Objectives: The accuracy of electrical impedance tomography was investigated.

Methods: The conductivities of the different compartments of the volume conductor were estimated by utilizing the boundary element method. The approach was tested for realistic head models with either 3 or 4 compartments. The impact of a geometrical error in the head model was investigated and the estimated conductivities were assigned to the compartments of the volume conductor used for the source imaging. The localization errors were quantified.

Results: The method used allowed the estimation of the conductivity of the compartments. The poor conductivity of the skull decreased the precision with which the conductivity of deeper structures could be estimated. A geometrical error in the head model was compensated by the estimated conductivities. However, the estimated conductivities did not cancel the geometrical error in the head model as localization errors of the order of 10–20 mm were obtained.

Conclusions: In principle, the conductivityestimation of the distinct regions in the head is possible. The application of conductivityestimation to increase the accuracy of sourcelocalization remains questionable.