Predicting Improved Daily Use of the More Affected Arm Poststroke Following Constraint-Induced Movement Therapy.

摘要: Background Constraint-induced movement therapy (CI therapy) produces, on average, large and clinically meaningful improvements in the daily use of a more affected upper extremity individuals with hemiparesis. However, individual responses vary widely. Objective The study objective was to investigate extent which characteristics before treatment predict improved arm following CI therapy. Design This retrospective analysis 47 people who had chronic (> 6 months) mild moderate hemiparesis were consecutively enrolled 2 randomized controlled trials. Methods An enhanced probabilistic neural network model predicted whether showed low, medium, or high response therapy, as measured Motor Activity Log, basis baseline assessments: Wolf Function Test, Semmes-Weinstein Monofilament Test touch threshold, Montreal Cognitive Assessment. Then, dynamic classification algorithm applied improve prognostic accuracy using most accurate combination obtained previous step. Results ability tactile sense improvement for activities intensive rehabilitation an nearly 100%. Complex patterns interaction among these predictors observed. Limitations fact that this sample size limitation. Conclusions Advanced machine learning/classification algorithms produce personalized predictions outcomes than commonly used general linear models.