Alzheimer’s disease is characterized by two hallmark neuropathologies: A plaques and hyperphosphorylated tau protein. A tends to arise in a diffuse manner throughout cortex, whereas tau pathology accumulates in a stereotyped pattern first in medial temporal lobe and then in selectively vulnerable neocortical regions as the disease progresses. The processes facilitating the propagation of tau pathology are still unknown, but converging cellular and neuroimaging evidence suggests that tau spreads transneuronally along pathways of functional connectivity throughout the brain. In this work, we utilize multimodal human neuroimaging data in older adults without Alzheimer’s disease to investigate the factors involved in the spread of tau at the earliest stages of pathology accumulation. First, we measure resting state functional connectivity between hippocampus and medial parietal cortex to show that greater connectivity strength along this pathway is associated with downstream tau pathology burden cross-sectionally, with consequences for memory function. Next, we identify networks of functional connectivity specific to either episodic memory or executive function, and show that these networks are altered by the presence of tau pathology before widespread neurodegeneration is evident. Finally, we show that baseline connectivity strength and AD pathology interact to predict longitudinal rate of tau pathology accumulation and rate of cognitive decline in a multicohort sample of cognitively unimpaired older adults. Together, this work reveals that functional connectivity can be used to better understand the extent and distribution of the earliest tau accumulation, and outlines what effects this accumulation has on changes in functional brain networks and the cognitive domains they support.

The major theory of causation for Alzheimer’s disease (AD) is deposition of the protein Aβ, however it is unclear how aging influences amyloid accumulation or why cognitive decline can occur in its absence. Evidence suggests that neurovascular dysfunction, such as blood-brain barrier disruption (BBBd), is related to neurodegeneration and cognitive impairment, and that this relationship may occur independently of AD pathology. Although there is evidence of BBBd during aging and AD, the direct relationship between breakdown of the BBB and AD biomarkers, as well as cognition, has not been thoroughly investigated in humans using multi-modal neuroimaging methods. As such, it is important identify the role, if any, that BBBd has in the development of AD pathology and cognitive decline in aging.

The first project investigates the regional pattern of BBBd, measured using dynamic contrast-enhanced MRI (DCE-MRI) during aging in a sample of cognitively normal older adults (OA) and young adults (YA). We found that OA had greater BBBd in regions primarily in the temporal lobe, with some involvement of the parietal and occipital lobes. This pattern of BBBd overlaps considerably with known topography of typical AD pathology. We also found that APOE4 carriers had significantly greater BBBd than non-carriers. Next, we examined whether BBBd in these regions was associated with AD biomarkers, measured using positron emission tomography (PET). BBBd was not associated with tau, however we did find a significant correlation between increased Aβ and increased BBBd in temporal and parietal brain regions. These results indicate important associations between aging, the spatial pattern of BBBd, and possible associations with AD pathology.

It is unclear whether BBBd is directly related to atrophy and cognition or if this relationship depends on AD biomarkers of Aβ and tau. Thus, the second project investigates the relationship between BBBd, atrophy, and episodic memory and non-memory performance. We found that participants with greater BBBd and smaller regional volume/thickness had lower episodic memory scores. Interestingly, this relationship was independent of Aβ and tau and specific for episodic memory, which suggests that the combined effect of greater BBBd and atrophy may operate through a separate pathway from the AD pathogenic cascade to affect memory performance.

There is conflicting animal model evidence concerning whether Aβ and tau lead to BBBd or the reverse; however, the temporal sequence for the relationship between BBBd, AD biomarkers, and cognition has not been investigated in humans using neuroimaging methods. The third project uses longitudinal data, collected prior to BBBd measurement, to examine the relationship between global Aβ accumulation, regional tau accumulation, regional atrophy, cognitive decline, and subsequent BBBd. We found that both EC tau accumulation and episodic memory decline were associated with greater HC BBBd, however these two factors were not independent of each other. There was a significant interaction, where the combined effect of increased EC tau accumulation and worse episodic memory decline was associated with greater HC BBBd, independent of global Aβ changes and regional atrophy. We did not find any associations between longitudinal Aβ accumulation or atrophy and BBBd, but further studies with larger samples are needed to confirm this.

Taken together, these results further our understanding of the potential role of neurovascular dysfunction in the AD pathway and cognitive decline. Our results indicate that during aging there is considerable overlap of brain regions with increased BBBd and regions that accumulate typical AD pathology. This BBBd has a complex relationship with AD biomarkers and cognition, but overall suggest that BBBd may be important in the pathogenesis of AD, but also may have an influence on memory performance in the absence of AD pathology.

An abundance of research has demonstrated that aging is associated with declines in cognition, particularly episodic memory. A similarly large body of work has highlighted the underlying age-related functional alterations that occur within the hippocampus and surrounding medial temporal lobes (MTL) that may give rise to memory impairment. However, aging is also associated with the accumulation of Alzheimer’s disease pathology, namely beta-amyloid plaques (Abeta) and tau neurofibrillary tangles, which have been associated with a host of neural and behavioral changes, even in those without the disease. As such, it is unclear to what extent our understanding of age-related decline is attributable to the accumulation of pathology. Using a combination of neuropsychological assessment, structural and functional magnetic resonance imaging, and positron emission tomography, I will present two projects which seek to quantify hippocampal function in cognitively normal older adults, assess how Abeta and tau interact to alter neural activity, and define a functional mechanism through which pathology leads to episodic memory decline.

The first project focuses on a function commonly attributed to the hippocampus called pattern separation. Here, using a lure discrimination memory paradigm, I demonstrate that cognitively normal older adults exhibit a behavioral shift from pattern separation towards pattern completion that is associated with elevated neural activity in the hippocampus and MTL. Additionally, I show that Abeta and tau are uniquely associated with two measures of neural function during memory encoding, suggesting that Abeta and tau pathology likely interact to produce aberrant neural activity in older adults. In the second project, I use a technique called representational similarity analysis to explore the relationship between memory encoding and retrieval. Here, I demonstrate that older adults exhibit greater neural similarity between encoding and retrieval for incorrect lures (i.e. false alarm) relative to correct lures. I also demonstrate that false alarm similarity is associated with tau, but only in older adults without Abeta. This suggests that, independent of Abeta, tau can disrupt memory function, specifically by altering the relationship between neural activity during encoding and retrieval. In sum, this research indicates that Alzheimer’s pathology alters hippocampal function in a way that leads to episodic memory impairment in cognitively normal older adults, and should be considered when addressing future questions about aging and memory.

Introduction

Method

Results

Discussion