as a useful quantitative phenotype for GWAS and other "omics" analyses. Methods:We included non-Hispanic White participants from four prospective community-based cohort studies (ROS, MAP, CHAP, EAS) who were non-demented at study entry and have at least two repeatedmeasures of cognition.Within each cohort, individual cognitive tests were combined to form aggregate measures of global cognition and episodic memory. Genotype data from each cohort was quality controlled and imputed using 1000Genomes reference panel. We used linear mixed effects models to characterize individual paths of change in cognition, modeling age, sex and education as fixed effects and intercept and slope as random effects. For initial screen, we extracted individual residual cognitive slope estimates from the models and fit a linear regression model for each single nucleotide polymorphism (SNP), adjusting for population stratification. Results were meta-analyzed across cohorts. SNPs with p< 5x10 -8 were considered genome-wide significant and p < 5x10 -5 were considered suggestive. Results: A total of 3534 subjects were included. The average age at enrollment in each study ranged from 72 to 81, follow-up ranged from 2-18 years and incidence of clinical diagnosis of AD ranged from 19% to 34%. The most significant loci for decline in both global cognition and episodic memory was APOE (p1⁄43.1x10 -34 and p1⁄42.1x10 -24, respectively). Although no other nonAPOE locus reached genome-wide significance, we found highly suggestive results at two other regions. First, for global cognition, near the EDAR gene with the top locus reaching a p1⁄48.9x10 -8 and second, on the RAB3GAP2 genewith the top locus reaching a p1⁄41.3x10 -7. Results were similar for episodic memory. We are currently analyzing data from two other cohort studies for validation. Top hits will be further investigated with full mixed effects models for all SNPs with p < 5x10 -4. Conclusions: Our findings to date suggest that cognitive decline may be a useful phenotype for identifying genomic variation associated with AD or other factors influencing cognitive decline.