CoilDrop-MRI: Self-supervised physics-guided MRI reconstruction with coil dropout

2026-05-25

AI Summary

This work addresses the challenge of lacking fully sampled data in accelerated magnetic resonance imaging (MRI) by proposing a novel self-supervised reconstruction method. By introducing a coil dropout strategy along the receiver coil dimension, the approach constructs self-supervised input–target pairs and integrates them into a SENSE/SPIRiT unrolling architecture to enable physics-guided reconstruction. This is the first method to partition data along the coil dimension, effectively exploiting inter-coil signal correlations and successfully extending to multi-shot diffusion MRI. Evaluated across multicenter, multi-field-strength (0.3T/0.55T/3T), and multimodal (T1/T2/T2-FLAIR/dMRI) datasets, the method achieves substantially higher reconstruction quality than existing self-supervised approaches—approaching supervised performance—while simultaneously enhancing generalization capability and data efficiency.

Abstract

Self-supervised deep learning-based methods have shown great promise for accelerated magnetic resonance imaging (MRI) reconstruction, achieving high image quality without requiring fully sampled data for training. These methods typically partition the acquired data into two disjoint subsets to construct input-target pairs for optimizing the reconstruction network. However, existing approaches perform this partition exclusively within the spatial frequency (k-space) domain, leaving the coil dimension unexplored. To enforce full exploitation of signal correlation across receiver coils, we propose CoilDrop-MRI, which applies coil-wise dropout to the input and uses the dropped data as training targets in a self-supervised framework. This method is integrated into unrolled architectures in both image-domain (SENSE) and k-space (SPIRiT) formulations. We further demonstrate its versatility by extending CoilDrop-MRI to multi-shot, phase-corrected diffusion MRI (dMRI) reconstruction. CoilDrop-MRI is extensively validated on multi-site, multi-field-strength (0.3T, 0.55T, and 3T), and multi-modality (T1-weighted, T2-weighted, T2-FLAIR, and dMRI) datasets and consistently outperforms state-of-the-art self-supervised methods, achieving quality comparable to supervised reconstruction methods without requiring fully sampled reference training data. Moreover, CoilDrop-MRI exhibits strong data efficiency and robust generalization across imaging conditions, establishing it as a practical and versatile framework for self-supervised parallel MRI reconstruction.

Problem

• MRI reconstruction
• self-supervised learning
• coil correlation
• parallel MRI
• accelerated MRI

Innovation

• CoilDrop-MRI
• self-supervised learning
• coil-wise dropout
• parallel MRI reconstruction
• physics-guided deep learning