Use of a priori information in estimating tissue resistivities--application to human data in vivo

Abstract

Accurate resistivity values are necessary to construct reliable numerical models to solve forward/inverse problems in EEG and to localize activity centres in functional brain imaging. These models require accurate geometry and resistivity distribution. The geometry may be extracted from high resolution images. The resistivity distribution may be estimated by using a statistically constrained minimum mean squared error estimator algorithm that has been developed previously by Baysal and Eyüboğlu. In this study, the data are obtained by EEG and MEG sensors during SEF/SEP experiments that involve nine human subjects. The numerical model is realistic, subject-specific and the scalp, the skull and the brain resistivities are estimated. By performing nine different estimations, we found average resistivities of 3.183, 64.559 and 2.833 omega m for scalp, skull and brain, respectively, all under 9% standard deviation. The discrepancies between these results and other works are discussed in detail.