also been reported (Matthews et al., 2007, Lydall et al., 2011). SNPs in this haplotype block have also been associated with drug dependence (Agrawal et al., 2006), suggesting that variability in replication may relate in part to differing co-morbidities among samples. The most commonly examined tag-SNP in this region, rs279858, a synonymous SNP in exon 5 of GABRA2, was the only candidate marker associated with AD in a genome-wide analysis of candidate genes (Olfson and Bierut, 2012). The molecular mechanism by which GABRA2 polymorphisms influence risk for AD is unknown. The absence of a linked coding variant suggests that AD-associated variation in this region is in linkage disequilibrium with an as-yet-unidentified functional variant that influences the developmentally regulated expression of GABRA2 or adjacent chr4p12 GABAA subunit genes. This could result in subtle differences in neural connectivity and subsequent behavioral effects. Support for this hypothesis comes from reports that SNPs in GABRA2 are associated with intermediate neural phenotypes, including fast beta frequency electroencephalographic (EEG) activity (Edenberg et al., 2004, Lydall et al., 2011), increased activation of the insular cortex during reward anticipation (Villafuerte et al., 2012), differences in activation of the medial frontal cortex and ventral tegmental area by alcohol cues
