Comparison of different source localization methods

Abstract

The localization of the neural generators of the brain's electromagnetic activity based an electrical or magnetic recordings requires the solution of an ill-posed inverse problem. In its general form, this inverse problem has no unique solution. The only way to deal with this problem, is to restrict the set of possible solutions by using either anatomical or mathematical constraints or a combination of both. On the basis of these restrictions, different reconstructions of the sources of EEG and MEG are possible, each providing a different representation of the electrical processes in the brain. Opposed to the more classical dipole localization methods, inverse procedures that produce distributed estimates of neuronal currents have received increasing attention in the last few years. In this paper, we compare different source localization methods using data from a pathological subject. Among these methods are: (1) spatio-temporal source models (BESA); (2) generalized minimum norm solutions (MNS) and (3) electromagnetic temporal tomography (EMTT). In order to assess their specific performance, each method is compared with the information contained in the CT.