Evaluation and Design of Non-Invasive, Wearable Musculoskeletal Monitoring Tools for Research, Occupational and Sport Applications

Abstract

The ability to estimate musculoskeletal dynamics (e.g., forces, stress and motion of individual muscles, bones and tendons) provides numerous scientific, societal, and clinical opportunities. Over the past few decades, new imaging and wearable measurement modalities have enhanced capabilities for long-term, widespread, and unconstrained monitoring of human movement outside the lab. However, there are key knowledge gaps regarding the validation, application, and interpretation of measurements from these modalities within the context of monitoring musculoskeletal dynamics. In this work, I first elucidate the limitations of two musculoskeletal monitoring approaches (using ultrasound to monitor muscle-tendon dynamics and using ground reaction force metrics to infer lower leg bone load and associated overuse injury risk). Second, I develop new approaches combining wearables, musculoskeletal biomechanics and machine learning to more accurately estimate loading on structures inside the body, specifically lower leg bone loading during running and low back loading during manual lifting.