Introduction
Face-to-face social interactions require rapid and accurate interpretation of both facial expressions and individual identity. Bruce and Young’s (1986) classic model of face perception suggests independent and parallel processing of these two cues, a proposal supported by studies in cognitive psychology (Calder et al., 2000, Campbell et al., 1996 and Young et al., 1986). Campbell and colleagues (1996) showed that judgments of lip-read speech, expression, and identity were not subject to interference by judgment-irrelevant factors (e.g., judgments of expression were not affected by changes in identity). Calder and colleagues (2000) presented three types of composite face stimuli produced by aligning the top half of one face with the bottom half of another face: same identity/different expression, different identity/same expression, and different identity/different expression. Adults were slower to identify the expression in the bottom half when the bottom half was aligned with a top half displaying a different emotional expression, but not when the bottom half was aligned with a top half of a different model displaying the same emotional expression. Similarly, adults’ reaction time when naming the identity of the bottom half of a face was impaired when the top half had a different identity, but not when the top half had the same identity but was displaying a different facial expression. Calder and colleagues concluded that holistic processing underlies recognition of both identity and facial expressions but that different information may be relevant for the two types of processing.
Further evidence for Bruce and Young’s (1986) model comes from cognitive neuropsychology. Prosopagnosics display impaired recognition of facial identity but intact recognition of facial expression, gender, and age (Tranel, Damasio, & Damasio, 1988), and following brain injury some individuals display impaired expression recognition but intact identity recognition, whereas others show the reverse pattern (Young, Newcombe, de Haan, Small, & Hay, 1993). Functional imaging studies have localized processing of facial identity to the lateral fusiform gyrus and processing of facial expression (and other changeable facial characteristics) to the superior temporal sulcus (see Haxby, Hoffman, & Gobbini, 2000, for a review). Based on this evidence, Haxby and colleagues (2000) proposed a distributed neural system for face perception where recognition of changeable and nonchangeable facial characteristics involves separable but overlapping neural structures.
Although processing of facial identity and expression appears to be dissociable, some integration must occur. Recognition of expression and identity may involve partially integrated representations, but the degree of functional integration may depend on information processing demands (Calder & Young, 2005). Indeed, the ability to integrate identity and expression cues allows individuals to recognize the same person in different affective states.
The purpose of our study was to investigate the integration of identity and expression cues in children. Although numerous studies have investigated the development of expert face recognition (Freire and Lee, 2001, Gilchrist and McKone, 2003, Mondloch et al., 2004, Mondloch et al., 2003, Mondloch et al., 2002 and Pellicano et al., 2006) or expression recognition (Camras and Allison, 1985, Kolb et al., 1992, Markham and Adams, 1992, Markham and Wang, 1996 and Vicari et al., 2000), no research to date has assessed the integration of expression and identity in children. To do so, we used a relatively new technique that is useful for probing representations underlying various perceptual abilities: face adaptation.
