A fronto-temporo-parietal network disambiguates potential objects of joint attention

We use the other’s gaze direction to identify her/his object of interest and to shift our attention to the same object, i.e. to establish joint attention. However, gaze direction may not be sufficient to unambiguously identify the object of interest as the other’s gaze may hit more than one object. In this case, the observer must use a priori information to disambiguate the object choice. Using fMRI, we suggest that the disambiguation is based on a 3-component network. A first component, the well-known ‘gaze following patch’ in the posterior STS is activated by gaze following per se. BOLD activity here is determined exclusively by the usage of gaze direction and is independent of the need to disambiguate the relevant object. On the other hand, BOLD activity revealing a priori information for the disambiguation and starting early enough to this end is confined to a patch of cortex at the inferior frontal junction. Finally, BOLD activity reflecting the convergence of both, a priori information and gaze direction, needed to shift attention to a particular object location is confined to the posterior parietal cortex.