Predicting Multiple Sclerosis From Gait Dynamics Using an Instrumented Treadmill: A Machine Learning Approach

Objective: Multiple Sclerosis (MS) is a neurological condition which widely affects people 50-60 years of age. While clinical presentations MS are highly heterogeneous, mobility limitations one the most frequent symptoms. This study examines machine learning (ML) framework for identifying through spatiotemporal and kinetic gait features. Methods: In this study, data during self-paced walking on an instrumented treadmill from 20 persons with age, weight, height, gender-matched healthy older adults (HOA) were obtained. We explored two strategies to normalize minimize dependence subject demographics; size-normalization (standard body size-based normalization) regress-normalization (regression-based normalization using scaling factors derived by regressing features multiple demographics); proposed ML based methodology classify individual strides (PwMS) controls. generalized both across different tasks subjects. Results: observed that improved accuracy pathological when compared size-normalization. When generalizing comfortable while talking, gradient boosting achieved optimal classification AUC 94.3 1.0, respectively generalization, multilayer perceptron resulted in best 80 xmlns:xlink="http://www.w3.org/1999/xlink">% 0.86, respectively, regression-normalized data. Conclusion: The integration may provide viable patient-centric approach aid clinicians monitoring MS. Significance: results have future implications way regression normalized be clinically used design ML-based disease prediction monitor progression PwMS.