A Numerical Study on Conductivity Estimation of the Human Head in the Low Frequency Domain Using Induced Current MR Phase Imaging EIT With Multiple Gradients

An accurate conductivity estimation of human brain tissues is important for the correct diagnosis and therapy of neurological diseases. These values are patient-specific and vary naturally with the frequency. Nevertheless, they are often approximated by a constant value. Induced current magnetic resonance-electrical impedance tomography (ICMR-EIT) is a possible technique for non-invasive conductivity reconstruction in the low frequency domain. This paper presents a novel ICMR-EIT based method that uses the difference of two MR phase images. These images are obtained by gradient echo sequences with and without switching an eddy-current induction gradient. We propose the use of multiple gradients with different time periods, so that only a single parameter per tissue needs to be estimated with conservation of the frequency dependence. A numerical study on a spherical head model with four tissues investigates the feasibility of estimating conductivity values and shows that the proposed technique can successfully reconstruct these conductivity values. Furthermore, the influence of the material model and the number of harmonics associated to the gradient is investigated, together with performance of the proposed technique in the presence of noise.