0603168 - ÚI 2026 NL eng J - Journal Article
Davoodi, Akbar - Holeňa, Martin - Brunovský, M. - Kathpalia, Aditi - Hlinka, Jaroslav - Bareš, M. - Paluš, Milan
Response prediction of antidepressants: Using graph theory tools for brain network connectivity analysis.
Biomedical Signal Processing and Control. Online 28 December (2024), č. článku 107362. ISSN 1746-8094. E-ISSN 1746-8108
R&D Projects: GA ČR(CZ) GF21-14727K; GA MŠMT(CZ) EH22_008/0004643
Institutional support: RVO:67985807
Keywords : Major depressive disorder * Electroencephalography * Antidepressant response * Machine learning * Graph theory * Partial ordering
Impact factor: 4.9, year: 2023 ; AIS: 0.805, rok: 2023
Method of publishing: Open access
Result website:
https://doi.org/10.1016/j.bspc.2024.107362DOI: https://doi.org/10.1016/j.bspc.2024.107362

Depression, particularly in its serious form as major depressive disorder, is rapidly becoming a global health concern and is expected to become a leading cause of disability worldwide. Despite the critical role of antidepressant treatment, its effectiveness varies greatly among patients. Early prediction of the response to specific antidepressant regimens is vital. In this paper, we rely on the predictive efficacy of changes in EEG signals after the first week of treatment, supported by existing evidence in the literature. We use EEG markers and graph theory to create and compare brain networks from two patient visits. Our approach involves the application of a comprehensive set of powerful classifiers. Additionally, we propose a robust model for comparing multiple classifiers on the data using statistical analysis and graph theory, ensuring a trustworthy evaluation.
Permanent Link: https://hdl.handle.net/11104/0360405