The Most Underrated Lever: A Steady Wake Time

You cannot decide when you fall asleep. You can lie there and issue the instruction, and the instruction goes nowhere; sleep arrives when the body is ready and not a minute earlier. Which raises a fair question. If the moment of falling asleep is out of your hands, which part of the night is actually yours to set?

The advice everyone gives

Sleep badly for a week and someone will tell you to get to bed earlier. It is kind advice, honestly meant, and it has the considerable appeal of sounding like a decision you can make on a Sunday evening. Most of us have tried it. Most of us have then lain in the dark at half past ten, wide awake, wondering why the early night has gone unrewarded.

The belief underneath the advice is that bedtime is the control surface. Fix the hour you get in, and everything else follows. The research keeps pointing at the other end of the night.

Two experiments, one comparison

The cleanest test comes from a pair of studies out of the same lab, using the same measure. That measure is dim light melatonin onset, or DLMO: the evening moment when melatonin begins to rise, and the standard marker for where a person’s internal clock is currently sitting.

In the first, fourteen healthy adults spent a fortnight on six-hour nights with their wake time held at the usual weekday hour, and a fortnight on nine-hour nights with the wake time pushed three hours later, in counterbalanced order. After the late-waking fortnight their melatonin onset had moved 2.4 hours later and melatonin offset 2.6 hours later.¹ The authors attributed the shift to the change in morning light exposure.

The mirror image ran in the same lab, with the same measure. Volunteers went to bed either at 22:00 or at 01:00, the same three-hour difference, while the lights came on at 07:00 for everyone. This time the clock moved by around 0.6 hours.²

Push your wake time three hours later and your body clock drifts nearly two and a half hours with it. Push your bedtime three hours later and it barely moves.

Where this gets less tidy

0.6 hours is a real shift, and it deserves saying plainly: bedtime does move the clock. It moves it far less. Anyone who tells you bedtime is irrelevant has travelled further than the evidence.

There is a second thing to be straight about. These were two separate protocols run years apart, rather than one head-to-head trial. Setting 2.4 hours against 0.6 hours across studies gives a rough sense of scale and carries less weight than a single experiment testing both arms at once. In the wake-time study, sleep duration also changed alongside wake time, six-hour nights against nine-hour ones, so the two are partly tangled together. Fourteen people is a small room, and the bedtime study’s ten is a smaller one.

Why mornings have the louder voice

The mechanism here is well characterised, and worth keeping separate from the evidence above. It explains why the effect ought to exist. It does not, on its own, prove that it does.

Light does not shift the body clock by some fixed amount. What it does depends on when it lands. In a careful laboratory study, twenty-one participants received a long pulse of bright light at varying points around their own internal clock, and the resulting response curve had a peak-to-trough range of roughly five hours.³ Light arriving before the low point of core body temperature, which falls in the small hours an hour or two before habitual waking, pushed the clock later. Light arriving after that point pulled it earlier. At one crossing point it did nothing at all.

Your wake time decides which side of that line your first light falls on. Rise at seven and the morning arrives in the territory that holds the clock steady or nudges it earlier. Stay in bed until ten and the same window passes in the dark. Two honest limits: the pulse used was considerably brighter than an ordinary morning, and the study tested light timing rather than wake schedules. It describes the machinery that the wake-time findings then act upon.

Outside the lab

Among people already keeping a steady schedule, the same pattern shows up in the correlations. In sixteen young adults sleeping at habitual times for a week, melatonin onset tracked wake time at r=0.77 and sleep midpoint at r=0.68, while its relationship with bedtime was weak enough to be statistically indistinguishable from noise (r=0.36).⁴ Onset fell about two hours before bed and roughly fourteen hours after waking. This is association, in a small group of young healthy people. The causal weight still sits with the experiments above.

A messier real-world study followed sixty-one undergraduates for thirty days with diaries, wrist actigraphy and light sensors. The irregular sleepers carried a clock around two and a half hours later than the regular ones, melatonin onset just after midnight against half past nine in the evening, and their light exposure was flatter across the day (a day-night swing of 102 lux against 179).⁵ Mathematical modelling attributed most of the phase difference to that light pattern rather than to anything resembling willpower. The two groups slept about the same number of hours, which is the useful part. This was timing. Regularity also tracked academic performance. It remains one observational student cohort, and what was measured was whole-schedule regularity rather than rise time in isolation.

At population scale, consistency carries a surprising amount of weight. Accelerometry from 60,977 UK Biobank participants, more than ten million hours of it, followed for a mean of 6.3 years. Set against the least regular sleepers, the four more regular quintiles showed all-cause mortality risk 20 to 48 per cent lower, with similar gradients for cancer and cardiometabolic death, adjusted for the usual demographic, lifestyle and health factors.⁶ Regularity outpredicted duration in equivalent models.

That finding needs its own asterisk. The regularity index combines bedtime and wake-time consistency into one number, so it cannot tell you that rise time in particular is what mattered. It supports the broad claim that regularity does real work; the wake-time-specifically claim rests on the two experiments earlier. It is also observational, so the possibility that illness scrambles sleep, rather than scattered sleep hastening illness, cannot be ruled out.

The thing worth sitting with

You cannot decide when you fall asleep. You can decide when light first reaches your eyes, and that is the one instruction the clock in your head reliably accepts. Which means your alarm is quietly voting on where tonight’s melatonin lands, roughly fourteen hours later.

What this might mean for your evening

Nothing dramatic, and nothing that needs a new regime. If your sleep has drifted later than you would like, the evidence suggests the more useful hour to hold steady is the one you get up, weekends included, since a long Sunday lie-in is where most of the drift is bought. Getting outside in the first hour, or at least to a window, helps the light land where it counts. Bedtime still matters. It simply appears to be the smaller dial, and it often settles by itself once the mornings stop moving.

Which makes the evening’s job smaller than we usually make it. Something to mark the end of a day rather than to engineer the sleep: reading, low light, a bath, the same quiet half-hour in the same order. If you would like a version of that from us, our interactive soundscape is free to play with, and the Mind · Sleep page sets out what we are building for the last hour of a day. Neither will change the time you get up. That part is yours.

← Back to The Lab