Abstract

Previous neuroimaging studies of scenarios such as the prisoner’s dilemma and ultimatum game show how brain systems which evolved to process rewarding and aversive
stimuli are activated differentially during adherence to and breaches of norms. One theory of how shared social norms evolve is that they arise from mutually rewarding
conventions which, through repeated execution, acquire a normative value that sustains social cohesion even when individual self-interest is not served. The authors report an fMRI study of a 2-player coordination game in which players must coordinate on an arbitrary convention (a left or right button press) to obtain monetary rewards. Once this
convention has been established, one of the participants is incentivized to deviate from the equilibrium with the offer of extra reward. During the period prior to decisions to
violate the convention, activity was observed in regions associated with reward processing, such as the midbrain, caudate, and orbitofrontal cortex. Activations in advance
of decisions to continue coordinating included the amygdala and anterior insula–inferior frontal gyrus. The data are discussed in the light of theories which propose the
existence of multiple interacting value-based decision-making systems in the human brain.