Referenceless distortion correction of gradient-echo echo-planar imaging under inhomogeneous magnetic fields based on...

Area of Science:

• Medical Imaging
• Neuroimaging
• Magnetic Resonance Imaging (MRI)

Background:

• Single-shot gradient-echo echo-planar imaging (GE-EPI) is crucial for high temporal resolution imaging.
• GE-EPI is prone to image distortions caused by magnetic field inhomogeneities.
• Current correction methods often require additional data acquisition or are ineffective at high magnetic fields.

Purpose of the Study:

• To develop a novel algorithm for correcting GE-EPI image distortions.
• To address the limitations of existing distortion correction techniques.
• To improve the accuracy and efficiency of GE-EPI imaging.

Main Methods:

• A deep convolutional neural network (CNN) was developed for distortion correction.
• Residual learning and a cascaded structure were incorporated to enhance CNN performance.
• The CNN was trained using a simulated dataset.

Main Results:

• The proposed CNN algorithm effectively corrected image distortions caused by magnetic field inhomogeneity.
• Simulated and experimental results validated the algorithm's performance.
• The method successfully corrected distortions without requiring additional acquisitions.

Conclusions:

• The developed CNN-based algorithm offers a robust solution for GE-EPI distortion correction.
• This approach eliminates the need for extra field mapping scans.
• The method holds promise for improving the reliability of high temporal resolution MRI applications.