Discussion
Previous research has shown that the SPN is robust to experimental variations of task, spatial attention, and visual memory load. Here we investigated whether the SPN is also robust to alcohol induced changes in mental state for the first time. A pilot Oddball task found a surprising alcohol induced SPN enhancement. In contrast, the new Oddball task showed a small alcohol induced SPN reduction. There was no effect of alcohol on SPN amplitude in the new Regularity task. It could be that task-relevant symmetry processing in the Regularity task is more robust and can thus survive moderate doses of alcohol.
However, exploratory analysis indicated that task differences are partly explained by individual differences in drinking behaviour. Participants who drink more show an alcohol induced SPN enhancement, participants who drink less show an alcohol induced SPN suppression. By chance, the participants in the Oddball task were lighter drinkers. This aspect of the results requires replication. However, it means we cannot overinterpret task differences.
Despite the remaining questions, we can confidently state that there was a large SPN in all tasks and conditions. The SPN is probably altered by alcohol in a complicated way, which depends partly on the task and partly on individual differences in drinking behaviour, but it was never substantially reduced, let alone abolished.
In contrast, N1 was consistently reduced by alcohol in the Pilot study, the Oddball task, and the Regularity task. We conclude that alcohol induced changes in N1 amplitude are more consistent and reliable than alcohol induced changes in SPN amplitude. The N1 is also generated by mid-level visual processes involved in perceptual grouping and gestalt formation, and these initial visual operations may be more directly sensitive to alcohol that the subsequent SPN.
Interestingly, stimulus regularity and alcohol consumption both affected N1 amplitude, but these effects did not interact. This suggests that early symmetry processing during the N1 time window is not strongly altered by alcohol. Moreover, there was some weak evidence that the P1 was enhanced by alcohol, and that the P300 in the Oddball task was reduced by alcohol. These two effects were much smaller than the alcohol induced N1 reduction.
Our reliable N1 results are broadly consistent with the attention-allocation model (Sayette, 1999). Alcohol may narrow attention, so potential distractions go unnoticed. This lets drinkers live in the moment, untroubled by stressful memories and duties. Distracting visual stimuli are also more likely to remain unconscious if activations that generate the N1 are dampened.
This project highlights the importance of replication in cognitive neuroscience. It would probably have been possible to publish the pilot results by rhetorically blurring the distinction between confirmatory and exploratory research. The narrative ‘alcohol enhances symmetry sensitivity’ is readily understandable and leads to interesting discussions about aesthetic experience. For instance, it has been suggested that alcohol makes faces look more beautiful because it makes them look more symmetrical (Halsey et al., 2010; Souto et al., 2008). Our pilot results were consistent with this. However, hasty publication of unexpected results from small sample experiments has a net negative effect on reproducibility (Bishop, 2019; Button et al., 2013; Munafò et al., 2017). We were keen to avoid this, and a strength of the current work is that we replicated the pilot experiment and expanded it with a new regularity task. While we are now less confident that the SPN is systematically altered by alcohol consumption, we have not polluted the literature with an eye-catching fluke!
The current work adds to previous studies that have found an SPN in oddball tasks (Höfel & Jacobsen, 2007, Makin et al., 2013). Here we went one step further and found an SPN in an Oddball task when participants were intoxicated. However, this automaticity may be qualified by known boundary conditions: Weaker regularities, such as symmetry embedded in noise, might not produce an SPN under these conditions (Makin et al., 2020). Moreover, SPNs generated by other kinds of visual symmetry, such as rotation, translation, and Glass patterns, might be more vulnerable to task and alcohol manipulations. Most importantly, these results only generalize to symmetry in the retinal image. This is an important limitation: Many objects are symmetrical, but they only project a symmetrical image onto the retina when viewed in the frontoparallel plane (Sambul et al., 2013; Sawada & Pizlo, 2008). Symmetrical objects do not project a symmetrical image when viewed in perspective. Perspective symmetry can be detected when it is task relevant (Bertamini et al., 2022) but may not be processed automatically when it is not relevant (Keefe et al., 2018; Makin et al., 2015; Rampone et al., 2019).