RESUMEN
PURPOSE: Knowledge on low frequency (LF) tissue conductivity is relevant for various biomedical purposes. To obtain this information, LF phase maps arising from time-varying imaging gradients have been demonstrated to create a LF conductivity contrast. Essential in this methodology is the subtraction of phase images acquired with opposite gradient polarities to separate LF and RF phase effects. Here we demonstrate how sensitive these subtractions are with respect to geometrical distortions. THEORY AND METHODS: The effect of geometrical distortions on LF phase maps is mathematically defined. After quantifying typical geometrical distortions, their effects on LF phase maps are evaluated using conductive phantoms. For validation, electromagnetic simulations of LF phase maps were performed. RESULTS: Even sub-voxel distortions of 10% of the voxel size, measured for a typical LF MR sequence, cause leakage of RF phase into LF phase of several milli-radians, leading to a misleading pattern of LF phase maps. This leakage is mathematically confirmed, while simulations indicate that the expected LF phase should be in order of micro-radians. CONCLUSION: The conductivity scaling of LF phase maps is attributable to the RF phase leakage, thus dependent on the RF conductivity. In fact, simulations show that the LF phase is not measurable. Magn Reson Med 76:905-912, 2016. © 2015 Wiley Periodicals, Inc.