<p> Electrophysiological behaviors in cardiomyocytes, such as the action potential and calcium transient, are emergent properties arising from the interaction of an ensemble of ion channels, transporters, and pumps. In this Thesis, I integrate mathematical modeling with experiments to gain new insight into cardiac electrophysiology. Cardiomyocyte models are probed using population-based parameter sensitivity analysis to comprehensively generate quantitative predictions of how key behaviors are determined by the levels of ion channels, transporters, and pumps. Experimental tests ground these predictions in reality and provide opportunities for model improvement when predictions differ from experiments. In Chapter 2, this approach was applied to the determinants of calcium transient amplitude in rat cardiomyocytes. Experiments validated the unexpectedly large predicted effect of the transient outward potassium current on calcium transient amplitude in epicardial cardiomyocytes, but others demonstrated that the sarco/endoplasmic reticulum calcium ATPase had a much larger impact than predicted. Further exploration revealed that model calcium fluxes were inaccurately balanced, which we corrected to yield an improved model accurately reflecting our experiments and previous reports. In Chapter 3, the determinants of action potential duration in guinea pig cardiomyocytes predicted by parameter sensitivity analysis were tested using dynamic clamp, which found generally larger experimental effect sizes than predicted. We adjusted the model using a genetic algorithm to match our results, which led us to show that the overly stable model action potential resulted from higher levels of the slow delayed rectifier current than in our experiments. Subsequent analysis revealed how this current more effectively stabilizes the action potential than a related current, the rapid delayed rectifier. Finally, in Chapter 4 I take a global approach to model analysis, exploring competing models of the rabbit cardiomyocyte by comparing patterns of variability and correlations between behaviors across a population of models with randomly varied parameters. This found key experimentally testable differences between the models, representing a novel potential method for assessing how well these mathematical models represent the electrophysiological system of these cells. Overall, this work adds to our understanding of cardiac electrophysiology and represents a potential new paradigm for combining modeling and experiments to understand complex behaviors.</p><p>