Abstract
Adults' expertise in face recognition has been related to their sensitivity to two cues to facial identity-the shape of individual features and the spacing among them. In previous studies we created 8 new versions of a single face: 4 differed from the original only in the spacing of features and 4 differed only in the shape of the eyes and mouth. Results for this feature set indicated little impairment by inversion (Mondloch et al., 2002), adult-like accuracy by 10-years of age (Mondloch et al., 2002) and sparing after a history of early visual deprivation from bilateral congenital cataract (Le Grand et al., 2001, 2003). However, these patterns might have resulted from the inadvertent choice of easily discriminated features, as suggested indirectly by adults' higher accuracy for upright faces in the feature set (M=.89) compared to the spacing set (M=.82). To address this issue we created 20 featural versions of a single face that were then paired to give 60 trials with different faces, intermixed among 60 trials with same faces. Adults' (n=24) accuracy in making same/different judgments was high across all pairings (M = .91) and did not differ as a function of orientation (p [[gt]] .20).
Furthermore, both 10-year-olds (n=24) and patients treated for bilateral congenital cataract (n=8) were as accurate as visually normal adults (all ps [[gt]] .20). This pattern held even for the six pairings for which adults' accuracy was less than .85. We conclude that adults are very sensitive to featural differences, that this sensitivity develops by 10 years of age and that sensitivity to features does not depend on either face orientation or visual input during the first few weeks of life.