Caffeine and deep sleep: how coffee changes your sleep stages

Key takeaways

• A 2026 review of 32 sleep-EEG studies found caffeine’s most consistent effect was suppressing the slow brain waves of deep sleep.
• Across studies the average loss is about 11 minutes of deep sleep; a single 400 mg dose cost about 15 to 30 minutes in one trial.
• Dose matters: 100 mg taken at least 4 hours before bed showed no significant effect in young men, though EFSA says 100 mg close to bedtime may affect some adults.
• Evidence on REM sleep is less consistent: some studies found REM started later, and one found less of it after caffeine.
• Because people do not reliably sense these changes, timing your caffeine is a better guide than how rested you feel.

What deep sleep is and why caffeine reduces it

Sleep researchers score sleep from brain-wave recordings (EEG) into lighter stages, deep sleep and REM sleep. Deep sleep is scored as N3 (some studies split it into two stages, N3 and N4) and is marked by large, slow brain waves, which is why it is also called slow-wave sleep.

Adenosine is widely accepted as one of the brain’s natural sleep-regulating substances, and caffeine reduces sleepiness by blocking the receptors adenosine acts on. Put simply, and this is a simplification, caffeine turns down a signal that is part of how the brain is pushed toward sleep, and the deepest stage is where that shows most clearly.

A 2026 systematic review of 32 sleep-EEG studies found that the effect seen most reliably was a damping of the slowest, lowest-frequency brain waves of non-REM sleep, the waves that define deep sleep. The review summarized studies in words rather than pooling their numbers, so it shows a clear direction rather than a precise size. An earlier systematic review pointed the same way: deep sleep and its slow waves usually decreased with caffeine, while the lightest stage of sleep and time awake went up.

How many minutes of deep sleep caffeine takes in trials

The size of the effect depends on how much caffeine and when. The table brings together the main measured results.

The 12-hour result stands out: in that trial, a 400 mg dose at 10 a.m. for a 10 p.m. bedtime would still have cost about 21 minutes of deep sleep. The 8-hour figure being smaller than the 12-hour one should not be over-read in a 23-person trial; the authors’ advice was simply to keep 400 mg at least 12 hours from bedtime.

For scale, 400 mg is a lot at once. EFSA considers single doses up to 200 mg free of safety concerns for healthy adults, but some large coffee-shop drinks get close to or beyond 400 mg: lab tests of one drink from the same shop found 259 to 564 mg over six days. Check your usual drinks, such as brewed coffee or espresso.

Caffeine and REM sleep: what is known so far

Evidence on caffeine and REM sleep is less consistent than for deep sleep. The 2026 EEG review found REM effects varied between studies, though some reported REM sleep starting later. In the 2015 lab study, 200 and 400 mg reduced the amount of REM sleep as well as slow-wave sleep in both young and middle-aged adults.

One 10-day study gave 20 healthy men 150 mg of caffeine three times a day. Total sleep and the mix of sleep stages did not differ significantly, but REM sleep started later, and the men found it harder to wake up and felt more tired on waking. The setup was unusual (an evening nap, then recordings that began about 5 hours after their usual bedtime), so the findings are hard to apply to a normal night.

So caffeine can affect REM sleep, but the size and pattern vary from study to study. The stronger, repeated finding is less deep sleep.

Why you can’t feel caffeine’s effect on deep sleep

You are unconscious during deep sleep, so losing some of it does not announce itself. In the 2025 trial, the gap between recorded and self-rated sleep led the authors to conclude that people may not accurately perceive how caffeine affects their sleep quality.

An earlier trial found the same thing from another angle: 400 mg taken 6 hours before bed cut sleep by more than an hour on a sleep monitor, but only the monitor clearly detected the change at that timing. Feeling rested is not proof that deep sleep was untouched, and feeling tired is not proof that caffeine was the cause. More on that gap in coffee before bed doesn’t affect me.

Does morning caffeine reduce deep sleep at night?

A large enough dose can. In the 2025 trial, 400 mg taken 12 hours before bed reduced deep sleep by about 21 minutes, and in the 1995 lab study, 200 mg taken at about 7 a.m. still reduced total sleep and the very slowest brain-wave activity of non-REM sleep that night.

A modest morning coffee looks different. In the 2025 trial, 100 mg taken 12 hours before bed had no significant effect. With a half-life of about 2 to 8 hours in healthy adults (EFSA), how much of a morning dose remains by bedtime depends heavily on the size of the dose and on how fast you clear it.

Does deep sleep adapt in daily coffee drinkers?

Not in any way that has been shown. The authors of the 2025 trial wrote that tolerance to caffeine-induced sleep disturbance remains unclear in habitual consumers, and their participants, who used under 300 mg a day, still lost deep sleep after 400 mg.

The 10-day study of daily caffeine found no significant change in sleep stages, which hints that regular daytime use may matter less than a single large dose. But its unusual design means it cannot settle the question, so the honest answer is that it is unknown.

How to protect deep sleep from caffeine

You cannot measure your deep sleep at home with lab precision, so set your caffeine by dose and timing instead. Getting more deep sleep in general is a wider topic covered by goodsleep.today; here is the caffeine part:

• Keep big doses early: the 2025 trial’s authors advise avoiding 400 mg within 12 hours of bedtime.
• Give a standard coffee room: a 2023 meta-analysis estimated a 107 mg coffee should be had at least 8.8 hours before bed to avoid losing sleep time.
• Mind small late doses too: EFSA says even 100 mg may affect sleep in some adults when taken close to bedtime.
• Allow for age: a systematic review found older adults’ sleep may be more sensitive to caffeine, and a lab study found 400 mg disrupted sleep more in middle-aged than in young adults.
• Use the caffeine cutoff calculator to turn your dose and bedtime into a latest time.

Common questions