Locating Cardiac Ischemia

The paper describes an optimization problem of identifying heart ischemia using ECG measurements on the torso surface. Heart ischemia is a condition of reduced blood supply to cardiac tissue and can lead to heart infarctions. In the simulations, we model the cardiac electrical activity by a stationary form of the bidomain equations. These are solved using an adaptive finite element method on a realistic geometrical model of the heart, lungs and torso, obtained from magnetic resonance imaging data. To model the ischemic region, we use a set of parameters in a level set framework. The ischemia is identified by minimizing a cost functional. Since this is a severly ill-posed problem, we investigate the effect of different regularization methods.