Invited comment on 'The law of reciprocity holds (more or less) for circadian-effective lighting', by M Rea, accepted for publication in LRT 2022

Delivering effective lighting in buildings to support circadian entrainment of the occupants is of growing interest and importance to lighting specifiers. Circadian-effective lighting is not simply about how much light to provide; it also depends upon how long that light is provided. Recent research suggests that the amount of light and the duration of light exposure are reciprocally related for entrainment in the morning hours (i.e. 08:00–12:00). This opens up a number of design options heretofore unaddressed by lighting specifiers INVITED COMMENT ON ABOVE PAPER In their paper Rea et al. add the exposure duration t to the circadian stimulus (CS) – describing the “instantaneous effective magnitude of the neural signals reaching the suprachiasmatic nuclei” – to define CSt, and from that a minimal dose CSd. Since CSd uses – in a first approximation – a linear combination of the magnitude of the light stimulus (expressed in CLA) and exposure duration, a lower level of circadian-effective light can be compensated by a longer exposure duration to result in the same circadian stimulus. In their conclusions, Rea et al. state that the law of reciprocity embedded in CSd creates design flexibility to find practical solutions for healthy lighting (for example by means of “bright days” and “dim nights”) by compensating lower light levels with longer exposure durations. We certainly recognize the elegancy of this finding for lighting engineers, but want to stress that CSd is not generally applicable, and cannot be extended to any light level of any duration at any time of day or night. Multiple findings in literature are not in line with such a generalized adoption of the CSd approach to specify circadian-effective lighting. For instance, in the late evening/early night, shorter bright light exposures (face illuminance ~10 klx) can more efficiently delay the circadian rhythm as compared to longer light exposures at the same illuminance: per minute of exposure a 0.2 h light exposure was reported to be at least twice as efficient as a 1 h light exposure and more than five times as efficient as a 4 h exposure 1. Moreover, for circadian phase advances (i.e. light exposures in the biological morning), a 5 h exposure to intermittent bright light (consisting of cycles with either 5 min of bright light and 20 minutes of darkness, or cycles with 46 min bright and 44 min dark) elicited an almost equal phase resetting response as a 5hr continuous bright light at the same illuminance (~10 klx) 2. In addition, a single 6.7 h exposure to bright light (8-10 klx) maximally produces a circadian phase advance and delay of 2.0 and -3.6 h, respectively, while for a 1h exposure this is 1.2 and -2.0 h, respectively 3. Here, the 6.7 h exposure result is definitely less than what one would expect from a linear combination of light level and exposure duration. These examples indicate that for circadian phase resetting responses to high magnitude morning or evening light stimuli, the reciprocity between stimulus magnitude and exposure duration fails. This justifies a further debate on whether the CSt and CSd approach (and the current minimum threshold of CSd=0.43 during morning hours as proposed by Rea et al.), provides the right framework to accurately describe “adequate circadian light exposure” under all circumstances.