Dawn light and the body clock

Why does the same bright light advance the body clock at dawn and delay it at dusk?

Light is the signal that keeps the body clock on the day, and everyone knows that morning light "helps". But notice how strange the arrangement is once you look at it squarely: the very same stimulus — bright light on the retina, identical in wavelength and intensity — pulls the clock earlier when it lands near dawn and pushes it later when it lands near dusk. A signal that reverses its own sign is not what we expect from a sense. So what kind of thing must the clock be for the answer to depend on when the message arrived rather than what it said?

Begin with what the clock is. Not a timer counting down, but an oscillator that runs on its own — a loop of gene products in cells of the suprachiasmatic nucleus, where certain proteins accumulate, shut off their own production, degrade, and let production restart. Around and around, with a natural period close to but not exactly twenty-four hours.

Ask what you can even do to a thing like that. You cannot set it, because there is no pointer to move; there is only a chemical cycle in progress. What you can do is nudge the cycle — add a push to whichever step is currently underway. And here is the whole answer in embryo: a push applied to a running cycle does not have a fixed effect. It has a sign, and the sign depends on where in the loop the push lands.

Make that concrete. Light reaching specialised retinal cells — distinct from the ones that build the image you see — signals directly to the pacemaker and induces one of the loop's key proteins. Consider two moments. In the early biological night that protein is climbing toward its peak; forcing more of it holds the loop where it is, prolonging the current phase, and the clock is delayed. In the late biological night it has fallen and the next rise is imminent; forcing it early starts the next cycle sooner, and the clock is advanced. Same push, opposite sign, decided by the state of the loop it met.

Does that give the observed pattern? It does, once you locate dawn and dusk on the internal cycle rather than on the wall. Dusk light arrives in your early biological night: delay. Dawn light arrives after the daily low point of core body temperature, late in your biological night: advance. Between them, through most of your biological day, light barely moves the clock at all — an insensitive stretch sometimes called the dead zone, which is why midday sun does far less for your timing than its brightness suggests.

Now one more question, because there is a reason evolution would want this. If the loop's own period were exactly twenty-four hours, no correction would be needed. It is not; in most people it runs slightly long. Left alone you would drift later every day. What holds you fixed is that a slightly-long clock, waking into morning light, receives a small advance each day that cancels the drift. The sign-flip is not a quirk — it is the mechanism by which an imperfect oscillator is held to an exact day.

Three refinements matter here.

First, this is the neighbour of Jet lag, and the two should not be run together. Jet lag is about two clocks disagreeing and the cost of re-aligning them; the question here is narrower and prior to it — why a single light pulse has a direction at all. That description, direction and magnitude as a function of internal phase, is the phase response curve, and it governs the shift-worker and the ordinary sleeper as much as the traveller.

Second, "dawn" and "dusk" are shorthand for internal phase, not clock time. In a person whose clock has already shifted — a late adolescent, a night worker, someone a few days into a new time zone — seven in the morning may fall in their early biological night, and the same light will then delay them. This is exactly the falsification test for the account: if the direction of a shift tracked external clock time, or tracked intensity, the phase-based story would be wrong. What is actually observed is that it tracks the individual's internal phase, which is why the practical advice must always be phrased relative to your own temperature minimum rather than to a time of day.

Third, and honestly: the shape of the human curve is well established, but its numbers are not something to quote confidently. Human phase response curves come from expensive laboratory protocols with small numbers of participants; the amplitude of the shift depends on intensity, duration, wavelength and on how much light you had in the preceding days, and the crossover point between delay and advance varies between people. The molecular story above is the standard mechanistic account rather than a settled matter in every detail — the mapping from protein induction to shift direction is better supported in some model organisms than in humans. And the widely-repeated "get bright light in the morning" advice is sound in direction while being much vaguer in dose than its confident phrasing implies.

• Why the retinal cells that report brightness to the clock are most sensitive to blue-ish light, and what that does and does not imply about screens at night.
• How timed melatonin shifts the clock with a curve roughly opposite in sign to the light curve.

Every term the collection defines is gathered in the glossary.