Implicit Action Anticipation in Children with and without ASD and Varying Intellectual and Language Impairments: Testing the Application of a Non-Verbal Eye-Tracking Paradigm in a Heterogeneous Sample

Friday, May 12, 2017: 5:00 PM-6:30 PM
Golden Gate Ballroom (Marriott Marquis Hotel)
S. Anns1 and S. B. Gaigg2, (1)Autism Research Group, School of Psychology, City, University of London, London, United Kingdom, (2)Psychology, City, University of London, London, United Kingdom

Explicit theory of mind has mostly been measured using verbal false belief paradigms and findings have been complicated by both the verbal nature of the tests and also by the pragmatic and structural language ability of the populations in question. To counteract this, Senju et al., (2010) built on Onishi & Baillargeon’s (2005) study of implicit false belief attribution and found that children with ASD (aged 6-8 years; n = 12) had difficulties with action anticipation (AA) compared to their typically developing (TD) counterparts matched on age and fluid intelligence.


The primary aim of the current study was to replicate Senju et al’s (2010) experiment with a larger cohort of children with ASD; including both low (ALI) and high (ALN) verbal abilities as well as those with idiopathic (no known cause) intellectual disabilities (ID) in comparison to TD children. A secondary objective was to test any associations of AA ability with VIQ and conceptual semantic knowledge.


Four groups were recruited in a between measures design: ALI (n=11;12-18 years; VIQ < 75 on WASI verbal subtests); ALN (n=15; 6-11 years; VIQ > 90); ID (n=11; 12-18 years; VIQ < 75; and TD (n=18; 6-11 years; VIQ > 90). Fluid intelligence (Ravens Coloured Progressive Matrices) and conceptual semantic knowledge (The Pyramids and Palm Trees Test; PPT) were also assessed.

Eye tracking data was collected using the same methodology as Senju et al., (2010) and Southgate et al., (2007). A video was presented of an actor watching a ball being hidden in one of two boxes. The object was then displaced when the actor was looking away. Total looking time to the correct versus incorrect location was then coded which represented children’s anticipation of the actor’s behaviour. This was only possible if children had attributed a false belief. Success at the fourth familiarisation stage was necessary for inclusion. Surprisingly 27 out of a total 55 participants (49%) were excluded (ALI = 5/11; ALN = 9/15; ID = 4/11 & TD = 10/18).

Additional tests of false belief attribution included an explicit question from the experimenter at the end of the video and a ‘Hidden Contents’ standard false belief task (SFB).


Due to the unexpectedly high participant exclusion rate analyses were conducted on the dataset before and after exclusion to explore the possibility that children may have taken longer than the 4 trials to familiarise. This was not the case. In both analyses there were no significant differences between any of the groups on implicit action anticipation (IAA), nor on the explicit AA question. This was also the case when combined ASD (ALI +ALN) versus Non-ASD (ID + TD) groups were taken into account. In addition there were no significant associations found between AA and other theory of mind measures (SFB and explicit question), as well as VIQ and PPT.


These findings call into question the validity and reliability of this experimental paradigm and invite further discussion as to how this may be remedied. Several explanations are offered.