Cardiac biomagnetic source estimation with a heart-torso model and a trained neural network

The intensity of the cardiac sources for normal adult subjects was estimate d from given magnetic field profiles with a trained neural network based on the relationship of the electrical activity of the heart to the cardiac ma gnetic fields. The input for training the neural network consisted of the m agnetic field profiles above the torso during the heartbeat. The outputs we re the dipole intensities which produced those magnetic field profiles. A b ack-propagating algorithm with bias and momentum was utilized for training. The measured and simulated torso magnetic held profiles and magnetocardiog rams were used for training the neural network. Estimation of the dipole in tensities was performed for unknown magnetic field profiles with the traine d neural network. The estimated cardiac dipole intensities were reasonably close to the true dipole intensities. These results show the feasibility of the estimation of cardiac dipole intensities with a trained neural network under a very restricted forward model of the cardiac magnetic fields. Gene ralization of the results to cover a large population base could be difficu lt because the activation isochrones are different from subject to subject.