EEG-based major depressive disorder recognition by neural oscillation and asymmetry

Xinyu Liu; Haoran Zhang; Yi Cui; Tong Zhao; Bin Wang; Xiaomeng Xie; Sixiang Liang; Sha Sha; Yuxiang Yan; Xixi Zhao; Ling Zhang

doi:10.3389/fnins.2024.1362111

EEG-based major depressive disorder recognition by neural oscillation and asymmetry

Front Neurosci. 2024 Feb 14:18:1362111. doi: 10.3389/fnins.2024.1362111. eCollection 2024.

Authors

Xinyu Liu^#^{1

2}, Haoran Zhang^#^{1

2}, Yi Cui³, Tong Zhao³, Bin Wang^{1

2}, Xiaomeng Xie^{1

2}, Sixiang Liang^{1

2}, Sha Sha^{1

2}, Yuxiang Yan³, Xixi Zhao^{1

2}, Ling Zhang^{1

2}

Affiliations

¹ Beijing Key Laboratory of Mental Disorders, National Clinical Research Center for Mental Disorders and National Center for Mental Disorders, Beijing Anding Hospital, Capital Medical University, Beijing, China.
² Advanced Innovation Center for Human Brain Protection, Capital Medical University, Beijing, China.
³ Gnosis Healthineer Co. Ltd., Beijing, China.

^# Contributed equally.

Abstract

Background: Major Depressive Disorder (MDD) is a pervasive mental health issue with significant diagnostic challenges. Electroencephalography (EEG) offers a non-invasive window into the neural dynamics associated with MDD, yet the diagnostic efficacy is contingent upon the appropriate selection of EEG features and brain regions.

Methods: In this study, resting-state EEG signals from both eyes-closed and eyes-open conditions were analyzed. We examined band power across various brain regions, assessed the asymmetry of band power between the hemispheres, and integrated these features with clinical characteristics of MDD into a diagnostic regression model.

Results: Regression analysis found significant predictors of MDD to be beta2 (16-24 Hz) power in the Prefrontal Cortex (PFC) with eyes open (B = 20.092, p = 0.011), beta3 (24-40 Hz) power in the Medial Occipital Cortex (MOC) (B = -12.050, p < 0.001), and beta2 power in the Right Medial Frontal Cortex (RMFC) with eyes closed (B = 24.227, p < 0.001). Asymmetries in beta1 (12-16 Hz) power with eyes open (B = 28.047, p = 0.018), and in alpha (8-12 Hz, B = 9.004, p = 0.013) and theta (4-8 Hz, B = -13.582, p = 0.008) with eyes closed were also significant predictors.

Conclusion: The study confirms the potential of multi-region EEG analysis in improving the diagnostic precision for MDD. By including both neurophysiological and clinical data, we present a more robust approach to understanding and identifying this complex disorder.

Limitations: The research is limited by the sample size and the inherent variability in EEG signal interpretation. Future studies with larger cohorts and advanced analytical techniques are warranted to validate and refine these findings.

Keywords: asymmetry; diagnostic regression model; electroencephalography; major depressive disorder; neural oscillation.

Grants and funding

The author(s) declare financial support was received for the research, authorship, and/or publication of this article. This work was funded by the Beijing Municipal Hospital Research and Development Project (PX2021068), the Advanced Innovation Center for Human Brain Protection Project (3500-12020137), the National Natural Science Foundation of China (82201701), and the China Postdoctoral Science Foundation (2022M723299). This work was also supported by the Sci-Tech Innovation 2030 – Major Project of Brain science and brain-inspired intelligence technology (2021ZD0200600).