Key takeaways
Three things to remember
- 01
Actigraphy estimates sleep from movement and is the method underlying every consumer wearable’s sleep data; it is excellent for trends across weeks but has a precision ceiling of roughly 10 to 15 percentage points for individual sleep stage measurements.
- 02
Actigraphy’s biggest advantage over lab sleep studies is its ability to capture 30 to 90 nights of continuous real-world data, which is how sleep regularity, chronotype drift, and behavioral change actually show up.
- 03
When wearable sleep data and subjective recovery consistently conflict, a clinical sleep study is the appropriate next diagnostic step, not more actigraphy.
How it works
The Mechanism
Every consumer sleep tracker, including Oura, WHOOP, Apple Watch, and Garmin devices, is fundamentally an actigraphy device with additional physiological inputs. Pure actigraphy devices (like those used in clinical sleep research) use only movement; modern consumer devices layer in heart rate, heart rate variability, skin temperature, and pulse oximetry to improve staging accuracy.
The core actigraphy algorithm treats sustained low movement as sleep and elevated movement as wakefulness. This works reasonably well for distinguishing sleep from wakefulness overall: actigraphy achieves roughly 85 to 90% accuracy for sleep vs. wake classification in healthy adults. The problem is specificity for sleep stages. Because movement is a poor proxy for brain state, actigraphy alone cannot distinguish N2 from N3, and REM sleep (which involves near-complete muscle paralysis) looks similar to any other immobile period. Modern wearables improve stage discrimination by adding heart rate variability patterns, but the fundamental accuracy ceiling for staging without EEG data remains well below polysomnography.
Where actigraphy genuinely outperforms lab studies is in longitudinal data collection. A sleep lab provides one or two nights of precise staging, which may or may not represent typical sleep. Actigraphy provides 30 to 90 nights of continuous data, capturing patterns like social jetlag, cumulative sleep debt trends, and the behavioral impact of schedule changes that a single lab night cannot reveal. For research into sleep regularity, chronotype, and the population-level effects of sleep timing on health outcomes, actigraphy has been indispensable. The UK Biobank studies on sleep regularity and mortality used wrist actigraphy data from hundreds of thousands of participants.
In practice
Why It Matters
Actigraphy is best for patterns across weeks, not precision on any given night.
Most people interact with actigraphy every day without realizing it: their wearable’s sleep data is an actigraphy-based estimate. Understanding what the method can and cannot do clarifies how to use wearable sleep data intelligently. Short-term stage data has meaningful noise. Long-term trend data, which is what actigraphy is best at, is genuinely useful for tracking behavioral changes and identifying chronic sleep patterns.
Common misconception
Because consumer wearables give you a detailed sleep stage breakdown nightly, many people treat each night’s data as a precise measurement. Actigraphy-based staging should be read as a probabilistic estimate with a noise range of roughly 10 to 15 percentage points per stage. A night showing 15% deep sleep versus 20% deep sleep on another night may reflect real variation, measurement noise, or both.
Signs it is disrupted
- Persistent discrepancy between how you feel and what your sleep data shows, which can indicate that the actigraphy estimate is missing something (common with certain sleep disorders like UARS where movement is normal but arousal is high)
- Very low movement during a period you were actually awake and still (reading in bed, for example), which actigraphy may incorrectly score as sleep
- Wearing the device inconsistently or on the non-dominant wrist rather than dominant, which reduces actigraphy accuracy
- Restless sleep with frequent position changes, which can cause actigraphy to overestimate wakefulness even if sleep quality is adequate
What moves it
How to Improve It
Wear consistently. Actigraphy reliability improves with consistent device placement and nightly use; gaps in the dataset remove the longitudinal context that makes the method valuable.
Read trends, not nights. Use 7 to 14 day rolling averages for sleep metrics rather than reacting to single-night values; this is how actigraphy was designed to be used in clinical sleep research.
Cross-check with subjective feel. Tracking perceived sleep quality alongside wearable data helps identify when the device is systematically misreading your sleep (e.g., consistently scoring high while you feel unrefreshed).
Consider clinical testing. When wearable data and subjective experience diverge chronically, a home sleep test or polysomnography provides the EEG-based accuracy that actigraphy cannot offer.
Where to see it
Which Devices Track It
- Oura Ring
- Oura adds skin temperature and HRV to its accelerometer data, making it one of the better-validated consumer actigraphy devices; published validation against PSG shows 69 to 76% epoch agreement for staging, with sleep vs. wake accuracy above 90%.
- WHOOP
- WHOOP uses a 3-axis accelerometer plus photoplethysmography (PPG) for heart rate; its staging accuracy is in the same range as Oura in independent studies, with both devices performing better at overall sleep architecture than precise stage durations.
- Apple Watch
- Apple Watch uses wrist accelerometry plus optical heart rate; staging accuracy from watchOS 9 onward is broadly similar to other consumer devices in the 65 to 75% epoch agreement range with PSG.
- Garmin
- Garmin devices use accelerometry plus optical heart rate and pulse oximetry; staging accuracy is comparable to other consumer wearables, with body battery and overnight monitoring as primary consumer outputs.
Go deeper