Multi-Contrast MRI Motion Correction via Parameter-Informed Disentanglement and Adaptive Experts

2026-05-29 · Source: Artificial Intelligence · Field: Science & Research — Health & Medical Research, Mathematics & Computational Sciences · Depth: Expert, quick

Summary

A novel unified framework addresses motion artifacts in multi-contrast MRI by combining parameter-informed contrast disentanglement with severity-aware adaptive correction. This method, detailed in "Multi-Contrast MRI Motion Correction via Parameter-Informed Disentanglement and Adaptive Experts," utilizes ScanCLIP, pretrained on over 30,000 MRI text-image pairs, to derive contrast embeddings from acquisition parameters, effectively disentangling contrast style from anatomical content. A Vision Transformer then estimates motion severity, routing features through a Mixture-of-Experts network for targeted artifact correction. A dual-pathway decoder reconstructs both the clean image and a residual artifact map, ensuring image-space consistency. On IXI and HCP benchmarks, the framework improves PSNR by 0.75 dB and SSIM by up to 0.0279 over existing approaches, showing larger gains at higher artifact severities and robust zero-shot generalization on unseen clinical data.

Key takeaway

For Machine Learning Engineers developing MRI reconstruction models, this framework offers a robust solution for motion artifact correction, particularly for multi-contrast and diverse severity scenarios. You should consider integrating parameter-informed disentanglement and severity-aware adaptive experts to enhance model generalization. This approach can significantly improve image quality, evidenced by PSNR gains of 0.75 dB, reducing the need for contrast-specific model retraining and improving diagnostic reliability in clinical applications.

Key insights

A unified MRI motion correction framework disentangles contrast and adapts to severity for robust, generalizable artifact removal.

Principles

• Disentangle contrast style from content.
• Adapt correction based on motion severity.
• Enforce image-space consistency via dual-pathway.

Method

ScanCLIP extracts contrast-free features; a Vision Transformer estimates motion severity; a Mixture-of-Experts network applies targeted correction; a dual-pathway decoder reconstructs the image and artifact map.

In practice

• Improve diagnostic reliability in MRI.
• Generalize across diverse MRI modalities.
• Correct artifacts in real-world clinical data.

Topics

• Magnetic Resonance Imaging
• Motion Correction
• Deep Learning
• Vision Transformers
• Contrast Disentanglement
• Medical Imaging

Best for: AI Scientist, Machine Learning Engineer, Research Scientist

Related on AIssential

• GloResNet: A lightweight 3D CNN with global topological features for preterm brain injury prediction — Lightweight 3D CNN with global topological features predicts preterm brain injury effectively.
• AI for Interoperability in Health Care: Philips’s Carla Goulart Peron — AI in healthcare augments clinicians to expand access and improve diagnostics, with interoperability identified as the most impactful global capability.
• China has approved the world’s first invasive brain-computer chip—here’s what’s next — China's NEO BCI, less invasive and government-backed, is the first globally approved for commercial use, offering hope for paralysis patients.
• ClinEnv: An Interactive Multi-Stage Long Horizon EHR Environment for Agents — ClinEnv reveals LLMs struggle with sequential, uncertain clinical decisions and information gathering in a multi-stage environment.
• Quantitative Movement Testing: Measuring Patient Movements from a Single Smartphone Video — QMT uses smartphone video and deep learning for accessible, accurate 3D kinematic biomarker extraction in clinical settings.
• Predicting the risk of colorectal anastomotic leak based on preoperative mapping of the blood supply of the bowel — An AI-driven system predicts colorectal anastomotic leak risk using CT imaging and deep learning for enhanced surgical planning.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.