Disrupted intrinsic functional brain topology in patients with major depressive disorder

Aberrant topological organization of whole-brain networks has been inconsistently reported in studies of patients with major depressive disorder (MDD), reflecting limited sample sizes. To address this issue, we utilized a big data sample of MDD patients from the REST-meta-MDD Project, including 821 MDD patients and 765 normal controls (NCs) from 16 sites. Using the Dosenbach 160 node atlas, we examined whole-brain functional networks and extracted topological features (e.g., global and local efficiency, nodal efficiency, and degree) using graph theory-based methods. Linear mixed-effect models were used for group comparisons to control for site variability; robustness of results was confirmed (e.g., multiple topological parameters, different node definitions, and several head motion control strategies were applied). We found decreased global and local efficiency in patients with MDD compared to NCs. At the nodal level, patients with MDD were characterized by decreased nodal degrees in the somatomotor network (SMN), dorsal attention network (DAN) and visual network (VN) and decreased nodal efficiency in the default mode network (DMN), SMN, DAN, and VN. These topological differences were mostly driven by recurrent MDD patients, rather than first-episode drug naive (FEDN) patients with MDD. In this highly powered multisite study, we observed disrupted topological architecture of functional brain networks in MDD, suggesting both locally and globally decreased efficiency in brain networks.