How Much Deep Sleep Do You Need? What the Number on Your Tracker Means

Short answer: healthy adults typically spend somewhere around a fifth of the night in deep sleep, and that share falls with age. But no sleep authority anywhere sets a target for it. The American Academy of Sleep Medicine and the National Sleep Foundation both recommend a total duration, 7 or more hours, and say nothing about how many minutes of any stage you should be getting.

That gap matters, because the number your ring shows you is being compared against a target that does not officially exist.

What deep sleep actually is

Deep sleep is stage N3, also called slow-wave sleep, named for the large slow oscillations that dominate the EEG during it. It concentrates in the first half of the night and is the stage from which waking feels worst.

It is genuinely important. It is when growth hormone secretion peaks, when the sharpest drop in blood pressure and heart rate occurs, and it is heavily implicated in memory consolidation. None of that is in dispute.

What is in dispute is whether you should be managing it as a number.

The typical amount, and why it is a range rather than a target

The reference point everyone cites is Ohayon and colleagues' 2004 meta-analysis in Sleep, which pooled 65 studies covering 3,577 healthy people aged 5 to 102.

It found that in adults, total sleep time, sleep efficiency, the percentage of slow-wave sleep and the percentage of REM all decrease with age, while time awake after falling asleep increases. A recent review summarising that dataset puts adult slow-wave sleep at roughly 18 to 22 percent of total sleep time, declining by about two percent per decade through adulthood.

Two caveats belong with those numbers rather than after them.

First, the percentages come from a review summarising Ohayon's tables, not from a single headline figure in the paper. Second, the finding is not unanimous: at least one later analysis restricted to modern scoring rules found the age-related decline in total sleep time without a significant reduction in slow-wave sleep specifically.

So the honest version is a range with real uncertainty attached, which is a different object from the target your app implies.

No authority sets a deep sleep target

This is the part worth being precise about, because it is the whole basis for the anxiety.

The AASM and Sleep Research Society joint consensus statement says: "Adults should sleep 7 or more hours per night on a regular basis to promote optimal health." The National Sleep Foundation recommends 7 to 9 hours for adults aged 18 to 64, and 7 to 8 for those 65 and over. Neither mentions stage quantities.

There is one near exception worth knowing. A separate National Sleep Foundation panel on sleep quality reached consensus in 2017 that, in all age groups except older adults, an N3 share of five percent or less does not indicate good sleep quality. That is a lower bound below which something may be wrong, and the same panel explicitly failed to reach consensus on target percentages for any stage.

A floor is not a goal. Nobody has established that more deep sleep is better, or that a given number of minutes is what you should aim for.

How well can a wrist or a finger detect it?

This is where the number on your screen meets its limits.

Chinoy and colleagues (2021) tested seven consumer devices against polysomnography, the laboratory standard, across 34 healthy adults for three nights each. Six of the seven produced stage estimates.

Their conclusion, stated plainly in the paper, was that device sleep stage assessments were inconsistent. All six differed significantly from polysomnography on light sleep. Three of them significantly overestimated deep sleep. Three underestimated REM.

The mechanism explains the result. Polysomnography reads brain activity directly through scalp electrodes. A wearable infers stages from heart rate, heart rate variability and movement, then applies a proprietary model. It is an educated guess derived from downstream signals, and the manufacturers do not publish the models.

Which does not make wearables useless. They are reasonably good at what they were built for: total sleep duration, bedtime consistency, and trends over weeks. Those are also the things that actually correspond to the health recommendations.

Device sleep stage assessments were inconsistent.

Chinoy et al., Sleep, 2021

What the tracker is good for, and what it is not

If your deep sleep number looks low

Work through this in order, because the first item resolves most cases.

Check your total sleep first. Deep sleep is a proportion of a whole. If you slept six hours instead of eight, the absolute minutes fall even when the architecture is completely normal. Fix duration before investigating stages.

Check your age expectation. Slow-wave sleep declines through adult life. A fifty-year-old comparing their number against a twenty-five-year-old's is comparing against a different biological baseline.

Check the timing, not just the amount. Deep sleep front-loads. Going to bed two hours later compresses the window in which most of it would have occurred.

Then check the obvious suppressors. Alcohol is the big one, and it is reliably visible on consumer devices. Late caffeine is next, and it acts long after the alertness fades. Both are covered in more detail in our pieces on caffeine's half-life and on what actually resets a body clock.

What genuinely raises it: consistent sleep and wake times, adequate total duration, aerobic exercise, a cool dark room, and less alcohol. Notice that this is the same list as "sleep well generally." There is no separate deep sleep protocol, because deep sleep is not a separate system.

The failure mode nobody warns you about

In 2017, Baron and colleagues published a case series in the Journal of Clinical Sleep Medicine describing three patients whose sleep complaints were driven by their sleep trackers. They coined the term orthosomnia: an unhealthy preoccupation with achieving perfect sleep data.

It is three patients, so treat it as a described phenomenon rather than an epidemiological finding. But the mechanism is not hard to believe. Anxiety about sleep is one of the better-established causes of insomnia, and a device that grades your night every morning is a reliable way to generate it.

If checking your score makes your sleep worse, the intervention with the best evidence behind it is to stop checking.

Questions people ask

How many hours of deep sleep do you need? There is no recommended number. Adults typically spend around 18 to 22 percent of the night in it, so roughly 80 to 110 minutes on an eight-hour night, but that is a description of what happens rather than a target to hit.

Is 45 minutes of deep sleep enough? Possibly, and the number may also be wrong. Check your total sleep time first. If you are getting seven or more hours, sleeping on a consistent schedule and waking without an alarm feeling rested, the stage breakdown is not the thing to act on.

Why is my deep sleep so low all of a sudden? The most common causes are a shorter night, alcohol, a later bedtime, illness, or a firmware update changing how your device scores stages. That last one is more frequent than people expect and is invisible unless you read release notes.

Does more deep sleep mean better sleep? Not established. No authority recommends maximising it, and the only stage-related consensus that exists is a lower bound below which quality is questionable.

Can I trust my Oura, Whoop, Garmin or Fitbit deep sleep number? For trends, cautiously. For an absolute minute count, no. Devices disagree with the laboratory standard and with each other.

This article covers sleep physiology for general information and is not medical advice. Persistent insomnia, loud snoring with daytime sleepiness, or suspected sleep apnoea should be assessed by a clinician rather than a wearable.