Dissociating effects of gaze direction and facial motion on memory of dynamic faces

Research on dynamic face processing often emphasizes facial movements while overlooking how social information conveyed by moving faces influences face perception and memory. Our study investigated how eye gaze, a critical social cue, affects the memory of dynamic faces and whether the effect generalizes across facial motion (e.g., rigid and elastic motion) and own- and other-race faces. We employed a recognition memory task across three experiments to address these questions. Experiment 1 compared memory following learning static or moving faces and showed better memory for static and elastically moving faces (i.e., with a direct gaze) than for rigidly moving faces (i.e., with varying gaze directions), more so for own- than other-race faces. Experiment 2 manipulated gaze direction in both elastic and rigid facial motion (i.e., direct vs. averted) and showed a direct-gaze advantage. Learning direct-gaze faces, with either rigid or elastic facial motion, produced consistently better memory performance than that after learning averted-gaze faces, which was more pronounced for own-race faces. Experiment 3 manipulated the congruency of gaze direction between learning and test phases for both rigid and elastic facial motion. The advantage of learning direct-gaze faces persisted, irrespective of gaze congruency between study and test faces. These results not only demonstrate the crucial role of social signals, like eye gaze, in face memory but also dissociate the influences of facial motion and its accompanying gaze directions on the memory of dynamic faces, which may help reconcile discrepant findings regarding the effect of facial motion on face perception and recognition.