The circadian system is the most biologically conserved timing mechanism in the human body, present in essentially every cell and coordinated by a central pacemaker that takes its primary input from light. Modern indoor lifestyles, with their compressed light dynamic range and electric evening illumination, sit far outside the conditions the system evolved under. The metabolic, cognitive, and mortality consequences of chronic circadian misalignment are increasingly well-characterized. This page lays out the biology, the dose-response of light, and the trial evidence on shift work and circadian disruption.
The SCN and the central pacemaker
The suprachiasmatic nucleus sits above the optic chiasm in the hypothalamus, two adjacent clusters of about 10,000 neurons each. It is the central circadian pacemaker for the human body. SCN neurons fire in coordinated 24-hour rhythms driven by a transcription-translation feedback loop involving the clock genes BMAL1, CLOCK, PER1/2/3, and CRY1/2. The intrinsic period averages 24.18 hours in healthy adults; without external time cues (in temporal-isolation studies of weeks to months), people drift toward later sleep and later wake by about 11 minutes per day.
The SCN does not run alone. Peripheral clocks exist in essentially every tissue (liver, muscle, adipose, kidney, immune cells), each with its own transcription-translation oscillator. The SCN coordinates the peripheral clocks via neural and humoral signals, but the peripheral clocks can be entrained directly by feeding cues, exercise, and temperature. This is the basis for the "feeding window" arguments in time-restricted eating: late-night feeding entrains liver and adipose clocks out of phase with the SCN, producing internal desynchrony.
The SCN takes its primary entrainment cue from light. Melanopsin-expressing intrinsically photosensitive retinal ganglion cells (ipRGCs), discovered in 2002, project directly to the SCN via the retinohypothalamic tract. ipRGCs respond best to short-wavelength blue light around 480 nm and integrate light over minutes to hours. They are anatomically separate from the rod and cone photoreceptors that drive vision; people with severe retinal damage to rods/cones can still have intact circadian light responses if their ipRGCs are preserved.
Light's phase-shifting power
Czeisler 1986 (Science) was the foundational human paper showing bright light can reset the circadian pacemaker independent of sleep timing ( Czeisler et al. 1986 ). Bright light (around 10,000 lux for several hours) administered at specific phases produced predictable phase advances or delays, mapped onto a phase-response curve (PRC). The shape of the human PRC is conserved with mammals: light in the early subjective night (around bedtime to mid-sleep) produces phase delays; light in the late subjective night and early subjective morning (around wake to mid-morning) produces phase advances; light in the middle of the subjective day has minimal phase-shifting effect.
The maximum achievable shift per day is about 2 hours under strong protocols (3+ hours of 10,000 lux at the optimal phase). For typical interventions (30 minutes of 10,000 lux on waking, or 5 minutes of 5,000 lux blue-enriched light), shifts of 30 to 60 minutes per day are typical. This is why jet lag recovery typically runs at about 1 hour per day eastward (advance) and slightly faster westward (delay): the human SCN is bounded by the achievable phase-shift rate.
Light intensity matters disproportionately. Outdoor daylight on a sunny day is 50,000 to 100,000 lux at noon; an overcast day is 5,000 to 20,000 lux. Indoor office lighting is typically 200 to 500 lux. The dynamic range humans evolved under (1,000:1 between bright morning sunlight and starlight) is much larger than the dynamic range modern indoor environments produce (5:1 between work-from-home morning desk and bedroom evening lamp). The implication is that even substantial indoor exposure produces a much weaker entrainment signal than 10 minutes of morning outdoor light.
The Wright 2013 camping experiment
Wright 2013 (Current Biology) quantified the effect of removing electric lighting on the circadian system ( Wright et al. 2013, n=8 ). Eight adults wore continuous light exposure monitors during a normal week of urban living, then went camping for one week with no electric lights, no flashlights, and no screens. Baseline ambient light averages were around 100 lux during waking hours; camping light averages were closer to 500 to 1,500 lux during daytime and near zero after sunset.
After one week of camping, dim-light melatonin onset shifted about 2 hours earlier, from around 22:00 in the city to around 20:00 camping. Melatonin onset aligned closely with sunset. Wake time also shifted earlier and aligned with sunrise. The magnitude of phase shift over a single week was striking and gives a quantitative anchor for the disruption that modern indoor lighting produces relative to a natural light-dark cycle.
The implication is not that humans should abandon electric light. The implication is that the systematic delay of circadian phase under modern conditions (10:00 PM melatonin onset, 6:30 AM wake) reflects a chronically delayed clock that would naturally run earlier given more outdoor light and less indoor evening light. This delayed phase is the proximate biology behind common complaints of difficulty falling asleep and difficulty waking at conventional times.
Melatonin: the dim-light hormone
Melatonin is the SCN's principal output hormone, synthesized by the pineal gland and released into circulation. Daytime serum melatonin is near zero (under 5 pg/mL); nighttime peak is 50 to 200 pg/mL in healthy young adults, declining with age (peak of 30 to 100 pg/mL in 50 to 70 year olds). The rise begins about 2 to 3 hours before habitual sleep onset, in dim light conditions; the term "dim-light melatonin onset" (DLMO) refers to this rise threshold and is the most reproducible circadian phase marker available in humans.
Melatonin secretion is suppressed by light, particularly in the 460 to 480 nm range. Even moderate room light (around 200 lux) acutely suppresses nighttime melatonin by 50 to 70%. Light at 5 lux (a single dim bedroom lamp) produces measurable suppression. Smartphone screens at typical viewing distance produce roughly 30 to 50 lux at the eye, with significant blue-light content.
Exogenous melatonin (commonly sold as 0.3 to 10 mg supplements) operates in two modes. At low doses (0.1 to 0.5 mg taken 4 to 6 hours before DLMO), it produces phase advances similar to evening light avoidance. At high doses (3 to 10 mg taken at bedtime), it produces sedation via partial agonism at MT1/MT2 receptors and an exaggerated nighttime peak. The phase-advancing application is mechanistically cleaner and at much lower doses than the typical retail product.
Cortisol and the awakening response
Cortisol shows the inverse phase pattern to melatonin. Levels are lowest around midnight, rise progressively through the second half of the night, peak 30 to 45 minutes after waking (the cortisol awakening response, CAR), then decline through the day to the midnight nadir. Healthy young adults typically run morning peaks of 250 to 500 nmol/L and afternoon levels of 75 to 150 nmol/L.
Adam 2017 meta-analyzed diurnal cortisol slope data and reported associations between flatter diurnal slopes (smaller morning peak, higher evening levels) and elevated mortality risk, depression, and metabolic dysfunction ( Adam et al. 2017 ). The slope is more informative than any single timepoint. Chronic stress, shift work, and certain depressive subtypes flatten the slope; healthy circadian alignment shows steep morning-to-evening descent.
The cortisol response is downstream of SCN signaling via the hypothalamic-pituitary-adrenal axis. Light exposure on waking acutely amplifies the CAR; this is part of the mechanism by which morning light tightens diurnal alignment. Conversely, night-shift workers show flattened slopes that partially recover during day-shift recovery weeks but rarely fully normalize while shift work continues.
Chronotype variation
Chronotype is the individual disposition toward earlier or later sleep timing. About 30 to 40% of adults are intermediate types ("hummingbirds"), 25 to 35% are morning types ("larks"), and 15 to 25% are evening types ("owls"). The variation is substantially genetic (twin heritability 40 to 50%) and partly determined by clock-gene polymorphisms (PER3, CLOCK, CRY1 variants). Chronotype shifts predictably across the lifespan: adolescents and young adults run later, middle-aged and older adults run earlier.
Chronotype is not absolute. The same individual can shift their habitual sleep timing by 1 to 3 hours over weeks with consistent light and behavioral entrainment. But pushing more than this often fails: extreme owls trying to maintain a 6 AM wake time consistently report worse sleep quality, lower morning cognitive performance, and poorer metabolic health than their habitual schedule allows. This is part of why "social jet lag" (the phase difference between work-day and weekend sleep timing) is associated with metabolic dysfunction in cohort studies; the phase difference is a sustained mismatch between intrinsic chronotype and imposed schedule.
The practical implication: chronotype is partially modifiable but not infinitely so. An adult identifying as a strong evening type may benefit more from finding work compatible with later wake times than from trying to force a 5 AM routine. The intervention literature on chronotype shifting is more about modest adjustments (1 to 2 hours) than about wholesale transformation.
Shift work: the strongest disruption
Shift work, defined broadly as work schedules that misalign with the circadian phase, is the cleanest natural experiment for circadian disruption. Vetter 2018 analyzed the UK Biobank cohort of about 270,000 adults and found usual or permanent night shift work associated with elevated type 2 diabetes incidence versus day workers ( Vetter et al. 2018, n=270000 ). The effect persisted after adjustment for BMI, smoking, alcohol, and physical activity, and was largely independent of genetic T2D risk score. Other UK Biobank analyses have reported similar associations with cardiovascular disease, certain cancers, and all-cause mortality.
The mechanisms cluster into three categories:
- Direct metabolic. Eating during the biological night (when insulin sensitivity is lower) produces postprandial glucose excursions 50 to 100% larger than the same meal eaten during the biological day. Chronic exposure to this pattern compounds.
- Sleep restriction. Shift workers sleep on average 1 to 2 hours less per 24 hours than day workers, with worse sleep quality. Sleep restriction independently elevates inflammatory markers, blood pressure, and insulin resistance.
- Internal desynchrony. SCN, peripheral clocks, and feeding clocks fall out of phase under rotating shifts. Hepatic gluconeogenesis, immune cell trafficking, and hormone secretion become decoupled from the central pacemaker.
The mortality cost of decades-long shift work is real; cohort estimates run roughly 5 to 15% elevated all-cause mortality after multi-decade shift exposure. Mitigation strategies (consistent light at the start of shift, blackout sleep environments, attention to feeding timing) help but do not fully eliminate the cost.
Practical synthesis
For an adult optimizing circadian alignment:
- Morning bright light. 10 to 30 minutes outdoor within 30 to 60 minutes of waking. On overcast days or in winter, supplement with 5,000 to 10,000 lux indoor light therapy. This is the highest-yield intervention.
- Evening dim light. Reduce indoor light below 100 lux for 2 to 3 hours before bed. Warm-temperature LEDs, dimmer switches, and avoidance of overhead lights matter more than blue-blocker glasses.
- Consistent wake time. Wake within a 30 minute window across weekdays and weekends. The wake-light combination is the tightest entrainment cue; weekend lie-ins drift the phase.
- Eating windows aligned with daylight. Last meal 2 to 3 hours before sleep at minimum. Late-night feeding entrains peripheral clocks out of phase with the central pacemaker.
- Melatonin at low dose. 0.3 to 0.5 mg taken 4 to 6 hours before habitual DLMO produces useful phase advances. Higher doses are unnecessary for most circadian applications.
- Shift workers: tactical light timing. Bright light during the start of the shift, dark commute home with sunglasses if returning during daylight, blackout sleep environment. Treat the schedule as the dominant constraint and minimize compounding factors.
