OHBM 2022 | 114 | Talk | Young Woo Park | Super-resolution in-vivo human MR imaging…

Title: Super-resolution in-vivo human MR imaging.
Session: Talk
Speaker: Young Woo Park
Increasing the spatial resolution of in-vivo human brain magnetic resonance imaging (MRI) to 100 µm/pixel or above could enhance the visualization of brain microstructure and microvasculature, thereby potentially improving the detection of microlesions in disorders like epilepsy or multiple sclerosis. It would also help further our understanding of the human brain anatomy, especially in small and complex regions such as the cerebellum. Here, we describe a technique based on super-resolution algorithms to generate higher resolution images, by acquiring and combining multiple lower resolution images. Super-resolution combines multiple lower-resolution images with sub-voxel shifts to yield a single high-resolution image. Previously, super-resolution had been used for MRI in scan time reduction applications. For this study, we used super-resolution to estimate the ultrahigh-resolution image. A comparison of high-resolution and super-resolution images shoe super-resolution yields comparable image quality to a high-resolution image with a couple of added benefits such as improved signal-to-noise ratio and contrasts as well as resistance to motion artifacts. Super-resolution can also be applied to various other contrast types such as T1, T2, and SWI to yield different anatomical information. Finally, the generation of high-resolution 3D volume from multiple 2D slices is possible with super-resolution in a single dimension. These demonstrate the possibility of more accessible high-resolution MRI acquisitions, which could enhance the clinical and scientific utilities of ultra-high field MR scanners.