MRI Before Electroconvulsive Therapy May Help Predict Treatment Success for Depression Interview with:

Dipl.-Psych. R. Redlich Neuroimaging Group Klinik und Poliklinik für Psychiatrie und Psychotherapie Westfaelische Wilhelms-Universitaet Muenster

Dr. Redlich

Dipl.-Psych. R. Redlich
Neuroimaging Group
Klinik und Poliklinik für Psychiatrie und Psychotherapie
Westfaelische Wilhelms-Universitaet Muenster What is the background for this study?

Response: Electroconvulsive therapy (ECT) is one of the most effective treatments for severe depression. The ability to advise psychiatrists and patients accurately regarding the chances of successful ECT is of considerable value, particularly since ECT is a demanding procedure and, despite having relatively few side effects, has a profound impact on patients. Therefore, the present study sought to predict ECT response in a psychiatric sample by using a combination of structural Magnetic Resonance Imaging data and machine-learning techniques. What are the main findings?

Response: In this non-randomized prospective study in patients with severe depression, we achieved a successful prediction of ECT response, with accuracy rates up to 78 % and remarkable sensitivity rates up to 100% using structural MRI obtained before therapy. Furthermore, the analysis of treatment effects on revealed a massive increase in hippocampal volume in patients treated with ECT indicating neuroplastic effects in patients treated with ECT. What should clinicians and patients take away from your report?

Response: We demonstrated specific brain plasticity effects of ECT. Electroconvulsive therapy treatment is associated with strong gray matter volume increases in hippocampal areas which might be one of the efficient causes of ECT. Further, to our knowledge this is the first study successfully predicting electroconvulsive therapy response by using structural brain data obtained before treatment. In future, neuroimaging and multivariate pattern classification techniques are promising tools to predict the therapeutic effectiveness in ECT and show promise as future diagnostic aid, not only for ECT outcome prediction. What recommendations do you have for future research as a result of this study?

Response: Besides replication in independent samples, future studies should aim to transfer a trained classifier to independent sites and samples, which would be a requirement for any potential usefulness in clinical practice. Hence, a successful transfer of trained patterns over several independent scanners and sites would be a crucial next step to demonstrate the generalizability regarding the prediction of  Electroconvulsive therapy response based on MRI data. Thank you for your contribution to the community.


Redlich R, Opel N, Grotegerd D, et al. Prediction of Individual Response to Electroconvulsive Therapy via Machine Learning on Structural Magnetic Resonance Imaging Data. JAMA Psychiatry. Published online May 04, 2016. doi:10.1001/jamapsychiatry.2016.0316.

Note: Content is Not intended as medical advice. Please consult your health care provider regarding your specific medical condition and questions.

More Medical Research Interviews on

[wysija_form id=”5″]



Last Updated on May 4, 2016 by Marie Benz MD FAAD