Abstract

Abstract Introduction During the wake to sleep transition, the EEG exhibits a reduction in the power in the beta (15-30 Hz) band and an increase in the power of the theta (4-8 Hz) band. In previous publications we reported that the log-ratio “ρ=10×log(β/θ)” quasi-monotonically decreases by an order of magnitude as sleep initiates. Methods We developed a closed-loop, real-time system that processes a single EEG signal (FPz-M2) to modulate the volume of (pink-noise) sound according to “ρ=10×log(β/θ)”. The volume was calibrated such that it progressively decreases as sleep initiates. The EEG was acquired using the Philips AliceTM PSG station connected to a laptop where the algorithm was implemented. The sound was played through a wearable headband connected to the laptop’s audio-output. The algorithm processes 6-second EEG windows to estimate: 1) a signal quality index, 2) the average “ρ”, and 3) the sleep stage using a deep-learning stager. The volume changes with “ρ” according to a sigmoidal model. From the time where N2 or N3 sleep has been continuously detected for 3 minutes, the volume decays to zero in an exponential fashion. Seven subjects without any sleep disorder diagnosis (3F/4M; 33.6 ± 8.7 years old) participated in a home-based trial and recorded 5 sleep sessions. The first familiarization session was followed by randomized 2-session blocks: Block 1: closed-loop volume modulation (active), and Block 2 open-loop (sham) constant volume decrease. Results A 2.2-minute decrease (p=0.1) in average sleep latency was found in the active condition (11.6 ± 5.0m) w.r.t. the sham condition (13.8 ± 6.1m). A 5.2-minute decrease (p=0.08) in average N3 latency was found in the active condition (29.3 ± 10.4m) w.r.t. the sham condition (34.5 ± 13.6m). The log-ratio decreased significantly faster (p<0.05) and more monotonously in the active condition suggesting a faster sleep-deepening due to the sound modulation. Conclusion Closed-loop modulation of the volume of pink-noise based on the EEG’s β/θ ratio may promote a faster sleep onset and a faster transition into deeper NREM sleep. The statistically trending results reported in this research grant further experimental validation with a larger number of subjects. Support Philips Sleep and Respiratory Care

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call