Félix Duplessis-Marcotte, Valentin Magnon, Marie-France Marin
Testosterone is commonly associated with risk-taking, yet empirical evidence remains mixed. The Somatic Marker Hypothesis proposes that affective physiological reactions to the consequences of a choice can later be reactivated in anticipation of subsequent decisions, thereby biasing decisions. The hypothesis was that testosterone would positively relate to risk-taking by moderating physiological responses occurring prior to risky decisions. Fifty-five adults (27 women) provided a salivary sample to assess testosterone levels before completing the Iowa Gambling Task. In this task, participants were instructed to maximize profits by selecting 100 cards from four decks, two risky (large gains and losses; negative net outcome) and two safe (smaller gains and losses; positive net outcome). Electrodermal activity was recorded to quantify physiological responses both prior to (anticipatory) and following (feedback-related) each card selection. Multilevel logistic models were conducted, and multiverse analyses were used to assess the robustness of the findings. Testosterone was not associated with risk-taking at the inter-individual level. The association between testosterone, anticipatory physiological responses and risk-taking was also not significant. As an exploratory objective, we investigated whether testosterone might relate to risk-taking by moderating feedback-related physiological responses. Multiverse analyses provided moderate evidence that testosterone was associated with feedback-related physiological responses and subsequent risk-taking, but only in men. In fact, in men with lower testosterone levels, physiological responses (following either gains or losses) did not relate to risk-taking, whereas in men with higher testosterone levels, feedback-related physiological responses positively correlated with risk-taking. Moreover, regardless of sex, individuals with higher within-sex testosterone levels exhibited greater risk-taking following losses, a pattern consistent with loss chasing. In line with recent meta-analytical results, testosterone was not associated with risk-taking propensity inter-individually. Yet, at an intra-individual level, testosterone appears to be associated with how feedback-related physiological responses are used to guide decision-making.