Distortion Correction in Diffusion Magnetic Resonance Imaging

Diffusion magnetic resonance imaging (dMRI) offers unparalleled insight into the microstructural organisation of biological tissues, most notably the white matter tracts of the human brain. However, the standard echo-planar imaging (EPI) sequences employed in dMRI are subject to geometric distortions arising from B₀ field inhomogeneities, eddy currents induced by rapid gradient switching and susceptibility differences at air-tissue interfaces. In addition, subject motion during the acquisition can interact with these artefacts, leading to signal dropout or misalignment that undermines quantitative analyses. Over the past two decades, a variety of retrospective correction strategies have been developed. These include field-mapping techniques that directly estimate B₀ inhomogeneity, reversed phase-encoding or blip-up/blip-down acquisitions that feed into optimisation algorithms, registration-based methods that align diffusion volumes to anatomical references, and model-driven approaches that jointly estimate motion, eddy currents and susceptibility fields. More recently, advances in computational power and algorithm design have enabled real-time or near-real-time correction, as well as integrated frameworks that address within-volume movement. Robust distortion correction underpins accurate estimation of diffusion metrics such as fractional anisotropy and mean diffusivity, and is essential for tractography, connectomics and clinical applications from neurosurgical planning to monitoring of neurodegenerative conditions.

Research from Nature Portfolio

A novel algorithm exploits the fact that a centric EPI acquires opposite phase-encoding polarities within a single k-space. By partitioning this k-space into two halves, the method generates paired images with reversed distortions, which are processed iteratively through automated brain masking and uniformity correction. This single-scan approach produces field maps that correlate closely with those from conventional dual-acquisition topup methods, and it extends seamlessly to multi-shot centric EPI, offering distortion correction without additional acquisitions.

Research from all publishers

PyHySCO introduces a GPU-accelerated hyperelastic susceptibility correction tool that applies a physical distortion model and a Chang–Fitzpatrick-based initialisation to achieve whole-brain EPI correction in seconds. Numerical validation on 3T and 7T data demonstrates accuracy comparable to leading reversed-gradient tools at a fraction of processing time.

An integrated framework has been proposed that models movement as a continuous function across slice acquisitions and simultaneously estimates susceptibility- and eddy-current-induced fields. Intra-volume motion correction within this framework restores the fidelity of diffusion tensor metrics and reduces variability between still and movement-affected scans.

A foundational retrospective method jointly optimises rigid-body motion parameters and higher-order eddy-current distortion models using mutual information registration. By correcting spatial misalignment and recalculating b-matrices in a single step, it achieves significant improvements in diffusion metric accuracy and has been widely adopted in large-scale connectome projects.

Distortion Correction in Diffusion Magnetic Resonance Imaging publication trend

The graph below shows the total number of articles in distortion correction in diffusion magnetic resonance imaging across all publications each year (not limited to Nature Index journals).

Technical terms

Echo-Planar Imaging (EPI): A rapid MRI acquisition technique that collects an entire image in a single excitation by traversing k-space in a zigzag pattern.

Topup Algorithm: An optimisation method that uses pairs of images with reversed phase-encoding directions to estimate and correct susceptibility-induced distortions.

Phase-Encoding Polarity: The direction of gradient blips that encodes spatial information and determines the orientation of distortions in EPI.

Eddy Currents: Spurious magnetic fields induced by rapid gradient switching that cause spatial warping in diffusion-weighted images.

Susceptibility Artefact: Geometric distortion arising from local variations in magnetic susceptibility, often at tissue–air boundaries.

Field Map: A spatial map of B₀ inhomogeneity estimated to correct geometric distortions in MRI.

Fractional Anisotropy: A scalar measure of diffusion directionality reflecting microstructural organisation within each voxel.

References

1. Distortion correction using topup algorithm by single k-space (TASK) for echo planar imaging. Scientific Reports (2023).
2. PyHySCO: GPU-enabled susceptibility artifact distortion correction in seconds. Frontiers in Neuroscience (2024).
3. Towards a comprehensive framework for movement and distortion correction of diffusion MR images: Within volume movement. NeuroImage (2017).
4. Comprehensive approach for correction of motion and distortion in diffusion‐weighted MRI. Magnetic Resonance in Medicine (2003).