UC Berkeley

Finite element analysis (FEA) is a computational method used to predict the behaviour (stresses, strains and deformation) of a structure under predefined loading conditions. It can be applied to different biological structures, such as bone, to study defined muscle-driven scenarios. However, as muscle is an extremely complex structure to model, evolutionary biologists usually model muscle forces indirectly. In 2007, the BONELOAD MATLAB routine was developed to distribute muscle forces on a surface defined by the user. This routine then had to be coupled with a pre-existing FEA software (e.g. Strand7) to perform the analyses and has been widely used ever since. In this manuscript, we present a new method to run muscle-driven finite element simulations on a bone by distributing muscle forces on their insertions area, all within a single environment. We apply this protocol in three different situations: two biting simulations (unilateral and bilateral) and a shoulder flexion simulation. We demonstrate how to prepare the mesh, delineate the muscle origins and insertions, define the constraints, adjust material properties, choose a loading scenario (uniform, tangential or tangential-plus-normal), and extract the results. Our automated script meshes the 3D model, defines the constraints and distributes muscle forces within a single simulation software: ‘Metafor’ (nonlinear solver, owned and distributed by Gesval S.A) or ‘Fossils’ (a new open-source linear static solver developed in the frame of this work). ‘Metafor’ and ‘Fossils’ can perform the entire protocol (from the meshing to the muscle-induced simulations) on high-resolution volumetric meshes (millions of tetrahedra) and rapidly, exceeding the processing time of other widely used software protocols by up to four times. We demonstrate that the results obtained from our protocol are highly congruent with brands such as Strand7. Thus, our protocol opens up the possibility to routinely and rapidly simulate the behaviour of high-precision muscle-driven FE models containing millions of tetrahedra.