CL-MRI: Self-Supervised contrastive learning to improve the accuracy of undersampled MRI reconstruction

Abstract

Deep learning (DL) methods have emerged as the state-of-the-art for Magnetic Resonance Imaging (MRI) reconstruction. DL methods typically involve training deep neural networks to take undersampled MRI images as input and transform them into high-quality MRI images through data-driven processes. However, deep learning models often fail with higher levels of undersampling due to the insufficient information in the input, which is crucial for producing high-quality MRI images. Thus, optimizing the information content at the input of a DL reconstruction model could significantly improve reconstruction accuracy. In this paper, we introduce a self-supervised pretraining procedure using contrastive learning to improve the accuracy of undersampled DL MRI reconstruction. We use contrastive learning to transform the MRI image representations into a latent space that maximizes mutual information among different undersampled representations and optimizes the information content at the input of the downstream DL reconstruction models. Our experiments demonstrate improved reconstruction accuracy across a range of acceleration factors and datasets, both quantitatively and qualitatively. Furthermore, our extended experiments validate the proposed framework’s robustness under adversarial conditions, such as measurement noise, different k-space sampling patterns, and pathological abnormalities, and also prove the transfer learning capabilities on MRI datasets with completely different anatomy. Additionally, we conducted experiments to visualize and analyze the properties of the proposed MRI contrastive learning latent space. Code available here.