Validation of a novel subject-specific musculoskeletal strength-scaling workflow using submaximal dynamic strength tests
Authors
Jensen, Bjørn Keller ; Jensen, Cecilie Bjørn ; Pedersen, Mads Kattrup
Term
4. semester
Education
Publication year
2019
Submitted on
2019-06-06
Pages
49
Abstract
Formål: at udvikle og afprøve en arbejdsgang, der gør det muligt at skalere muskelstyrke i muskuloskeletale (muskel-knogle) computermodeller til den enkelte person, og at validere disse styrkeskalerede modeller med isometriske (kraft uden bevægelse) ledmålinger. Metode: 28 voksne (21 mænd, 7 kvinder) fik estimeret deres one-repetition maximum, 1RM (den maksimale vægt man kan løfte én gang), i 10 øvelser ved felttest. Disse 1RM-estimater blev brugt i en optimeringsalgoritme til at beregne styrkefaktorer, som skalerede alle relevante muskler i 10 øvelsesspecifikke modeller. Derudover blev maksimale ledmomenter målt isometrisk med et dynamometer (måleudstyr) for albuebøjning/-strækning, knæbøjning/-strækning og ankel plantarfleksion. Resultater: Den optimeringsbaserede styrkeskalering reducerede den gennemsnitlige normaliserede root mean square error (et mål for forskellen mellem model og data) fra 48,39% (±22,99%) til 28,13% (±15,47%) sammenlignet med standard skalering. Optimeringsrutinen var hurtigere end tidligere anvendte metoder og mere nøjagtig end standard styrkeskalering. Udfordringer med simple muskelbaner, der bøjer rundt om knæ- og ankelled, gjorde sammenligningen med dynamometerdata ikke mulig. Konklusion: Den foreslåede optimeringsrutine forbedrer helekrops muskel-knogle-modeller og kan nemt udvides med flere øvelser. Simple modeller er dog ikke egnede til at estimere og sammenligne topmomenter tæt på leddenes yderstillinger.
Purpose: to develop and test a workflow that personalizes musculoskeletal computer models to an individual’s muscle strength and to validate these strength-scaled models using isometric (force without movement) joint measurements. Methods: Twenty-eight adults (21 men, 7 women) completed field strength tests in 10 exercises to estimate their one-repetition maximum, 1RM (the heaviest weight a person can lift once). These 1RM estimates were used in an optimization algorithm to compute strength factors that scaled all relevant muscles in ten exercise-specific models. We also assessed peak joint torques isometrically with a dynamometer (a device that measures torque) for elbow flexion/extension, knee flexion/extension, and ankle plantarflexion. Results: Optimization-based strength scaling reduced the mean normalized root mean square error (a measure of disagreement between model and data) from 48.39% (±22.99%) to 28.13% (±15.47%) compared with standard scaling. The optimization routine was faster than previously used methods and more accurate than standard strength scaling. However, issues with simple muscle paths that wrap around the knee and ankle made comparison with the dynamometer data infeasible. Conclusion: The proposed optimization improves whole-body musculoskeletal models and can be extended with additional exercises. Simple models, however, are not suitable for estimating and comparing peak joint torques near end-range joint angles.
[This abstract was generated with the help of AI]
Keywords
Documents