Network dynamics in post-stroke cognitive impairment: insights from effective connectivity analysis

Abstract This study investigates post-stroke cognitive impairment (PSCI) by utilizing spectral dynamic causal modeling (spDCM) to examine changes in effective connectivity (EC) within the default mode network, executive control network, dorsal attention network, and salience network. Forty-one PSCI patients and 41 demographically matched healthy controls underwent 3D-T1WI and resting-state functional magnetic resonance imaging on 3.0T MRI. The study compared the EC among eight representative regions of interest using spDCM and analyzed the correlation between the altered EC and cognitive tests. Subgroup analysis was conducted based on lesion location. The study found a significant increase in EC in the PSCI group, specifically from the posterior cingulate cortex (PCC) to the left and right dorsolateral prefrontal cortex (L_DLPFC and R_DLPFC, respectively), and from the right insula to R_DLPFC (p = 0.01, dNBS-corrected). The study found that these changes were significantly negatively correlated with cognitive scores. Subgroup analysis also showed increased executive control in PSCI patients with left anterior circulation lesions. Validation through low-order functional connectivity analysis revealed abnormalities dominated by large-scale intra- and inter-network functional connectivity increases in patients with PSCI. The study suggests that there is an increase in effective connectivity between networks, particularly within the triple network model. The findings implicate the PCC and R_DLPFC in the pathophysiology of PSCI, which sheds light on its cognitive implications. This study emphasizes the importance of understanding network changes in PSCI from various perspectives and enhances our understanding of the neural mechanisms underlying PSCI, establishing a foundation for future research.