| Summary: | Cardiovascular disease is a leading cause of death throughout the world. In order to understand the mechanisms by which the heart functions, computational models of a variety of cardiac myocytes have been developed. Although these models provide substantial information on the activity of single heart cells, there is limited understanding of the interactions of these cells at the tissue level. This thesis is concerned with the integration of such single cell models into larger, more complex tissue structures. Firstly, this thesis introduces and reviews a number of single cell models. These individual models were integrated into a structurally detailed 2D slice of heart with a histological geometry. This model was used to investigate the cardiac sequence of a rabbit heart, and a number of biophysical parameters were compared with, and found to closely reproduce experimental data. To further investigate, a 3D rabbit heart geometry was obtained and segmented into atrial, sino-atrial nodal, atrio-ventricular nodal, Purkinje fibre and ventricular tissue, with each ventricle divided into endocardial, epicardial and mid-myocardial regions, with an aim to investigate the effect of a number of pro- and anti-arrhythmic effects on the rabbit heart. Analysis of the 3D model predicted huge computational load related to the resolution of millions of cells. Currently, large scale computational simulations of cardiac activity are carried out using dedicated supercomputing facilities, but such solutions can be costly and unwieldy. To overcome this, investigation was carried out into the use of graphical programming to decrease the computational time required to resolve such equations. Investigation was carried out into the problem size at which graphical computation outperforms traditional computation, and found to give up to a hundred-fold increase in computational speed at 10 6 cells. These results pave the way for a new style of cardiac programming, allowing quick and detailed investigation into a multitude of problems using cheap, easily available hardware.
|
|---|
