Aberrated Multidimensional EEG Characteristics in Patients with Generalized Anxiety Disorder: A Machine-Learning...

Area of Science:

• Neuroscience
• Computational Psychiatry
• Biomedical Engineering

Background:

• Psychiatric disorders, including generalized anxiety disorder (GAD), are increasingly linked to abnormal brain communication.
• Research on electrophysiological disconnectivity in GAD patients is limited, highlighting a need for advanced diagnostic tools.

Purpose of the Study:

• To develop an automated framework for GAD detection using multidimensional electroencephalography (EEG) feature extraction and machine learning.
• To investigate aberrant EEG characteristics in GAD patients compared to healthy controls.

Main Methods:

• Acquired 10-minute resting-state EEG data from 45 GAD patients and 36 healthy controls (HC).
• Extracted multidimensional EEG features: univariate power spectral density (PSD), fuzzy entropy (FE), and multivariate functional connectivity (FC).
• Utilized statistical comparisons to identify aberrant features, fused them, and applied machine learning for classification.

Main Results:

• GAD patients showed increased beta rhythm and decreased alpha1 rhythm in PSD.
• Reduced long-range functional connectivity (FC) between frontal and other brain areas was observed across all frequency bands.
• The developed framework achieved high classification performance: 97.83% accuracy, 97.55% sensitivity, 97.78% specificity, and 97.95% F1 score.

Conclusions:

• Findings support the hypothesis of brain disconnectivity in psychiatric disorders, specifically in GAD.
• The study identified specific spatio-spectral EEG alterations in GAD patients.
• The developed framework shows potential for the automatic diagnosis of generalized anxiety disorder.