Opposite to Zack et al. (
2011), who reported alcohol consumption prolonged SSRT in a group of male problem drinkers, we failed to find such a main effect in a sample with males and a similar number of females without a problematic drinking history. Specific factors of the present study might explain this difference. To clarify those potential factors, we compiled a list of existing studies where the effect of acute alcohol use on the stop-signal task performance was tested (see Table S1a) and followed up with some preliminary analyses (see results in Table S1b). We compared studies that did find the impairing effect of alcohol on the stop-signal task performance with studies that did not in terms of sample characteristics, task parameters, the dosage of alcohol administered, and the study design. Regarding sample characteristics, gender ratio and typical alcohol use are possible moderators. It is suggested that males are more vulnerable to the acute effect of alcohol than females (Fillmore & Weafer,
2004), and heavy drinkers are hypersensitive to the short-term effect of alcohol compared to light drinkers (Field et al.,
2010). However, both assumptions have very limited empirical and theoretical support. As to task parameters, the most relevant one is the modality of the stop signal. It was stated that studies using auditory stop signals report statistically significant differences in SSRT compared to studies using visual stop signals (Guillot, Fanning, Bullock, McCloskey, & Berman,
2010). The underlying reasons remain unclear, except that auditory stop tones are perceived as more intense than visual stop cues (van der Schoot, Licht, Horsley, & Sergeant,
2005). Regarding the dosage of alcohol administered, in principle, a high dose of alcohol was more likely to cause impaired inhibition than a smaller amount (0.8 g/kg vs. 0.4 g/kg, Caswell, Morgan, & Duka,
2013). However, exceptions exist such that even a high dose failed to impair response inhibition (BAC: 0.10%: Guillot et al.,
2010; 0.8 g/kg: Dougherty, Marsh-Richard, Hatzis, Nouvion, & Mathias,
2008), and a low dose was sufficient to cause stopping impairment (0.4 g/kg, de Wit, Crean, & Richards,
2000; Nikolaou, Critchley, & Duka,
2013; Reynolds, Richards, & de Wit,
2006). The study design mainly refers to whether alcohol and placebo manipulation is a between-subject or within-subject factor and whether there is a baseline/pre-drink measure. These design options are relevant as individuals differ in their response to alcohol and there is a day-to-day variance of inhibition performance (Campbell, Chambers, Allen, Hedge, & Sumner,
2017). The fact that the three groups were matched in terms of demographics, typical alcohol use, and especially sensitivity to alcohol, made these concerns less vital in our study. Overall, Table S1b revealed that none of these potential factors had a significant effect on research findings (i.e., positive/negative). Therefore, the absence of an alcohol effect on SSRT amongst the population that usually does not drink too much might not readily be attributed to participants’ low typical alcohol use, the use of visual rather than auditory stop signals, the low amount of alcohol administered, and/or between rather than within-subject design without a baseline measure. Note that, statistical analyses that simultaneously take multiple factors into consideration might be more appropriate than
t test for study comparison.
In sum, the effect of alcohol consumption on stimulus-driven inhibition was less robust as one might have expected. In fact, nearly half of the studies that used the stop-signal task failed to identify a significant main effect of alcohol (see Table S1a here, and Table 5 in Bartholow et al.,
2018). By contrast, studies used the cued go/no-go task (Marczinski, Abroms, Van Selst, & Fillmore,
2005) all confirmed the acute alcohol effect (Bartholow et al.,
2018). A potential reason is that the prepotency/urgency of stopping is increased by invalid go cues in the cued go/no-go (Bartholow et al.,
2018). Furthermore, alcohol may influence inhibitory control only during the decreasing limb of BAC (Bartholow et al.,
2018), which helps explain the less apparent effect when the whole BAC curve was considered. As a next step, researchers can consider adding (in)valid cues into the stop-signal task and investigate why alcohol influences inhibition as a function of the BAC curve.