To approach or avoid? The role of risk and ambiguity in threat–reward decisions

Introduction

Avoidance behaviour is a core mechanism of anxiety and depression and may contribute to symptom maintenance by preventing learning and reinforcing threat expectancy. Understanding the factors that influence approach-avoidance decisions is therefore important for clarifying mechanisms underlying anxiety-related psychopathology. Decision-making under uncertainty, particularly risk and ambiguity, may play an important role in these behaviours. Although previous research has linked altered risk processing to anxiety and depression, it remains unclear whether individual differences in risk proneness and ambiguity aversion can predict behaviour in approach-avoidance conflict situations. The present study investigated whether behavioural tendencies measured in the Wheel of Fortune task (WoF) predict approach and avoidance behaviour in the Approach-Avoidance Conflict Task (AACT).

Methods

Adults from the general population in Sweden were recruited online (N = 570, 86% women, mean age ± SD = 39.05 ± 12.42 years) and completed a battery of behavioural tasks and self-report questionnaires assessing decision-making processes and affective symptoms. In the AACT, participants decided whether to approach rewarding yet threatening situations (1, 5, or 20 points) or avoid threat at the cost of reward (0 points). Additionally, they completed the WoF, designed to assess attitudes towards risk and ambiguity in a gambling paradigm. Individual estimates of risk and ambiguity were obtained with a hierarchical Laplace-Gibbs sampler. To examine whether risk and ambiguity attitudes predicted approach behaviour in the AACT, generalised linear mixed-effects models were applied with participant ID as a random intercept and including reward, risk, ambiguity, and their interaction effects as predictors of approach decisions.

Results

The approach proportion in the AACT was significantly associated with reward, with higher rewards increasing the likelihood of approaching the aversive stimuli. This relationship was moderated by individual differences in both risk and ambiguity attitudes measured in the WoF. Specifically, individuals displaying greater risk-prone behaviour showed increased sensitivity to reward, reflected in a positive interaction between reward and risk on approach decisions (reward × risk interaction: OR = 1.56, 95% CI [1.25, 1.94], p < .001). Similarly, individuals displaying greater ambiguity aversion showed increased sensitivity to reward, reflected in a positive interaction effect (reward × ambiguity interaction: OR = 1.26, 95% CI [1.06, 1.51], p = .009). These results indicate that the sensitivity to reward in AACT increases with risk proneness and ambiguity aversion in WoF.

Conclusions

The findings suggest that, rather than serving as a direct predictor, individual differences in risk proneness and ambiguity aversion moderate sensitivity to reward during approach–avoidance conflict. These results support the idea that cognitive and behavioural mechanisms underlying uncertainty processing are relevant for understanding maladaptive avoidance patterns associated with anxiety and depressive symptoms. Identifying such decision-making tendencies may improve our understanding of the mechanisms contributing to psychopathology and help inform future computational or behavioural models of anxiety-related avoidance.

References

No references