Comparison of biomechanical analysis results using different musculoskeletal models for children with cerebral palsy

Abstract

Introduction: Musculoskeletal model-based simulations have gained popularity as a tool for analyzing human movement biomechanics. However, when examining the same gait, different models with varying anatomical data and assumptions may produce inconsistent biomechanical results. This inconsistency is particularly relevant for children with cerebral palsy, who often exhibit multiple pathological gait patterns that can impact model outputs. Methods: The aim of this study was to investigate the effect of selecting musculoskeletal models on the biomechanical analysis results in children with cerebral palsy. Gait data were collected from multiple participants at slow, medium, and fast velocities. Joint kinematics, joint dynamics, and muscle activation were calculated using six popular musculoskeletal models within a biomechanical simulation environment. Results: The degree of inconsistency, measured as the root-mean-square deviation, in kinematic and kinetic results produced by the different models ranged from 4% to 40% joint motion range and 0%–28% joint moment range, respectively. The correlation between the results of the different models (both kinematic and kinetic) was good (R (Formula presented.) 0.85, P (Formula presented.) 0.01), with a stronger correlation observed in the kinetic results. Four of the six models showed a positive correlation between the simulated muscle activation of rectus femoris and the surface EMG, while all models exhibited a positive correlation between the activation of medial gastrocnemius and the surface EMG (P (Formula presented.) 0.01). Discussion: These results provide insights into the consistency of model results, factors influencing consistency, characteristics of each model’s outputs, mechanisms underlying these characteristics, and feasible applications for each model. By elucidating the impact of model selection on biomechanical analysis outcomes, this study advances the field’s understanding of musculoskeletal modeling and its implications for clinical gait analysis model decision-making in children with cerebral palsy.