Enhanced computer modeling of cardiac action potential dynamics using experimental data-based feedback

Mathematical models of cardiac action potential (AP) dynamics are useful for studying the formation of dynamically significant patterns such as alternans and conduction block. A closed-loop observer is an augmented version of a mathematical model, in which experimental data are supplied to the model through feedback. In this study, tools for observer analysis were applied to a two-variable Karma model of AP dynamics. For a single-cell system, it was confirmed that membrane potential data could be used to reconstruct the system state, and that Luenberger feedback could stabilize the observer. Next an observer with a 1D geometry was tested with microelectrode membrane-potential data from a 2.1cm in vitro canine Purkinje fiber. It was shown that the observer produced more accurate AP duration (APD) estimates than the model by itself. These reconstructed quantities could be used to provide enhanced information to anti-tachyarrhythmic stimulus protocols that depend on real-time measurements.