A novel system for automatic classification of upper limb motor function after stroke: an exploratory study

In the early post-stroke phase, when clinicians attempt to evaluate interventions and accurately measure motor performance, reliable tools are needed. Therefore, the development of a system capable of independent, repeated and automatic assessment of motor function is of increased importance. This manuscript explores the potential of a newly designed device for automatic assessment of motor impairment after stroke. A portable motion capture system was developed to acquire three-dimensional kinematics data of upper limb movements. These were then computed through an automatic decision tree classifier, with features inferred from the Functional Ability Score (FAS) of the Wolf Motor Function Test (WMFT). Five stroke patients were tested on both sides across five selected tasks. The system was compared against a trained clinician, operating simultaneously and blinded. Regarding performance time, the mean difference (system vs clinician) was 0.17s (sd=0.14s). For FAS evaluation, there was agreement in 4 out of 5 patients in the two tasks evaluated. The prototype tested was able to automatically classify upper limb movement, according to a widely used functional motor scale (WMFT) in a relevant clinical setting. These results represent an important step towards a system capable of precise and independent motor evaluation after stroke. The portability and low-cost design will contribute for its usability in ambulatory clinical settings and research trials.