Sleep tracking has become one of the most common personal health practices. Consumer wearables estimate sleep stages with meaningful accuracy for some and poor accuracy for others. Author on what the research shows, what the clinically significant stages are, and what actually improves them.
“Wearables detect slow-wave sleep — where your body releases growth hormone and clears brain waste — with sensitivity between 37 and 82 percent depending on the device. The most clinically significant stage is the one consumer technology measures least reliably. This does not mean stop tracking. It means stop optimising against data that may be a detection error.”
Sleep is not a uniform state. The clinical model of sleep architecture describes a cyclical pattern that repeats approximately four to six times across a full night's sleep, each cycle lasting between seventy and one hundred and ten minutes.
Within each cycle, sleep progresses through three stages of non-rapid eye movement sleep (NREM 1, 2, and 3, also known as slow-wave or deep sleep) before entering rapid eye movement (REM) sleep. Slow-wave sleep predominates in the first two cycles, while REM sleep predominates in the later cycles, which means that the most restorative physically and the most restorative cognitively are concentrated in different parts of the night.
NREM stage 3 - slow-wave sleep - is characterised electroencephalographically by delta waves, high-amplitude oscillations at frequencies below four hertz. It is during this stage that the majority of physical restoration occurs: human growth hormone release is concentrated here (approximately seventy to eighty percent of daily growth hormone secretion), protein synthesis and tissue repair are elevated, and the glymphatic system - the brain's waste clearance mechanism, which uses cerebrospinal fluid to flush metabolic byproducts including beta-amyloid and tau proteins associated with neurodegeneration - is most active.
REM sleep - characterised by rapid eye movements, desynchronised EEG activity, and muscle atonia - is the stage most associated with emotional memory processing, procedural memory consolidation, and the integration of new information with existing knowledge. The popular association of REM with dreaming is correct, but the functional significance extends to the offline consolidation of the day's learning and the emotional recalibration of affectively charged experiences.
Consumer wearable devices estimate sleep stages through accelerometry and photoplethysmography (PPG, optical heart rate monitoring). The validation research against polysomnographic gold standard shows consistent patterns: wearables detect the transition from wake to sleep with reasonable accuracy and detect REM sleep with moderate accuracy. They have consistently poor accuracy for slow-wave sleep detection, which is the most clinically significant stage.
A 2023 systematic review and meta-analysis in SLEEP found that sensitivity for slow-wave sleep detection ranged from thirty-seven to eighty-two percent across devices. The practical implication: if your wearable shows low slow-wave sleep, it may be correct or it may be a detection error. If it shows adequate slow-wave sleep, this is more likely to be accurate.
What actually improves slow-wave sleep, with clinical evidence: aerobic exercise; consistent sleep timing; room temperature between sixteen and nineteen degrees Celsius; alcohol avoidance in the three to four hours before sleep (alcohol suppresses slow-wave sleep and fragments REM despite producing faster sleep onset); and magnesium glycinate or magnesium threonate supplementation, which has clinical evidence for improving sleep quality in magnesium-deficient adults.
The clinically validated intervention for significant sleep disorders is cognitive behavioural therapy for insomnia (CBT-I), which has a stronger evidence base for long-term insomnia management than any pharmacological intervention currently available.
Used in isolation as a basis for optimisation decisions, wearable sleep data can produce orthosomnia — anxiety about sleep quality generated by sleep tracking data, which itself disrupts the sleep the tracking was intended to improve.
BY OONA CHANEL
