IMI: The Role of Light in Refractive Development and Myopia: Evidence from Animal and Human Studies

Abstract

Spending time outdoors is consistently associated with delayed myopia onset in children and has been incorporated into prevention programs. While the underlying mechanisms remain under investigation, substantial evidence supports sunlight exposure as a key contributing factor. This review evaluates the evidence supporting this association. Animal studies demonstrate that light characteristics-such as intensity, chromaticity, and photoperiod-can influence refractive development, often postulated to function through modulation of the dopaminergic system. However, translating these findings to humans is challenging due to limited data. Evidence remains insufficient regarding how specific light properties-including intensity thresholds, exposure durations, spectral composition, and temporal patterns-affect human myopia. Consequently, although clinical recommendations for outdoor time (e.g., 2 hours daily) are well supported by epidemiological studies and widely endorsed, current literature does not yet support evidence-based guidelines concerning specific characteristics of light exposure. Addressing this gap requires further randomized controlled trials using standardized wearable technologies to better quantify children's visual environments and identify light-related cues relevant to myopia development. Interest in light-based therapies is growing, but most interventions remain in early stages of development/testing. Due to limited efficacy data or unresolved safety concerns, no clinical recommendations can currently be made. Interpreting light-related research-especially from across animal models-requires caution. Species-specific differences in ocular transmittance and opsin distribution/tuning complicate the translation of chromatically related findings between species and to humans. Moreover, the term "white light" can be misleading, as artificial sources vary spectrally from each other and from sunlight, resulting in varied patterns of opsin activation. Studies should therefore attempt to report not only light intensity, but also the radiant power emitted at each wavelength to enable meaningful comparisons.

Figure 1.

Normalized photopigment absorbance spectra and ocular transmittance across species. Absorbance curves were calculated as described by Govardovskii et al. using λ max values for each species. Ocular media transmittance values (cornea/lens/aqueous humor/vitreous humor), as best known, are displayed as dotted lines for each species (A) human,^, (B) rhesus monkey,^, (C) tree shrew,^,, (D) guinea pig,^, (E) mouse,^, and (F) chicken^71,109). For reference, the visual spectrum for humans is displayed as a color bar under each species panel. OPN3 and OPN5, as well as the effects of ellipsoid membranes and oil droplets are not plotted due to limits in the characterization of their absorbance spectra across all species. It is also important to note that λ max and transmittance data are currently limited for some species, and, as such, these values may be subject to revision as additional data become available. Key: SWS, short-wavelength sensitive cone opsin (blue and violet lines); OPN4, melanopsin (light blue line); Rho, rhodopsin (black line); MWS, medium-wavelength sensitive cone opsin (green line); DC, double cone opsin, chick only (yellow line); LWS, long-wavelength sensitive cone opsin (red line). Figure adapted from Reference .

Figure 2.

Spectral distribution of different “white light” sources. The irradiance at wavelengths between 300 and 800 nm is presented for (A) approximately 90,000 lux sunlight (as measured in the early afternoon during summer in Canberra, Australia; correlated color temperature (CCT) = approximately 6000 K), (B) 500 lux fluorescent light (CCT = approximately 4000), (C) 500 lux incandescent light (CCT = approximately 2700), (D) 500 lux halogen light (as used in initial studies examining the role of light in experimental myopia^,; CCT = approximately 3000), (E) 500 lux “cool white” using light emitting diodes (LEDs; CCT = 6000–6500), (F) 500 lux “warm white” light using LEDs (CCT = 3000–3500), (G) 500 lux “neutral white” light using LEDs (CCT = approximately 4500), and (H) 500 lux white light generated from red-green-blue LEDs (CCT = approximately 4000). For reference, the human visual spectrum is displayed as a color bar under each panel. Figure adapted from Reference .

Figure 3.

Differences in total photons presented to the retina across species for two different light sources. The total photons presented to each classical photoreceptor type were calculated between 300 and 800 nm for (A) human,^,,, (B) rhesus monkey,^,,, (C) tree shrew,^–, (D) guinea pig,^,,, (E) mouse,^, and (F) chicken.^,, Total photons available to the retina were calculated for (i) approximately 90,000 lux sunlight (as measured in the early afternoon during summer in Canberra, Australia; CCT = approximately 6000 K) and (ii) 500 lux fluorescent light (CCT = approximately 4000 K). The ratio of total photons available to each cone photoreceptor type is presented in the top right of each panel. Ratios were calculated by normalizing to the photons presented to the LWS cone, which is given a value of 1. Ratios are presented for tetrachromatic (SWS1:SWS2:MWS:LWS), trichromatic (SWS:MWS:LWS), and dichromatic (SWS:LWS) species. Total available photons were calculated as described by Wilby and colleagues, taking into account photoreceptor spectral sensitivity (Fig. 1), ocular light transmission, and the multiplication factor of the (eye's aperture pupil area:retinal area ratio) for each species (references given for each species when first listed above). The peaks depicting total effective photons for rods (rhodopsin) and MWS/LWS cones overlap in some species and may not be visible. For reference, the human visual spectrum is displayed as a color bar under each panel. Key: Rhodopsin (black line), short-wavelength sensitive cone opsin (blue and violet [chicken only] lines); medium-wavelength sensitive cone opsin (green line); and long-wavelength sensitive cone opsin (red line). Figure adapted from Reference .

Figure 4.

Differences in cone opsin activation patterns in response to four different light sources as demonstrated for human s and chicks. Total effective photon estimates were calculated for (A) humans and (B) chicks exposed to four different light sources: (i) approximately 90,000 lux sunlight (as measured in the early afternoon during summer in Canberra, Australia; CCT = approximately 6000 K), (ii) 500 lux fluorescent light (CCT = approximately 4000 K), (iii) 500 lux “warm white” LED light (CCT = approximately 3000–3500 K), and (iv) 500 lux “cool white” LED light (CCT = approximately 6000–6500 K). Total effective photon estimates were calculated between 300 and 800 nm following the protocol described by Wilby and colleagues. These calculations took into account photoreceptor spectral sensitivity (Figs. 1, 3), ocular light transmission, the multiplication factor of the eye's aperture (pupil area:retinal area ratio), photoreceptor cell light transmission (representative of the total photons able to reach a photopigment)^, and the relative ratios^, of each cone type (to take into account the differences in cone numbers). For reference, the human visual spectrum is displayed as a color bar under each panel. The ratio of total photons available to each cone photoreceptor type is presented in each panel title within the brackets. Ratios were calculated by normalizing to the photons presented to the LWS cone, which is given a value of 1. Key: Short-wavelength sensitive cone opsin (blue and violet [chicken only] lines), medium-wavelength sensitive cone opsin (green line), and long-wavelength sensitive cone opsin (red line). Figure adapted from Reference .

Figure 5.

Differences in the spectral power distribution (SPD) of two “white light” sources with matched illuminance, and their corresponding cone opsin activation profiles in the chick retina. Presented between wavelengths of 300 and 800 nm are the SPDs of 2 light sources matched for illuminance (approximately 20,000 lux): (A.i.) LED “neutral white” light (CCT = approximately 4500 K) and (B.i.) halogen light (CCT = approximately 3000 K). In both panels, the gray dotted line represents the human photopic luminosity function (peaking at 555 nm), which is used to calculate illuminance in lux. Although both sources emit approximately 20,000 lux, the halogen light delivers more than twice the number of photons (LED = 1.3 × 10¹⁴ photons versus Halogen = 3.35 × 10¹⁴ photons). Due to differences in SPD, the spectral distribution of photons reaching the retina varies between sources (A.ii. for LED and B.ii. for halogen). Total effective photons for chicken cone opsins were calculated as described in Figure 4. For reference, the human visible spectrum is shown as a color bar beneath each panel. The relative photon availability for each cone type is shown in the top right of each panel, normalized to the long-wavelength sensitive (LWS) cone, which is assigned a value of 1. Key: Short-wavelength sensitive cone opsins (blue and violet lines), medium-wavelength sensitive cone opsin (green line), and long-wavelength sensitive cone opsin (red line). Figure adapted from Reference .