Naps: Friend or Foe to Your Night?

It is a little after two. The afternoon has gone flat, the sofa is right there, and somewhere at the back of your mind a voice says that if you lie down now you will be looking at the ceiling at midnight. So you make another coffee instead and push on through.

The voice is not making things up. It is repeating something with a genuine physiological basis, and then stretching it further than the measurements go.

The belief, stated fairly

The warning is simple: naps ruin your sleep. It gets handed around kindly, usually by people who have watched someone doze off at five and then rattle about the house at one in the morning. It is also the one piece of sleep advice most people can give a reason for, which is rare and worth respecting. The reason is sleep pressure.

The idea is that wakefulness accumulates something. The longer you are up, the heavier the drive to sleep becomes, and that drive is not merely a feeling. It shows up on an EEG as slow-wave activity, the deep, low-frequency waves that dominate the first hours of a night. Sleep discharges the pressure. A nap discharges some of it early.

Where the belief comes from

This has been watched directly. In a laboratory study combining polysomnography with mathematical modelling, healthy young men given an early-evening nap took longer to fall asleep that night, showed reduced non-REM EEG power in the delta and theta bands, built up less slow-wave activity, and showed a flattened decline of that activity across the successive non-REM episodes. The authors concluded that homeostatic mechanisms, operating under reduced sleep pressure, largely accounted for what they saw.¹

That is about as clean a demonstration of the folk belief as the literature offers. Two things temper it. The sample was small, entirely male and confined to a sleep laboratory, and the nap tested was an early-evening one, which is the worst case anybody could have chosen. It tells you what happens when you sleep two hours before you sleep. It says nothing about a quarter of an hour at two in the afternoon.

The ledger

The strangest finding in this literature is also the oldest. A group working in the mid-eighties measured, by computer rather than by eye, the integrated amplitude of the slowest EEG waves across a nap and the night that followed. The total came out conserved. Near enough the same quantity of deep activity, whether it arrived in one block overnight or was split between the afternoon and the night.²

It resolved something that had puzzled people: why a late nap seemed to cost a disproportionate amount of visually scored deep sleep later. It was not being destroyed. It had already been spent.

The brain, on this reading, keeps a ledger. What you take in the afternoon is quietly subtracted from the night’s account. Not lost, not added, moved. The books balance.

Honesty requires saying that the metaphor is tidier than the physiology. The same authors noted that the adjustments involved changes in the periods, durations and distributions of the delta waves that no existing model had predicted. Something is being conserved. Nobody has fully explained how.

Then the awkward study

If naps really did wreck nights, a properly instrumented trial should show it. One went looking. Thirty-two healthy adults aged 55 to 85 were studied across two laboratory sessions with polysomnography and performance testing, with and without an afternoon nap. Napping had little effect on the quality or duration of the following night’s sleep. It produced a significant increase in total sleep across the twenty-four hours, and improved cognitive and psychomotor performance both immediately afterwards and through the next day. The authors suggested that adding daytime sleep to the daily quota “seems worthy of consideration”.³

That is a direct hit on the folk belief, and it should not be waved away. It should also not be over-read: the participants were older adults, whose sleep architecture differs from a twenty-five-year-old’s, and the whole thing happened in a laboratory.

A second study pushed harder, enforcing a ninety-minute nap every afternoon for seventeen days, fourteen of them at home and three in the laboratory. In the laboratory nights, nocturnal sleep efficiency fell by 2.4 per cent and wake times shifted earlier. In the at-home diaries, total sleep across the twenty-four hours rose by around 38 minutes; the laboratory arm showed no such increase. Evening performance improved slightly, and the timing of the body clock was unchanged.⁴ There were nine participants, all elderly, so it settles nothing. But note the direction of the result. A daily ninety-minute enforced nap cost about two per cent of night-time sleep efficiency. That is a real cost. It is not a wrecked night.

Timing, rather than length

The most useful study for ordinary life is also the least controlled. Sixty-two healthy young adults wore actigraphs for eight days while their napping was tracked. Those who napped frequently, three or more times across the eight days, had significantly more fragmented night sleep than moderate nappers and than people who did not nap at all. Late naps were associated with poorer night-time sleep than early ones, though the paper sorts them by proximity rather than by the clock: a nap counts as late if it falls within seven hours of that night’s sleep. Nap duration was not associated with any significant difference in sleep outcomes.⁵

Read that last sentence twice, because it cuts against half of what everyone believes. In the one study here that watched people in their actual lives, when you napped mattered and how long you napped did not.

The caution is the obvious one. This is observational, it used wrist movement rather than EEG, and the arrow could point either way. People who sleep badly at night have every reason to nap more the next day.

How long, then

Nap length does matter enormously for something else: how you feel afterwards. In a within-subject study, twenty-four young adults on restricted overnight sleep were given naps of exactly five, ten, twenty or thirty minutes, or none. Five minutes did very little. Ten minutes produced immediate improvements in alertness, fatigue, vigour and cognitive performance that lasted around two and a half hours. Twenty minutes worked, but the benefits took about thirty-five minutes to arrive. Thirty minutes made people measurably worse before it made them better, the thick-headedness known as sleep inertia.⁶

Two honest qualifications. The participants had been deliberately sleep-deprived, so this describes what a nap does for a tired person. And the study measured daytime alertness only. It never looked at the following night, so it cannot speak to the question this essay is asking.

A meta-analysis pooling eleven studies and 381 working-age participants found cognitive performance improved after napping, most clearly for alertness, with the benefit mostly holding for about two hours; naps taken before one o’clock outperformed later ones, although the paper gives two estimates for that comparison and the confidence interval on one of them crosses zero, which makes it a direction more than a settled result.⁷ Two further flags. Despite the word “short” in the title, the average nap in the pool ran to about 55 minutes, so it is not evidence for the ten-minute power nap. And the authors themselves note the studies are almost exclusively laboratory work, with one real-workplace component, and call for research done in the world.

Nap sleep and night sleep come out of the same account. Ten minutes at two o’clock barely registers. Ninety minutes at six is a withdrawal you’ll feel at bedtime.

Mechanism is not proof

It is worth keeping these two things apart, because most nap advice quietly merges them. The mechanism is in good shape: sleep pressure is real, it is measurable on an EEG, and discharging it early demonstrably changes the shape of the night that follows. The claim built on top of that mechanism, that napping meaningfully damages ordinary sleep, rests on a handful of small laboratory studies, one observational week of wrist data, and at least one well-run trial that found the opposite. A mechanism tells you what could happen. It does not tell you how much, for whom, or whether it matters.

What this might mean for your evening

Nothing here is a prescription, and none of it replaces knowing your own body. But the shape of the evidence is fairly readable.

If you nap early and briefly, the studies give you very little to worry about. Ten to twenty minutes in the early afternoon is a small withdrawal from an account that has all day to refill, and it is the pattern most consistently associated with feeling better afterwards. If you nap late, the mechanism is squarely against you, and it is the one variable that showed up in real lives as well as in the laboratory. A long sleep at six in the evening is borrowing directly against the night. And if you are a person who naps most days and also sleeps poorly, the honest reading is that nobody knows which is driving which, and that untangling it is a better use of your attention than banning the nap outright.

What most of this leaves you with is the ordinary business of the last hour: a room that gets darker rather than brighter, an evening that slopes downward, a bit of pressure left in the tank when you finally get in. That hour is what we spend our time on. If you want to hear what we make for it, the sleep and wind-down door is the place to start.

← Back to The Lab