Music-evoked emotion classification using EEG correlation-based information.

Area of Science:

• Neuroscience
• Psychology
• Music Cognition

Background:

• The relationship between music and emotions is a long-standing research area.
• Previous studies often used short music clips and specific tasks, limiting ecological validity.
• Understanding brain responses to music appreciation is crucial for affective computing.

Purpose of the Study:

• To investigate emotional states during music appreciation using electroencephalogram (EEG) signals.
• To develop and validate a method for analyzing EEG features related to emotional responses to music.
• To explore the potential of EEG-based emotion recognition for affective applications.

Main Methods:

• Recorded EEG signals from 15 healthy adults during full music listening sessions.
• Extracted Band Power Change (BPC) and Higher Order Crossing (HOC) features from EEG data.
• Employed a correlation-based feature analysis to identify relevant features across time, frequency, and channel domains.

Main Results:

• A novel correlation-based feature analysis approach was proposed and implemented.
• The combined BPC and HOC features, analyzed via this method, achieved an average accuracy of 67.2% for classifying high and low valence.
• Identified specific brain activity patterns associated with emotional states during music appreciation.

Conclusions:

• The study demonstrates the feasibility of using EEG signals to infer emotional states during music appreciation.
• The proposed feature analysis method effectively identifies relevant EEG markers for emotion classification.
• Further understanding of brain emotional patterns can advance the development of intelligent affective applications.