Download files
Abstract
A variety of cardiac action potentials (AP) and ionic currents from heterogeneous cardiac myocytes have been studied quantitatively using whole-cell current and voltage-clamp techniques. Such data provide essential information concerning the electric activity of cardiac tissue. At the same time, mathematical representations are required to provide a quantitatively deeper understanding of the underlying mechanisms of cardiac electric activity. Existing cardiac ionic models which are able to accurately reproduce a range of electrophysiological behaviours, are highly complex, with ever increasing number of variables and parameters, which render these computationally expensive when integrating into higher-dimensional tissue or whole heart simulations, Moreover, detailed mathematical models may accurately reproduce physiological ionic mechanisms only under the experimental conditions they were based on, limiting the predictive utility of these models. Simple models are computationally practical because of their highly-simplified nature. However, despite the successful application of previous generic models in cardiac modeling, their inherent over-simplicity restricts their further utility.
It is therefore desirable to develop computationally simple, robust ionic models and corresponding parameter optimisation approach, which can accurately simulate or predict action potentials as well as other underlying physiological mechanisms in a variety of cardiac myocytes. The model structure should be flexible and modular. By optimising the appropriate model parameters the model should be able to fit action potentials recorded from different myocytes under different experimental conditions.
Towards this aim, in this study, a simplified Hodgkin-Huxley-type generic model was used to accurately reconstruct central SAN cells, peripheral SAN cells, as well as right and left atrial action potentials recorded from intact rabbit cardiac tissue preparations. More importantly, different types of multiple dataset fitting improving parameter reliability were applied using a custom least squares optimisation algorithm. Given appropriate experimental datasets, many of the known physiological ionic currents could be effectively reconstructed. Also, the optimised model was able to predict additional experimental recordings that were not used in the optimisation process. The generic model will be useful in higher dimensional whole-heart simulations to investigate cardiac arrhythmias and possible pharmacological treatment.