Social Engagement Does Not Modulate Object Processing in Young Children with Autism Spectrum Disorder (ASD): An Electrophysiological Investigation

Background:  Many studies have demonstrated that typically developing infants and children learn more successfully when information is presented in a social context, with learning defined as the differential response to stimuli based on their social presentation. This learning can be quantified through a neural response to stimuli using event related electrophysiology (EEG). Because impairment in social interaction represents a core deficit in children with autism spectrum disorders (ASD), we were interested in understanding the effect of social interaction on information processing of this population. 

Objectives:  We used a novel EEG paradigm that couples a live interaction exposure, in which children are exposed to toys in both a social and non-social context, with an EEG test phase in which these objects presented socially or non-socially are then presented passively. Our aims were to understand the effect of social interaction on information processing in children with ASD by examining the association of the child’s social behavior toward the examiner with electrophysiological markers of learning. We hypothesized that children with ASD would demonstrate less engagement during the interactive exposure phase and, therefore, would not demonstrate a differential response to objects based on their social saliency.

Methods:  In the exposure phase, children (2-6 years old) were presented with toys by an adult examiner in either a social or non-social manner. Child and examiner behaviors were coded per trial using both likert scales and quantifiable measures of sociability and level of interaction. In the test phase, children viewed discrete images of the toys while high-density EEG was recorded. EEG was recorded using a 128-electrode Hydrocel Geodesic Sensor Net System (EGI Inc) and adequate data was gathered on 14 children in each group. Analysis was focused on the differential neural response to the social and non-social conditions, with variables of interest including the frontal N1, Pb and Nc.

Results:   TD children were overall more social than the ASD group. The ASD group, however, did show evidence of having detected the difference in the examiner’s behavior between the two conditions, as they were significantly less engaged in the non-social condition when the examiner withdrew interaction with the child.  In the EEG analysis, TD children demonstrated differentiation of conditions based on their left Nc amplitude, with a larger (more negative) response to the non-social condition, suggesting that greater attention resources were required to process objects lacking social salience. In contrast, children with ASD did not differentiate the conditions. Finally, in the TD group, attempts to initiate eye contact correlated with N1 amplitude difference by condition, suggesting that interaction with a responsive social partner facilitates learning in typical development. 

Conclusions:  These results suggest that social cues do not seem to facilitate object processing in young children with ASD, and they illustrate a disconnect between overt behavior and information processing in this population. This type of investigation will be critical to determine if learning deficits not only predict response to specific interventions but also whether they can be modified through treatments targeting joint attention and social engagement.