| dc.description.abstract | Distortion and other artefacts caused by an uneven magnetic field affect MR images acquired with the rapid technique echo-planar imaging (EPI). This study investigates the effectiveness of reverse gradient susceptibility correction methods on spin echo (SE) and gradient echo (GE) EPI. In addition, the effects of varying bandwidth, SENSE factor and slice thickness on the corrected images were measured. Undistorted, anatomically accurate images are necessary in relating EPI images to anatomical MRI scans. This is particularly important for diffusion weighted imaging, perfusion imaging and fMRI. Due to the ever-increasing need for better temporal resolution with existing hardware, effective artefact correction in post-processing is necessary, as is appropriate sequence optimisation. This investigation was accomplished using normalised mutual information to compare the images pre and post correction to an anatomical image as a measure of anatomical accuracy. It was found that the correction of geometric distortions is equally effective in GE-EPI and SE-EPI, and that geometric correction can reduce the impact of the aforementioned scan parameters on the anatomical accuracy of the images. Three correction methods were compared in this investigation; FSL TOPUP, EPI-EPIC and ACID HySCO. It was found that, in most situations, the EPIC susceptibility correction tool was the most effective tool at correcting geometric distortions without the generation of new artefacts. Susceptibility artefact corrections were also successfully conducted on 7T contrast enhanced single shot EPI images of a prostate tumour xenograft (nude mouse). These findings could help establish clinical and research protocol for susceptibility artefact post-processing, as well as sequence optimisation. | eng |