Reducing the Complexity of Musculoskeletal Models Using Gaussian Process Emulators

Musculoskeletal models (MSKMs) are used to estimate the muscle and joint forces involved in human locomotion, often associated with onset of degenerative musculoskeletal pathologies (e.g., osteoarthritis). Subject-specific MSKMs offer more accurate predictions than their scaled-generic counterparts. This accuracy is achieved through time-consuming personalisation manual tuning procedures that suffer from potential repeatability errors, hence limiting wider application this modelling approach. In work we have developed a methodology relying on Sobol’s sensitivity analysis (SSA) for ranking muscles based importance determination contact (JCFs) cohort older women. The thousands data points required SSA generated using Gaussian Process emulators, Bayesian technique infer input–output relationship between nonlinear limited number observations. Results show there pool whose has little effects JCFs, allowing reduced but still representation system within shorter timeframes. Furthermore, subject-specific generic influenced by different sets muscles, suggesting existence model-specific component analysis.