Task difficulty adjustment in biocooperative rehabilitation using psychophysiological responses

Abstract

This study presents a biocooperative feedback loop where the difficulty of an upper extremity rehabilitation task is adjusted based on four psychophysiological measurements: heart rate, skin conductance, respiration and skin temperature. They are used both by themselves and in combination with task performance and biomechanics. Different variants of linear discriminant analysis are used for data fusion, including a variant that can adjust the fusion rules online and thus gradually adapt to the subject. Both healthy subjects and hemiparetic patients participated in the study. The accuracy rate of the biocooperative controller was defined as the percentage of times it matched the subjects' preferences. Psychophysiological measurements yielded a relatively low accuracy rate by themselves (76.4% for healthy subjects and 68.2% for patients). Task performance, on the other hand, yielded an accuracy rate of approximately 82%. Combining task performance with psychophysiology increased the accuracy rate to 84.7% for healthy subjects and 89.4% for patients. Psychophysiology can thus provide additional information, but factors such as the increased cost and complexity of the system should also be taken into account.