Titolo della tesi: Phase-targeted auditory stimulation during human sleep: a home-based paradigm
Phase-targeted auditory stimulation (PTAS) during slow-wave sleep (SWS) is a promising technique for modulating neural rhythms and enhancing memory consolidation. Despite its potential, PTAS has primarily been restricted to complex, supervised laboratory settings, which limits ecological validity and scalability for multi-night interventions. The present thesis sought to develop and validate a comprehensive, end-to-end paradigm for at-home, unsupervised, and multi-night PTAS to address existing methodological and logistical barriers.
The research was organized into three sequential studies. Chapter 2 details the validation of a minimal, home-based acquisition system employing a wearable headband (ZMax) modified with a single fronto-mastoid EEG channel, evaluated over 35 nights. Automated ensemble-based sleep scoring was benchmarked against expert consensus from co-recorded polysomnography (PSG). Chapter 3 presents an unsupervised PTAS protocol tested in 27 healthy adults, utilizing a predictive model-based controller to target four canonical slow oscillation phases (0°, 90°, 180°, 270°) with 50-ms pink-noise stimuli. The accuracy of phase targeting and the phase-dependent specificity of cortical responses were assessed. Chapter 4, the overall system and approach were integrated into a multi-night (five-night), between-subjects pilot study to evaluate scalability, robustness, EEG response stability, and effects on sleep macrostructure and memory.
The ensemble-based automated scoring pipeline, when applied to the single-channel data acquired from the minimal, at-home setup, achieved almost perfect accuracy in sleep staging (accuracy = 88.83%, Cohen’s κ = 0.841) compared to the PSG consensus, establishing the reliability of this combined approach (wearable setup + analysis pipeline) for home-based sleep measurement. Moreover, we demonstrated that the unsupervised predictive controller demonstrated high accuracy in targeting all four phase points and high sleep stage specificity. Physiological analyses confirmed a strong phase dependency of the cortical response, with the up-state identified as the optimal period for maximizing the EEG response. Finally, the multi-night pilot study demonstrated the paradigm's robustness and scalability. The electrophysiological response to stimulation was stable and replicable across intervention nights, with no evidence of habituation. Preliminary findings suggested that the protocol was safe and well-tolerated, with no significant changes in sleep macrostructure, daytime vigilance, and memory performance.
Taken together, the outcomes of this Thesis highlighted the successful development and validation of an end-to-end methodological platform for at-home, unsupervised, and multi-night PTAS. Demonstrating reliable operation with minimal hardware addresses a central methodological gap and enables future large-scale, longitudinal studies in naturalistic settings. The established paradigm bridges laboratory experimental control and ecological validity, facilitating new directions for translational research on sleep neuromodulation.