Answer

How accurate are sleep trackers for REM and deep sleep?

Sleep sleep sleep-tracker wearables

Verdict

Leaning for

Sleep trackers measure sleep versus wake well but label REM and deep sleep poorly.

+16 32 sources: 8 positive, 18 partly, 6 negative

32 sources: 8 positive, 18 partly, 6 negative

By Arun AgrahriLast reviewed

Based on what researchers say in public interviews on YouTube. Educational only, not medical advice. Ask a qualified clinician before changing a supplement, medication, treatment, or testing plan.

Short answer

Not very, for stages. Sleep trackers tell sleep from wake well enough for trends. Their REM and deep-sleep labels are much less reliable, because they infer stages from body signals instead of measuring brain waves. Sleep researcher Michael Grandner advises watching long-term patterns, not one night's percentages or a proprietary sleep score.

Leaning for

Sleep is the most important pillar of health and longevity.

The verdict on Sleep overall. 8 sources: 4 positive, 4 partly

What the evidence says

positive partly neutral negative Each researcher and each study counts once per claim, however often they come up.

  1. Leaning for

    Sleep trackers measure sleep versus wake well but label REM and deep sleep poorly.

    32 sources: 8 positive, 18 partly, 6 negative.

    Researchers say

    1 positive

    Studies show

    7 positive, 18 partly, 6 negative

    31 papers we read

    7 more not counted · see why
    The study cited could not be identified. 3
    The study is not yet read in full. 2
    We could not confirm who is speaking. 1
    We could not tell where the evidence comes from. 1
    Show the receipts from 32 sources

    1. Russell Foster, on Rangan Chatterjee's channel, at 1:37:16 Video · positive · Opinion

      “they say oh you've had lots of good deep sleep or you've haven't had enough REM sleep and and for that they're completely useless”

    2. Trial or observed group, Sleep, 2015, PMID 26158896 Study · positive

      Jawbone UP shows good agreement with polysomnography in measures of total sleep time and wake after sleep onset in adolescent boys and girls.

    3. Other primary study, Nat Sci Sleep, 2020, PMID 33149712 Study · positive

      Generally speaking, commercial sleep technologies displayed lower error and bias values when quantifying sleep/wake states as compared to sleep staging durations.

    4. Other primary study, Nat Sci Sleep, 2021, PMID 33623459 Study · positive

      EBE analyses demonstrated excellent sleep-wake accuracies, specificities, and sensitivities - between 0.88 and 0.89 across all TIBs.

    5. Other primary study, Sensors (Basel), 2022, PMID 36016077 Study · positive

      Analyses regarding the two-state categorisation of sleep indicate that all six devices are valid for the field-based assessment of the timing and duration of sleep. However, analyses regarding the multi-state categorisation of sleep indicate that all six devices require improvement for the assessment of specific sleep stages.

    6. Other primary study, Annu Int Conf IEEE Eng Med Biol Soc, 2023, PMID 38083143 Study · positive

      The Apple Watch performed well for identifying sleep-wake states but had difficulty identifying the sleep stages compared to the reference PSG system.

    7. Other primary study, Sci Rep, 2025, PMID 40108409 Study · positive

      Sleep/Wake distinction of the Oura and SleepOn rings reached similar performance as previously reported for healthy individuals (~ 85% accuracy), but was worse for the Circul ring (~ 65% accuracy).

    8. Trial or observed group, Sensors (Basel), 2025, PMID 40807979 Study · positive

      The device showed limited precision in classifying sleep stages, often misclassifying deep and REM sleep as light sleep.

    9. Other primary study, Behav Sleep Med, 2019, PMID 28323455 Study · partly

      From EBE analysis, ŌURA ring had a 96% sensitivity to detect sleep, and agreement of 65%, 51%, and 61%, in detecting "light sleep" (N1), "deep sleep" (N2 + N3), and REM sleep, respectively.

    10. Trial or observed group, Chronobiol Int, 2018, PMID 29235907 Study · partly

      Fitbit Charge 2™ shows promise in detecting sleep-wake states and sleep stage composition relative to gold standard PSG, particularly in the estimation of REM sleep, but with limitations in N3 detection.

    11. Other primary study, J Sleep Res, 2019, PMID 30407680 Study · partly

      These results suggest FBA-HR cannot replace EEG-based measurements of sleep and wake in the diagnostic assessment of suspected CDH, and that improvements in device performance are required prior to adoption in clinical or research settings.

    12. Other primary study, J Clin Sleep Med, 2019, PMID 30853052 Study · partly

      Also, it failed to discriminate between non-rapid eye movement sleep stages and did not detect the rapid eye movement sleep stage.

    13. Other primary study, J Clin Sleep Med, 2019, PMID 31383243 Study · partly

      Sleep detection accuracy by S+V1 (93.3%), S+V2 (93.8%), and actigraphy (96.0%) was high; wake detection accuracy by each (69.6%, 73.1%, and 47.9%, respectively) was low.

    14. Other primary study, J Sleep Res, 2020, PMID 31626361 Study · partly

      FBA also showed high specificity when identifying N3, and rapid eye movement, though sensitivity was modest.

    15. Other primary study, PLoS One, 2020, PMID 33306678 Study · partly

      GV4 detected sleep with high accuracy (0.90) and sensitivity (0.98) but low specificity (0.28).

    16. Other primary study, JMIR Mhealth Uhealth, 2021, PMID 33533726 Study · partly

      REM sleep was detected with 0.92 specificity, 0.60 sensitivity, and 0.84 accuracy.

    17. Other primary study, Chronobiol Int, 2021, PMID 33792456 Study · partly

      Our results are in line with a general trend in the literature, suggesting better performance for the recently introduced multi-sensor devices compared to motion-only devices, although further developments are needed to improve accuracy in sleep stage classification and wake detection.

    18. Other primary study, Biosensors (Basel), 2021, PMID 34201016 Study · partly

      For WHOOP-AUTO, sensitivity for sleep, wake, and agreement for two-stage and four-stage categorisation of sleep were 90%, 60%, 86% and 63%, respectively.

    19. Other primary study, J Med Internet Res, 2021, PMID 34609317 Study · partly

      Nevertheless, the generally wide limits of agreement hamper the precision of quantifying individual sleep episodes.

    20. Other primary study, Sleep Breath, 2022, PMID 34741243 Study · partly

      Most of the applications tested in this study showed poor validity, while wearable devices mildly correlated with PSG.

    21. Other primary study, PLoS One, 2022, PMID 36256631 Study · partly

      This research shows that FBC cannot replace PSG thoroughly in the quantification of sleep variables and classification of sleep stages in Chinese patients with chronic insomnia; however, the user-friendly and low-cost wearables do show some comparable functions.

    22. Other primary study, Sensors (Basel), 2022, PMID 36502241 Study · partly

      If precise data on sleep stages and parameters are required, their use is limited.

    23. Trial or observed group, JMIR Mhealth Uhealth, 2023, PMID 37917155 Study · partly

      More specifically, the highest macro F1 score was 0.69, while the lowest macro F1 score was 0.26.

    24. Other primary study, Sleep Health, 2024, PMID 38087674 Study · partly

      This was followed by the iteratively improved consumer trackers: Oura (2-stage kappa=0.64, 4-stage kappa=0.55-0.70) and Fitbit (2-stage kappa=0.58, 4-stage kappa=0.45-0.60) which had comparable total sleep time and sleep efficiency discrepancies that outperformed accelerometry-only Actigraph (2-stage kappa=0.47).

    25. Other primary study, Sensors (Basel), 2024, PMID 39460013 Study · partly

      In adults with healthy sleep, all the devices were similar to PSG in the estimation of sleep duration, with the devices also showing moderate to substantial agreement with PSG-derived sleep stages.

    26. Trial or observed group, Sleep Med, 2026, PMID 42258963 Study · partly

      Across devices, agreement with PSG was generally weak with wide LOA.

    27. Trial or observed group, J Affect Disord, 2017, PMID 28448949 Study · negative

      The FBF demonstrated significant limitations in quantifying sleep and wake, relative to PSG.

    28. Other primary study, Sensors (Basel), 2019, PMID 31557952 Study · negative

      Surprisingly, all devices had poor agreement with the PSG gold standard.

    29. Other primary study, J Clin Sleep Med, 2020, PMID 32048595 Study · negative

      Basis B1 specificity was poor, and Basis B1 is not useful for the assessment of wake.

    30. Other primary study, Sleep Med, 2024, PMID 38382312 Study · negative

      The Oura Ring Gen3 with OSSA 2.0 shows good agreement with PSG for global sleep measures and time spent in light and deep sleep.

    31. Trial or observed group, J Clin Sleep Med, 2024, PMID 38450553 Study · negative

      FB-estimated total sleep and wake times and sleep stage times were in good agreement with PSG, with intraclass correlations of 0.79-0.96.

    32. Other primary study, Sleep Adv, 2025, PMID 40303381 Study · negative

      Nevertheless, all wearables demonstrated a higher percentage of correctly identified epochs for deep sleep and rapid eye movement sleep compared to wake (W) and LS.

Where the sources disagree

Leaning for

Sleep trackers measure sleep versus wake well but label REM and deep sleep poorly.

Supports it 8 sources

Researchers

Studies

Partly 18 sources
Against it 6 sources

Precis is not affiliated with or endorsed by YouTube or by any researcher, show or publisher we quote. We take no sponsorship, ad or affiliate money, and every verdict is our own review of what was said. How we stay independent

Detailed Answer

Sleep trackers measure sleep versus wake well but label REM and deep sleep poorly. Telling whether you were asleep or awake is driven largely by movement and is reasonably reliable. The impressive-looking breakdown of REM, deep and light sleep is where they are weakest.

Consumer wearables infer sleep stages from indirect body signals rather than measuring brain activity, which a clinical sleep study (polysomnography) does with EEG. Michael Grandner, a sleep researcher and guest on Rhonda Patrick's show, describes the stage breakdown as a rough estimate, not a precise nightly measurement.

Be wary of the score

The "sleep score" or "readiness" number is often proprietary and not clinically validated, Grandner says, so it may not map onto any real health outcome. Treating one number as a verdict on your health can mislead. Watch long-term trends instead: your typical total sleep time, how regular your schedule is, and signals like resting heart rate.

What a tracker is good for

Used that way, a tracker can still help. It can reassure you that you slept more than it felt like, or surface a pattern of broken sleep that explains daytime fatigue. Persistently irregular patterns can be a nudge to look at your bedroom or to ask a clinician about something like sleep apnea.

What it cannot do is diagnose you; the stage percentages are a trend tool, not a medical measurement.

Related Questions

Are Oura, Whoop, and Apple Watch accurate for sleep stages?
They are good at telling sleep from wake but only a rough estimate for REM and deep sleep, which they infer from body signals rather than brain waves. Use them for trends, not exact nightly stage percentages.
Should I trust my sleep score?
Be cautious. Grandner notes these scores are often proprietary and not clinically validated, so they may not reflect a health outcome. Long-term trends mean more than any single night's number.
What sleep metric is actually worth watching?
Consistent trends, such as typical total sleep time, schedule regularity and resting heart rate, are more useful than stage breakdowns. Resting heart rate in particular is a strong signal of sleep quality.
Can a tracker diagnose a sleep disorder?
No. A tracker can flag persistently irregular patterns worth checking, but diagnosing a condition like sleep apnea takes a clinical evaluation, not a wearable's estimate.

More Questions About Sleep

Can a sleep tracker make insomnia worse? Why does alcohol help me fall asleep but ruin m... Does magnesium help with sleep?

Precis membership

Get the exact protocol for Sleep

The answer above is the whole answer: the verdict, every counted mention, where the sources disagree, and the receipt behind each one. Nothing on this page is held back.

A free account adds

  • Each dose named on the record, and the form named with it
  • Timing, and what to take it with or keep it away from
  • The interactions flagged by name

Pro adds

  • Chat answers drawn from 3,700+ hours of recorded interviews with 380 researchers and doctors, plus 48,500+ research studies, each with a timecode
Go Pro — $9/mo or $79/yr

Full Sleep evidence report

The verdict on every claim, and where the sources disagree

How Precis checks a claim · Report an error

All answers