Can Personalized Light Timing Improve Sleep in Parkinson's?

Most light-therapy advice starts with a clock: use a bright light at 7 a.m., for example. Your circadian system does not own a clock like that. It responds to light according to its internal phase. A 2024 Phase 2 trial in Parkinson's disease found that light delivered at chronotype-personalized times improved polysomnography-measured N3, or deep sleep, in both treatment groups. The result gives a plausible explanation for why earlier Parkinson's light trials produced a muddle of results. It does not prove that timing, rather than the light spectrum or the daily routine itself, caused the improvement.

That distinction matters. A fixed early-morning light session could occur after biological wake time for one person, near-neutral for another, and at a phase that delays the clock for a late chronotype. The same lamp can therefore deliver different circadian signals to different people.

Why a light box can work differently at the same clock time

Light is a zeitgeber: an environmental cue that helps set circadian timing. Its effect depends on where it lands on the body's phase-response curve. At some phases it advances the clock, at others it delays it, and at others it has little effect. In an observational study of healthy adults, Zerbini et al. in 2021 found that dim-light melatonin onset, or DLMO, varied by more than four hours among individuals and also shifted with season and week. DLMO marks the evening rise in melatonin and is a physiologic measure of circadian phase.

That does not show that a particular person should buy a particular lamp. It does show why a universal instruction such as "use it at 7 a.m." rests on a weak biological assumption. A questionnaire such as the Morningness-Eveningness Questionnaire can estimate chronotype, while DLMO measurement is more direct but less convenient. The Parkinson's trial used chronotype-personalized timing, although the published account available here does not specify the instrument used to assign each person's timing window.

A Parkinson's trial designed around circadian biology

Feigl et al.'s 2024 Phase 2 trial, conducted at Queensland University of Technology in Australia between February 2021 and December 2022, enrolled people with Parkinson's disease who had melanopsin-containing intrinsically photosensitive retinal ganglion cell dysfunction at baseline. These retinal cells, often called ipRGCs, help transmit light information to the brain's central clock.

The randomized, double-blind, parallel-group study compared two active light conditions for four weeks, followed by a two-week washout. The Day Mel arm used a spectrum designed to match natural daylight's melanopsin excitation. The Enhanced Mel arm provided supraphysiological melanopsin stimulation. Both conditions were delivered at each participant's chronotype-aligned time, and they were matched for visible color appearance and photometric luminance. That matching matters because participants could not simply identify one condition as the "special" light from what it looked like.

The primary endpoint was N3 deep sleep measured by polysomnography. Both arms showed statistically significant improvement from baseline. The study found no statistically significant difference between the two arms, though Enhanced Mel showed a nonsignificant trend in its favor. It also reported improvements in motor function, wellbeing, and caregiver-rated sleep quality.

The article does not provide the exact sample size, N3 change, confidence intervals, or between-group effect estimates in the reported material available here. Those missing numbers prevent a judgment about clinical importance. Statistical significance versus baseline says that the groups changed during treatment; it does not establish how large, durable, or specifically treatment-caused that change was.

The timing hypothesis fits the result—but the trial did not test it alone

The two arms received the same chronotype-matched schedule and had similar N3 outcomes. That pattern is consistent with timing being an important ingredient. It is not a clean test of timing as the primary ingredient, because the study changed several things at once: both groups received biologically directed light, both followed a structured daily exposure routine, and neither group received a sham intervention.

A sham arm would have helped separate a light-specific effect from expectancy, closer study contact, or the consequences of performing the same routine every day. A third arm using chronotype-matched light without melanopsin targeting would have helped isolate spectrum. An arm receiving the same spectrum at a non-personalized time would have tested timing directly. The trial included neither comparison.

The defensible conclusion is narrower than "timing beats spectrum." In this trial, two chronotype-timed active spectra improved deep sleep similarly, so the data support timing as a plausible contributor but cannot identify it as the cause.

Parkinson's disease provides a reasonable setting to test that idea because circadian photoreception itself can be impaired. A 2024 exploratory meta-analysis by Steiner et al. reported reduced post-illumination pupil responses in Parkinson's disease. This response is used as an in-vivo indicator of melanopsin-related retinal function. The finding supports the biological rationale for targeting melanopsin-sensitive pathways; it does not prove that impaired ipRGC signaling caused an individual participant's sleep problem or that a stronger melanopsin signal will overcome the impairment.

Earlier trials leave a real prior against easy claims

Before this study, two meta-analyses covering five randomized trials of light therapy for Parkinson's disease sleep found heterogeneous, inconclusive results. As Feigl et al. describe, protocols ranged from 100 to 10,000 lux and used white, blue, orange, or red light, with arbitrary delivery times and no individualized circadian timing. The prior trials also did not use a spectrum designed to activate rods, cones, and ipRGCs in daylight-like proportions.

Those results do not prove that generic light therapy fails for every person with Parkinson's disease. They do mean that light therapy remains investigational rather than established care for Parkinson's sleep disturbance. The newer study is a more biologically reasoned attempt, not a final settlement of a previously failed question.

Sleep disruption in Parkinson's disease also has more than one pathway. In a separate randomized, double-blind, sham-controlled study, Wu et al. in 2024 assigned 67 people with Parkinson's disease to 10 consecutive days of 1-Hz repetitive transcranial magnetic stimulation over the right dorsolateral prefrontal cortex and assessed sleep with polysomnography. The trial found objective improvement in sleep architecture versus sham. Its mechanism is cortical rather than retinal or circadian, and its sham comparator makes it a useful reminder that light therapy is one experimental approach among several.

Timing can matter outside Parkinson's disease, too, without making every timed intervention interchangeable. In a 12-week randomized trial of 58 sedentary men, Shen et al. in 2025 compared morning aerobic exercise, evening aerobic exercise, and no exercise. Both exercise groups reduced body fat; the study reported different sleep and cardiometabolic effects by exercise timing. That supports the broad idea that timing can alter a physiologic intervention's outcome. It does not test light therapy or Parkinson's disease.

What the study cannot answer yet

The trial was single-center and Phase 2, so it needs larger, multicenter replication before it can guide clinical practice. It excluded participants with apnea-hypopnea index above 15, periodic limb movement arousal index above 15 per hour, severe cognitive impairment, and deep-brain stimulation. Those exclusions leave out many people with complex Parkinson's sleep complaints.

It also offers no evidence that benefits persist after treatment stops. Parkinson's disease is chronic, while this intervention lasted four weeks and washout lasted two. Nine COVID-19-related protocol deviations occurred during the study period; lockdowns can alter daylight exposure, daily schedules, and sleep. A longitudinal study by Esquivel-Mendoza et al. documented pandemic-related changes in sleep and circadian markers, which makes that disruption relevant to a circadian intervention.

Finally, the trial confirmed ipRGC dysfunction but did not stratify outcomes by its severity. Someone with more severe retinal dysfunction may respond differently than someone with milder impairment. No pharmacologic comparator was included, so the study cannot benchmark its deep-sleep change against existing Parkinson's sleep treatments.

A sensible implication for light-therapy users

For people without Parkinson's disease, this trial is not evidence that chronotype-timed light therapy improves sleep. It is a proof of concept in a specific clinical population. Still, it identifies a practical flaw in one-size-fits-all light routines: clock time is a poor substitute for circadian phase.

If a light routine has done nothing for your sleep, changing intensity or chasing a fashionable wavelength is not the only variable worth examining. First consider whether the exposure occurs at a biologically sensible time for your own schedule and chronotype. Treat that as a hypothesis to test, not a diagnosis or a promise. The evidence supports better trial design; it has not yet earned a universal prescription.