A frequency domain inverse solution in electrocardiography

Recent advances in the development of detection schemes for myocardial ischemia which use frequency-domain parameters as detection statistics open up the possibility of using frequency-domain methods in the inverse problem of electrocardiography. The application of the minimum relative entropy (MRE) reconstruction principle to this problem is considered and its theoretical justification is given. The MRE method is compared to a standard time-domain least-squares method, constrained least squares (CLS). Initial results indicate that the MRE method promises to be able to overcome limitations of the time-domain least-squares methods which tend to make the latter difficult to apply in a clinical setting.<>