---
title: "How Accurate Is Your Wearable's Sleep Staging — For You?"
description: "Published studies give wearable sleep staging a group average — Apple Watch scores highest, but none of them describe your wrist. Our three-night home sleep study records EEG, so you can compare your own device against measured sleep stages, night by night."
url: https://www.nocturnehealth.com/wearable-sleep-tracking-accuracy
markdown_url: https://www.nocturnehealth.com/wearable-sleep-tracking-accuracy.md
site: Nocturne Health
last_reviewed: 2026-07-01
---

Your Wearable vs. Your Brain Waves

# How accurate is your wearable's sleep staging for you?

The published studies can only tell you how it performed on someone else. Our three-night home sleep study records EEG — so for once, you can hold your own wearable up against measured sleep stages.

Medical content reviewed by the sleep medicine physicians at Nocturne Health · Last updated July 2026

## Short answer: good at sleep, mediocre at stages, unknown on you

Apple Watch has the most rigorous published validation of any consumer wearable, so it’s the example we use throughout this page — the same questions apply to whatever you’re wearing. Apple Watch is good at telling whether you’re asleep or awake: Apple’s validation work reports sensitivity for sleep near 98%. It’s less reliable at telling which stage you were in. Against attended polysomnography, independent researchers put its four-stage agreement at a Cohen’s kappa of 0.53 — the highest of six wearables tested in the same study, and still only _moderate_: clearly better than chance, clearly not a measurement.

Those are averages across a study population, not a description of your wrist. The only way to find out how your wearable performs on you is to record your sleep stages properly on the same nights it’s estimating them — and that requires EEG.

## What the research actually found

A 2025 validation study in _SLEEP Advances_ put six consumer wearables against attended in-lab polysomnography, epoch by epoch. Apple Watch Series 8 results:

| Stage | Agreement with EEG | Typical error |
| --- | --- | --- |
| Light sleep (Core, N1+N2) | 83% of epochs matched | Overestimated by about 59 minutes a night |
| Deep sleep (N3) | 51% of epochs matched | Underestimated by about 25 minutes a night |
| REM sleep | 69% of epochs matched | Underestimated by about 13 minutes a night |
| Wake after falling asleep | 52% of epochs matched | Underestimated by about 21 minutes a night |

The right-hand column is a pattern, not a set of unrelated errors. The watch pulls time out of deep sleep, REM, and wake, and deposits it into light sleep. If you have ever looked at a night where you know you lay awake for an hour and the watch showed you peacefully in Core, that is this effect.

## How other wearables compared

The same study tested five other wearables against the same PSG reference, scored on the same four stages. Apple Watch came out on top; the rest ranged from moderate down to only fair:

| Wearable | 4-stage kappa vs. PSG | Agreement |
| --- | --- | --- |
| Apple Watch Series 8 | 0.53 | Moderate — highest of the six |
| Fitbit Sense | 0.42 | Moderate |
| Fitbit Charge 5 | 0.41 | Moderate |
| Whoop 4.0 | 0.37 | Fair |
| Withings ScanWatch | 0.22 | Fair |
| Garmin Vivosmart 4 | 0.21 | Fair |

Whatever you’re wearing, the logic is the same: a published kappa is a group average, not a fact about your sleep. Wear whichever device you own on your study nights — the comparison below works the same way regardless of brand.

## Why the average doesn't answer your question

A kappa of 0.53 is a summary of a distribution. Somewhere in that study were people whose watch tracked their EEG closely and people whose watch was nearly uninformative, and the paper cannot tell you which one you are. Several things push individuals toward the tails:

- Validation cohorts skew young, healthy, and free of the sleep disorders that bring people to a sleep physician
- The algorithm infers sleep stage from wrist movement and pulse. People who lie unusually still, or who move a lot during light sleep, end up on opposite ends of that error range
- Untreated sleep apnea fragments sleep in ways an accelerometer systematically misreads
- Fit, wrist, band tightness, tattoo ink over the sensor, atrial fibrillation, and beta blockers all change the pulse signal the model depends on
- A study average is a single number across hundreds of people. Measuring your own wrist is the only way to answer the question for you

This isn’t a knock on Apple Watch, or any other wearable. It just means the question “how accurate is it?” has no useful answer until you attach a person to it.

## How the comparison works

Most home sleep apnea tests cannot do this. A standard Level III test records airflow, breathing effort, oxygen, and heart rate — no EEG, which means no sleep stages and no way to score the very thing your wearable is claiming to measure. Nocturne Health uses a Level II study instead: it records EEG, so sleep stages come from brain activity rather than being inferred from movement and heart rate, and it runs for three nights rather than one.

- A sleep medicine physician reviews your intake and orders a home sleep study if testing is clinically indicated
- An FDA-cleared Level II device ships to your door. It records EEG, so sleep stages are measured from brain activity rather than inferred from movement
- You wear your wearable — Apple Watch or otherwise — to bed as you normally would, on all three recording nights
- The study is scored and interpreted by a sleep medicine physician, typically within 72 hours
- Bring your wearable's nightly sleep-stage breakdown to your follow-up visit and we can line it up against the EEG record, night by night

Three nights matters here. A single night gives you one comparison, and any wearable can get one night right by luck. Three nights gives you a trend instead.

## What you can learn from it

- Whether your wearable's deep-sleep number tracks your measured N3 (deep sleep) or drifts from it — and in which direction
- Whether the nights your wearable calls bad are the nights your sleep was actually fragmented
- How much wake time your wearable is missing
- Whether your total sleep time is roughly right even when the stage breakdown is not — a common pattern
- Whether the wearable is worth trusting as a week-to-week trend line for you, once you know its offset

The useful outcome is usually not “the wearable is wrong.” It is knowing the shape of its error on you, which turns a number you half-trust into one you can read properly — or one you can stop worrying about.

If what brought you here was a notification rather than curiosity, start with [the Apple Watch sleep apnea notification](https://www.nocturnehealth.com/apple-watch-sleep-apnea-notification). If you want the broader picture of what home testing does and doesn’t cover, read [home sleep test vs. in-lab study](https://www.nocturnehealth.com/home-sleep-test-vs-in-lab).

## Find out what your sleep actually looks like

Start with a virtual consultation ($199) with a sleep medicine physician. If a home sleep study is indicated, it's $169 — three nights, EEG-based sleep staging, results typically within 72 hours. Available to patients in Arkansas, California, Massachusetts, Ohio, and Pennsylvania.

## Wearable sleep accuracy: common questions

### How accurate is wearable sleep tracking compared to a sleep study?

It depends on the device. Apple Watch has the most data behind it: Apple's own validation work reports sensitivity for telling sleep from wake around 98%, and in a 2025 independent validation against attended polysomnography, Apple Watch Series 8 topped six wearables tested with a four-stage Cohen's kappa of 0.53 — moderate agreement, matching about 83% of light-sleep epochs, 69% of REM epochs, and 51% of deep-sleep epochs. The other five devices in that study ranged from 0.42 down to 0.21, moderate to only fair. Those are all group averages; individual results vary considerably, whichever device you use.

### Do wearables measure brain waves?

No consumer wearable does, Apple Watch included. They estimate sleep stages from wrist motion and pulse data, then apply a model trained against EEG recordings. None has electrodes on your head or measures brain activity directly. Sleep staging is defined by EEG, eye movement, and muscle tone, so anything without those signals is producing an estimate, however good the model behind it.

### How is your home sleep study different from a typical home sleep apnea test?

Most home sleep apnea tests are Level III studies: they record airflow, breathing effort, oxygen, and heart rate, but no EEG, so they cannot tell sleep from wake or score sleep stages. Nocturne uses a Level II study, which adds EEG-based sleep staging, and records for three nights rather than one. That is what makes a direct comparison against your wearable possible — there is a measured sleep stage to compare against. The study is $169 when a physician determines testing is indicated.

### Can I order the study just to benchmark my wearable?

Being curious about how well your wearable tracks your sleep is a fine reason to book a consultation. A sleep medicine physician still reviews your intake and orders the study, and the comparison comes along with it at no extra cost. If something in your intake points to a separate issue like sleep apnea, the physician will raise that too.

### Which Apple Watch models support sleep staging?

Sleep stages arrived with watchOS 9 in 2022 and are supported on Apple Watch Series 4 and later, Apple Watch SE, and Apple Watch Ultra, provided the watch and iPhone are on a supported OS version and Sleep Focus and sleep tracking are switched on. The separate sleep apnea notification feature is newer and runs on a narrower set of models.

### Is the comparison a formal part of my report?

The clinical report is the physician's interpretation of your sleep study — that is what testing decisions are based on, and it stands on its own. The wearable comparison is something you and the physician can walk through together in your follow-up visit, using your own wearable's app data alongside the study record. It is context, not a diagnostic finding.

## References

1. [Schyvens AM, Peters B, Van Oost NC, et al. A performance validation of six commercial wrist-worn wearable sleep-tracking devices for sleep stage scoring compared to polysomnography. SLEEP Advances, 2025;6(2):zpaf021.](https://academic.oup.com/sleepadvances/article/6/2/zpaf021/8090472)
2. [Apple Inc. Estimating Sleep Stages from Apple Watch (technical white paper), October 2025.](https://www.apple.com/health/pdf/Estimating_Sleep_Stages_from_Apple_Watch_Oct_2025.pdf)
3. [Berry RB, et al. The AASM Manual for the Scoring of Sleep and Associated Events. American Academy of Sleep Medicine.](https://aasm.org/clinical-resources/scoring-manual/)
4. [Kapur VK, et al. Clinical Practice Guideline for Diagnostic Testing for Adult Obstructive Sleep Apnea. Journal of Clinical Sleep Medicine, AASM, 2017.](https://jcsm.aasm.org/doi/10.5664/jcsm.6506)

---

_HTML version of this page: https://www.nocturnehealth.com/wearable-sleep-tracking-accuracy_

_Nocturne Health is a direct-to-consumer virtual sleep medicine practice — virtual consultations with sleep medicine physicians and FDA-approved at-home sleep apnea testing. Overview for AI agents: https://www.nocturnehealth.com/llms.txt · Machine-readable actions: https://www.nocturnehealth.com/mcp-actions.json_