Abstract
When high and low luminance contrast images are presented separately to each eye in conditions of binocular rivalry, the percept of the stronger (high luminance contrast) image tends to dominate. We tested for parallels to this contrast dependence in faces where the “contrast” of the stimulus corresponded to the face's “identity strength.” Faces are thought to be represented in a multidimensional face-space that is likely to be neuronally instantiated in the ventral temporal cortex. According to this hypothesis an individual face is represented by how it deviates from an average face (the origin of the face-space), and face-contrast varies as the magnitude of the deviation of a given face from the prototype. We asked whether more distinctive faces would be more likely to dominate in rivalry. Methods: Distinctiveness was varied by distorting an original face by locally expanding or contracting the facial features relative to a midpoint on the nose. These distortions had the advantage of preserving many of the low-level image features while varying how atypical the faces appeared. Observers viewed the display through a stereoscope and were presented with the original face to one eye and a distorted (sibling) face to the other eye to create a condition of binocular rivalry. During continuous viewing they reported whether they perceived the average face, the distorted face or neither. Results: Observers perceived the distinctive faces for longer periods of time than the average faces. Conclusion: The results suggest that more distinctive faces may tend to have a higher effective contrast in rivalry, suggesting in turn that the rivalry may partly depend on the encoding of the configural characteristics defining the faces.