The use of dynamic cues in self and familiar face recognition

Familiarity plays an important role in face processing. The importance of familiarity is increased when facial form cues are degraded, so that a person must rely primarily on movement (dynamic) information to identify someone. It is, however, unclear which dynamic cues are used for face recognition of both familiar and unfamiliar faces. Furthermore, little work has been done on dynamic self-face recognition, and none has focused on the type of movement that facilitates this process. The current study used motion capture cameras to record and isolate facial movements in order to test recognition of self, familiar and unfamiliar faces. Participants completed a 2AFC same/different face-matching task involving point-light displays of natural motion (i.e. both rigid and non-rigid motion), rigid motion only (e.g. nodding/shaking), non-rigid motion only (e.g. mouth/eyebrow motion) and still images to determine whether differences in familiarity resulted in the use of different movement cues. The manner (style) in which someone is speaking may also impact on whether they can be easily identified from dynamic cues. Consequently, speech style was either matched or mismatched between video clips. We found that matching performance was more accurate overall when speech style was matched than mismatched. Familiar face matching appears to use rigid, non-rigid and natural movement cues equally, but unfamiliar and self-face matching are more accurate for rigid than natural motion when speech style differs between clips. These results are discussed in relation to previous research on dynamic face recognition, and possible implications for current face processing models.