Is a 90-minute nap good? What happens when you wake up
Updated October 19, 2026 · Sources: Nat Neurosci 6(7):697-698 · National Heart · Sleep Med Rev 4(4):341-353 · Sleep 29(6):831-840 · DozeClock
Short answer
90 min
Good when you have an open afternoon and want a nap that finishes a sleep cycle instead of stopping in the middle. In a Harvard study, naps of 60 to 90 minutes holding both deep and REM sleep matched a night of sleep on one learning task. It is the longest nap here, and the one that most affects tonight.
General information from population studies (NIH, CDC, NSF, AASM). It is not medical advice and does not replace a clinician who knows you.
01The Harvard study: a nap that held a full set of stages
The reason to give a nap 90 minutes is the cycle, so start with how a cycle works. The National Heart, Lung, and Blood Institute says the sleep cycle starts over every 80 to 100 minutes, with usually 4 to 6 cycles a night. Each cycle runs through non-REM stages and then REM sleep. A 90-minute nap is long enough to run that sequence once, which is why the site data rates waking from it as low for grogginess.
The study behind this page is Mednick, Nakayama and Stickgold (2003) in Nature Neuroscience. It tested a texture-discrimination task, a visual skill that earlier work showed depended on sleep after training. The abstract reports that this learning can come from brief 60 to 90 minute naps that contain both slow-wave and REM sleep. The gains resembled those after an 8-hour night in magnitude, in dependence on sleep stage and in retinotopic specificity, which means the learning stayed tied to the part of the visual field that was trained.
The result also stacked. The nap-dependent improvement was additive to the improvement from the next night's sleep, so across 24 hours performance showed as much learning as normally follows twice that length of sleep. The authors summed it up as: from the view of behavioral improvement, a nap is as good as a night of sleep for learning on this perceptual task.
Tassi and Muzet (2000) explain why the end of a cycle is a kinder place to wake. Abrupt waking from slow-wave sleep produces more inertia than waking from stage 1 or 2, and REM falls in between. A 90-minute nap is meant to end after the deep stage has passed, but sleepers differ, so the alarm may still land in a deeper stage. Nothing in the abstracts says how many people wake without fog after exactly 90 minutes.
How the conditions compared
| Condition | What the authors reported |
|---|---|
| 30-minute nap | A period of impaired alertness, then gains lasting up to 155 minutes (Brooks and Lack 2006). |
| 60-minute nap | Reaches the 60 to 90 minute range in which the Mednick study found learning gains when deep and REM sleep were both present. |
| 90-minute nap | About one cycle; learning on a visual task comparable to an 8-hour night in the Mednick study. |
The learning result is for one visual task, in a study whose abstract gives no sample size, no figure for alertness and no result for the next night.
The abstracts for Mednick and Tassi do not give the odds of waking in light sleep after 90 minutes, and the cycle length varies between people and across the night.
A sleep cycle runs about 80 to 100 minutes, so 90 minutes of sleep is one cycle, not a guarantee of waking at its end. The alarm on this page sits 105 minutes after you lie down. Lie down at 1:00 pm with an 11:00 pm bedtime and the nap uses 1 hour 45 minutes of the 10 hours between them.
Bars show nap length; colour shows the grogginess rating from the site nap data (Mednick et al. 2003, Tassi and Muzet 2000, NHLBI).
02When to lie down: 12 pm to 5 pm
The Mednick nap had no start time in the abstract, so no row is marked. The table lists every half-hour start from 12:00 to 5:00 pm for a 90-minute sleep, the alarm that allows for falling asleep, and the awake time left before an 11:00 pm bedtime.
| Lie down | Alarm | Awake time before 11:00 pm bed | Against the 8-hour guide |
|---|---|---|---|
| 12:00 pm | 1:45 pm | 9 h 15 min | Inside the guide |
| 12:30 pm | 2:15 pm | 8 h 45 min | Inside the guide |
| 1:00 pm | 2:45 pm | 8 h 15 min | Inside the guide |
| 1:30 pm | 3:15 pm | 7 h 45 min | Inside the guide |
| 2:00 pm | 3:45 pm | 7 h 15 min | Inside the guide |
| 2:30 pm | 4:15 pm | 6 h 45 min | Inside the guide |
| 3:00 pm | 4:45 pm | 6 h 15 min | Inside the guide |
| 3:30 pm | 5:15 pm | 5 h 45 min | Later than 3:00 pm: expect less sleep pressure tonight |
| 4:00 pm | 5:45 pm | 5 h 15 min | Later than 3:00 pm: expect less sleep pressure tonight |
| 4:30 pm | 6:15 pm | 4 h 45 min | Later than 3:00 pm: expect less sleep pressure tonight |
| 5:00 pm | 6:45 pm | 4 h 15 min | Later than 3:00 pm: expect less sleep pressure tonight |
Alarm = lie-down time + 15 min to fall asleep (site planning figure) + 90 min of sleep · the 8-hour guide is a sleep-hygiene rule of thumb, not a measured threshold (data/naps.json)
The alarm rings 105 minutes after you lie down. A 12:00 start ends at 1:45 pm with 9 hours 15 minutes before bed. A 3:00 pm start, the last inside the 8-hour guide, ends at 4:45 pm with 6 hours 15 minutes left.
A 5:00 pm start ends at 6:45 pm and leaves 4 hours 15 minutes before an 11:00 pm bedtime. The abstracts give no figure for sleep that night, and the site rule says late and long naps can reduce sleep pressure then, so this row is the one to avoid.
03Who this length suits
It suits you if
You have a block of 105 minutes with nothing right after it and want to wake near the end of a cycle.
You are learning something visual or skill-based and want a nap after practice. That is the case the Mednick study covers, and only that.
You work nights or are making up a long sleep debt, and can lie down by 3:00 pm for an 11:00 pm bedtime.
Pick another length if
Pick 20 minutes if you only have half an hour, or if you must be sharp soon.
Pick 10 minutes if tonight's sleep is fragile. The 10-minute nap takes the least from the evening.
Do not stop at 45 or 60 minutes. The site data rates those lengths moderate to high for grogginess, and the 90-minute alarm exists to avoid waking mid-cycle.
Questions people ask about a 90-minute nap
- Is a 90 minute nap good?
- It is a good choice when you can sleep a whole cycle. The site data rates waking from it as low for grogginess, and the Mednick study found 60 to 90 minute naps with deep and REM sleep matched a night on one learning task.
- Is a 90 minute power nap better than 20 minutes?
- It is a different nap. Twenty minutes suits a quick lift. Ninety covers a full cycle, but takes 105 minutes on this site and costs more of tonight.
- What happens during a 90 minute nap?
- You pass through non-REM stages and then REM sleep, one cycle. The NHLBI puts the cycle at 80 to 100 minutes and says usually 4 to 6 happen in a night.
- Does a 90 minute nap affect sleep that night?
- The studies opened for this page do not report that. This site recommends lying down at least 8 hours before bed, as a rule of thumb, not a measured threshold.
Sources
- Sleep-dependent learning: a nap is as good as a night — Nat Neurosci 6(7):697-698, 2003 (Mednick S., Nakayama K., Stickgold R.) — PubMed 12819785
- How Sleep Works - Sleep Phases and Stages — National Heart, Lung, and Blood Institute (NIH)
- Sleep inertia — Sleep Med Rev 4(4):341-353, 2000 (Tassi P., Muzet A.) — PubMed 12531174
- A brief afternoon nap following nocturnal sleep restriction: which nap duration is most recuperative? — Sleep 29(6):831-840, 2006 (Brooks A., Lack L.) — PubMed 16796222
- DozeClock nap data: 8-hour late-nap guide and 15-minute sleep-onset planning figure — DozeClock (data/naps.json, data/cycles.json) — rules of thumb, not measured thresholds
General information from population studies (NIH, CDC, NSF, AASM). It is not medical advice and does not replace a clinician who knows you.