An Eye-Scanning Approach to Measuring Receptive Language Abilities In Infants at High and Low Risk for Autism

Background: Language difficulties are a primary feature of autism.  Consequently, there is interest in understanding the developmental origins of these difficulties.  In order to study the development of language difficulties in children with autism, several studies have focused on infants who are at either high risk (HR) or low risk (LR) for autism based on whether or not they have an older sibling already diagnosed with autism.  These studies suggest that by 12 months of age, children who ultimately develop autism demonstrate language delays.  Results with HR infants who do not develop autism are mixed but suggest they may also have language delays (see review, Rogers, 2010). Receptive language ability in infants is typically measured by use of assessments such as the Mullen and parental vocabulary checklists like the McArthur Bates Communicative Developmental Inventory (MB-CDI). These measures rely on parental judgment which may limit their accuracy. With the advent of eye-tracking technologies, it may be possible to objectively assess receptive language abilities.  Thus, the current study was designed to use eye-tracking as a novel methodology to objectively measure receptive language and categorization abilities in high and low risk 16- month old infants.

Objectives: To use eye-tracking technology to assess the performance of low and high risk infants in their language and categorization abilities.

Methods: Sixteen month old infants at high and low risk for ASD were shown stimuli that had realistic photographs of 8 different objects.  Four of the 8 objects belonged to a single super-ordinate category (e.g., animals) and the other 4 objects belonged to a second category (e.g., food items).  The objects chosen were commonly used words found in tests such as the PPVT, the CDI, and the Mullen. While the infants were presented with the 8-object stimulus, they heard one of the objects (the target word) verbally labeled several times in short sentences such as “see the (target)?” Across all stimuli, infants heard twelve target words while their eye movements were recorded by a Tobii x120 eye tracker. Percentage of looking towards the target word (e.g., dog) and the target category (e.g., animals) was calculated as the dependent measures.

Results:  LR infants looked at the target category (t = 3.19, p <0.05) and target word (t = 3.20, p <0.05) significantly above chance. This indicates that they recognized the target word and its category.  In contrast, the HR infants did not perform above chance for either the target category or the target word.

Conclusions: At 16 months, low risk infants were able to not only recognize the target word, but they were also able to recognize the target category. This ability was not demonstrated by HR infants. This finding suggests that not only are there receptive language delays in HR infants, but that these delays may be related to an inability to categorize objects.  Results also suggest that eye-tracking technology may provide a more objective way of assessing receptive language compared to parental measures as used in the Mullen or the MB-CDI.