Reducing susceptibility distortion related image blurring in diffusion MRI EPI data

Abstract

Diffusion magnetic resonance imaging (MRI) is an increasingly popular technique in basic and clinical neuroscience. One promising application is to combine diffusion MRI with myelin maps from complementary MRI techniques such as multi-parameter mapping (MPM) to produce g-ratio maps that represent the relative myelination of axons and predict their conduction velocity. Statistical Parametric Mapping (SPM) can process both diffusion data and MPMs, making SPM the only widely accessible software that contains all the processing steps required to perform group analyses of g-ratio data in a common space. However, limitations have been identified in its method for reducing susceptibility-related distortion in diffusion data. More generally, susceptibility-related image distortion is often corrected by combining oppositely phase-encoded images (blip-up and blip-down) using the arithmetic mean (AM), however, this can lead to blurred images. In this study we sought to (1) improve the susceptibility-related distortion correction for diffusion MRI data in SPM; (2) deploy an alternative approach to the AM to reduce image blurring in diffusion MRI data when combining blip-up and blip-down EPI data after susceptibility-related distortion correction; and (3) assess the benefits of these changes for g-ratio mapping. We found that the new processing pipeline, called consecutive HySCO, improved distortion correction when compared to the standard approach in the ACID toolbox for SPM. Moreover, using a weighted average (WA) method to combine the distortion corrected data from each phase-encoding polarity achieved greater overlap of diffusion and more anatomically faithful structural white matter probability maps derived from minimally distorted multi-parameter maps as compared to the AM. Third, we showed that the consecutive HySCO WA performed better than the AM method when combined with multi-parameter maps to perform g-ratio mapping. These improvements mean that researchers can conveniently access a wide range of diffusion-related analysis methods within one framework because they are now available within the open-source ACID toolbox as part of SPM, which can be easily combined with other SPM toolboxes, such as the hMRI toolbox, to facilitate computation of myelin biomarkers that are necessary for g-ratio mapping.

Conduction velocity along a key white matter tract is associated with autobiographical memory recall ability

“…Diffusion data are also susceptible to distortions which can particularly affect the uncinate fasciculus. We addressed this challenge by using a new technique that improves distortion correction in this region (Clark et al, 2021b). The absence of fornix findings in our study could echo those relating to hippocampal volume (Clark et al, 2020; Clark et al, 2021a), whereby the structure is widely acknowledged to be involved in autobiographical memory recall, and damage impedes retrieval (Aggleton et al, 2000; D’Esposito et al, 1995; Gaffan and Gaffan, 1991; Tsivilis et al, 2008), but microstructural variations have limited impact within the healthy adult population.…”

“…The diffusion MRI data were processed using the ACID toolbox (www.diffusiontools.com) within SPM12 (www.fil.ion.ucl.ac.uk/spm). The weighted average consecutive HySCO pipeline described in Clark et al (2021b) was followed, with the addition of multi-shell Position-Orientation Adaptive Smoothing (msPOAS; Becker et al, 2012) and Rician bias correction (André et al, 2014). In brief, the blip-up and blip-down data were first separately corrected for motion and eddy current artefacts.…”

Reducing Susceptibility Distortion Related Image Blurring in Diffusion MRI EPI Data

“…We would also like to thank Lars Rutthotto for helpful discussions. This manuscript has been released as a pre-print at Biorxiv (Clark et al, 2021).…”