Children with ASD can use gaze to map new words

The conclusion that children with autism spectrum disorders (ASD) do not use eye gaze in the service of word learning is based on one-trial studies.

To determine whether children with ASD come to use gaze in the service of word learning when given multiple trials with highly reliable eye-gaze cues.

Fifteen children with ASD with a mean age of 59 months (range = 36-92 months) and 15 typically developing (TD) peers with a mean age of 37 months (range = 16-92 months), and matched to the ASD group on receptive vocabulary raw scores, participated in four conditions formed by crossing-gaze load (high versus low) and attention load (high versus low). The high eye-gaze load condition required the children to shift attention to the examiner and follow her gaze to fast map new words correctly. The low-gaze load did not require shift and follow behaviours. The high-attention condition involved three (as opposed to one) distracter objects.

As compared with the TD group, a lower proportion of the ASD group shifted and followed the examiner on the initial trial of the high-gaze load condition, but there was not a significant difference between groups when shift and follow behaviours were averaged over subsequent trials, nor was there a difference between groups in fast-mapping performance. Fast-mapping outcomes were correlated with gaze shift and follow behaviours in the high-gaze load condition.

The finding that the children with ASD altered their looking behaviour over the course of the experiment suggests that children with ASD were sensitive to statistical regularities present in the examiner's gaze cues and used this information to alter their looking behaviour over the course of the experiment.