التفاصيل البيبلوغرافية
| العنوان: |
Determination of electrophysiological biomarkers to diagnose depression from alpha band using machine learning methods. |
| المؤلفون: |
Nassehi, Farhad, Eken, Aykut, Par, Asuhan, Yetkin, Sinan, Eroğul, Osman |
| المصدر: |
Anatomy: International Journal of Experimental & Clinical Anatomy; 2023 Supplement, Vol. 17, p14-14, 1/2p |
| مصطلحات موضوعية: |
MACHINE learning, EYE tracking, ELECTROPHYSIOLOGY, FEATURE selection, SUPPORT vector machines, K-nearest neighbor classification |
| الشركة/الكيان: |
WORLD Health Organization |
| مستخلص: |
Objective: Depression, a prevalent psychiatric disorder, affects millions worldwide according to the World Health Organization. Currently, depression diagnosis relies on clinical questionnaires interpreted by experts. Neurophysiological imaging techniques, like electroencephalography (EEG), are increasingly used for diagnosing and studying depression. This study aims to identify neurophysiological biomarkers for depression diagnosis using Alpha band spectral features in resting-state EEG signals. Methods: This study included 22 diagnosed depression patients and 25 age-gender matched healthy individuals. During a 5-minute resting EEG session with eyes closed, EEG signals were recorded from 19 electrodes based on the 10-20 system. Simultaneously, Electrooculography (EOG) signals detected eye movements and removed their influence through regression analysis. Using a signal-slicing approach, data augmentation resulted in 132 epochs from patients and 150 epochs from the control group. EEG signals were bandpass filtered in the 0.5-64 Hz range and cleaned from eye movement artifacts using EOG signals and regression analysis. Power Spectral Density (PSD) was calculated using the Welch method, generating 76 features in the 8-13 Hz Alpha band, such as mean, total power, maximum, and relative alpha power. The obtained features were used as input for K-Nearest Neighbors (KNN), Support Vector Machines (SVM), AdaBoost, and Multilayer Perceptron (MLP) classifiers. Results: AdaBoost showed the highest performance with 95% Area Under the Curve (AUC) and 87.25% accuracy. Using the ReliefF feature selection method, 28 relevant features were selected. When provided as input, AdaBoost achieved the best performance with 96% AUC and 90.41% accuracy. The selected 28 features primarily consisted of mean and total power values from different electrodes, consistent with findings from existing statistical studies. Conclusion: These results suggest that the Alpha band's spectral mean and total power can serve as neurophysiological biomarkers for depression diagnosis. [ABSTRACT FROM AUTHOR] |
|
Copyright of Anatomy: International Journal of Experimental & Clinical Anatomy is the property of Deomed Publishing and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) |
| قاعدة البيانات: |
Complementary Index |
