Can motion capture improve task-based fMRI studies of motor function post-stroke? A systematic review

Abstract

AbstractBackgroundVariability in motor recovery after stroke represents a major challenge in its understanding and management. While functional MRI has traditionally been used to address post-stroke motor function in relation to clinical outcome, it lacks details about movement characteristics linked to observed brain activations. Combining fMRI with detailed information of motor function by using motion capture (mocap) might provide clinicians with additional information about mechanisms of motor impairment after stroke.ObjectivesWe aimed to identify fMRI and mocap coupling approaches and to evaluate their potential contribution to the understanding of motor function post-stroke.MethodA systematic literature review was performed according to PRISMA guidelines, on studies using fMRI and mocap in post-stroke individuals. We assessed the internal, external, statistical, and technological validity of each study. Data extraction included study design and analysis procedures used to couple brain activity with movement characteristics.ResultsOf the 404 studies found, 23 were included in the final review. The overall study quality was moderate (0.6/1). The majority of studies focused on the upper limb, using a wide variety of motor tasks. Half of the studies performed a statistical analysis between movement and brain activity by either using kinematics as variables during group or individual level regression or correlation. This permitted establishing a link between motor characteristics and brain activations. Mocap was also integrated without statistical confrontation, to compare results between fMRI and kinematics, or to incorporate real-time movement information to supply external devices, like motor feedback.ConclusionOur review suggests that the simultaneous use of fMRI and Mocap provides new insights compared with conventional fMRI analysis. It allows a better understanding of post-stroke motor function, although being still subject to practical and technological limitations. Further research is needed to optimize and standardize both data measurement and processing procedures.