Acute stress alters individual risk taking in a time-dependent manner and leads to anti-social risk

Abstract

Decision-making processes can be modulated by stress, and the time elapsed from stress induction seems to be a crucial factor in determining the direction of the effects. Although current approaches consider the first post-stress hour a uniform period, the dynamic pattern of activation of the physiological stress systems (i.e., the sympathetic nervous system and hypothalamic-pituitary-adrenal axis) suggests that its neurobehavioural impact might be heterogeneous. Here, we evaluate economic risk preferences on the gain domain (i.e., risk aversion) at three time points following exposure to psychosocial stress (immediately after, and 20 and 45 min from onset). Using lottery games, we examine decisions at both the individual and social levels. We find that risk aversion shows a time-dependent change across the first post-stress hour, evolving from less risk aversion shortly after stress to more risk averse behaviour at the last testing time. When risk implied an antisocial outcome to a third party, stressed individuals showed less regard for this person in their decisions. Participants' cortisol levels explained their behaviour in the risk, but not the antisocial, game. Our findings reveal differential stress effects in self- and other-regarding decision-making and highlight the multidimensional nature of the immediate aftermath of stress for cognition.

Keywords: behavioural economics; behavioural neuroscience; decision-making; resilience; stress.

Figure 1

The stress induction protocol successfully induced a stress response in stress‐group subjects. Subjects performed either a TSST‐G stress procedure (TSST‐1 and TSST‐2 denote the respective interview and mathematical portions of the TSST‐G stressor) or control procedure and then performed tasks as outlined (A). Subjects in the stress group reported significantly higher levels of subjective stress during each saliva measurement (B), exhibited increased heart rates (C) and cortisol levels (D) during the experiment compared to controls. A portion of subjects exhibited a pattern of cortisol response such that they could be classified as either responders or non‐responders, with non‐responders demonstrating similar cortisol levels to control groups, and responders exhibiting significantly higher levels. Data are presented as mean and standard error, except for (E) which depicts the means. [Colour figure can be viewed at wileyonlinelibrary.com].

Figure 2

A schematic representation of the game used to assess risk preferences, under both standard risk and anti‐social conditions. Participants (denoted here as Agents) provided a switching probability ‘p(switch)’ at which they would accept a minimum probability of winning to enter a lottery for 20 CHF (Option B) over the certain gain of 10 CHF (Option A).

Figure 3

Stress caused a fluctuation of risk aversion over time that is antisocial. Switching probabilities are plotted for decisions taken at three different time points after stress induction for both stressed and non‐stressed control subjects. Higher probabilities indicate higher risk aversion. Error bars represent Standard Error of the Mean (SEM).