Experimental and theoretical investigation of cardiac restitution and memory: A comprehensive approach using the restitution portrait

Restitution, the relationship between action potential duration (APD) and the preceding diastolic interval (DI), has been implicated as a source of instability in cardiac response, possibly leading to ventricular fibrillation. However, assessment of the role of restitution in the onset of abnormal rhythms has produced mixed results. This is due in part to an incomplete understanding of restitution, particularly the role of memory (the dependence of APD on activation history) in cardiac response.;In this dissertation, memory was investigated by applying a theory-derived pacing protocol to bullfrog ventricular tissue. This protocol simultaneously measured dynamic, S1-S2, and constant-basic cycle length (CB) restitution. The results were presented in a restitution portrait (RP), which concurrently displayed these responses, thus emphasizing the distinctions between them. The same protocol was applied to the Fox-McHarg-Gilmour ionic model of cardiac membrane, yielding an RP that was qualitatively similar to the experimental one.;Mapping models of cardiac dynamics were also investigated to determine whether they reproduce the responses seen in the RP. It was determined that a mapping model must have at least three variables (APD is a function of the previous APD and two previous DIs) to reproduce the qualitative features of the experimental and ionic model RPs. A graphical analysis was performed to explain these features. The analysis pointed to limitations in conventional restitution protocols and the importance of CB restitution. Two three-variable models, the Chialvo-Michaels-Jalife model and the David G. Schaeffer model, were fit to RP data from the experiment and the ionic model. Qualitative and partial quantitative agreement was obtained. However, stability criteria based on the three-variable mapping model failed to predict the onset of alternans. These results suggest that more memory or another model form may be more appropriate to describe cardiac response.;The unified framework provided by the RP is a powerful tool for investigating the multidimensional aspects of cardiac response in experiments and for evaluating models of cardiac dynamics. This work elucidated the relationship between restitution and memory, and proved that stability criteria and mapping models developed to date are not adequate. Future studies should be directed towards addressing these deficiencies.