Abstract
Recent studies indicate that memory for familiar/famous faces can be improved when the faces are viewed in motion. For relatively unfamiliar faces, however, the role of motion in face memory is less clear. Several studies have yielded inconsistent results, suggesting the possibility that different types of motion may affect recognition accuracy in different ways. We examined the effects of non-rigid facial speech movements (Exp. 1) and rigid face rotation movements (Exp. 2) on recognition accuracy.
In Experiment 1, faces were presented and tested as either moving video clips or as still frontal images. The moving video clips were non-audible frontal faces talking directly to the viewer. Both the video and static presentations were matched for exposure time. The results indicated that face recognition was less accurate for the moving faces than for the static faces.
In Experiment 2, recognition accuracy was assessed in four conditions. In the first condition, participants viewed video clips of individuals rotating their head, from left to right profile, through nine equally spaced viewpoints. In the second condition, participants viewed still frontal images of faces. To control for the additional view information available in the motion clips, a third and fourth group of participants learned faces from the nine static images taken from viewpoints used in the video. These were presented in sequential or random order. Although recognition memory for moving faces was better than memory for the static frontal images, the availability of additional view information accounted for this advantage.
In conclusion, not all facial motions have similar effects on recognition. A plausible interpretation of the recognition disadvantage found with facial speech is that processing socially relevant facial motions may distract people from paying attention to the identity of a face.
Supported by a DOD/DARPA grant to A.O & H.A.