New MRI Sequence Enables Efficient Distortion Correction for Ultra-High-Resolution Brain Imaging
19th June 2026
Seong Dae Yun*, Patricia Pais-Roldán, Jeongbeen Lee and N. Jon Shah
Functional MRI (fMRI) commonly uses echo-planar imaging (EPI), a fast imaging method that is highly susceptible to geometric distortions caused by magnetic field inhomogeneities. These distortions can make it difficult to accurately align functional images with high-resolution anatomical scans, limiting the precise localisation of brain activity.
One way to overcome this challenge is to acquire anatomical images using an EPI-based sequence, known as MP2EPI, which shares the same distortion pattern as functional scans and therefore allows straightforward image alignment. However, because the anatomical images, measured by MP2EPI, are also distorted, they require additional scans with reversed phase encoding to correct these distortions, doubling the overall imaging time.
INM-4 researchers have now developed a new MP2EPI sequence that simultaneously acquires the reversed phase-encoding data needed for distortion correction within a single acquisition. The sequence was technically implemented and validated for submillimetre-resolution fMRI at 7 Tesla (7T), where image distortions can be particularly challenging. The approach is intended to facilitate increased anatomical fidelity which is comparable to standard MP2RAGE in the submillimetre fMRI analysis, while maintaining an efficient acquisition protocol.

The figure above shows the results of the anatomical ROI parcellation. The parcellated cortical ROIs, obtained from uncorrected (top) and distortion-corrected (bottom) MP2EPI scans, are overlaid on the MP2RAGE scan. The anatomical ROIs, shown in distinct colours, are more accurately aligned with the MP2RAGE scan in the distortion-corrected case when compared to the uncorrected one. This is particularly evident in the indicated regions: the superior frontal, middle frontal, precentral, and cingulate gyri.
This work provides a practical advance for ultra-high-field MRI and may support future high-resolution functional imaging studies by simplifying distortion correction without requiring additional scan time.
Original publication