Multi muscle – magnetic resonance elastography (MM-MRE)

We are interested in understanding the mechanical behavior of individual muscles during coordinated joint motion. Currently, methods such as sEMG and ultrasound elastography can only be used to quantify individual muscle forces in a limited set of superficial muscles. In our research, we use magnetic resonance elastography (MRE) to obtain shear wave velocities from all muscles in the forearm with measurements of joint torque and angles to quantify forces of individual muscles during isometric tasks. Our MR-compatible wrist robot, the MRE-bot, allows us to safely measure wrist joint torques at different joint postures during magnetic resonance imaging (MRI). Our work focuses on quantifying individual muscle forces in the forearm to evaluate their role in the control of wrist joint tasks. The work on this project is done in collaboration with the Mechanical Neuroimaging Lab. In this protocol, individuals are cued to reach and hold a desired wrist torque target in either the flexion/extension or radio/ulnar deviation directions, while constrained in a desired wrist posture. Simultaneously, a pneumatic actuator and MRE scan are used to image the propagation of shear waves through forearm muscle tissue.
Fig. 1. - <p>Setup for multi-muscle magnetic resonance elastography. (a) Subjects are positioned headfirst and prone with their right arm placed in a custom MRE driver and perform isometric contractions cued by (b) the visual feedback system that allows the subject to visualize and maintain applied torque. (c) The MRE driver is connected to a pneumatic actuator.</p>

We use an anatomical image collected during scanning to segment muscle-specific regions of interest to obtain individual muscle values. Measurements of shear wave speed squared (SWSS) are normalized using the value of each muscle during the rest condition (torque = 0 Nm). In our first experiment, we collect MRE data during a 1-degree-of-freedom task where individuals applied wrist torques about the flexion/extension direction with varying magnitudes.

Fig. 3. - <p>(Left) Normalized shear wave speed squared speed (SWSS) values for all muscles during agonist and antagonist contractions, at both 0.5 and 1.0 N·m. Averages are connected via black lines while individual muscle data are connected via gray lines. (Right) Map of significant differences between activation states for agonist and antagonist muscles. Statistical significance (<italic>p</italic> &lt; 0.05) of post-hoc tests is denoted by *.</p>

MRE-based Muscle Force Estimation

Once we obtain the muscle-specific shear wave velocity squared measurements. We apply a linear regression model to obtain estimates of individual muscle forces. We define a linear relationship between shear wave velocity squared and muscle force that depends on muscle specific parameters of γ and κ as the slope and intercept parameters, respectively. It is infeasible to obtain non-invasive measurements of muscle forces, as a proxy, we implement the joint torque measurements to calibrate and evaluate the accuracy of muscle force estimation using MRE measurements. Using a musculoskeletal model obtained via OpenSim, we acquire the moment arm matrix and combine with measurements of joint torque and shear wave velocity squared to obtain the muscle-specific coefficients for all thirteen forearm muscles.

Fig. 7. - <p>(a) Estimated torque obtained through muscle force estimation. Average torque estimated across repeats and posture from each participant is reported via a small marker, filled with the color and shape corresponding to the cued contraction state (large markers). The ellipsoids mark the 2D distribution of estimated torque. (b) The average ordinary least squares fit per subject and (c) least squared error per contraction state help provide an overall picture of the models fit.</p>

Publications on this topic

C. A. Helm, F. Sergi, “Model-Based Estimation of Active and Passive Muscle Forces Using MRE in Forearm Muscles During 2-DOF Wrist Tasks”, doi: 10.1101/2024.02.15.580561, 2024, pre-print.

D. R. Smith, C. A. Helm, A. Zonnino, M. D. J. McGarry, C. L. Johnson, F. Sergi, “Individual Muscle Force Estimation in the Human Forearm Using Multi-Muscle MR Elastography (MM-MRE)”, IEEE Transactions on Biomedical Engineering, vol. 70, no. 11, pp. 3206-3215, Nov. 2023, pre-print, available online.

A. Zonnino, D. R. Smith, P. L. Delgorio, C. L. Johnson, F. Sergi, “MM-MRE: a new technique to quantify individual muscle forces during wrist isometric contractions using MR Elastography”, 16th International Conference on Rehabilitation Robotics, 2019, pre-print, available online.

Print Friendly, PDF & Email