Facial Emotions Elicit Atypical Arousal and Visual Attention Patterns in 14-Month-Old Infants at High Risk for Autism
Many individuals with autism have difficulty understanding emotions in themselves and others (Nuske et al., 2013). Further, many individuals with autism experience clinically significant levels of anxiety (White et al., 2009). However, little is known about the developmental roots of these difficulties. In early development, combining information about arousal state with social signals from caregivers is central to learning to regulate and understand emotions (e.g. Thompson, 2011). In autism, early atypicalities in face processing or in arousal responses to emotional situations could thus compromise this critical developmental process. Here, we test this model by examining visual attention and arousal responses to emotion faces in a group of infants with older siblings with autism. In the chosen paradigm, infants viewed faces displaying emotional expressions (happy, fearful, sad, angry) with direct or averted gaze. This paradigm was chosen because typically developing infants show robust preferences for happy faces with direct gaze, and fearful faces with averted gaze (Rigato et al., 2013). Atypicalities may indicate a diminished ability to interpret the communicative intent of facial expressions.
To examine whether infants at high risk for autism display atypical gaze and arousal responses to facial emotion, and whether atypicalities are linked to high levels of temperamental fear.
Participants were infants with older siblings with autism (n=56) or typical development (n=25) tested at 8- and 14-months. Gaze was recorded with TOBII eye-tracking technology and processed using Matlab. Key dependent variables include first look direction and pupil dilation responses during each slide.
There were no group differences at 8 months. At 14 months, low-risk infants showed the expected pattern of modulation of emotion processing by gaze direction. High-risk infants showed typical (though attenuated) preferences for happy faces with direct gaze; this indicates that they were able to discriminate direct and averted gaze. However, whilst low-risk infants showed very strong preferences for fear faces with averted gaze, high-risk infants showed no significant preferences in this condition. Analysis of pupil dilation indicated that the low-risk group showed greater arousal to faces displaying approach emotions (happy, angry) whilst the high-risk group showed greater arousal to faces displaying avoidance emotions (fear, sad). Finally, the high-risk group showed significantly greater levels of temperamental fear. Within the high-risk group, showing high levels of temperamental fear was associated with an atypical combination of arousal and visual attention responses.
Taken together, these data suggests that infants at high risk for autism show atypical visual attention to fearful faces; show atypically high levels of arousal in response to fearful and sad faces; and show particularly high levels of temperamental fear that were associated with an atypical combination of visual attention and arousal responses. This is consistent with the proposal that atypical cognitive and arousal responses to emotion faces are present in infants at high risk for autism, and may be related to early behavioral manifestations of increased anxiety. Follow-up of this cohort is ongoing; further analyses will examine the relation between these early atypicalities and symptoms of autism at 24 months.