Temporal Properties of Compensation for Positive and Negative Spectacle Lenses in Chicks

Abstract

purpose. Chicks’ eyes rapidly compensate for defocus imposed by spectacle lenses by changing their rate of elongation and their choroidal thickness. Compensation may involve internal emmetropization signals that rise and become saturated during episodes of lens wear and decline between episodes. The time constants of these signals were measured indirectly by measuring the magnitude of lens compensation in refractive error and ocular dimensions as a function of the duration of episodes and the intervals between the episodes.

methods. First, in a study of how quickly the signals rose, chicks were subjected to episodes of lens-wear of various durations (darkness otherwise), and the duration required to cause a half-maximum effect (rise-time) was estimated. Second, in a study of how quickly the signals declined, various dark intervals were imposed between episodes of lens-wear, and the interval required to reduce the maximum effect by half (fall-time) was estimated.

results. The rise-times for the rate of ocular elongation and choroidal thickness were approximately 3 minutes for positive and negative lenses. The fall-times had a broad range of time courses: Positive lenses caused an enduring inhibition of ocular elongation with a fall-time of 24 hours. In contrast, negative lenses caused a transient stimulation of ocular elongation with a fall-time of 0.4 hour.

conclusions. The effects of episodes of defocus rise rapidly with episode duration to an asymptote and decline between episodes, with the time course depending strongly on the sign of defocus and the ocular component. The complex etiology of human myopia may reflect these temporal properties.

During early postnatal life, ocular growth is modulated by the visual input from the retina, resulting in the correction of refractive errors (reviewed in Refs. ^{1 2 3 4} ). Eyes use two compensatory mechanisms to reduce defocus: When the eye wears a positive lens, which would put the images of distant objects in front of the photoreceptors (myopic defocus), it slows its rate of elongation and thickens the choroid, pushing the retina forward toward the image plane. When the eye wears a negative lens, which would put the images of distant objects behind the photoreceptors (hyperopic defocus), it accelerates its rate of elongation and thins the choroid, pulling the retina back toward the image plane. Compensation by changing the eye size has been found in fish, ⁵ chicks, ^{6 7} guinea pigs, ⁸ tree shrews, ^{9 10} cats, ¹¹ marmosets, ¹² and rhesus monkeys. ¹³ Compensation by changing choroidal thickness has been shown in chicks, ¹⁴ tree shrews, ⁹ guinea pigs, ¹⁵ marmosets, ¹⁶ and macaque monkeys. ¹⁷  

In normal life, of course, the retina does not exclusively experience continuous myopic or hyperopic defocus. Rather, most parts of the retina experience frequent episodes of myopic and hyperopic defocus depending on the spatial layout of the environment, the distance of the objects viewed, and the eye’s refraction and its accommodative state. How does the retina sum these episodes of defocus over time to determine the direction of the eye’s growth toward emmetropia? 

The simplest model of integration would be that the defocus is summed linearly (i.e., the resulting compensation is proportional to the average amount of defocus that the retina experiences summed over a period; see Fig. 1A ). If the eye were experiencing myopic defocus, for example, the emmetropization signal favoring hyperopic growth would keep rising as long as the defocus is present. When there is no defocus (e.g., in darkness), the signal would remain stable. When the sign of defocus reverses, the signal would switch direction. 

Despite its appealing simplicity, there is evidence against this linear model. When the chicks had the same total amount of lens wear (28 minutes) each day given in various numbers of episodes with darkness between, ^{1 18} many episodes of lens-wear (14 episodes, 2 minutes each) were more effective than a single episode (of 28 minutes) in causing compensation for either positive or negative lenses, in agreement with findings that several episodes of normal vision each day were more effective than a single episode of the same total duration in reducing form-deprivation myopia in chicks. ¹⁹ These results imply either that the hypothesized internal emmetropization signal reaches saturation within an episode or that it declines between episodes. However, very brief episodes were ineffective at causing compensation, even if repeated very often (2-second episodes given every 2 minutes for a total duration of 28 minutes), suggesting that the signal must accumulate to threshold within each episode before downstream effects are initiated. 

In light of the evidence, we hypothesize that episodes of myopic and hyperopic defocus are integrated in a nonlinear fashion, as proposed by Flitcroft ²⁰ : Lens wear causes one or several internal emmetropization signals to rise and eventually become saturated during episodes of lens wear and decline slowly between episodes (Fig. 1B) . 

To begin exploring this hypothesis, we studied the time constants of the signals regulating compensation for positive and negative lenses. Because we do not know the nature of these signals, we cannot measure them directly, but can measure their effects on the rate of ocular elongation and choroidal thickness. Because a single brief episode of lens wear would not generally have a measurable effect, we estimated the rise and fall-times of these inferred signals by measuring the eye before and after 3 days of many episodes of viewing through positive or negative lenses: (1) We estimated how quickly the hypothesized signals rise during each episode by giving frequent episodes of lens wear of a range of durations, reasoning that the ocular responses will increase with increasing duration up to an asymptote. (2) We estimated how quickly the signals decay between episodes by giving long (saturating) episodes of lens wear with a range of intervals between the episodes, reasoning that the ocular responses will decline with increasing intervals. 

We find that the signals for the effects of both positive and negative lenses on both the rate of ocular elongation and choroidal thickness rise at a similar rate (within minutes), but they decline at slower, very different rates, with the signal regulating the rate of ocular elongation declining the slowest in the case of positive lens wear, and the fastest in the case of negative lens wear. Some of these results have been presented in an abstract (Zhu X, et al. IOVS 2004;42:ARVO E-Abstract 4285). 

Materials and Methods

Animals

White Leghorn chicks were obtained from either Cornell University (Cornell K-strain; Ithaca, NY) or Truslow Farms (Hyline-W98-strain; Chestertown, MD; group 14 and part of groups 8, 11, and 13); because we did not find any statistically significant differences between the strains in the degree of compensation, results were pooled for analysis. Before the experiments started, chicks were reared in heated brooders under fluorescent lighting (lights off from 10 PM to 8 AM). During the experiments the chicks were housed in a heated, sound-attenuated chamber (76 × 61 cm), with the lights off from 8:00 to 8:30 PM (depending on the experimental group) to 8 AM. The timing of illumination for the experimental groups is shown in Table 1 . The birds were kept in darkness between episodes. Food and water were provided ad libitum. Care and use of animals adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Experiments started when the chicks were 1 week old and lasted for 3 days. 

Lenses

PMMA plastic lenses (with a back optic radius of 7 mm) or glass lenses (not conspicuously curved) of +7 and −7 D were used. The lenses (diameter = 12 mm) were glued between rigid plastic rings and Velcro rings attached to mating Velcro rings glued to the feathers around the chicks’ eyes. Lenses were cleaned at least twice daily. In both experiments, the chicks had monocular lens wear, leaving the other eye uncovered. 

Measurements

Measurements of refractive error and ocular dimensions were conducted on chicks anesthetized with 1.5% halothane or isofluorane. The birds were measured at the same time of day before and after the 3 days of lens wear. Refractive error was measured with a modified Hartinger refractometer. ²¹ A-scan ultrasonography was used to measure internal ocular dimensions. ²² Ocular length was measured from anterior cornea to posterior sclera (the sum of anterior chamber depth, lens thickness, vitreous chamber depth, and the thickness of retina, choroid, and sclera), with appropriate sound velocities used for each ocular component. ²¹ This method differs from conventional axial length measurement, from cornea to retina, which is affected by changes in choroidal thickness and therefore confounds ocular length with choroidal thickness, which, as we will show, are controlled separately. 

Protocols

Treatments and lens-wearing paradigms for experiments 1 and 2 are listed in Table 1 . 

Experiment 1: Estimation of Rise-Time.

To characterize how fast the hypothesized internal emmetropization signals rise, we estimated the episode duration that produced lens compensation half as effective as continuous lens wear (called the rise-time), by giving the birds repeated episodes of light with lens wear for various durations, with darkness in between (see Fig. 2Afor the schematic and Table 1for details). Chicks wearing lenses were given episodes of various durations either every 10 min (those wearing −7-D lenses and groups 1 and 5 of those wearing +7-D lenses) or every hour (all other +7-D lens groups). We used the more frequent intervals for the birds wearing −7-D lenses because we found, in a preliminary experiment, that even the fellow eyes to those wearing −7-D lenses became myopic (−2.9 D) with thinned choroids (−139 μm over 3 days) when the interval was 1 hour, thereby limiting the additional thinning that the negative lenses could produce. Two control groups of chicks wore positive (group 8) or negative (group 13) lenses continuously under a normal 14:10 light–dark cycle. 

Experiment 2: Estimation of Fall-Time.

To characterize how fast the hypothesized internal emmetropization signals fall, we studied the interval in the dark between episodes that produced lens compensation half as effective as continuous lens wear (called the fall-time), by imposing long (30 minutes) episodes of light with +7 or −7 D lens wear with various dark intervals between (see Fig. 2Bfor the schematic and Table 1for details). Since experiment 1 showed that the rise-times for both positive and negative lenses were 5 minutes or less, we imposed 30-minute episodes of viewing through lenses on the chicks to ensure a saturating effect of each episode, with various dark intervals from 0.5 hour (one episode every hour) to 47.5 hours (one episode every 2 days). When only one episode of lens wear was imposed every day, it took place at the same time each day (between 12 and 2 PM). When two daily episodes were imposed, they took place from 8:00 to 8:30 AM and 8:00 to 8:30 PM, resulting in a dark interval of 11.5 hours. When more frequent episodes were imposed, the first was at 8 AM the last at 8 PM. 

Statistics and Calculations

Changes in experimental eyes and fellow eyes over the 3-day-long experiments are summarized in Table 1 . Because we were specifically interested in the changes during the course of the treatment, we used relative change (the change in the experimental eye over the course of the experiment minus the change in the fellow eye) to represent the net effect of the lens wear. The relative changes that were significantly different from 0 by paired, two-tailed Student’s t-tests are shown in Table 1 . Relative change is plotted against either the duration of the lens-wearing episodes for rise-time in experiment 1 (Fig. 3)or the dark interval between lens-wearing episodes for fall-time in experiment 2 (Fig. 4) , both on logarithmic scales, so that different time points are nearly evenly separated. Analysis of variance (ANOVA) with LSD post hoc tests was used to compare the relative changes among various episode durations (for experiment 1) or dark intervals between episodes (for experiment 2). 

Data of the individual birds of each experiment were fitted (using Igor Pro version 5.02; WaveMetrics, Inc, Lake Oswego, OR) with a sigmoidal curve of the form y = y _o + Δy/(1 + exp^{(x ₀−x)/z} ), with x being the logarithm of the episode duration or interval. The coefficient y _o is the y value at small x, (y _o + Δy) is the y value at large x, x ₀ is the log of the x value at which y is midway between the low and high y asymptotes (i.e., the rise- or fall-times), and z is the rate of rise or fall. The 95% confidence intervals (CIs) for x ₀ were also calculated. In some cases, the unconstrained curve-fitting yielded implausible confidence intervals. Therefore, in all cases, we performed three successive fits to yield x ₀, the primary parameter of interest: First, we limited the ranges of y _o and Δy and let the algorithm fit these 2 parameters to yield the fitted values for y _o and Δy; second, we fixed y _o and Δy at the their fitted values and let the algorithm fit the optimal value for z; and third, we fixed y _o, Δy, and z at their fitted values and let the algorithm obtain the value for x ₀ and its 95% CI. Because x ₀ was calculated as a logarithm, the 95% CI was asymmetric when converted into minutes or hours. 

Results

The findings presented here support our hypothesis that lens wear causes rapidly saturating effects that decay during subsequent darkness. When the chicks viewed the world through defocusing lenses for episodes lasting 1 minute or less, no effect was seen, but with episodes lasting a few minutes the effects rose rapidly and became as great as if the chicks had worn the lenses continuously in normal illumination. When long episodes of viewing through lenses were followed by intervals of darkness, the effects differed dramatically in their persistence: The inhibition of ocular elongation caused by positive lens wear persisted for many hours, whereas the stimulation of ocular elongation caused by negative lens wear was gone after only half an hour of darkness. For simplicity, we will present the main results in terms of the changes over the experiment in the lens-wearing eye minus the change in the fellow eye, termed relative change. The changes in ocular length, choroidal thickness, vitreous chamber depth, and refractive error for lens wearing and fellow eyes are summarized in Table 1 . 

Experiment 1: Estimation of Rise-Time

As expected, compensation by changes in ocular elongation (Fig. 3A)and choroidal thickness (Fig. 3B)was not simply a linear function of the duration of lens wear, but rather was a sigmoidal function. The rise-times were very similar for both ocular elongation and choroidal thickness when either positive or negative lenses were worn, ranging from 1.3 to 4.4 minutes. 

Ocular Length.

The degree to which the rate of ocular elongation was affected by lens wear was determined by the episode duration: When viewing through +7-D lenses was given 10 seconds every 10 minutes (group 1, Table 1 ), little ocular compensation was found (Fig. 3A) ; vision for 30 seconds or 1 minute every hour (groups 2 and 3) resulted in small but consistent ocular inhibition, and further increasing the duration to 2 minutes or more every hour caused the maximum inhibition of ocular elongation, like that with continuous positive lens wear (relative change: −183 μm vs. −195 μm, group 4 vs. 8). Durations of 2 minutes or longer showed a significantly greater inhibition than durations of 1 minute or shorter (P < 0.05, ANOVA with LSD post hoc test; Fig. 3A ). The rise-time for ocular inhibition caused by wearing positive lenses was 1.3 minutes (95% CI: 0.8–1.9 minutes), whether we used the data from the hourly episodes (group 4), or the every-10-minute episodes (group 5), suggesting that the lens-wearing frequency did not have a big impact on the rise-time. 

Viewing through −7-D lenses for 10 seconds or 1 minute every 10 minutes (groups 9 and 10) caused 30% and 45% of the maximum increase in the rate of ocular elongation seen in continuous negative lens wear (group 13), respectively; further increasing the duration of lens wear to 5 minutes caused the maximum increase in the rate of ocular elongation. The two longest durations (5 and 10 minutes) had a significantly greater effect than did 10 seconds (relative change, 5 or 10 minutes vs. 10 seconds; P < 0.05, ANOVA with LSD post hoc test). The rise-time for ocular elongation caused by negative lens wear was 1.6 minutes (95% CI: 0.5–5.2 minutes). (Curiously, viewing through −7-D lenses for 2 minutes every 10 minutes [group 11] had no effect on the rate of ocular elongation in two separate experiments with a total of 15 birds. The negative lens-wearing eyes elongated less than the fellow control eyes in 10 of 15 birds [mean ± SEM: −7 ± 22 μm, relative change]. Including this group in the curve-fitting changed the rise-time to 3.7 minutes [95% CI: 2.5–5.4 minutes; Fig. 3A , dashed line]). 

Choroidal Thickness.

Similar to their inhibition of ocular elongation, short durations of viewing through positive lenses caused little choroidal expansion (10 seconds to 2 minutes, groups 1–4), but a longer duration (5 minutes, group 6) caused a robust choroidal expansion (relative change, 58 μm; Table 1 , Fig. 3B ). When the duration was further increased to 10 minutes, choroidal thickening reached the maximum (group 7, 132 μm relative change), significantly more than that found when durations were 2 minutes or shorter (P < 0.01, ANOVA with LSD post hoc test, see Fig. 3B ). The rise-time for choroidal thickening was 4.4 minutes (95% CI: 2.3–8.4 minutes) with the data from hourly episodes (group 4), or 5.0 minutes with the data from the every-10-minute episodes (group 5). 

For negative lens wear, although the shortest duration did not cause any choroidal thinning (10 seconds/10 minutes, group 9), longer durations (1–5 minutes, groups 10 to 12) caused from 25% to 73% of maximum choroidal thinning (group 13). The rise-time of choroidal thinning with negative lens wear was 2.9 minutes (95% CI: 1.1–8.2 minutes). 

In summary, the signals modulating both the rate of ocular elongation and choroidal thickness for both positive and negative lens wear rose at a similar speed, with the rise-time ranging from 1.3 to 4.4 minutes. No significant difference was found. 

The rise-times for vitreous chamber depth and refractive error were both less than 4 minutes for both positive and negative lens wear (data not shown). The sum of the changes in vitreous chamber depth and in choroidal thickness correlated well with the changes in ocular length (slope = 0.92, r ² = 89.7% for relative change). No difference was found between paired eyes for either anterior chamber depth or lens thickness. 

Experiment 2: Estimation of Fall-Time

Similar to the rise-times found in experiment 1, the decay of the signals controlling the rate of ocular elongation (Fig. 4A)and choroidal thickness (Fig. 4B)was a sigmoidal function of the dark interval between lens-wearing episodes. Unlike the rise-times, the signals declined not only more slowly, but also with dramatically different patterns, depending on the sign of defocus and the ocular component. The fall-times ranged from 0.4 hour to 24.4 hours. Episodes of positive lens wear had a more enduring effect on both the rate of ocular elongation (fall-time 24.4 hours) and choroidal thickness (6.7 hours) than did episodes of negative lens wear (fall-times for the rate of ocular growth and choroidal thickness: 0.4 and 3.2 hours, respectively). 

Ocular Length.

One of the most striking findings was that the signal regulating the rate of ocular elongation with positive lenses was very enduring: Viewing through positive lenses for 30 minutes with dark intervals from 1.5 to 11.5 hours all caused more than 90% of the maximum inhibition of ocular elongation found with birds wearing positive lenses continuously (Fig. 4A ; groups 14–17). When the dark interval was increased to 23.5 hours (i.e., one episode per day), ocular inhibition was still 70% of the maximum (group 18), significantly less than the inhibition with the dark intervals 3.5 hours or shorter (P < 0.05, ANOVA with LSD post hoc test; Fig. 4A ). Even viewing through positive lenses once every 2 days (dark interval, 47.5 hours, group 19) still caused about one third of the maximum inhibition, significantly less than the inhibition with most of the dark intervals of 23.5 hours or less (P < 0.05, ANOVA with LSD post hoc test, see Fig. 4A ). The fall-time for the rate of ocular elongation with positive lenses was 24.4 hours (95% CI: 20–30 hours). 

In contrast to the enduring signal found in positive lens wear, the signal for negative lens wear was transient: Dark intervals of only 0.5 hour significantly reduced the rate of ocular elongation by two thirds, compared with continuous lens wear (group 20 vs. 13; P < 0.01, ANOVA with LSD post hoc test); longer dark intervals eliminated the effect of the lens wear (all groups were significantly less than that with continuous lens wear or with a 0.5-hour dark interval). The fall-time for the rate of ocular elongation with negative lenses was 0.4 hour (95% CI: 0.2–0.9 hour). This fall-time was so much shorter than that for positive lens wear that even their 99% CIs do not overlap. 

Choroidal Thickness.

Viewing through positive lenses every 2 to 6 hours (dark intervals, 1.5–5.5 hours; groups 14–16) caused approximately 80% of maximum choroidal expansion, but with dark intervals 11.5 hours or longer (groups 17–19), no choroidal expansion was evident (relative change, continuous lens wear, 3.5 or 5.5 hours vs. 11.5 or 27.5 hours; P < 0.05, ANOVA with LSD post-hoc test). The fall-time for choroidal thickening was 6.7 hours (95% CI: 5.0–8.8 hours). 

Viewing through negative lenses every 4 hours or less often (dark intervals, 3.5 hours or longer; groups 22–24) reduced choroidal thinning, but, not significantly different from viewing through negative lenses continuously. The fall-time for choroidal thinning was 3.2 hours (95% CI: 1.8–5.4 hours), a value not significantly different from the fall-time for positive lens wear. 

Given that the fall-time was greater with positive than for negative lens wear for both ocular length and choroidal thickness (Fig. 4) , it is not surprising that the fall-times for vitreous chamber depths followed this same pattern, with the fall-time of 17.2 hours (95% CI: 14.4–20.6 hours) for positive lens wear, significantly longer than that for negative lens wear (2.2 hours, 95% CI: 1.7–2.9 hours). On the other hand, for refractive error, the fall-time for negative lens wear (17.4 hours, 95% CI: 9.3–32.5 hours) was longer than the fall-time for positive lens wear (5.5 hours, 95% CI: 5.3–5.8 hours), although we are inclined to doubt this finding because the confidence interval for the negative lens wear was much larger than that for any of our other measurements. 

Again, the sum of the changes in vitreous chamber depth and in choroidal thickness correlated well with the changes in ocular length (slope = 0.93, r ² = 87.4% for relative change). No difference was found between paired eyes for either anterior chamber depth or lens thickness. 

Discussion

Our results are consistent with the existence of emmetropization signals that increase to saturation during episodes of defocus and then decline between episodes. Furthermore, there seem to be separate signals for the ocular elongation and choroidal components of lens compensation, as well as differences between positive and negative lenses, suggestive of separate signals for myopic and hyperopic defocus. The signals regulating the rate of ocular elongation and choroidal thickness with both positive and negative lens wear rise at a similar rate (1–4 minutes required to increase it by 50%), and fall at drastically different rates: The signal for inhibition of ocular elongation by positive lens wear is the most enduring (a 24.4-hour dark interval is necessary for a 50% decrease), and the signal for acceleration of ocular elongation by negative lens wear is the most transient (only a 0.4-hour dark interval is necessary for a 50% decrease); the signals controlling choroidal thickness are intermediate. 

Relation to Previous Studies

The results presented here generally confirm the results from a previous study in which only a few durations and intervals were tested (plotted as a and b in Figs. 3 4) . ¹⁸ Specifically, the fall-times for positive lenses were nearly identical with respect to both choroidal thickness and ocular elongation, although our results yield both rise-times for positive lenses and fall-times for negative lenses that are somewhat longer for choroidal thickness and slightly shorter for ocular elongation than those of Winawer and Wallman. ¹⁸ Our results also confirm those of another study, ²³ in which chicks were given a single 10-minute episode of positive lens wear, which produced robust choroidal expansion, consistent with the 4.4-minute rise-time reported in our study. This choroidal expansion lasted for 6 hours, consistent with the 6.7-hour fall-time reported in our study. 

If we compare our results with those of a previous study in which chicks had normal visual experience between the episodes of positive lens wear (Figs. 3 4 , triangles), ²⁴ we find that the rise-time of choroidal expansion was identical, but the fall-time seemed a bit longer than in the birds with intervals of normal vision. Unfortunately, the changes in ocular length in that study were so small (because of the preponderance of normal vision) that one cannot infer the rise- and fall-time from the two points of data available. 

Our finding that the effect of negative lenses is very transient whereas that of positive lenses is surprisingly enduring may explain the temporal asymmetries between wearing positive and negative lenses observed by others. First, only 1 to 4 hours of unrestricted vision each day completely cancelled out the effect of wearing negative lenses or diffusers the remainder of the day in chicks, ^{25 26} tree shrews, ²⁷ and monkeys ^{28 29} (see Ref. ²⁸ for a comparison of the effect of unrestricted vision on myopic development across species). Second, wearing positive lenses for just 1 hour each day, with unrestricted vision the rest of the day, still caused hyperopic development. ^{24 26}  

Furthermore, although brief periods of wearing either positive or negative lenses produced approximately equal amounts of compensation, alternating positive and negative lens wear or alternating myopic and hyperopic defocus favored the compensation for positive lenses ¹⁸ or myopic defocus, ³⁰ even if the duration of positive lens wear (four 2-minute episodes in chicks ²⁴ ) was much shorter than that of negative lens wear (the remainder of the day). A similar, but less consistent, effect was caused by an episode of 45 minutes of positive lens wear in tree shrews. ³¹ And when hyperopic and myopic defocus is alternated rapidly by an optical trick in chicks, the balance between the efficacy of the myopic and hyperopic defocus depends strongly on the frequency of alternation: At low frequencies, the myopic defocus dominates; at high frequencies (2 Hz), the myopic and hyperopic defocus is equally effective. ³⁰ Our findings that the inferred emmetropization signals of both components of compensation for positive lens wear have longer fall-times than those for negative lens wear (in case of ocular length, 50 times as long) may help explain why positive lens wear can so potently reverse the effects of negative lens wear. 

Separate Signals Regulating the Rate of Ocular Elongation and Choroidal Thickness

Our findings add to the evidence that choroidal expansion is controlled separately from ocular elongation, in that for negative lenses the fall-time for choroidal thinning is eight times that of ocular elongation, and for positive lenses the fall-time for choroidal thickening is one quarter that of the inhibition of ocular elongation. Furthermore, there are conditions under which we find robust compensation for one of the components and no compensation for the other. For example, with 30-minute episodes, twice per day (dark interval, 11.5 hours, group 17), only inhibition of ocular elongation (by 204 μm), but not choroidal thickening, occurred in the case of positive lens wear, whereas the opposite was found in the case of negative lens wear (a substantial choroidal thinning of 54 μm, without ocular elongation, group 23). Therefore, it appears that ocular and choroidal compensation are driven by different signals. 

The dissociation of ocular and choroidal compensation has been reported previously: (1) brief periods of strobe light at dawn and dusk in eyes wearing translucent diffusers ^{32 33} and brief periods of light during the night in eyes wearing negative lenses ³² reduce the rate of ocular elongation with little effect on choroidal thinning; (2) twice daily 30-minute episodes of spectacle lens wear (with darkness in between) produces an inhibition of ocular elongation without choroidal thickening in the case of positive lens wear, and choroidal thinning without ocular elongation in the case of negative lens wear, ¹⁸ similar to our findings; (3) interrupting all-day-long binocular negative lens wear with brief periods of positive lens wear on one eye and plano lens wear on the other eye produces greater inhibition of ocular elongation in the positive-lens wearing eye without any choroidal thickening, ²⁴ ; and (4) wearing a weak diffuser on top of a positive lens reduces choroidal expansion and enhances the inhibition of ocular elongation, compared with having the fellow eye wear a positive lens alone. ³⁴  

Nickla et al. ³⁵ have shown that the inhibition of ocular elongation in several different visual conditions is always accompanied by a transient choroidal thickening. Our present findings that viewing through positive lenses separated by long intervals of darkness can cause ocular inhibition without choroidal thickening are not necessarily in conflict with these results, because transient choroidal expansions may have occurred with no cumulative effects. 

Possible Mechanisms Underlying the Range of Time Courses

It is striking that only minutes of vision are necessary to produce near-maximum responses to either sign of lens and that the responses last at least 10 times as long in darkness. We speculate that the rise-time is set by retinal processes, possibly with the same processes affecting changes in both ocular elongation and choroidal thickness. Once the changes are initiated, however, they may set into motion in other tissue layers processes whose endurance is determined by different mechanisms for each response. Thus, we can imagine that the stimulation of ocular elongation may be under the control of growth factors, such as TGFβ, ³⁶ the biological availability of which could be rapidly modulated, yielding short fall-times, while initiating the transcription of the genes for the enzymes that produce a growth-inhibiting molecule like retinoic acid ³⁷ or glucagon ³⁸ may inhibit ocular elongation either directly in the sclera or indirectly via other tissues long after the initiating molecule was gone. If the initiating molecule also initiated choroidal changes, it could result in one molecule’s causing effects with two different time courses. Furthermore, the expansion of the choroid may involve active transport of osmotically active molecules into the lacunae of the choroid, whereas the thinning may result from either control of the lymphatic valves from lacunae to veins or by passive transport of water and ions out of the lacunae, which may have a very different time course. 

Effects of Darkness

Because the chicks in this study were kept in darkness for much of the time, we should discuss whether the darkness vitiated our results. To be sure, the ideal way to assess the temporal characteristics of the lens wear would be to keep the birds between episodes in a neutral condition that had no influence on the ocular parameters. Unfortunately, darkness is the condition closest to neutral that we are aware of. In the case of positive lenses, it is possible to see effects of brief episodes of lens wear when these episodes are separated by periods of normal vision, ²⁴ although the effects are much smaller than those shown. As mentioned in the Results section, from that study we could estimate the effect on the choroid rise-time, which was the same as measured in our study, and on the choroid fall-time, which may be a few hours shorter. For negative lenses, however, even brief periods of normal vision would eradicate the effect of the episodes of lens wear. Thus, we can estimate the effect of the darkness on the fellow eyes, but if darkness interacted with the time course of the response to the lenses, it would not be discernible in our results. 

In spite of this obstacle, we estimated the effect of darkness, per se, dissociated from the effect of the lenses, by comparing the fellow eyes of the different groups of experiment 2 (Fig. 5A) . (Changes in the fellow eyes in experiment 1 are comparable to those in eyes in experiment 2.) Clearly, the choroids are progressively thinned in longer periods of darkness. This thinning may limit the extent of further choroidal thinning with negative lenses, but would be less of a problem for eyes wearing positive lenses, which cause the choroids to expand. We are encouraged by the fact that in the birds wearing negative lenses with the longest dark interval (47.5 hours, group 25), the fellow eyes thinned by only 23 μm. Nonetheless, this dark-induced thinning may be responsible for the fact that, even after subtracting the changes in the two eyes, the choroid thinning did not fall to zero with long intervals between episodes (Fig. 4B) . On the other hand, long periods of darkness caused only a small increase in the rate of ocular elongation (Fig. 5B) . 

Effects of Other Sources of Variation

In addition to responding to the lenses we imposed, the eyes may have adjusted their choroidal thickness and ocular length as part of their natural emmetropization process. Thus, eyes that had been myopic or hyperopic before our experiment may have thickened or thinned the choroids in response and may be readjusting the choroids back to normal thickness (Fig. 6) . Not surprisingly, this effect was particularly robust when the birds were kept in the darkness between episodes (slope = −0.8, filled symbols in Fig. 6 ), and significantly weaker when the birds were given light between episodes and thus had stronger visual influences on choroidal thickness (slope = −0.6, open symbols, data from Ref. ²⁴ ); the slopes of these two groups are significantly different). A similar normalizing was seen with respective to refractive error (data not shown), although the starting choroidal thickness and refractive error did not correlate (r ² = 6.9%). 

Furthermore, even though it is well established that, although the two eyes of chicks respond substantially independently to lens wear, there is some degree of yoking of the two eyes, such that the choroidal thickness and rate of ocular elongation of the fellow eye changes in the same direction, but to a much smaller degree (10%–30%, in our data), as does the lens-wearing eye (Figs. 5C 5D) . ³⁹  

As an estimate of the relative contribution of these two factors, plus that of darkness, on the choroidal thickness of the fellow eyes, an ANOVA indicated that hours of darkness account for much less of the variance (F = 15) than does the initial thickness of the choroids (F = 111), but more than the changes in choroidal thickness of the lens-wearing eyes (F = 5.1). 

In the course of this study, we discovered a new form of yoking between the eyes: Although there is a positive correlation between the fellow eye and the lens-wearing eye, the mean of the fellow eye changes in the opposite direction from the mean of the lens-wearing eye. Thus, although the choroids are, of course, thicker in eyes wearing positive lenses than those wearing negative lenses, the choroids in the fellow eyes of those wearing negative lenses are significantly thicker than those in the fellow eyes of those wearing positive lenses (P < 0.01, Student’s t-test; Figs. 5A 5C) . A colleague has discovered the same phenomenon in eyes compensating for lenses under dim illumination. ⁴⁰ We speculate that this may be a reflection of a process that normally drives the two eyes toward a common refraction (i.e., an isometropizing influence). This “anti-yoking” phenomenon was not found in the rate of ocular elongation, in that the fellow eyes of positive and negative lens-wearing eyes changed by roughly the same amount (Figs. 5B 5D) . 

Implications for Human Myopia

The development of myopia in school-age children has long been associated with the presumed hyperopic defocus that the eyes experience during reading. ^{41 42} One of the puzzles of the etiology of human myopia is that although there is much epidemiologic evidence drawing a strong association of reading or other near-work with the incidence of myopia, studies explicitly looking for such evidence within a population by correlating rates of myopic progression with hours spent at near-work have yielded generally weaker associations. ^{43 44 45 46 47} Given the distinctively different fall-times of ocular compensation caused by positive and negative lens wear in chicks, it is likely that the total amount of near work does not capture all the temporal information used for integration of the defocus signals. Our findings that the effect of episodes of lens wear saturates after a few minutes and that the endurance of the effect of episodes of myopic defocus from wearing positive lenses is greater than that of negative lenses may provide a basis for the hypothesis that it is not the total duration of hyperopic or myopic defocus that is important, but the particular temporal pattern of the defocus. 

Indeed, without waiting for scientific verification, there is widespread encouragement of children to take frequent breaks from reading and to view distant objects, particularly in Singapore and China. One can speculate that the existence of a transient myopia after periods of near-work ⁴⁸ may mean that such breaks have something in common with the positive lenses used in our study. 

The recent growing awareness that ocular elongation is influenced not only by vision at the fovea, but by the entire retina (reviewed in Refs. ⁴ , ⁴⁹ ) also gives heightened significance to temporal factors, in that, in daily life, any given retinal region would experience an alternation of myopic and hyperopic defocus, as the objects imaged on that region happen to be nearer or farther from the focal plane, depending on the level of accommodation, which is determined largely by the fovea. Furthermore, even if we ignore these spatial inhomogeneities, whatever it is about reading or other near work that causes myopic progression will be present in a complex temporal pattern as the child interrupts reading by viewing distant objects. We view this study as a step toward understanding the temporal properties of emmetropization. Given that a study of this sort could not be performed on children, we suggest that correlating the pattern of reading and breaks with the rate of progression of myopia may prove useful. 

 Supported by National Institutes of Health Grants EY-02727 and RR-03060.

 Submitted for publication March 31, 2008; revised June 27 and August 7, 2008; accepted October 27, 2008.

 Disclosure: X. Zhu, None; J. Wallman, None

 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.

 

Schematics showing the linear model (A) and the nonlinear model (B) of the hypothesized internal emmetropization signal. (A) In the linear model, the magnitude of the signal is proportional to the sign and duration of defocus experienced over time, and the signal stays the same when no defocus is present. (B) In the nonlinear model, the signal rises and becomes saturated during defocus, and it declines slowly between episodes of defocus.

Table 1.

 

Summary of Treatment Protocols and Measured Changes in Experimental Eyes (ΔX) and Fellow Eyes (ΔN)

Experiment 1: Estimation of Rise-Time
Episode Duration	Interval	Group	n	+7 D Lens Treatment															Interval	Group	n	−7 D Lens Treatment
								Ocular Length (μm)			Choroidal Thickness (μm)			Vitreous Chamber Depth (μm)			Refractive Error (D)									Ocular Length (μm)			Choroidal Thickness (μm)			Vitreous Chamber Depth (μm)			Refractive Error (D)
								ΔX	ΔN	ΔX	ΔN	ΔX	ΔN	ΔX	ΔN							ΔX	ΔN	ΔX	ΔN	ΔX	ΔN	ΔX	ΔN
10 sec	10 min	1	9	202 ± 27	218 ± 24	−35 ± 15	−45 ± 17	111 ± 35	155 ± 26	−1.1 ± 0.6	−0.8 ± 0.5	10 min	9	10	223 ± 24	185 ± 16	−48 ± 19	−50 ± 15	129 ± 21	104 ± 16	−4.5 ± 0.4	−1.5 ± 0.6
30 sec	1 h	2	10	176 ± 21	253 ± 26	−46 ± 14	−44 ± 8	103 ± 25	183 ± 25	−1.0 ± 0.4	−1.7 ± 0.3
1 min	1 h	3	15	154 ± 24	222 ± 23	−3 ± 14	−27 ± 14	43 ± 19	134 ± 29	0.4 ± 0.5	−0.5 ± 0.5	10 min	10	10	292 ± 24	235 ± 24	−29 ± 18	−16 ± 25	216 ± 28	121 ± 30	−3.5 ± 0.7	−0.7 ± 0.7
2 min	1 h	4	11	99 ± 48	281 ± 30	−16 ± 15	−42 ± 16	−22 ± 36	192 ± 28	0.5 ± 0.6	−0.2 ± 0.9
2 min	10 min	5	7	−12 ± 42	199 ± 27	5 ± 19	−33 ± 19	−40 ± 50	192 ± 19	4.2 ± 0.6	0.6 ± 0.4	10 min	11	15	252 ± 34	259 ± 25	−17 ± 10	10 ± 14	161 ± 28	115 ± 19	−1.7 ± 0.6	0.3 ± 0.6
5 min	1 h	6	9	52 ± 21	250 ± 21	28 ± 18	−29 ± 29	−73 ± 31	163 ± 16	4.0 ± 0.8	−0.4 ± 0.4	10 min	12	10	318 ± 23	190 ± 25	−53 ± 17	−12 ± 13	261 ± 20	66 ± 22	−5.8 ± 0.5	−1.2 ± 0.6
10 min	1 h	7	7	54 ± 20	207 ± 24	104 ± 35	−28 ± 16	−162 ± 37	118 ± 27	6.3 ± 0.5	0.9 ± 0.4
Continuous		8	13	−19 ± 31	176 ± 28	89 ± 17	−10 ± 16	−160 ± 30	65 ± 22	5.5 ± 0.7	−0.8 ± 0.5		13	14	312 ± 37	187 ± 21	−58 ± 26	−3 ± 14	276 ± 16	61 ± 19	−3.6 ± 0.5	−0.2 ± 0.4
Experiment 2: Estimation of Fall-Time
	Episode Duration	Dark Interval	Group	n	+7 D Lens Treatment															Dark Interval	Group	n	−7 D Lens Treatment
									Ocular Length (μm)			Choroidal Thickness (μm)			Vitreous Chamber Depth (μm)			Refractive Error (D)									Ocular Length (μm)			Choroidal Thickness (μm)			Vitreous Chamber Depth (μm)			Refractive Error (D)
									ΔX	ΔN	ΔX	ΔN	ΔX	ΔN	ΔX	ΔN							ΔX	ΔN	ΔX	ΔN	ΔX	ΔN	ΔX	ΔN
	30 min												0.5 h	20	8	348 ± 32	301 ± 19	−18 ± 12	43 ± 22	263 ± 33	99 ± 19	−3.8 ± 0.5	0.2 ± 0.7
	30 min	1.5 h	14	6	10 ± 40	242 ± 43	81 ± 15	10 ± 24	−192 ± 23	63 ± 32	NA			1.5 h	21	18	256 ± 22	260 ± 18	−45 ± 15	31 ± 13	235 ± 23	99 ± 20	−3.3 ± 0.4	−0.3 ± 0.3
	30 min	3.5 h	15	23	33 ± 19	243 ± 18	101 ± 16	−5 ± 11	−159 ± 29	116 ± 19	6.2 ± 0.6	−0.4 ± 0.3	3.5 h	22	17	290 ± 27	289 ± 17	−26 ± 10	16 ± 13	217 ± 23	133 ± 24	−3.0 ± 0.5	0.0 ± 0.3
	30 min	5.5 h	16	7	103 ± 26	279 ± 13	58 ± 26	−26 ± 17	−56 ± 36	189 ± 17	2.7 ± 0.7	−1.8 ± 0.5
	30 min	11.5 h	17	7	89 ± 46	293 ± 34	−8 ± 16	−4 ± 12	12 ± 36	184 ± 22	1.6 ± 0.7	−1.3 ± 0.4	11.5 h	23	24	217 ± 21	227 ± 17	−71 ± 13	−30 ± 11	182 ± 16	142 ± 13	−4.6 ± 0.8	−1.3 ± 0.4
	30 min	23.5 h	18	21	137 ± 17	276 ± 21	−18 ± 13	−37 ± 9	49 ± 22	199 ± 16	1.2 ± 0.5	−1.0 ± 0.4	23.5 h	24	10	253 ± 28	269 ± 23	−39 ± 14	−5 ± 14	153 ± 25	112 ± 20	−1.9 ± 0.6	0.1 ± 0.3
	30 min	47.5 h	19	8	100 ± 18	163 ± 28	−33 ± 8	−62 ± 11	22 ± 17	116 ± 21	2.0 ± 0.6	−0.4 ± 0.5	47.5 h	25	7	160 ± 38	172 ± 39	−46 ± 19	−23 ± 14	109 ± 40	101 ± 34	−2.0 ± 0.6	−0.1 ± 0.5

 Data shown are changes in the treated and fellow eyes over the course of 3-day experiments. All values shown for ΔX and ΔN are mean ± SEM. The changes that were significantly different from each other by paired, two-tailed Student’s t-tests are shown in bold and underscored. In experiment 1, episodes were repeated every hour or every 10 minutes. In experiment 2, 30-minute episodes were repeated with frequencies from hourly to every other day.

Figure 2.

 

Schematics showing the rationale for determining the rise-time (A) and the fall-time (B) of the hypothesized emmetropization signals. (A) Imposing repeated lens-wearing episodes with different durations at a fixed frequency allowed an estimation of how fast the signal rises. (B) Imposing saturating lens-wearing episodes of a fixed duration but at various intervals, allowed an estimation of how fast the signal declines.

Figure 3.

 

The rise-times of the hypothesized internal emmetropization signals for positive and negative lens wear for ocular length (A) and choroidal thickness (B). Values plotted are relative changes (ΔX–ΔN). See Table 1for the statistical significance of the values. Continuous positive and negative lens wear was plotted at 60 and 10 minutes, respectively. Overlapping data points have been shifted slightly left or right for clarity. The dashed curve in (A) is fit to all data; the solid curve omits an outlying group in which birds wore lenses for 2 minutes every 10 minutes. Data points marked with † or § differ significantly from those marked with †† or §§, respectively, by ANOVA with LSD post hoc tests. Error bars, SEM.

Figure 4.

 

The fall-times of the hypothesized emmetropization signals for both positive and negative lens wear for ocular length (A) and choroidal thickness (B). Values plotted are relative changes (ΔX–ΔN). See Table 1for the statistical significance of the values. Continuous spectacle lens wear was plotted at 0.01 hour. Overlapping data points have been shifted slightly left or right for clarity. The negative lens wear data points fell into three clusters that were significantly different from one another by ANOVA with LSD post hoc tests and are marked with †, ††, or †††; the positive lens wear data points marked with § significantly differ from those marked with §§ by ANOVA with LSD post hoc tests. The horizontal line indicating 95% CI for ocular length is obscured by the adjacent data point. Error bars, SEM.

Figure 5.

 

The effects of darkness and yoking on fellow eyes. Shown are the mean changes in (A) choroidal thickness and (B) ocular length of the fellow eyes versus the amount of total darkness per day; the mean changes in (C) choroidal thickness and (D) ocular length of the fellow eyes versus those of the treated eyes, for experiment 2. Long periods of daily darkness caused progressively more choroidal thinning (A), but had little consistent effect on the rate of ocular elongation (B) in the fellow eyes (continuous lens wear is shown at 10 hours of darkness for eyes wearing positive (solid diamond) or negative (open diamond) lenses and their fellow eyes; third-order polynomial fit to the means from each group). Even though the choroidal changes in the fellow eyes correlated positively with those in the treated eyes (C), the choroids of the fellow eyes of birds wearing negative lenses became thicker on average than those of birds wearing positive lenses (A, C). This anti-yoking was not observed for the rate of ocular elongation (B, D). Squares are averages; open circles are birds wearing negative lenses; filled circles are birds wearing positive lenses. Error bars, SEM.

Figure 6.

 

Changes in choroidal thickness of the fellow eyes as a function of initial choroidal thickness. The correlation was greater in eyes with multiple periods of darkness each day than in eyes in normal light between episodes of lens wear.