Dynamic topological changes of the motor network after stroke

Background: Previous studies indicated that motor stroke is characterized by changes in the motor system's resting-state functional connectivity and functional topology. Furthermore, recent reports have shown that time-varying connectivity among motor areas is crucial for motor impairment and its recovery. Yet, it is unknown to what extent the topological organization of the motor network exhibits temporal dynamics that might be clinically relevant for the motor deficit after motor stroke. Objective: We combined a graph-theoretic approach and dynamic functional connectivity MRI to identify dynamic central motor nodes, that is motor areas whose time-varying topological properties are associated with motor impairment, and to characterize the dynamic interactions among regions of the motor system after stroke. Methods: Resting-state fMRI data were collected from a cohort of twenty acute right-hemispheric stroke patients (17 ischemic/3 hemorrhagic) exhibiting NIHSS scores ranging from 1 to 22 (mean=10.05; SD=5.58). Dynamic functional connectivity was estimated using a sliding window approach applied to regions of the motor network. Next, time-varying nodal betweenness centrality, defined as the portion of all shortest paths in the network involving such a node, was computed at each sliding window. Then, dynamic central motor nodes were characterized by correlating the amount of time that a given node exhibited high centrality (i.e., high centrality mode) with the degree of the upper limb impairment. Finally, the time-varying topological interactions within the motor network were investigated by characterizing its shortest paths. Results: A dynamic central motor node was identified in a region located within the ipsilesional primary cortex, namely the anterior wall of the ventral central sulcus (vCS). Specifically, severely impaired patients exhibited shorter stays in high centrality mode than less affected patients. Furthermore, upper limb impairment was associated with a dynamic network profile characterized by low functional connections among such a dynamic central motor node and a set of regions located in the central sulcus and supplementary motor area of the left hemisphere, as well as in the right cerebellum. Conclusions: The current results indicate that acute motor stroke with upper limb impairment affects the time-varying topological properties of functional interactions within the motor network. Therefore, these findings may contribute to understanding motor deficits after stroke.