Motivation: 3D structural brain MRI and accurate region segmentation are important for neurological research and clinical applications, but they are time-consuming in both acquisition and processing. Goal(s): This study aims to achieve simultaneous high-resolution 3D T1W brain imaging and precise whole-brain segmentation in just 36 seconds. Approach: A dual-task transformer-based network was developed to perform both image enhancement and brain region segmentation using data acquired in 36-second on a 5.0T MRI system. Results: The proposed method produced high-quality images and accurate segmentation (Dice 0.83） comparable to 3-min 36-second standard scans, offering an ultra-fast solution for clinical applications. Impact: This study implemented a dual-task transformer-based network on a 5.0T MRI platform, enabling high-resolution 3D T1W brain imaging and accurate segmentation in 36 seconds. With performance comparable to conventional 3-min 36-second scans, this approach offers a promising ultra-fast clinical solution.
Li et al. (2025) studied this question.