Nap length

Why can a twenty-minute nap leave you sharp while a forty-minute one leaves you groggier than before?

There is something odd about the complaint. More sleep is supposed to leave you better rested, and yet the person who slept forty minutes wakes worse off than the person who slept twenty, and worse off than if they had not slept at all. Sleep is not a substance that turns harmful past a dose. So the extra twenty minutes cannot be the cost in itself — something else must have changed in that interval, and the obvious candidate is not how much sleep happened but what kind of sleep you were interrupted in.

Start by noticing that sleep is not one state held for a while. Fall asleep and you descend through recognisable stages: a brief drowsy transition, then light sleep, then — some tens of minutes in — slow-wave sleep, so called because the electrical activity of the cortex settles into large, slow, synchronised waves. Then the whole thing cycles. If the stages differ, then "how long did you sleep" and "what were you doing when the alarm went" are two different questions, and the second is the one an alarm actually answers.

Now ask what those slow waves are. Very large numbers of cortical neurons oscillating together, rather than firing in the fine-grained, independent way that computation seems to need. That synchrony is the deepest departure from waking the brain performs. And here is the reasoning step that matters: a state that far from waking cannot be exited instantly. Whatever produces the synchrony has to be undone before the cortex resumes ordinary business, and the undoing takes time it does not take from lighter sleep.

That is sleep inertia — the transient period of impaired alertness, slowed reaction and poor judgement immediately after waking. It is measurable, it is not simply the feeling of wanting more sleep, and the crucial finding is that its severity tracks the stage you were woken from, not the hours you slept. Wake someone abruptly out of slow-wave sleep and you get the worst of it. The brain does not switch on; parts of it, particularly the frontal regions that handle deliberate control, come back online noticeably behind the rest, and slow electrical activity characteristic of sleep persists into nominal wakefulness for a while.

So reconstruct the two naps. Twenty minutes rarely gets you past light sleep. The nap has drawn down some of the pressure to sleep that builds across a waking day, and you exit from a state near enough to waking that the return is quick. Forty minutes typically lands you inside slow-wave sleep and the alarm cuts across it. You have paid the exit cost, and the modest extra restoration you bought does not cover it for the first stretch after waking.

Does that account make a prediction we can test? It makes a strong one. If the culprit were duration — more sleep, more grogginess — then grogginess should climb steadily with nap length. If the culprit is the exit stage, it should not climb steadily at all: extend the nap far enough that the sleeper comes back up out of slow-wave sleep before the alarm, and the grogginess should fall again. It does. A full cycle of roughly an hour and a half is generally reported as leaving people clearer than forty minutes, which no monotonic dose account can produce. That non-monotonicity is the whole case.

Two refinements, and then an honest limit.

First, the twenty and the forty are conveniences, not thresholds in nature. How quickly you reach slow-wave sleep depends on how much sleep you are owed, the time of day, your age, and simple individual variation. A badly sleep-deprived person can be in deep sleep well inside twenty minutes; a well-rested one may not reach it in forty. The mechanism is about stages, so any figure in minutes is a proxy for the stage and inherits all that variability.

Second, there is a separate cost that is easy to confuse with inertia. A long nap also discharges a good deal of the accumulated pressure to sleep, and that pressure is what carries you into sleep at night. Grogginess in the hour after the nap and difficulty falling asleep that evening have different causes and different time courses, and only the first is sleep inertia.

The honest limit is the evidence base. That inertia is worst on waking from slow-wave sleep is well established. The practical nap advice built on top of it is much softer: nap studies typically run small samples in laboratory conditions, and published figures for how long inertia lasts range widely with the task used to measure it, so I will not quote one. The mechanism is firmer than the prescription — and the prescription's fame greatly exceeds what has been shown about naps taken in ordinary life.

• Why the afternoon dip in alertness is a rhythm in its own right rather than an effect of lunch, and how it changes what a nap at that hour does.
• Whether caffeine taken immediately before a short nap genuinely blunts sleep inertia, and how well that finding has held up outside the laboratory.

Every term the collection defines is gathered in the glossary.