Dynamic facial expressions are processed holistically, but not more holistically than static facial expressions.

There is evidence that facial expressions are perceived holistically and featurally. The composite task is a direct measure of holistic processing (although the absence of a composite effect implies the use of other types of processing). Most composite task studies have used static images, despite the fact that movement is an important aspect of facial expressions and there is some evidence that movement may facilitate recognition. We created static and dynamic composites, in which emotions were reliably identified from each half of the face. The magnitude of the composite effect was similar for static and dynamic expressions identified from the top half (anger, sadness and surprise) but was reduced in dynamic as compared to static expressions identified from the bottom half (fear, disgust and joy). Thus, any advantage in recognising dynamic over static expressions is not likely to stem from enhanced holistic processing, rather motion may emphasise or disambiguate diagnostic featural information.

Dynamic composite faces are processed holistically.

Holistic processing is considered one of the hallmarks of face recognition. Recent studies using the composite task claim to show a lack of holistic processing for dynamic faces, however they only presented moving faces in the learning phase and tested with static composite images. So while previous research has addressed the question of whether moving faces influence the processing of subsequently viewed static faces, the question of whether moving faces are processed holistically remains unanswered. We address that question here. In our study participants learned faces in motion and were tested on moving composite faces, or learned static faces and were tested on static composite faces. We found a clear composite effect for both upright static and dynamic faces, with no significant difference in the magnitude of those effects. Further, there was no evidence of composite or motion effects in inverted conditions, ruling out low level or other motion signal properties as explanations of performance in upright faces. Together, these results show that upright moving faces are processed holistically, in a similar manner to static faces, extending decades of research with static faces and confirming the importance of holistic processing to familiar face recognition.