The perception of dynamic human faces

Abstract

Faces are at once simple and complex objects. Every face contains the same basic set of features organized in the same basic pattern or configuration. Individual identity, however, is conveyed by subtle variations of these features and configurations. Our ability to identify individuals based on their facial characteristics, a skill we often take for granted, continues to puzzle and impress scientists who are interested in understanding the human mind and/or replicating its functions in a machine. Our ability to process individual identity is even more impressive when we consider that faces present a moving target. That is, when we smile, speak or gesture approval, our faces change shape. The current research examines how this motion affects the processing and representation of facial identity. In one series of experiments, an explicit matching paradigm was used to compare performance in matching two successively presented faces as a function of whether the first face (prime) was a single still image or a short video clip of a smile or a frown. Performance was significantly enhanced when a moving prime face had been seen and an identity decision involved generalizing across a change of expression or orientation. This finding is consistent with work on object recognition which suggests that motion may aid in the construction of view-invariant representations. In a second series of experiments, a representational momentum paradigm was used. Previous work with representational momentum has shown that the final position or configuration of a transforming object is misremembered as being further forward in the direction of motion or change. When we see a smiling or frowning face, will we misremember the stopping point of that expression as being more intense than it really was, consistent with previous representational momentum findings using nonface stimuli? Results indicate no evidence for such forward memory bias. Instead, a consistent pattern emerged in which memory was shifted back towards the starting point of the expression. Results from both sets of experiments indicate that motion can influence the processing and representation of faces. It is hoped that such findings will help direct attention to facial dynamics as an important area for future research.