Atypical Value Driven Attention in Young Children with ASD

Background: Poor attention to social targets such faces constitute one of the defining features of ASD in young children and has been documented in the laboratory and real-world contexts. Given the centrality of information conveyed through faces and facial gestures, poor attentional attunement to this class of stimuli during prodromal and early syndromal stages of the disorder, contributes to the variability in social, adaptive, and cognitive outcomes observed in preschoolers with ASD. We hypothesize this phenomenon is linked with disruption in the neural network involved in the signaling of value-based attentional priorities.

Objectives: To examine value learning for face and fractal stimuli and its effect on selective attention in preschoolers with ASD, and in developmentally delayed (DD) and typically developing (TD) controls. We hypothesized that unlike TD and DD controls, children with ASD will show selective impairment in value learning in Face but not in Fractal condition.

Methods: Children with ASD (n=55), DD (n=36) and TD (n =38) groups underwent value learning (VL) training in social (faces) and nonsocial (fractals) conditions implemented on a gaze-contingent eye-tracking platform. The VL task consists of the Baseline, Training, and Choice Test phases and represents an extension of our prior VL work by our team (Wang et al., 2018). The task measures whether reinforcement for attending to a given stimulus during Training biases visual attention toward the stimulus during subsequent Choice Test. The primary outcome variable was the proportion of looking time at the reinforced (high-value) stimulus during Choice Test.

Results: There were no pre-existing biases for specific stimuli during Baseline in either group and all groups completed a similar number of Training trials. During Choice Test, children with ASD showed evidence for learning in the Fractal (p<.001) but not in the Face (p=.809) condition. In contrast, the DD and TD groups, showed evidence for learning in the Face (p=.001, p=.002) but not in the Fractal (p=.703, p=.906) condition. Performance on the task was correlated with severity of autism symptoms.

Conclusions: Consistent with our hypothesis, children with ASD showed limited value learning of faces; however their leaning of fractal values was enhanced compared to the mental and chronological age-matched groups. Results suggest that atypical value learning in ASD facilitates attentional selection of nonsocial stimuli and hampers selective attention of faces. This pattern contrasts with that observed in controls matched for chronological age and developmental level. If present early in life, atypical value learning may affect selection for processing of social stimuli that are most relevant to adaptation and thus, may play a formative role in emergence of autism symptoms. The study replicates and extends our prior work and highlights the role of value-driven attentional network in atypical attention to faces in ASD.