REM vs deep sleep statistics: the key numbers
REM vs deep sleep has six traceable figures in the published literature. An older NHLBI guide describes each stage as roughly one-fifth of total sleep. A 2025 lab study found that a popular wearable correctly identified about half of deep-sleep epochs. These are descriptions and device results, not personal targets.
- About one-fifth of total sleep is described as deep sleep in a historical NHLBI guide. That is a broad educational estimate, not a required amount. (National Heart, Lung, and Blood Institute, Your Guide to Healthy Sleep, 2011)
- 50.66% of deep-sleep epochs were correctly identified by the Apple Watch Series 8 in a 2025 lab study. That means about half were misclassified. (Schyvens et al., SLEEP Advances, 2025)
- 7 or more hours of total sleep per night is the consensus recommendation for healthy adults. This is a duration target, not a stage-specific one. (Watson et al., Journal of Clinical Sleep Medicine, 2015)
7 figures from 6 sources, data years 2004–2025, mostly US and global. The full explanation of what these numbers mean is in our stage-amounts section of the REM vs deep sleep article.
These figures describe groups of people and device-performance results, not you personally. They are for information, not medical advice.
What do typical REM and deep sleep amounts describe?
An NHLBI guide describes deep sleep and REM sleep as each making up about one-fifth of a night. The same guide notes that healthy adults need 7 or more hours of total sleep. These are broad historical descriptions, not personal quotas. A 2004 pooled study found that stage proportions shift with age. No single percentage fits everyone.
About one-fifth of total sleep is deep sleep (N3), according to a broad historical educational description for adults
About one-fifth of total sleep is REM sleep, according to the same NHLBI guide describing later childhood and adulthood
7 or more hours per night is the consensus sleep duration recommendation for healthy adults aged 18 to 60
What this means for you: The one-fifth figures are a historical starting point for understanding sleep. They are not a nightly target to match. Total sleep duration has a consensus recommendation: 7 or more hours. Stage percentages do not. To see what these amounts mean for your sleep, read the stage-amounts section of the REM vs deep sleep article.
How do REM and deep sleep proportions change with age?
A 2004 pooled study of healthy individuals found that three measures all decrease as people get older. Those measures are slow-wave sleep (deep sleep), REM sleep, and the time before REM first appears. This means the one-fifth descriptions are not fixed at any age. The finding is a group-level pattern, not a personal benchmark or a target to reverse.
Significant age-related decrease in slow-wave sleep, REM sleep, and REM latency — all three fell as age increased in a pooled study of healthy individuals
What this means for you: If you are older, you are likely to see lower deep-sleep and REM figures than the historical one-fifth descriptions suggest. That is a normal group-level pattern. A single wearable percentage cannot measure this shift in your own case. If sleepiness is affecting your daily life, that is what warrants attention. The REM vs deep sleep answers page addresses the most common “is my number OK?” questions.
How accurately do wearables classify REM and deep sleep?
A 2025 study tested six wrist-worn wearables against clinical polysomnography in a lab. Results varied by device and by stage. No device was equivalent to polysomnography. The figures below show how often each device matched the clinical standard. They are not overall accuracy scores. Do not use them as a ranking to decide which device to buy.
50.66% of deep-sleep (N3) epochs were correctly identified by the Apple Watch Series 8 in a 2025 lab study, compared with polysomnography
68.57% of REM epochs were correctly identified by the Apple Watch Series 8 in the same 2025 lab study, compared with polysomnography
69.63% of deep-sleep (N3) epochs were correctly identified by the Whoop 4.0 in the 2025 study — the best N3 result among the six devices tested, compared with polysomnography
What this means for you: Wrist wearables infer sleep stages from movement and optical pulse signals. A clinical sleep study (polysomnography) records brain waves, eye movements, and muscle tone. These are different kinds of measurement. A 2024 review noted that wearables can label quiet wakefulness as sleep. A 2026 synthesis found limitations in capturing sleep stages accurately across consumer devices. None of this means your wearable is useless. It means the stage label it shows is an estimate, not a clinical finding.
To understand what your tracker number can and cannot tell you, the tracker-limits section of the REM vs deep sleep article explains the full picture. If sleepiness is affecting your daily life, a care route starts with the REM vs deep sleep compass.
| Device | Stage measured | Correct classification | Usable wearer-nights | Comparator · timepoint |
|---|---|---|---|---|
| Apple Watch Series 8 | N3 (deep sleep) | 50.66% | 20 | Polysomnography · single lab night |
| Apple Watch Series 8 | REM | 68.57% | 20 | Polysomnography · single lab night |
| Whoop 4.0 | N3 (deep sleep) | 69.63% | 40 | Polysomnography · single lab night |
REM vs deep sleep statistics table
| What it measures | Figure | Who / based on | Compared with · when | Source |
|---|---|---|---|---|
| Deep sleep share of a night’s sleep (historical educational description) | About one-fifth of total sleep | Broad adult educational description · sample not reported | No comparator · a typical night · data year not independently reported; guide revised 2011 | National Heart, Lung, and Blood Institute, Your Guide to Healthy Sleep, 2011 |
| REM share of a night’s sleep (historical educational description) | About one-fifth of total sleep | Later childhood and adulthood · sample not reported | No comparator · a typical night · data year not independently reported; guide revised 2011 | National Heart, Lung, and Blood Institute, Your Guide to Healthy Sleep, 2011 |
| Recommended total sleep duration for healthy adults | 7 or more hours per night | Healthy adults aged 18–60 · consensus recommendation | No comparator · regular nightly sleep · literature assessed through 2015 | Watson et al., Journal of Clinical Sleep Medicine, 2015 (doi:10.5664/jcsm.4758) |
| Age-related change in deep sleep, REM sleep, and REM latency | Significant decrease in all three measures with increasing age | Healthy individuals, childhood to old age · pooled meta-analysis cohorts · sample not reported in abstract | Younger versus older ages · published 2004 | Ohayon et al., Sleep, 2004 (doi:10.1093/sleep/27.7.1255) |
| Apple Watch Series 8: correct identification of deep-sleep (N3) epochs vs. polysomnography | 50.66% (epoch-classification sensitivity) | Adult sleep-laboratory participants · 20 usable Apple Watch Series 8 wearer-nights · exact N3 epoch denominator not reported | Polysomnography N3 labels · single laboratory night · published 2025 | Schyvens et al., SLEEP Advances, 2025 (doi:10.1093/sleepadvances/zpaf021) |
| Apple Watch Series 8: correct identification of REM epochs vs. polysomnography | 68.57% (epoch-classification sensitivity) | Adult sleep-laboratory participants · 20 usable Apple Watch Series 8 wearer-nights · exact REM epoch denominator not reported | Polysomnography REM labels · single laboratory night · published 2025 | Schyvens et al., SLEEP Advances, 2025 (doi:10.1093/sleepadvances/zpaf021) |
| Whoop 4.0: correct identification of deep-sleep (N3) epochs vs. polysomnography (best N3 result among six devices tested) | 69.63% (epoch-classification sensitivity) | Adult sleep-laboratory participants · 40 usable Whoop 4.0 wearer-nights · exact N3 epoch denominator not reported | Polysomnography N3 labels · single laboratory night · published 2025 | Schyvens et al., SLEEP Advances, 2025 (doi:10.1093/sleepadvances/zpaf021) |
Why the numbers differ
The one-fifth figure for deep sleep and the 50.66% figure for Apple Watch N3 sensitivity look like they describe the same subject. They do not. The one-fifth figure comes from an NHLBI guide revised in 2011. It is a broad historical description of how sleep time is distributed. It does not come from a clinical accuracy study. The Apple Watch figure comes from a 2025 lab study that compared stage labels with polysomnography, the clinical standard. One figure describes how much deep sleep adults tend to get. The other shows how often a wearable gets the label right. They answer different questions.
The two one-fifth figures (deep sleep and REM) appear the same but describe different stages in the same source. The NHLBI guide uses each figure as a rough illustration. The 2004 Ohayon pooled study found that stage proportions change across the lifespan. So the one-fifth descriptions do not hold equally at every age. The Ohayon study used historical PSG scoring methods. Its pooled sample sizes by outcome were not reported in the abstract. Any comparison between the 2011 educational estimate and a personal wearable reading must account for these gaps.
Among the wearable figures, the Apple Watch Series 8 reached 50.66% for N3 and 68.57% for REM. The Whoop 4.0 reached 69.63% for N3 in the same study. These differences reflect how each device processes movement and optical pulse signals. They are not a device ranking. A 2024 review and a 2026 synthesis each noted that performance varies across devices and populations. No single figure represents all wearables.
What the numbers can’t tell you
The one-fifth figures do not come with a sample size or an age breakdown. They describe a central tendency for a broad adult population at an unspecified point in time. They cannot tell you how much deep sleep or REM you personally need. They also cannot tell you whether your tracker reading is low for your age. And they cannot tell you whether a deficit explains how you feel.
The wearable accuracy figures come from one study of six devices. Each device had a small number of usable wearer-nights in that lab setting. These results cannot tell you by how many minutes your own device’s stage estimate might be wrong on any given night. A review-level finding and a percentage from a single device study are not the same thing. Neither is a personal prediction interval.
No reliable figure exists in this page’s sources for the share of adults worried about their stage readings. There is also no reliable figure for average wearable use among people with sleep complaints. Nor is there one for whether acting on stage estimates changes sleep outcomes. Those are genuine gaps. When no reliable figure exists for a question, “no reliable figure” is the honest answer.
How we compiled this page: we gathered figures from published studies and official sources below. We recorded each figure’s outcome, population, comparison, timepoint, and year exactly as published. We ran no surveys of our own. The original figures belong to their publishers. Our compilation and wording may be reused with credit under CC BY 4.0.
See how REM and deep sleep connect to the bigger picture
Numbers are a starting point. To understand what they mean for your sleep, the REM vs deep sleep mind map shows how the stages, amounts, measurement limits, and next steps link together. More context lives in the Sleep Science pillar.
See how it all connects →Sources
- Watson et al. — Recommended Amount of Sleep for a Healthy Adult: A Joint Consensus Statement of the American Academy of Sleep Medicine and Sleep Research Society (2015) — Tier 1 consensus statement; used for: 7 or more hours total sleep duration recommendation.
- National Heart, Lung, and Blood Institute — Your Guide to Healthy Sleep (2011 revision) — Tier 1 official education; used for: deep sleep approximately one-fifth; REM approximately one-fifth.
- Ohayon et al. — Meta-Analysis of Quantitative Sleep Parameters From Childhood to Old Age in Healthy Individuals, Sleep (2004) — Tier 1 meta-analysis; used for: age-related decrease in slow-wave sleep, REM, and REM latency.
- de Zambotti et al. — State of the science and recommendations for using wearable technology in sleep and circadian research, Sleep (2024) — Tier 1 state-of-science review; used for: wearable misclassification of wakefulness in context paragraphs.
- Klier et al. — Consumer-oriented (wearable) sleep technology: a systematic SWOT analysis and recommendations for athletes and fitness enthusiasts, Frontiers in Sports and Active Living (2026) — Tier 1 systematic qualitative synthesis; used for: 2026 synthesis limitations in wearable sleep-stage measurement in context paragraphs.
- Schyvens et al. — A performance validation of six commercial wrist-worn wearable sleep-tracking devices for sleep stage scoring compared to polysomnography, SLEEP Advances (2025) — Tier 1 primary device-validation study; used for: Apple Watch Series 8 N3 sensitivity 50.66%; Apple Watch Series 8 REM sensitivity 68.57%; Whoop 4.0 N3 sensitivity 69.63%; context paragraphs on how wrist devices estimate stages.
Figures checked · 7 figures from 6 sources · Next review
How we made this page: figures gathered and checked from the sources above, written up with AI assistance and edited by Aamir Bashir, Founder & Editor. Read our editorial policy.
