Sustained Posterior Negativity
The SPN was analysed in two a priori intervals (200-400 ms and 400-1000 ms). In the Oddball Task, SPN was reduced in the alcohol block (opposite of the pilot results). In the Regularity task, The SPN was similar in alcohol and placebo blocks. In both tasks, the SPN peaked at around 300 ms then declined.
Mixed ANOVA found a strong main effect of Interval (F (1,50) = 57.519, p < .001, ηp2 = 0.535). The main effect of Drink on SPN amplitude was not significant (F (1,50) = 1.171, p = .284, ηp2 = 0.023). There was no Drink X Interval interaction (F (1,50) = 0.288, p = .594, ηp2 = 0.006) and no Interval X Drink X Task interaction (F (1,50) = 1.676, p = .201, ηp2 =0.032).
There was also no Drink X Task interaction (F (1,50) = 3.601, p = .064, ηp2 = 0.067). Despite this, we report the separate pre-registered analysis on each task: There was a main effect Drink in the Oddball task (F (1,25) = 5.724, p = .025, ηp2 = 0.186), but not in the Regularity task (F (1,25) = 0.272, p = .607, ηp2 = 0.011).
Bayesian ANOVA confirmed the main effect of Interval (BF include > 100). Other effects and interactions were mostly inconclusive (BF include between 0.134 and 0.498). Bayesian pairwise comparisons between alcohol and placebo in the late interval are instructive, given that this was the important window in the pilot study. There was no support for presence or absence of an alcohol effect in the Oddball task (BF10 = 0.814). However, analysis supports theabsence of an alcohol effect in the Regularity task (BF01 = 4.631).
Overall, alcohol has no consistent effect on SPN amplitude. There was no effect of alcohol on SPN amplitude in the Regularity task, and the evidence from Oddball tasks is mixed, especially when considering the contradictory pilot results. In contrast, alcohol reliably reduces N1 amplitude in all cases.