State, not trait, neuroendocrine function predicts costly reactive aggression in men after social exclusion and inclusion.

Social exclusion increases aggressive behaviour, and the possible neuroendocrine underpinnings of the effect are largely unknown. Here, we examined the extent to which testosterone and cortisol responses to social exclusion would predict subsequent reactive aggression. Men were randomly assigned to a social exclusion (SE) or inclusion (SI) condition of 'Cyberball', a computer ball-toss game. Aggression was then measured using the Point Subtraction Aggression Paradigm (PSAP). Saliva was collected at three points for the measurement of testosterone and cortisol. Regression analyses indicated that testosterone concentrations 10-min into the PSAP (controlling for pre- and post-Cyberball testosterone) were positively correlated with aggressive behaviour, irrespective of SI/SE. Post hoc analyses for the conditions separately, however, suggested the relationship was stronger for SI men (R(2)(change)=13.3%, F(1,)(29)=5.28, p=0.03) than for SE men (R(2)(change)=1.8%, F(1,)(26)=0.49, p=0.49). Aggressive behaviour was also positively correlated with cortisol concentrations 10-min into the PSAP (controlling for pre- and post-Cyberball cortisol) irrespective of SE/SI. When both hormones were included in the regression model, the interaction of baseline 'Cortisol'×'Testosterone'×'Experimental Group' approached significance (R(2)(change)=5.4%, F(1,)(55)=3.53, p=0.07), but no significant effects were observed in either group alone. The findings add to evidence that individual differences in state neuroendocrine function map onto variability in human social behaviour.