Reconfiguring brain networks via lightweight dynamic connectivity framework: An EEG-based stress validation

Area of Science:

• Neuroscience
• Computational Neuroscience
• Signal Processing

Background:

• Electroencephalographic (EEG) analysis combined with Artificial Intelligence (AI) and Machine Learning (ML) is increasingly used in stress research.
• Static functional connectivity measures often overlook temporal and directional brain influences crucial for understanding dynamic neural processes.

Purpose of the Study:

• To propose a lightweight dynamic brain connectivity framework for estimating Time-Varying Directed Transfer Function (TV-DTF) in EEG signals.
• To evaluate the discriminative capability of TV-DTF features for stress detection using various ML models.
• To compare the performance of dynamic TV-DTF features against static connectivity measures.

Main Methods:

• Utilized EEG recordings from the 32-channel SAM 40 dataset during mental arithmetic tasks.
• Estimated dynamic effective connectivity using a novel TV-DTF framework across different frequency bands.
• Validated TV-DTF features using Support Vector Machine (SVM), Random Forest (RF), Gradient Boosting (GB), Adaptive Boosting (AdaBoost), and Extreme Gradient Boosting (XGBoost).

Main Results:

• Alpha-band TV-DTF features demonstrated the strongest discriminative power, achieving 89.73% accuracy (3-class) with SVM and 93.69% accuracy (2-class) with XGBoost.
• Dynamic alpha-TV-DTF and beta-TV-DTF features outperformed static measures (absolute power, phase locking) across all tested ML models.
• Feature importance analysis revealed significant frontal-parietal and frontal-occipital information flow, indicating frontal lobe regulatory roles under stress.

Conclusions:

• The lightweight TV-DTF framework robustly captures spatiotemporal brain dynamics and directional influences in EEG signals.
• Dynamic connectivity measures, specifically TV-DTF, offer significant advantages over static methods for stress level classification.
• Findings highlight the utility of TV-DTF in revealing neural mechanisms underlying stress responses and its potential for developing advanced diagnostic tools.