Short Communication
Effects of hydrocortisone and yohimbine on decision-making under risk

https://doi.org/10.1016/j.psyneuen.2020.104589Get rights and content

Highlights

  • We investigated the effect of stress hormones on decision-making under risk.

  • Yohimbine and hydrocortisone were administered before a decision-making task.

  • Hydrocortisone reduces risk taking in gain-only trials.

  • Yohimbine has no effect.

  • Hydrocortisone might decrease reward processing.

Abstract

Introduction

Many studies have investigated the influence of stress on decision-making. However, results are equivocal and the exact role of increased noradrenaline and cortisol after stress remains unclear. Using pharmacological manipulation, we investigated the influence of noradrenergic and glucocorticoid activity on risky decision-making in a gambling task that included mixed-gamble trials (gains and losses are possible) and gain-only trials.

Methods and materials

One hundred-and-four healthy young men participated in our randomized, double-blind, placebo-controlled, between-group study. Participants were randomly assigned to one of four groups: (A) yohimbine, (B) hydrocortisone, (C) yohimbine and hydrocortisone, or (D) placebo. Frequency of risky choices, i.e., monetary risk taking, was the dependent variable. We also investigated the influence of hydrocortisone and yohimbine on loss aversion, which is the tendency to overweigh losses compared with gains.

Results

Participants chose the risky option less often after receiving hydrocortisone compared with no hydrocortisone. This effect was strongest in the gain-only trials. Yohimbine had no effect. Loss aversion was not affected by hydrocortisone or yohimbine.

Discussion

Decreased reward processing may explain the reduction of risk taking by hydrocortisone in gain-only trials. The effects of stress hormones on different decision-related constructs and processes hence require further investigation.

Introduction

Cumulative prospect theory states that individuals translate an objective value into a subjective value through their perception (Tversky and Kahneman, 1992). Values are processed as gains or losses, and losses are typically overweighed compared with gains of the same amount – a phenomenon called loss aversion. Loss aversion influences decision-making under risk, which refers to making decisions when the outcome is uncertain but the probabilities of possible outcomes are known. In addition, choice consistency and attitude towards risk (i.e. outcome variability) influence decision-making under risk (Sokol-Hessner et al., 2016).

There is evidence that acute stress alters risky decision-making. However, results are contradictory and the majority of studies did not disentangle the different processes involved in decision-making, e.g. loss aversion and attitude towards risk. Stress might lead to more reward-seeking, and thus more risk taking (Nowacki et al., 2019; Starcke and Brand, 2016). Furthermore, several pharmacological studies have shown that hydrocortisone (Putman et al., 2010) and fludrocortisone (Deuter et al., 2017) are associated with riskier decisions. However, less gambling under stress (Cahlíková and Cingl, 2017; Porcelli and Delgado, 2017) and no changes in gambling (Von Dawans et al., 2012) have also been described. In addition, factors like gender (Lighthall et al., 2012) and riskiness of the decision (Von Helverson and Rieskamp, 2013) might moderate these effects.

The stress response is mainly characterized by two stress neuromodulators. First, stress leads to a rapid release of noradrenaline and adrenaline through the sympathetic nervous system (Hermans et al., 2014). Second, a slower release of glucocorticoids via the hypothalamic-pituitary-adrenal (HPA) axis takes place (Hermans et al., 2014).

Two brain networks are related to adaptive decision-making and are influenced by these stress neuromodulators: the executive network, which includes the prefrontal cortex (PFC), responsible for probabilistic reasoning (Domenech and Koechlin, 2015), and the salience network, which includes the amygdala, responsible for processing losses (Margittai et al., 2018). These networks are affected differently by glucocorticoid and noradrenergic activity (Hermans et al., 2014), so stress can affect decision-making via at least two possible pathways (Margittai et al., 2018). Hydrocortisone decreases executive network activity (Oei et al., 2007), whereas both stress neuromodulators increase salience network activity (Hermans et al., 2014). According to the ‘salience-of-losses’ hypothesis, combined noradrenaline and cortisol amplify loss aversion by reallocating resources to the salience network, promoting fear and consequently enhancing salience of losses at the expense of the executive control network (Margittai et al., 2018).

In contrast, according to the ‘alignment’ hypothesis, loss aversion is diminished because reward and threat susceptibility are aligned (Margittai et al., 2018). Here, combined yohimbine and hydrocortisone administration increase reward processing by activating the ventral striatum, especially in subjects with very high cortisol levels. Margittai and colleagues (2018) observed a significant reduction in loss aversion when 20 mg hydrocortisone was combined with 20 mg yohimbine compared with the administration of either drug alone.

The aim of the current study was to examine whether the findings of Margittai and colleagues (Margittai et al., 2018) hold with using lower dosages of yohimbine and hydrocortisone but a similar gambling task including mixed-gamble (gains and losses are possible) and gain-only trials. Since the majority of studies hint toward riskier decisions irrespective of dosage and stress induction, we hypothesized similar results as described by Margittai and colleagues (Margittai et al., 2018), that is, reduced loss aversion after combined administration of yohimbine and hydrocortisone. This pattern would be in line with the ‘alignment’ hypothesis and the majority of studies hinting towards riskier decision-making after stress induction. According to our hypothesis, participants were expected to choose the gambling option more frequently in the mixed-gamble trials because of decreased loss aversion.

Section snippets

Participants

One hundred-and-four healthy men with body mass index < 30 between 18 and 35 years old (mean [standard deviation] age, 24.1 [3.5] years) currently enrolled at a university participated in the study. All participants were native German speakers and unmedicated. Exclusion criteria included: CNS or somatic diseases, metabolic or endocrine diseases, current infection or autoimmune diseases, and current or past psychiatric disorders screened as per axis I of DSM-IV (based on SCID I Screening

Sample characteristics

There were no significant differences in demographics (apart from BMI) between treatment groups. Obesity was an exclusion criterion and the number of overweight participants did not differ between groups (Chae et al., 2019) (for details see Table S1).

Decision-making task

Since BMI differed between groups, we included BMI as a covariate in all further analysis.

Choice frequency: There was a significant main effect of hydrocortisone (F1, 99 = 9.62, p =.005, η2 =.07) and a significant interaction effect between

Discussion

We investigated the influence of noradrenergic and glucocorticoid activity on decision-making under risk. After receiving hydrocortisone, participants chose less often to gamble in the gain-only trials, indicating less risk-seeking for gains.

The aim of the current study was to examine whether the findings of Margittai and colleagues (Margittai et al., 2018) hold with using lower dosages of yohimbine and hydrocortisone. Contrary to the ‘alignment’ hypothesis, we did not find an interactive,

Declaration of Competing Interest

None.

References (21)

There are more references available in the full text version of this article.

Cited by (16)

  • Stressed individuals exhibit pessimistic bursting beliefs and a lower risk preference in the balloon analogue risk task

    2022, Physiology and Behavior
    Citation Excerpt :

    This translates into increased activity of regions such as the insula or the amygdala, which are key nodes in the detection and response to negative or threatening stimuli [13, 25, 44]. Thus, the other prominent hypothesis on how stress could influence our decisions is the salience-of-losses hypothesis [25, 28], according to which the psychological impact of negative stimuli would be amplified and, in line with our results, more conservative behaviors would be promoted. This was also observed by Pabst et al. [34], where 5 and 18 min after the original TSST onset, participants took fewer risks in the GDT; in contrast, when the cortisol peak was reached at 28 min, risk propensity was reversed and riskier decisions were found.

  • Sunk costs under stress: Acute stress reduces the impact of past expenses on risky decisions

    2022, Psychoneuroendocrinology
    Citation Excerpt :

    Although the investment task was performed under stress-induced elevations of cortisol concentrations and individual cortisol reactivity also predicted the strength of the reduction of the sunk-cost effect, we cannot rule out that the effect is mediated by other stress-related factors. Future studies could use pharmacological manipulations (see, e.g., Metz et al., 2020) to test more directly whether cortisol causally contributes to this effect. Furthermore, given that the influence of stress can be time-dependent (Hermans et al., 2014; Joëls et al., 2011), earlier or later stages of the cortisol response (or rapid sympathetic action) could be associated with different effects, mediated through different mechanisms (e.g., non-genomic vs. genomic glucocorticoid actions).

  • Early stages of the acute physical stress response increase loss aversion and learning on decision making: A Bayesian approach

    2021, Physiology and Behavior
    Citation Excerpt :

    This result supports our hypothesis that stress would increase loss aversion. In turn, this could be in line with the salience of losses hypothesis [28]. This hypothesis states that stress produces a relocation of the brain resources in favor of the amygdala and in detriment of the prefrontal cortex.

View all citing articles on Scopus
View full text