Up to 40% of all dementia cases, particularly Alzheimer’s Disease (AD), are associated with modifiable lifestyle-related risk factors, such as alcohol consumption, obesity, and hypertension, among others. As exposure to most of these risk factors begins decades before dementia onset, interventions must be implemented in midlife. However, the indicators of the disease process in mid-life are poorly understood. Therefore, a necessary preliminary step to future large-scale interventions is the development and validation of sensitive biomarkers of early Alzheimer’s disease process. Two previous studies – the Nun Study and the Precursor Study – found that the language output of young adults (i.e., idea density in written language) was associated with Alzheimer’s disease in old age. However, brain imaging was not performed, so it’s impossible to know whether idea density related to fundamental brain differences/early pathology in individuals who developed AD relative to those who did not. To address this gap, this study will investigate whether language output relates specifically to risk of late life AD (i.e., family history of dementia, genetic, cardiovascular risk), and to risk-related changes in cognition and the brain. In healthy middle-aged individuals (N = 179) assessed with cognitive and neuroimaging tests, we will investigate whether language output differentiates those at high vs low AD risk cross-sectionally, and whether it tracks AD risk and incipient neuropathology longitudinally, at 2-year follow-up. Our first hypothesis is that lower language output scores will be related to higher AD risk and brain changes in regions associated with AD neuropathology. Our second hypothesis is that the coupling between language output and AD risk will be stronger over time, showing a longitudinal change over and above the initial cross-sectional effect. This work will shed light on whether language output could be used to predict future Alzheimer’s disease, 20–30 years prior to symptom onset.
