Tuning daytime and evening light did not improve older adults' sleep

The idea is appealing and increasingly marketed: give older people plenty of bright, blue-rich light by day and dim the blue at night, and their fragmented sleep should consolidate. A small but carefully instrumented crossover trial in the Journal of Sleep Research, published 2 September, put that idea to an objective test and did not find the sleep benefit [s1].

What the trial did

The CleverLights study was a double-blind, randomised, controlled crossover trial in older adults with self-reported poor sleep [s1]. Twelve participants (mean age 65.6 ± 8.7 years) each spent two separate 72-hour stays in a laboratory, two weeks apart, under two lighting conditions [s1]. During the day, one condition delivered blue-enriched light (melanopic equivalent daylight illuminance, or EDI, of 843.6 lux) and the other standard light (melanopic EDI 215.8 lux) [s1]. For three hours before habitual bedtime, illuminance was lowered under both, with greater blue attenuation in the intervention (melanopic EDI 49.5 lux) than the control (60.8 lux) [s1]. On the third night, sleep was measured with high-density EEG — a far more sensitive readout than a wrist tracker [s1].

The primary result was null

The pre-specified primary outcome was wake after sleep onset (WASO), the total time spent awake after first falling asleep. It did not differ between conditions: 136.1 ± 52.6 minutes under blue-enriched light versus 117.2 ± 43.9 minutes under standard light (p = 0.338) [s1]. If anything the point estimate ran the wrong way, though with twelve people the intervals are wide and the difference is not statistically meaningful [s1]. The authors' conclusion is plain: daytime enrichment and evening depletion of blue light did not improve objective sleep in this group [s1].

Two secondary signals did move. Blue-enriched light increased sigma-band activity (12–16 Hz), a marker associated with sleep spindles, during N1 (p = 0.010) and N2 (p = 0.017) sleep [s1]. And the distal-proximal skin temperature gradient — a surrogate for the peripheral heat loss that accompanies sleep onset — shifted under the intervention (−1.21 °C ± 1.44 versus −1.85 °C ± 1.39, p < 0.001) [s1]. The authors label the EEG spectral analyses exploratory, which is the correct caution: these were not the question the trial was built to answer, and with this many comparisons some will reach significance by chance [s1].

Reading it fairly

That evening light can disturb the sleep system is not in doubt — controlled work has shown ordinary room light before bed suppresses and shortens melatonin secretion [s2]. What CleverLights tests is narrower and more product-shaped: whether engineering the spectrum of ambient light across the day measurably improves the sleep of older people who already sleep poorly. On its primary, objective endpoint, it did not [s1].

The units the trial reports in — melanopic EDI — are worth a word, because they are what make this more rigorous than typical lighting claims. Melanopic equivalent daylight illuminance weights light by how strongly it stimulates the eye's melanopsin-containing cells, the ones that drive the circadian system, so a daytime figure of 843.6 lux versus 215.8 lux is a genuine difference in circadian "signal," not just perceived brightness [s1]. That the intervention nonetheless failed to cut night-time wakefulness is therefore a reasonably clean negative: the light really was different in the way that should matter, and objective sleep still did not improve [s1]. The one internally consistent secondary signal — more sigma-band activity, the frequency band of sleep spindles — is a thread worth pulling in a bigger study, since spindles are implicated in sleep stability and memory, but the authors are right to label it exploratory rather than a benefit [s1].

The obvious limitation is size. Twelve participants is a proof-of-concept, not a definitive answer, and a real effect could exist that this trial was underpowered to detect [s1]. The stay was also three days in a laboratory, which is not how anyone lives, and the participants were defined by self-reported poor sleep rather than a formal insomnia diagnosis [s1]. The authors explicitly call for further investigation rather than claiming light does nothing [s1].

What to watch

The honest takeaway is a caution against overselling spectral lighting as a fix for age-related sleep fragmentation. The spindle and thermoregulation signals are interesting leads for a larger, longer, at-home trial powered on an objective sleep endpoint — but until such a trial reports, a tuned-light regimen is a hypothesis, not a treatment.

This article is informational and does not constitute medical advice.