Neurophysiological predictors of non-response to rTMS in depression

• Main Authors: Martijn Arns, Wilhelmus H. Drinkenburg, Paul B. Fitzgerald, J. Leon Kenemans
• Format: Article
• Language: English
• Published: Elsevier 2012-10-01
• Series: Brain Stimulation
• Subjects: rTMS; EEG; ERP; Personalized medicine; Depression; Alpha frequency
• Online Access: http://www.sciencedirect.com/science/article/pii/S1935861X11001720

Background: The application of rTMS in Depression has been very well investigated over the last few years. However, little is known about predictors of non-response associated with rTMS treatment. Objective: This study examined neurophysiological parameters (EEG and ERP) in 90 depressed patients treated with rTMS and psychotherapy and sought to identify predictors of non-response. Methods: This study is a multi-site open-label study assessing pre-treatment EEG and ERP measures associated with non-response to rTMS treatment. Results: Non-responders were characterized by 1) Increased fronto-central theta EEG power, 2) a slower anterior individual alpha peak frequency, 3) a larger P300 amplitude, and 4) decreased pre-frontal delta and beta cordance. A discriminant analysis yielded a significant model, and subsequent ROC curve demonstrated an area under the curve of 0.814. Conclusions: Several EEG variables demonstrated clear differences between R and NR such as the anterior iAPF, fronto-central Theta and pre-frontal cordance in the Delta and Beta band (representative of increased relative pre-frontal perfusion). The increased P300 amplitude as a predictor for non-response requires further study, since this was the opposite as hypothesized and there were no correlations of this measure with clinical improvement for the whole sample. Combining these biomarkers in a discriminant analysis resulted in a reliable identification of non-responders with low false positive rates. Future studies should prospectively replicate these findings and also further investigate appropriate treatments for the sub-groups of non-responders identified in this study, given that most of these biomarkers have also been found in antidepressant medication studies.