Electrical anisotropy of fascicles in a realistic neck model for phrenic nerve stimulation

Purpose The purpose of the paper is to analyze and compare a novel three-stage macromeso-micro (M3) modeling framework for simulating phrenic nerve stimulation. This stimulation might be used to improve ventilation in intensive care unit patients. The focus is on a detailed finite element analysis of the first two M3 steps using a realistic neck model. Design/methodology/approach A finite element method for stationary electric current flow was used to compute extracellular electric potential distributions along intra-fascicular pathways of the phrenic nerve, essential for nerve activation prediction. A 14-tissue, electrically isotropic realistic neck model was used in the first stage (macro model). Subsequently, in the meso stage, a geometric phrenic nerve model with three electrically anisotropic fascicles representing bundles of axon fibers was analyzed. In this step, the electric potential values calculated in the macro stage were used to define Dirichlet boundary conditions on the nerve surface. The pathways along which the extracellular electric potential was calculated run centrally within the fascicles. To assess the accuracy of the extracellular electric potential calculations, a full macro model including electrically anisotropic fascicles was additionally developed. The potential values calculated based on this model are used for validation and error calculations. Findings The results demonstrate changes in the courses of the extracellular electric potential with implemented anisotropic electrical conductivity. Specifically, these changes are larger at the boundaries of the fascicles compared to the center of the fascicles. The use of separate macro- and meso-geometric models can significantly increase computation time. Therefore, it is recommended to use (if computational resources allow) only one, full macro-geometric model that accounts for the electrical anisotropy of fascicles. Originality/value Development and comparison of modeling approaches for simulating phrenic nerve stimulation.