Frequency{domain Localization of Intracerebral Dipole Sources First Author and Second Author Third Author Fourth Author the Theory of Neuromagnetic Fields Introduction

The localization of intracerebral dipole sources for detecting pathological events is one object of magnetoencephalography (MEG). To facilitate interactive evaluation of MEG data arti-cial neural network models are used. Trained on simulated examples with superimposed noise they are able to predict the so-called \goodness of t" of real data in contrast to minimization algorithms, which are iterative and more time consuming. This improves the results for backpropagation algorithms. We conclude that these neural networks are highly suitable for fast processing of neuromagnetic data. The activity of a nerve cell in the human brain causes a movement of ions and thereby a change in the magnetic eld. Because of this a variation of potential (electroencephalographic or EEG) and a variation of the magnetic eld (magnetoencephalographic or MEG) is caused at the surface of the head. Both phenomenons can be measured outside the head. In the following we will focus on the magnetic eld, because it comes undistorted through the tissue physique. Another reason for considering the magnetic eld is that remote activities overlie less of the magnetic eld. The measuring of the magnetic eld close to the scalp makes it possible to get information about the kind of sources which generate this magnetic eld and their position in the brain. In the following we will investigate so called