18151
Eye-Tracking Measurements of Language Processing: Developmental Differences for Infants at High Risk for Autism
Children with ASD have impaired online processing of language, a predictor for later language development (Venker, Eernisse, Saffran & Weismer, 2013). Since in some aspects of their development, children at high risk for autism (who have a sibling with ASD) are more similar to children with ASD, and in others are more similar to typically developing children, this study investigates which of the two is true of online language processing.
Objectives:
Three components of online processing of language were explored: comprehension (measured through accuracy and reaction time), sustained attention (to the language input as well as to the language referent), and visual reception (of the referent associated with the spoken language).
Methods:
A paradigm similar to that of Fernald and colleagues (2006) was used to investigate online speech processing at three different ages: 18, 24 and 36 months. 125 children participated in the study, 57 children at high risk for autism and 68 children at low risk. 16 children from the high-risk group were diagnosed with autism at 36 months, and their data was analyzed separately.
In each session, a child saw 20 image pairs, 10 depicting early acquired nouns and 10 depicting later nouns. A blank screen was shown for one second, followed by a pair of pictures that the child was allowed to explore freely for three seconds, after which a recorded voice instructed the child to “Look at the (target noun)!”. The pictures remained on the screen for two additional seconds. For higher time resolution, and in order to circumvent coder reliability issues, we used not a camera but an automated infrared eye-tracker, the TOBII T60, to acquire the gaze data.
Results:
After noun onset, the percentage of trials in which the children were looking at the target image as opposed to the distractor image was calculated as a function of time.
The results show that children at high risk for autism demonstrate significantly lower accuracy (percentage looking at target image) than the low risk children for later developing words at 36 months (p=0.02) but not for early developing words. Also, at 36 months, low risk children show a significantly better sustained attention to the target image than children at high risk for autism (p<0.001). Performance on the eye-tracking task is significantly correlated with the visual reception skills as measured by the Mullen test at 36 months for the high-risk children (p=0.005), but not for the low-risk children. The results for children with ASD show a similar pattern to those of children at high risk.
Conclusions:
Performance on this eye-tracking language task captures subtle differences between children at high risk for autism and children at low risk. These differences, which seem to emerge between 24 and 36 months, might have broader and longer-term implications for their development, for example delayed acquisition of spoken language. These differences might also be reflective of social-communicative impairments and the link between performance on this task and visual reception for high risk children might point towards a compensatory mechanism.