Stochastic Biological Modeling: The Forward Problem of Electrocardiography

by
Sarah Geneser

Advised by
Mike Kirby

The accuracy of computaional biology simulations is highly dependent upon realistic model parameters. These parameters often vary largely depending on the system of interest and can have a number of different underlying statistical distributions (uniform, gaussian, etc). While the majority of biological simulations are treated deterministically, randomly distributed parameters result in stochastic systems. Such systems can be solved with Monte Carlo methods, but the computational time to generate reliable means and standard deviations for the simultaions is unreasonably large. Polynomial Chaos (PC) is an alternative and significantly more efficient method, which produces higher-order statistics for the responses of stochastic systems. We applied polynomial chaos (PC) to the forward problem of electrocardiography to assess the importance of accurately determining the conductivity of organ tissues in a two-dimensional torso model.