PROJECT SUMMARY/ABSTRACT This highly responsive R03 from a new investigator proposes to apply contemporary causal in- ference methods to research in Alzheimer's disease (AD) and Alzheimer's disease related dementia (ADRD), which are complex conditions whose causes and risk factors over a person's life time are not well understood by the scientific community. This one will be a demonstration project using causal inference methods to investigate the complex relationship between time-varying mid-life exposures and late-life cognitive outcomes that are potentially related to Alzheimer's disease, in the presence of time-varying confounders, and with time-varying mediators. As illustration we will consider the mid- life alcohol exposure collected from the Honolulu Heart Program (HHP) and the cognitive outcomes collected from the Honolulu-Asia Aging Study (HAAS), where a number of time-varying confounders and time-varying mediator variables were also collected in both studies. Under this project we will develop and implement statistical approaches to study: 1) the causal effect of time-varying mid-life exposure on late-life cognitive outcomes; 2) the effects of time-varying mid-life exposure on late-life cognitive outcomes mediated by time-varying physical well being variables; 3) the causal effect of time-varying mid-life exposure on time to late-life moderate or severe cognitive impairment; 4) the effects of time-varying mid-life exposure on time to late-life moderate or severe cognitive impairment mediated by time-varying physical well being variables. Successful application of these methods will shed light on the research field of causal effects of various life style and other exposures on the out- comes of Alzheimer's disease and Alzheimer's disease related dementia, where the literature currently presents inconsistent findings using traditional statistical approaches such as direct applications of tra- ditional regression models. Statistical approaches developed under this proposal will be published in scientific journals. In cases where software for certain analysis are not yet available, open source R software packages will be implemented and made freely available to the public.