Magnetic Resonance Imaging of the Brain in Prenatal Alcohol Exposure and Advances in Measuring Cortical Microstructure

• Author / Creator

  Little, Graham Taylor

• Fetal alcohol spectrum disorder (FASD) encompasses a large spectrum of physical, cognitive and behavioral deficits resulting from prenatal alcohol exposure. Studies using magnetic resonance imaging (MRI) have shown structural and functional brain alterations in children and adolescents with FASD. This thesis addresses the following three absences in the neurodevelopment / FASD literature. 1) The most common differences reported in structural brain imaging studies of FASD is reductions in regional brain volumes associated with prenatal alcohol exposure. However, most studies of regional brain volume have examined individual brain regions separately, ignoring complex relationships between structures that may be able to better discriminate individuals with FASD. 2) Previous studies of brain function in FASD populations have largely employed task-based functional MRI (fMRI) with few studies using resting-state fMRI (rs-fMRI) a method that measures intrinsic functional connectivity between brain structures at rest. The few studies of resting-state functional connectivity in FASD have used samples limited in size and have reported contradictory findings of increased / decreased functional connectivity in FASD justifying the need for studies of larger cohorts of children/adolescents with FASD. 3) Most MRI studies of neurodevelopment of the human cortex rely on indirect measures of brain structure (e.g. thickness) and function (e.g. blood oxygen level dependent signal), results of which infer but do not explicitly measure microstructural changes. In-vivo studies using diffusion MRI of the adult human cortex have shown promise in differentiating regions with known microstructural differences but have relied on long-acquisitions (~1 hour) along with anatomical registration to a T1-weighted image prone to registration errors. This thesis aims to address these fundamental challenges by first using machine learning to classify children/adolescents with FASD from controls and investigate whether patterns (i.e. multivariate analysis) of regional volumetric brain reductions can better discriminate those with FASD relative to analysis of any individual brain region. Secondly, resting-state functional connectivity is investigated in a large multisite cohort of children/adolescents with FASD and compared to controls. This project also required an additional analysis of between-site reliability and correction of functional connectivity measurements using data acquired from the same 8 individuals at the same 4 sites used in the multisite FASD cohort. Thirdly, a framework for diffusion analysis of the cortex is proposed that segments the cortex in native diffusion tensor imaging space (removing the need for additional anatomical imaging/registration) and applied to diffusion data acquired at 1.5mm isotropic resolution.
  
  Data presented in this thesis present novel information about patterns of regional brain volume differences in FASD, demonstrating that an FASD classification model generated from multiple regional brain volumes can more accurately discriminate individuals with FASD from controls (accuracy 77%, sensitivity 64%, specificity 88%) when compared to models generated from any brain region independently. Additional evidence is provided showing differences in FASD related regional brain volume change between males and females, suggesting that FASD classification models should be created separately for males and females. Functional connectivity analysis revealed decreased internetwork connectivity between regions primarily associated with the salience network, frontal-parietal network and language network and suggests underlying deficits in the functional network brain architecture are associated with prenatal alcohol exposure. Both the classification and functional connectivity findings provide evidence of altered brain structure and function in individuals with FASD and this knowledge has the potential to inform diagnostic and medical interventions for this population in the future. Furthermore, this thesis proposes a framework for in-vivo analysis of diffusion tensor imaging (DTI) of the cortex. Results here show accurate segmentation of the cortex on native DTI. Additionally, surface-based diffusion measurements extracted from the cortex revealed values in line with other higher-resolution (1.0 mm - 1.25 mm isotropic) acquisitions. Given the short scan time (~3.5 minutes) the proposed technique could be a useful approach for the study of clinical or neurodevelopmental populations. Overall, this thesis provides novel contributions to the understanding of the effects of prenatal alcohol exposure on the developing brain and provides methodological advances for future studies of the development of cortical microstructure.

• Subjects / Keywords

  • FASD
  • MRI
  • Machine Learning
  • Functional Connectivity
  • Cortex

• Graduation date

  Fall 2020

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-79vt-2912

• License

  Permission is hereby granted to the University of Alberta Libraries to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. Where the thesis is converted to, or otherwise made available in digital form, the University of Alberta will advise potential users of the thesis of these terms. The author reserves all other publication and other rights in association with the copyright in the thesis and, except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatsoever without the author's prior written permission.