To maximize power, the analyses above used all available data. However, we also conducted a secondary, but more stringent (and, because of sample size, less powerful) analysis in the smaller group of 7 subjects in whom we had both alcohol and placebo imaging data. In this case, we analyzed the [AO > NApO] and [AO > ApCO] difference images, submitting each to a paired-t test under alcohol and placebo conditions.
