Abstract
During the past few years it has been demonstrated that active areas in the human brain can be located by measuring the magnetic fields arising from the electric currents in the neurons. An established model for the conductivity geometry of the head in these studies is the layerwise homogeneous sphere. If, however, the measurement grid is too large or the local radius of curvature of the head is changing rapidly in the measurement area, this simple model may become inadequate. The authors investigate the feasibility of replacing the conducting sphere by a homogeneous body having the shape of the brain. The authors show by a semi-quantitative argument that the homogeneous body model is justified. A numerical procedure for the calculation of the magnetic field is presented with examples of the accuracy that can be obtained. An example of significant differences between the predictions of the homogeneous and sphere models is discussed.