Emotion-dependent integration and segregation in EEG functional brain networks revealed by data-driven sparsity...

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Neuroscience

Background:

• Functional connectivity analysis using electroencephalogram (EEG) is crucial for understanding emotional processing networks.
• Traditional methods for constructing brain networks often use empirical thresholds, hindering objective identification of emotion-specific patterns.

Purpose of the Study:

• To develop and validate a data-driven sparsity optimization framework for objectively identifying emotion-discriminative EEG connectivity patterns.
• To explore emotion-specific network topology differences across neutral, sad, and happy states.

Main Methods:

• Constructed functional networks using Pearson Correlation Coefficient across five frequency bands.
• Systematically varied network sparsity and employed ensemble learning to find optimal sparsity for emotion classification.
• Analyzed graph-theoretic properties (global efficiency, local efficiency, clustering coefficient) of optimized networks.

Main Results:

• Achieved high classification accuracies (94.28% in Beta, 94.44% in Gamma bands) using sparsity-optimized networks, outperforming fully connected networks.
• Identified distinct network topologies for different emotional states: happy state showed higher global efficiency, while neutral states exhibited higher local efficiency and clustering.
• Demonstrated that sparsity optimization enhances emotion recognition and reveals emotion-dependent brain network organization.

Conclusions:

• The proposed sparsity optimization framework objectively identifies emotion-discriminative EEG connectivity patterns.
• Sparsity-optimized networks improve emotion recognition and highlight key differences in brain network integration and segregation across emotions.
• This principled framework advances the study of emotion-related brain network organization and emotional regulation mechanisms.