Forward-dynamics tracking simulation of coupled multibody and finite element models: application to the tongue and jaw

Finite-element (FE) models of muscle tissue are gaining popularity in musculoskeletal simulation research. In contrast to the traditional line-based Hill-type models, volumetric representations permit a detailed analysis of internal tissue deformation, contact conditions, and force transmission in broadly inserting muscles [1]. However, analysis techniques for FE models are less developed than those for traditional line-based muscles. One technique that has achieved widespread use in multibody musculoskeletal simulators (e.g. OpenSim [2]) is Computed Muscle Control (CMC) [3]. CMC predicts muscle excitations required for a forward-dynamics simulation to track a target kinematic trajectory. While CMC tracking-based simulation has been applied to strand-based musculotendon models [4], quasi-static FE-based muscles for the face [5] and line-based muscles for the tongue [6], it has yet to be applied to dynamically coupled rigid-body and FE muscle models. Doing so is needed for validating and applying volumetric muscle models in dynamic settings. We have extended the CMC algorithm for use with fully-coupled rigid skeletal bodies and FE-based muscle models. Our approach offers improved accuracy and convenience over the more standard practice of using multibody skeletal simulation to prescribe boundary conditions for use with a separate FE simulator (e.g. FEBio [7]). Our trackingbased simulation allows for flexible specification of targets, but nominally permits skeletal kinematics as input and FE-muscle activations and deformations as output. We demonstrate our approach with a jaw-tongue model (Figure 1) that tracks electro-magnetic midsagittal articulographic (EMMA) recordings of tongue and jaw movements in speech.