EEG-Based Emotion Recognition Using Spatial-Temporal Graph-Aware Network with Channel Selection.

Area of Science:

• Neuroscience
• Artificial Intelligence
• Biomedical Engineering

Background:

• Electroencephalogram (EEG)-based emotion recognition is crucial for human-computer interaction and brain-computer interfaces.
• High dimensionality and redundancy in EEG data lead to computational challenges and performance degradation.
• Current channel selection methods lack frequency-specific adaptability and inter-channel modeling, causing information loss.

Purpose of the Study:

• To develop a novel framework for EEG-based emotion recognition that enhances both performance and efficiency.
• To address the limitations of existing channel selection methods in EEG data analysis.
• To improve the accuracy and reduce the computational cost of emotion recognition systems.

Main Methods:

• A novel framework combining discriminative channel selection with hierarchical spatial-temporal modeling.
• Preprocessing using wavelet coherence and mutual information for adaptive, multi-frequency channel selection.
• Utilizing a Spatial Temporal Graph-aware Network (STG-Net) for spatial and temporal feature extraction and fusion.

Main Results:

• The proposed framework achieved superior recognition accuracy compared to state-of-the-art methods.
• The method demonstrated enhanced model efficiency through effective dimensionality reduction.
• Adaptive channel selection and advanced modeling improved the capture of emotional state dynamics.

Conclusions:

• The developed framework offers a significant advancement in EEG-based emotion recognition.
• The combination of discriminative channel selection and STG-Net effectively addresses EEG data complexity.
• This approach holds promise for more efficient and accurate intelligent human-computer interaction systems.