A novel consciousness emotion recognition method using ERP components and MMSE

Area of Science:

• Neuroscience
• Computational Psychology
• Signal Processing

Background:

• Electroencephalogram (EEG) based emotion recognition often suffers from low accuracy due to the complexity of EEG signals.
• Traditional methods primarily use time and frequency domain features, neglecting valuable event-related potential (ERP) components.
• Manual identification of ERP components is time-consuming and requires expert knowledge.

Purpose of the Study:

• To develop an automated method for emotion recognition by integrating ERP components and nonlinear features from EEG signals.
• To enhance the accuracy of classifying consciousness and unconsciousness emotions.
• To provide a novel approach for analyzing nonlinear time series in emotion recognition.

Main Methods:

• Automated identification and extraction of ERP components (N200, P300, N300) using the shapelet technique.
• Variational mode decomposition and wavelet packet decomposition to process EEG signals and extract modified multi-scale sample entropy (MMSE) features.
• Fusion of extracted ERP components and MMSE features into a new vector for classification using a random forest model.

Main Results:

• The proposed method achieved high average classification accuracies: 94.42% for happiness, 94.88% for horror, and 94.95% for anger.
• Demonstrated the effectiveness of fusing ERP components with nonlinear features for emotion recognition.
• Successfully classified consciousness and unconsciousness emotions with improved accuracy.

Conclusions:

• The fusion of event-related potential (ERP) components and modified multi-scale sample entropy (MMSE) offers a more effective approach for consciousness and unconsciousness emotion recognition.
• This study presents a novel, automated method for EEG-based emotion recognition, reducing reliance on manual analysis.
• The findings open new research avenues in nonlinear time series analysis for affective computing.