Abstract
Conventional biomechanical modelling approaches involve the solution of large systems of equations that encode the complex mathematical representation of human motion and skeletal structure. To improve stability and computational speed, being a common bottleneck in current approaches, we apply machine learning to train surrogate models and to predict in near real-time, previously calculated medial and lateral knee contact forces (KCFs) of 54 young and elderly participants during treadmill walking in a speed range of 3 to 7 km/h. Predictions are obtained by fusing optical motion capture and musculoskeletal modeling-derived kinematic and force variables, into regression models using artificial neural networks (ANNs) and support vector regression (SVR). Training schemes included either data from all subjects (LeaveTrialsOut) or only from a portion of them (LeaveSubjectsOut), in combination with inclusion of ground reaction forces (GRFs) in the dataset or not. Results identify ANNs as the best-performing predictor of KCFs, both in terms of Pearson R (0.89–0.98 for LeaveTrialsOut and 0.45–0.85 for LeaveSubjectsOut) and percentage normalized root mean square error (0.67–2.35 for LeaveTrialsOut and 1.6–5.39 for LeaveSubjectsOut). When GRFs were omitted from the dataset, no substantial decrease in prediction power of both models was observed. Our findings showcase the strength of ANNs to predict simultaneously multi-component KCF during walking at different speeds—even in the absence of GRFs—particularly applicable in real-time applications that make use of knee loading conditions to guide and treat patients.
Highlights
Musculoskeletal diseases, along with natural age-related sensorimotor decline, affect the lower limbs’ soft tissue homeostasis and/or skeletal integrity, resulting in pain [1], muscle loss [2] and functional decline [3,4] along with elevated fall [5,6] and fracture risk [7]
We examined two of the most popular methods for non-linear regression, the artificial neural networks [59] and the support vector regression [60]
The current study demonstrates for the first time a mocap-agnostic, machine learning (ML)-empowered framework for prediction of multi-component knee contact forces (KCFs) during different speeds of walking and showcases the advantages of using supervised learning coupled with musculoskeletal modeling in mapping the kinematic to joint force space
Summary
Musculoskeletal diseases, along with natural age-related sensorimotor decline, affect the lower limbs’ soft tissue homeostasis and/or skeletal integrity, resulting in pain [1], muscle loss [2] and functional decline [3,4] along with elevated fall [5,6] and fracture risk [7]. Even though knowledge discovery has been immense during recent decades, the scientific community has been unsuccessful in treating musculoskeletal diseases, and current prevalence, incidence, and socioeconomic burden still impose a significant threat to healthcare systems [8,9] Measures to counter their adverse effects include pharmacological and rehabilitation interventions, mainly utilized when health status has already diminished, and the patient seeks medical help. Exercise has proven a lifetime successful surrogate strategy for prevention and treatment of numerous pathologies [10], since it can trigger an anabolic response to muscle and bone matrix and improve neuromotor function, followed by improvement in key cardiovascular indicators [11] and overall quality of life [12]. Research interest has focused on characterizing the mechanical loading exerted during exercise in important skeletal sites, in order to elucidate the interplay between external forces and biological response of the human skeleton [13,14,15].
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
