Investigating the shape bias in typically developing children and children with autism spectrum disorders

Young typically developing (TD) children have been observed to utilize word learning strategies such as the noun bias and shape bias; these improve their efficiency in acquiring and categorizing novel terms. Children using the shape bias extend object labels to new objects of the same shape; thus, the shape bias prompts the categorization of object words based on the global characteristic of shape over local, discrete details. Individuals with autism spectrum disorders (ASDs) frequently attend to minor details of objects rather than their global structure. Therefore, children with ASD may not use shape bias to acquire new words. Previous research with children with ASD has provided evidence that they parallel TD children in showing a noun bias, but not a shape bias (Tek et al., 2008). However, this sample was small and individual and item differences were not investigated in depth. In an extension of Tek et al. (2008) with twice the sample size and a wider developmental timespan, we tested 32 children with ASD and 35 TD children in a longitudinal study across 20 months using the intermodal preferential looking paradigm. Children saw five triads of novel objects (target, shape-match, color-match) in both NoName and Name trials; those who looked longer at the shape-match during the Name trials than the NoName trials demonstrated a shape bias. The TD group showed a significant shape bias at all visits, beginning at 20 months of age while the language-matched ASD group did not show a significant shape bias at any visit. Within the ASD group, though, some children did show a shape bias; these children had larger vocabularies concurrently and longitudinally. Degree of shape bias elicitation varied by item, but did not seem related to perceptual complexity. We conclude that shape does not appear to be an organizing factor for word learning by children with ASD.