An iterative method for indirectly solving the inverse problem of electrocardiography

Solution of the inverse problem of electrocardiography is an ill-posed problem and noise sensitive, leading to non-physiological solutions under noise conditions. We propose the use of an indirect iterative method for solving the inverse problem in terms of multiple moving dipoles. A database of surface potentials generated by single dipoles was generated by solving the forward problem. The inverse problem was solved by iteratively searching the combination of surface potentials in our database most similar to the observed surface potentials. Simulated surface potentials were used to test the algorithm. Single dipoles were accurately located in 92% simulations and in 90% simulations in which noise (SNR=6dB) was added. Double dipoles were accurately located in 78% simulations, and 72% when 6dB noise was added. The present method has shown to be accurate and robust against noise.