Real-Time Identification Performance of Fine Motor Movements for Traumatic Brain Injured Patients

During the assistive rehabilitation for neurologically disordered patients with motor impairment, identifying the patients' motor intents is an important issue according to the theory of brain plasticity. Motivated by examining the real-time identification performance of fine motor movements for traumatic brain injured (TBI) patients in this study, four male TBI subjects with mild to moderate motor impairment were included, and whose motor control performance in terms of five real-time performance metrics as well as one offline performance metric were evaluated. The results show similar offline motion classification accuracies (all above 95%) from all subjects, but obviously different in real-time scenario. And the real-time performance was found to be related with the specific motor impairment level of each subject that assessed by clinical evaluation scales. In addition, the learning ability of subjects was observed by analyzing the unsuccessful motion tasks. These findings suggest that it is necessary and important to conduct subject-specific physical exercise training based on the pattern recognition technique.