To the Editor:

We enthusiastically read Gorka et al. [1]. Earlier laboratory work from both their group and authors of this letter (e.g., refs. [2, 3]) highlight stressor uncertainty as a key moderator of alcohol’s long reported but inconsistently observed stress response dampening effects. Importantly, this research identifies exaggerated stress reactivity to uncertain stressors as either a consequent “disease marker” of repeated alcohol use (i.e., drug induced neuroadaptation; [3]) or, as suggested by Gorka et al. [1], a pre-morbid risk factor for alcohol use disorder. Gorka et al. [1] provided critical adjudication between these competing hypotheses using a rigorous longitudinal approach and a well-validated psychophysiological task. Though the sample’s preexisting alcohol exposure limits evidence for premorbidity, Gorka et al.’s [1] study provides intriguing data for the premorbid risk factor hypothesis which may inform intervention efforts for excessive alcohol use.

Gorka et al. [1] speculated that those “hyper-reactive to [uncertain threat] express vulnerability to coping oriented motives, or ‘self-medication,’” and, “The subjective… relief gained from intoxication can contribute to [alcohol] expectancies … and promote coping-motives for alcohol use.” Although speculation about potential clinical implications of basic research is common, empirical confirmation of even a tenuous connection between implicit measures (e.g., fMRI & startle) with drinkers’ subjective perceptions, cognitions, and subsequent drinking is sparse. Therefore, we agree with Gorka et al. [1] that reactions to uncertainty “can be captured at multiple units of analysis and similarly relate to real-world drinking behaviors.” and we propose this relationship will only be established if the units of analysis include subjective expectations and experience ([2, 4]). This will not only better connect neurobiology to real-world behavior, but also inform research in related sub-fields (e.g., expectancy).

Demonstrating this in an undergraduate sample similar to Gorka et al.’s [1], we used a novel experimental affective forecasting task (https://osf.io/pskg7/) wherein 158 drinkers (Mean (SD) age = 19.7 (1.8); 63% female, 1% nonconforming; past-month binge-drinking = 67.1%;) reported expectations about their anxiety and alcohol’s effects during imaginal stressor exposures based on social anxiety measures (e.g., ref. [5]). Exposure presentation was counterbalanced, once with a relatively uncertain negative outcome and once with a relatively certain one (see https://osf.io/pskg7/ for details). Participants forecasted (1–100) significantly less anxiety when they imagined being intoxicated relative to sober for uncertain scenarios (ηp² = 0.81,b = −21.58, t(156) = −26.31, p < 0.001) than for certain scenarios (ηp² = 0.82, b = −19.76, t(156) = −25.24, p < 0.001); interaction (ηp² = 0.07, b = −1.82, t(156) = −3.41, p < 0.001; Fig. 1). Neither gender nor binge-drinking moderated results (p’s > 0.645). This complements Gorka et al. 2023 and related work (e.g., refs. [2, 3]) by suggesting that youth indeed expect alcohol to lower their anxiety more in relatively uncertain compared to certain social situations. Such nuance has not been a focus in alcohol expectancy research and its discovery was motivated by neuropsychopharmacology research like Gorka et al.’s [1].

Fig. 1: Effects of stressor certainty and imagined intoxication.
figure 1

Confidence bars represent + one standard error for point estimates from the GLM. GLM coefficients show the simple effect of imagined intoxication for each stressor type. (*p < 0.05). Figure © 2023 Daniel E. Bradford, Amelia S. Dev, Angelica DeFalco, Freya Whittaker, Kiara R. Timpano under Creative Commons Attribution 4.0 International Public License CC-By.

Full size image

These results are an initial step towards connecting Gorka et al.’s and our own lab’s findings to drinking motivations. Only through simultaneously measuring varied response domains (e.g., neural; self-reported expectations; overt choice), phenomena that predicts drinking (e.g., salience and craving [6]), manipulating additional stressor characteristics (e.g. uncontrollability [2]), in and outside the laboratory (e.g., mobile psychophysiology and ecological momentary assessment) we will strengthen these connections and build a robust foundation for transformative understanding of problematic alcohol use. This may support therapeutic development targeting neurotransmitter systems involved in uncertain stressor reactivity and/or (re)appraisal skills for decreasing stressor uncertainty specifically rather than stress-coping skills broadly.