Current study
Our first aim was to replicate the pilot study in a new experiment. We refer to this new experiment as the Oddball task . We conservatively assumed that the true effect size of the late alcohol-induced SPN difference is half that found in the pilot (dz = 0.58) and collected a larger sample accordingly (N = 26, power = 0.8, alpha = 0.05, two-tailed).
As well as increasing sample size, we increased the number of trials to 100 per condition and increased the number of oddball trials to 60 (from 80 and 32 in the pilot study). This improved signal quality, and the additional oddball trials allowed us to measure a more reliable P300 component. Furthermore, to reduce variability between participants, the absorption period of alcohol was kept to 10 minutes.
Given the pilot study results, we predicted that alcohol would have no effect on positive P1 peak, reduce the negative N1 dip, and enhance the SPN in the 400-1000 ms window (https://aspredicted.org/hj9eh.pdf). We also predicted that alcohol would reduce the P300 response to oddballs. Analysis of the Oddball task confirmed our predictions regarding P1, N1 and P300. However, the late SPN was slightly reduced, rather than enhanced, in the alcohol session (the opposite of the pilot results).
Following those results, we completed another new experiment, with another group of 26 participants. These participants discriminated symmetry from random trials and ignored colour. We refer to this as theRegularity task . This is an important comparison. It might be that automatic symmetry responses are more vulnerable to alcohol induced changes. Here we predicted alcohol would have no effect on P1, reduce N1, and have no effect on the SPN. These predictions were confirmed. We also predicted that the SPN would be larger overall in the Regularity task than the Oddball task, although this was not confirmed (https://aspredicted.org/yn3vq.pdf).
Results of the Oddball and Regularity tasks are presented together and as separate groups in a mixed design.