Auto-tuning of Cardiac Myocyte Models
Armen Kherlopian, Cornell University
In the study of cardiac function and dysfunction, computational models of cardiac myocytes provide an important mechanistic view. However, established models with ionic parameters based on limited experimental data often underperform when trying to reproduce or predict phenomena outside the original model scope. Individual myocytes exhibit different behavior both when compared to one another and to a generic model. Apparent myocyte differences depend on factors that include the pacing protocols used, the presence of mutations or drugs, and the expression levels of voltage-gated ion channels, pumps, and transporters. Using nominal and drug block cases, we show that using dynamic pacing protocols in tandem with a genetic algorithm is both effective and efficient in determining the underlying ionic parameters for cardiac myocytes.
Abstract Author(s): Armen Kherlopian