Brain network features based on theta-gamma cross-frequency coupling connections in EEG for emotion recognition

Area of Science:

• Cognitive Neuroscience
• Interpersonal Interaction
• Brain-Computer Interfaces

Background:

• Emotion recognition relies heavily on Electroencephalography (EEG) feature selection.
• Current methods often use simple channel synchronization, neglecting complex brain communication.
• Cross-frequency coupling (CFC) is increasingly recognized for its role in advanced cognitive processes.

Purpose of the Study:

• To explore the efficacy of Cross-frequency coupling (CFC) for updating brain network connections in EEG-based emotion recognition.
• To investigate the impact of combining global, local, and dynamic network features on emotion classification accuracy.
• To provide novel insights into feature selection for advanced cognitive activities at the functional connectivity level.

Main Methods:

• Reconstructing brain networks using Cross-frequency coupling (CFC) to model neural communication.
• Extracting both global and local features from dynamic brain network structures.
• Utilizing continuous time-windows for dynamic feature extraction.

Main Results:

• EEG networks constructed with CFC demonstrated superior performance in emotion classification compared to traditional synchronization methods.
• The integration of global and local brain network features significantly enhanced recognition accuracy.
• Dynamic network features further improved the precision of emotion recognition.

Conclusions:

• Cross-frequency coupling (CFC) offers a more effective approach for modeling brain connectivity in emotion recognition.
• Combining diverse network features and dynamic analysis is crucial for advancing EEG-based emotion recognition.
• This research pioneers a new direction for feature selection in cognitive neuroscience and brain-computer interface applications.