Time to Change: Deciding When to Switch Action Plans during a Social Interaction

Building on the extensive cognitive science literature on the subject, this paper introduces a model of the brain mechanisms underlying social interactions in humans and other primates. The fundamental components of the model are the “Action Observation” and “Action Planning” Systems, dedicated respectively to interpreting/recognizing the partner’s movements and to plan actions suited to achieve certain goals. We have implemented a version of the model including reaching and grasping actions, and tuned on real experimental data coming from human psychophysical studies. The system is able to automatically detect the switching point in which the Action Planning System takes control over the Action Observation System, overriding the automatic imitation behaviour with a complementary social response. With such computational implementation we aim at validating the model and also at endowing an artificial agent with the ability of performing meaningful complementary responses to observed actions in social scenarios.