Cardiac myocyte model parameter identification via dynamic electrophysiology protocols and automated search algorithms

Cardiac myocyte models integrate data on ionic processes underlying physiological and pathophysiological cell behavior. The associated electrophysiological activity stems from the biophysical action of ion channels, pumps, exchangers, and transporters in the context of the cellular milieu and the experimental stimuli used. Although there has been progress in identifying and incorporating new model components, including an increasing number of distinct and disease related currents, parameter estimation remains a challenge. Selection of values for cell-level parameters associated with each ionic component, such as the maximal ion channel conductance, remains difficult due to the variable experimental conditions and disparate nature of the source data. As cardiac myocyte models are built upon data from multiple cells, cell-specific variability is not accounted for in the inherently population-based models used presently. Furthermore, the standard approach for identification of model parameters involves the use of simple voltage step protocols, which do not probe the wide range of dynamics associated with cardiac arrhythmia.;This work uses a combination of dynamically rich electrophysiology protocols with an automated search procedure to simultaneously tune multiple ionic parameters. First, we validate our approach by conducting the first global optimization of a cardiac model for a real-time electrophysiology experiment. We show that it is possible to rapidly estimate a mouse electrophysiology model as to enable mouse-to-human cell conversion for a wide range of initial experimental morphologies. Next, we investigate the relationship between predicted electrophysiological behavior and estimation of underlying parameter estimates. We consider arrhythmia pertinent protocols and a multi-segment voltage clamp protocol designed to target current components in models and experiments. Finally, we conduct simulations of multiple drug block configurations and corresponding parameter estimations to disentangle effects of multi-target drug action on total current contributions.;This body of work enables the transition from generic to specific models, furthering the utility of simulation engaging experiment in cardiac electrophysiology. The biological system of a cardiac myocyte is complex and by studying component interactions in the context of a fully functional cell, we allow for component interactions under novel protocols. In particular, the automated tuning of ionic parameters presented provides a useful tool in the model building process for discovering parameter combinations corresponding to a dynamic range of whole cell behavior.