EEG Internal Volatility: A New Insight for Depression Evaluation via Nonlinear Dynamic Analysis

• Published in: Journal of medical and biological engineering Vol. 45; no. 2; pp. 166 - 176
• Main Authors: Chen, Feifei, Zhao, Lulu, Cai, Zhipeng, Li, Jianqing, Zhang, Hongxing, Liu, Chengyu
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.04.2025; Springer Nature B.V
• Subjects: Biological Techniques; Biomedical and Life Sciences; Biomedical Engineering and Bioengineering; Biomedical Engineering/Biotechnology; Biomedicine; EEG; Electroencephalography; Evaluation; Fractal geometry; Mental depression; Mental disorders; Nonlinear dynamics; Original Article; Regenerative Medicine/Tissue Engineering; Remission; Signs and symptoms; Statistical analysis; Time signals; Variability; Volatility; Signal internal volatility; Spontaneous gamma variability; Depression; GosFME
• ISSN: 1609-0985; 2199-4757
• DOI: 10.1007/s40846-025-00939-2

Purpose Mental disorders could cause variations in neuronal signals. Nonlinear electroencephalogram (EEG) analysis could quantify this variability and provide potential physiological mechanisms. This study analyzed the gamma internal volatility by measuring the nonlinear index offset within time signals. Methods Prefrontal EEG signals from 30 healthy subjects (HC) and 54 depression patients (DP). According to Hamilton Depression Rating Scale (HDRS), DP was divided into 21 depressed patients with remission symptoms (RD, HDRS ≤ 7) and 33 depressed patients with mild symptoms (MD, 7 < HDRS < 17). The gamma band was extracted from each EEG signal. Two novel features, the offset of fuzzy measure entropy (gosFME) and the offset of Higuchi fractal dimension (gosHFD), were proposed. The internal fluctuation of nonlinear features within individual gamma rhythm was assessed. Statistical analysis was performed to investigate the index change trends among three groups. The gosFME stability under different time scales was investigated in this study. Depression recognition models were compared to explore which time scale is most conducive to distinguishing different depressive states. Results The DP individual exhibits more smooth internal fluctuation of the nonlinear indexes within the gamma band, indicating the prefrontal neural variability is suppressed by depression. The feature distribution among different groups under different time scales shows consistency, and HC exhibits more flexible gamma variability. The highest classification accuracy of 94.1% was achieved when the time scale of 7.5s was selected to discriminate three groups. Conclusion Signal internal volatility could provide a novel perspective to evaluate depression, and gamma variability could reflect different EEG characteristics caused by different depression severity levels. The spontaneous gamma variability may be a potentially reliable tool in future depression monitoring.