磁共振成像圖像的準確性受到磁場不均勻引起的失真問題影響。為了解決這些失真問題，許多EPI校正方法被廣泛應用。傳統的EPI校正方法通常需要使用參考圖像，例如反相圖像或額外的校準掃描，以校正EPI序列中的失真。然而，這些方法的缺點是需要額外的掃描，從而增加了成像的時間和複雜性。
本研究中，我們提出了一種名為Merge-D的深度學習方法，旨在解決EPI失真問題。該方法不僅不需額外的參考圖像，還能夠自動檢測EPI校正的失真方向。我們的模型通過以下四個步驟來實現EPI校正：（1）輸入失真影像：三維的DWI影像，（2）失真方向感知：包括AP-PA和LR-RL，（3）估計位移場：用於表示每個像素位移量大小及（4）位移場應用：將位移場應用於失真的DWI圖像，以獲得校正後的EPI圖像。通過跨機構驗證，我們的模型不僅能夠準確識別失真方向，還能顯著減少EPI圖像中的幾何失真，並改善與高分辨率T1w圖像的對位。這些結果表明，我們的方法在提高MRI圖像準確性方面具有潛力，同時降低了成像過程的複雜性和時間成本。

EPI correction addresses distortion issues caused by magnetic field inhomogeneity, ensuring the accuracy and reliability of MRI images. Traditional EPI correction methods rely on independent reference images, such as reverse phase images or additional calibration scans, to correct distortions in EPI sequences. However, these methods require additional calibration scans or reference images, increasing imaging time and complexity. In this study, we propose a deep learning method called Merge-D to address EPI distortion problems. This method offers a non-reference image and automatically detects distortion direction for EPI correction. The EPI correction process in the network model involves four steps: (1) Inputting three-dimensional DWI data, (2) Detection of distortion direction, including AP-PA and LR-RL, (3) Estimating the displacement field that represents the magnitude of displacement for each pixel in the image and (4) Applying the displacement field to unwrap the distortion in the distorted DWI images results in corrected EPI images.Through cross-institutional validation, the results demonstrate the accurate identification of distortion directions, reduction in geometric distortions in EPI images, and improved registration with high-resolution T1w images.

[1] S. Warach, J. Gaa, B. Siewert, P. Wielopolski, and R. R. Edelman, "Acute human stroke studied by whole brain echo planar diffusion-weighted magnetic resonance imaging," Ann Neurol, vol. 37, no. 2, pp. 231-41, Feb 1995, doi: 10.1002/ana.410370214.
[2] P. J. Basser, J. Mattiello, and D. LeBihan, "Estimation of the effective self-diffusion tensor from the NMR spin echo," J Magn Reson B, vol. 103, no. 3, pp. 247-54, Mar 1994, doi: 10.1006/jmrb.1994.1037.
[3] B. Biswal, F. Z. Yetkin, V. M. Haughton, and J. S. Hyde, "Functional connectivity in the motor cortex of resting human brain using echo-planar MRI," Magn Reson Med, vol. 34, no. 4, pp. 537-41, Oct 1995, doi: 10.1002/mrm.1910340409.
[4] F. Farzaneh, S. J. Riederer, and N. J. Pelc, "Analysis of T2 limitations and off-resonance effects on spatial resolution and artifacts in echo-planar imaging," Magn Reson Med, vol. 14, no. 1, pp. 123-39, Apr 1990, doi: 10.1002/mrm.1910140112.
[5] C. Liu, M. E. Moseley, and R. Bammer, "Simultaneous phase correction and SENSE reconstruction for navigated multi-shot DWI with non-cartesian k-space sampling," Magn Reson Med, vol. 54, no. 6, pp. 1412-22, Dec 2005, doi: 10.1002/mrm.20706.
[6] K. P. Pruessmann, M. Weiger, M. B. Scheidegger, and P. Boesiger, "SENSE: sensitivity encoding for fast MRI," Magn Reson Med, vol. 42, no. 5, pp. 952-62, Nov 1999.
[7] D. K. Sodickson and W. J. Manning, "Simultaneous acquisition of spatial harmonics (SMASH): fast imaging with radiofrequency coil arrays," Magn Reson Med, vol. 38, no. 4, pp. 591-603, Oct 1997, doi: 10.1002/mrm.1910380414.
[8] M. A. Griswold et al., "Generalized autocalibrating partially parallel acquisitions (GRAPPA)," Magn Reson Med, vol. 47, no. 6, pp. 1202-10, Jun 2002, doi: 10.1002/mrm.10171.
[9] P. Jezzard and R. S. Balaban, "Correction for geometric distortion in echo planar images from B0 field variations," Magn Reson Med, vol. 34, no. 1, pp. 65-73, Jul 1995, doi: 10.1002/mrm.1910340111.
[10] D. Holland, J. M. Kuperman, and A. M. Dale, "Efficient correction of inhomogeneous static magnetic field-induced distortion in Echo Planar Imaging," Neuroimage, vol. 50, no. 1, pp. 175-83, Mar 2010, doi: 10.1016/j.neuroimage.2009.11.044.
[11] M. O. Irfanoglu, P. Modi, A. Nayak, E. B. Hutchinson, J. Sarlls, and C. Pierpaoli, "DR-BUDDI (Diffeomorphic Registration for Blip-Up blip-Down Diffusion Imaging) method for correcting echo planar imaging distortions," Neuroimage, vol. 106, pp. 284-99, Feb 1 2015, doi: 10.1016/j.neuroimage.2014.11.042.
[12] J. L. Andersson, S. Skare, and J. Ashburner, "How to correct susceptibility distortions in spin-echo echo-planar images: application to diffusion tensor imaging," Neuroimage, vol. 20, no. 2, pp. 870-888, 2003.
[13] S. T. M. Duong, S. L. Phung, A. Bouzerdoum, and M. M. Schira, "An unsupervised deep learning technique for susceptibility artifact correction in reversed phase-encoding EPI images," Magn Reson Imaging, vol. 71, pp. 1-10, Sep 2020, doi: 10.1016/j.mri.2020.04.004.
[14] A. Babayan et al., "A mind-brain-body dataset of MRI, EEG, cognition, emotion, and peripheral physiology in young and old adults," Sci Data, vol. 6, p. 180308, Feb 12 2019, doi: 10.1038/sdata.2018.308.
[15] L. M. Alexander et al., "An open resource for transdiagnostic research in pediatric mental health and learning disorders," Sci Data, vol. 4, p. 170181, Dec 19 2017, doi: 10.1038/sdata.2017.181.
[16] M. F. Glasser et al., "The minimal preprocessing pipelines for the Human Connectome Project," Neuroimage, vol. 80, pp. 105-24, Oct 15 2013, doi: 10.1016/j.neuroimage.2013.04.127.