Semantic Priming in Children with High-Functioning Autism: An Eye-Tracking Study

Background:  Children with high functioning Autism (HFA) have problems processing language (e.g., Goldstein, Minshew, & Siegel, 1994).  Some of their problems may arise from abnormalities in semantic categorisation (e.g., Dunn & Bates, 2005; Gaigg, Gardiner, & Bowler, 2008; Kamio, Robins, Kelley, Swainson, & Fein, 2007). Priming aids language processing; it relies on implicit memory and results in increased sensitivity to stimuli due to prior experience. Priming effects have been previously studied in children with typical development (TD) using eye-tracking technology (e.g., Heuttig & Altmann, 2005); a prime may facilitate a faster response to a target item.  Children with autism have also been found to benefit from priming, with support found for ‘global processing’ (e.g., Ozonoff, Strayer, McMahon, & Filloux, 1994). However, eye tracking research with young children with HFA using semantic categories for priming has not previously been reported. In order to better understand language processing problems in autism, an on-line method (eye tracking) is advantageous in identifying to what extent young children with autism respond similarly to children with TD as they listen to and integrate verbal input.

Objectives:  The main objective of the study was to use eye tracking to investigate if priming of semantic categories equally benefited 5-7 year old children with TD and children with HFA.

Methods:  The semantic task comprised two conditions, priming and non-priming. Participants heard a total of 24 items (16 test items: eight primed and eight non-primed) and eight fillers whilst they viewed pictures displayed on the eye-tracker. For each test sentence, one of four pictures was the target; the target was named in the test sentence.  Four versions of the stimuli were prepared in order to counterbalance which items were primed or not primed and the order of presentation.  Children’s eye movements were recorded by the Tobii V2.2.8. Eye-gaze was coded from the critical points, including (1) the onset of the prime (e.g., ‘fruit’) to determine if the prime influenced which items children looked at, and; (2) the onset of the target noun (e.g., ‘apple’) to determine if the prime facilitated a faster response to the target in comparison to targets presented in the non-priming condition. Proportion of looking time to each of the four displayed items was calculated for 2 second time periods (per 100ms time frames) following the critical point. Children’s language and attention were assessed using standardised tests.

Results:  Preliminary results for 20 of the children tested (10 TD and 10 HFA) showed large variability but a significant main effect of priming, as predicted, with a greater proportion of looking time to the target items in the priming condition. The paper will present the results by group for the total sample and in relation to the attention and language abilities of the participants.

Conclusions:  Preliminary findings indicate that semantically related words are processed and accessed faster following a prime than a neutral word. We discuss the group differences in relation to the Weak Central Coherence Theory.