The Impact of Distinctiveness on Face Recognition

Abstract

Face recognition depends on both the integration of facial information into a unified perceptual representation and the processing of individual facial features. Facial distinctiveness also contributes to recognition, as distinctive faces are generally more easily discriminated and remembered than typical faces. However, relatively little research has examined whether the distinctiveness of individual facial regions influences performance within a paradigm associated with holistic face processing. The present study addressed this question using a composite-face matching task. A total of 104 participants completed a computerised task involving aligned composite faces constructed from well-known celebrity identities. Participants judged whether the upper or lower halves of two simultaneously presented composite faces were the same or different while ignoring the other half. Task-relevant regions had been independently classified as distinctive or non-distinctive, whereas task-irrelevant regions were consistently non-distinctive. Participants completed separate Top and Bottom matching blocks, and performance was assessed using perceptual sensitivity, correct reaction time, and error rate. The results showed that participants discriminated distinctive target halves more sensitively than non-distinctive target halves, and this advantage was comparable for upper and lower facial regions. Contrary to the response-time prediction, correct judgements were slower for distinctive targets. Error rates did not vary systematically across the experimental conditions. Distinctiveness therefore enhanced discrimination sensitivity without producing a uniform improvement across all measures of performance. Together, these findings reveal a trade-off between discrimination sensitivity and response speed and demonstrate that the diagnostic value of local facial information remains behaviourally influential within an aligned whole-face context. More broadly, the results support a flexible and feature-sensitive account of face processing in which the contribution of local information varies according to its distinctiveness.