Conductivity reconstructions with polynomial fitting using subvoxel FOV shifts

Motivation: Conductivity reconstructions are highly sensitive to noise and errors at tissue interfaces due to the calculation of derivatives (e.g., using finite differences kernels). These lead to poor quality conductivity maps. Goal(s): Increase spatial sampling of the B1+ field by subvoxel field-of-view shifts to improve fitting-based conductivity reconstructions. Approach: Acquisition of multiple low resolution scans, shifted in one direction with a subvoxel-shift. The increased spatial sampling is used to improve reconstructions, while the lower resolution is needed for increased SNR. Results: The results show an improved noise robustness in comparison to conventional sampling, enabling the use of smaller kernel sizes, which increases effective resolution. Impact: With multiple low-resolution Cartesian acquisitions using subvoxel-shifts, reconstructed conductivity is more robust to noise, in comparison to conventional methods for similar scan times. The higher quality reconstructions can be important for clinical application of EPT in the future.