Author ORCID Identifier

Date Available


Year of Publication


Degree Name

Doctor of Philosophy (PhD)

Document Type

Doctoral Dissertation




Behavioral Science

First Advisor

Dr. Gregory A. Jicha

Second Advisor

Dr. Charles D. Smith


Introduction: Alzheimer’s Diseases (AD) & cerebral small vessel disease associated with cardiovascular risk factors (cSVD) frequently coexist, differentially affecting both imaging and clinical features associated with aging and dementia. We hypothesized that Magnetic Resonance Imaging (MRI) can be used in novel ways to identify relative contributions of AD & cardiovascular risks to cSVD and brain atrophy, generating new biomarkers & insights into mixed disease states associated with cognitive decline and dementia.

Methods: Three experiments were conducted to address the overarching hypothesis. First, we visually rated the clinical MRI of 325 participants from a community-based cross-sectional sample to elucidate the relative association of age, AD (visualized as hippocampal atrophy) and cSVD (visualized as white matter hyperintensities; WMH) with global brain atrophy in experiment 1. In experiment 2, we analyzed cross-sectional MRI scans from 62 participants from the University of Kentucky Alzheimer’s Disease Center (UKADC) with available clinical data on cardiovascular risk and cerebrospinal fluid (CSF) beta-amyloid levels as a marker of AD. Voxel wise regression was used to examine the association of white matter hyperintensities with AD and/or cardiovascular risk (hypertension). Experiment 3, examined longitudinal MRI changes in WMH volumes in 377 participants from the Alzheimer’s Disease Neuroimaging Initiative 2 (ADNI 2). Subjects were categorized into three groups based on WMH volume change, including those that demonstrated regression (n=96; 25.5%), stability (n=72; 19.1%), and progression (n=209; 55.4%) of WMH volume over time. Differences in brain atrophy measures and cognitive testing among the three group were conducted.

Results: In the first experiment, logistic regression analysis demonstrated that a 1-year increase in age was associated with global brain atrophy (OR = 1.04; p = .04), medial temporal lobe atrophy (MTA; a surrogate of AD) (OR = 3.7; p < .001), and WMH as surrogate of cSVD (OR = 8.80; p < .001). Both MTA and WMH were strongly associated with global brain atrophy in our study population, with WMH showing the strongest relationship after adjusting for age. In the second experiment, linear regression as well as mediation and moderation analyses demonstrated significant main effects of hypertension (HTN; the strongest risk factor associated with cSVD) and CSF Aβ 1-42 (a surrogate of AD) on WMH volume, but no significant HTN×CSF Aβ 1-42 interaction. Further exploration of the independence of HTN and Aβ using a voxelwise analysis approach, demonstrated unique patterns of WM alteration associated with either hypertension or CSF Aβ 1-42, confirming that both independently contribute to WMH previously classified as cSVD. Extending this work into a longitudinal model rather than focusing on purely cross-sectional associations, we demonstrated that spontaneous WMH regression is common, and that such regression is associated with a reduced rate of global brain atrophy (p = 0.012), and improvement in memory function over time (p = 0.003).

Conclusion: These data demonstrate that both AD and cSVD frequently coexist in the same brain, contributing differentially to alterations in brain structure, subcortical white matter injury, and cognitive function. These effects can be disentangled using MRI, and while we currently lack therapeutic interventions to halt or reverse AD, the dynamic WMH change evident in our data clearly suggests that the ability to reverse cSVD exists today.

Digital Object Identifier (DOI)

Funding Information

This study was funded by NIH P30 AG028383, UH2 NS100606, NR014189, and R01 AG042419.