Biophysical modeling of white matter in magnetic resonance imaging

Thesis

Biophysical modeling can be combined with magnetic resonance imaging (MRI) to measure microstructural features of brain white matter (WM). Non-invasive measurements of WM microstructure is important whether we are trying to study pathology, monitor progression of disease or track drug efficacy. This DPhil project investigates the relationship between features of a biophysical model of WM and predicted/measured MRI signal properties.

An explicit multi-compartment model of WM is developed in Chapter 2. This model is based on the magnetic susceptibility of myelin and captures microstructural compartments in terms of their size, shape and physical properties. We use this model to examine the role of myelin content (g-ratio and axon density) by calculating microstructure-driven field perturbations and forward predicting the gradient echo (GRE) signal.

Chapter 3 focuses on myelin geometry and its relation to the GRE signal using the framework developed in Chapter 2. Current models of WM assume idealized packings of nested cylinders as axons. In reality, axons exist in varying geometries. We explore the role of geometry at the single axon and axon bundle level through simulation and by incorporating realism based on electron microscopy. We then apply this model to study demyelination (loss of healthy myelin, characteristic of many neurodegenerative diseases) by comparing simulation predictions with measurement collected from an animal model of demyelination. Overall, our results suggest that myelin geometry has a significant effect on the GRE signal and that estimates of microstructural features may be biased if myelin shape is not appropriately considered.

Chapter 4 extends the models developed in Chapters 2-3 to examine WM microstructure and its relation to the balanced steady state free precession (bSSFP) signal. We focus especially on asymmetries in the bSSFP profile and explore how these attributes could provide useful biomarkers for tissue health. We fit our biophysical model of WM to bSSFP measurements to quantify features specific myelin content. Our results demonstrate promise for the extraction of clinically relevant features of WM from in vivo bSSFP data as well as challenges in the current method.

Authors: Xu, T

• DOI: 10.5287/ora-qzxbnkj8y
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English