Effects of low intensity ambient lighting on refractive development in infant rhesus monkeys (Macaco mulatto)

Abstract

Studies in chickens suggest low intensity ambient lighting causes myopia. The purpose of this experiment was to examine the effects of low intensity ambient lighting (dim light) on normal refractive development in macaque monkeys. Seven infant rhesus monkeys were reared under dim light (room illumination level: ~55 lux) from 24 to ~310 days of age with otherwise unrestricted vision. Refractive error, corneal power, ocular axial dimensions, and choroidal thickness were measured in anesthetized animals at the onset of the experiment and periodically throughout the dim-light-rearing period, and were compared with those of normal-light-reared monkeys. We found that dim light did not produce myopia; instead, dim-light monkeys were hyperopic relative to normal-light monkeys (median refractive errors at ~155 days, OD: +3.13 D vs. +2.31 D; OS: +3.3 ID vs. + 2.44 D; at ~310 days, OD: +2.75D vs. +1.78D, OS: +3.00D vs. +1.75D). In addition, dim-light rearing caused sustained thickening in the choroid, but it did not alter corneal power development, nor did it change the axial nature of the refractive errors. These results showed that, for rhesus monkeys and possibly other primates, low ambient lighting by itself is not necessarily myopiagenic, but might compromise the efficiency of emmetropization.

1. Introduction

Recent studies have consistently found that spending more time outdoors reduces the risk of myopia genesis in children (Dirani et al., 2009; French et al., 2013; Guggenheim et al, 2012; Jones et al, 2012; Rose et al., 2008). Although the exact mechanism remains to be elucidated, the higher lighting intensities that are common in outdoor environments might underlie this protective effect. For instance, elevated laboratory lighting attenuated the reduction of hyperopia that is normally associated with emmetropization (Ashby et al., 2009; Ashby & Schaeffel, 2010; Cohen et al., 2011,2012; Karouta & Ashby, 2015; Siegwart Jr. et al., 2012). Moreover, rearing animals under elevated lighting reduced the degree of form-deprivation myopia in chicks (Ashby et al., 2009), tree shews (Siegwart Jr. et al., 2012), and rhesus monkeys (Smith III et al., 2012). Finally, elevated lighting reduced the degree of negative-lens-induced myopia in guinea pigs (Li et al., 2014), slowed its development in chicks (Ashby & Schaeffel, 2010) and tree shrews (Siegwart Jr. et al., 2012), although no obvious effects were seen in monkeys (Smith III et al., 2013). These observations provided strong evidence that elevated lighting levels protect animal eyes against some forms of experimentally induced myopia, supporting the role of brighter outdoor lighting in reducing the risk of myopia in children.

A logical extrapolation of the protective effects of elevated lighting is that low intensity ambient lighting encourages myopia genesis and promotes myopia progression (Norton & Siegwart, Jr., 2013). Low intensity ambient lighting has been found to alter the ocular morphology of chickens during postnatal development in a way that resembles myopic ocular changes. The first report of a possible low-light effect on eye growth was from Harrison and McGinnis (Harrison & McGinnis, 1967). They found that rearing chickens under low intensity (0.06~0.09 μW/cm²*nm), blue diurnal lighting caused excessive ocular axial elongation, and their eyes became substantially myopic. However, despite some successful replications of their refractive outcome (Bercovitz, Harrison, & Leary, 1972; Harrison, Bercovitz, & Leary, 1968), the role of light intensity remained confounded by the spectral composition of light, until Lauber and Kinnear (Lauber & Kinner, 1979) induced eye enlargement in three different sub-species of chicks using low-intensity white light. This finding suggested that the eye enlargements in the earlier studies were likely lighting intensity, rather than wavelength, associated. It is noteworthy that the eye enlargements observed in these studies did not involve severe corneal flattening and increased intraocular pressure, indicating that the observed phenomena were distinct from constant-light-induced buphthalmia, a glaucomatous condition marked by eye enlargement, high intraocular pressure and severe corneal flattening (Jensen & Matson, 1957; Lauber, Boyd, & Boyd, 1970; Lauber, Shutze, & McGinnis, 1961). Instead, the low-lighting alterations shared certain biometrical characteristics with other vision-induced experimental models of myopia (specifically, axial elongation due to excessive growth of the vitreous chamber) (for a review see Troilo et al., 2019), suggesting that these two phenomena might have common, if not identical, regulatory mechanisms.

The above pioneering works were conducted before the introduction of practical and precise in vivo biometry measurements and commonly used animal models of myopia (e.g. form-deprivation myopia; Raviola & Wiesel, 1978, 1985), and therefore limited compared with more recent refractive studies. In this regard, Ashby et al. (Ashby et al., 2009) examined the refractive effects of low intensity lighting using an avian model of form-deprivation myopia. They found that 6-hour daily exposures to low-intensity white lighting (average intensity = 50 lux) for 5 days did not exacerbate form-deprivation myopia in chicks, nor did it make the fellow, untreated eyes more myopic. Similarly, Feldkaemper et al. (1999) reported that reducing ambient lighting levels (550 lux) by 2 log units (5.5 lux) for 9 days failed to produce myopia in chickens reared with unrestricted vision (Feldkaemper, Diether, Kleine, & Schaeffel, 1999). Interestingly, these findings were somewhat contrary to those of Cohen et al.’s, in which rearing chickens with unrestricted vision in dim light for prolonged treatment periods (90 days) caused reduced corneal power and axial myopia (Cohen et al., 2011, 2012). Strengthened by the long observation period and periodic biometric measures, the clear distinction between dim- and normal-light emmetropization patterns in the latter study strongly suggested that lower ambient lighting intensity could cause myopia.

Although the studies of Cohen et al.’s (2011, 2012) associated low illumination levels with increased risk of myopia, the translational application of the results maybe limited due the nature of the ocular component changes (e.g., reduced corneal power) that were associated with the myopic refractive errors. To the best of the authors’ knowledge, the effects of low-intensity lighting on refractive development have only been studied in chicks, which possess species-specific ocular anatomical features and light-response mechanisms that might influence refractive development. For example, in contrast to the rod-dominated retinas of humans and non-human primates (cone-to-rod ratio = 1:20), the retinas of chickens are cone-dominated (cone-to-rod ratio = 3:2) (Wisely et al, 2017). These differences may be important because mice without functional rod pathways do not emmetropize, nor do they develop form-deprivation myopia (Park et al., 2014), which indicates that rods can play a role in vision-dependent refractive development. Considering that the two photoreceptor populations function under different ambient lighting levels, the differences in rod-cone ratio might affect how refractive developments proceeds under different ambient lighting levels. Moreover, the chicken cornea is subject to constant-dark-(Troilo & Wallman, 1991) and constant-light-induced corneal flattening (Cohen et al., 2008; Li et al., 1995), the latter of which is a light-intensity-dependent phenomenon (Cohen et al., 2008) that has not been observed in non-human primates (Smith III et al., 2001; Smith III et al., 2003). Finally, rearing animals under quasi-monochromatic lighting appears to affect refractive development in chicks (Foulds, Barathi, & Luu, 2013; Seidemann & Schaeffel, 2002) and rhesus monkeys (Hung et al., 2018; Smith III et al., 2013, 2015) in a qualitatively different manner, suggesting that ocular mechanisms influenced by ambient lighting in chicks might not be identical to those in primates. In contrast, rhesus monkeys are similar to humans with respect to ocular anatomy, visual physiology (Harwerth & Smith III, 1985), the course of refractive development, and the nature of vision-induced refractive errors (Qiao-Grider et al., 2007; also see the review by Troilo et al., 2019). It is likely that the refractive development of rhesus monkeys reared under low-intensity ambient lighting is etiologically similar to that in humans, making them a promising animal model for the study of dim-light effects on refraction. The purpose of this experiment, therefore, was to examine the effects of low intensity lighting on the refractive development of rhesus monkeys.

2. Methods

2.1. Animal subjects and intervention strategy

Seven infant rhesus monkeys (Macaca mulatta) acquired at 2 weeks of age were the primary subjects. Prior to the onset of the experiment, these monkeys were housed in a primate nursery illuminated by “white” fluorescent lights (GE Ecolux® Starcoat® T8 F32T8/SP35/ECO, General Electric Co., Boston, MA) on a 12-hour light/12-hour dark cycle. The spectral output of the fluorescent lighting was multi-peaked with maximum intensities located at 612 nm and 550 nm (Figure 1A). Lighting intensities ranged between 312 – 860 lux as measured in the middle of each caging area with the light sensor facing the ceiling light panel (“normal” light, mean ± standard deviation of ambient lighting intensity = 504 ± 168 lux, correlated color temperature = 3170K).

Figure 1.

The spectral irradiance (Spectrophotometer CL-500A, Konica Minolta Sensing America, Inc. Ramsey, NJ, USA) of two representative measurements for the normal (corresponding illuminance level: 459.31 lux) and dim ambient lighting levels (corresponding illuminance level: 30.7 lux). Panel A. Spectral irradiance for normal lighting (black line) compared to that for dim lighting (red line). Panel B. Spectral irradiance for the dim lighting condition plotted on an expanded y-axis to illustrate the similarities in the spectral composition of the light under the dim and normal ambient lighting conditions.

At 24 ± 2 days of age, monkeys were transferred to another nursery room with reduced diurnal lighting and reared in that room until the end of the dim-light-rearing period (310 ± 21 days of age). This observation period was longer than that employed in our previous investigations of ambient lighting effects, which ended at approximately 150 days of age (Smith III et al, 2013). We employed an extended observation period for the dim-light monkeys because the myopic refractive changes observed in chickens reared under dim light were slow to develop and appeared to require long treatment periods to become obvious (Cohen et al., 2011).

The experimental lighting was maintained on the same diurnal cycle as the preexperiment environment. During the light phase, the environment was illuminated by filtering the fluorescent lights through an aluminum-deposited, clear polyester film (Grafix™ Metalized Dura-Lar®, Silver, 0.05mm-thick; Grafix, Maple Heights, Ohio) that was tightly fitted onto the ceiling lighting panels. The resulting room illumination was approximately 55 lux as measured directly under the light panels at waist-level, without significant alterations in spectral composition (Figure 1B). Light levels measured at the front of individual cages with the light meter facing horizontally out of the cage ranged from 7 to 36 lux. Aside from the monkey housing area, the group-socialization area and the connecting area to the housing room were also equipped with fluorescent lights dimmable to the described illumination level. Lighting conditions in these areas remained highly stable throughout the experimental period.

We chose the current ambient lighting level because it facilitates comparisons to recent experiments in chickens (Ashby et ah, 2009; Cohen et ah, 2011, 2012) and because it should have been sufficient to maintain normal circadian rhythms (Duffy & Czeisler, 2009). Our rearing environment was very dim in comparison to typical outdoor lighting levels. In this respect, our light levels were just above twilight light levels typically encountered outdoors (Koomen et ah, 1952). Moreover, our lighting levels were dim in comparison to indoor lighting standards. For example, the lowest maintained illuminance levels recommended by the International Organization for Standardization (ISO 8995-1: 2002, Lighting of Indoor Work Spaces) for general building areas, educational buildings (student common room), and offices areas for cleric workers are 100, 200, and 500 lux, respectively. In support of our chosen lighting level, a recent intervention trial found that increasing classroom lighting levels from 70 – 100 lux to 440 – 550 lux significantly reduced the magnitude of myopic refractive shifts, slowed axial ocular growth, and lowered the incidence of myopia in children (Hua et al., 2015).

To ensure that the dim-light subjects were not exposed to typical laboratory lighting due to human activity, the following measures were implemented. Animal care personnel who accessed the animals were instructed to dim the light in the connecting area before entering the dim-light monkey housing area. Animals in the dim-light group never left the low-illumination area except for the periodically scheduled measurement sessions. For the measurement sessions, researchers followed the same instructions for accessing the animals and covered the anesthetized animal’s head with light-blocking cloth before transferring them to the lab. During data collection, the lab was illuminated only by projecting a desk lamp onto a wall remote from the animals. Auxiliary lighting was used to facilitate contact lens insertion/removal during optical coherence tomography (OCT) measurements by shining a penlight on the examined eye from the side.

Control data were obtained from age-matched monkeys reared under typical laboratory lighting without visual restriction (normal-light-reared monkeys). Four of these control monkeys were reared during the dim-light experiment, whereas the rest were from previous studies and their data have been published and discussed (Hung, Arumugam, Ostrin, et al., 2018; Hung, Arumugam, She, et al, 2018; Qiao-Grider et al., 2007; Smith III et al., 1999, 2003, 2010, 2013, 2015). The rearing environments for the normal-light monkeys were similar to the pre-experiment environment described above. The general animal husbandry procedures and data collection methods (see below) for these monkeys were identical to those for the dim-light monkeys.

2.2. Data collection

Refractive error, corneal power, ocular axial dimensions, and sub-foveal choroidal thickness were measured for both eyes of the dim-light monkeys at the onset of the experiment and periodically throughout the treatment period (every two weeks for the first seven months, then monthly until the end of the experiment). To prepare for data collection, cycloplegia and mydriasis were induced by 1% tropicamide (Akom Pharmaceuticals, IL, USA) instilled 25 and 20 minutes before the measurements. Immediately prior to measurement, monkeys were anesthetized with an intramuscular injection of ketamine hydrochloride (15 – 20 mg/kg) combined with acepromazine (0.15 – 0.20 mg/kg). Topical anesthesia (1% tetracaine ophthalmic solution, Bausch & Lomb Incorporated, Bridgewater, NJ, USA) was applied as needed. Refractive error was measured using retinoscopy by two experienced examiners and was reported as the mean spherical-equivalent of the spectacle-plane refractive correction for a 14-mm vertex distance. Corneal power in the 3-mm central region was measured along the pupillary axis with a hand-held keratometer (Alcon Auto-keratometer: Alcon, Inc., St. Louis, MO, USA). Three readings were obtained, and the spherical-equivalent corneal powers were averaged. In the case of steep corneas that were outside the measurement range of the keratometry (about 5% occurrence rate among 2-week-old monkeys), a corneal topographer (EyeSys 2000; EyeSys Vision, Inc. Houston, TX, USA) was used (95% limits of inter-instmment agreement = 0.49 to −0.37 D) (Kee et al. 2002). Anterior chamber depth, lens thickness, vitreous chamber depth, and total axial length were measured using A-scan ultrasonography along the normal to the cornea apex with a 13 MHz transducer (OTI-Scan 1000, Ophthalmic Technologies Inc., Downsview, Ontario, Canada). For the calculation of axial separations between acoustic interfaces, ultrasound velocities in monkey ocular tissue were assumed to be identical to those in the human eye (cornea and lens: 1641 m/s, aqueous and vitreous: 1532 m/s) (Byrne & Green, 2002). Ten separate measurements were taken, and then the calculated dimensions were averaged. The intra-session standard deviations of the A-scan ultrasonography measurements ranged from ± 0.04 to ± 0.08 mm between axial dimension components. Finally, choroidal thickness was measured using spectral-domain optical coherence tomography (SD-OCT; Spectralis, Heidelberg Engineering Inc., Heidelberg, Germany). Specifications for the choroidal thickness analysis have been described previously (Hung et al. 2018). Choroidal thickness data were available from all dim-light monkeys and 7 normal-light monkeys.

All rearing and measurement procedures followed the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were reviewed and approved by the University of Houston’s Institutional Animal Care and Use Committee.

2.3. Statistical methods

Statistical analyses were performed using STATA (MP 14; StataCorp, College Station, TX, USA). Mann-Whitney U test and Wilcoxon signed rank test were used respectively for between group and interocular comparisons of refractive error. Paired t-tests were used for the interocular comparisons of ocular parameters. Student’s T-tests were employed for the between-group comparisons of ocular parameters and interocular differences in refraction. These cross-sectional analyses were performed for data collected at ages corresponding to the onset of the experiment, at 155 days of age (at the approximate end of the rapid phase of emmetropization in monkeys and the midpoint of the dim-light-rearing period), and at the end of the dim-light exposure period (~310 days of age). Due to differences in the length of follow-up, the availability of normal control data at these two time-points differed (at 155 days of age, n = 41; at ~310 days of age, n = 32). Multi-level, mixed-effect model analyses were used to compare the longitudinal development of refractive error, corneal power, and choroidal thickness of the right eyes between dim-light and normal-light monkeys. Pearson correlation and linear regression were used to characterize the relationship between refractive error and vitreous chamber to corneal radius ratio (VC/CR ratio; both vitreous chamber depth and corneal radius specified in millimeters). The use of VC/CR ratios in these analyses provides a relatively simple, but more sensitive assessment of the contribution of the vitreous chamber to refractive error. In comparison to vitreous chamber depth alone, VC/CR ratios compensate, at least in part, for individual differences in corneal power and their expected effects on refractive error during emmetropization. For statistical inference, the significance level was set to 0.05.

3. Results

3.1. General observations

Dim-light-reared monkeys showed normal, age-related physical development. Body weight gain at ages of 155 ± 6 days were comparable with age-matched, normal-light-reared monkeys of the same birth year (dim-light vs. normal-light, 0.64 kg vs. 0.63 kg, t = −0.33, p = 0.74). Daily observations did not reveal any abnormal behaviors among the dim-light monkeys. After moving from normal-light housing to dim-light housing, the dim-light monkeys consistently showed higher activity levels during the daily lights-on period and became quiescent during the lights-off period, suggesting that, after transitioning to the dim-light environment, the infant monkeys maintained their light-entrained circadian rhythms that were obtained under the normal intensity, diurnal lighting condition.

Pupil diameters of the dim-light and normal-light monkeys from the same birth year were measured from photographs taken at the fronts of their cages. Only photographs showing the subjects’ eyes pointing in a horizontal direction were used (i.e., similar in orientation to the way room illumination was measured at the fronts of individual cages). The diameters in the vertical meridian were reported with the assumption that all pupils were circular. Pupil diameters of the dim-light monkeys observed at approximately 7 weeks of age were 4.98 ± 0.62 mm and 4.96 ± 0.74 mm in the right and left eyes, respectively, and were significantly larger (t = 7.51 and 6.22 for right and left eyes, respectively, p < 0.001) in comparison to age-matched normal-light-reared monkeys (left eye pupil diameter: 3.58 ± 0.43 mm). During the course of the experiment, the pupils of the dim-light monkeys remained larger than those of the normal-light-reared monkeys (dim light vs. normal light at 25 weeks of age, OD: 5.33 ± 0.37 mm vs. 3.99 ± 0.42 mm, t = 7.21, p < 0.001; OS: 5.34 ± 0.40 vs. 4.06 ± 0.42 mm, t = 6.62, p < 0.001). We speculate that the primary effects of larger pupils were to increase retinal illumination and decrease the depth-of-focus (Charman & Whitefoot, 1977; Green, Powers, & Banks, 1980). Using equation (9) from Green, Powers, and Banks (Green et al., 1980) and visual acuity estimates from Boothe, Dobson and Teller (7 cycles/degree for 7-week-old infant monkeys) (Boothe, Dobson, & Teller, 1985), the depth-of-focus at 7 weeks of age for the dim-light- and normal-light-reared monkeys was 0.20 D and 0.28 D, respectively. Note that the dioptric differences in depth-of-focus between the dim-and normal-light monkeys would likely be smaller after accounting for any reductions in visual acuity due to lower ambient lighting (van Ness & Bouman, 1967). Although the pupil-related changes in depth-of-focus could potentially reduce the efficiency of optical defocus signals, the dioptric differences in estimated depth-of-focus were relatively small. Therefore, it seems unlikely that the differences in pupil size could significantly alter the emmetropization process. It should be noted, however, that because peripheral vision can dominate central refractive development, these calculations may be an oversimplification of the effects of pupil size on emmetropization.

3.2. Refractive development

Dim-light monkeys were moderately hyperopic at the onset of the low-ambient-lighting exposure (24 ± 2 days). At this age, their median refractive errors were +3.38 D and +3.50 D in the right and left eyes, respectively (z = −1.36, p = 0.17). These starting refractive errors were not significantly different from those of age-matched, normal-light-reared monkeys (OD: +3.63 D, z = 0.46, p = 0.65; OS: + 3.38 D, z = 0.25, p = 0.81) (Table 1).

Table 1.

Refractive errors, corneal powers, and ocular axial dimensions of monkeys reared under dim light and normal light at baseline, 155 days of age and 310 days of age. The asterisks indicated significant between-group differences.

	Baseline				Midpoint of the dim-light-rearing period				End of the dim-light-rearing period
	Dim Light (24 ± 2 days, n = 7)		Normal Light (24 ± 4 days, n = 41)		Dim Light (155 ± 6 days, n = 7)		Normal Light (148 ± 10 days, n = 41)		Dim Light (310 ± 21 days, n = 7)		Normal Light (306 ± 10 days, n = 32)
	OD	OS	OD	OS	OD	OS	OD	OS	OD	OS	OD	OS
Median RE (D)	+ 3.38	+ 3.5	+ 3.63	+ 3.38	+3.13*	+ 3.31*	+ 2.31*	+ 2.44*	+ 2.75*	+ 3.00*	+1.78*	+1.75*
Corneal power (mean ± SD, D)	61.52 ± 1.15	61.46 ± 1.12	61.52 ± 2.03	61.56 ± 1.89	55.76 ± 1.76	55.61 ± 1.66	55.80 ± 1.66	55.88 ± 1.72	54.02 ± 1.52	54.07 ± 1.78	54.10 ± 1.76	54.05 ± 1.78
Anterior chamber depth (mean ± SD, mm)	2.54±0.13	2.54 ± 0.11	2.65 ± 0.29	2.64 ± 0.30	3.11 ± 0.11	3.08 ± 0.09	3.06 ± 0.28	3.09 ± 0.29	3.21 ± 0.12	3.18 ± 0.14	3.17 ± 0.22	3.14 ± 0.26
Lens Thickness (mean ± SD, mm)	3.73 ± 0.10	3.72 ± 0.11	3.50 ± 0.29	3.52 ± 0.00	3.66 ± 0.13	3.67 ± 0.10	3.63 ± 0.21	3.61 ± 0.21	3.67 ± 0.13	3.70 ± 0.17	3.60 ± 0.12	3.62 ± 0.13
Vitreous chamber depth (mean ± SD, mm)	8.51 ± 0.34	8.5 ± 0.32	8.64 ± 0.31	8.64 ± 0.31	9.66 ± 0.50	9.61 ± 0.46	9.85 ± 0.31	9.85 ± 0.31	10.22 ± 0.63	10.25 ± 0.59	10.53 ± 0.37	10.55 ± 0.35

^*:Significant between-group difference.

Whereas the normal-light monkeys exhibited age-related reductions in hyperopia that are associated with emmetropization, the dim-light monkeys showed little evidence of emmetropization. Figure 2 illustrates the refractive errors of individual dim-light monkeys as a function of age. The filled and open symbols represent refractive errors for the right and left eyes, respectively. The refractive errors of the right eyes of individual normal-light monkeys are represented by the grey lines in each plot. From these plots, we see that the refractive development of the dim-light monkeys was highly variable, but there was no evidence for systematic myopic changes. Specifically, monkeys 692 (Figure 2A) showed qualitatively normal age-associated reductions in hyperopia in both eyes; the data for this animal remained in the middle of the range of refractions for the normal control monkeys throughout the observation period. On the other hand, the remaining monkeys showed significant deviations from normal-light monkeys in their refractive development. Both monkeys 740 and 734 (Figure 2B and 2C) showed small reductions in hyperopia over the first 6 weeks of treatment, however, both of these monkeys then maintained roughly the same degree of hyperopia for the rest of the treatment period so that at the end of the observation period these monkeys were more hyperopic than the majority of normal monkeys. Interestingly, monkeys 694, 735, 741 and 709 (Figure 2, bottom row) exhibited hyperopic shifts for much of the observation period.

Figure 2.

Refractive errors of individual dim-light monkeys plotted as a function of age. Filled and open symbols represent the refractive errors in the right and left eyes, respectively. Refractive errors for the right eyes of normal-light monkeys (n = 41) are plotted in the same manner as thin, solid lines. Only one dim-light monkeys showed age-related reductions in hyperopia that were consistently near the middle of the normal range (panel A). Other monkeys either maintained the same degree of hyperopia (panels B and C) or developed progressive hyperopic shifts in refractive error (second row).

Figures 3A and 3B compare the average relative changes in refractive error between dim-light and normal-light monkeys up to 155 days of age. Multi-level, mixed-effect model analysis showed that rearing under dim light increased the variability in refractive change and shifted the mean refractive change trajectory in the less myopic/more hyperopic direction (linear- and quadratic-effect coefficients, OD: z = 3.24, p < 0.01 and z = −2.39, p = 0.02; OS: z = 3.36, p < 0.01 and z = −2.14, p = 0.03). At the midpoint of the rearing period (155 days), the dim-light monkeys were significantly more hyperopic than age-matched, normal-light-reared monkeys (Figure 3C; dim-light vs. normal-light, OD: + 3.13 D vs. + 2.31 D, z = −2.24, p < 0.03; OS: +3.3ID vs. + 2.44 D, z = −2.49, p = 0.01). Continued exposure to dim light did not change the comparative refractive status between the dim-light and normal-light monkeys, nor did it produce any instances of myopia. At the end of the dim-light exposure period (Figure 3D, 310 days), six of the seven dim-light monkeys exhibited refractive errors that were more hyperopic than 95% of the normal-light monkeys (95% confidence interval for the right eyes: + 1.56D - + 2.18D, n = 32), and the dim-light monkeys remained significantly more hyperopic than age-matched normal-light monkeys (dim-light vs. normal-light, OD: +2.75 D vs. +1.78 D, z = −2.78, p < 0.01; OS: + 3.00 D vs. + 1.75 D, z = −2.93, p < 0.01).

Figure 3.

Panels A and B: Mean refractive-error changes relative to baseline values plotted as a function of age for the right (filled symbols) and left eyes (open symbol) of the normal-light and dim-light monkeys, respectively. Dashed lines represent zero change in refractive error. Error bars represent ±1 standard deviation from the mean refractive-error change. Normal-light monkeys (panel A) exhibited age-related myopic shifts in both eyes, whereas dim-light monkeys developed highly variable refractive changes that averaged between 0 ~ −0.5D throughout the experiment (panel B). In comparison to the age-matched normal-light monkeys, these changes resulted in more hyperopic refractive states in the dim-light monkeys at both ~155 days (Panel C) and ~310 days of their age (Panel D).

Refractive development in the dim-light monkeys was associated with larger than normal interocular differences (IOD) in refractive error. This characteristic can be seen from the course of refractive development, particularly in monkeys that failed to exhibit an emmetropization-associated reduction of hyperopia. Specifically, for monkeys 734, 694, 735 and 741 (Figures 2C – 2E), obvious interocular difference in refractive errors (≥ 0.5 D) manifested on multiple occasions during the dim-light exposure period. For monkey 709 (Figure 2G), binocular hyperopic shifts were accompanied by a concurrent, progressive increase in anisometropia throughout the dim-light exposure period. The anisometropias in the dim-light monkeys frequently fell outside the ±2 SD range of those for the normal-light monkeys over the course of the experiment (Figure 4A). The frequency distribution of the magnitude of anisometropia for the dim-light monkeys (Figure 4C) was skewed towards larger interocular differences and less leptokurtic than that for the normal-light monkeys (Figures 4B), indicating that dim-light monkeys more frequently developed larger anisometropias. At the age of 155 ± 6 days, the mean magnitude of anisometropia was significantly greater in dim-light monkeys than in normal-light monkeys (0.38 ± 0.26 D vs. 0.18 ± 0.18 D, t = 2.45, p = 0.02). Note that, despite the apparent struggle in maintaining interocular balance, refractive errors in the two eyes of dim-light monkeys remained highly correlated (Pearson correlation, r = 0.97, p< 0.001).

Figure 4.

Panel A: Interocular differences (IOD) in refractive error of individual dim-light monkeys plotted as function of age. The grey area represents the ± 2 standard deviation range of IODs for the normal-light monkeys. Panels B and C: Frequency distributions for the magnitude of IODs in refractive error for the normal-light and dim-light monkeys, respectively. The first columns in these panels represent the percentage of samples that had zero interocular difference.

3.3. Corneal power and ocular axial components

Comeal power development appeared unaffected by reduced ambient lighting. At ages corresponding to the onset of the experiment, corneal powers of the dim-light monkeys were well-matched in their two eyes (OD: 61.52 ± 1.15 D, OS: 61.46 ± 1.12 D, t = 0.28, p = 0.79) and similar to those of the normal-light monkeys (OD: 61.52 ± 2.03 D, OS: 61.56 ± 1.89 D, t = 0.24, p = 0.80; between-group comparisons, OD: t = 0.01, p = 0.99; OS: t = 0.14, p = 0.89). Figure 5A plots the corneal powers of the dim-light and normal light monkeys as a function of age. As the animals grew, corneal powers in both the dim-light and normal-light monkeys decreased exponentially without any discemable between-group differences in the rate of change (z = 1.08, p = 0.28 and z = −0.56, p = 0.58 for the linear- and quadratic-effect coefficients, respectively). There were no significant differences in average corneal powers between dim-light and normal-light monkeys at either the midpoint (dim-light vs. normal-light, OD: 55.76 ± 1.76 D vs. 55.80 ± 1.66 D, t = 0.07, p = 0.95; OS: 55.61 ± 1.66 D vs. 55.88 ± 1.72 D, t = 0.40, p = 0.69. Figure 5B) or the end of the dim-light exposure period (OD: 54.02 ± 1.52 D vs. 54.10 ± 1.73 D, t = 0.12, p = 0.91; OS: 54.07 ± 1.78 D vs. 54.06 ± 1.76 D, t = −0.01, p = 0.99). These results show that reduced ambient lighting did not exacerbate nor attenuate the age-associated corneal flattening in infant rhesus monkeys.

Figure 5.

A. Corneal powers for the right (filled symbols) and left eyes (open symbols) of the dim-light monkeys plotted as a function of age, along with the corneal powers of the right eyes of normal control monkeys (represented by the thin solid lines). With one exception, all dim-light monkeys’ corneal power development fitted well into the pattern observed in normal-light monkeys. B. Corneal powers of the dim-light monkeys at 155 days of age compared with those of age-matched, normal-light monkeys. Symbols above and below the error bars represent the outliers. The horizontal line inside the boxplot indicates the median corneal power, whereas the horizontal dashed line across the panel represents the mean corneal power of normal monkeys. There was no statistically significant difference in corneal power between the two groups at 155 days of age.

In both dim-light and normal-light monkeys, age-related changes in refractive error were associated with alterations in ocular axial dimensions. Ocular dimensions measured at the onset of the experiment, the midpoint of the dim-light period (155 days), and the end of the dim-light exposure period (310 days) are compared between dim-light and normal-light monkeys in Table 1. In brief, the ocular dimension components were similar in the dim-light and normal-light monkeys at ages corresponding to the onset of the experiment (p > 0.05). At 155 and 310 days of age (Table 1), the anterior chamber depths and lens thicknesses of the dim-light monkeys were similar to those of age-matched, normal-light-reared monkeys (p > 0.05). The average vitreous chamber depth, in contrast, was approximately 0.2 mm shorter in the dim-light monkeys than in the normal-light monkeys at 155 days of age, although this difference was not statistically significant (dim-light vs. normal-light, OD: 9.66 ± 0.50 vs. 9.85 ± 0.31 mm, t = 1.41, p = 0.17; OS: 9.61 ± 0.46 vs. 9.85 ± 0.31 mm, t = 1.82, p = 0.08). Note that the magnitude of the between-group difference in vitreous chamber depth was optically substantial in the context of monkey eyes, and the relative differences in vitreous chamber depth agreed with the relative differences in refractive status between these two groups (Table 1). In addition, refractive errors observed at 155 days were significantly correlated with axial length (r² = −0.27, p = 0.04) and vitreous chamber depth (r² = −0.43, p < 0.001). These observations suggested that vitreous chamber depth was plausibly the determinant for the variation in refractive errors in dim-light monkeys.

As noted in the methods, a more sensitive assessment on the axial nature of refractive error can be achieved using the VC/CR ratio. For the dim-light monkeys, changes in the VC/CR ratios were highly correlated in changes in vitreous chamber depth (r = 0.77, p < 0.01), suggesting that this metric was a valid representation of axial eye growth. In addition, the changes in VC/CR ratios were inversely correlated with changes in refractive error (r = −0.75, p < 0.001). More importantly, as illustrated in Figure 7, at the end of the regular experiment period, there was a strong linear correlation between refractive error and the VC/CR ratio in dim-light monkeys (r = −0.89, p < 0.01); eyes with higher degrees of hyperopia had lower VC/CR ratios, which accounted for 80% of the variance in refractive error (linear regression, r² = 0.8, p < 0.01). These data, in association with the corneal power changes described above, indicated that dim-light monkeys’ refractive states were mostly determined by ocular axial elongation.

Figure 7.

Refractive error for the right (filled symbols) and left eyes (open symbols) of the dim-light monkeys at 155 days of age plotted as a function of the vitreous chamber depth to corneal radius ratio (VC/CR ratio). Both vitreous chamber depth and corneal radius in the calculation of the VC/CR ratio are specified in millimeters. The two eyes from the same monkey are represented by the same shaped symbol.

3.4. Choroidal thickness changes

Rearing monkeys under dim light caused sustained choroidal thickening in the sub-foveal region. At ages that correspond to the onset of the experiment, sub-foveal choroidal thickness was not significantly different between dim-light and normal-light monkeys (average ± SEM choroidal thicknesses for dim-light vs. normal-light monkeys: OD: 122.76 ± 8.24 μm vs. 121.91 ± 7.30 μm, t = −0.08, p = 0.94, OS: 118.28 ± 9.02 vs. 119.75 ± 6.02, t = 0.14, p = 0.89). Multi-level, mixed-effect model analysis on the right eye data showed that, in both dim- and normal-light monkeys, the choroid underwent age-related thickening (average thickening rate ± SEM = 0.17 ± 0.04 μm/day from the onset of experiment, z = 4.33, p < 0.001). Greater choroidal thickness increases were also associated with more hyperopic refractive errors (3.3 ± 1.5 μm per diopter of relative hyperopia, z = 2.28, p = 0.02). When the influences of age and refractive error to choroidal thicknesses were accounted for, dim-light rearing caused additional relative choroidal thickening in comparison to normal-light rearing, of which the rate was 0.19 μm per day at the onset of the experiment (SEM = ± 0.06, z = 3.11, p = 0.002) and gradually decreased as dim-light rearing continued (z = −2.67, p = 0.01). The resulting choroidal thickness changes are illustrated in Figure 8, in which the trajectories of mean choroidal thickness changes of the dim-light and normal-light monkeys became roughly parallel following a rapid departure after the onset of the experiment. At 155 days of age, the changes in choroidal thicknesses were greater in dim-light monkeys than in the age-matched normal-light monkeys (dim-light vs. normal-light: 167.1 ± 6.2 μm vs. 146.9 ± 6.43 μm, t = 2.21, p = 0.04).

Figure 8.

Changes in sub-foveal choroidal thickness relative to the onset of the experiment for dim-light (red symbols) and normal-light (black symbols) monkeys plotted as functions of age. For a given animal the data for the right and left eyes were averaged. The symbols in the figure represent group averages; the error bars represent ±1 standard error of the mean. Compared with normal-light monkeys, dim-light monkeys showed sustained choroidal thickening in the sub-foveal region.

4. Discussion

We examined the effects of low intensity ambient lighting on normal emmetropization and the underlying ocular development in infant rhesus monkeys. We found that exposing infant rhesus monkeys to reduced ambient lighting did not cause myopia; instead, it caused considerable inter-subject and inter-ocular variability in refractive error and reduced the probability that monkeys would emmetropize from the normal, moderate levels of hyperopia found at infancy. Low intensity ambient lighting increased sub-foveal choroidal thickness, but did not cause systematic alterations in corneal power, anterior chamber depth, or lens thickness. The depth of vitreous chamber remained the primary determinant of refractive error.

4.1. Effects of dark-rearing versus dim-light-rearing

Although rearing animals in complete darkness has also been reported to alter refractive development, the alterations produced by constant darkness (dark rearing) and dim-light-rearing strategies are qualitatively different. In chickens, dark rearing greatly increases the between subject variability in refractive error, most commonly resulting in hyperopia. The hyperopic errors in dark-reared chickens are interesting because they are associated with increases in axial length and vitreous chamber depth (Gottlieb, Fugate-Wentzek, & Wallman, 1987; Troilo & Wallman, 1991; Yinon & Koslowe, 1986), which would normally result in relative myopic shifts. The manifest hyperopia in dark-reared chick eyes can be attributed primarily to dramatic corneal flattening (Gottlieb et al., 1987; Troilo & Wallman, 1991; Yinon & Koslowe, 1986). In contrast, as documented above, rearing chickens in dim diurnal ambient lighting consistently produced relative myopic refractive errors that were associated with corneal flattening, but dominated by comparatively much larger increases in vitreous chamber depth (Cohen et al., 2011).

Although there is only a small amount of data available for non-human primates, the most obvious effect of dark rearing in monkeys appears to be an increase in between subject variability in refractive errors. Raviola and Wiesel reported that the refractive errors for the control eyes of two dark-reared, monocularly form-deprived monkeys were +2 D at the end of 9- and 12-month treatment periods (corneal powers appeared to be normal) (Raviola & Wiesel, 1978). As illustrated by the normal control data in Figure 2, normal monkeys typically exhibit low degrees of hyperopia at comparable ages. For example, at about 300 days of age, our average normal monkey had a refractive error of +1.87 ± 0.86 D, which suggests that dark rearing does not alter refractive development in monkeys. However, in the only longitudinal study of dark rearing in monkeys, Guyton et al. reported that 2 of 5 dark-reared monkeys developed large degrees of myopia, 2 showed large shifts in the hyperopic direction, and 1 monkey exhibited relatively stable hyperopic refractive errors throughout the treatment period (Guyton et al., 1989). The report did not include ocular parameter measurements, thus the nature of these dark-rearing-induced refractive errors is not known. In contrast to these dark-rearing results, the majority of the dim-light-reared monkeys in this study exhibited relative hyperopic errors that were associated with shorter vitreous chambers; no dim-light monkeys showed relative myopic shifts in refractive error.

The observed differences in the effects of dark rearing and dim-light rearing in both chickens and monkeys are probably not surprising. In particular, the absence of visual input under dark-rearing conceptually precludes emmetropizing responses, a process known to be visually regulated. In this respect, the overall pattern of refractive development in both chickens and monkeys is consistent with unregulated, open-loop behavior. In addition, dark-rearing deprives the eye of the visual signals necessary to maintain a number of ocular circadian rhythms that are important for normal ocular growth and refractive development (for reviews, see Chakraborty et al., 2018; Nickla, 2013).

4.2. Inter-species comparisons of the effects of dim light on refractive development

The refractive outcomes in our dim-light monkeys were different from those previously observed in chickens (Bercovitz et al., 1972; Cohen et al., 2011,2012; Lauber & Kinner, 1979). In the early studies involving chickens, animals reared under very low intensity lighting (~0.12 lux to ~0.34 lux) developed larger equatorial diameters and longer axial lengths, and became myopic relative to those reared under relatively “normal” illumination levels (Bercovitz et al., 1972; Lauber & Kinner, 1979). Note, however, that the “normal” lighting levels in these experiments (Bercovitz et al., ~ 31 lux; Lauber et al., ~ 6.8 lux) approximated the “dim” ambient lighting level in the later studies of Cohen et al. (50 lux), in which the association between lower lighting level and more myopia was also observed. Together, these studies showed that the myopiagenic effect of low ambient lighting in chickens was consistent over a relatively large range of the light-intensity continuum. The myopiagenic effects of dim ambient lighting were also qualitatively consistent and robust across subjects. For example, dim-light-reared chickens became relatively less hyperopic than those reared under 500 lux lighting by the 20^th – 30^th treatment day and by the 90^th treatment day, all but one of 13 dim-light-reared chickens had developed absolute myopia (Cohen et al., 2011,2012). For our dim-light-reared monkeys, however, systematic myopic changes were not observed at any point during the experiment. It is unlikely that the failure to observe myopia in our monkeys was related to the length of exposure to dim lighting. We deliberately employed a long treatment period in order to increase the possibility that we would detect any potential myopic shifts. Taking into account interspecies differences in the relative rates of ocular growth (Troilo et al., 2019), the duration of our treatment period substantially exceeded the 90-day dim-light exposure that consistently produced myopia in chickens (Cohen et al., 2011, 2012). Given that the illumination levels in our study were very similar to those in Cohen et al.’s, our results indicate that low ambient lighting by itself does not always induce myopia, at least not in rhesus monkeys.

4.3. Effects of dim, elevated and typical laboratory lighting on emmetropization in monkeys.

Based primarily on the results in chickens, it has been proposed that ambient light intensity quantitatively changes the course and/or endpoint of emmetropization with dim and elevated lighting promoting myopia and hyperopia, respectively (Norton & Siegwart, Jr., 2013). However, that may not be case in monkeys. We previously showed that elevated ambient lighting did not alter emmetropization in monkeys reared with unrestricted vision, nor did it alter the time course or the degree of compensation to imposed hyperopic defocus in treated eyes or the course of emmetropization of the fellow, untreated eyes, i.e., monkeys reared under elevated ambient lighting emmetropized normally (Smith III et al., 2013). On the other hand, the present study showed that monkeys reared with unrestricted vision under dim light largely failed to emmetropize. The absence of consistent emmetropizing responses in dim-light-reared monkeys suggests that there might be an ambient lighting intensity threshold for the initiation and regulation of emmetropization. Supra-threshold ambient lighting (i.e., typical laboratory lighting and elevated lighting) seems to ensure a high probability of normal emmetropization, speculatively by allowing accurate encoding and/or proper amplification of signals. In contrast, sub-threshold lighting levels may reduce the accuracy in signal encoding and/or the gain in signal processing, leading to a dampened emmetropization process. A possible consequence of such alterations is that hyperopic and myopic defocus, which are thought to regulate refractive development (Hung, Crawford, & Smith III, 1995; Schaeffel, Glasser, & Howland, 1988; for a review see Troilo et al., 2019), must be optically stronger (higher in nominal power) in order to produce functionally adequate “go” and “stop” signals under dim light. In this respect, it is possible that the larger-than-normal fluctuations in anisometropia observed in the dim-light monkeys reflected an increase in the threshold defocus level required to “trigger” mechanisms that are responsible for maintaining isometropia or overcoming anisometropia.

4.4. The effects of dim light on corneal power

Reduced ambient lighting had little effect on corneal power in rhesus monkeys. The degree and time course of the age-related reduction in corneal power were similar in monkeys reared under reduced, normal (Qiao-Grider et al., 2007), and elevated ambient lighting levels (Smith III et al., 2013). Such stability is important because a relatively small amount of defocus is sufficient to consistently induce compensatory refractive changes in rhesus monkeys (Hung, Crawford, & Smith III, 1995). On the other hand, in the studies that reported dim-light-associated myopic shifts in chickens, the myopic shifts appeared to be associated with corneal flattening (Bercovitz et al., 1972; Cohen et al., 2011; Harrison et al., 1968; Harrison & McGinnis, 1967). Specifically, in the study of Cohen et al., substantial differences in corneal power (2.2D flatter) were observed between chickens reared under normal- and dim-light on as early as the 20^th day of exposure and clearly preceded any meaningful differences in axial length (Cohen et al., 2011). Following the reasoning in section 4.3, these early reductions of corneal power might have augmented the reduced regulatory signals by increasing the magnitude of hyperopic defocus, thereby triggering the emmetropization process that might otherwise be quiescent at low lighting levels. It is possible that these animals subsequently developed myopia because once axial elongation was triggered, the resulting low levels of myopic defocus were not sufficient to stop axial elongation. If this is correct, we would predict that optically imposed defocus might similarly result in higher than expected degrees of myopia in infant monkeys.

4.5. Possible role of dopamine in light-intensity-induced refractive changes

Animal studies suggest that the protective effects of elevated lighting against myopia in children is associated with retinal dopamine (Rose et al., 2008). It has been established that the synthesis, release, and turnover of retinal dopamine can be induced by ambient light (Iuvone, Galli, Garrison-Gund, & Neff, 1978) in an intensity-dependent manner (Brainard & Morgan, 1987; Proll, Kamp, & Morgan, 1982). In chicks and rhesus monkeys, elevated laboratory lighting inhibits form-deprivation myopia (Ashby and Schaeffel 2010; Smith III, Hung, and Huang 2012). The role of dopamine in this process was evident in that intravitreal injection of spiperone, a D₂ receptor antagonist, abolished these protective effects in chicks (Ashby and Schaeffel 2010). Cohen et al. further demonstrated that, in chicks, both refractive development and the production of vitreal 3, 4-dihydroxyphenylacetic acid (DOPAC, the primary metabolite of dopamine) were light-intensity dependent. Specifically, higher ambient-light intensities were associated with less myopia and higher dopamine levels (Cohen et al., 2012). Largely based on this observation, Norton and Siegwart (Norton & Siegwart, Jr., 2013) proposed a working model, in which light-intensity dependent changes in retinal dopaminergic activity allows ambient light level to quantitatively alter the end point for refractive development.

There are, however, challenges to this theory. For example, Stone et al. did not find correlations between dopamine/DOPAC, outdoor rearing, and the transient inhibitory effects of outdoor rearing on form-deprivation myopia (Stone et al., 2016). With respect to the light intensity dependency of retinal dopamine, a recent study by Landis et al. (Landis et al. IOVS 2019;60:ARVO E-Abstract 3152) showed that, although retinal DOPAC production in mice was dependent on ambient light intensity, under long-term exposure to altered ambient illumination levels (0.005 lux, 50 lux, and 15,000 lux during the light phase of the diurnal lighting cycle), retinal dopamine levels remained similar across the different lighting levels. These findings suggested that, under chronic exposures to different ambient illuminations, adaptive mechanisms might respond and maintain a relatively constant retinal dopamine level, thus retinal dopamine availability might not always be light-intensity dependent. With respect to the potential light dependency of refractive development, we found in this and in a previous study (Smith III et al., 2013) a lack of apparent correlation between ambient lighting intensity and the degree of refractive error in rhesus monkeys reared without visual restrictions. Both lines of evidence suggest that the effects of lighting level on the degree of refractive error may not be explained by the light-intensity dependency of a single molecule. It is possible that retinal dopamine is part of a more complex molecular mechanism that mediates the lighting-intensity effect on refractive development.

4.6. Dim-light-induced choroidal thickness changes

To the best of the authors’ knowledge, this is the first report of sustained, dim-light-induced choroidal thickening in non-human primates. In the early works of Harrison and McGinnis (Harrison & McGinnis, 1967), choroidal thickness increased dramatically (3-4 folds) in low intensity blue-light-reared chickens, although the results might be confounded by the narrow spectral compositions of the ambient lighting. In this respect, a subsequent study conducted by the same group not only failed to replicate these choroidal changes, but rather found thinner choroids in comparison to the normal-light-reared chickens (Harrison et al., 1968). More recently, Lan el al. found that the chicken choroid slightly thinned after moving from normal (500 lux) to bright ambient lighting (15,000 lux). When these chickens were subsequently removed from bright light, placed under normal lighting for 2 hours, and then exposed to darkness for another 2 hours, their choroidal thicknesses increased (Lan, Feldkaemper, & Schaeffel, 2013). In a somewhat analogous manner, in humans, nightly exposure to 1,000 lux lighting instead of 150 lux lighting was reported to induce thinning of the sub-foveal choroid (Ahn et al., 2017), whereas dark adaptation was reported to increase sub-foveal choroidal thickness (Alagöz et al., 2016). Despite the substantial methodological differences between these studies, these studies suggest a link between relative choroidal thickening and lower ambient lighting levels.

Did dim-light-associated choroidal thickening influence refractive development? In many species, the choroid thickens in response to myopic defocus and thins in response to hyperopic defocus (Nickla and Wallman 2010). At least in chicks, the magnitude of defocus-induced choroidal thickness change was sufficiently large to serve as an intermediate-acting mechanism for reducing refractive error (Wallman et al., 1995; Wildsoet & Wallman, 1995). For our dim-light monkeys, however, it is unlikely that the relative increases in choroidal thickness were induced by defocus, as the animals were hyperopic throughout the experiment. It is also unlikely that the direct optical effects induced by relative choroidal thickening constituted a significant dioptric force for refractive regulation. Although the choroid of the dim-light monkeys rapidly thickened relative to those of the normal-light monkeys, the mean difference in thickness observed at 155 days of age (~20 μm) remained small and was not sufficient to substantially alter the eye’s effective refractive state. In fact, even if such a difference took place in 24-day-old monkeys (in which the increase in choroidal thickness would effectively augment the natural hyperopia) (Qiao-Grider et al., 2007), the resulting relative hyperopia (< + 0.25 D) associated with the observed degree of choroidal thickening would appear to be negligible considering the level and between-subject variability in refractive error that is naturally present early in life (in our experiment, the mean ± standard deviation of refractive error at baseline = +3.82 ± 1.83 D). Therefore, the dim-light-associated relative choroidal thickening did not appear to contribute significantly to the dioptric changes observed in the dim-light monkeys. Although the changes in choroidal thickness appeared to be predictive of the hyperopic shifts observed over the course of the treatment period in some animals, the underlying mechanisms and implications remain to be investigated.

4.7. Clinical implications, limitations, and future directions

Although low ambient lighting is not always myopiagenic for primates, it does appear to compromise normal emmetropization in rhesus monkeys. From a refractive error management standpoint, if a light-intensity threshold for emmetropization exists, determining this level has important implications. For rhesus monkeys, this critical level appears to be within a narrow range between about 50 and 500 lux, (approximately the average “dim” and “normal” lighting levels in our laboratory), which is somewhat alarming because humans can encounter comparable low light levels indoors (Ostrin, 2017). If human refractive development responds to dim ambient lighting in the same way as rhesus monkeys, reduced indoor lighting may indeed be a risk factor of abnormal refractive development.

A limitation of our study was that the illumination level, as well as the duration of exposure, was not representative of real-world scenarios. In addition, whereas transitioning between relatively lower and higher ambient lighting frequently takes place in daily life; our subjects were deprived of such opportunities. Transitioning between ambient lighting conditions might be physiologically impactful because temporal contrast might serve as an additional trigger for retinal dopamine release. These limitations suggest that the observed refractive effects of dim lighting might be largely exaggerated. Nonetheless, our study highlighted the importance of proper indoor lighting, especially for young children who are in the early stages of emmetropization.

From our findings, it is not clear from a mechanistic perspective how dim ambient lighting influences refractive development. Specifically, our data did not show whether the ocular regulatory mechanisms were not detecting defocus signals or that these mechanisms were simply not responding to appropriately encoded signals. This issue has practical significance because current optical interventions for myopia are either based on or thought to be related to the response of regulatory mechanisms to imposed defocus signals (for a review see Troilo et al., 2019). For example, if dim ambient lighting causes growing eyes to misinterpret the nominal sign or magnitude of optical regulatory signals, one would expect that dim lighting has the potential to render optical treatment strategies ineffective. In this respect, it will be important to investigate the effects of dim ambient lighting on the various forms of experimentally induced refractive errors in non-human primates.

Figure 6.

Ocular axial parameters for the dim-light and normal-light monkeys at 155 days of age. Data from the right (red, filled symbol) and left eyes (red, open symbol) of the dim-light monkeys are plotted on the right side of each panel. Data from the right eyes of individual normal-light monkeys (grey filled symbols) are plotted on the left in each panel. The symbols with error bars to the right of the individual data represent the mean and ±1 standard deviation from the mean for the corresponding ocular parameter. There were no statistically significant differences in ocular axial parameters between the dim-light and normal-light monkeys.