Efficient Representations of Cardiac Spatial Heterogeneity in Computational Models

TL;DR

This study explores how to efficiently model spatial heterogeneity in cardiac tissue by using coarse parameter grids, enabling accurate simulations of complex electrical behaviors with less computational effort.

Contribution

It introduces a method to represent cardiac tissue heterogeneity using coarse spatial grids, maintaining accuracy during complex dynamical states.

Findings

Coarse grid spacing of 1.0-1.6 cm yields accurate action potential duration profiles.

Piecewise-constant and piecewise-linear functions perform similarly in parameter assignment.

Efficient modeling of heterogeneous cardiac tissue is feasible with coarse spatial parameterization.

Abstract

It is generally assumed that all cells in models of the electrical behavior of cardiac tissue have the same properties. However, there are differences in cardiac cells that are not well characterized but cause spatial heterogeneity of the electrical properties in tissue. Optical mapping can be used to obtain experimental data from cardiac surfaces at high spatial resolution. Variations in model parameters can be defined on a coarser grid than considering each single pixel, which would allow a representation of heterogeneous tissue to be obtained more efficiently. Here, we address how coarse the parameterization grid can be while still obtaining accurate results for complicated dynamical states of spatially discordant alternans. We use the Fenton-Karma model with heterogeneity included as a smooth nonlinear gradient over space for more model parameters. To obtain the more efficient representations, we set parameter values everywhere in space based on the assumption that the exact parameter values are known at the points of the coarser grid; we assume the parameter values could be obtained from experimental data. We assign parameter values in space by fitting either a piecewise-constant or piecewise-linear function to the spatially coarse known data. We wish to identify the maximal grid spacing of such points to obtain good agreement with spatial profiles of action potential duration during complex states. We find that coarse grid spacing of about 1.0-1.6 cm generally results in spatial profiles that agree well with the true profiles for a range of different model parameters and different functions of those parameters over space. In addition, the piecewise-constant and piecewise-linear functions perform similarly. Our results to date suggest that matching the output of models of cardiac tissue to heterogeneous experimental data can be done efficiently, even during complex dynamical states.