A Case Study of Error Fields: A Three-Dimensional Personalized Robotic Therapy

Abstract

Treating post-stroke individuals with physical devices (mechanoceuticals) has produced mild but significant recovery. While it is widely recognized that therapy should focus on movement error, there is currently no model of relearning that can inform the design of longer-term treatment. We present a case study dubbed Error Fields, a personalized robotic therapy for stroke survivors, based on a mathematical model of each person's unique response to error, used to diagnose and address specific deficits. After three days of Error Fields treatment, errors significantly decreased compared to baseline demonstrated by shift in error distributions. In contrast, three days of practice with the robotic system without treatment showed no significant error reduction. Comparing Error Fields sessions to null treatment sessions also revealed significant improvements in error levels.