Optimizing acoustic stimulation timing for enhanced deep sleep: an analytical approach using polysomnography data

Abstract

Acoustic stimulation, a common neuromodulation technique, has been shown to enhance slow oscillations (SOs) and slow-wave activity (SWA) during deep non-rapid eye movement (NREM) sleep. This enhancement is associated with improved immune function, autonomic regulation, cognitive health, memory consolidation, and restfulness upon awakening. Prior work indicates that stimulus efficacy depends on timing relative to ongoing cortical rhythms, especially SO upstates. Given the close connection between central and peripheral systems during sleep, particularly at the cortico-cardiac level, we aimed to investigate how cortical and cardiac phases shape responses to stimulation and to determine optimal timing strategies. We analyzed overnight polysomnography data from 133 adolescents, during which auditory tones (80 dB, 1000 Hz, 50 ms) were presented at random 15–30 s intervals. Instantaneous phases of EEG SOs (∼0.8 Hz) and heart-rate (HR) low-frequency (LF; 0.04–0.15 Hz) and high frequency (HF; 0.15–0.4 Hz) components were extracted, and three complementary analyses were performed: (1) comparison of tone-evoked responses during upstate vs. downstate phases of cortical SOs and HR components; (2) evaluation of phase-locking strategies, contrasting unimodal (SO-only or HR-only) and combined EEG–HR phase-locking against non–phase locked and no-tone conditions to assess the net effect of phase-locking and tones; and (3) continuous 360° phase analysis to identify optimal timing locations beyond two broad phase categories. Stimulation during HR-LF upstates and HR-HF downstates produced significantly larger SO amplitudes (up to 16% increase), more SO events (up to 22%), stronger SWA (up to 25%), and marked HR oscillatory responses (up to 56%) compared to opposite phases (p < 0.05). Relative to non–phase-locked stimulation, phase-locking to HR components enhanced SO amplitude by ∼22 μV and SWA by 12%, while phase-locking exclusively to SO upstates increased SO amplitude by ∼18 μV and SWA by 19%. Maximal effects were observed when optimal HR phases coincided with SO upstates, yielding 38 μV increases in SO amplitude and 32% increases in SWA. Continuous phase analysis further revealed that responses peaked just before the HR-LF up-peak and HR-HF down-peak. These findings demonstrate that cortical and peripheral oscillatory phases provide robust timing cues for closed-loop auditory stimulation. Combining EEG and HR phase information enables more precise targeting, highlighting brain–heart coupling as a mechanism for refining adaptive stimulation strategies to enhance deep sleep and its associated physiological and cognitive benefits.

Date	22 Dec 2025
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	Mohamad Forouzanfar (Supervisor) & Massimiliano De Zambotti (Co-supervisor)