Dreem headband performance in Insomnia Disorder: An in-lab comparison to polysomnography

Capturing sleep data historically required complex, expensive, and labour-intensive equipment, with the gold standard being overnight polysomnography (PSG). Newer portable devices, such as the Dreem 3 headband, provide a novel opportunity to collect field-based data and have demonstrated accuracy in healthy sleepers compared to PSG. However, this device's performance has not been assessed in Insomnia Disorder, despite sleep-tracking technologies traditionally performing poorly in disordered sleepers. This study aimed to evaluate the performance of Dreem 3 against PSG on key sleep outcomes.

Thirty-one adults (M_age = 45.9 years, 16 males) with Insomnia Disorder participated in an overnight sleep study wearing Dreem 3 and PSG simultaneously. Sensitivity and specificity were calculated using an epoch-by-epoch analysis. Bland-Altman plots further assessed performance related to sleep stages and continuity variables: total sleep time (TST), sleep efficiency (SE), sleep latency (SL), and wake after sleep onset (WASO).

Dreem 3 showed the highest sensitivity for REM sleep (89.88 %) and the lowest sensitivity for N1 (29.79 %). N3 sensitivity was also notably low (65.65 %). Specificity was >90 % for all stages, except N2 (83.39 %). Dreem accurately summarised SL and WASO but significantly overestimated TST and SE.

The Dreem 3 headband can accurately evaluate and report most sleep staging and continuity variables in individuals with Insomnia Disorder. Users should be mindful of the strengths and limitations of the device when deciding whether the Dreem 3 is suitable for their needs. If applied and interpreted correctly, this device could facilitate large-scale, longitudinal sleep studies and assessment of sleep in Insomnia Disorder.