Image distortion and image mis-registration in low frequency current density imaging

Abstract

Current density imaging (CDI) is a technique that uses magnetic resonance imaging (MRI) to measure volume current density distributions in tissue. CDI is used to measure current pathways through tissue which adds a much needed tool to electrophysiological research such as in the area of defibrillation research. CDI maps magnetic fields, produced by an externally applied current, onto the phase images of an MRI data set. Current density is computed from the curl of the these magnetic fields. Two CDI artifacts, image distortion and image mis-registration, are studied in this article. Spatial encoding of MR images is achieved by a set of magnetic field gradients. The nonlinearity of these gradient fields causes image distortion. This article reports on the measurement of this distortion using a phantom consisting of a 3D rectangular array of point sources and the subsequent correction of this distortion using feature mapping and interpolation. Image distortion in CDI also causes mis-registration of overlying data sets. Mis-registration leads to incorrect computation of current density due to violation of Maxwell's equations. In simulation, mis-registration was also found to cause current density and the curl of current density to exhibit nonzero values in locations where proper registration gives zero current density.