However, using all pairs with usable ERP data and quasi-continuous outcome measures for greater statistical power, we correlated within-pair difference scores of ERP variables with within-pair differences in alcohol use/abuse (self-reported alcohol use measures and alcohol dependence criteria symptom counts). Results show that the 3 novel P3 amplitude values correlated negatively, and statistically significantly, with 1 measure of alcohol use. The difference scores of novel P3 frontal (r = −0.28, p < 0.01), central (r = −0.23, p < 0.01), and parietal (r = −0.20, p < 0.05) amplitudes correlated negatively with difference scores of high-density alcohol use, measured by MAX-D24. High-density drinking was associated with low amplitude novel P3. Within-pair differences in this lifetime measure of high-density drinking were as closely associated with within-pair differences in novel P3 amplitudes as were the correlations of MAX-D24 across twins as individuals (see column 3 of Table 4): intrapair difference correlations matched the magnitude of interindividual correlations of these variables. The association was not reduced by controlling for between family confounds. The other 3 self-report measures of alcohol use we used, including self-reported alcohol problems at age 18½ and 25, showed not such robust associations with any P3 variable.
