Related Experiment Video
Updated: Jan 9, 2026
High-resolution Functional Magnetic Resonance Imaging Methods for Human Midbrain
Published on: May 10, 2012
Motion-Informed Deep Learning for Human Brain Magnetic Resonance Image Reconstruction Framework
Zhifeng Chen1,2,3, Kamlesh Pawar1, Kh Tohidul Islam1
1Monash Biomedical Imaging, Monash University, Clayton, Victoria, Australia.
Abstract:
Motion artifacts in magnetic resonance imaging (MRI) are one of the frequently occurring artifacts due to patient movements during scanning. Motion is estimated to be present in approximately 30% of clinical MRI scans; however, motion has not been explicitly modeled within deep learning image reconstruction models. Deep learning (DL) algorithms have been demonstrated to be effective for both the image reconstruction task and the motion correction task, but the two tasks are considered separately. The image reconstruction task involves removing undersampling artifacts such as noise and aliasing artifacts, whereas motion correction involves removing artifacts including blurring, ghosting, and ringing. In this work, we propose a novel method to simultaneously accelerate imaging and correct motion. This is achieved by integrating a motion module into the DL-based MRI reconstruction process, enabling detection and correction of motion. We model motion as a tightly integrated auxiliary layer in the DL model during training, making the DL model "motion-informed". During inference, image reconstruction is performed from undersampled raw k-space data using a trained motion-informed DL model. Experimental results demonstrate that the proposed motion-informed DL image reconstruction network outperformed the conventional image reconstruction network for motion-degraded MRI datasets.
Insights
This study introduces a novel deep learning method to simultaneously accelerate magnetic resonance imaging (MRI) and correct motion artifacts. The "motion-informed" deep learning model improves image quality in scans with patient movement.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Biomedical Engineering
Background:
- Motion artifacts affect approximately 30% of clinical MRI scans, degrading image quality.
- Current deep learning models address image reconstruction and motion correction separately.
- Existing methods fail to explicitly model patient motion within deep learning reconstruction frameworks.
Purpose of the Study:
- To develop a novel deep learning method for simultaneous MRI acceleration and motion artifact correction.
- To integrate motion detection and correction directly into the deep learning reconstruction process.
- To create a "motion-informed" deep learning model for enhanced MRI data acquisition.
Main Methods:
- A novel deep learning architecture was developed, integrating a motion module as an auxiliary layer.
- The model was trained to be "motion-informed," enabling it to learn and correct for motion during reconstruction.
- Image reconstruction was performed using undersampled k-space data with the trained motion-informed deep learning model.
Main Results:
- The proposed motion-informed deep learning network demonstrated superior performance compared to conventional reconstruction methods.
- Experimental results confirmed the network's effectiveness in reconstructing high-quality MRI data from motion-degraded datasets.
- The method successfully addressed undersampling artifacts alongside motion-induced artifacts like blurring, ghosting, and ringing.
Conclusions:
- The developed motion-informed deep learning approach effectively corrects motion artifacts during MRI reconstruction.
- This integrated method offers a promising solution for accelerating MRI scans while maintaining high image quality.
- The findings suggest a new paradigm for deep learning-based MRI reconstruction that accounts for patient motion.
