Genotyping the Understanding America Study to generate novel opportunities for research on cognitive functioning and dementia

Project summary / abstract Cognitive impairment and dementia are prevalent and cause significant morbidity and substantial financial and social burden. With the rising number of cases of dementia in the U.S. and worldwide, there is an urgent need to identify opportunities for preventing or delaying its onset.

In this infrastructure proposal, we propose to make use of recent advances in genetics by genotyping the Understanding America Study (UAS) and constructing “polygenic scores” (PGSs), indexes that aggregate the small effects of millions of genetic variants from across the genome, for use in social-science studies of factors

that increase or mitigate the risk of Alzheimer’s Disease and Related Dementias (ADRD). The UAS, a probability- based Internet panel housed at the Center for Economic and Social Research (CESR) at the University of Southern California, longitudinally tracks a sample of approximately 10,000 adults in the U.S. (growing to at least

20,000 by 2026). It combines several sources of information, including from surveys, wearable devices, administrative linkages, and contextual data, and has several unique features: it provides the opportunity for on- demand data collection on short notice; it allows for the collection of data at higher frequencies and for the

possibility of initiating new data collection in response to major societal events (such as the COVID-19 pandemic), or triggered by events in the lives of respondents (such as “burst surveys” fielded when there is an important change in the life of a panel member); it can be used to take advantage of natural experiments; it

allows for frequent collection (once or twice a month) of paradata (computer user-behavior from surveys, e.g., errors and processing speed gleaned from keystrokes) which is predictive of cognitive functioning. Ours is not a genome-wide association study (GWAS). Instead, we will use genetic variants (SNPs) identified

from existing large, replicated GWASs, to create polygenic scores (PGSs), and exploit unique UAS capabilities, afforded by its Internet mode of operation, to better understand ADRD risk in a nationally representative sample. We will use PGSs, as well as APOE-ε4 status, together with longitudinal health, cognitive, behavioral, and

environmental measures, to: (i) identify populations at risk of cognitive decline, (ii) collect new data for causal inferences of the effects of ADRD risk/protective factors on cognition by genetic ADRD risk, and iii) study the role of genetics in the resilience to adverse life events affecting cognitive functioning.

By making publicly available a large number of genetic measures for ADRD, cognitive decline, and associated protective/risk factors (e.g., physical activity, cardiovascular risk [diabetes, obesity, smoking and hypertension], diet, sleep, pollution, and education, among others), and through our own research, we seek to

stimulate the use of unique UAS capabilities in economic and social-science research of ADRD, cognitive impairment, and cognitive decline.