How much deep sleep do you need? The main adult sleep recommendation sets no deep sleep target; it asks for 7 or more hours of total sleep. In one laboratory reference study, deep sleep made up a median 20.7% of sleep in adults aged 30 or under and 14.9% in adults over 60. In validation studies, several consumer trackers misjudged deep sleep by 15 to 44 minutes on average.
That matters because a disappointing deep sleep score can send you chasing a number your device may not measure well.
What is deep sleep?
Deep sleep is stage 3 of non-REM sleep, also called slow-wave sleep or N3. The NHLBI describes sleep as cycles of non-REM and REM (rapid eye movement) sleep that restart every 80 to 100 minutes, usually four to six times a night. It notes that you spend more time in deep sleep early in the night and have more REM sleep later. REM sleep, by contrast, is when dreaming usually happens, the brain is active and the muscles go limp.
In a sleep laboratory, deep sleep is defined by brain waves, not by how rested you feel. A 2025 analysis in the journal Sleep summarizes the American Academy of Sleep Medicine scoring rule: slow waves have a frequency of 0.5 to 2.0 Hz and a minimum amplitude of 75 µV, and a 30-second stretch of recording counts as N3 when at least 20% of it is slow waves.
That definition comes from polysomnography, the overnight sleep study that records brain waves, eye movements and muscle activity; a watch or ring records none of these directly.
How much deep sleep do you need each night?
No figure for how much deep sleep you need is set by the main adult guidance. The American Academy of Sleep Medicine and Sleep Research Society consensus states that adults should sleep 7 or more hours per night on a regular basis to promote optimal health. It does not set a target for any sleep stage.
What research offers instead is a description of typical sleep. In an Austrian study of 100 healthy sleepers recorded with polysomnography to AASM standards, the median share of deep sleep was 20.7% at age 30 or under and 14.9% over 60. As arithmetic only, 20% of a 7-hour night is 84 minutes.
The ranges matter more than the medians. Healthy people in that study ranged from 15.2% to 37.5% deep sleep at 30 or under, and from 2.4% to 35.6% over 60. A figure well below the median can still belong to a healthy sleeper.
How much deep sleep is normal by age?
Deep sleep declines through adulthood. NHLBI notes that slow-wave sleep peaks in early childhood, drops sharply in the teenage years and continues to decrease through adulthood, and that older people may not have any slow-wave sleep at all.
| Age group | Deep sleep share of total sleep | Source |
|---|---|---|
| 30 or under | Median 20.7% (range 15.2% to 37.5%) | Austrian lab study, 20 adults in this group |
| Over 60 | Median 14.9% (range 2.4% to 35.6%) | Same study, 24 adults in this group |
| Older age | May be none at all | NHLBI |
For men, the decline looks real rather than a scoring quirk. The 2025 Sleep analysis of 2,913 participants in the Sleep Heart Health Study found that men showed an age-related decrease in deep sleep, and that the decrease held when the researchers relabeled the recordings with automated methods.
The same paper points out that the 75 µV threshold was established using data mainly from young men. Because brain-wave amplitude varies with age and sex, the authors suggest some reported differences between men and women in deep sleep may come from the scoring rule rather than physiology.
Are sleep trackers accurate for deep sleep?
Sleep trackers are good at detecting that you are asleep and weaker at spotting wake time and sleep stages, including deep sleep. Two recent studies compared popular devices with polysomnography on the same night.
| Device | Deep sleep compared with polysomnography | Study |
|---|---|---|
| Apple Watch Series 8 | Underestimated by 43 min; by 25 min | Robbins 2024; Schyvens 2025 |
| Fitbit Sense 2 | Underestimated by 15 min | Robbins 2024 |
| Garmin Vivosmart 4 | Overestimated by 44 min | Schyvens 2025 |
| Whoop 4.0 | Overestimated by 31 min | Schyvens 2025 |
| Oura Ring Gen3 | No significant difference | Robbins 2024 |
In the Robbins study of 35 healthy adults aged 20 to 50, all three devices detected sleep against wake with sensitivity of 95% or higher. The Apple Watch underestimated deep sleep by 43 minutes and the Fitbit by 15 minutes, while the Oura ring was not different from polysomnography. The study was funded by Oura Ring Inc, and the lead author sits on Oura's medical advisory board.
In the independently funded Schyvens study of 62 adults, some referred with suspected sleep apnea, the Garmin Vivosmart 4 overestimated deep sleep by 44 minutes, the Whoop 4.0 overestimated it by 31 minutes and the Apple Watch underestimated it by 25 minutes. The Fitbit Sense and Charge 5 showed no significant deep sleep bias. Every device detected more than 90% of sleep but correctly identified only 29% to 52% of time awake. Overall agreement with the sleep study, measured by Cohen's kappa, ranged from 0.21 for the Garmin to 0.53 for the Apple Watch.
Both studies covered one night and specific models; newer devices and algorithms were not tested.
Why is a nightly deep sleep score a weak target?
A single night's deep sleep figure combines a hard-to-measure stage, a device error that can run to tens of minutes and natural variation between healthy people. Chasing it can backfire.
Sleep researchers have described orthosomnia, a preoccupation with improving or perfecting wearable sleep data. In the three cases reported, patients spent excessive time in bed trying to raise their tracked sleep, which the authors suggest may have worsened their insomnia.
More deep sleep on one night is not automatically better sleep either. A review of alcohol and the sleeping brain describes large amounts of alcohol before bed increasing slow-wave sleep in the first half of the night, then disrupting the second half with more wakefulness. Heavy drinking can raise the number you are watching while worsening the night.
The Schyvens authors concluded that the better-performing devices could be used to track prolonged and significant changes in sleep architecture. That is the sensible use: long-term trends, not single nights.
What matters more than your deep sleep number?
Total sleep is the target with clear guidance behind it: the 7 or more hours the consensus statement recommends. Consistent bed and wake times support it; the guide to a consistent sleep schedule shows how to build that around a real working week.
How you feel and function during the day is better evidence than a stage chart. If you sleep enough hours and still feel unrefreshed, read tired despite enough sleep. If you wake in the early hours and cannot get back to sleep, waking up at 3 a.m. covers the common causes.
Checks before you act on a deep sleep score
1. Look at the trend, not the night.
Compare averages across weeks. One low night tells you very little given device error and normal variation.
2. Know your device's track record.
Check whether your model has been compared with polysomnography and in which direction it tends to err. A tracker that overestimates deep sleep can make a poor night look fine.
3. Check total sleep and timing first.
If total sleep is short or your bedtime moves by hours across the week, fix that before worrying about the deep sleep share.
4. Note alcohol on the same log.
A night after drinking can show more early deep sleep and a broken second half. Without the context, the score misleads.
5. Weigh daytime function above the app.
Record sleepiness, concentration and mood; a clinician will ask about those.
When should you see a doctor about deep sleep?
See a doctor about how you feel, not about the deep sleep percentage. Book an appointment if you are persistently unrefreshed despite enough time in bed, sleepy during the day, or if someone has noticed loud snoring, gasping or pauses in your breathing; sleep apnea signs in men explains what to report.
In the US, start with your primary care physician; in the UK or Australia, your GP. Bring a few weeks of total sleep and wake times, your symptoms and the tracker model. Ask whether a formal sleep study is warranted. Only polysomnography measures deep sleep directly.
Frequently asked questions
How much deep sleep is normal?
In one laboratory study of 100 healthy adults, deep sleep made up a median 20.7% of total sleep at age 30 or under and 14.9% over 60, with wide ranges in both groups. As arithmetic only, 20% of a 7-hour night is 84 minutes. The main adult sleep recommendation sets no deep sleep target.
Does deep sleep decrease with age?
Yes. The NHLBI says slow-wave sleep continues to decrease through adulthood and that older people may not have any at all. In a 2025 analysis of 2,913 participants in the Sleep Heart Health Study, the age-related decline in men held up when recordings were rescored with automated methods.
Is my Apple Watch or Oura deep sleep number accurate?
Only roughly. In one study the Apple Watch Series 8 underestimated deep sleep by 43 minutes and in another by 25 minutes. An Oura-funded study found the Oura Ring Gen3 did not differ significantly from a sleep study, while a Garmin Vivosmart 4 overestimated deep sleep by 44 minutes in independent testing.
Is deep sleep the same as REM sleep?
No. Deep sleep is stage 3 of non-REM sleep, defined by slow brain waves, and it is concentrated early in the night. REM sleep is when dreaming usually happens, the brain is active and the muscles go limp; you have more of it later in the night. Both recur in cycles.
When does deep sleep happen during the night?
Mostly early. The NHLBI says you spend more time in deep sleep early in the night, while REM sleep increases later. Sleep runs in cycles that restart every 80 to 100 minutes, usually four to six times a night, so the early cycles hold more of it.
The short version
The main adult recommendation sets no deep sleep target; it asks for 7 or more hours of total sleep on a regular basis. Deep sleep typically makes up a smaller share of sleep with age, and healthy people vary widely. Consumer trackers can misjudge deep sleep by tens of minutes, so use them for long-term trends, not nightly verdicts. Bring symptoms, not a stage chart, to your doctor.
This article is general education, not medical advice. It does not replace your own physician.
