Abnormal large-scale resting-state functional networks in drug-free major depressive disorder.

Major depressive disorder (MDD) is associated with aberrant function and interaction encompassing large parts of cortical, subcortical and limbic regions that always organized into integrative networks implicated in specific tasks. And cumulative evidence suggests that MDD can be understood as a disorder of dysregulated network. Our study used resting-state fMRI and independent component analysis (ICA) to investigate intrinsic functional connectivity (FC) within and between resting-state networks (RSNs) in 27 drug-free MDD patients and 54 healthy control subjects (HCs). Granger causality analysis (GCA) was further used to identify the direct functional interaction between RSNs. We identified sixteen independent components (ICs) as meaningful RSNs. Compared with HCs, the MDD had significantly decreased intra-FC within lateral visual network (VN), parietal network (PN) and posterior default mode network (pDMN), decreased inter-FC between fronto-parietal network (FPN) and subcortical network, between pDMN and anterior DMN, and increased inter-FC between salience network and FPN, and enhanced effective connectivity from VN to PN and to cerebellum network. The functional synchronization of pDMN was negatively correlated with Hamilton Depression Rating Scores. The relatively small number of MDD, the use of medication and the application challenges of GCA on fMRI data may limit the interpretability. These findings indicated that MDD is indeed a disorder of dysregulated network, especially in the functional networks implicated in self-referential activities and emotional visual processing.