Indirect Detection of Axonal Architecture With Q-Space Imaging

Loading...
Thumbnail Image
Degree type
Doctor of Philosophy (PhD)
Graduate group
Bioengineering
Discipline
Subject
MRI
white matter
mouse
diffusion
q-space simulation
spinal cord
Biomedical Engineering and Bioengineering
Funder
Grant number
License
Copyright date
Distributor
Related resources
Contributor
Abstract

Evaluating axon morphology would provide insights into connectivity, maturation, and disease pathology. Conventional diffusion MRI can provide metrics that are related to axon morphology, but cannot measure specific parameters such as mean axon diameter (MAD) and intracellular fraction (ICF). Q-space imaging (QSI) is an advanced diffusion MRI technique that may be able to provide more information on axon morphology. However, QSI has several limitations that affect its implementation and accuracy. The main objective of this dissertation was to address these limitations and to evaluate the potential of QSI to accurately assess axon morphology. First, a custom-built high-amplitude gradient coil was used to address the limitations in the maximum gradient amplitude available with commercial systems. Second, to understand the relationship between axon morphology and QSI, simulations were used to investigate the effects of the presence of both extracellular and intracellular signals (ECS and ICS) as well as variation in cell size and shape. Third, three QSI-based methods were designed provide specific measures of axon morphology which have not been reported before. The maximum amplitude of the custom gradient coil was 50 T/m that, for the first time, allowed for sub-micron displacement resolution while fulfilling the short gradient approximation. This enabled near-ideal QSI experiments to be performed. QSI experiments on excised mouse spinal cords showed good correlation with histology, but overestimated MAD. Simulations showed that axon morphology was the dominant effect on QSI and suggested that the presence of ECS and ICS signals may complicate interpretation. Three methods were designed to account for signal in ECS and ICS: two relied on a two-compartment model of the displacement probability density function and the echo attenuation at low q-values, and a third varied the gradient duration to differentiate diffusion in ECS from ICS. All three methods provided estimates of MAD and ICF that showed better agreement with histology than QSI. The methods were also evaluated implementation on a clinical scanner. This dissertation demonstrated the sensitivity of QSI to axon morphology and showed the feasibility of three methods to accurately estimate MAD and ICF. Further investigation is warranted to study future applications.

Advisor
Felix W. Wehrli
Date of degree
2011-05-16
Date Range for Data Collection (Start Date)
Date Range for Data Collection (End Date)
Digital Object Identifier
Series name and number
Volume number
Issue number
Publisher
Publisher DOI
Journal Issue
Comments
Recommended citation