Off‐resonance artifact correction for MRI: A review

Haskell, Melissa W.; Nielsen, Jon‐Fredrik; Noll, Douglas C.

doi:10.1002/nbm.4867

Cited by 14 publications

(8 citation statements)

References 120 publications

(253 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Latent signal models improved significantly utilizing a fully sampled phase estimate in comparison to a low‐resolution estimate when reconstructing EPTI data in Figures 8 and S7. Thus, future work will explore improved B 0 estimation and refinement 48,63,64 techniques to improve phase estimates for latent signal model EPTI reconstructions.…”

Section: Discussionmentioning

confidence: 99%

“…Sequences that acquire multiple k‐space lines after excitation or inversion across a GE or spin‐echo train concatenate data from all echoes into a single time‐point, often of dimension

M\times N\times P\times C

, where

M,N,P

represent read‐out, phase‐encode, and partition dimensions, respectively; and

C

represents the number of coils. Reconstructions produce a static image that suffers from blurring 47 or distortion 48 artifacts due to signal decay and phase evolution in the collapsed temporal evolution.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Latent signal models: Learning compact representations of signal evolution for improved time‐resolved, multi‐contrast MRI

Arefeen

Zhang

et al. 2023

Magnetic Resonance in Med

View full text Add to dashboard Cite

Purpose: To improve time-resolved reconstructions by training auto-encoders to learn compact representations of Bloch-simulated signal evolution and inserting the decoder into the forward model. Methods: Building on model-based nonlinear and linear subspace techniques, we train auto-encoders on dictionaries of simulated signal evolution to learn compact, nonlinear, latent representations. The proposed latent signal model framework inserts the decoder portion of the auto-encoder into the forward model and directly reconstructs the latent representation. Latent signal models essentially serve as a proxy for fast and feasible differentiation through the Bloch equations used to simulate signal. This work performs experiments in the context of T 2 -shuffling, gradient echo EPTI, and MPRAGE-shuffling. We compare how efficiently auto-encoders represent signal evolution in comparison to linear subspaces. Simulation and in vivo experiments then evaluate if reducing degrees of freedom by incorporating our proxy for the Bloch equations, the decoder portion of the auto-encoder, into the forward model improves reconstructions in comparison to subspace constraints.Results: An auto-encoder with 1 real latent variable represents single-tissue fast spin echo, EPTI, and MPRAGE signal evolution to within 0.15% normalized RMS error, enabling reconstruction problems with 3 degrees of freedom per voxel (real latent variable + complex scaling) in comparison to linear models with 4-8 degrees of freedom per voxel. In simulated/in vivo T 2 -shuffling and in vivo EPTI experiments, the proposed framework achieves consistent quantitative normalized RMS error improvement over linear approaches. From qualitative evaluation, the proposed approach yields images with reduced blurring and noise amplification in MPRAGE-shuffling experiments. Conclusion:Directly solving for nonlinear latent representations of signal evolution improves time-resolved MRI reconstructions.

show abstract

Section: Discussionmentioning

confidence: 99%

“…Sequences that acquire multiple k‐space lines after excitation or inversion across a GE or spin‐echo train concatenate data from all echoes into a single time‐point, often of dimension

M\times N\times P\times C

, where

M,N,P

represent read‐out, phase‐encode, and partition dimensions, respectively; and

C

Section: Methodsmentioning

confidence: 99%

Latent signal models: Learning compact representations of signal evolution for improved time‐resolved, multi‐contrast MRI

Arefeen

Zhang

et al. 2023

Magnetic Resonance in Med

View full text Add to dashboard Cite

show abstract

“…In fMRI studies, distortion of the source images is transmitted downstream, distorting all derived research findings and clinical maps [1]. Previously state-of-the-art methods employed static distortion correction techniques that depend on the acquisition of a separate field map image [5, 6].…”

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

“…Functional MRI (fMRI) data acquired using echo planar imaging (EPI) sequences are prone to local image distortions due to magnetic field inhomogeneities (B0) arising from differences in magnetic susceptibility, particularly across air-tissue interfaces [1]. The orbitofrontal and inferior temporal cortices suffer the largest distortion due to their proximity to the sinuses, mastoids, and ear canals [2], but distortion is present to varying degrees across the brain.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Framewise multi-echo distortion correction for superior functional MRI

Van,

Montez,

Laumann

et al. 2023

Preprint

View full text Add to dashboard Cite

Functional MRI (fMRI) data are severely distorted by magnetic field (B0) inhomogeneities which currently must be corrected using separately acquired field map data. However, changes in the head position of a scanning participant across fMRI frames can cause changes in the B0 field, preventing accurate correction of geometric distortions. Additionally, field maps can be corrupted by movement during their acquisition, preventing distortion correction altogether. In this study, we use phase information from multi-echo (ME) fMRI data to dynamically sample distortion due to fluctuating B0 field inhomogeneity across frames by acquiring multiple echoes during a single EPI readout. Our distortion correction approach, MEDIC (Multi-Echo DIstortion Correction), accurately estimates B0 related distortions for each frame of multi-echo fMRI data. Here, we demonstrate that MEDIC’s framewise distortion correction produces improved alignment to anatomy and decreases the impact of head motion on resting-state functional connectivity (RSFC) maps, in higher motion data, when compared to the prior gold standard approach (i.e., TOPUP). Enhanced framewise distortion correction with MEDIC, without the requirement for field map collection, furthers the advantage of multi-echo over single-echo fMRI.

show abstract

B₀‐insensitive image navigators for prospective motion‐corrected MRS with localized second‐order shimming

Adanyeguh,

Park,

Henry

et al. 2024

Magnetic Resonance in Med

View full text Add to dashboard Cite

PurposeLocalized shimming in single‐voxel MRS often results in large B0 inhomogeneity outside the volume‐of‐interest. This causes unacceptable degradation in motion navigator images. Switching back and forth between whole‐brain shim and localized shim is possible for linear shims, but not for higher‐order shims. Here we propose motion navigators largely insensitive to B0 inhomogeneity for prospective motion‐corrected MRS with localized higher‐order shimming.MethodsA recent fast high‐resolution motion navigator based on spiral‐in/out k‐space trajectories and multislice‐to‐volume registration was modified by splitting the readout into multiple shot interleaves which shortened the echo time and reduced the effect of B0 inhomogeneity. The performance of motion correction was assessed in healthy subjects in the prefrontal cortex using a sLASER sequence at 3T (N = 5) and 7T (N = 5).ResultsWith multiple spatial interleaves, excellent quality navigator images were acquired in the whole brain in spite of large B0 inhomogeneity outside the MRS voxel. The total duration of the navigator in sLASER remained relatively short even with multiple shots (3T: 10 spatial interleaves 94 ms per slice; 7T: 15 spatial interleaves 103 ms per slice). Prospective motion correction using the multi‐shot navigators yielded comparable spectral quality (water linewidth and metabolite SNR) with and without subject motion.ConclusionB0‐insensitive motion navigators enable prospective motion correction for MRS with all first‐ and second‐order shims adjusted in the MRS voxel, providing optimal spectral linewidth.

show abstract

Off‐resonance artifact correction for MRI: A review

Cited by 14 publications

References 120 publications

Latent signal models: Learning compact representations of signal evolution for improved time‐resolved, multi‐contrast MRI

Latent signal models: Learning compact representations of signal evolution for improved time‐resolved, multi‐contrast MRI

Framewise multi-echo distortion correction for superior functional MRI

B₀‐insensitive image navigators for prospective motion‐corrected MRS with localized second‐order shimming

Contact Info

Product

Resources

About

Off‐resonance artifact correction for MRI: A review

Cited by 14 publications

References 120 publications

Latent signal models: Learning compact representations of signal evolution for improved time‐resolved, multi‐contrast MRI

Latent signal models: Learning compact representations of signal evolution for improved time‐resolved, multi‐contrast MRI

Framewise multi-echo distortion correction for superior functional MRI

B0‐insensitive image navigators for prospective motion‐corrected MRS with localized second‐order shimming

Contact Info

Product

Resources

About

B₀‐insensitive image navigators for prospective motion‐corrected MRS with localized second‐order shimming