principal component scores estimated from genotypic data [Price et al., 2006]. Success of the principal component score-based correction was assessed by examining the GC correction factor. Furthermore, neurophysiological endophenotypes are known to vary between males and females and by age and thus age and sex were also incorporated as covariates in the linear regression model. Case/control status was not used as a covariate in the main analysis, despite observed theta ERO deficits among the GWAS cases, in order to detect genetic variants that may contribute to the susceptibility of alcohol dependence through their effects on electrophysiology. If a gene affects brain function, as indexed by electrophysiological measures, and these differences in brain function also contribute to risk of alcohol dependence, then correcting for case/control status partially removes the effects of the very genes we are most interested in localizing. However, secondary analyses were performed in which case/control status was used as a covariate, producing negligible differences among the top-ranking SNPs as compared to the main GWAS results. The full set of P-values that emerged from the PLINK association analysis were loaded and visualized in Haploview ver. 4.1 [Barrett et al., 2005] and WGAViewer ver. 1.25. These two programs were used
