Network‐level disconnectivity tracks poststroke depressive symptom improvement

Aims Poststroke depression is the most common psychological condition following a stroke and can severely impact outcome. However, the neurobiological biomarkers associated with recovery of poststroke depression remain unclear. In this longitudinal observational study, we set out to investigate the effect of lesion‐induced focal damage and network‐level disconnectivity, and longitudinal structural connectivity on poststroke depression and its recovery. Methods Sixty‐two participants underwent a psychological assessment utilizing the Montgomery–Asberg Depression Rating Scale (MADRS‐SIGMA) at 3 and 12 months post first‐ever stroke. First, we evaluated the relationship between lesions and MADRS scores at the 3 months using voxel‐based lesion‐symptom mapping and lesion network mapping. Next, we assessed the longitudinal relationship between structural connectivity and MADRS scores (change from 3 to 12 months) using multi‐shell diffusion‐weighted MRI data ( n = 29) collected at both timepoints. Results Patient lesions mapped to a connected brain network centered on the left dorsolateral prefrontal cortex (P FDR <0.05). Longitudinal diffusion MRI analysis revealed that increased quantitative anisotropy in the identified structural network was associated with improved depressive symptoms longitudinally ( P = 0.014), independent of demographic and clinical covariates. Significant associations between the lesions and focal brain structures were not identified. Conclusions Depressive symptoms and their improvement mapped to a specific structural brain network. Knowledge of the integrity of this network could prove useful for predicting poststroke depression. Clinical Trial Registration START‐PrePARE Australian New Zealand Clinical Trials, www.anzctr.org.au , Registry number: ACTRN12610000987066. EXTEND ClinicalTrial.gov identifier: NCT00887328.