Personalized sleep-aiding music intervention for insomnia: A closed-loop neurofeedback approach

Insomnia is a common sleep disorder that severely impacts health and quality of life. Although music therapy is a widely used non-pharmacological treatment for sleep improvement, its lack of personalization limits its effectiveness.

We proposed a novel EEG neurofeedback intervention for sleep-aiding music therapy. A sleep-aiding style transfer model was developed to convert traditional five-element music into a sleep-aiding style. Subsequently, we established a neurofeedback framework linking an EEG parameter, the sleep state, and musical features. Leveraging the EEG alpha/theta power ratio, our system dynamically adjusted the tempo and volume of sleep-aiding five-element music in real time.

EEG neurofeedback intervention significantly improved ISI, PSQI, and sleep latency compared with the non-neurofeedback intervention. Notably, potential biomarkers were identified: reduced prefrontal-parietal functional connectivity in the alpha band during wakefulness (W) and increased functional connectivity from the central region to the central, occipital, and temporal regions in the theta band during light sleep (N1).

These findings highlight the potential of personalized, closed-loop neurofeedback music as an effective non-pharmacological intervention for insomnia. By dynamically adapting music parameters based on real-time EEG, this approach offers a novel method for optimizing sleep regulation and improving treatment outcomes.