Effect of altering movement metrics identified by predictive models on stroke recovery outcomes

Project: Research


  • Yazan Abdel Majeed (PI)


Patient recovery after stroke is evaluated using a host of clinical assessments, yet there is no clear line from the patients' movement performance to their functioning on these clinical tests. More effective therapy can be designed and customized for patients if we know the relationship between movement metrics and the standard clinical tests. Not every metric that changes is relevant when explaining clinical results, and identifying those important for prediction is a key advancement in our understanding of neurological injury and interventions that can improve recovery. We examined the results of our framework predicting clinical changes in a three-week bimanual reaching study. We used patient movement and demographic information to predict changes in two commonly-used clinical outcomes, as well as ranked the predictive features to identify potential targets for therapy. Our models identified movement speed as a sound target for therapy. Here, we first identify the most effective method to shape movement speed. Next, We determine how interventions designed to target movement speed affect clinical outcomes.
Award amount$26,844.00
Award date07/01/2018
Program typePredoctoral Fellowship
Award ID18PRE34080333
Effective start/end date07/01/201806/30/2019