Jenny Chen1, Benjamin Ades-aron1, Hong-Hsi Lee1, Durga Kullakanda1, Saurabh Maithani1, Dmitry S. Novikov1, Jelle Veraart1, and Els Fieremans1
1Radiology, NYU School Of Medicine, New York, NY, United States
Synopsis
Diffusion
MRI is prone to various artifacts such as noise, eddy current artifacts, and
Gibbs ringing. This study compares diffusion tensor imaging (DTI) and
diffusional kurtosis imaging (DKI) parameter estimates among healthy subjects
in their 20s to 80s using a minimal diffusion pre-processing approach from
Human Connectome Project (HCP) and two DESIGNER (Diffusion parameter EStImation
with Gibbs and NoisE Removal) pipelines, which corrects for additional imaging
artifacts HCP pipeline does not account for. Our results show that
preprocessing quantitatively impacts parameter estimation as well as alters
observed age correlations.
Introduction
Diffusion
MRI (dMRI) probes brain microstructure non-invasively and offers quantitative
biomarkers that are of interest to study in development and disease. To provide
reliable diffusion parameter estimates, dMRI pre-processing is required and various pipelines have been proposed1,2,3 to correct for typical artifacts including distortion from
EPI, eddy currents, and Gibbs ringing and thermal noise. Such pipelines differ
in terms of algorithmic methods and it remains unclear how
these differences manifest in dMRI parameters and, ultimately, impact population-wise statistical findings.
In
this work, we focus on targeted methods for noise reduction using eigenvalue shrinkage4, and Gibbs5,6 removal. Here we study DTI and DKI parameters in white
matter of 89 subjects to evaluate the impact of three integrated pipelines (Figure 1) on parameter estimation and on age associations, a
well-studied biological effect7,8,9.Methods
A
retrospective IRB-approved study included a cohort of 89 control subjects (61 females)
with ages ranging from 23- to 85-year-old, selected out of a total dataset of 3754
subjects who presented for clinical MRI on a Magnetom Prisma 3T (N=39) or Skyra
3T (N=50). dMRI was acquired as follows: 5 b = 0 images, b = 250 s/mm2
– 4 directions, b = 1000 s/mm2 – 20 directions, b = 2000 s/mm2
– 60 directions, TE = 70ms (N=13) or 95ms (N=76), TR = 3.7s, 50 slices,
resolution = 1.7x1.7x3mm3, 6/8 Partial Fourier. One b = 0 image was
acquired with reverse phase-encoding direction for EPI distortion correction
10,11.
dMRI were
pre-processed using three different pipelines:
-
Human Connectome Project (HCP)1, a standard
processing pipeline which targets eddy current, EPI distortion, and motion
correction;
- DESIGNER (DV1.0)2, a processing
pipeline involving the same correction as HCP pipeline along with Gibbs correction6, denoising12, and Rician bias correction13.
- DESIGNER (DV2.0) pipeline, a further development of DESIGNER (DV1.0) that integrates recent developments in Gibbs correction and denoising: (a) RPG (Removal of Partial-fourier induced Gibbs ringing)5 and (b) MP-PCA nonlocal-patch denoising with eigenvalue shrinkage4.
dMRIs were also
pre-processed using DV1.0 and DV2.0 without Rician bias correction to study
effect of the noise floor on parameter estimates. The
pre-processed dMRI output from each pipeline was used to estimate DTI (MD – mean
diffusivity, RD – radial diffusivity, AD – axial diffusivity, and FA –
fractional anisotropy) and DKI (MK – mean kurtosis, RK – radial kurtosis, AD –
axial kurtosis) parameter maps. Kurtosis tensor was estimated with iteratively
reweighted linear least squares (IRWLLS) fitting, which includes outlier
detection and removal
14. A nonlinear warp was applied to JHU white matter atlas labels
15 to bring the labels to each subject’s FA space using FSL FNIRT
16,17. Mean
regional value in 48 white matter ROIs was extracted from the DKI and DTI
parametric maps, followed by linear correlations with age characterized by
the Pearson correlation coefficient.
Results
Figure
2 shows example parametric maps of AK, MK, and RK for a 27-year-old male. The HCP-derived
kurtosis maps appear most noisy with notable black voxels neighboring corpus
callosum (CC) and cerebrospinal fluid (CSF) boundary, while the DV2.0-derived kurtosis
maps have the least black voxels. In Figure 3, noticeably higher MD, RD, and
lower FA is observed in CC and PLIC using HCP pipeline compared to DV1.0 and
DV2.0 pipeline, while this is not the case for the ACR and the SLF. Figure 4 shows
notably lower MK and RK using HCP in comparison to DV1.0 and DV2.0 pipeline.
Additionally, as shown in Figure 5, we observed significantly decreased AK with
Rician bias correction, while there was not much effect in other metrics. As
for observed correlations with age: all significant correlations were
consistent across all pipelines and expected, except for positive correlation
between AD and age in splenium and body CC and SLF only being observed for HCP.Discussion and Conclusion
Our
results illustrate how preprocessing affects diffusion parameter estimation
quantitatively, as well as change the outcome of observed statistically
significant trends. At the level of parametric maps, the DV2.0 pipeline appears
most robust with the least outliers. This improvement, particularly in the AK,
can be explained by the targeted removal of Gibbs ringing, particularly near
sharp edges such as CC/CSF boundary, even in case of Partial Fourier.
At the population
level, Gibbs ringing may also explain the biases in both DTI and DKI metrics in
CC and PLIC (Figure 3-4), while the (opposite) effect of Rician bias is much
smaller and only noticeable in the AK (Figure 5). Consistent among all
pipelines, positive correlations in RD, MD and negative correlations in FA are
observed with increasing age, which agrees with previous studies on aging using DTI/DKI7,8,9. Remarkably, positive
correlations of AD increasing with age are observed in CC for HCP only, which
seem spurious and related to increased Gibbs ringing, potentially due to age-related
atrophy. On the other hand, differences between DV1.0 and DV2.0 are much smaller,
yet also systematic, potentially due to further differences in Gibbs correction5,6 and improved denoising4.
Future work will explore
reproducibility of these results, comparing DV1.0 and DV2.0 with the same
denoising or Gibbs correction method to distinguish Gibbs correction effects
from denoising effects, and studying correlation effects in grey matter
regions.Acknowledgements
Research was supported by the National Institute of Neurological Disorders and Stroke of the NIH under awards R01 NS088040 and R01 EB027075, and by the Hirschl foundation and was performed at the Center of Advanced Imaging Innovation and Research (CAI2R, www.cai2r.net), a Biomedical Technology Research Center supported by NIBIB with the award P41 EB017183.
References
- Glasser MF, Sotiropoulos SN, Wilson JA, Coalson
TS, Fischl B, Andersson JL, Xu J, Jbabdi S, Webster M, Polimeni JR, Van Essen
DC, Jenkinson M; WU-Minn HCP Consortium. The minimal preprocessing pipelines
for the Human Connectome Project. Neuroimage. 2013 Oct 15;80:105-24. doi:
10.1016/j.neuroimage.2013.04.127. Epub 2013 May 11. PMID: 23668970; PMCID:
PMC3720813.
- Ades-Aron, B., Veraart, J., Kochunov, P.,
McGuire, S., Sherman, P., Kellner, E., Novikov, D. S., & Fieremans, E.
(2018). Evaluation of the accuracy and precision of the diffusion parameter
EStImation with Gibbs and NoisE removal pipeline. NeuroImage, 183, 532–543. https://doi.org/10.1016/j.neuroimage.2018.07.066
- Pierpaoli C, Irfanoglu LW,MO, Barnett A, Basser P, Chang L-C, Koay C, Pajevic S, Rohde G, Sarlls J, and Wu M, 2010. TORTOISE: an integrated software package for processing of diffusion MRI data ISMRM, Stockholm, Sweden.
- M. Gavish and D. L. Donoho, "Optimal Shrinkage of Singular Values," in IEEE Transactions on Information Theory, vol. 63, no. 4, pp. 2137-2152, April 2017, doi: 10.1109/TIT.2017.2653801.5.
- Lee, H. H., Novikov, D. S., & Fieremans, E. Removal of Partial Fourier-Induced Gibbs (RPG) Ringing artifacts in MRI. Magnetic Resonance in Medicine. 2020. Under review.
- Kellner E, Dhital B, Kiselev VG, Reisert M. Gibbs-ringing artifact removal based on local subvoxel-shifts. Magn Reson Med. 2016 Nov;76(5):1574-1581. doi: 10.1002/mrm.26054. Epub 2015 Nov 24. PMID: 26745823.
- Maximov, II, Alnæs, D, Westlye, LT. Towards an optimised processing pipeline for diffusion magnetic resonance imaging data: Effects of artefact corrections on diffusion metrics and their age associations in UK Biobank. Hum Brain Mapp. 2019; 40: 4146– 4162. https://doi.org/10.1002/hbm.24691
- Falangola, M. F., Jensen, J. H., Babb, J. S., Hu, C., Castellanos, F. X., Di Martino, A., Ferris, S. H., & Helpern, J. A. (2008). Age-related non-Gaussian diffusion patterns in the prefrontal brain. Journal of magnetic resonance imaging : JMRI, 28(6), 1345–1350. https://doi.org/10.1002/jmri.21604
- Beck D, de Lange AG, Maximov II, Richard G, Andreassen OA, Nordvik JE, Westlye LT. White matter microstructure across the adult lifespan: A mixed longitudinal and cross-sectional study using advanced diffusion models and brain-age prediction. Neuroimage. 2020 Oct 9;224:117441. doi: 10.1016/j.neuroimage.2020.117441. Epub ahead of print. PMID: 33039618.
- J.L.R. Andersson, S. Skare, J. Ashburner. How to correct susceptibility distortions in spin-echo echo-planar images: application to diffusion tensor imaging. NeuroImage, 20(2):870-888, 2003.
- S.M. Smith, M. Jenkinson, M.W. Woolrich, C.F. Beckmann, T.E.J. Behrens, H. Johansen-Berg, P.R. Bannister, M. De Luca, I. Drobnjak, D.E. Flitney, R. Niazy, J. Saunders, J. Vickers, Y. Zhang, N. De Stefano, J.M. Brady, and P.M. Matthews. Advances in functional and structural MR image analysis and implementation as FSL. NeuroImage, 23(S1):208-219, 2004.
- Veraart, J., Novikov, D. S., Christiaens, D., Ades-Aron, B., Sijbers, J., & Fieremans, E. (2016). Denoising of diffusion MRI using random matrix theory. NeuroImage, 142, 394–406. https://doi.org/10.1016/j.neuroimage.2016.08.016
- Koay CG, Basser PJ, 2006. Analytically exact correction scheme for signal extraction from noisy magnitude MR signals. Journal of Magnetic Resonance 179, 317–322.
- J. Veraart, J. Sijbers, S. Sunaert, A. Leemans, B. Jeurissen, Weighted linear least squares estimation of diffusion MRI parameters: strengths, limitations, and pitfalls. NeuroImage 81, 335-346 (2013).
- Hua et al., Tract probability maps in stereotaxic spaces: analysis of white matter anatomy and tract-specific quantification. NeuroImage, 39(1):336-347 (2008)
- Andersson JLR, Jenkinson M, Smith S (2010) Non-linear registration, aka spatial normalisation.
- M.W. Woolrich, S. Jbabdi, B. Patenaude, M. Chappell, S. Makni, T. Behrens, C. Beckmann, M. Jenkinson, S.M. Smith. Bayesian analysis of neuroimaging data in FSL. NeuroImage, 45:S173-86, 2009