Privileged Role of Symbolic Number Sense in Mediating Math Abilities in Children with Autism

Background:  

Despite reports of enhanced mathematical ability in individuals with autism, little is known about basic number processing abilities in children with the disorder. In typically developing (TD) children, both non-symbolic (the ability to compare quantities of dots) and symbolic (the ability to compare Arabic numerals) number sense have been linked to individual differences in math ability. Based on known strengths in individuals with autism, including local processing bias and hyperlexia, both non-symbolic and symbolic processing could be important components of number sense that contribute to math skills in this group. Understanding basic quantity processing abilities and their relationship to math ability has the potential to inform the design of educational interventions and employment opportunities for individuals with the disorder.

Objectives:  

We first investigated whether non-symbolic and symbolic number sense are spared or enhanced in children with ASD. We then investigated whether children with ASD show the same pattern of relations between number sense and math ability as their TD peers.

Methods:  

We examined number sense in 36 high functioning children with ASD and 61 matched controls (aged 7-12). In separate tasks, participants identified the larger value in pairs of either Arabic numerals or arrays of dots. Number sense acuity was measured using the Weber fraction (w), which captures the smallest ratio that an individual can reliably discriminate. We examined the relationship between w and math ability, as measured by participants’ math composite score on the Wechsler Individual Achievement Test-II. Finally, we used a mediation model to assess the extent to which symbolic number sense mediates the relationship between non-symbolic number sense and math ability.

Results:  

Children with ASD had intact symbolic estimation skills but showed significant impairments in non-symbolic (dot) estimation relative to TD children. For both groups, symbolic acuity was a stronger predictor of math performance than non-symbolic acuity, but this effect was more pronounced in children with ASD. Finally, we found that symbolic comparison ability mediated the relationship between non-symbolic comparison and math performance in children with ASD, but not in TD children.

Conclusions:  

Contrary to previous suggestions of superior non-symbolic number sense in ASD, our results indicate that children with the disorder are impaired in non-symbolic estimation, and relative to their TD peers, this capacity plays a weaker role in the development of math skills in children with ASD. Intact symbolic number sense and its strong correlation with math achievement scores together suggest a privileged role for symbolic processing in the acquisition of mathematics proficiency in ASD. Building on symbolic strengths in children with ASD may facilitate acquisition of mathematical skills in a way that non-symbolic quantity information may not. Our findings further suggest that symbolic notation may be a useful tool for organizing imprecise, continuous information in children with ASD.