Alcohol did not affect working memory task performance in Study 1 or automobile driving simulator performance in Study 2. Values for theta, alpha and beta band power and standard deviation variables were not affected by alcohol in Study 1, but were affected in Studies 2 (p<.01) and 3 (p<.01). These differences in EEG ANOVA significances between the three studies may merely reflect chance variations in the small subject populations. In Study 3, post-hoc tests showed significantly greater power and standard deviations for the alpha and beta variables, consistent with prior studies [9]–[13], [18], [25], [31]. Thus, the equation generated on EEG data with a weak alcohol effect performed well when applied in the two independent studies with the same BAC level. This suggests that the equation indeed quantified a salient aspect of the neurophysiological effect of alcohol.
