Polysomnography (PSG)
The gold-standard clinical measurement of sleep
Plain English
Polysomnography (PSG) is a comprehensive sleep study conducted in a lab (or increasingly at home) that records brain waves, eye movements, muscle activity, heart rate, respiratory effort, airflow, and blood oxygen simultaneously throughout the night. It is the reference standard against which all consumer wearable sleep tracking is measured.
The Mechanism
A full polysomnography study records multiple physiological channels at once. Electroencephalography (EEG) measures the electrical activity of the brain using electrodes placed on the scalp; the pattern of these waves is what defines each sleep stage. Slow, high-amplitude delta waves mark slow-wave sleep (N3). Sleep spindles and K-complexes appear in N2. Rapid eye movement (REM) sleep is identified by a combination of active brain waves, near-complete muscle paralysis, and rapid conjugate eye movements tracked by electrooculography (EOG). Electromyography (EMG) records chin and limb muscle activity, detecting the muscle atonia of REM and the limb movements of conditions like restless leg syndrome.
Respiratory channels capture airflow at the nose and mouth, respiratory effort via chest and abdominal belts, and blood oxygen saturation via pulse oximetry. This combination allows the scoring of respiratory events: apneas (complete cessation of airflow) and hypopneas (partial reductions). The number of these events per hour of sleep is the apnea-hypopnea index (AHI), which is the primary metric for diagnosing obstructive sleep apnea (OSA). An AHI above 5 meets the threshold for mild OSA; above 15 is moderate; above 30 is severe.
Sleep staging from a PSG is performed manually by a trained technician using American Academy of Sleep Medicine (AASM) scoring rules, reviewing the EEG record in 30-second epochs. Consumer wearables estimate staging from surrogate signals (heart rate, movement, skin temperature) without any EEG data, which is why accuracy for individual stage percentages is limited to within 10 to 15 percentage points of PSG staging on average.
Why It Matters
Without brain waves, you are estimating sleep, not measuring it.
Polysomnography is the diagnostic standard for sleep apnea, narcolepsy, parasomnias, and other clinical sleep disorders. Understanding what PSG measures explains the ceiling on consumer wearable accuracy: without EEG, wearables cannot directly observe brain state and must infer it from cardiovascular and movement proxies. When sleep issues are suspected but wearable data is ambiguous, a home sleep test or in-lab PSG is the diagnostic next step.
Common Misconception
Many people assume wearables give them their own polysomnography-equivalent data every night. Consumer devices are validated against PSG and perform reasonably well at overall stage proportions and sleep onset detection, but their accuracy for precise stage durations in any given night is limited. A 10-minute difference in your deep sleep reading is often within the measurement error range, not a meaningful signal.
Signs It Is Disrupted
- Persistent fatigue despite adequate sleep time, where no wearable explanation is apparent and a clinical evaluation for OSA or other disorders is warranted
- Wearable data showing consistently low HRV, elevated resting heart rate, or fragmented sleep across weeks despite behavior optimization, suggesting an undetected sleep disorder
- Witnessed apneas, loud snoring, or gasping during sleep, which are strong clinical indications for a PSG or home sleep test
- Excessive daytime sleepiness despite 7 to 8 hours of sleep, particularly with cataplexy or sleep paralysis, which would point toward a PSG evaluation for narcolepsy
How to Improve It
Which Devices Track It
Oura Ring
Oura is validated against PSG for overall sleep staging with moderate accuracy (roughly 69% agreement on epoch-by-epoch staging); it performs best at distinguishing sleep vs. wake and worst at precisely separating N1 from N2. Published validation studies (de Zambotti et al.) show sensitivity for deep sleep detection is lower than REM detection.
WHOOP
WHOOP uses a similar cardiac and accelerometer-based staging approach and shows comparable accuracy to other consumer wearables in independent validation, with per-stage accuracy ranging from 60 to 80% depending on the stage and study.
Apple Watch
Apple Watch sleep staging is available from watchOS 9 and uses heart rate, respiratory rate, and movement. Published accuracy is similar to other consumer devices; it provides useful summary data but not clinical-grade staging.
Garmin
Garmin devices use a pulse oximeter and accelerometer for staging and offer body battery and sleep score composites; staging accuracy is broadly similar to other wearables at 65 to 75% epoch agreement with PSG in consumer validation studies.
3 Things to Remember
Polysomnography measures sleep using direct brain wave recording (EEG), which is why it is the clinical gold standard and consumer wearables, which use heart rate and movement, cannot fully replicate its staging accuracy.
The apnea-hypopnea index (AHI) from PSG is the primary diagnostic metric for sleep apnea: above 5 is mild, above 15 is moderate, above 30 is severe.
If your wearable shows persistently low sleep quality despite behavior optimization, or if you have risk factors for sleep apnea, a home sleep test or in-lab PSG is the appropriate next step, not another month of app data.
Appears In
Related Terms
Protocol
Turn what you've learned into daily practice
Protocol pulls your wearable and nutrition data together into a daily health score, morning brief, and AI coaching. All in one place.
Get started free