Principal Investigators: Peter van Zijl, PhD, Xu Li, PhD, Hye-Young Heo, PhD
Co-investigators: Hyeong-Geol Shin

The definition of a Quantitative Imaging Biomarker (QIB) is very broad and can be rightfully interpreted as relating to any imaging parameter that can report on normal or abnormal tissue status or on changes in tissue status related to a treatment. This is perfect for magnetic resonance (MR) technology, which can probe tissue composition and structure non-invasively on microscopic, mesoscopic, and macroscopic spatial scales, and can also access molecular information. Using different types of radiofrequency and magnetic field gradient based pulse sequences, such characteristics can be probed through a multitude of MRI-based parameters that reflect water content and dynamics, water proton relaxation, water proton exchange transfer, and water magnetic susceptibilities, which can all be seen as candidate biomarkers. The absolute value of tissue magnetic susceptibility is affected by its myelin and iron contents, as well as blood oxygenation (Y), parameters of great interest to our CPs. The magnetic susceptibility of a voxel is also affected by tissue structural anisotropy, e.g. oriented myelin sheaths in the white matter. The overall goal of this TRD is to design pulse sequences and analysis approaches to efficiently quantify these parameters and use them to assess tissue composition, structure and physiology in the many disorders studied by our CPs.

We are developing susceptibility source separation methods for mapping tissue composition and oxygenation (aim 1), susceptibility tensor imaging (STI) based mapping of crossing fibers with submillimeter (mesoscopic) resolution to assess white matter connections and their myeloarchitecture (aim 2), and multi-parametric MR fingerprinting (MRF) frameworks that can efficiently assess the candidate biomarkers mentioned above (aim 3). The MRF reconstruction and analysis will be combined with deep learning models by utilizing a multilayer, nonlinear parameter-mapping approach to relate MR fingerprints to the underlying tissue properties. 

Contents: