Višeskalni računarski model mišića zasnovan na makromodelu konačnih elemenata i Hakslijevom mikromodelu

Abstract

The study of the muscle behavior based on precisely defined computer models is one of the greatest challenges in the field of applied science and engineering. Changes in the structural and functional characteristics of muscles during some diseases or disorders, require modeling of biophysical processes on several spatial and temporal scales. Multiscale muscle models can implement different phenomenological or biophysical muscle models within a microscale. The implementation of phenomenological micromodels contributes to the lower complexity of the multiscale model, but such models are not able to accurately predict transient muscle behavior under non-isometric conditions. To improve these shortcomings, a multiscale muscle model based on the finite element macromodel and the Huxley micromodel was developed as part of the thesis. The finite element method (FEM) integrates the active and passive material characteristics of the muscles into a continuum mechanics on the macroscale, while a modified Huxley’s cross-bridge model is used to calculate active muscle tension and instantaneous stiffnessin muscle fibers on the microscale. All predictions generated by the FE-Huxley multiscale model were verified by comparison with experimental results and with simulation results obtained by spatially explicit molecular model (MUSICO). The possibilities of using the FE-Huxley model in simulations of complex muscles are presented on a 2D model of the human tongue. Also, the use of the FE-Huxley model in simulations of certain muscle diseases is presented. Thanks to the Mexie platform for parallel execution simulations of multiscale muscle models, computationally demanding simulations of the FE-Huxley model are performed in a reasonable time frame, which makes the model usable for a variety of research and clinical applications.