Single-trial EEG-based emotion recognition using kernel Eigen-emotion pattern and adaptive support vector machine.

Area of Science:

• Neuroscience
• Computer Science
• Psychology

Background:

• Single-trial electroencephalography (EEG) offers rapid assessment of emotional states.
• Existing EEG-based emotion recognition methods require improved classification accuracy for valence and arousal.
• Imbalanced datasets pose a challenge for machine learning models in emotion recognition.

Purpose of the Study:

• To introduce a novel feature extraction method, Kernel Eigen-emotion Pattern (KEEP), for EEG signals.
• To develop an adaptive Support Vector Machine (SVM) to handle imbalanced EEG data.
• To enhance the accuracy of emotion recognition using EEG data.

Main Methods:

• Emotion induction using the International Affective Picture System (IAPS) dataset.
• Application of the Kernel Eigen-emotion Pattern (KEEP) for EEG feature extraction.
• Implementation of an adaptive SVM classifier for valence and arousal classification.

Main Results:

• KEEP demonstrated superior classification performance compared to traditional EEG frequency band power features.
• The adaptive SVM significantly improved classification accuracy over standard SVM classifiers.
• Combined KEEP and adaptive SVM achieved average valence and arousal classification rates of 73.42% and 73.57%, with peaks of 80% and 79%.

Conclusions:

• The proposed KEEP feature extraction method and adaptive SVM offer a promising approach for accurate EEG-based emotion recognition.
• These novel methods effectively address the limitations of existing techniques, particularly with imbalanced datasets.
• The findings suggest a significant advancement in the field of affective computing and brain-computer interfaces.