Non-contact radar simultaneously estimates sleep duration and apnea severity with clinical-grade accuracy
Millimeter-wave radar (MWR) can simultaneously estimate both sleep-disordered breathing severity and total sleep time without any body contact, achieving clinically acceptable agreement with gold-standard polysomnography (PSG), according to a new study in Sleep Biological Rhythms.
The finding addresses a growing clinical challenge: patients with comorbid insomnia and sleep apnea (COMISA) are difficult to assess with conventional tools. PSG is cumbersome and expensive, while wearable devices require skin contact that many patients — particularly those with insomnia — find disruptive or intolerable.
What they found
Hoshi and colleagues from Nihon University, Kyoto University, and affiliated institutions tested MWR against overnight PSG in 62 adults. The radar recorded respiratory movements and body motion through the air, producing two derived metrics: a respiratory event index (REI) and an estimated sleep duration (ESD).
The key methodological insight was that post-movement data exclusion windows needed to be optimized differently for each metric:
- For sleep-disordered breathing (REI vs. AHI): A 3-minute exclusion window after large body movements produced near-zero bias (mean difference +0.1 events/hour) with 95% limits of agreement from -8.2 to +8.4 events/hour. The mean REI (18.9 events/h) closely matched PSG-derived AHI (19.0 events/h).
- For sleep duration (ESD vs. total sleep time): A 1-minute exclusion window yielded the best agreement, with a mean bias of -26 minutes and 95% limits of agreement from -72 to +20 minutes. The radar slightly underestimated sleep duration (396 vs. 422 minutes), but the concordance was clinically acceptable.
Why it matters
COMISA affects an estimated 30-40% of sleep clinic patients, but standard diagnostic pathways often miss one of the two conditions. Patients with chronic insomnia may refuse or poorly tolerate CPAP titration studies, while sleep apnea screening tools overlook sleep continuity problems. A non-contact device that captures both dimensions simultaneously in the patient’s home could transform screening and monitoring workflows.
The radar requires no electrodes, straps, or wearable sensors. Patients simply sleep in a room with the device positioned nearby, reducing the first-night effect and improving ecological validity compared to laboratory PSG.
Limits
The study was relatively small (n=62) and conducted as a single-night laboratory comparison. Generalizability to multi-night home monitoring — where body position, room environment, and movement patterns differ — requires further validation. The post-movement exclusion heuristic also needs testing against varied sleep architectures, including in patients with periodic limb movements or parasomnias.
Bottom line
Millimeter-wave radar can simultaneously estimate sleep apnea severity and sleep duration with clinically relevant accuracy, using metric-specific post-movement exclusion windows. The technology offers a low-burden, contact-free pathway for screening and monitoring COMISA and other sleep disorders in home settings.
Source: Hoshi M, Kozu Y, Okumura S, et al. Simultaneous estimation of sleep duration and sleep apnea-hypopnea using millimeter-wave radar. Sleep Biol Rhythms. 2026;24(3):367-376. DOI: 10.1007/s41105-026-00644-w