What happens to your brain after a night without sleep?

The wakefulness clock

Drag from 0 to 72 hours. The bars show two real forces — sleep pressure and the circadian signal — while the brain below gradually loses stability. Hours are not a diagnosis: they shift risk and average performance; they don't produce identical symptoms in everyone.

A full day awake You are inside the strongest modern evidence: sustained attention is already measurably impaired.

• 0 h — You woke upSleep pressure is at its lowest. An ordinary day lies ahead.
• 16 h — End of a normal dayFor many adults the day would have ended here. Research territory begins.
• 17 h — The alcohol line — 0.5‰In a classic experiment, performance had reached the level seen at 0.05% BAC. Details below.
• 24 h — A full day awakeYou are inside the strongest modern evidence: sustained attention is already measurably impaired.
• 36 h — Into the second nightFrom here the evidence comes mostly from old experiments, hard to repeat today. Language gets more cautious.
• 48 h — Two daysPerceptual disturbances grow more common in long studies — but there is no universal hour when they appear.
• 72 h — Three daysSome historical experiments describe disorganised thinking. Individual variability is enormous.

Two forces pull at you

Sleepiness doesn't rise smoothly. It is the meeting point of two processes.

The diagram is conceptual: it shows the direction of the forces, not your values.

17 hours awake and the famous alcohol comparison

This result is quoted everywhere — and usually wrong. Here is the correct version.

Important

You have no alcohol in your blood. This is an analogy between performances on particular laboratory tasks, not proof that sleep loss and alcohol affect the brain identically.

There are no published anchor points allowing extrapolation to 36, 48 or 72 hours. Any curve continuing past 24 h is invented.

How stable is your attention?

Inspired by the psychomotor vigilance tests used in sleep research. This version is much shorter and is not a medical test.

3 practice trials, then 20 timed trials. Wait for the signal, then press as fast as you can — click, tap or the Space key.

A single night already counts

The strongest modern evidence comes from something gentler than a full all-nighter: a single night of only 2–6 hours of sleep.

• subjective sleepiness
  SMD 0.986
• sustained-attention reaction times
  SMD 0.512
• attention lapses
  SMD 0.489

Sustained attention looks like one of the first systems to give way. The brain doesn't simply get “dumber” at everything.

In the same meta-analysis, other domains — choice reaction time, working memory, inhibition — showed no equally clear deterioration.

Why do you get sleepy?

During wakefulness, neuronal metabolism and signalling are associated with rising extracellular adenosine in relevant brain regions. Adenosine participates in the homeostatic regulation of sleep.

ATP → AMP → adenosine

As hours awake grow, so does the adenosine-linked sleepiness signal.

+ Coffee

Caffeine doesn't erase the hours you've been awake. It blocks adenosine receptors, reducing part of the sleepiness signal.

Does the brain “wash itself” in sleep? The story is more interesting.

You may have heard that during sleep, cerebrospinal fluid “washes toxins out of the brain”. There is real science behind the idea. But the simple version is not settled science.

1. 2013 — the famous mouse experiment MICE

   In natural sleep or anaesthesia, interstitial space grew by about 60%, and exchange between cerebrospinal fluid and tissue — along with β-amyloid clearance — ran faster. Xie et al., Science.

   The evidence

   Xie et al., Science 2013 — mice, microscopy. Limitation: mice, not humans.

2. 2019 — human sleeping brains HUMANS

   In NREM sleep, large cerebrospinal-fluid oscillations appear, coupled to slow electrical activity and blood volume. Fultz et al., Science.

   The evidence

   Fultz et al., Science 2019 — humans, EEG (n=13) + fast MRI. Limitation: fluid waves don't prove “toxin” washout.

3. 2021 — one night awake HUMANS · SMALL

   After one night without sleep, tracer clearance from the brain was impaired — and the difference persisted after the recovery night. Eide et al., Brain.

   The evidence

   Eide et al., Brain 2021 — n=7 deprived vs 17, intrathecal gadobutrol. Limitation: neurology patients, small sample, invasive method.

4. 2024 — a direct challenge MICE · CONFLICTING RESULT

   Directly measuring fluorescent molecules in tissue, clearance was lower during sleep and anaesthesia than during wakefulness. Miao et al., Nature Neuroscience.

   The evidence

   Miao et al., Nature Neuroscience 2024 — male mice, fluorescent tracers. Limitation: contradicts the 2013 reading; active controversy.

5. 2024 — another mechanism MICE

   Synchronised neuronal activity can generate ionic waves linked to fluid perfusion and molecular transport. Jiang-Xie et al., Nature.

   The evidence

   Jiang-Xie et al., Nature 2024 — mice. Limitation: proposed mechanism; translation to humans open.

Sleep dramatically changes fluid dynamics in the brain. Exactly how those changes contribute to clearing different molecules — and how well mouse mechanisms translate to humans — remains an active research question.

One all-nighter and β-amyloid

A PET experiment tested 20 healthy adults after normal sleep and after one night of sleep deprivation. After the sleepless night, β-amyloid tracer signal was higher in the right hippocampus and the thalamus.

Preliminary mechanistic evidence, not “one all-nighter gives you Alzheimer's”. n=20.

24 → 48 → 72 hours

Here the language has to get more careful. Modern ethics make extreme total-sleep-deprivation experiments nearly impossible — so after 24 h the character of the evidence changes.

0–24 h: strong modern experimental evidence

24–72 h: mostly older prolonged-deprivation experiments

• Between 24 and 48 h, historical studies describe rising irritability, anxiety, temporal disorientation and visual distortions.
• Between about 48 and 90 h, complex hallucinations and disorganised thinking were reported increasingly often.
• After ~72 h, some historical experiments describe delusional thinking.

Perceptual disturbances grew more common in studies beyond two days — but individual variability is enormous, and much of this evidence comes from old experiments that would be hard to repeat today.

The problem isn't only how slowly you react

Sometimes the reaction is missing entirely. Watch the dot for about 15 seconds.

Microsleeps are especially dangerous when driving, cycling in traffic, operating machinery and in any safety-critical work. People can't reliably detect their own microsleeps.

Don't drive after a sleepless night.

Myths going around — and what the evidence says

• “After 24 h awake you equal 0.5‰.”

  The classic result was 17 h → 0.5‰ and 24 h → 1.0‰ — and only for certain laboratory performance measures.

• “The brain clears toxins only in deep sleep.”

  Sleep profoundly alters fluid dynamics, but how much that changes molecular clearance is still actively debated.

• “Coffee removes tiredness.”

  Caffeine can mask part of the adenosine signal. It doesn't substitute for sleep.

• “After an all-nighter the brain runs at 50%.”

  There is no single percentage for “brain function”. Cognitive systems respond differently, and performance becomes above all unstable.

• “I'll sleep 16 hours tomorrow and repay exactly the 8 lost.”

  Recovery isn't hour-for-hour accounting. Physiological and cognitive effects return at different rates.

• “At 48 h you're guaranteed to hallucinate.”

  Risk rises with prolonged deprivation, but there is no universal hour when a symptom appears.

How long can a human go without sleep?

There is no scientifically useful “safe maximum”. Guinness stopped monitoring the record in 1997 — because of the danger.

• Randy Gardner stayed awake about 264 hours (11 days) in 1964, under observation.
• The Guinness-listed record: Robert McDonald, 453 h 40 min in 1986 — before the category was discontinued.
• Extreme sleep deprivation is dangerous.
• Brief microsleeps make continuous wakefulness nearly impossible to verify without physiological recording.

What fails first?

Pick a system. Each says what researchers measure, what tends to change, and how strong the evidence is.

Questions

Sources

1. Dawson D, Reid K. Fatigue, alcohol and performance impairment. Nature. 1997;388:235.
2. Williamson AM, Feyer A-M. Occup Environ Med. 2000;57:649–55.
3. Lamond N, Dawson D. J Sleep Res. 1999;8:255–62.
4. Wuest et al. Sleep Med Rev. 2024;76:101940.
5. Hudson AN, Van Dongen HPA, Honn KA. Neuropsychopharmacol. 2020;45:21–30.
6. Doran SM, Van Dongen HPA, Dinges DF. Arch Ital Biol. 2001;139:253–67.
7. Reichert CF, Deboer T, Landolt H-P. J Sleep Res. 2022;31(4):e13597.
8. Elmenhorst D et al. J Neurosci. 2007;27:2410–5.
9. Xie L et al. Science. 2013;342:373–7.
10. Fultz NE et al. Science. 2019;366:628–31.
11. Eide PK et al. Brain. 2021;144:863–74.
12. Miao et al. Nat Neurosci. 2024.
13. Jiang-Xie et al. Nature. 2024.
14. Shokri-Kojori E et al. PNAS. 2018.
15. Waters et al. Front Psychiatry. 2018.
16. NIOSH/CDC — training on fatigue, alcohol and performance.
17. Guinness World Records — longest time awake (1997/2023).
18. Cleveland Clinic — microsleep.

Sleep is not the time when the brain does nothing.

One night without it shows almost the opposite: the brain keeps working, but finds it ever harder to work steadily. Attention slips. Sleep pressure rises. The circadian rhythm can temporarily hide the problem. And some of the most interesting processes — from adenosine to cerebrospinal-fluid movement — remain active research subjects.

The dangerous part of fatigue isn't always feeling incapable. It's that you can keep feeling capable exactly when your performance has become less safe.

Retake the test after sleeping