Transient neural oscillations reveal new biomarkers for therapeutic transcranial magnetic stimulation for comorbid MDD and PTSD

Abstract

Abstract Repetitive transcranial magnetic stimulation (TMS) is a therapy for major depressive disorder (MDD) and post-traumatic stress disorder (PTSD). TMS effectiveness varies across patients and optimal protocols are unknown. Electroencephalography (EEG) examines brain dynamics associated with TMS often in highly averaged data (Zandvakili et al, 2019). In non-averaged signals, EEG oscillations may emerge as high power transients, coined “events” (Jones, 2016; Shin et al, 2017). We applied methods to analyze event features to describe biomarkers of TMS and their relationship to clinical outcomes. We collected resting 8-channel EEG before and after therapeutic TMS for 23 patients with comorbid MDD and PTSD targeting the left dorsolateral prefrontal cortex. Using our open-source toolbox (https://github.com/jonescompneurolab/SpectralEvents), we tested if resting state oscillations can be quantified as events in non-averaged data using 2-second windows. We assessed if event characteristics – event rate, power, duration, frequency span – changed after TMS or correlated with symptom improvement. Averaged omnibus power spectral density revealed peaks in frequency bands including delta/theta (1-6 Hz), alpha (7-14 Hz), and beta (15-29 Hz). Transient high-power events emerged from non-averaged data. Overall, frontal delta/theta and alpha event number and medial delta/theta event duration decreased after TMS (all p < 0.05). Pretreatment event characteristics correlated with subsequent clinical response. Frontal delta/theta event frequency span negatively correlated with MDD & PTSD improvement; medial alpha event power and duration positively correlated with MDD improvement, whereas medial alpha event rate and duration positively correlated with PTSD improvement; and frontal beta duration negatively correlated with PTSD improvement (all r2 > 0.15). Novel methods to characterize spectral events in non-averaged signals identified periods of high power activity in resting EEG data and revealed changes in event features potentially predictive of TMS response. EEG-recorded events ultimately may provide new biomarkers of response and a better mechanistic understanding of TMS. Keywords: Transcranial magnetic stimulation, Resting state EEG, Major depressive disorder, Post-traumatic stress disorder