Discriminating bipolar disorder from major depression using whole-brain functional connectivity: A feature selection analysis with SVM-FoBA algorithm

It is known that both bipolar disorder (BD) and major depressive disorder (MDD) indicate depressive symptoms, especially in the early phase of illness. Therefore, discriminating BD from MDD is a major clinical challenge due to the absence of biomarkers. Feature selection is especially important in neuroimaging applications, yet high feature dimensions, low sample size and model understanding present huge challenges. Here we propose an advanced feature selection algorithm, “SVM-FoBa”, which enables adaptive selection of informative feature subsets from high dimensional brain functional connectives (FC) resulted from fMRI. With 38 significant FCs chosen from 6,670 ones, classification accuracy between BD and MDD was achieved up to 88% with leave-one-out cross validation. Further, by conducting weight analysis, the most discriminative FCs were revealed, which adds our understanding on functional deficits and may serve as potential biomarkers for mood disorders.