Genome-wide association tests were performed on 1,064 frontal theta ERO measurements and corresponding genotypic data using a standard quantitative trait measured genotype method [Boerwinkle et al., 1986] implemented in the new GWAS software package PLINK [Purcell et al., 2007a]. Under a simple fixed effects additive measured genotype (MG) model, SNP genotypes are coded 0 for heterozygotes, −1 for one homozygote, and 1 for the other homozygote and then the variation of the trait mean by genotype is assessed via a general linear regression. Population stratification, a well-known source of confounding for case-control association studies, was corrected using two methods: computation of genomic-control (GC) adjusted significance values [Devlin and Roeder, 1999], which accommodates over-dispersion of the chi-square statistic due to heterogeneity through the estimation of an inflation statistic; and adjustment of genotypes and phenotypes for population ancestry through linear regressions of 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
