Neurocomputational Dynamics of Cooperation

Problems in cooperating with others define antagonism, a transdiagnostic dimension of psychopathology expressed in personality and externalizing disorders. People high on antagonism are prone to aggressive, callous, and exploitative behaviors that damage relationships and incur substantial costs in the form of productivity loss, substance misuse, and criminality. While clinical theories link antagonism to deficits in social cognition, it is hard to deny that exploiting others requires an understanding of their mind. This apparent contradiction can be resolved by distinguishing between two facets of antagonism: callousness and exploitativeness. Prior work by the investigators and preliminary data inform the hypothesis that exploitativeness reflects an enhanced capacity to learn about the mental states and behaviors of others coupled with an increased sensitivity to personal, rather than shared, rewards. In contrast, callousness reflects a broader insensitivity to social feedback. The investigators hypothesize that these dissociable social learning processes are instantiated in canonical circuits that expanded during anthropogenesis to support mentalizing and self-directed thought. These learning signals, which track behaviorally salient features of the social environment, are integrated in the posterior hub of the default network (DN) where they drive adaptive cooperation. This study proposes to test these hypotheses in two samples elevated on antagonism. At the behavioral level, participants are characterized using a battery of personality and psychopathology measures, social and reward-based decision-making tasks, and corresponding hierarchically estimated reinforcement learning (RL) models (Aim 1). In a model-based functional magnetic resonance imaging (fMRI) study, model-derived learning signals and facets of antagonism are linked to underlying neural activity during cooperative decision-making (Aim 2). Leveraging the multi-timescale design of our study, the investigators plan to relate neurocomputational signatures of callousness and exploitativeness to mentalizing abilities, multi-informant reports of momentary interpersonal behavior, and prospective stress generation. Certain elements of study design are withheld from this record to protect the scientific integrity of the study design.