Stroke causes significant disability, and recovery is often incomplete. In animal models of stroke, robust upper extremity (UE) motor recovery can be elicited if high doses of functional training are given early. In humans, however, the optimal training dose is unknown, because no quantitative dose-response trials have been undertaken in the first months after stroke. This deficiency stems from a lack of measurement instruments that can accurately and easily quantify UE functional training dose and recovery. To address this gap, this proposal will generate two new measurement tools to enable quantitative stroke recovery research. The first tool (Aim 1) will quantify the number of functional movements made during stroke rehabilitation, measuring UE training dose. The second tool (Aim 2) will quantify the abnormality of movements, measuring UE recovery and response to interventions. We will combine wireless motion capture and computational methodologies to create objective, precise, and user-friendly tools. Inertial measurement units, worn by individuals with stroke and healthy controls performing various activities, will capture upper body motion.
In Aim 1, machine learning algorithms will be trained to identify and count functional movements in activities normally practiced during rehabilitation.
In Aim 2, functional principal components analysis will quantify movement impairment and compensation in standardized motions. Validity will be determined by correlating tool outcomes with current gold standards. The proposed study will be conducted at New York University, in collaboration with investigators from the NYU Center for Data Science, Columbia University, and Washington University-St. Louis. Each have complementary expertise in machine learning, functional data analysis, and functional movement identification. This K02 Independent Scientist Award will provide the candidate with skill in advanced motion capture and analytical methodologies needed to study stroke rehabilitation and recovery. The career development plan includes personalized tutorials and coursework combined with longitudinal oversight of data analysis, providing an excellent foundation for launching an independent research career. Ultimately, the developed tools have the potential to immediately impact neurorehabilitation research, facilitating the rigorous dose-response trials so critically needed to change clinical practice and improve stroke outcomes.
The proposed research is relevant to public health, because it will enable the precise and feasible measurement of rehabilitation and recovery after stroke. The generated measurement tools will facilitate important stroke recovery research, which is expected to inform stroke rehabilitation practices and reduce the burden of stroke disability.
