A spatial filtering forward and inverse model of EEG and MEG

The electroencephalogram (EEG) and the magnetoencephalogram (MEG) recorded outside the human head are the result of underlying neural activity in the cortex. The process of determining the neural sources from the measured EEG and/or MEG is known as the inverse problem. Given the source distribution, the forward problem may be solved by electromagnetic field theory, which relates the sources to the electric potential and magnetic field by Green´s functions. Inversion of the Green´s function is then necessary to perform the inverse procedure. Since this inversion generally results in instability, the inverse Green´s function must be regularized. Examination of the regularization of the inverse problem reveals important information about the spatial resolving power, the accuracy of the inverse calculation and its sensitivity to noise