Drawing Ability in Autism: A Window into the Imagination

Article excerpt

Abstract: This study investigated imagination via drawing tasks, in 15 children with autism and 15 children with Asperger Syndrome, compared to verbal mental age matched normal children and children with moderate learning difficulties (MLD). Experiment 1 used the Draw an Impossible Man Task. While children with autism were impaired relative to the normal group, they were not impaired relative to the children with MLD. In order to probe for an imagination deficit, Experiment 2 employed a more challenging measure of imaginative drawing, a task involving mixing categories to produce drawings of real or unreal entities (e.g., drawing half-fish/half-mouse). This revealed an autism-specific deficit. Experiment 3 confirmed this was not due to difficulties in combining elements per se. Experiment 4 required subjects to transform a picture (e.g., a cloud into a swan) and again found an autism-specific deficit. Children with Asperger Syndrome were only impaired when required to make such transformations spontaneously.

Autism is characterized by a "triad" of features. A great deal of research has been conducted into two of these: the social and communication difficulties. However, the third symptom in the triad remains somewhat neglected: that of limited imaginative ability and restricted interests (1). The few experimental investigations into this have mostly been concerned with the deficit in spontaneous pretend play (e.g., 2-5). These suggest an imaginative deficit, but few studies have addressed imagination directly.

Our earlier study began to do this by assessing the ability of children with autism to imagine unreal objects (6). This study was based on Karmiloff-Smith's (7) paradigm, where normally developing children were asked to draw a house and a man, and then asked to draw a man that doesn't exist, and a house that doesn't exist. For brevity, we refer to this as the Draw An Impossible Man Task. Karmiloff-Smith found that when drawing things that didn't exist, older normal children (8-10 yrs) introduced changes that included adding elements from other conceptual categories. Even younger normal children (4-6 yrs) deleted elements or changed the size and shape of elements. In contrast, we found that most children with autism had marked difficulties producing drawings of unreal objects. In fact, their drawings of real and unreal objects were remarkably similar (6).

The studies reported here had four aims: First, to test if our earlier results (6) replicate. Secondly, to test if the deficit on the Draw an Impossible Man Task is a developmental phenomenon (for example, a function of mental age). Thirdly, to test how children with autism perform on related tasks requiring imagination. Finally, to test if results differ if the diagnosis is Asperger Syndrome or autism. In Experiment 1, we therefore tested if the Scott and Baron-Cohen (6) findings replicated in a sample of children with autism with a higher verbal mental age, and if they extended to children with Asperger Syndrome (AS). Subsequent experiments (2-4) were planned to investigate production of imaginative drawings by children with autism and AS.

In the experiments reported here, rather than attempting to measure how imaginative a response is, we focus on drawings of unreal entities (e.g., a walking candle) and unreal transformations (e.g., turning a cloud into a swan). The contents of such drawings are, we think, unambiguously imaginary.


Four groups of children took part in the study. Their details are summarized in Table 1. The first was a group of 15 children with autism, all of whom met the standard diagnostic criteria (1). The second was a group of 15 children with Asperger Syndrome (AS). AS was defined as meeting the criteria for autism but with no history of general cognitive or language delay. Children in both of these groups were diagnosed by independent clinicians and were attending special schools in Merseyside or Cambridgeshire. …