Leveraging Graph Neural Networks to Decode Music-Induced Emotions from EEG

Decoding emotion processing from electroencephalogram (EEG) signals is challenging yet promising. We investigate the use of Graph Neural Networks (GNNs) for interpreting EEG data. Specifically, we explore the neural correlates of music perception and enjoyment by classifying emotional states from EEG recordings of participants listening to music. Our focus lies in distinguishing neural activity patterns associated with enjoyment and disenjoyment during music exposure. Using the Naturalistic Music EEG Dataset-Tempo (NMED-T), we demonstrate that GNNs outperform traditional Convolutional Neural Networks (CNNs), achieving 87% accuracy in classifying previously unseen EEG data. These findings underscore the potential of GNNs to decode intricate brain signals related to emotion processing. In the future, GNNs could also enable the identification of key brain regions involved in processing music-induced emotions, providing deeper insights into the neural basis of affective responses to music.