Growing studies have reported an elevated risk of violence in patients with depression, yet the neurobiological underpinnings remain poorly understood. The present study explored the resting-state electroencephalogram (EEG) features in major depressive disorder (MDD) patients with violent offenses to identify potential neurological markers for violence prediction and intervention.

Twenty-nine MDD patients who committed violent offenses (violent depression [VD] group), 27 MDD patients without violent behaviors (nonviolent depression [NVD] group), and 25 healthy controls (HCs) were included. Resting-state EEGs were recorded for at least 5 min. EEG microstates, functional connectivity (FC), and graph theory metrics were analyzed and compared between groups.

First, the VD group had increased microstate A, more microstates A-B transition, but lower microstates B-D and C-D transition. Second, the VD group exhibited two enhanced functional brain networks compared to NVD and HCs, and three weakened functional brain networks compared to HCs, which were primarily distributed in the frontal and frontoparietal networks. Third, the VD group specifically exhibited reduced nodal efficiency (aNe) in the superior parietal lobe and increased aNe in the middle occipital gyrus.

MDD patients with violent offenses exhibited alterations in EEG microstates, FCs in the frontal lobe and frontoparietal network, and disrupted aNe in specific parietal and occipital lobes. These alternations are closely associated with deficits in emotional regulation, executive function, and inhibitory control, which may subserve as potential neurobiomarkers for violence risk assessment in patients with depression.