Neurophysiological contributors to advantageous risk-taking: an experimental psychopharmacological investigation

Abstract

The ability to learn from experience is critical for determining when to take risks and when to play it safe. However, we know little about how within-person state changes, such as an individual's degree of neurophysiological arousal, may impact the ability to learn which risks are most likely to fail vs succeed. To test this, we used a randomized, double-blind, placebo-controlled design to pharmacologically manipulate neurophysiological arousal and assess its causal impact on risk-related learning and performance. Eighty-seven adults (45% female, Mage = 20.1 ± 1.46 years) took either propranolol (n = 42), a beta-adrenergic receptor blocker that attenuates sympathetic nervous system-related signaling, or a placebo (n = 45). Participants then completed the Balloon Emotional Learning Task, a risk-taking task wherein experiential learning is necessary for task success. We found that individuals on propranolol, relative to placebo, earned fewer points on the task, suggesting that they were less effective risk-takers. This effect was mediated by the fact that those on propranolol made less optimal decisions in the final phase of the task on trials with the greatest opportunity for advantageous risk-taking. These findings highlight that neurophysiological arousal supports risk-related learning and, in turn, more advantageous decision-making and optimal behavior under conditions of risk.

Keywords: arousal; beta-adrenergic blockade; learning; propranolol; risk-taking.

Fig. 1.

Points earned by placebo and propranolol groups across the task. There was a main effect of drug, F(1, 84) = 4.85, P = 0.030, partial η² = 0.055, such that participants on propranolol (M = 337.74, SD = 72.50) earned fewer points across the task than those on placebo (M = 373.96, SD = 82.44). Error bars are standard errors.

Fig. 2.

Pumps across task phases, split by drug group, within the certain-long balloon. The difference in pumps between drug groups was only significant for the certain-long balloon in the final phase of the task, F(1, 84) = 4.39, P = 0.039, η² = 0.050, and not the first two task phases (P > 0.05), such that the propranolol group (M = 66.29, SD = 29.72) pumped less in the final third of the task than did the placebo group (M = 79.13, SD = 29.93). Error bars are standard errors.

Fig. 3.

Mediation model. The link between drug group and higher overall BELT performance scores (total accumulated points across the entire task) was mediated via a significant indirect (a*b) effect of pumps made in the late phase with the certain-long balloon, estimated between 95% CIs [−61.78, −2.17]. Because zero was not within the 95% CI, the indirect effect was significantly different from zero at P < 0.05. Note all paths are significant but see “Results” for specific details.