Neural mechanisms of recognizing scene configurations from multiple viewpoints

Using functional magnetic resonance imaging (fMRI), the present study examined the neural mechanisms involved in recognizing spatial configurations of a scene from multiple viewpoints. Prior to scanning, participants were instructed to learn a desktop array of seven objects relative to an intrinsic direction that was different from the participants' viewpoint. During scanning, participants recognized triplets of objects from the previously memorized scene and from a mirror reflection of the scene at different perspectives. Half of the triplets included two objects located along the instructed intrinsic direction (intrinsic triplets) and the other half did not (non-intrinsic triplets). Consistent with previous mental rotation studies, bilateral intraparietal sulcus and bilateral middle frontal gyrus showed increasing activation with the angular disparity between the test view and the study view. The right intraparietal sulcus was more activated to the non-intrinsic triplets than the intrinsic triplets. The anterior cingulate cortex was more deactivated in recognizing non-intrinsic triplets and novel views. These findings are consistent with the behavioral results that recognition was easier for intrinsic triplets than for non-intrinsic triplets and easier for the familiar view than for novel views (Mou et al., 2008a).