Previous investigations of gaze processing in autism have demonstrated a pattern of intact and impaired performance. Although individuals with autism are capable of discriminating another's gaze, they fail to interpret gaze direction, especially within the context of sociocommunicative (i.e., mentalistic) interactions. Extending this general line of inquiry, we explored whether typical children and children with autism spectrum disorder (ASD) were influenced by gaze direction in a task that demanded a core person-related judgment-namely, sex categorization. The results revealed that typically developing school-aged children were faster to classify faces by sex when targets displayed direct rather than averted gaze, or when the eyes were closed. This was not the case, however, for children with ASD, whose responses were unaffected by gaze direction. These findings suggest that difficulties in gaze processing in autism extend beyond sociocommunicative inferences to include basic person-perception judgments.
Detecting and interpreting another's direction of gaze is pivotal to interpersonal interactions. Mutual eye gaze, in particular, serves to signal the likely flavor of a social interaction, be it nasty or nice, to initiate and regulate social communication, and to provide important information about the mental states of others (Nummenmaa & Calder, 2009). It is perhaps of no surprise, then, that adults are exquisitely sensitive to gaze direction (von Grünau & Anston, 1995), that such sensitivity emerges early in ontogenesis (Farroni, Mansfield, Lai, & Johnson, 2003), and that atypicalities in gaze processing are a defining feature of autism (American Psychiatric Association [APA], 2000), a developmental condition characterized by profound difficulties in social reciprocity and communication.
In typical development, infants appear to be born with a preparedness to respond to gaze cues. Newborns prefer looking at faces showing direct gazes over those showing averted gazes (Farroni, Csibra, Simion, & Johnson, 2002) or those with closed eyes (Bakti, Baron-Cohen, Wheelwright, Connellan, & Ahluwalia, 2000). Threemonth- olds smile less when an adult looks away and resume smiling when the adult reinitiates eye contact (Hains & Muir, 1996), and 4-month-olds rapidly shift their own gaze toward a target only when such shifts are preceded by a period of mutual eye gaze (Farroni et al., 2003). Recent work has revealed that this early sensitivity to eyegaze direction, particularly mutual gaze, holds special significance during infant-adult interactions. Senju and Csibra (2008) examined the gaze following of 6-monthold infants as they observed a person looking toward a toy located on either the left or the right. Infants followed the adult's eye movement only when it was preceded by direct eye contact, suggesting an early awareness of the ostensive nature of gaze cues. In another study, 9-month-olds observed a face either always looking toward an object or always looking away from it. Infants preferred to look at an adult making object-directed shifts in eye gaze but, again, only when such shifts were preceded by direct gaze (Senju, Csibra, & Johnson, 2008). Both findings indicate an apparent "eye-contact effect" in which perceived eye contact influences subsequent processing of another's sociocommunicative intentions.
Direct eye gaze clearly affords advantages for children's developing mentalizing skills (e.g., Baron-Cohen, 1995). Yet research with adults and children suggests that mutual gaze also affords additional benefits. Notably, it shapes more basic aspects of the person-perception process. Macrae, Hood, Milne, Rowe, and Mason (2002) demonstrated that gaze direction moderates the efficiency of person construal. Specifically, adults are faster to judge the gender of a face when the target's eyes are directed straight ahead than when they are averted or closed. Also, they are more accurate in recognizing previously seen faces that display direct gaze (Mason, Hood, & Macrae, 2004), a memorial effect that also has been reported in 4-month-olds (Farroni, Massaccesi, Menon, & Johnson, 2007) and children (Hood, Macrae, Cole-Davies, & Dias, 2003). …