Abstract
Human visual perception begins with light entering the eyes and subsequently stimulating the receptors in the retina. The pupil has the crucial task of regulating the amount of incident light by dilating or constricting (between approx. 8 and 2 mm in diameter, with the average pupil diameter decreasing with age). Pupil size not only influences the quality of the retinal image but also modulates non-visual effects of light (e.g., melatonin suppression by light). In controlled laboratory experiments, it has been established that steady-state pupil diameter is primarily influenced by the activity of the intrinsically photosensitive retinal ganglion cells (ipRGCs) expressing the short-wavelength- sensitive photopigment melanopsin (λmax = 480 nm) under photopic conditions. However, there is a lack of investigations on pupil size regulation in dynamic, real-life lighting conditions across the lifespan. Methods: Informed by previous laboratory studies, we set out to test the following confirmatory hypotheses under everyday lighting conditions: (a) melanopic equivalent daylight illuminance (melanopic EDI) predicts steady-state pupil size, (b) melanopic EDI predicts steady-state pupil size better than photopic illuminance, and (c) pupil size decreases as a function of age. We integrated a wearable infrared video-based eye tracker (Pupil Labs GmbH) with a small-scale, calibrated research-grade spectroradiometer (Ocean Insight). Both devices were attached to a bespoke, 3D-printed, and adjustable head mount and connected to a miniature, battery-driven control computer (Raspberry Pi), enabling simultaneous sampling of pupil size and spectral irradiance in the near-corneal plane at 10-second intervals. We measured the natural variation in pupil size during a 60 min protocol during which healthy participants from a broad age range (n = 80 , age: 18–87 years, 51% female) engaged in various activities of daily living indoors and outdoors under naturally varying light conditions. Results: Our data demonstrate that under photopic real-world conditions, pupil size is strongly determined by light intensity and declines with higher age, yielding steeper age effect slopes in dimmer light conditions. Bayesian statistical analysis shows decisive evidence for the superiority of melanopic EDI compared to photopic illuminance in predicting pupil size. In line with the sluggish properties of melanopsin signalling, our exploratory analysis revealed that averaging the preceding 60 seconds of melanopic EDI values further improved pupil size prediction over using simultaneous melanopic EDI samples. Our real-world dataset yields no evidence for sex, iris colour or reported caffeine consumption significantly affecting steady-state pupil size. Conclusions: Our study sheds light on the factors influencing human pupillary physiology in real-life conditions, confirming findings from prior laboratory experiments. Taken together, the data provide strong evidence for considering age in personalized lighting solutions and extending the use of melanopsin-weighted light measures to evaluate real-world light conditions.
