Dynamic connectivity predicts acute motor impairment and recovery post-stroke

Objective Thorough assessment of cerebral dysfunction after acute brain lesions is paramount to optimize predicting short- and long-term clinical outcomes. The potential of dynamic resting-state connectivity for prognosticating motor recovery has not been explored so far. Methods We built random forest classifier-based prediction models of acute upper limb motor impairment and recovery after stroke. Predictions were based on structural and resting-state fMRI data from 54 ischemic stroke patients scanned within the first days of symptom onset. Functional connectivity was estimated using both a static and dynamic approach. Individual motor performance was phenotyped in the acute phase and six months later. Results A model based on the time spent in specific dynamic connectivity configurations achieved the best discrimination between patients with and without motor impairments (out-of-sample area under the curve and 95%-confidence interval (AUC{+/-}95%-CI): 0.67{+/-}0.01). In contrast, patients with moderate-to-severe impairments could be differentiated from patients with mild deficits using a model based on the variability of dynamic connectivity (AUC{+/-}95%-CI: 0.83{+/-}0.01). Here, the variability of the connectivity between ipsilesional sensorimotor cortex and putamen discriminated the most between patients. Finally, motor recovery was best predicted by the time spent in specific connectivity configurations (AUC{+/-}95%-CI: 0.89{+/-}0.01) in combination with the initial motor impairment. Here, better recovery was linked to a shorter time spent in a functionally integrated network configuration in the acute phase post-stroke. Interpretation Dynamic connectivity-derived parameters constitute potent predictors of acute motor impairment and post-stroke recovery, which in the future might inform personalized therapy regimens to promote recovery from acute stroke.