Prediction of rTMS Treatment Response in Depression Using a Frequency-Based EEG Biomarker

Major depressive disorder (MDD) is a leading cause of disability worldwide, and a significant proportion of patients do not respond to standard pharmacological and psychotherapeutic treatments. Repetitive transcranial magnetic stimulation (rTMS) provides a non-invasive neuromodulatory therapy for treatment-resistant depression (TRD); however, the absence of reliable biomarkers limits its precision. In this study, resting-state EEG was recorded prior to treatment from 80 MDD patients undergoing a 20-session rTMS protocol. Spectral features across canonical frequency bands were extracted from 19 scalp electrodes, and a regularized feature selection technique identified a single high-frequency component at electrode F4 as the primary discriminator between responders (R) and non-responders (NR). This feature was normalized relative to an empirically derived reference cohort of optimal R to create a differential index. Combined with Support Vector Machine (SVM) and k-Nearest Neighbor (k-NN) classifiers under leave-one-out cross-validation, this index separated R from NR with 100% accuracy. These findings indicate that a succinct, interpretable spectral marker can predict individual rTMS outcomes and may inform personalized neuromodulation strategies in future studies.