Clinical Spasticity Assessment Assisted by Machine Learning Methods and Rule-Based Decision.

The Modified Ashworth Scale (MAS) is commonly used to assess spasticity in clinics. The qualitative description of MAS has resulted in ambiguity during spasticity assessment. This work supports spasticity assessment by providing measurement data acquired from wireless wearable sensors, i.e., goniometers, myometers, and surface electromyography sensors. Based on in-depth discussions with consultant rehabilitation physicians, eight (8) kinematic, six (6) kinetic, and four (4) physiological features were extracted from the collected clinical data from fifty (50) subjects. These features were used to train and evaluate the conventional machine learning classifiers, including but not limited to Support Vector Machine (SVM) and Random Forest (RF). Subsequently, a spasticity classification approach combining the decision-making logic of the consultant rehabilitation physicians, SVM, and RF was developed. The empirical results on the unknown test set show that the proposed Logical-SVM-RF classifier outperforms each individual classifier, reporting an accuracy of 91% compared to 56-81% achieved by SVM and RF. A data-driven diagnosis decision contributing to interrater reliability is enabled via the availability of quantitative clinical data and a MAS prediction.

Elucidating factors influencing machine learning algorithm prediction in spasticity assessment: a prospective observational study.

The Machine Learning Model (MLM) has garnered popularity in rehabilitation, ranging from developing algorithms in outcome prediction, prognostication, and training artificial intelligence. High-quality data plays a critical role in algorithm development. Limited studies have explored factors that may influence the MLM algorithm performance in predicting spasticity severity level. The objectives of this study were to train and validate a MLM algorithm for spasticity assessment and determine the algorithm's prediction performance in predicting ambiguous spasticity datasets. Forty-seven persons with central nervous system pathology that fulfilled the inclusion and exclusion criteria were recruited. Four biomechanical properties of spasticity were obtained using off-the-shelf wearable sensors. The data were analyzed individually, and ambiguous datasets were separated. The acceptable inertial data were used to train and validate MLM in predicting spasticity. The trained and validated MLM algorithm was later deployed to predict the ambiguous spasticity datasets. A series of MLM were applied, including Support Vector Machine, Decision Tree, and Random Forest. The MLM's performance accuracy of the validation data was 96%, 52%, and 72%, respectively. The validated MLM accuracy performance level predicting ambiguous datasets reduces to 20%, 23%, and 23%, respectively. This study elucidates data biases and variances of disease background, pathophysiological and anatomical factors that have to be considered in MLM training.