Stochastic modelling of cardiac cell structure

Abstract

Anatomically realistic and biophysically based computational models of the heart have provided valuable insights into cardiac function in health and disease. Nevertheless, these models typically use a "black-box" approach to describe the cellular level processes that underlie the heart beat. We are developing techniques to stochastically generate three-dimensional models of mammalian ventricular myocytes that exhibit salient characteristics of the spatial organisation of key cellular organelles in cardiac cell excitation and contraction. Such anatomically detailed models will facilitate a deeper understanding of cardiac function at multiple scales. This paper presents an important first step towards understanding and modelling the spatial distribution of two key organelles in cardiac cell contraction - myofibrils and mitochondria. The sarcolemma, myofibrils and mitochondria were segmented from transmission electron micrographs of ventricular cells from a healthy wistar rat. The centroids of the myofibrils and mitochondria were calculated, and various spatial statistical techniques for characterising the centroid distribution and inter-point interactions were investigated and implemented using the R spatstat package. Techniques for modelling the observed spatial patterns were also investigated, and preliminary results indicate that the Strauss Hard-core model best captures the interaction observed. We intend to confirm these results with larger sample of cells.