Background: People with autism suffer from impaired executive functions.  There are various executive functions (e.g., mental flexibility and inhibitory control) but all are not equally affected in autism, possibly because these different functions are implemented in different brain circuits.  More research is necessary to characterize which executive functions are affected most, which are intact, and, possibly, which are superior.

Objectives: We aim to refine our understanding of executive impairments in children with autism.  We used the task-switching paradigm, which offers a good assessment of executive functions, in particular of the skill to switch attention between tasks, and the skill to inhibit irrelevant information.

Methods: Two versions of a task-switching paradigm were used.  The first version administered was a standard cued task-switching paradigm. The second version used was an easier version which does not rely strongly on working memory.

Results: The ASD group performed poorly in the standard task-switching paradigm, with several children working at chance level.  However, in the low-memory load version of the task, they performed more accurately than typically developed adults (TDA), and relative switch costs were smaller (marginally significant: p=.051).  Further, in the low-memory version, inhibition of irrelevant stimulus information was similar in both groups.

Conclusions: Children with ASD have intact skills of rapidly switching task between rules and inhibiting irrelevant stimulus features, as long as this switching does not require the use of working memory.  As soon as it does, the children fail.  This indicates that problems with working memory may help to explain, at least shifting difficulty in autism, if not other various executive deficits.  At the same time, the finding of better performance than TD adults in the low-memory condition suggests that the task-switching paradigm is useful for studying the savant skills observed in about 10% of children with autism.