1H-Guided Unsupervised Super-Resolution Reconstruction of 23Na in Simulcast X-nuclei MRI

Motivation: Simulcast multi-nuclei MRI, when acquiring images from multiple nuclei simultaneously, face challenges with non-proton nuclei displaying lower resolution compared to proton nuclei. Conventional methods enhance resolution via k-space zero-padding and cropping, albeit with detail loss. Goal(s): To enhance the resolution of non-proton nuclear images in Simulcast X-nuclei MRI without compromising texture details. Approach: We proposed a deep learning-based method for unsupervised super-resolution reconstruction of X-nuclei images, guided by 1H. Results: Our method enhance the resolution of non-proton nuclear images while preserving texture details as much as possible. Impact: Our deep learning approach markedly improves non-proton nuclei image quality in simulcast multi-nuclei MRI, advancing rigorous scientific research in simulcast multi-nuclei MRI, especially in settings with limited access to advanced multi-nuclei simultaneous imaging technologies.