LAUSR

Abstract This study aims to develop and evaluate an automated system for upper-limb motor function assessment of stroke patients. The proposed system contains one motion tracking subsystem (to measure the kinematic data of participants through one Kinect V2) and one motor function assessment subsystem (to realize the automated assessment based on a feed-forward neural network (FFNN)-based assessment model). For validation, 16 stroke patients and 10 healthy subjects were recruited to perform 4 WMFT-FAS tasks, and 5 evaluation metrics were used. The experimental results showed that the proposed system could present satisfactory performance (accuracy: 0.87–0.96, F1-score: 0.83–0.93, specificity: 0.94–0.98, sensitivity: 0.87–0.95, and AUC: 0.93–1.00), and the FFNN-based assessment model could also present promising comprehensive performance (top two in all tasks in terms of accuracy and F1-score).