Automatic Assessment of the Wrist Movement Function in a Haptic Virtual Environment for Home-based Stroke Rehabilitation

The functional assessment of post-stroke patients is crucial in the rehabilitation theoretical process. In the study presented in this paper, an instrument which allows the patients to perform rehabilitation exercises was developed for the functional assessment without a need of assistance of nurses. This instrument included a desk-top device along with a classifier and can be operated by the patients at home. The grade for the functionality of the limp in the context of post-stroke patients is the Brunnstrom stage assessment system that is popular in clinic. In this paper, the wrist coordination functionality of upper limb (Grade 6 in particular) was taken as an example to demonstrate the effectiveness of this instrument along with the classifier. 16 patients and 10 health persons were involved in the development of the classifier. The instrument along with the classifier was tested, and the result indicated that the instrument can achieve a high classification accuracy (98.68%), sensitivity (92.31%), specificity (100%), F-score (96%) and the area under curve (AUC) of the receiver operating characteristic (ROC) (0.99878). Therefore, the instrument along with the classifier can be used in clinic with high confidence.