Sleep Tracking Accuracy: Wearables vs Clinical Polysomnography

Your watch gives you a sleep score every morning. Here’s what that number can and can’t actually tell you — and where consumer sleep tracking stands against the clinical gold standard.

Sleep tracking has become a standard feature on smartwatches, fitness bands, and rings, and millions of people now wake up to a detailed breakdown of their night: time asleep, sleep stages, a score out of 100. The data looks authoritative. The question worth asking is how accurate it really is — and what you should and shouldn’t do with it.

The honest answer is nuanced. Consumer sleep trackers have genuinely improved and are useful for some things, but they measure sleep indirectly and have real, well-documented limitations compared with the clinical methods they’re often implicitly compared to.

The Gold Standard: Polysomnography

Polysomnography (PSG) is the clinical reference standard for measuring sleep, performed in a sleep lab (or, in limited forms, at home). It directly measures the things that define sleep: brain activity via EEG, eye movement via EOG, muscle activity via EMG, plus heart rhythm, breathing, blood oxygen, and more. Trained technologists score the night into stages — wake, light sleep, deep (slow-wave) sleep, and REM — based primarily on brain-wave patterns.

This is the key distinction: PSG measures sleep by reading the brain directly. It’s the basis for diagnosing sleep disorders like sleep apnea, narcolepsy, and others — and it’s what any consumer device is being measured against when researchers test accuracy.

How Consumer Trackers Actually Work

Wearables don’t read your brain. They infer sleep from proxies — primarily movement (accelerometry), heart rate and heart-rate variability, and sometimes skin temperature, breathing rate, and blood oxygen. Algorithms combine these signals to estimate whether you’re asleep and which stage you’re in.

This indirect approach has predictable consequences. The earliest movement-only trackers (and basic ones today) essentially detected stillness, and tended to overestimate sleep — if you lie still while awake, the device may score you as asleep. Adding heart-rate and HRV data improved things substantially, because those signals shift with sleep stages, which is why modern devices estimate stages at all. But it remains an estimate built from the body’s downstream signals, not a direct reading of the brain.

What the Research Shows About Accuracy

Studies comparing consumer devices to PSG have found a consistent pattern, and it’s worth understanding rather than dismissing:

• Detecting “asleep vs awake” is the strongest area. Modern wearables are reasonably good at estimating total sleep time and at telling broadly whether you slept — often within a tolerable margin for general tracking.
• Stage classification is much weaker. Distinguishing light, deep, and REM sleep accurately is genuinely hard without EEG. Devices have improved but still misclassify stages meaningfully compared with PSG, so the precise “2h 14m of deep sleep” figure is best treated as a rough estimate, not a measurement.
• Detecting wake after falling asleep is a known weak point. Brief awakenings and quiet wakefulness are easy for movement-based systems to miss or misread.
• Accuracy varies by device, population, and night. Performance differs between brands and is generally lower in people with sleep disorders — exactly the group most interested in the data.

The reasonable summary: today’s better wearables are useful for tracking broad patterns and total sleep over time, less reliable for the specific stage breakdowns they display most prominently.

What Sleep Trackers Are Genuinely Good For

• Spotting trends over time. Even if a single night’s numbers are imprecise, consistent patterns — your typical sleep duration, how weekends differ, the effect of late caffeine or alcohol — are where the data earns its keep.
• Behavioural awareness and motivation. Seeing the effect of a regular bedtime, or of screens before bed, can reinforce better habits. For many people that nudge is the real value.
• Flagging things worth a closer look. Some devices estimate breathing disturbances or oxygen dips that may prompt someone to seek a proper evaluation — useful as a flag, not a diagnosis.

Where They Can Do Harm

There’s a recognized downside worth naming: orthosomnia — anxiety about sleep driven by tracker data. People can become preoccupied with optimizing their scores, stressed by a “bad” reading, or convinced they slept poorly because the app said so, even when they feel fine. Because the stage data in particular is an estimate, treating a low deep-sleep number as a medical fact can create a problem that didn’t exist. If your tracker is making you more anxious about sleep rather than less, that’s a sign to step back from the numbers.

Equally important: a consumer tracker cannot diagnose a sleep disorder. If you have symptoms of sleep apnea — loud snoring, gasping or pauses in breathing, severe daytime sleepiness — a wearable’s reassuring (or alarming) numbers are not a substitute for a clinical evaluation. See a doctor.

How to Use a Sleep Tracker Sensibly

• Trust trends, not single nights. Look at weekly and monthly patterns rather than reacting to one night’s score.
• Treat stage breakdowns as estimates. Total sleep time is more reliable than the deep/REM split.
• Use it to test behaviours — bedtime consistency, caffeine timing, alcohol — rather than to chase a perfect score.
• Believe how you feel over what the app says. If you feel rested but got a low score, the score is probably wrong, not your body.
• Escalate real symptoms to a doctor. A tracker is a starting point for a conversation, not a diagnosis.

The Bottom Line

Consumer sleep trackers infer sleep from movement and heart-rate signals rather than measuring the brain directly, which is why they’re reasonably good at estimating whether and how long you slept, and much weaker at the stage-by-stage breakdowns they display so confidently. Polysomnography remains the only method that actually measures sleep architecture and diagnoses sleep disorders. Used for trends and behavioural awareness, a tracker is genuinely useful; used as a precise medical instrument — or as a source of sleep anxiety — it’s the wrong tool. And no wearable replaces a doctor when real symptoms are present.

This article is general information, not medical advice. If you have symptoms of a sleep disorder, talk to a healthcare professional about proper evaluation.