A posterior-alpha ageing network is differentially associated with antidepressant effects of venlafaxine and rTMS

Major depressive disorder (MDD) is a highly prevalent psychiatric disorder, but chances for remission largely decrease with each failed treatment attempt. It is therefore desirable to assign a given patient to the most promising individual treatment option as early as possible. We used a polygenic score (PGS) informed electroencephalography (EEG) data-driven approach to identify potential predictors for MDD treatment outcome. Post-hoc we conducted exploratory analyses in order to understand the results in depth. First, an EEG independent component analysis produced 54 functional brain networks in a large heterogeneous cohort of psychiatric patients (n = 4,045; 5-84 yrs.). Next, the network that was associated to PGS for antidepressant-response (PRS-AR) in an independent sample (n = 722) was selected: an age-related posterior alpha network that explained >60 % of EEG variance, and was highly stable over recording time. Translational analyses were performed in two other independent datasets to examine if the network was predictive of psychopharmacotherapy (n = 535) and/or repetitive transcranial magnetic stimulation (rTMS) and concomitant psychotherapy (PT; n = 186) outcome. The network predicted remission to venlafaxine (p = 0.015), resulting in a normalized positive predicted value (nPPV) of 138 %, and rTMS + PT - but in opposite direction for women (p = 0.002) relative to men (p = 0.018) - yielding a nPPV of 131 %. Blinded out-of-sample validations for venlafaxine (n = 29) and rTMS + PT (n = 36) confirmed the findings for venlafaxine, while results for rTMS + PT could not be replicated. These data suggest the existence of a relatively stable EEG posterior alpha aging network related to PGS-AR that has potential as MDD treatment predictor.