REM Sleep Disruption and Its Potential Uses in Depression Management

This was a single-center, randomized, parallel-group superiority study conducted in adults aged 18-65 years with clinically significant depressive symptoms. A total of 208 participants were randomized in a 1:1 ratio to an active intervention group or a monitoring-only control group. Participants used an iPhone application together with an Apple Watch companion application during sleep. The system continuously processed wrist accelerometry and photoplethysmography data and classified consecutive sleep epochs as wake, REM sleep, or non-REM sleep. In the active group, detection of probable REM sleep triggered brief vibrotactile stimulation at the wrist. The sleep-stage estimate was reassessed after each stimulation event and additional stimulation could be delivered according to a predefined escalation algorithm if REM classification persisted. The control group underwent the same sleep monitoring but did not receive REM-contingent night-time stimulation. A morning Smart Alarm haptic function was available in both groups. The main clinical objective was to determine whether REM-contingent haptic stimulation was associated with a reduction in depressive symptom severity compared with monitoring alone. The primary outcome was the baseline-adjusted between-group difference in final Montgomery-Åsberg Depression Rating Scale (MADRS) score. The Beck Depression Inventory-II (BDI-II) served as a secondary clinical outcome and was also assessed repeatedly during the intervention period. Additional analyses evaluated wearable-derived sleep architecture, REM-related measures, intervention delivery, adherence, and tolerability. Outcome assessors were blinded to treatment allocation.