A theoretical model of the spatial distribution of EEG/MEG for correlated and uncorrelated sources

In this paper a study is made of the relationship between the correlation of EEG and MEG signals and the coherence of brain activity. The quantification of this relationship is complicated by the fact that electrical activity of a current source spreads out in space before it is measured with EEG electrodes or MEG sensors. In this paper a mathematical model is derived which expresses the covariance of the EEG/MEG sensors into the statistics of the sources. When both the volume conductor and the distribution of the sources are spherically symmetric, the spatial covariance can be expressed as a simple series of Legendre polynomials. For the EEG the variance is studied as a function of electrode distance, for several source correlation distances. Due to the large amount of smearing of the skull these functions are hardly different showing that the EEG is not a very sensitive instrument to distinguish between correlated and uncorrelated sources. For the MEG the model is tested experimentally by plotting the coherence as a function of gradiometer distance. It appears that these data are in good agreement with the model