Automated Upper Limb Motor Functions Assessment System Using One-Class Support Vector Machine

Stroke is a worldwide spread disease that causes movement disability for survivors preventing them from carrying out their daily activities in a normal manner. Stroke survivors have a long journey of rehabilitation, Fugl-Meyer Assessment (FMA) is a widely used assessment tool for evaluating the upper extremity motor function during rehabilitation. However, the FMA is time consuming and repetitive for both the patient and the physiotherapist which must be performed at a hospital or a clinic. This may not be possible sometimes due to time or place limitations. Thus, an automated alternative method was proposed which is not limited to a time or a place in which the patient preforms assessment exercises using the sensors. The system automatically scores the movement according to the FMA scoring system. In the proposed system, data is collected on the exercise movement through the sensors and then features are extracted and fed to a machine learning model, which after training classifies the movement done. The results show that the system is reliable and consistent, with an accuracy of around 87% in scoring the exercises done. Thus, the automated FMA portable assessment system could be potentially used to assess the upper extremity motor function during the long journey of post stroke rehabilitation removing the time and place constraints which may prevent the patient from following on their rehabilitation, also removing the repetitive and time-consuming burden from physiotherapists allowing them to only monitor and intervene when needed.