Simulating cooperative interactions to investigate the neural correlates of joint attention

Joint attention is a fundamental aspect of social cognition, enabling effective intersubjective experiences. However, due to lacking ecological measures, we currently know little about its neural representation, or how this diverges in developmental disorders such as autism. Joint attention is conceptualised as involving two distinct functions. Responding to joint attention (RJA) involves converging with someone else's focus of attention. Initiating joint attention (IJA), involves guiding others to converge with one's own attentional focus. The Parallel Distributed Processing (PDP) model accounts for the overlapping function, and developmental divergence of RJA and IJA. Consistent with this model, we anticipated that sources of activation associated with RJA and IJA would reflect partially independent regions, existing in a distally distributed, posterior and anterior network. Inspired by the second person perspective (2PP) approach to social cognition research, we developed a co-operative, virtual reality task for the acquisition of fMRI and MEG data. Using eye-tracking techniques, subjects interacted with an onscreen avatar, which responded contingently to their gaze. Subjects completed the task under the impression that they were interacting with a confederate, outside the scanner, who was controlling the avatar's eye movements. Our analyses have focused on comparing neural activation during the execution of RJA and IJA behaviour, as well as during conditions where participants complete the same task 'alone'. Our fMRI univariate whole brain analyses revealed temporoparietal and anterior networks associated with gaze detection, attention orienting and mentalising processes. An MEG case study, using the same task and a novel saccade-related beamforming analysis was also conducted to localise activation associated with preparing saccades for the execution of RJA and IJA. These data and critical issues in designing interactive paradigms will be discussed.