Doctoral Thesis: MRI techniques for quantitative and microstructure imaging


Event Speaker: 

Zijing Dong

Event Location: 

via Zoom, see details below

Event Date/Time: 

Thursday, August 5, 2021 - 1:00pm


Mapping brain microstructures such as axonal fibers and myelin content is critical to understand human brain organization and study neurological diseases. Quantitative MRI (qMRI) is a safe, non-invasive imaging method that has been proven to provide high sensitivity to imaging microstructures in in-vivo human studies. However, the need to acquire multiple images for biophysical model fitting requires a long scan time in conventional qMRI acquisition, leading to low SNR, limited spatial resolution, severe image artifacts, and high vulnerability to motion.

This thesis aims at overcoming these challenges and providing efficient microstructure imaging for the human brain with higher speed, SNR, resolution, and motion robustness. Novel MRI acquisition and reconstruction methods were developed to fully exploit the strength of spatiotemporal encoding, hardware, and low-rank priors, which provide significant improvement on three important qMRI fields, including diffusion MRI (dMRI), myelin water imaging, and MR relaxometry. Specifically, in the first part of this thesis, we developed fast diffusion imaging methods that can provide 30-40% higher SNR efficiency, robustness to physiological and field variations, as well as the capability to resolve multi-echo images free from distortions and artifacts commonly seen in conventional methods. The second part presents a novel acquisition method for myelin water imaging with >10× acceleration, which enables a 5-minute fast scan at 1.5-mm isotropic resolution, and allows us to push the current limit of spatial resolution and acquire a first-ever submillimeter in-vivo myelin water imaging at 600-um. The third part of this thesis shows an ultra-fast MR relaxometry method with navigated motion correction, which was demonstrated to produce fast and repeatable quantitative measurement even in presence of large motion. These techniques solve the major challenges of qMRI and should provide important improvements in detecting brain microstructure to better understand human brain organization in both health and diseases.
Thesis Supervisor(s): Kawin Setsompop, Elfar Adalsteinsson (co-advisor)
To attend this defense, please contact the doctoral candidate at zijingd at mit dot edu