People with congenital prosopagnosia (CP) have difficulties recognizing familiar individuals, for example, acquaintances or celebrities, by their face alone, which has been assumed to be due to disruptions of holistic face processing. Showing a face stimulus in an upright position enables holistic face processing, whereas showing it inverted makes observers rely on individual features. One would assume that upright faces are thus more salient for observers with intact holistic face processing than inverted faces and that this difference is much smaller in CP.
Binocular rivalry (BR) can be used to measure the saliency of visual stimuli since salient stimuli tend to be more perceptually dominant. Thus, we used BR to implicitly measure the relative saliency of upright vs. inverted faces for observers with and without CP (N per group = 21).
In Exp. 1, we further tested whether the processing of facial emotions was intact in CP by using faces displaying different emotions (fearful, happy, neutral), which was the case. In Exp. 2, we tested whether introducing the Thatcher effect (inverted eyes and mouth are strongly noticed in an upright but not in an inverted face by people with intact face processing) could further differentiate between observers with and without CP. This latter hypothesis was not confirmed, but we could show in both experiments that faces were more dominant and the face inversion effect more pronounced in the control group than in the CP group.
BR can thus serve as a sensitive tool to detect processing differences between groups.