Comparison of Cardiac Activation Models Using 3D MCG Data

Abstract

In the study of the inverse problem in biomagnetism, the measured signals are used to localize the source that generates a given magnetic field distribution. The problem has no unique solution, since the same magnetic field map (MFM) can be generated by an infinite number of possible source configurations. To restrict the set of possible solutions, physiological information can be used. In magnetocardiography (MCG), the knowledge of the heart functionality may help in making a suitable choice to model the current source in the human heart. A possible choice for the current model is a single dipole, with varying position, orientation and amplitude inside the heart volume, the so-called source space. An alternative solution can be a multiple dipole model, characterized by a limited number of dipoles, with varying positions and orientations, or a limited range of variability for the position inside the source space. A more general description of the wavefront propagation can be provided by a distributed sources model, which considers an elevated number of dipolar sources, located in fixed positions inside the heart volume region, having varying amplitudes and directions, to model the current flow during the heart cycle. A single dipole model is a suitable one only for short segment of the MCG signals, whereas a distributed model allows estimating the current density in the source space. We performed the source localizations of MCG signals using different source models on the same data sets, to evaluate whether the differences in the source models descriptions can be minimized in the MCG data sets, or in some of them. The analyses are repeated on normal subjects and WPW patients, using normal components data and vector component data of the measured MFMs.