Magnetic Resonance Imaging–Measured Muscle Parameters Improved Knee Moment Prediction of an EMG-Driven Model

Abstract

Listen

Translate article

Acquisition of muscle anatomic parameters is essential for the development of a musculoskeletal model to estimate muscle forces and joint kinetics and can be derived in three ways: 1) use of a generic anatomic model, 2) scaling of a generic model based on anthropometric measures, and 3) direct in vivo measurements using various imaging techniques. The purpose of this study was to investigate how incorporating direct measurements of muscle anatomic parameters using magnetic resonance imaging (muscle volumes and moment arms) influences knee moment predictions when compared with generic and scaled models. Joint moment predictions of the three modeling approaches were examined by comparing the net knee moments calculated by each model with standard net joint moment measurements (inverse dynamics calculations and dynamometry) obtained while seven subjects (three females, four males) performed a drop landing and isokinetic knee extension task. The coefficient of multiple correlation and mean absolute difference were calculated to examine the prediction error and agreement of each model with standard net knee moment measurements. For both tasks, the model incorporating direct measurements of muscle volumes and moment arms had a higher coefficient of multiple correlation and smaller mean absolute difference than the generic and scaled models (effect size range = 0.99-1.37). The scaled model had a lower coefficient of multiple correlation and greater mean absolute difference than the generic model (effect size = 1.36). Our findings demonstrate that knee moment predictions from an EMG-driven model can be improved with direct measurements of muscle anatomic parameters. Knee moment predictions did not improve when scaling a generic anatomic model. Musculoskeletal models that incorporate direct measures of muscle anatomic parameters may provide more accurate assessments of joint kinetics when compared with generic and scaled models.