Date of Award

2019

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Neurology

College

College of Graduate Studies

First Advisor

Leonardo Bonilha

Second Advisor

Jens H. Jensen

Third Advisor

Julius Fridricksson

Fourth Advisor

Joseph Helpern

Fifth Advisor

Chris Rorden

Sixth Advisor

Steven Kautz

Abstract

In 1909 German neurologist Korbinian Brodmann wrote “functional localization of the cerebral cortex without the lead of anatomy is impossible... In all domains, physiology has its firmest foundations in anatomy [1”. While histology is the current gold standard for studying brain microstructure, it is primarily a post-mortem technique that has an average resolution of one micrometer making it impractical for studying the entire brain. Diffusion Magnetic Resonance Imaging (dMRI) is ideally suited to study whole-brain tissue microstructure by sensitizing the MRI contrast to water diffusion, which has a length scale on the order of micrometers. Even though dMRI is applied clinically for the detection of acute ischemia, the relation between tissue microstructure and the dMRI signal is complex and not fully understood. The focus of this dissertation was the validation and development of a new biophysical model of the dMRI signal. Notwithstanding, it is important to keep in mind the potential clinical applications of these models, so in parallel we studied the relationship between white matter integrity and language impairments in post-stroke anomia. This application is of interest since response to language treatment is variable and it is currently difficult to predict which patients will benefit. A better understanding of the underlying brain damage could help inform on functionality and recovery potential. Our work resulted in 9 peer-reviewed papers in international journals and 13 abstracts in proceedings at national and international conferences. Using data collected from 32 chronic stroke patients with language impairments, we studied the relation between baseline naming impairments and microstructural integrity of the residual white matter. An existing dMRI technique, Diffusional Kurtosis Imaging (DKI), was used to assess the tissue microstructure along the length of two major white matter bundles: the Inferior Longitudinal Fasciculus (ILF) and the Superior Longitudinal Fasciculus (SLF). The frequency of semantic paraphasias was strongly associated with ILF axonal loss, whereas phonemic paraphasias were strongly associated with SLF axonal loss. This double dissociation between semantic and phonological processing is in agreement with the dual stream model of language processing and corroborates the concept that, during speech production, knowledge association (semantics) depends on the integrity of ventral pathways (ILF), whereas form encoding (phonological encoding) is more localized to dorsal pathways (SLF). Using a smaller dataset of 8 chronic stroke subjects whom underwent speech entrainment therapy, we assessed if naming improvements were supported by underlying changes in microstructure. Remarkably, we saw that a decrease in semantic errors during confrontational naming was related to a renormalization of the microstructure of the ILF. Together, these two studies support the idea that white matter integrity (in addition to regional gray matter damage) impacts baseline stroke impairments and disease progression. Acquiring accurate information about a patient’s linguistic disorder and the underlying neuropathology is often an integral part to developing an appropriate intervention strategy. However, DKI metrics describe the general physical process of diffusion, which can be difficult to interpret biologically. Different pathological processes could lead to similar DKI changes further complicating interpretation and possibly decreasing its specificity to disease. A multitude of biophysical models have been developed to improve the specificity of dMRI. Due to the complexity of biological tissue, assumptions are necessary, which can differ in stringency depending on the dMRI data at hand. One such assumption is that axons can be approximated by water confined to impermeable thin cylinders. In this dissertation, we provide evidence for this “stick model”. Using data from 2 healthy controls we show that the dMRI signal decay behaves as predicted from theory, particularly at strong diffusion weightings. This work validated the foundation of a biophysical model known as Fiber Ball Imaging (FBI), which allows for the calculation of the angular dependence of fiber bundles. Here, we extend FBI by introducing the technique Fiber Ball White Matter (FBWM) modeling that in addition provides estimations for the Axonal Water Fraction (AWF) and compartmental diffusivities. The ability to accurately estimate compartment specific diffusion dynamics could provide the opportunity to distinguish between different disease processes that affect axons differently than the extra-axonal environment (e.g. gliosis). Lastly, we were able to show that FBI data can also be used to calculate compartmental transverse relaxation times (T2). These metrics can be used as biomarkers, aid in the calculation of the myelin content, or be used to reduce bias in diffusion modeling metrics. Future work should focus on the application of FBI and FBWM to the study of white matter in post-stroke anomia. Since FBWM offers the advantage of isolating the diffusion dynamics of the intra- and extra- axonal environments, it could be used to distinguish between pathological processes such as glial cell infiltration and axonal degeneration. A more specific assessment of the structural integrity underlying anomia could provide information on an individual’s recovery potential and could pave the way for more targeted treatment strategies. The isolation of intra-axonal water is also beneficial for a technique known as dMRI tractography, which delineates the pathway of fiber bundles in the brain. dMRI tractography is a popular research tool for studying brain networks but it is notoriously challenging to do in post-stroke brains. In damaged brain tissue, the high extra-cellular water content masks the directionality of fibers; however, since FBI provides the orientational dependence of solely intra-axonal water, it is not affected by this phenomenon. It is important to understand that caution should be taken when applying biophysical models (FBWM/FBI vs. DKI) to the diseased brain as the validation we provided in this work was only for healthy white matter and these experiments should be repeated in pathological white matter.

Rights

All rights reserved. Copyright is held by the author.

Download

Share

COinS