Parameter fitting using multiple datasets in cardiac action potential modeling

A multiple dataset model fitting approach for improving parameter reliability in action potential modeling is presented. A robust generic cardiac ionic model employing membrane currents based on two-gate Hodgkin-Huxley kinetics is described. Its generic nature allows it to accurately reproduce action potential waveforms in heterogeneous cardiac tissue by optimizing parameters governing ion channel kinetics and magnitudes. The model allows a user-defined number of voltage and time-dependent ion currents to be incorporated, in order to reproduce and predict multiple action potential waveforms recorded in intact cardiac myocyte. In total 12Nc+2 parameters were optimized using a curvilinear gradient method, where Nc is the user-specified number of time-dependent currents. Given appropriate experimental datasets, many of the known physiological membrane currents could be effectively reconstructed. Also, the optimized models were able to predict additional experimental action potential recordings that were not used in the optimization process.