Investigating Human Brain Function and Human Brain Organization Using Functional Magnetic Resonance Imaging at 4.7 T

• Author / Creator

  Hrybouski, Stanislau

• Understanding the brain-cognition association has been a major goal of neuroscientists for more than 50 years. The discovery of functional magnetic resonance imaging (fMRI) blood-oxygen-level-dependent (BOLD) contrast by Ogawa and colleagues (1990, 1992) has fundamentally transformed the field of human neuroscience, allowing researchers to non-invasively measure function of the human brain. Since that initial discovery, imaging methodology has improved substantially, and new approaches for studying brain function are continuously developed. A recent shift towards high-field (> 3 T) scanners enabled researchers to probe brain structure and function with unmatched anatomical precision. However, the advantages of high-field fMRI datasets can be challenging to realize in practice, since greater spatial resolution comes at a cost: reduced contrast-to-noise ratio, increased vulnerability to head motion artifacts, and geometric distortions caused by longer readout times. High-field fMRI data are also more sensitive to cardiac and respiratory signals of no-interest, which can weaken or bias most statistical inferences, especially in resting-state functional connectivity work. Consequently, the primary objective of this thesis was to develop methodology that is capable of taking the full advantage of high-field fMRI to study brain function and brain organization. We accomplished this aim by combining precise anatomical localization on ultra-high-resolution structural MRI in native space, with an extensive set of fMRI denoising techniques as well as task-specific statistical models of each structure’s hemodynamic response. To test our approach for studying brain function using high field high-resolution fMRI we investigated functional aspects of small medial temporal lobe (MTL) structures, which are notoriously difficult to study because of lower vascular density, susceptibility artifacts, and signal contamination from larger drainage veins.
  The first two experiments of this thesis examine functional properties of the amygdala subnuclei and hippocampal subfields. To activate the amygdala subnuclei we employed negative emotional stimuli that have been shown to elicit amygdala response in other fMRI studies, and the hippocampus was engaged by a computerized adaptation of a standardized clinical memory battery, that is capable of testing neurobiological processes responsible for item, spatial, and associative memory. In the third experiment, we combined a 4.7 T acquisition with a data-driven network parcellation to study age differences in the brain’s functional architecture. We investigated network topography, network amplitude, and inter-network communication for the entire connectome. Inter-network functional connectivity was estimated using a novel sparse graphical estimation procedure that aims to uncover true graph structure with edges representing direct connections only. Simulation work by Allen and colleagues (2012) suggests that fMRI data with greater BOLD contrast-to-noise ratio (e.g., > 4 T) should produce network parcellation that in less biased and more sensitive to differences when performing statistical comparisons between groups.
  In the MTL experiments, emotional stimuli elicited differential engagement of the amygdala subnuclei demonstrating the necessity of studying these small grey matter nuclei separately from each other. Similarly, our memory paradigm revealed a complex pattern of intra-hippocampal specialization within both the anterior and posterior hippocampal subfields in memory encoding and memory retrieval. Finally, our investigation into age effects on the brain’s functional architecture revealed widespread BOLD signal reduction among old adults, affecting every major brain system. Connectivity analyses, however, showed a high degree of age-invariance in the brain’s functional architecture with some subtle differences among age group. At 4.7 T, we were able to obtain a finer network splitting than is commonly reported in neuroimaging literature, highlighting the advantages of high-field acquisitions for studying the brain’s connectome using data-driven network detection techniques. Together, these results revealed clear advantages of high-field fMRI for studying the brain’s structure-function relationships and for investigating its network properties.

• Subjects / Keywords

  • High-field fMRI
  • Amygdala connectivity
  • Aging
  • Emotions
  • Hippocampus
  • Hippocampal subregions
  • Amygdala subnuclei
  • Hippocampal subfields
  • Memory encoding
  • Functional connectivity
  • Sparse graphical models
  • High-resolution fMRI
  • Memory retrieval
  • Resting-state fMRI

• Graduation date

  Fall 2020

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-fa1e-da08

• 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.