EEG Data Predictive of Response to TMS for PTSD, Major Depressive Disorder

A machine learning algorithm trained on electroencephalographic coherence can predict transcranial magnetic stimulation treatment outcome for patients with major depressive disorder and posttraumatic stress disorder.

Electroencephalography (EEG) data may be used to predict clinical response to transcranial magnetic stimulation (TMS) in patients with comorbid major depressive disorder and posttraumatic stress disorder (PTSD), according to study data published in the Journal of Affective Disorders.

Investigators conducted a prospective, unblinded trial of 5 Hz TMS at the Veterans Affairs Medical Center and Butler Hospital, both located in Providence, Rhode Island. All participants met Diagnostic and Statistical Manual of Mental Disorders, 5th Edition, criteria for major depressive disorder and PTSD. Resting-state 8-channel EEG was collected before and after TMS. TMS was administered at 5 Hz to the left dorsolateral prefrontal cortex for a maximum of 40 daily sessions. PTSD and major depressive disorder severity were measured with the checklist for Diagnostic and Statistical Manual of Mental Disorders, 5th Edition, PTSD and the Inventory of Depressive Symptomatology, Self-Report, respectively, at baseline and within 72 hours of the final TMS treatment. A 50% reduction in clinical scale scores from baseline to final assessment was classified as clinical response. Area under Receiver Operating Characteristic curves (AUC), sensitivity, and specificity were calculated per this definition of clinical response. Least absolute shrinkage and selection operator regression and Support Vector Machine were used to train a linear classifier to assign EEG observations as pre- or post-TMS.

Related Articles

The final analysis used data from 29 individuals. TMS was associated with significant improvement in both PTSD and depression severity (all P <.05). Posttreatment clinical response was achieved on the depression and PTSD self-report scales in 13/29 and 12/29 patients, respectively. Using regression, TMS treatment outcome could be predicted from pretreatment EEG Alpha band coherence. The model using Alpha coherence values predicted clinical response with an AUC of 0.83 (95% CI, 0.69-0.94) for depression and 0.71 (95% CI, 0.54-0.87) for PTSD. Theta coherence was also a significant predictor of improvement in depressive symptoms, with an AUC of 0.69 (95% CI, 0.51-0.86). The optimal sensitivity for Alpha-trained regression in predicting clinical response was 100% for depression and 94% for PTSD. Specificity was lower, with 46% for depression and 50% for PTSD.

The Support Vector Machine was able to accurately classify EEG recordings as pre- or posttreatment above chance (>50%) for all frequency bands (all P <.001). Maximum classifier performance for Alpha, Beta, Theta, and Delta bands was 75.40%±1.47%, 77.44%±1.44%, 73.81%±1.47%, and 78.57%±1.42%, respectively.

Limitations of this work include lack of sham condition and the small size of the cohort.

These data suggest that pretreatment EEG coherence may be used to predict TMS response in patients with depression and PTSD.

“Given the affordability and accessibility of EEG and application of data-driven approaches, this approach holds promise in designing treatment regimens, devices and measurements to make screening and personalizing treatment possible in the office setting,” investigators concluded.

See source for complete funding and disclosure information.


Zandvakili A, Philip NS, Jones SR, Tyrka AR, Greenberg BD, Carpenter LL. Use of machine learning in predicting clinical response to transcranial magnetic stimulation in comorbid posttraumatic stress disorder and major depression: a resting state electroencephalography study. J Affect Disord. 2019;252:47-54.