Motion Artifact-Aware Self-Supervised Representation Learning for 3D Brain MRI Motion Artifact Reduction

2026-08-10

AI Summary

This work addresses signal loss, blurring, and geometric distortions in brain MRI caused by patient motion by proposing the SSRL-MAR framework, which achieves motion artifact correction through a three-stage self-supervised learning strategy without requiring paired clean-corrupted data or k-space information. The method first extracts motion representations via 3D patch-based contrastive learning, then constructs a motion artifact synthesis network, and finally leverages the learned degradation model to guide a generator in restoring high-quality images. Notably, it is the first approach to enable motion-aware self-supervised representation learning and image-domain artifact removal without motion labels or paired data, further enhanced by unsupervised domain adaptation for improved generalization. Experiments show that on simulated data, it achieves a PSNR of 23.81 dB and SSIM of 91.55%; on real MR-ART data, domain adaptation yields a 2.0 dB PSNR gain and reduces volumetric errors in key brain structures by over 50%, approaching the performance of supervised methods that rely on paired data.

Abstract

Patient motion remains a source of image degradation in brain MRI, leading to signal loss, blurring, and geometric distortion that compromise quantitative analysis. Existing deep learning methods for motion correction typically rely on paired clean-corrupted data or k-space acquisitions, which are rarely available in clinical settings. We propose SSRL-MAR, a motion artifact-aware unpaired representation learning framework for motion artifact reduction that requires neither paired training data nor explicit motion labels. SSRL-MAR employed a three-stage training strategy: (1) contrastive learning on 3D patches to extract motion representations by contrasting clean and synthetically corrupted images, (2) a motion artifact-aware synthesis network to generate motion artifacts from clean scans, and (3) a motion artifact-aware generator to restore clean volumes using the learned degrader for self-supervised supervision. On in-silico dataset, SSRL-MAR achieved PSNR 23.81dB, SSIM 91.55%, and NMSE 0.79%. On in-vivo MR-ART dataset, the pretrained model reduced motion distortion, and unsupervised domain adaptation further improved anatomical fidelity. Against a source-only supervised model trained on the same simulated pairs, SSRL-MAR improved PSNR by up to 2.0 dB on MR-ART after unsupervised domain adaptation, and remained within 0.25-0.47 dB of an oracle supervised model that requires real paired data unavailable in practice. At the milder motion level, volumetric error in structures such as the corpus callosum and ventricular system decreased by more than 50%, confirming improved neuroanatomical consistency. These results indicate that SSRL-MAR provides a robust and scalable image-domain solution for 3D brain MRI motion correction, enabling reliable structural quantification in large-scale neuroimaging studies without requiring prospectively acquired pairs or acquisition-specific calibration.

Problem

• motion artifact
• 3D brain MRI
• image degradation
• quantitative analysis
• patient motion

Innovation

• self-supervised learning
• motion artifact reduction
• unpaired representation learning
• 3D brain MRI
• contrastive learning