Topological Alterations of Functional Brain Networks in Post-Stroke Cognitive Impairment: a Graph-Theoretical Study.

Abstract

Objective: To investigate topological alterations of functional brain networks in patients with post-stroke cognitive impairment (PSCI) using resting-state functional magnetic resonance imaging (rs-fMRI), and to explore the relationship between network organization and post-stroke cognitive performance.

Materials and methods: This study was conducted with a prospective enrollment of 45 patients with ischemic stroke, including 21 patients with PSCI and 24 patients with post-stroke non-cognitive impairment (PSNCI), coupled with the recruitment of 30 age-, sex-, and education-matched healthy controls (HC). All participants underwent brain rs-fMRI and cognitive function assessment. This study further employed comparative analyses to clarify the inter-group differences of global and nodal topological metrics, and the correlation between different brain regions and cognitive scores.

Results: Compared with the HC group, both PSCI and PSNCI groups exhibited significant topological alterations of functional brain networks, including increased characteristic path length (Lp), reduced global efficiency (Eg) and local efficiency (Eloc), and disrupted small-worldness (σ). However, PSCI and PSNCI groups exhibited no significant differences in global network metrics. At the nodal level, PSCI and PSNCI patients showed increased nodal clustering coefficient (NCp) and local efficiency (NLe) in the left medial and paracingulate cortices, left caudate nucleus, and right paracentral lobule. Pearson correlation analysis revealed that Eloc (r = 0.580, P = 0.006), σ (r = 0.513, P = 0.017) and normalized clustering coefficient (γ) (r = 0.581, P = 0.006) were positively correlated with MoCA scores in PSCI group. Clustering coefficient (Cp) (r = -0.492, P = 0.015) was negatively correlated with the MMSE score in PSNCI group.

Conclusions: The global and node topological properties of the brain networks in patients with PSCI have changed. This is manifested as impaired information transmission efficiency and decreased integration ability in the entire brain. The abnormal global properties are related to cognitive dysfunction, providing valuable insights from the imaging perspective for understanding the neural cognitive mechanism of PSCI.