Changes in Contrast Sensitivity Induced by Defocus and Their Possible Relations to Emmetropization in the Chicken

Abstract

purpose. To test whether the level of contrast adaptation (CA) relates to refractive development in the chicken. (CA refers to a spatial frequency-selective increase of suprathreshold contrast sensitivity after exposure to low-contrast patterns).

methods. CA was determined in individual chicks by comparing their optomotor gain in response to drifting low-contrast stripe patterns before and after treatment with spectacle lenses. The amount of CA was compared with the loss of contrast predicted from defocus at the tested spatial frequency. The reversion of CA and recovery from deprivation myopia were studied while the retinal image features were controlled by forcing the animals to watch spatially filtered digital video clips.

results. CA was induced by wearing positive and negative lenses for 1.5 hours, both without and with cycloplegia, but was less pronounced in the case of positive lenses when accommodation was intact. The amount of CA at a tested spatial frequency was predicted from the loss of contrast calculated from the modulation transfer function for a defocused optical system. Watching low-pass–filtered video clips induced deprivation myopia and inhibited recovery from it. It also prevented the reversal of CA that was previously induced by deprivation. Both recovery from deprivation myopia and recovery from CA occurred with sharp video clips, although less so than with normal visual exposure.

conclusions. CA changes with retinal image sharpness and occurs even when accommodation is intact. Because CA correlates with myopia induced by frosted occluders, negative lenses, and low-pass–filtered video clips, and its reversal correlates with recovery from myopia, it is possible that shifts in CA may represent a signal related to refractive error development.

A major question of emmetropization is how the optimal refractive state is defined in the presence of continuous shifts of the position of the focal plane with viewing distance and state of accommodation. A retinal mechanism that quantifies image focus with a long temporal integration is necessary. A possible candidate is the level of contrast adaptation (CA), which changes continuously with focus and contrast of the retinal image with relatively long time constants (several minutes in humans). ¹ Because animal experiments have shown that emmetropization is largely guided by retinal image processing, ² it must be postulated that CA occurs as early as the retinal level and not exclusively in the central visual system. Initially, using pattern electroretinograms, Brigell et al. ³ could not find contrast adaptation in the retina, whereas Odom and Norcia ⁴ suggested a retinal contribution of 30%. Recently, the evidence for retinal contrast adaptation has further increased. ^{5 6 7} CA was observed in chickens after 1.5 days of treatment with frosted eye occluders, and recovery to baseline levels was found after 1.5 days of normal vision. ⁸ Both reserpine and atropine, which suppress development of myopia, were found to increase suprathreshold contrast sensitivity to its maximal level, ⁸ leaving the possibility open that emmetropization becomes “blind” to changes in retinal focus. 

To make it likely that CA acts as a retinal signal for enhanced eye growth, a number of conditions have to be met: (1) CA must be inducible by physiologically relevant amounts of retinal image defocus and also in the presence of functional accommodation. (2) The optimum refractive state (emmetropia) should be a condition in which CA is at a minimum—that is, equally powered lenses with different signs should produce similar amounts of CA in an emmetropic animal. (3) The amount of CA at a given spatial frequency should be predictable from the loss of contrast calculated from the modulation transfer function of a defocused optical system. (4) Artificial visual stimulation with low-pass–filtered digital video clips should prevent reversion of CA and recovery from deprivation myopia, but stimulation with unfiltered sharp video clips should trigger reversion of CA and permit recovery from myopia. (5) Artificial stimulation with spatially low-pass–filtered digital video clips should cause deprivation myopia. In the present study, all these hypotheses were tested. 

Methods

Animals

Male white Leghorn chickens of an egg strain originating from a local hatchery in Suppingen, Germany, were obtained at day 1 after hatching. They were raised in groups of 7 to 11 animals in the animal facilities of the institute. The treatment of the chickens was approved by the animal welfare commission of the university (reference: AK 3/99) and adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Treatment

Forty-two chickens with intact accommodation and 24 chickens with paralyzed accommodation wore lenses of −7.4-D (n = 36), −4-D (n = 5), −13.3-D (n = 9), or +6.9-D (n = 16) power for 1.5 hours. Lenses were attached to small leather hoods, as previously described, ⁹ and were binocularly applied. Binocular cycloplegia was induced by eye drops containing 0.1% vecuronium bromide, as previously described. ¹⁰ Because it was intended to study how accommodation affects the process of contrast adaptation induced by lens wear (see Fig. 1 ), cycloplegia was induced before the lenses were put on. It had been shown that cycloplegia lasts up to 3 hours, ¹⁰ which ensured that cycloplegia would persist over the 1.5-hour period of lens wear. Before and after the lens treatment, contrast sensitivity was determined (described later). The number of chickens for each of the optomotor experiments varied, because not all animals were found to be cooperative. Those who tended to sleep after they were placed in the drum were excluded (the number of chickens contributing to the data are shown in the figures). 

Twenty-nine chickens wore binocularly frosted occluders for 3 days. After occluder removal, they were either exposed to normal cage experience for 2 hours a day over a period of 4 days, or they were forced to watch a phase-reversing square wave grating or video clips (described later) for 2 hours a day over a period of 4 days. For the remaining time of the day they were in complete darkness. Measurements were performed before occluder attachment, after occluder removal, and after the visual-stimulation period. To investigate the effect of artificial visual stimulation on emmetropization, 10 untreated chickens were exposed to video clips for the same period, and changes in refraction were determined. The effects of controlled low-pass filtering of the retinal image on contrast adaptation, on recovery from myopia, or on refractive development were tested in six, five, and five chickens, respectively. The animals were individually placed between two computer monitors that were set up in a dark room approximately in the direction of the optical axes of the eyes (65° with respect to the rostrocaudal axis ¹¹ ) and at a distance of approximately 5.9 D from their eyes. On both monitors, the same digitally created color video clip (described later) was shown in an infinite loop. Chicks were sitting in a box that restrained their head movements so that the viewing distance was controlled to approximately± 5 mm (equivalent to approximately 0.1 D, at a viewing distance of 20 cm, see Fig. 3A ). Head rotations were not restricted. We did not observe whether they accommodated continuously at the monitors, but this seems unlikely, given that accommodation was observed to be very fast and transient during refraction measurements. The start-up refractive states of the untreated chickens ranged from +2 to +4 D (see Fig. 5 ), which is typical at this age. All chickens were tested between 11 and 18 days of age. 

Techniques

Measurement of Contrast Sensitivity.

The measurement procedure has been described in detail. ⁸ Briefly, individual chicks were placed in a box in the center of a drum with a diameter of 66 cm and a height of 48 cm (Fig. 1 , top). Drifting low-contrast stripe patterns with 8% contrast were presented at a angular speed of 48 deg/sec. The spatial frequencies were 0.2, 0.5, or 1.4 cyc/deg. The stripes were projected on the inner wall of the drum and caused smooth pursuit of the head, interrupted by saccades in the opposite direction (optokinetic head nystagmus). The angular speed of the head rotation was automatically determined by a video image-processing program that followed the movements of the head. It had been found that the ratio of angular head speed to stripe speed (gain) correlated with the suprathreshold contrast sensitivity. Gain approached a maximum between 0.5 and 1 cyc/deg of stripe frequency, and there was no difference at angular stripe speeds between 30 and 60 deg/sec. ⁸ In particular, it was shown that the optomotor gain correlated approximately linearly with the stripe contrast over a range of 8% to 17%. ⁸  

Selectivity of contrast adaptation for the adapted spatial frequency was also tested by adapting to a 0.5 cyc/deg vertical stripe pattern for 10 minutes and measuring contrast sensitivity at 1.4 cyc/deg. The adapting patterns reversed phase at 2 Hz. They were generated on two computer monitors (as in Fig. 3 ) by a short program written in Borland C++, which used the standard graphic tools. Contrast was more than 90%, in that pixel luminance values of 0 and 255 were used at 8-bit gray-scale resolution. 

Prediction of Contrast Adaptation.

The loss of contrast at a given spatial frequency with defocus can be estimated from the modulation transfer function of a diffraction-limited optical system, the first Bessel function. ¹² The loss of contrast at the tested spatial frequency for a given defocus was calculated and compared with the change in gain in the optomotor experiment. 

Generation and Filtering of Digital Video Clips.

Digital video tapes of a group of chickens in their cage environments were recorded (DCR-TRV10E PAL camcorder; Sony, Tokyo, Japan). Two hundred frames (equivalent to 10 seconds of video with the camcorder frame rate of 25 Hz) were individually filtered with a circular aperture with a low-pass–filtered edge, using image-management software (Photoshop; Adobe, Mountain View, CA; see Fig. 3B ), and were recompiled to a digital video file with a 500 × 375-pixel frame size. The video file was played in an infinite loop, using a standard computer video player. The file is available from the authors on request. Because the computer monitors were in the dark and the circular edges of the frames were heavily low-pass filtered, the spatial frequency content of the active video window inside the aperture could be controlled. A version with the full spatial spectrum was generated (referred herein to as “focused video clip,” see Fig. 3B , top) and another version in which each individual frame was low-pass filtered, using the low-pass–filtering function of the software (Photoshop; Adobe; referred to herein as “low-pass–filtered video clip,” see Fig. 3B , bottom). The spatial frequency content of the frames of the two video clips was compared (see Fig. 3C ) using NIH Image, ver. 1.61 (provided in the public domain by the National Institutes of Health, Bethesda, MD, and available at http://www.nb.nih.ncbi.gov). 

Measurements of Refractive State.

Refractive state was measured by automated infrared photoretinoscopy. ¹³  

Statistics

Pre- and posttreatment optomotor gains were compared by paired t-tests. Interindividual and intergroup comparisons were not applied, because the baseline optomotor gain was different in each animal. 

Results

Contrast Adaptation with Positive and Negative Lenses, without and with Cycloplegia

Chickens under cycloplegia that were treated with negative lenses with a refractive power of −7.4 D for 1.5 hours showed a change in optomotor gain from 0.75 ± 0.03 to 0.82 ± 0.04 (P = 0.003; Fig. 1B ). Without cycloplegia, the change in optomotor gain remained very similar (from 0.78 ± 0.02 to 0.85 ± 0.03; P = 0.004; Fig. 1C ). With positive lenses (+6.9 D), the change in optomotor gain was very pronounced under cycloplegia (0.70 ± 0.05 to 0.80 ± 0.07, P = 0.014; Fig. 1D ) but was reduced when accommodation was intact (0.77 ± 0.04 to 0.80 ± 0.03, P = 0.021; Fig. 1E ). 

Symmetry of CA with Lenses of Similar Power but Different Sign

Without accommodation, CA was very similar for negative and positive lenses (increase in gain: 0.06 ± 0.05, with −7.4-D lenses, versus 0.10 ± 0.10, with +6.9-D lenses, respectively; NS; Figs. 1B 1D ). By contrast, with intact accommodation, it was more different, although the difference just missed significance (increase in gain: 0.06 ± 0.03, with −7.4-D lenses, versus 0.03 ± 0.03, with +6.9-D lenses; P = 0.06; Figs. 1C 1E ). 

Prediction of CA from the Loss of Contrast at a Given Spatial Frequency

Using the modulation transfer function of a defocused optical system, the loss of contrast at a given spatial frequency can be predicted from the optical power of the used lens. The chickens were tested with drifting stripe patterns at the respective spatial frequency. Lens powers, tested spatial frequencies, modulation transfer, and measured changes in optomotor gain are shown in Figure 2A . Figure 2B shows that the change in optomotor gain was predicted from the decrease in contrast in the case of negative lenses without cycloplegia. For defocus imposed by negative lenses under cycloplegia, only two data points are available (change in optomotor gain: 58% modulation transfer [MT]: from 0.76 ± 0.03 to 0.82 ± 0.04; P < 0.01; 92% MT: 0.70 ± 0.06 to 0.68 ± 0.10; NS). Nevertheless, these two data points indicate that the function was steeper in the absence of functional accommodation. If the chicken is able to clear the imposed defocus, the actually existing loss of contrast seems to be smaller than the predicted loss of contrast. 

That CA is selective for the adapted spatial frequency was also tested. Chicks were forced to watch high-contrast square-wave patterns of 0.5 cyc/deg for 10 minutes on the computer monitors in the setup shown in Figure 3A . Patterns were phase reversed at 2 Hz. The optomotor gain was significantly reduced by the stimulation if the adapted and tested spatial frequency were identical (0.72 ± 0.07 before and 0.62 ± 0.14 after, n = 8, P = 0.03). However, there was no change in gain if suprathreshold contrast sensitivity was determined at another spatial frequency (1.4 cyc/deg; 0.74 ± 0.05 vs. 0.74 ± 0.05, n = 6, NS). 

Effects of Video Clip Watching on Contrast Adaptation and Recovery from Deprivation Myopia

In line with previous findings, ⁸ it was found that wearing frosted occluders increased the optomotor gain with high significance (Fig. 4A , left, 1 vs. 2: before: 0.55 ± 0.09, after: 0.69 ± 0.07, P = 0.015). Normal visual exposure recovered contrast sensitivity to the start-up value (Fig. 4A , left, 2 vs. 3: gain after recovery: 0.52 ± 0.05, P = 0.0003). If the chickens watched low-pass–filtered video clips during the 4-day recovery period, contrast sensitivity no longer returned to the baseline level (start-up gain 0.66 ± 0.05, after deprivation 0.76 ± 0.04, P = 0.0006; after video watching 0.72 ± 0.03; Fig. 4A , middle, 2 vs. 3, NS). If they were exposed to focused video clips, there was a significant recovery of contrast sensitivity toward the baseline level (start-up gain 0.64 ± 0.07, after deprivation 0.72 ± 0.07, P = 0.0016; after video watching 0.63 ± 0.09; Fig. 4A , right, 2 vs. 3, P = 0.0126). 

Refractive development of chickens that were deprived for 3 days and then allowed to recover is shown in Figure 4B . Three days of occluder wear caused significant deprivation myopia (Fig. 4B 1 vs. 2). Four days of normal visual exposure for 2 hours a day induced recovery to normal refractions (left, 2 vs. 3). If the chickens were exposed to low-pass–filtered video clips during the recovery period, myopia was significantly enhanced (Fig. 4B , middle, 2 vs. 3: −6.48 ± 1.9 D to −9.95 ± 2.56 D, P < 0.001) rather than reduced. In contrast, if they watched focused video clips, they recovered at least by approximately 40% (Fig. 4B , right, 2 vs. 3:− 6.68 ± 2.12 D to −3.18 ± 1.94 D, P = 0.0077). Because the amount of the preinduced deprivation myopia was very similar for the chickens watching the low-pass–filtered video clips and those watching the focused video clip, uncorrected refractive errors could not have had an effect on this difference in refractive development. We observed that if, in the same experimental setup, chickens were stimulated with a square-wave grating at a spatial frequency of 0.2 cyc/deg rather than with a video, they did not recover from deprivation myopia (2 vs. 3: −9.78 ± 2.54 D to −9.6 ± 3.0 D; NS). 

Induction of Deprivation Myopia with Low-Pass–Filtered Video Clips

If previously untreated chickens were forced to watch 2 hours of video per day for 4 days at a 5.9-D distance, they all made a significant shift toward myopia (Figs. 5A 5B , left, 1 vs. 2). However, the amount of myopia induced by the short viewing distance to the computer monitors was low. With spatially unfiltered video clips, the average shift in myopic direction was only− 1.29 ± 0.99 D (Fig. 5B , left, 1 vs. 2, P = 0.003). If the videos were low-pass filtered, the shift toward myopia was increased (Fig. 5A 1 vs. 2: −2.87 ± 1.85 D; difference compared with focused video clip: P = 0.028, unpaired t-test). The results show that some deprivation myopia could be induced by externally controlled spatial object features but that short viewing distances did not cause the expected amount of myopia. 

Discussion

Sensitivity of CA

We found that CA could be sensitive enough to provide a meaningful signal for eye growth. Significant CA was induced by lenses of +6.9 or− 7.4-D power, both with and without functional accommodation. It is known that the chicken eye responds to as little as 1 D of imposed defocus. ¹⁴ Can this small effect also be mediated by CA? Certainly, CA could not be detected in our setup after 1 D of defocus, because additional experiments (results not shown) made clear that the threshold for significant CA with intact accommodation was approximately 4 D of defocus (0.72 ± 0.05 to 0.76 ± 0.04; spatial frequency: 1.4 cyc/deg). Figure 2 shows that a decrease in contrast of at least 40% was necessary at the tested spatial frequency to induce CA in the behavioral paradigm. At 1 D of defocus, a comparable decline would occur only at spatial frequencies higher than 4 cyc/deg, which cannot be generated in our setup for technical reasons. ⁸ In conclusion, nothing argued against the assumption that CA can also mediate growth changes with only 1 D of imposed defocus. The compensatory changes in refraction and eye growth with 7 D of imposed defocus are in the linear range of the response curve of the chicken eye. ¹⁵ Short-term exposure to optical defocus of only 1 D improves visual acuity in humans, a finding supporting our assumption. ¹⁶ After defocus, in the same study, ¹⁶ investigators found no changes in human contrast sensitivity at low and high spatial frequencies and a decrease in contrast sensitivity at midrange spatial frequencies. ¹⁶ However, threshold and suprathreshold contrast sensitivity represent two different functions. In the cited study the former was determined, in the present study, the latter. 

CA with intact accommodation shows clearly that chickens do not completely refocus their retinal images when they wear lenses, a result consistent with observations of Nau et al. ¹⁷ Because there is definitely an error signal present at the retinal level, it is not necessary to assume a role of the accommodative feedback loop in emmetropization to explain refractive changes. 

Set Point of CA and Emmetropization

A surprising finding is that CA is very similar for equally powered lenses of both signs. Given that the average viewing distance of the chicken is probably not at infinity but rather in the cage environment at perhaps 3 D, much more CA would be expected with negative than with positive lenses—in particular, when accommodation is paralyzed. Apparently, the minimum of CA is calibrated for zero refraction (infinity) and not for average viewing distance. With an average viewing distance of 3 D, the eyes with the −7-D lenses should experience 10 D of defocus and those with +7 D lenses only 4 D. If CA were only a response to a decrease in contrast at a given spatial frequency, this asymmetry could not be explained. The decrease in contrast should then be differently weighted for positive and negative defocus. To decide on which side of zero refraction the loss of contrast occurs, some information about the sign of defocus would have to be available at the level of CA. A possible alternative explanation is that CA was in saturation with both types of lenses without the availability of accommodation, in which case no differences in CA would be expected either. 

Another striking observation is that accommodation has less effect on CA with negative lenses than with positive lenses. With intact accommodation, CA seems to be, at a minimum, more close to the average viewing distance than without accommodation. Shifts in the set point of emmetropization have previously been induced by continuous light ¹⁸ or high doses of pirenzepine ¹⁹ or occur in response to optic nerve section. ²⁰ In these cases, the effects of lenses on refractive development are fully preserved but are superimposed to a much more hyperopic baseline (approximately +8 to +10 D). 

The resting refraction of the chicks measured approximately +3 D (Figs. 4 5) but it is likely that the animals were, in fact, emmetropic. The difference can be explained by the small-eye artifact of retinoscopy. ²¹ Its exact magnitude is unknown but, theoretically, it could be as large as 6 D (for a retinal thickness of 230 μm and an anterior focal length of 6 mm). 

Prediction of CA by Loss of Contrast at a Given Spatial Frequency

There was a correlation between the calculated loss of contrast at a given spatial frequency and the CA measured in the optomotor experiment (Fig. 2B) . However, because there were only a few combinations of lens power and spatial frequency, it could be that the correlation is potentially confounded by spatial frequency. It could also be that gain changes occur preferentially for high spatial frequencies. 

If the chickens were adapted to a certain spatial frequency, CA was not detected at another spatial frequency, indicating that CA is spatial frequency selective. This finding is in line with those in previous psychophysical studies. ^{22 23 24}  

Correlation between CA and Myopization in Recovery Studies with Video Clips

The correlations between CA and induction of and recovery from deprivation myopia with artifical stimulation were consistent with the idea that CA is an error signal for changes in eye growth and refraction. The experiments with the video clips are the first demonstration that eye growth can be controlled by an artificially produced and externally controlled spatial frequency spectra. The technique also permits study of the role of motion in emmetropization. An uncertain variable at present is the level of accommodation that the chicks maintain when they look at the video clips. Because the accommodation response of the chicks was observed to be very fast and transient during measurements of their noncycloplegic refraction, it seems likely that the same was true for their accommodation during the video stimulation. Nevertheless, no matter whether the video was low-pass filtered or not, the chicks all showed a shift toward myopia (Fig. 5) , indicating that the eyes responded also to the proximity of the target and, hence, had the distance information. The effect of low-pass filtering on refractive development was superimposed. Refractive development is then hard to explain by differences in accommodation for both types of video clips. 

The animals did not completely adapt their refractions to the distance of the computer monitors during video stimulation. This finding is in line with previous data showing that artificial restriction to short viewing distances does not produce the expected amount of myopia (chicken ²⁵ and monkeys ²⁶ ). However, it is well documented that accommodation is also incomplete. ^{27 28 29}  

Conclusion

The data are in line with the idea of a role for CA in emmetropization. However, our experiment cannot rule out that CA is just an epiphenomenon of defocus with no role at all in information processing for eye growth control. In addition, two issues remain unresolved. 

First, where does CA occur? During recent years, there has been increasing evidence that not only cortical but also retinal mechanisms contribute to CA. ^{5 6 7} A promising candidate for the site of retinal CA is the bipolar pathway. ^{6 30} To guide emmetropization, CA should be retinal. A direct demonstration by Heinrich and Bach of a retinal contribution in humans was recently successful. ⁷ Their pattern ERG (PERG) recordings showed significant shifts in the PERG phase after adaptation to low contrasts. Another approach is to study the function by which CA changes with contrast. They also found that the PERG displays a linear contrast-transfer function (amplitude versus linear contrast), whereas the visual evoked potential (VEP) displays a sigmoid contrast-transfer function (amplitude versus log contrast). With more data on CA with different adaptation contrasts, a decision can be made about which function provides a better match to the data in the present experiments in chickens. 

Second, the contrast of objects changes continuously with ambient illuminance and object features. Accordingly, contrast sensitivity is adapted to the respective conditions. It would not make sense to control eye growth with illuminance-dependent changes in CA. The difference between defocus-induced and illuminance-induced CA is the spatial frequency response function. Therefore, to get a useful signal for emmetropization out of CA, the contrast at different spatial frequencies must be compared. 

 Supported by the German Research Council Grant Scha 518/1-11, the Max-Planck Award, and Fortune Program Project 532-0-0 of the University of Tübingen.

 Submitted for publication March 8, 2001; revised June 18, 2001; accepted July 26, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

(A) The optomotor drum used to measure suprathreshold contrast sensitivity of the chickens. Drifting square-wave patterns were projected at the wall inside a large drum. The head-turning movements of individual chicks were recorded by a video camera, a frame grabber, and an image-processing program that tracked two reflectant dots on the head of the chicks at a 25-Hz sampling rate. The optomotor gain is a measure of contrast sensitivity. After negative (B, C) or positive (D, E) lens wear for 1.5 hours, the optomotor gain was increased. The increase, referred to as contrast adaptation (CA), was also significant if the chicks were not under cycloplegia and could have refocused their retinal images by accommodation. CA was reduced when accommodation was intact and the chicks wore positive lenses.

Figure 2.

 

(A) Calculated modulation transfer for a given lens power and spatial frequency and comparison with the measured changes in optomotor gain of the chickens after they had worn the lenses for 1.5 hours. (B) Correlation between the loss of contrast in the retinal image with negative lenses (as predicted from the modulation transfer function) and the induced change in contrast sensitivity. Lenses were worn for 1.5 hours; accommodation was intact. Lens powers and tested spatial frequencies are indicated in the figure. Each data point was an experiment with an individual chick. cycl, under cycloplegia.

Figure 3.

 

(A) Setup for presenting spatially filtered video clips to the chickens; viewing distance, 5.9 D. The computer monitor frames were covered by black cardboard and the whole setup was in the dark. Chickens were individually placed in a small box midway between both monitors for 2 hours each day. To generate the data shown in Figures 4 and 5 , the same videos were shown on both monitors for 2 hours each day for a period of 4 days. Chicks were in their cages in the dark for the remaining time. Lateral head movements were restricted to approximately 0.1 D (5 mm), but head rotations were unrestricted. (B) Appearance of the unfiltered (focused) video frames (top) and the low-pass–filtered video frames (bottom). (C) Comparison of the spatial frequency spectra of the focused and low-pass–filtered video clips, as determined by computer (NIH Image, ver. 1.61).

Figure 4.

 

(A) Changes in optomotor gain induced by wearing frosted occluders for 3 days (1 vs. 2) and its reversal (2 vs. 3) by normal visual exposure in the cages for 2 hours a day (left), watching low-pass–filtered video clips for 2 hours a day (middle), and watching focused video clips for 2 hours a day (right) after removal of the occluders. Before and after the stimulation periods, the animals were in the dark. There was no significant recovery of CA if the low-pass–filtered video clips were presented (middle), but both normal visual environments and focused video clips caused significant reversal of the changes in optomotor gain and, hence, reversal of CA. (B) Changes in refraction under the same conditions as in (A). Deprivation myopia increased if the low-pass–filtered video clips were presented, but at least some recovery occurred with presentation of focused video clips. 1, start-up gain-refraction; 2, gain-refraction after 3 days of occluder wear; 3, gain-refraction after 4 days of visual stimulation for 2 hours a day after removal of the occluder.

Figure 5.

 

Induction of myopia by video viewing at a 5.9-D distance. (A) With low-pass–filtered video clips, the chicks became approximately 3 D more myopic (left: group data; right: individual data). (B) Chicks also became somewhat more myopic with the focused video clips, but this refractive change is attributable to the short viewing distances. However, they became significantly more myopic with the low-pass–filtered than with the focused video clips. 1, start-up gain-refraction; 2, gain-refraction after 4 days of visual stimulation.

The authors thank Michael Bach for helpful discussions.