Performance of three consumer sleep-tracking devices compared with Actigraphy and Polysomnography

To compare the performance of three consumer wearables (Apple Watch Series 7, Fitbit Charge 5, Polar Vantage M2) against home polysomnography (PSG) and actigraphy using complementary metrics of accuracy, agreement, and reliability.

This prospective observational study was conducted at King Abdulaziz University Hospital on healthy adults aged 18-45 years. Participants were allocated to one wearable (Fitbit n = 20; Apple Watch n = 19; Polar n = 15). PSG performed during a home visit and a research-grade actigraph was worn concurrently (7 days) for ambulatory reference. Primary outcomes were wearable-PSG differences for time in bed (TIB), total sleep time (TST), and stage durations (Rapid Eye Movement (REM), light, deep). Agreement was assessed using intraclass correlation coefficients (ICCs) and Bland-Altman analyses.

Across devices, agreement with PSG was generally weak with wide LOA. Compared with PSG, Fitbit overestimated light sleep (+57.7 ± 80.5 min, P = .009). Apple Watch underestimated light sleep (-73.8 ± 61.6 min, P < .001) and overestimated REM (+30.3 ± 36.3 min, P = .003). Polar showed a positive REM bias (+32.3 ± 43.6 min, P = .026) without significant differences for other parameters. Overall ranking favored Apple Watch, followed by Fitbit and Polar.

Apple Watch Series 7, Fitbit Charge 5, and Polar Vantage M2 demonstrated parameter-specific bias and limited agreement with PSG and actigraphy. These devices may be appropriate for longitudinal self-monitoring rather than clinical-grade assessment of sleep architecture.