Abstract
Facial movements enhance the quality of our everyday social interactions. A raised eyebrow, sidelong glance, or nodding head can add meaningful content to a spoken message and may enrich our perception of the structure of a face. The effect of facial movement on memory for faces, however, is unclear. Recent studies have employed a variety of methods and facial motions, but have yielded inconsistent findings. We propose a model that synthesizes and consolidates these divergent psychological findings into the neural framework outlined by Haxby et al. (2000).
In combination, the psychological data we review suggests that: 1) dynamic information contributes more to recognition under non-optimal viewing conditions (e.g., poor illumination or recognition from a distance); 2) dynamic information contributes more as a viewer's familiarity with a face increases; and 3) facial motion can contribute to a perceptual structure-from-motion analysis by bootstrapping the encoding of the invariant facial structure. Poor viewing conditions may increase the contribution of the structure-from-motion analysis, but familiarity with the face is probably not a factor.
Haxby et al.'s (2000) “distributed” model of the neural substrates of face processing can accommodate these findings with minor modifications. In this model, the identification information available from static features is processed in the lateral fusiform gyrus/inferotemporal (IT) cortex, whereas socially-relevant facial movements (e.g., gaze, expression) are processed in superior temporal sulcus (STS). We hypothesize that dynamic face recognition is mediated by processes in both the dorsal and ventral visual streams. The dorsal stream processes social communication signals and person-specific “dynamic facial signatures”. Dorsal processing of structure-from-motion information may contribute subsequently to the quality of a static face representation in the ventrally-based IT (Sary et al, 1993).
Supported by a DOD/DARPA grant to A.O. & H.A.