Harley Glassman Daniel Dwyer Nicia John Denis Laesker Matthew So


Introduction: Emotion regulation is an integral part of mental health, dynamically impacting brain function, as one’s emotions change continuously throughout the day. Impairments in emotion regulation are associated with a range of psychiatric disorders. Although the implications of emotion regulation are crucial to mental health, few studies have examined training emotion regulation strategies with respect to the brain. Thus, we propose an affective brain-computer music interface (aBCMI) prototype for emotion regulation that continuously generates music by estimating emotions from real-time electroencephalography (EEG) signals.

Methods: In this proposal, we describe our prototype consisting of an emotion classifier that detects the expression of emotions from EEG signals, and a music generator that generates music reflective of those emotions. We evaluate our prototype in three separate studies. In study 1, we test the accuracy of the music generator. In study 2, we test the accuracy of the emotion classifier by assessing its correlation with real-time, self-reported emotions. In study 3, the generative music algorithm is used to explore emotion regulation strategies.

Discussion: The proposed BCMI is expected to accurately estimate emotions, provide musical feedback of participants’ emotions, and enable users to intentionally modulate their emotions from musical feedback. This involves capturing the listener’s emotions in real-time using EEG signals, providing the opportunity to regulate one’s emotional state with musical feedback. Thus, in addition to enabling greater neurofeedback training of emotions, our prototype can enhance the understanding of affective computing and emotions with EEG and machine learning.

Conclusions: Clinical applications of this prototype may have a tremendous impact as a neurofeedback tool in music therapy for training emotion regulation. Future research may benefit from using the proposed BCMI as a neurofeedback treatment in mood disorders.

Abstract 1056 | PDF Downloads 406


Research Protocol