (1974) for the study of human infants was used. The method is based on the observation that, given the choice between a square-wave high-contrast grating and uniform gray surface of matched average luminance, the infants prefer to stare at the grating (Fantz, 1958). Based on the infants behaviour, a naive observer is able to guess the side of location of the grating. This method is used extensively for the study of visual development in human infants (cf Gwiazda et al., 1978; for a review, see Dobson and Teller, 1978). and more recently in infant monkeys V.R.?J&D
(Teller er al., 1978; for a complete discussion of the technique see Teller, 1979). I felt encouraged to adapt this method for kittens by the recent observation that kittens as young as If14 days orient reliably toward visual stimuli presented not further than 45 deg in their peripheral visual field (Sireteanu and Maurer, 1982). METHODS
The subjects were 26 healthy kittens from 7 litters, aged from 2 to 10 weeks. Two kittens were tested longitudinally over most of this period. Twelve other kittens were tested extensively over periods of 2-3 consecutive weeks. The remaining 12 kittens were tested at only one age. Since there was no difference whether the kittens were tested longitudinally or cross-sectionally, the results shall be presented together. The kittens were tested daily, five times per week. During one experimental session, the kitten was held gently by an experimenter on a white table. A second experimenter drew the cat’s attention to the midline by noises or a centering visual stimulus. When she judged that the gaze of the kitten was properly centered, a white cardboard containing the stimuli was presented at 19cm in front of the kitten. The acuity cards were 40 cm wide (93 deg) and I5 cm high (40 deg). This covered most crthe effective visual field of the youngest kittens (Sireteanu and Maurer, 1982). The stimuli were square-wave vertical gratings of high contrast (close to I), paired with a gray surface of matched mean luminance (58cd/m’). Both the gratings and the gray surfaces were circular, 21 deg in diameter, and centered at 33.5 deg from the midline (see Fig. 1). 781
For all kittens . gratings of SIX different spatial frequencies were used: 0.10, 0.10. O.-RI.0.80, I .60 and 3.20~ deg. During each session. each spatial frequency was presented eight times. four times on the right side and four times on the left. The sequence of presentation of the gratings was randomized (method of constant stimuli). Each session contained 48 trials and lasted about 30 min. The observer was not aware of the side of presentation of the gratings. After each trial, she was forced to guess the location of the grating, basing her judgement on the kitten’s behaviour. In some instances, the card containing the stimuli was moved slowly from left to right. This maximized the chances of the stimuli to fail on the retinal area of highest spatial resolution, and gave the kittens a chance to follow the stimuli, thus providing an additional cue for the observer. Trial-by-trial feed-back was provided by the experimenter holding the kitten. For the youngest ages (2-4 weeks), IO-12 kittens were tested for each week of development; Ior older age groups, at least four kittens were tested for each week. Four additional kittens 14-20 days old (3rd week) were tested with the stimuli centered at 20 deg, instead of 33.5 deg. During this control experiment, the experimenter holding the kitten was prevented from seeing the stimuli by opaque goggles (doubleblind procedure). The responses were pooled over all kittens for each spatial frequency and each week of development. The descending part of the resulting psychometric functions was fitted by straight lines. The curves were fitted by eye. The interpolated spatial frequency yielding a correct response of 70% was taken as a measure of visual acuity. This way of processing the data was chosen to enable a direct comparison with a recently completed study on human infants (Sireteanu et al., 1984).
toire increased. and their Interest in other ~ispe~t< tit the experiment made them less suited for rhis !est. Even for kittens of the same age. the responses depended on their physical condition. The srrongrst kittens were also the best experimental subjects. Part of this variability might be due to the {presumable) different gestational ages of the different litters. HOLVever. clear ditrerences were seen even between iitermates of obviously diKerent physical strength. Pq&ametric
fk-lions
(a) Individuaf kittens. Figure 2 shows a few representative psychometric functions for individual kittens of different ages. For each kitten, the data were pooled over one week of development. It is obvious from this figure that kittens of increasing age respond to gratings of progressively higher spatial frequencies. The visual acuity of individual kittens can be estimated by the spatial frequency which yields 709); correct responses from the observer (shown by a dotted iine in Fig. 2). Figure 3 shows the acuity of a part of the tested kittens, plotted as a function of age. Each symbol indicates the acuity of a kitten, tested daily over one week. Kittens tested longitudjna~ly are indicated by the same symbols. Although there are some inconsistencies, it is clear that, on the average, visual acuity increases with age. This increase occurs very rapidly during the first weeks of testing, and more slowly after 6 weeks of age. Between 2 and 10 weeks of age, the averaged acuity develops by a factor 15, from 0.12 to 1.80c/deg (shown by star symbols in Fig. 6). (b) Cun~ufff~e~reszrlrs. To provide a more reliable estimate of the kittens visual acuity at various ages, 100
1
RESULTS
Behavioural observations Most kittens easily completed the 48 daily trials. However, their responses varied, according to their age and physical condition. During the third week of life, the kittens responsiveness to gratings was poor, although they clearly oriented toward auditory stimuli and sometimes toward high-contrast moving objects presented in their peripheral visual field. However, some kittens definitely oriented toward gratings with the lowest spatial frequencies (see Figs 1 and 3). With increasing age, the gratings became increasingly effective; at 7-8 weeks, the gratings were SO com~I~ng that some of the kittens litera&+ jumped on them, ignoring the pairing gray surface. fiowever, even at these ages, the responses rarely exceeded 85-90x, probably because they were not reinforced. For the oldest kittens, the overall responsiveness to gratings decreased slightly; their behavioural reper-
m 0 40.
30
0.1
0.2
SPATIAL
g.4
0.B
FREQUENCY
1.6 ( c ldag
3 3.2
)
Fig. 2.
Examples of psychometric functions for kittens of different ages. For each kitten, data were accumulated over one week of development. Number of trials for each data point: week IV (solid circles): 24; week V (open triangles): 40; week VI (solid triangles): 32; week VIII (open squares): 32; week X (solid squares): 32. The age of the kitten is shown
by numerals next to each curve.
Fig. 1. Experimental arrangement. The kitten shown in this figure was 3 w&ks otd at the time of the experiment.
783
The development of visual acuity in very young kittens
During the fourth week of life, the kittens oriented clearly toward gratings of the three lowest spatial frequencies, but practically randomly toward gratings of the highest spatial frequencies (Fig. 4, solid circles). In the following weeks, the effectiveness of the lower spatial frequencies increased progressively. Higher spatial frequencies also became increasingly effective. At IO weeks of age, the kittens exceeded the acuity threshold for ail but the highest spatial frequency used in this experiment (see Fig. 5, open squares).
0.Y
^
4
1
785
i
Visual acuity development AOC
trook01
Fig. 3. Visual acuity of individual kittens as a function of age. Each point indicates the spatial frequency which yietds 70% correct observer’s response, For each kitten, data were accumulated over one week of development. Kittens tested longitudinally are indicated by the same symbols.
in the following analysis the results are pooled over all kittens tested in each age group. Figures 4 and 5 show the percentage of correct observer’s responses for each spatial frequency and each week of deveiopment. During the third week of development, the pooled responses did not exceed the acuity threshold of 70% for any of the spatial frequencies used in this experiment. However, the kittens’ orienting toward the two lowest spatial frequencies was definitely better than chance. The extrapolated visual acuity of these kittens was about 0. IO c/deg (see Fig. 4, open circles). Similar results were obtained in four additional kittens tested in a control experiment with the stimuli centered at 20deg, instead of 33.5 deg. This shows that the low responsiveness of these very young kittens was not due to their restricted visual field. WEEK III 12 kmens
1.0
Visual acuity, taken as the interpolated spatial frequency which yielded a 70% correct observer’s from about response, increased monotoni~lly O.lOc/deg during the third week to 1.60 c/deg at 10 weeks of age (see Fig. 6, solid circles). These values come very close to the mean visual acuity obtained from individual psychometric functions (Fig. 6, star symbols). Using a less conservative criterion (58x), acuity of the cumulated functions developed from 0.75 to 6.4cfdeg (Fig. 6, open circles). DISCUSSION
Comparison with other behavioural studies A comparison of the developmental curve obtained here with the developmental curve described by Mitchell el al. (1976) using a jumping stand technique is shown in Fig. 7. Although a direct comparison is difficult, since the luminances and contrasts of the stimuli were not identical, and the psychophysical techniques were also different (constant-stimuli vs staircase}, it appears that the jumping stand tends to produce acuities about one octave higher than the
WEEK Ip 10 kittens
WEEK P 6 Iuttens
I
i
Jb
1
I
al
a2
u
SPATIAL
u
u
12
FREOUENCV
al.1uuuu22
OF
THE
L1uuoJu2.2
GRATING
STIMULUS
tc/deg)
Fig. 4. ~ychomet~c funetions of kittens aged 2-5 weeks. Each point was based on 207 trials (week III, open circles), 198 trials (week IV, sotid eireles) and 184 trials (week V, open triangles).
RUXANDRA SIRETEA~-L WEEK 900
!Zl
WEEK PI] 5 ktttens
4 kIttens
I
a1
cz
Cl
SPATIAL
*.I
i‘
11
a.,
FREOUEUCY
c*
I4
OF
WEEK
0..
I.#
THE
12
,,
a*
GRATING
Pm
a1
u
WEEK X 4 klt!ens
t.
1.2
STIMULUS
u
0.2
w
01
1.6
1.2
(c/des)
Fig. 5. Psychometric functions of kittens aged 610 weeks. Each point was based on 112 trials (week VI,
solid triangles). 96 trials (week VII. open diamonds), 120 trials (week VIII, solid diamonds) and 96 trials (week X, open squares).
method of forced-choice preferential looking (compare the solid symbols in Fig. 7). One reason for this difierence might be that the two methods tap acuity at different neural levels: the jumping stand might require the kittens to make a decision based on cortical activity, while the visual orienting reaction (on which preferential looking is based) might be mediated by the superior colhculus; visual acuity of single celis in the colliculus is much lower than the acuity of cortical cells (Bisti and Sireteanu, 1976). Another reason might be that on the jumping stand. the kittens fixate the patterns with their area centralis, while here their decision is probably based on a more peripheral position of the patterns. Finally, it might be that the jumping stand, with its relatively high penalty for failure (after a
Fig. 6. Development of preferential-looking acuity of kittens aged 2-10 weeks. Solid citi acuity d&red as the spatial frequency which yields too/, cormct observer’s responses. The bars indicate one standard error of the mean. Open circles: acuity deftned by the Sg% observer correct criterion. Data from Figs 3 and 4.
miss, the kittens were let to fail 4Ocm to the floor) is more likely to tax the kittens at their behavioural limit than the compkte freedom of the preferentiailooking situation. Note, however, that this difference holds only for the 70% correct observer’s criterion. For a 58% criterion, the acuity values obtained with the FPL procedure are actually higher than those obtained with the jumping stand (compare Fig. 2 in Mitchell et al., 1976, with Figs 4 and 5 in the present paper). Comparison with electrophysiologicul studies The development of the kittens’ visual acuity has been studied with several ekctrophysiological methods: using the cortical visually evoked potential in anaesthetized (Freeman and Marg, 1975) or awake kittens (Fiorentini et al., 1983), the responses of single cortical cells (Derrington, 1977) or of cells in the lateral geniculate nucleus (Ikeda and Tremain, 1978) and the retina (Ikeda, 1979). The results of these studies are shown in Fig. 7. Although all show a clear development in the first postnatal weeks, these results again cannot be compared directly, hecause the stimuli had different luminances and contrasts, and sometimes also different wave-forms, Freeman and Marg (1975) used square-wave gratings, while Derxington (1977), Ikeda and Tremain (1978), Ikeda (1979) and Fiorentini et al. (1983) used sinusoidal gratings. The acuity thresholds were also obtained differently: by extrapolation of the evoked potential (Freeman and Marg, 1975; Fiorentini er al., 1983), as the resolution of the highest-resolving cortical cell (Derrington, 1977), or as a mean of the resolution of the special class of “sustained” retinal and LGN celts from the area centralis (Ikeda and Tremain, 1978; Ikeda, 1979).
787
The development of visual acuity in very young kittens
Freeman and Marg. 75 Fiorontini et al.. 93 Derrington, 77 Ikeda Jnd fromain. 79 Ikeda. 79 Mitchell et al.. 79
0
Sir*teanu.
94
- 0 _ q I “EP 1 ;
,ingle
l l
* I
cells
I behauiwr
I 0
1
2
3
4
5
6
7
g
9
10
*12
AGE ( weeks ) Fig. 7. Visual acuity as a function of age. Solid circles: forced-choice preferential looking acuity, taken as the spatial frequency which yields a 70”/, correct observer’s responses (data replotted from Fig. 6). Small closed squares: acuity determined with the jumping stand (700/, correct); data from Mitchell er al.. 1976. Small open circles: acuity determined with the visually evoked potential in anaesthetixed kittens (data from Freeman and Marg, 1975). Small open squares: acuity estimated by extrapolation of the contrast sensitivity function estimated with steady-state visually evoked potentials in awake kittens (data from Fiorentini et al., 1983). x-symbols: optimal acuity of x ganglion cells in the area centralis of the retina (data from Ikeda, 1979). Crosses: mean spatial resolution of “sustained” cells in the LGN (data from Ikeda and Tremain, 1978). Star symbols: spatial resolution of the highest-resolving cortical cell (data from Derrington, 1977). The two sets of acuity values obtained
with visually
evoked potentials agree with each other for the younger age groups; for older kittens, VEP acuity is higher in awake kittens (open squares) than in anaesthetized kittens (open circles). The acuity obtained with the jumping stand (solid squares) is close to the VEP acuity of awake kittens (open squares). While the mean acuity of “sustained” retinal and the LGN ceils with inputs from the area centralis (X-symbols and crosses) agrees with the jumping stand acuity, the acuities determined for single best-resolving cortical cells (star symbols) are, understandably, much higher. Preferential-looking acuity, defined as the spatial frequency which yields a 70% correct observer response, is lower than the acuity determined will all the other methods. Again, if the 58% criterion is considered, the discrepancy decreases, and the acuity of the very young kittens comes close to the values reported for single cortical cells (compare Figs 6 and 7). It is remarkable that the different behavioural and electrophysiological methods tend to produce developmental curves which run parallel to each other; apparently, these methods probe different aspects of the same biological function. It is thus possible, by using a conversion factor, to pass from one developmental curve to the other. This is of importance for the very young age groups, for which acuity testing is particularly frustrating.
Possible neural substrate of acuity development
The results presented in the previous section suggest that the low visual acuity of the young kitten is determined at or before the level of the retinal ganglion cells. In the adult cat, acuity is inversely related to the size of the dendritic field of a particular class, of ganglion cells (the beta-cells described by Boycott and Wlssle, 1974); these cells have sustained, or X-type physiological properties (Levick, 1975). In the young kitten, the receptive fields of ganglion cells are difficult to classify, and their sizes are abnormally large (Russoff and Dubin, 1977). Surprisingly, however, the dendritic fields of identified beta-cells from the area centralis are equal or smaller than in the adult (Russoff and Dubin, 1978). It follows that the low resolution of these cells has to be determined at the preganglionar level. Factors such as poor optics or the smaller size of the eyeball could play a role. However, the kitten’s optic media are cloudy only until about four weeks after birth (Thorn ef al., 1976), and according to Bonds and Freeman (1978), poor optics is not a limiting factor of visual acuity. On the other hand, the combined growth of the retina and of the eyeball accounts for only a factor 1.5 increase in spatial resolution between 3 weeks and adulthood (Russoff, 1979). Therefore, as Russoff and Dubin (1978) point out, the abnormally large size of the receptive fields of retinal ganglion cells may be due to an abnormally extensive extrasynaptic excitatory input or to the lack
of inhibitory synapses onto these cells in the young kitten. The postnatal maturation of these mcchanisms might constitute the neural basis of the development of acuity described in this paper. COSCLUSION
The orienting response toward a patterned stimulus in preference over a uniform surface seems to be a universal property of young mammals: human babies, infant monkeys and kittens show this response. in ail three species, the visual acuity estimated on the basis of this behaviour develops rapidly during the early postnatal periad. The results of the present experiment show that the preferential looking procedure provides a method that can be used in kittens over a wide age range, including ages at which no other behaviourai methods are availabIe. am grateful to Professor Wolf Singer, whose generous and interested support made this research possible. Thanks are due to Angefika Sachse for help with testing tke kittens, to Ulrich P&me, Hedwig Thomas anti Lisa Kannenbrock for preparing the illustrations and to Gisela Knott for typing the manuscript. Acknowledgements-I
REIXRENCES Bisti S. and Sireteanu R. (1976) Sensitivity to spatial frequency and contrast of visual cells in the cat superior collicutus. Vtsion Res. 16, 247-251. Bonds A, B, and Freeman R. D. (1978) Development of optical quality in the kitten eye. Ytrion Res. t8, 391-398. Boycott B. B. and W%sle N, (1974) The morpholo&al types of @nglion c@ls of the domestic cat’s retina, J. Physiot. 240, 397-419.
Derrington A. M. (1977) Deveiopment of selectivity in kitten striate cortex. J. Phy&& Z76, 46p47P. Dobson V. and Teller D. Y. (1978) Visual acuity in human infants: a review and comparison of behavioural and electrophysiological studies. Vision Res. 18, 1469-1483. Fantz R. L. (1938) Pattern vision in young infants. Psychof. Rec. I&43-47. Fiorentini A., Pirchio M, and Spinelli D. (1983) Development of retinal and cortical responses to pattern reversal in infants: a selective review. B&M. Bruin Res. 1% 99-H&5
Freeman D. N. and Marg E. (1975) Visual acu~rk de;&opment coincides with the sensitive period in kittens ‘valure 254, 614-615. Gwiazda J.. Briii S., Xohindra f. and Held R. (1978) Infant visual acuity and its meridional variation. i&ion &.Y. t8,
1557-1564. lkeda ‘cf.and Tremain K. E. (1978) The development of spatial resolving power of lateral geniculatc neuroses in kittens. E.rpL Brain Res. 31, 193-206. ikeda H. (1979) Physiological basis of visual acuity and its development
in kittens.
Child
Care
&t&k
ftecl
S,
375-383.
Levick W. R. (1975) Form and function af cat retinal
ganglion cells. Narure 254, 659-662. Mitchell D. E., Gifin F., Wilkinson F., Anderson P. and Smith M. L. (1976) Visual resolution in young kittens. Vision Res. f6, 363-366.
Nortan 7. T. (1981) Development of the visual system and visually guided behaviour. In ~#efo~~e~r of Perception. Voi. 2. pp. i13-$56. Academic Press, New York. RussofT A. (1979) Development of ganglion cells in the retina of the cat. In Developmental Neurobiatogy of Vision (Edited by Freeman R. D.) Plenum Press, New York. Russoff A. C. and Dubin M. W. (1977) Development of receptive field properties of ret&al gang&on cells in kittens. 2. ~e~~o~k~s~ot. 40, 5188-t 198. Russoff A. C. and’ l&bin M. W. f1978) Kitten ~n~lj~n cells: dendritic field site at 3 weeks of age and correlation with receptive field size. Irtncest. Opkrhal. &suat Sci. 17, 819-821.
Sireteanu R., Kellerer G. and Bargen K.-P. (1984) The development of peripheral visual acuity in human infants. A preljminary study. Hunt, Netrrobiol. 3, 81-85.
Sireteanu R. and Maurer D. 0982) The develoament of the kitten’s visual field. Vis~on‘Res.~22, 1105-l if l. Teller D. Y. (1979) The forced-choice preferential looking procedure: a psychophysical technique for use with human infants. Infanr Behau. Dent. 2, 135-153.
Teller D. Y., Morse R., Borton R. and Regal D. (1974) Visual acuity for vertical and diagonal gr&ngs in human infants. Vision Res. 14, 143s$439.
Teller D. Y., Regal D. M., Videen T. 0. and pufos E. (fQ78) Development of visual acuity in infant monkeys (Macuca nemesrrina) during the early postnatal weeks. Vision&x 18, X1-566. Thorn F., Golfender M. and Erickson P. (1976) The develooment of the kitten’s visual ootic. @ion Res. 16.
lirls-1149. Timney B. (1981) Development of binocular depth percep rion in kittens. !nuesr. Uphihnl: rzswl Sci. 21, 493-496.