Are problems of pretend play in autism due to mentalising or executive difficulties?

Background: Individuals with autism show difficulties in pretence, but the cause of this deficit remains unclear. It may relate to a failure to grasp the metarepresentational nature of pretence, which in turn links to individuals' problems in theory of mind. Alternatively, executive impairments might make it difficult for individuals to inhibit the true state of affairs in pretence. Objectives: Experiment 1 varied the theory of mind requirements of pretend tasks, while Experiment 2 manipulated executive demands by comparing comprehension of object substitution and the attribution of absent properties in pretence; object substitution has a greater executive load because it involves inhibiting objects' true identity. Methods: Experiment 1 presented 37 typically developing children, and 34 children with autism, with two tests of comprehension of pretence. Both involved scenarios in which protagonists held different pretend views, however only one condition required the participant to represent their mental states (ToM condition). Experiment 2 compared 16 typically developing children, and 16 children with autism, on tests of pretence comprehension involving object substitution or attribution of absent properties. Results: In Experiment 1, individuals' performance on the ToM condition was dependent on their theory of mind status, but theory of mind status had no effect on performance in the non-ToM condition. Once theory of mind status was accounted for, the two groups did not differ across the two pretend conditions, p = .98. In Experiment 2 there was a trend for an interaction, p = .08, such that individuals with autism had greater problems with object substitution. Conclusions: Experiment 1 shows that many pretend tasks do not require theory of mind. In contrast, there is evidence that executive difficulties make certain aspects of pretence particularly difficult for individuals with autism.