Principles of quantitative magnetic resonance imaging and relevance to clinical applications

Lecturer: Claudia Wheeler-Kingshott (UCL Queen Square Institute Of Neurology)

In this lecture I will present how Magnetic Resonance Imaging (MRI) can be used for assessing health and disease of brain tissue by measuring properties of the brain in a magnetic field.

The learning objectives of the lecture will be to:
1) understand why MRI is clinically important and what makes MRI so powerful compared to other imaging techniques;
2) understand what do we mean by quantitative MRI and biophysically meaningful feature extraction from MRI scans;
3) to assess how we can use MRI for brain mictrostructure, function and physiology properties assessment;
4) to evaluate challenges to clinical translation and generatability and finally
5) to discuss new frontiers and limits of MRI.

In particular, I will explain how we can use sensitivity to water diffusion in tissue can help us assessing tissue microstructure integrity and properties. Exploiting the sensitivity of MRI to changes in magnetic susceptibility associated to blood oxygenation status during brain function, we can detect areas of functional activations and resting state networks. As example of using MRI to assess physiological properties of brain tissue I will present sodium imaging and sodium ion concentration quantification. This is really far from clinical adoption but could have great impact if technical challenges are met.

Examples of clinical applications will be presented to make the lecture more practical, including in multiple sclerosis (Toschi et al, Neuroscience, 2019), neurodegeneration (Castellazzi et al, Frontiers in Neuroscience, 2014) and stroke (Särkämö et al. Frontiers in Human Neuroscience 2014). Discussion of emerging areas of development go from the use of high field human scanners (e.g. 7T), to going beyond what we already know and find ways to challenge our approach to study design, to multi-modal applications for example to understand brain dynamics and obviously to big data analysis using artificial intelligence approaches.

Author: Neil Oxtoby

POND group, UCL CMIC and CS