Increased resting-state brain entropy (BEN) in mild-to-moderate depression was decreased by cognitive-behavioral therapy and neurofeedback.
Abstract
Brain entropy (BEN) reflects temporal complexity of spontaneous neural activity, quantified using sample entropy of fMRI signals. The neural mechanisms underlying nonpharmacological treatments for depression remain incompletely understood. We used BEN to characterize treatment-related changes in neural complexity in mild-to-moderate depression. Resting-state fMRI data from 29 medication-free patients and 20 matched healthy controls (HCs) were analyzed from two public datasets (ds003007, ds002748). Among patients, 14 received cognitive-behavioral therapy (CBT, n = 8) or real-time fMRI neurofeedback (NFB, n = 6), while 15 remained untreated. Baseline BEN differences between patients and HCs were identified using two-sample t-tests, and significant regions served as regions of interest (ROIs). Repeated-measures ANOVA examined longitudinal BEN changes. Compared with HCs, patients showed higher baseline BEN in left dorsomedial prefrontal cortex (DMPFC), left amygdala/putamen, and right hippocampus/parahippocampal cortex. Following treatment, the treated group showed reduced BEN across all ROIs and decreased Beck Depression Inventory scores, whereas untreated patients showed no significant changes. Post-treatment BEN in the left DMPFC and left amygdala/putamen was no longer significantly different from HCs. These findings suggest CBT and NFB are associated with altered neural complexity within prefrontal-limbic network; BEN may represent an exploratory neuroimaging marker for tracking treatment-related neural changes in depression.