Date of Award
Doctor of Philosophy (PhD)
Functional MRI (fMRI) data, acquired through echo-planar imaging (EPI), are consistently distorted by main (B0) magnetic field inhomogeneities which are driven by differences in tissue susceptibility. Traditionally, these distortions are corrected using separately acquired field map data, which provide an estimate of the B0 inhomogeneity. However, factors such as head position changes during scans, corrupted field maps due to movement, or even the absence of field map data can impede this correction process. In this work, we present two new solutions to address these challenges and evaluated them using several quality metrics (Chapter 2). First, Synth, a field map-less correction method, has been developed to facilitate distortion correction and cross-modal image registration without the need for field map data (Chapter 3, published as Montez, D.F, Van, A.N et al., Using synthetic MR images for distortion correction). By leveraging T1w and T2w anatomical images, Synth constructs a synthetic reference image resembling the contrast properties of EPI data. This provides an individual-specific distortion-free reference for distortion correction. Using a multiple fMRI data sets, we demonstrate that Synth performs on par with, or even surpass, conventional field map correction techniques. Secondly, we present another distortion correction method, MEDIC (Multi-Echo DIstortion Correction), which harnesses phase information from multi-echo (ME) fMRI data to dynamically account for distortion arising from fluctuating magnetic field inhomogeneity across an fMRI time series (Chapter 4). With the ability to accurately estimate distortions at each frame of an EPI image, MEDIC offers enhanced alignment to anatomy and diminishes the effect of head motion on resting-state functional connectivity maps. MEDIC furthers the advantage of multi-echo fMRI over single-echo fMRI and removes the need for an additional field map scan during data acquisition. Our findings indicate that both Synth and MEDIC improve the quality of fMRI studies through superior distortion correction, in addition to providing fMRI researchers with greater flexibility in their data acquisition and analysis.
Available for download on Monday, November 10, 2025