Compensation to Positive as well as Negative Lenses Can Occur in Chicks Reared in Bright UV Lighting

Abstract

An earlier report describing a lack of compensation to imposed myopic and hyperopic defocus in chicks reared in UV lighting has led to the belief that the spatial resolving power of the UV cone photoreceptor network in chicks is not capable of decoding optical defocus. However this study used dim light rearing conditions, of less than 10 lux. The purpose of the current study was to determine if emmetropization is possible in young chicks reared under higher luminance, UV lighting conditions. Young, 4 day-old chicks were reared under diurnal near UV (390 nm) illumination set to either 20 or 200 lux while wearing a monocular defocusing lens (+20, +10, −10 or −20 D), for 7 days. Similarly treated control groups were reared under diurnal white lighting (WL) of matching illuminance. The WL and UV LED sources were set to equivalent illuminances, measured in “chick lux”, calculated from radiometer readings taken through appropriate narrow band interference filters, and a mathematical model of the spectral sensitivity of the chick visual system. High resolution A-scan ultrasonography was undertaken on days 0 (before lenses were fitted), 2, 4, and 7 to track ocular dimensions and refractive errors were measured by retinoscopy on days 0 and 7. Compensation to negative lenses was unaffected by UV illuminance levels, with near full compensation being achieved under both conditions, as well as under both WL conditions. In contrast, compensation to the positive lenses was markedly impaired in 20 lux UV lighting, with increased instead of decreased axial elongation along with a myopic refractive shift being recorded with the +10 D lens. Compensation under both WL conditions was again near normal for the +10 D lens. However, with the +20 D lens, myopic shifts in refractive error were observed under both dim UV and WL conditions. The spatial resolving power of the UV cone photoreceptor network in the chick is sufficient to detect optical defocus and guide the emmetropization response, provided illumination is sufficiently high. However, compensation to imposed myopic defocus may be compromised, when either the amount of defocus is very high or illumination low, especially when the wavelength is restricted to the UV range.

1.1 Introduction

Emmetropization describes a process by which neonatal refractive errors are corrected during early development. Animal experiments using defocusing lenses to artificially create refractive errors have clearly demonstrated the existence of an active feedback mechanism by which axial growth rates are altered to compensate for existing focusing errors (Schaeffel, Glasser & Howland, 1988, Wildsoet, 1997). This process requires decoding of both the sign and magnitude of retinal defocus, and although the cues used in this process remain poorly understood (Wallman & Winawer, 2004, Wildsoet, 1997). Longitudinal chromatic aberration has been postulated as a possible cue to decode the sign of defocus, providing a signal for emmetropization. Supporting experimental results show a differential effect of red and blue light on choroid and axial length responses to defocus (Rucker & Wallman, 2008, Rucker & Wallman, 2009).

In one of the earliest studies to address the question of whether chromatic aberration was an essential cue to emmetropization, Rohrer and colleagues (Rohrer, Schaeffel & Zrenner, 1992) reported that chicks reared in dim long wavelength red (665 nm, 3.92 lux) lighting compensated normally to defocus imposed with spectacle lenses while similarly treated chicks reared in dim ultraviolet (UV)(383 nm, 2.49 lux) lighting showed no response to imposed defocus. Eye growth was found to be similar for the two eyes of chicks fitted with bilateral low powered lenses of opposite sign (+4 and −4 D) under the UV lighting conditions in the above study. Unlike higher order mammals and primates, many birds including chicks, have UV-sensitive cone photoreceptors (Bowmaker, Heath, Wilkie & Hunt, 1997). Thus this finding was interpreted as evidence that UV cones do not participate in emmetropization. Note however, that inspection of graphed axial length data suggests close correspondence between patterns of growth for the lens treatments under the UV lighting conditions and the −4 D lens treatment under white light conditions. Nonetheless, overall these data combined with data from other studies involving chicks reared under monochromatic conditions (589 nm, 140 lux (Schaeffel & Howland, 1991)/550 nm, 33 lux (Wildsoet, Howland, Falconer & Dick, 1993)) suggest that chromatic aberration per se is not essential for emmetropization but that input from UV cone photoreceptors alone may not be sufficient. However, the influence of monochromatic lighting also appears complex. In a more recent study, altered emmetropization responses were observed in chicks wearing defocusing lenses reared in blue lighting (460 nm, 0.67 lux) (Rucker & Wallman, 2008). Bidirectional axial growth responses appeared to remain intact for both positive and negative lenses in blue light; however choroidal responses were not observed. The converse was observed using the same lens paradigms in red light, with negative lenses inducing choroidal thinning, and positive lenses, choroidal thickening, with no significant changes in eye length noted. The blue lighting conditions would have stimulated both UV and short-wavelength sensitive cones, while the red lighting would have stimulated double (D), medium (M), and long (L) wavelength cones.

The lighting level used in rearing also appears to have an important influence on emmetropization. In one early study (Feldkaemper, Diether, Kleine & Schaeffel, 1999), different outcomes were observed when retinal light levels were reduced with neutral density (ND) filters covering the eyes, to around 5.5 lux, compared to an equivalent reduction in ambient illumination; low myopia instead of emmetropia was observed with the ND filter condition only. In another related, but independent study (Moore, Irving, Sivak & Callender, 1998), the combination of 2ND filters and defocusing lenses (+10 & −10 D), resulted in myopia with both lens types, implying impaired emmetropization. Unfortunately, the lighting conditions used in the latter study were not specified. In a more recent study, high luminance levels was shown to retard and enhance the rate of compensation to negative lenses and positive lenses respectively, although in both cases, the emmetropization endpoints were not affected (Ashby & Schaeffel, 2010, Siegwart, Ward & Norton, 2012, Smith, Hung & Huang, 2012). This effect of high light levels on lens-induced myopia has also been described in tree shrews and monkeys (Ashby & Schaeffel, 2010, Siegwart et al., 2012, Smith et al., 2012).

In the study reported here, we further investigated the effect on emmetropization of rearing chicks in UV lighting, specifically to explore its possible light intensity-dependency. To this end, we compared the emmetropization responses to two different levels of imposed myopic and hyperopic defocus of young chicks reared under either bright or dim UV lighting, or equivalent bright or dim white (WL) diurnal lighting.

1.2 Materials and Methods

1.2.1 Animals

One day-old White Leghorn chicks (Gallus gallus domesticus) were obtained as hatchlings from a commercial hatchery (Privett Hatchery, New Mexico). Throughout experiments, chicks were provided with food and water ad libitum. Care and use of the animals were in compliance with an animal use protocol approved by the Animal Care and Use Committee of the University of California-Berkeley, and adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.

1.2.2 Experimental design

Chicks were housed in temperature and light-proof chambers for this study. For the first 3 days, all were exposed to diurnal white lighting of 800 lux set to a 12:12 hour light/dark cycle. On the afternoon of day 3, baseline refractive error and biometry measurements were made and the chicks were randomly allocated to one of 4 monocular lens treatment groups (−20, −10, +10 or +20 D), and lenses fitted. Their fellow (contralateral) untreated eyes served as controls. The lenses were worn for next 7 days. Chicks in each of the four lens groups were randomly allocated to one of 4 different lighting conditions for rearing during the lens-wearing period: either 390 nm UV or white lighting (WL) set to an illumination level of either 20 or 200 lux. The number of chicks assigned to each of the 12 experimental groups is summarized in Table 1. The effects of the treatments were tracked with biometry measurements made between 1 and 3 pm (5–7 hours after lights-on), on days 0 (baseline), 2, 4 and 7, and refractive error measurements made on days 0 and 7, immediately before biometry. Care was taken during measurements and lens cleaning to ensure that chicks were exposed only to the experimental lighting conditions to which they had been assigned.

Table 1.

Number of birds in each lens treatment group, for each of the four lighting conditions used in this study.

Lighting conditions	Lens treatment
	+20 D	+10 D	−10 D	−20 D
Ultraviolet, 200 lux	6	10	7	8
Ultraviolet, 20 lux	6	5	6	8
White Light, 200 lux	5	5	4	5
White Light, 20 lux	5	5	5	5

1.2.3 Ocular biometry & refractive error measurements

All measurements were made under general anesthesia (isoflurane; 1–2% in oxygen). High frequency A-scan ultrasonography (Nickla, Wildsoet & Wallman, 1998, Schmid & Wildsoet, 1996) was used to determine axial ocular dimensions, including anterior chamber depth (ACD), vitreous chamber depth (VCD), and lens thickness, as well as the thicknesses of the ocular layers at the posterior pole of the eye, i.e. retinal thickness, choroidal thickness (CT), scleral thickness. Optical axial length (OAL) was derived from the sum of ACD, lens thickness and VCD. In reporting results, emphasis is given to parameters significantly affected by the treatments. Refractive errors (RE) were measured using streak retinoscopy, with the average of values for the two principal meridians used in data analyses.

1.2.4 Light sources and luminance measurements

An IL1700 research radiometer (International Light, Inc, USA) was used to measure lighting levels. Two triphosphor fluorescent lamps provided the white lighting, under which all chicks were reared for the initial 3 pre-experimental days. The provided illumination averaged 800 lux at the level of the cage floor.

For the experimental white lighting conditions, four 110V, white phosphorous LED bulbs, which have wide bandwidth (410 – 760 nm), with emission peaks at 450 nm and 540 nm were used. The UV lighting was provided by four 110V, UV gallium nitride LED bulbs, a peak emission at 390 nm and a 25 nm full width-half maximum. Thus, although the latter lighting cannot be classified as monochromatic, the emission spectrum was narrow, limited to 365 to 415 nm, as assessed using appropriate interference filters in conjunction with the radiometer. Illumination, controlled by an in-line rheostat, was set to either 20 or 200 “chick lux”. Measurements in this case were made using a flat response filter (F-filter), with 78% transmittance at 390 nm, attached to the sensor. Thus a correction for the 22% absorbance at 390 nm was applied to readings. To account for the spectral sensitivity of the chick visual system we derived a mathematical model of the chick photopic sensitivity function based on previously reported ERG responses to photopic stimuli (Chen & Goldsmith, 1984). A 2^nd order polynomial was fitted to the data, corresponding to wavelengths less than or equal to 450 nm (V1 =−5⁻⁰⁵λ² + 0.041λ − 8.282). Our UV light source was considered to be monochromatic for the purpose of these relative sensitivity calculations, yielding a V1 of 0.103. The latter value was used to convert irradiance data to chick lux. The calculation of equivalent chick lux for the white light was done similarly, with a second function (V2 = −1⁻¹⁰λ⁵ + 4⁻⁷λ⁴ − 0.0001λ³ + 0.244λ² − 66.01λ + 7094) used for wavelengths above 450 nm. The output of the white LEDs was measured using 10 nm full-half width maximum interference filters spanning the range of 400 – 690 nm in combination with a radiometer. The latter radiances were used to calculate the overall illuminance in chick lux, using Simpson’s method of numerical integration with the summed values representing the area under the curve.

1.2.5 Data analyses

Differences between treated and untreated fellow eyes in ACD, CT, OAL and RE were derived for each bird at each time point. These data were subjected to one-way ANOVA analyses, with post-hoc Sidak testing to look for lighting-related differences in responses to each of the lens treatments. OAL, CT and RE interocular difference data are also shown graphically, normalized to baseline values. In addition, we performed mixed ANOVA analysis of fellow eyes to look for additional effects of the lighting conditions.

1.3 Results

Overall, we found that the lighting conditions used in rearing affected only the responses to imposed myopic defocus (positive lenses), and these effects were also limited to the dim (20 lux) lighting conditions. Described in detail below are the effects on optical axial length and choroidal thickness, the contributions to refractive error changes of anterior ocular components (ACD & LT) being first ruled out with a mixed ANOVA analysis. We also found no significant differences in either ACD or LT changes between lighting conditions. An additional one-way ANOVA analysis was conducted on data for each time-point to confirm this result and avoid the possibility of introducing a type II statistical error.

1.3.1 Effect of UV illumination on ocular growth and refractive changes induced by negative lenses

With a negative lens in place, the imposed hyperopic defocus, when correctly decoded, triggers compensatory increased axial elongation accompanied by choroidal thinning in young chicks. This growth response pattern was observed in all groups treated with −10 D lenses (Figs. 1A – C), irrespective of the lighting conditions (UV and WL & both 20 and 200 lux), with no statistical differences in lens treatment effects between the 4 lighting groups. Likewise, the lighting conditions did not significantly affect the response to the −20 D lenses (Figs. 1D – F). Over the 7 day treatment period, the −10 and −20 D lens treatments resulted in mean increases in OAL of 0.636 ± 0.033 mm and 0.892 ± 0.055 mm respectively (table 2). Both lens treatments also induced choroidal thinning (Figs. 1B & E), and here also, there was no significant difference in these responses between lighting conditions (−10 D: p = 0.447; −20 D: p = 0.121). These dimensional changes resulted in near complete refractive compensation to the −10 D lens after 7 days of lens treatment, as reflected in the combined average change in refractive error for all groups of −11.04 ± 2.12 D. This was not the case for the −20 D lens treatment groups, which recorded an average change in refractive error of −16.20 ± 1.26 D, i.e. compensation was incomplete.

Figure 1.

The effect of lighting conditions used in rearing on ocular responses to imposed hyperopic defocus. Mean interocular differences between lens-treated and fellow eyes, plotted against day of lens wear for both −20 and −10 D lenses. Error bars are standard errors of the means. No statistically significant, lighting-related differences observed (p < 0.05).

Table 2.

Group mean interocular differences (treated eye – fellow eye, with SDs; mm) in anterior chamber depth (ACD) vitreous chamber depth (VC) and choroidal thickness (CH) and optical axial length (OAL), recorded at the end of the 7 day lens treatment period.

Lighting Conditions	Parameter	Lens Treatment
		+20 D	+10 D	−10 D	−20 D
UV 20 lux	ACD	0.380±0.147	0.124±0.048	0.113±0.090	0.088±0.110
	VC	0.898±0.183	0.385±0.133	0.646±0.119	0.648±0.067
	CH	0.114±0.017	0.007±0.022	−0.074±0.036	−0.023±0.053
	OAL	0.942±0.171	0.419±0.082	0.585±0.124	0.691±0.055
WL 20 lux	ACD	0.139±0.074	−0.067±0.129	0.135±0.053	0.227±0.138
	VC	0.382±0.132	−0.384±0.070	0.295±0.099	0.762±0.317
	CH	−0.120±0.046	0.025±0.042	−0.041±0.015	−0.061±0.023
	OAL	0.315±0.186	−0.298±0.094	0.637±0.062	0.943±0.137
UV 200 lux	ACD	0.229±0.131	0.025±0.026	0.038±0.061	0.213±0.063
	VC	0.139±0.198	−0.209±0.040	0.500±0.093	0.899±0.131
	CH	0.186±0.129	0.155±0.069	−0.061±0.049	−0.039±0.017
	OAL	−0.099±0.111	−0.258±0.052	0.682±0.076	0.838±0.083
WL 200 lux	ACD	0.425±0.108	0.074±0.095	0.198±0.092	0.065±0.075
	VC	0.136±0.086	−0.222±0.124	0.338±0.041	0.327±0.167
	CH	0.352±0.089	−0.352±0.089	−0.009±0.016	0.021±0.024
	OAL	−0.194±0.144	−0.194±0.144	0.544±0.076	1.006±0.105

1.3.2 Effect of UV illumination on ocular growth and refractive error changes induced by positive lenses

The typical response to positive lenses in young chicks includes choroidal thickening and slowed axial elongation, the former changes occurring more rapidly than the latter scleral changes. However, the pattern of response was found to vary with both the lighting conditions used in rearing as well as the magnitude of imposed myopic defocus. With both 20 and 200 lux WL conditions, a significant increase in choroidal thickness was observed with the +10 D lens after only 48 h of lens wear; however, choroidal thickening was maintained out to 7 days only under the brighter 200 lux WL condition (table 2). These near normal responses contrast with the choroidal response to the +10 D lens under the UV conditions, which was attenuated relative to the response under the equivalent WL condition, for the brighter (200 lux) UV condition, and nonexistent in the case of the dimmer (20 lux) UV condition (Fig. 2B). No significant change in choroidal thickness beyond 2 days was observed with the +10 D lens under either of the dim (20 lux) conditions (UV or WL; Fig. 2E). With the +20 D lens, only the 200 lux WL group showed the usual choroidal thickening. The choroids of the 200 lux WL group were significantly thickened after 48 h of lens wear and this change was sustained over the lens treatment period. Under the equivalent, 200 lux UV lighting condition, a trend towards choroidal thickening was also evident after 48 h of lens wear (0.129 ± 0.106 mm), and although this trend was maintained over the remainder of the lens wearing period, no statistically significant differences were observed between this treatment group and any other, for any day except day 7, when values proved to be significantly larger than those recorded for the 20 lux WL and UV groups. The absence of statistical significance reflects the large variability in these data. The 20 lux WL and 20 lux UV groups both showed choroidal thinning instead of thickening, with recorded changes being significantly different from those of the 200 lux WL group.

Figure 2.

The effect of lighting conditions used in rearing on ocular responses to imposed myopic defocus. Mean interocular differences between lens treated and fellow eyes plotted against day of lens wear for both +20 and +10 D lenses. Error bars are standard errors of the means. Symbols denote significant lighting-related differences (p<0.05).

In addition to the above changes in choroidal thickness, eyes wearing +10 lenses recorded reduced OALs compared to their fellow eye (Fig. 2A), with corresponding hyperopic shifts in refraction (Fig. 2C), under all but the 20 lux UV condition, under which the same lens treatment elicited the opposite response - a significant increase in OAL after 4 days (Fig. 2A), with a corresponding myopic shift in refractive error. When the magnitude of imposed defocus was increased to +20 D, only groups reared under the two brighter lighting conditions displayed the expected response patterns, with similar OAL and RE changes observed in both 200 lux UV and WL groups, over the course of the experiment, despite differences in their CT responses. The two +20 D lens groups reared under the 20 lux conditions both exhibited abnormal responses, with increased instead of decreased axial elongation, and myopia instead of hyperopia.

1.3.3 Treatment effects on fellow eyes

The light conditions used in rearing had no significant effect on the ocular growth patterns of the fellow untreated eyes. Mixed ANOVA and one-way ANOVA analyses of biometric data (AC, LT, VCD, OAL), collected from the fellow eyes of those treated with positive and negative lenses did not reveal any significant differences attributable to the lighting conditions.

1.4 Discussion

Most noteworthy of the findings in the study reported here is the contrasting influence of the wavelength and intensity of the lighting used in rearing on compensation (emmetropization) to imposed hyperopic and myopic defocus. With imposed hyperopia (negative lenses), compensation was near normal under both our WL and UV lighting conditions, even when lighting was lowered to 20 lux and irrespective of whether a moderate (−10 D), or a large (−20 D) focusing error was imposed. In contrast, with imposed myopic defocus (positive lenses) and UV conditions, near normal compensation was limited to moderate level of defocus (+10 D lens), and the brightest (200 lux) illuminance. Compensation to imposed myopic defocus was superior under the WL conditions although restricted to the brightest (200 lux) illuminance for the larger (+20 D) focusing error.

What is the significance of our results for emmetropization? The apparent lack of sensitivity of the response to imposed hyperopic defocus to lighting wavelength and intensity may be expected if the default response to retinal blur is increased growth, as seen with form deprivation where growth regulation appears to be open loop (Schaeffel & Howland, 1991). Such default growth responses, which underlie the blur hypothesis model proposed by Schaeffel et al, (Schaeffel & Howland, 1991) will nonetheless progressively attenuate focusing errors imposed by negative lenses, avoiding the need to decode the direction of imposed defocus, leading to the impression that emmetropization is functioning normally. This same model predicts impaired compensation to imposed myopia, when the direction of the imposed defocus is inaccurately decoded or ambiguous. The increased growth observed with very large amounts of myopic defocus is consistent with this notion (Nevin, Schmid & Wildsoet, 1998), and may contribute to the apparent saturation of responses with increasing myopic defocus seen in another study (Irving, Sivak & Callender, 1992). It is plausible that decoding is also inaccurate under monochromatic conditions, because chromatic aberration cues are absent, although in a much earlier study, appropriate compensation was reported under red but not UV lighting (Rohrer et al., 1992). While these data were interpreted as evidence that emmetropization does not use input from UV system, nonetheless, the typical ‘dark-rearing syndrome’ (Gottlieb, Fugate-Wentzek & Wallman, 1987) was not observed by Rohrer (Rohrer et al., 1992) in the birds raised in UV lighting, implying that the UV cones were actively participating in ocular growth regulation, even if not able to elicit defocus-driven responses. The results of our current study, showing that compensation for imposed defocus is possible in bright UV lighting call for a reconsideration of the role of UV-cones in emmetropization in chicks.

In interpreting the current results, it is important to establish whether only UV-sensitive cones (“UV cones”) would have been stimulated by our UV lighting conditions, given our use of a narrow-band rather than monochromatic source. Specifically, our UV source has a peak emission at 390 nm, with 25 nm full width-half maximum. Activation of more than one cone type would potentially provide the necessary chromatic input to decode the direction of defocus, and so to guide emmetropization. Also did cones other than UV-sensitive ones mediate the observed compensatory responses to imposed myopia? Of possible cone types activated by our UV source, S- and double (D)- cones appear the mostly likely candidates. However, given the emission of our source at wavelengths above 415 nm is near zero, the sensitivity of S-cones would seem too low; the λ_max of their visual pigment is approximately 455.2 nm (Bowmaker et al., 1997), and their inner segments include C-type droplets, which act as cut-off filters (λ_cut 445–450 nm), effectively displacing their peak sensitivity to longer wavelengths. The plausibility that D-cones could be active under our UV conditions rests with the strength of the evidence that accessory cones contain oil droplets, as the spectral sensitivity maximum for the long wave-sensitive (LWS) pigment located in the D-cones is in the yellow region of the spectrum (λ_max of 567 nm). While there is general agreement that the principal member contains an oil droplet (P-type), with a λ_cut of 430 nm, electron microscopy data have called into question the presence of oil droplets in accessory cones (Morris, 1970). However, more convincing data obtained using phase contrast microscopy applied to unpreserved, flat mounted retina indicate the presence of yellow-green (C-type) oil droplets in accessory cones (Meyer & May, 1973). We note also that functional ERG data recorded from the chick show relative minima in the spectral sensitivity function at 445 and 425nm (Chen & Goldsmith, 1984) (Rohrer et al., 1992), corresponding approximately to the λ_cut’s of S- and D-cones. Thus S- and D-cones do not appear to have sufficient sensitivity in the short wavelength region to have contributed to the visual responses under our UV lighting conditions, which we argue would have exclusively excited UV-cones. We can also rule out a contribution from rod photoreceptors, whose activity is limited to night-time hours due to the presence of a rod-cone switch in the chick retina (Schaeffel, Rohrer, Lemmer & Zrenner, 1991).

Previous studies have reported near normal lens compensation in chicks reared under monochromatic yellow (589 nm)(Schaeffel & Howland, 1991) and red (665 nm)(Rohrer et al., 1992) lighting conditions, but abnormal responses under both blue (460 nm)(Rucker & Wallman, 2008) and UV (383 nm) (Rohrer et al., 1992) lighting conditions. To our knowledge, we are the first to observe near normal compensation in chicks reared under UV (390 nm) lighting conditions i.e. similar to that seen under photopic white lighting. We attribute the difference in our study outcome from the previous one involving UV lighting to critical differences in experimental design. In the earlier study (Rohrer et al., 1992), binocular treatments with low power lenses of opposite sign (+/−4 D) were applied, and thus abnormal responses that did not preserve the expected response direction would have been difficult to detect. Our study used monocular treatments as well as higher power lenses (±10D & ±20 D vs. ±4 D). We also included two lighting levels, one in the same range as that used in the earlier study (~10 vs. 20 lux illuminance), and another that was one log unit brighter (200 lux). The latter condition combined with the lower of the two lens powers (+10 D), proved critical to demonstrating the ability of UV-cones to drive the compensatory response to imposed myopic defocus.

In the most recent study to examine the effects of lighting wavelength on emmetropization, Rucker and Wallman (Rucker & Wallman, 2008) drew the intriguing conclusion that short and long wavelengths are able to modulate eye length and choroidal thickness respectively. This conclusion was based on their observation that chicks reared in blue light (460 nm) of low luminance (0.67 lux) and treated for 3 days with +6 D lenses showed retarded axial growth, as expected, but no choroidal thickening. Conversely, when the equivalent luminance red light (620 nm) was substituted for the blue light, the choroidal response was preserved; with +8 D lenses, 73% of observed axial length changes were attributable to choroidal thickening. Interestingly, our UV conditions also resulted in attenuated choroidal thickening responses to imposed myopia (+10 & +20 D lenses). However, subjective evaluation of our graphed data suggests less attenuation under the brighter, 200 lux, compared to 20 lux UV lighting condition. One plausible, alternative explanation for the latter observation and the discrepancies between our study and the earlier one by Rohrer et al (Rohrer et al., 1992), is that there is a threshold luminance requirement for the choroidal thickening response to myopic defocus, and that it is higher for short wavelengths than long wavelengths. In an independent study, dim (2 lux) white lighting was shown to impair the choroidal thickening response to +8 D lenses (Roberts, Zhu & Wallman, 2003), and reducing illumination further to 0.2 lux resulted in reduced growth retardation, by 25% relative to that recorded in birds reared in 400 lux conditions. This finding also has a parallel in the current study, in that the response to the higher power positive lenses was impaired under our dimmer white lighting condition, and more broadly, for the UV lighting, the only near normal response to imposed myopic defocus was recorded under our 200 lux condition. In dim lighting, presumably functional alterations in the retinal circuitry, including increased spatial summation (e.g. Dowling, 1991(Dowling, 1991)), as well as reductions in the numbers of participating photoreceptors under reduced luminance, especially when combined with UV wavelengths, likely contribute to the deteriorating performance of the emmetropization mechanism, due at least in part to the increasing depth of focus.

Interest in the possible protective effect of sunlight against myopia in humans has lead to a number of recent studies involving much higher light levels than used in the current study. In a study involving chicks by Ashby and colleagues (Ashby & Schaeffel, 2010), exposure to very bright (15,000 lux) lighting was found to enhance or retard the rate of lens compensation, depending on whether positive or negative lenses were worn but the refractive end points were unaffected. Similar trends with negative lenses have since been reported for monkeys and tree shrews (Siegwart et al., 2012, Smith et al., 2012), and for these animals, as for chicks, bright light has a more dramatic inhibitory effect on form deprivation myopia (Ashby, Ohlendorf & Schaeffel, 2009). In another study involving chicks, high illuminance levels were found to inhibit form deprivation myopia, with this effect being prevented when a D₂ dopaminergic antagonist was administered prior to exposure (Ashby et al., 2009). This result is consistent with the well-established link between reduced retinal dopamine levels and form deprivation myopia (Weiss & Schaeffel, 1993). While the role of dopaminergic mechanisms in compensatory lens responses is not fully resolved (Rohrer, Spira & Stell, 1993, Schmid & Wildsoet, 2004, Weiss & Schaeffel, 1993), dopamine turnover is expected to decrease under reduced light levels. Thus these conditions need not impair the responses to the negative lenses, especially if the visual conditions become more akin to form deprivation, for example, when the mechanism for decoding of the sign of defocus appears to fail under the dimmer (20 lux) UV conditions. That altered retinal dopamine turnover also explains the impaired compensatory response to positive lenses observed with decreased luminance is more speculative, although select dopamine agonists are known to cause transient choroidal thickening (Nickla, Totonelly & Dhillon, 2010). Much is yet to be learnt about the retinal circuitry mediating the compensatory responses to imposed defocus before the modulatory influences of lighting levels can be understood. Nonetheless, collectively these various studies point to an operating range above and below which altered compensation may be expected, to either or both negative and positive lenses.

In summary, it appears the input from UV cones may be used by the emmetropization mechanism in chicks. Our results further indicate that accurate decoding of the sign of optical defocus, as required for compensation to myopic defocus requires a higher luminance threshold to be reached under UV lighting. Compensation to hyperopic defocus does not have the same requirement, perhaps because the default growth response to visually-degraded conditions is coincidently in the same direction as that required for compensation in this case.

Highlights.

• We examined if young chicks can use UV cones to guide emmetropization.

• Compensation to negative lenses was unaffected by illuminance levels.

• Compensation to plus lenses is possible in bright UV illumination.

• The UV photoreceptor network in the chick can guide the emmetropization response.