Social neuroendocrinology of competition
Abstract
The relationship between testosterone concentrations and aggressive behaviour in studies
of people has produced very inconsistent findings. However, one consistent fmding that
has emerged is that competitive and aggressive interactions potentiate testosterone release
in both human and non-human species. It has been argued that socially-induced
alterations in testosterone concentrations may function to influence ongoing and/or future
social behaviour. Nonetheless, few studies have empirically tested this hypothesis. The
current series of experiments was designed to address the extent to which competitioninduced fluctuations in testosterone concentrations were associated with ongoing and/or
subsequent social behaviour. In Study 1, men (n = 38) provided saliva samples prior to,
and at the conclusion of, the Point Subtraction Aggression Paradigm (PSAP). Although
baseline testosterone concentrations were not related to aggressive behaviour, there was a
positive correlation between change in testosterone and aggressive behaviour such that
men who were most aggressive on the PSAP demonstrated the largest increase in
testosterone concentrations. Furthermore, a rise in testosterone during the PSAP predicted
willingness to choose a subsequent competitive task. In Study 2, men and women
provided saliva samples prior to and after competing against a same-sex opponent on the
Number Tracing Task (NTT). The outcome of the competition was rigged such that half
of the individuals won most of the races, while the other half lost most of the races, thus
experimentally creating a winner and loser in the laboratory. Following the competitive
interaction, men and women played the PSAP with their same-sex partner. Results
indicated that men selected the aggressive response (but not reward or protection
responses), more frequently than women. For men assigned to the loss condition, an increase in testosterone concentrations in response to the NTT predicted subsequent
aggressive behaviour. For men assigned to the win condition, an increase in testosterone
concentrations in response to the NTT predicted subsequent aggressive behaviour, but
only among those men who scored high on trait dominance. Change in testosterone and
trait dominance did not predict aggressive behaviour in women. In Study 3, men provided
saliva samples prior to, during, and at the end of the PSAP. They were randomly assigned
to one of four experimental conditions that differed in the extent to which they were
provoked and whether they received reward for behaving aggressively (i.e., stealing
points). Results indicated that baseline testosterone concentrations did not correlate with
aggression in any of the experimental conditions. Consistent with Study 1, there was a
positive correlation between change in testosterone and aggressive behaviour among men
who were provoked, but did not receive reward for aggression (i.e., reactive condition).
Men who were provoked but did not receive reward for aggression enjoyed the task the
most and were more likely to choose the competitive versus non-competitive task relative
to men assigned to the other experimental conditions. Also, individual differences in
aggressive behaviour among these men were positively correlated with the extent to
which they enjoyed the task. Together, these studies indicate that testosterone dynamics
within the context of competition influence subsequent competitive and aggressive
behaviours in humans and that testosterone may be a marker of the intrinsically rewarding
nature of costly aggressive behaviour.