Suppressed Attentional Response to Dyadic Social Cues in Infants with Autism

Background: In typical development, the unfolding of social and communicative skills hinges upon the ability to allocate and sustain attention towards people. Deficits in social attention have been documented in autism, though the underlying mechanisms are poorly understood.

Objectives: To examine effects of context on attention to people and faces in toddlers with ASD using a free-viewing eye-tracking paradigm.

 Methods: Participants were 13- to 25-month-old toddlers with autism (AUT; n=54), developmental delay (DD; n=22) and typical development (TD; n=48). To identify the factors responsible for limited attention to faces we manipulated the presence of salient social (child-directed speech (CDS) and eye contact (EC)) and nonsocial (distractor toys) cues.  Four conditions were tested: actress making a sandwich (Sandwich, no CDS or EC), actress attempting to engage viewer through CDS and EC (DyadicCue), actress initiating joint attention (JointAttention, CDS and EC limited), and actress looking at moving toys in background (MovingToys, no CDS or EC). A generalized deficit in social attention would result in limited attention to person and enhanced attention to toys across conditions; elementary sensitivity to the context in which people appear in their visual field would produce condition-specific deficits. The results were analyzed using linear mixed models.

 Results:   In conditions devoid of EC and CDS, the distribution of attention between key features of the scene in all groups was comparable (Sandwich, MovingToys). However, when explicit dyadic cues were introduced (DyadicCue), AUT group showed suppressed attention to the entire scene (p<.01) and, when they looked at the scene, spent less time looking at the speaker’s face (p<.01) and monitoring her mouth (p<.05) compared to the control groups. In the JointAttention, they tended to spend more time looking away from the entire scene (p<.01).. 

Conclusions:   The mere presence of a person within the visual field does not appear to disturb the general looking patterns in toddlers with autism.  Neither did the presence of toys and objects. It was only when child-directed speech and eye contact were introduced that differences between autism and control groups became pronounced.  Thus, as a group, toddlers with autism show abnormal visual responses to social scenes, which appear context-specific and do not reflect a generalized social attention deficit or prepotent preference for objects.  A combination of EC and CDS represent the prototypical bid for dyadic attention, the most elementary and perhaps most salient social behavior, to which a keen sensitivity is already present in newborns. An attenuated attentional bias for this class of social stimuli early in life is likely to have a profound and debilitating effect on the development of social-cognitive skills and language in autism. Considering marked inter-individual variability observed in response to Dyadic Cue condition in the autism sample, further investigation into potential presence of subtypes amongst toddlers with ASD is warranted.  Similarly, the relative contribution of dyadic cues and perceptual characteristics of potential distractors to atypical scanning patterns in autism needs to be further clarified.