Correcting eddy current and motion effects by affine whole‐brain registrations: Evaluation of three‐dimensional distortions and comparison with slicewise correction

Abstract

Eddy-current (EC) and motion effects in diffusion-tensor imaging (DTI) bias the estimation of quantitative diffusion indices, such as the fractional anisotropy. Both effects can be retrospectively corrected by registering the strongly distorted diffusion-weighted images to less-distorted T2-weighted images acquired without diffusion weighting. Two different affine spatial transformations are usually employed for this correction: slicewise and whole-brain transformations. However, a relation between estimated transformation parameters and EC distortions has not been established yet for the latter approach. In this study, a novel diffusion-gradient-direction-independent estimation of the EC field is proposed based solely on affine whole-brain registration parameters. Using this model, it is demonstrated that a more distinct evaluation of the wholebrain EC effects is possible if the through-plane distortion was considered in addition to the well-known in-plane distortions. Moreover, a comparison of different whole-brain registrations relative to a slicewise approach is performed, in terms of the relative tensor error. Our findings suggest that for appropriate intersubject comparison of DTI data, a whole-brain registration containing nine affine parameters provides comparable performance (between 0 and 3%) to slicewise methods and can be performed in a fraction of the time.Magn strong gradients used for diffusion encoding (in the following referred to as "diffusion gradients") create ECs in the conductive parts of the magnet, causing geometric distortions in the DW images that depend on the amplitude and direction of the diffusion gradients (11-14). The EC effects depend on the pulse sequence (15-18), scanner calibration, and scanner design. The EC and motion effects can be retrospectively corrected by registering the strongly distorted DW images to the less distorted images acquired without diffusion weighting (14,19). However, the EC effects cannot be completely corrected by retrospective registration. Thus, after retrospective EC correction residual image artifacts (e.g., image blurring caused by the time-dependent components of the EC field) will remain and reduce the comparability of DTI data. This is particularly important for longitudinal DTI studies, during which the scanner adjustment and, thus, the EC effects might vary.One possibility for assessing the comparability of DTI data is to quantify EC effects by estimating the induced EC field. Rohde et al. (20) presented a new registration approach for EC and motion correction, using a whole-brain transformation. They showed that the EC field induced by one diffusion gradient can be retrospectively estimated from the transformation parameters of a 14-parameter, nonlinear transformation that corrects the distortions of the corresponding DW image. The DTI data, however, consists of at least six DW images, each corresponding to different diffusion directions and containing EC distortions that are corrected by a specific set of EC parameters. Each respective set ...

Analysis of White Matter Damage in Patients with Multiple Sclerosis via a Novel In Vivo MR Method for Measuring Myelin, Axons, and G-Ratio

“…23 Images were then denoised by using multishell position-orientation adaptive smoothing based on the propagation-separation approach. 24,25 This algorithm does not assume a specific model and can be used for any diffusion model including NODDI.…”

Neurodegeneration in the Spinal Ventral Horn Prior to Motor Impairment in Cervical Spondylotic Myelopathy

“…27 For diffusion data, we used the ACID toolbox (www.diffusiontools.com) for eddy current, motion, and physiological artefacts correction 28,29 and for robust tensor fitting 29 . Mean diffusivity (MD) and fractional anisotropy (FA) were calculated and normalized to the MNIPoly-AMU template 30 using the FA voxel-based statistics toolbox.…”

Neuroimaging in Susac's syndrome: Focus on DTI

“…The diffusion weighted images were normalized by the "Muenster Neuroimaging Evaluation System" (Eval 3.0) by means of the "SPM normalization toolbox for voxel-based statistics on fractional anisotropy images" [9]. The Eval 3.0 system provides an optimized image pre-processing, including a newly developed eddy current correction for diffusion weighted MR images [10]. The voxelby-voxel FA and MD differences were evaluated using statistical parametric mapping (SPM8, www.fil.ion.ucl.ac.uk/spm/software/ spm8/).…”