UC Berkeley

The cognitive spectrum between normal aging and dementia is broad. Many terms including mild cognitive impairment (MCI) have been developed to identify a group of people at the transitional phase for early detection of Alzheimer disease (AD). The lack of biomarker based criteria and the dependence on the sociocultural context result in great variability in case definition.

Cerebrospinal fluid (CSF), positron emission tomography (PET) and magnetic resonance imaging (MRI) serve as three important tools to track biological changes in AD. The Alzheimer's Disease Neuroimaging Initiative (ADNI) provides the infrastructure for investigators to examine the longitudinal patterns of CSF, PET and MRI biomarkers at different cognitive stages. The dissertation first delineated the biomarker changes over time in relation to cognitive decline in ADNI and found that the trajectories support a hypothetical sequence of AD pathology, suggesting that biomarker prediction for cognitive change is stage dependent.

Missingness is common but often overlooked in longitudinal studies of AD. The mechanism of missing data is often assumed to be missing completely at random. The second aim of the dissertation is to test this assumption. The missing biomarker data in ADNI were found not completely at random but rather conditional on certain clinical features. Understanding the missing data structure may help in the design of future longitudinal studies and clinical trials in AD.

Cognitive reserve has been proposed to account for the discordance between cognitive performance and AD pathology. The long held viewpoint is that cognitive reserve affects the clinical expression but has no direct effect on AD pathology. This viewpoint was re-examined in the dissertation. The results showed that higher cognitive reserve indexed by education and other proxies was associated with slower rates of AD pathological deterioration, particularly among cognitively normal elderly people. These findings suggest that the pathological course of AD can be modified by cognitive reserve.

Many cardiovascular risk factors increase the risk of AD. Vascular dysfunction reduces brain reserve or threshold of cognitive impairment. Whether the underlying mechanism also involves impairment of cerebral amyloid clearance remains controversial. Vascular burden, indexed by cardiovascular risk profile and MRI white matter hyperintensities, was not significantly associated with rates of AD biomarker changes, suggesting that typical AD pathology, presumably reflective of amyloid accumulation, appears to be independent of vascular burden.

In conclusion, CSF and imaging markers change over time at different rates in aging and dementia and the missing data are conditional on certain clinical features during follow-ups. Education and other cognitive reserve surrogates may have direct effects on AD pathological progression while vascular burden may influence cognitive function via its own pathway independent of amyloid deposition. Considering the longitudinal effect of cognitive reserve and the potential to control vascular risks, AD can be a preventable disease.