Testosterone reduces the threat premium in competitive resource division

Abstract

Like other animals, humans are sensitive to facial cues of threat. Recent evidence suggests that we use this information to dynamically calibrate competitive decision-making over resources, ceding more to high-threat individuals (who appear more willing/able to retaliate) and keeping more from low-threat individuals. Little is known, however, about the biological factors that support such threat assessment and decision-making systems. In a pre-registered, double-blind, placebo-controlled, cross-over testosterone administration study ( n = 118 men), we show for the first time that testosterone reduces the effects of threat on decision-making: participants ceded more resources to high-threat (versus low-threat) individuals (replicating the 'threat premium'), but this effect was blunted by testosterone, which selectively reduced the amount of resources ceded to those highest in threat. Thus, our findings suggest that testosterone influences competitive decision-making by recalibrating the integration of threat into the decision-making process.

Keywords: bargaining; contest; dominance; signalling; social perception.

Figure 1.

Experimental timeline. White numbers represent minutes from the start of the study. (Online version in colour.)

Figure 2.

Examples of the low- and high-threat stimuli used in the current study. Edited versions of photographs reproduced with permission from [25]. (Online version in colour.)

Figure 3.

Bar graph showing the threat premium—the tendency to cede more resources to high than to low threat individuals—as a function of drug administration. Values on the y-axis indicate the amount of additional resources ceded to high relative to low threat faces. This tendency to cede additional resources to high relative to low threat faces was reduced by 57% after testosterone compared with placebo administration. *p < 0.05. (Online version in colour.)

Figure 4.

Line graphs showing the threat premium—the tendency to cede more resources to high-threat (versus low-threat) individuals—under placebo (a) versus testosterone (b). Each line represents the change in resources ceded, from the low- to the high-threat version of a given face identity (i.e. each of the 20 lines corresponds to a different face identity; 20 face identities total). Stronger increases from the low- to the high-threat version of a given face (steeper positive slopes) indicate larger threat premiums, whereas weaker increases (moderate positive slopes) or decreases (negative slopes) indicate smaller threat premiums or threat penalties. (c) Shows the results of the robust multilevel model, with the two lines representing the amount of resources ceded to the high- versus low-threat versions of the faces after placebo (blue) versus testosterone (yellow) administration. The testosterone slope is not as steep as the placebo slope, in part because testosterone selectively decreased resource ceding to the high-threat versions of the faces, thus reducing the threat premium. *p < 0.05. (Online version in colour.)