Emotion recognition based on group phase locking value using convolutional neural network.

Area of Science:

• Neuroscience
• Computer Science
• Human-Computer Interaction

Background:

• Electroencephalography (EEG)-based emotion recognition is crucial for advanced human-computer interactions.
• Neuromarketing utilizes group EEG to analyze collective emotional states, but individual EEG analysis limits scalability.
• Existing methods struggle with multi-user emotion estimation, necessitating improved data processing techniques.

Purpose of the Study:

• To identify a data processing method that enhances the efficiency of emotion recognition.
• To compare the accuracy of emotion recognition using individual versus group EEG data.
• To explore the potential of group EEG analysis for understanding collective emotional responses.

Main Methods:

• Utilized the DEAP dataset, containing EEG signals from 32 participants watching emotionally diverse videos.
• Developed and applied a convolutional neural network (CNN) model for emotion recognition.
• Analyzed differences in phase locking value (PLV) across EEG frequency bands correlating with distinct emotional states.

Main Results:

• Achieved up to 85% emotion recognition accuracy using group EEG data with the proposed CNN model.
• Demonstrated that group EEG data processing significantly improves emotion recognition efficiency.
• Identified distinct phase locking value (PLV) patterns in different EEG frequency bands associated with various emotions.

Conclusions:

• Group EEG data analysis effectively enhances emotion recognition accuracy and efficiency.
• The proposed model and methodology offer a viable approach for estimating group human emotional states.
• Findings contribute to advancing neuromarketing and human-computer interaction research through multi-user emotion analysis.