Hyperconnectivity and altered dynamic interactions of a nucleus accumbens network in post-stroke depression

Background Post-stroke depression (PSD) is a common complication after stroke. To date, no consistent locus of injury is associated with this complication. Here, we probed network dynamics in four functional circuits tightly linked to major depressive disorder and investigated structural alterations within these networks in PSD. Methods Forty-four participants with recent stroke and 16 healthy volunteers were imaged with 3T structural, diffusion and resting-state functional MRI and completed the Geriatric Depression Scale (GDS). Associations between GDS and functional connectivity were investigated within networks seeded from nucleus accumbens (NAc), amygdala and dorsolateral prefrontal cortex. In addition, the default mode network (DMN) was identified by connectivity with medial prefrontal cortex (mPFC) and posterior cingulate cortex (PCC). Circuits that exhibited altered activity associated with GDS were further investigated by extracting within-network volumetric and microstructural measures from structural images. Results Functional connectivity within the NAc-seeded network and DMN correlated positively with depressive symptoms. Normal anticorrelations between these two networks were absent in patients with PSD. PCC grey matter volume as well as microstructural measures in mPFC and the medial forebrain bundle, a major projection pathway interconnecting the NAc-seeded network and links to mPFC, were associated with GDS scores. Conclusions Depression after stroke is marked by reduced mutual inhibition between functional circuits involving NAc and DMN as well as volumetric and microstructural changes within these networks. Aberrant network dynamics present in patients with PSD are therefore likely to be influenced by secondary, pervasive alterations in grey and white matter, remote from the site of injury.