Introduction: Intermittent Explosive Disorder (IED) is a psychiatric syndrome characterized by aggressive behavior (APA, 2013). Despite the distress and impairment associated with the disorder, little is known about what factors lead to IED. However, anger-based aggression is one of the primary symptoms of IED (APA, 2013), so research on aggression may inform our understanding of IED. We hypothesized that people with high levels of anger-proneness and aggression may have a bias toward perceiving anger. That is, relative to other people, they may be faster to detect faces exhibiting anger versus other emotions.
Method: We conducted an emotion-perception study on Amazon Mechanical Turk (AMT). Participants were 85 highly-rated AMT workers. After providing informed consent, participants watched 30, 10-second videos of faces morphing from neutral into one of five emotions: anger, contempt, fear, joy, and sadness. They clicked as soon as they identified an emotion. The face then disappeared and they identified the emotion. Finally, participants completed self-report measures of anger-proneness, reactive aggression, proactive aggression, and social anxiety symptoms.
Results: We used mixed linear models to examine anger-proneness, reactive aggression, and proactive aggression as predictors of emotion detection speed for correctly identified trials, controlling for social anxiety symptoms. Higher anger proneness predicted slower reaction speed to faces exhibiting anger (b=.31, p=.007), fear (b=.34, p=.002), and sadness (b=.28, p=.003), but not contempt or joy (ps=.012-.860). Proactive aggression predicted faster reaction speed to faces exhibiting anger (b=-.43, p < .001), contempt (b=-.45, p < .001), fear (b=-.51, p < .001), joy (b=-.23, p < .001), and sadness (b=-.39, p < .001). Reactive aggression did not predict emotion detection speed (ps=.07-.35). We used multiple regression to examine anger proneness, reactive aggression, and proactive aggression as predictors of emotion detection accuracy. Higher anger-proneness predicted lower accuracy for joy (b=.34, p=.012) but not other emotions (ps= .26-.86). Proactive aggression predicted lower accuracy for anger (b=-0.04, p<.001), contempt (b=-.03, p=.05), fear (b=-.06, p<.001), joy (b=-.04, p < .001), and sadness (b=-.04, p<.001). Reactive aggression did not predict emotion detection accuracy (ps=.265-.927).
Discussion: Inconsistent with our hypothesis, we found that people with high anger proneness are slower to detect anger, fear, and sadness, whereas people with high proactive aggression are faster to detect all emotions but are less accurate. These results suggest that anger proneness and proactive aggression have distinct cognitive correlates and thus may require different treatment approaches.