Capturing characteristic features in the human cortical gray matter and hippocampus in vivo using submillimeter diffusion MRI

Limited Access
This item is unavailable until:
2024-09-16

Date

2022

Journal Title

Journal ISSN

Volume Title

Repository Usage Stats

67
views
0
downloads

Abstract

Alzheimer's disease (AD) accounts for 60%-80% of dementia. AD patients start by having mild memory, language, and thinking difficulties, then gradually lose more critical abilities, such as dressing, bathing, or walking. AD not only degrades patients’ life quality but also burdens caregivers and the health system. Specifically, there are 6.5 million AD cases in the U.S. today, and the annual health costs for 2022 are estimated to be $321 billion. AD diagnosis has been evolving in the past 30 years. The criteria established in 1984 recommended that AD cannot be identified until a post-mortem neuropathological test is performed. Recently, more biomarkers have gradually been discovered, such as brain atrophy, Positron Emission Tomography (PET) measures of glucose hypometabolism, and cerebrospinal fluid (CSF) and PET measures of pathological amyloid-beta and tau. However, these biomarkers lack the specificity to probe the damage in the neuronal microstructure that directly causes the disease, and they only provide late diagnoses when the AD progression is no longer reversible. Since the neuronal damages are believed to begin 20 years or more before symptoms start, biomarkers that can detect abnormalities in the neuronal microstructure would enable the diagnosis of AD at the very early stage of neurodegeneration, years before the onset of symptoms, and they could thus potentially enable better treatment outcomes since neuronal damage at the early stage could be reversible.Diffusion tensor imaging (DTI) is a magnetic resonance imaging technique that can noninvasively probe the microstructure of the human brain in vivo. Some regions in the cortical gray matter and hippocampus are known to experience early neurodegeneration in AD, and changes in DTI metrics in these regions could reflect the early stage of AD. However, the cortex is folded and is made of different cortical layers and cortical regions and the hippocampus is made of different subfields that have distinct neuronal populations with a specific microstructure. Additionally, neurodegeneration does not necessarily occur at the same time across different cortical depths or regions in the cortex or across different subfields in the hippocampus. As such, the development of early diagnostic biomarkers would require the ability to probe the neuronal microstructure at specific cortical depths and in specific cortical regions and hippocampal subfields in vivo. However, doing so with DTI has been challenging because the average cortical thickness is only 2.5 mm and the average hippocampal volume is only 2.84 mL. Therefore, a high-resolution DTI acquisition within a reasonable scan time is needed. In this dissertation, we first aim to develop DTI acquisition and reconstruction methodologies to acquire high-resolution (0.9-mm to 1.0-mm isotropic) whole-brain DTI images. Specifically, we used an efficient multi-band multi-shot echo-planar imaging sequence and a multi-band multiplexed sensitivity-encoding reconstruction. Furthermore, we aim to develop a data analysis pipeline that can quantitatively probe the microstructure and capture characteristic features: 1) in the cortex by performing a column-based cortical depth analysis of the diffusion anisotropy and radiality; and 2) in the hippocampus by investigating intra-hippocampal fiber tracts and connectomes, with the long-term goal of enabling the early diagnosis of AD. In the cortex, a column-based cortical depth analysis that samples the fractional anisotropy (FA) and radiality index (RI) along radially oriented cortical columns was performed to quantitatively analyze the FA and RI dependence on the cortical depth, cortical region, cortical curvature, and cortical thickness across the whole brain. We first studied young healthy subjects to optimize the data acquisition and analysis pipeline and to investigate the consistency of the results. The results showed characteristic FA and RI vs. cortical depth profiles, with an FA local maximum and minimum (or two inflection points) and a single RI maximum at intermediate cortical depths in most cortical regions, except for the postcentral gyrus where no FA peaks and a lower RI were observed. These results were consistent between repeated scans from the same subjects and across different subjects. They were also dependent on the cortical curvature and cortical thickness in that the characteristic FA and RI peaks were more pronounced i) at the banks than at the crown of gyri or at the fundus of sulci and ii) as the cortical thickness increases. We then performed a preliminary clinical study in a small cohort of AD patients and age-matched healthy controls (HC) to further examine if this methodology could be applied to detect differences in the FA and RI vs. cortical depth profiles between the AD and HC groups. The FA and RI at each cortical depth and in different regions of interest (ROIs) were sampled and compared between these two groups to look for any significant differences. Additionally, based on the results from the young healthy subjects, we minimized the dependence of these DTI metrics (FA and RI) on structural metrics such as cortical thickness and cortical curvature. The results showed significant differences (p < 0.05) in the FA and RI profiles between the AD and HC groups for specific cortical depths, curvature subsets, and ROIs. To generate intra-hippocampal fiber tracts and connectomes, the hippocampus of all subjects was registered to a common template and deterministic fiber tracking was performed. The fiber orientations across hippocampal subfields were investigated, and the connectivity among subfields was quantified. The results showed characteristic fiber orientations in different hippocampal subfields that were generally consistent between repeated scans and across all subjects: right/left in the middle of the CA4/dentate gyrus subfield and the inferior part of the subiculum; anterior/posterior in CA2/CA3; superior/inferior in the medial and inferior parts of the molecular layer and subiculum. These in vivo fiber orientations aligned with those obtained from an ex vivo specimen scanned over 21 hours at a 0.2-mm isotropic resolution. However, the ex vivo scan delineated the C-shaped molecular layer, which was not shown in the in vivo scans. The in vivo connectomes were generally consistent between repeated scans and across all subjects. The in vivo and ex vivo connectomes both showed more connectivity within the head than within the body of the hippocampus; however, the in vivo and ex vivo connectivity ranking across pairs of subfields was not exactly the same, which could be explained by altered diffusion properties in the ex vivo sample due to fixation or by the higher resolution in the ex vivo scan. In conclusion, the proposed high-resolution whole-brain DTI acquisition, column-based cortical depth analysis of the diffusion anisotropy and radiality, and intra-hippocampal fiber tracking captured characteristic features of FA and RI vs. cortical depth profiles in the cortex and characteristic fiber orientations and connectivity strengths across different subfields of the hippocampus, which were consistent between repeated scans from the same subjects and across different subjects. In addition, the cortical analysis applied in a preliminary clinical study of AD patients vs. HC showed significant differences in the FA and RI profiles between these two groups, showing the potential of this methodology to generate biomarkers for the early diagnosis of AD.

Description

Provenance

Citation

Citation

Ma, Yixin (2022). Capturing characteristic features in the human cortical gray matter and hippocampus in vivo using submillimeter diffusion MRI. Dissertation, Duke University. Retrieved from https://hdl.handle.net/10161/25805.

Collections


Dukes student scholarship is made available to the public using a Creative Commons Attribution / Non-commercial / No derivative (CC-BY-NC-ND) license.