Objective Analysis of Muscle Spasticity Level in Rehabilitation Assessment

Abstract

In current practice, the assessment of upper limb spasticity is subjectively evaluated based on the experience and perception of therapists. This leads to inconsistency in assessment and could affect the efficacy of rehabilitation process. Thus, the aims of this paper are to study and extract relevant information from the torque and angle signal measured from the muscle of the arm and to select independent features in order to classify the level of spasticity of the muscle based on Modified Ashworth Scale (MAS) assessment tool. Data were collected from twenty-five subjects that met the criteria with consent. The data went through pre-processing stage and analyzed before the features extracted. The seven features extracted from the data forming the dataset which later used to train and feed into suitable classifier to classify the level of spasticity. One-way ANOVA test was run in order to evaluate the statistical significant differences among the level. Based on the results from the test, four features were selected out from seven. Linear Support Machine (SVM) based classifier accorded the highest performance with 84% accuracy compared to other classifiers.