Item

Abstract

The question of whether non-rigid motion of the face has any impact on the processing
of identity has so far led to contradictory results. Particularly with unfamiliar faces, it is
not yet clear, whether non-rigid motion facilitates recognition. Most studies to-date
have adopted old/new recognition tasks to investigate the impact of non-rigid motion,
with some finding an advantage and others not. These differential outcomes raise
concerns about the applicability of such a method. That is, an explicit recognition
paradigm, such as an old/new recognition task, may encourage compressed or abstract
coding and may not be feasible to explore the dynamic aspects of face recognition.
Here, we tried a different approach. Subjects were familiarized with two faces, one
moving, the other one a static picture, and later had to recognize them in a visual
search task. During the search task all faces were presented as static. We found a clear
reaction time advantage for faces learned as a dynamic sequence. The search slopes
for both faces were the same, indicating that the search itself was equally efficient,
whereas the process of identification was more robust in the dynamic case. This
advantage seems to be due to the motion itself rather than any additional information
provided by the moving sequence. These results not only show that non-rigid motion
plays a role in identity processing and facilitates later recognition. They also further
question the suitability of old/new recognition paradigms to study face recognition.