Perhaps you're reading this an hour after a coffee that cleared your head and has since stopped doing it. So what was it hiding?
Every hour you spend awake adds to a pressure for sleep1, and sleep is what drains it. Most of that draining happens in the deep sleep of the first hours of the night16. All day, your body clock pushes back, keeping you alert while the pressure climbs. It pushes hardest late in the day, just as the pressure is highest3. Around your usual bedtime it eases off, and the pressure the day has built is free to send you to sleep.
Tonight's sleepiness, in other words, is made during today. Once you can picture that store filling by day and emptying by night, the coffee that hides it for a few hours and then lets it back starts to make sense, and so does the slow Sunday night after sleeping in. If you've come here worn out and still not sleepy, that question has its own answer below, in Why can I be exhausted and still not sleepy?
What is sleep pressure, and how do we know it's there?
Sleep pressure is what researchers call the part of sleepiness that depends on how long you've been awake. Researchers can't see it directly, so they follow the trace it leaves in the large, slow brain waves of deep sleep. The longer people have been awake, the more of this slow-wave activity their sleep has, and across a night it falls away1. It peaks in the first sleep cycle, and each cycle after rises less high16. Most of the draining is done early.
But how could anyone tell it's the hours awake doing this, and not simply the time of day? The answer came from volunteers who agreed to lose track of time. For weeks they lived with no clue to the hour, on days stretched so that their sleep fell at every hour of the body clock in turn1. Whatever hour their sleep fell at, how deep it ran followed how long they'd been awake, and hardly at all the time of day1. In every sleep, whenever it came, the slow waves faded as it went on1. That's the store, filling with each waking hour and emptying in sleep.
Is it adenosine building up, like a fuel gauge?
It’s often said
Adenosine builds up in your brain all day, and that's what makes you sleepy.
Why it can feel that way
It's a tidy story, and much of it is right. In animals, adenosine levels outside brain cells rise during waking, fall during sleep and climb further when sleep is held off5. Coffee blocks adenosine and plainly keeps people awake. The link feels close to obvious.
What’s going on underneath
The fuel gauge is a good picture as far as it goes. Adenosine is the best-studied candidate for the substance behind sleep pressure, and researchers broadly accept it as one of the body's own sleep-regulating substances5. The direct measurements come from rats and cats. In people, the evidence comes mostly by way of caffeine. And the researchers who know adenosine best stop short of calling it the whole store. The measurements so far don't show it following the smooth daily rise and fall of deep-sleep activity, and no single 'homeostat' has been found56. Changes in the connections between brain cells, and signals from support cells called astrocytes, are among the other candidates56. How they all add up to the urge to sleep is still an open question56. So adenosine is probably one part of the picture, and the fuel gauge is a fair sketch of that part. The rest of the machinery is still being worked out.
Why am I so sleepy at 3pm, and wide awake at 10pm?
Two forces are at work. One is the pressure, rising steadily from the moment you wake. The other is your body clock, which pushes for wakefulness by day and lets go at night. Researchers draw the two together in what they call the two-process model26. It's a model, a way of setting two forces against each other, and no one has found them as two separate parts of the brain. Even so, it has predicted the timing and depth of sleep well across many experiments for more than thirty years26. In it, the clock pushes hardest for wakefulness late in the day, when pressure is highest, and hardest for sleep late in the night, once pressure has run down37.
One of the neatest ways to see those two forces at work was also one of the most tiring to take part in. Twenty-two volunteers, after a night without sleep, lay down for seven minutes every twenty minutes and tried to sleep, around the clock, for a day or more8. With a whole night's sleep missing, you might expect them to drop off every time they lay down. Sleep came unevenly instead. They fell asleep most often at two points in the body clock's cycle: the large one at night, and a smaller one in the middle of the afternoon8. Another laboratory found the same two peaks in fifteen people living in time isolation, the second about nine to ten hours after the first9.
The evening was stranger still. In the few hours before their usual bedtime, the volunteers found sleep unusually hard to reach, even with all that pressure behind them, because the clock's alerting signal was at its strongest810. Researchers came to call that stretch the 'forbidden zone'. For those volunteers it ran from about 8 to 10pm8, and yours comes wherever your own clock puts it, a few hours before your usual bedtime.
So if 3pm is your sleepiest hour and 10pm your clearest, both forces are doing what they do. The second wind is real. It comes from your clock, which keeps you alert through those last hours and then lets the day's pressure through. The clock has its own page, What is the body clock, and how does it decide when I'm sleepy?. If bedtime often finds you alert, there's Why am I wide awake at bedtime?
Why does coffee work, and why does the sleepiness come back?
When you drink a coffee, the caffeine blocks the receptors adenosine works through, so its signal can't get through5. It doesn't take adenosine away, and it can't undo the hours you've been awake. Going by how caffeine works, then, the pressure you've built goes on building out of sight, and it's still there when the caffeine fades. The same cup can also work differently from one day to the next. How much it lifts you depends on how much pressure you've built and how long it's been since your last coffee5, which suggests a short night changes what that cup can do for you. A coffee in the afternoon works on the sleepiness you have then. One taken late reaches into tonight as well: too close to bed, caffeine typically delays sleep, shortens it and makes it shallower, with less deep sleep511. And your cup may not do to you what it does to the person across the table, because people differ a great deal in how caffeine affects them, partly because of their genes511.
So why does the sleepiness come back around four in the afternoon? Nobody has yet measured a same-day 'crash' as such5. What we do know is how caffeine works, and that points to the simplest explanation: the pressure was waiting behind the coffee all along. If you drink coffee every day, something else may add to it. Going without caffeine leaves regular drinkers drowsy, tired and less alert, usually from twelve to twenty-four hours after their last dose12. Researchers who study adenosine suggest that some of the sleepiness regular drinkers feel during the day may be a mild form of that withdrawal, an idea they haven't yet tested513. How long caffeine lingers is the subject of How long does caffeine stay in your system?
Why does a doze on the sofa make bedtime harder?
A nap in the evening spends pressure you'd otherwise have taken to bed with you. In a lab study of healthy young men, sleep after an early-evening nap took longer to arrive and had less deep-sleep activity4. There wasn't enough time between the nap and bedtime for the pressure to build back up14. That study's men took a proper nap in a lab. A twenty-minute doze in front of the television hasn't itself been measured, so its exact cost is still an open question. The picture suggests it spends at least some of what you'd built for the night.
Earlier in the day, and after too little sleep, a short nap can do real good. After nights cut to about five hours, young good sleepers who took a ten-minute nap at 3pm were less sleepy, more alert and performed better straight away, with some of the benefit lasting more than two hours14. If naps are part of your days, Are naps good for you? goes into when to take them and for how long.
Why is Sunday night so hard after sleeping in?
Getting up at ten instead of seven leaves you three fewer hours awake by your usual bedtime, and so less pressure behind you when you lie down. No study of sleeping in has measured that, but it follows directly from how the pressure builds.
Sleeping in seems to move the clock later too, and researchers have measured that part in teenagers. When 15- to 17-year-olds went to bed an hour and a half later and woke three hours later at the weekend, the evening rise of their melatonin had moved about forty-five minutes later by Sunday15. Adults' clocks haven't been tested this way. If they shift too, the clock would keep sleep away until later on Sunday evening.
If your morning in bed follows a week of early starts, a night up with a child or a run of late shifts, you'll know why you took it. It may help on Sunday night to know why sleep is slow to come, and that dread of Monday isn't the only thing at work.
Here the same day is lived three ways, up at 7am and in bed at 11pm, hours that are only an example. The body clock is left out of the drawing, though sleeping in moves it too. The solid line is the pressure for sleep, building while you're awake and draining while you sleep1, drawn at the usual rates of the two-process model2. An early-evening nap and sleeping in each end the day short of the dotted line. The nap's effect was seen in healthy young men in a lab4; sleeping in is reasoning from the model.
Drawn as an illustration of the pattern
See this as a table
| Day | Up | Nap | Pressure at bedtime |
|---|---|---|---|
| An ordinary day | 7am | None | As an ordinary day leaves it |
| A doze in the early evening | 7am | 6pm to 7pm | Lower than an ordinary day |
| Sleeping in until ten | 10am | None | Lower than an ordinary day |
| Part | What it shows |
|---|---|
| Solid line | Sleep pressure |
| Shaded stretch | Asleep |
| Dotted line | Where an ordinary day leaves it |
After a terrible night, why don't I sleep twice as long?
The body makes up for a bad night mainly by sleeping deeper, above all early in the night16, and only partly by sleeping longer. That's why you can sleep soundly the next night and still wake not much later than usual. In one trial, volunteers spent five nights in a sleep lab with four hours in bed each night. Then came a single recovery night, and some of them were given ten hours in bed16. Those given ten hours slept about nine, longer and deeper than usual, and yet their alertness still hadn't fully returned16.
Over a longer run of short nights, how you feel and how well you manage can drift apart. In a two-week trial, people kept to four or six hours in bed did steadily worse on attention tests, while their own sleepiness ratings rose at first and then barely moved17. After the first few days, people stopped feeling much sleepier, even as their attention kept slipping. The trial measured only those two things, attention and how sleepy people said they felt. In the two-process picture, the pressure goes on building underneath after the feeling levels off. What brings alertness back is the fuller nights that follow, perhaps more than one16. For the day itself, there's How do I get through the day after a bad night?
Why can I be exhausted and still not sleepy?
Part of the answer is in the word 'tired', which covers two different things. Sleepiness means being likely to drop off. Fatigue means feeling drained. The two are related but different, though everyday speech calls both of them 'tired'18. A day can leave you with plenty of fatigue and too little sleepiness. That's how you can lie there exhausted and wakeful at once. If that's a familiar night, Why am I wide awake at bedtime? looks at what may be going on. If it has gone on for a while, Do I have insomnia? is the place to turn.
A closer look How do researchers measure something no one can see?
Mostly they take sleep away and watch what comes back. Researchers keep volunteers awake and record the brain's slow waves in the sleep that follows1. To pull the pressure apart from the clock, they use a method called forced desynchrony, putting people on twenty- or twenty-eight-hour 'days' so that sleep falls at every hour of the body clock1. In one such study, eight healthy volunteers lived this way for thirty-three to thirty-six days1. In another, seven healthy women napped on separate days at two-hour steps from morning to late evening, and the slow-wave activity in the first stretch of each nap rose with the hour1, tracing the store as it filled.
Slow-wave activity is a marker of the pressure, the trace it leaves, and no single structure that stores the pressure has been found6. Most of the studies above were of small groups of young, healthy sleepers in laboratories, several of them men only3413, so they describe that kind of sleep best. The two trials of short nights were larger. The recovery-night trial followed 159 healthy adults aged 22 to 45 through twelve days in a lab16, and the two-week trial followed forty-eight healthy adults aged 21 to 3817. The balance shifts as we age. Compared with young adults, older adults in laboratory studies had a weaker evening alerting signal, with more sleep and sleepiness in the late afternoon and evening, less of the deepest sleep and a weaker build-up of pressure710.
The sleep of a labouring man is sweet, whether he eat little or much: but the abundance of the rich will not suffer him to sleep.
Every hour you spend awake today goes into tonight's sleep. Sleep spends it most freely in the first hours of the night1, and by morning the store is low again.

