Iterative approaches to the solution of the bioelectric inverse problem

Iterative approaches to the solution of the Bioelectric and the Biomagnetic inverse problems have been the subject of several recent studies. The unifying feature of these methods is that the estimate of a source based on noisy sensor measurements can be improved considerably by reducing the region of brain that needs to be searched for a solution. Although ancillary data might be used for this purpose, for example PET or functional MRI, recent studies are concerned with "stand alone" methods in which the search space it reduce iteratively based only on the measured evoked potential field or evoked magnetic field. The iterative methods are explored here using a realistic simulation. Anatomical MRI data from a human subject are used to model the head, the boundary element method is used to calculate potential fields on the scalp, and samples of background EEG are used as the perturbing noise.