Data-driven generation of Targeted Biomarkers for Comprehensive Motor Function Assessments: The Network-Information Cluster Framework

To empower motor function assessments, new mechanistic approaches with greater explanatory depth are required. Here we address this need by synthesising network- and information-theoretic tools and machine-learning into a fully end-to-end biomarker generation framework. Showcasing our approach, we perform a comprehensive spatio-spectral decomposition of muscle activations into functionally diverse muscle networks. Then, by incorporating rigorous feature selection and our newly developed clustering algorithm, we identify motor features optimally associated with a chosen clinical measure and cluster participants in a targeted, clinically meaningful way across scales. Framework applications illustrate the mechanistic insights provided into the underlying physiological constructs of any clinical measure, uncovering data-driven milestones of aging and post-stroke impairment chronicity and neurocomputational motor characteristics. This adaptable framework bridges the underutilised large-scale motion data of clinical labs to the assessment tools they currently rely upon, offering in-depth characterisations of individual motor (dis)abilities, thus representing a powerful new assessment methodology.