Global Brain Connectivity Alterations After Stroke: A Microstate-Based Community Analysis

Quantifying the dynamic changes in brain function after stroke is essential for understanding its neural mechanisms. Stroke disrupts global brain connectivity and alters the spatiotemporal properties of brain network interactions. However, most research focuses on time-frequency methods to study dynamic connections within localized brain regions, leaving a gap in understanding the global topological complexity of brain networks after stroke. This study integrates resting-state EEG microstate analysis to build a microstate brain network and then propose a new community detection model based on microstate brain network (CDM-MBN) with the Louvain-PSI algorithm focusing on quasi-stability and network topology. We validated its effectiveness through comparison, and applied it to explore how the brain processes and integrates information after stroke. Simulation results showed that the Louvain-PSI algorithm outperforms other methods in detecting communities in complex networks, making it more suitable for brain network analysis. Further experimental results showed that the stroke patient group had stronger connectivity in the occipital lobe but weaker connectivity in the frontal lobe, while the elderly group showed similar connectivity to the young group. Differences in community structure were also observed, with the patient group displaying a more chaotic network, especially in the occipital lobe. In the healthy population, the elderly group had fewer communities than the young group. The community division results indicate that age-related changes speed up brain damage and hinder recovery in stroke patients. Abnormal brain activity patterns reflect the stroke's impact on brain function. The correlation with rehabilitation metrics underscores the clinical importance of these findings, offering a clearer understanding of brain function in stroke patients.