How Do Individuals with ASD Process and Copy Ambiguous Figures?

Background: The ability to switch between two interpretations of an ambiguous figure (e.g. rabbit/duck) is a well-documented perceptual phenomenon. Children with Autism Spectrum Disorder (ASD) can process these ‘reversals' when explicitly told about the ambiguity, however they tend not to do so spontaneously (Mitchell and Ackroyd, 2003; Sobel, Capps and Gopnik, 2003). One way to implicitly examine the understanding of multiple representations is to ask participants to copy the same ambiguous figures under different contextual conditions. If a figure is labelled a ‘duck', one might draw something ‘duck-like' even if the target stimulus is equally representative of a rabbit. Given the tendency for individuals with ASD to be unbiased by contextual information (‘weak central coherence'; Frith, 1989), an alternative possibility is that they may create accurate pictures which more closely represent the target, regardless of context. Objectives: The purpose of the present study was to 1) determine if the ability to experience a spontaneous reversal affects drawing accuracy; 2) explore if labeling affects drawing accuracy; and 3) replicate research suggesting that children with ASD are unlikely to make spontaneous reversals. Methods: Twelve adolescents with ASD (CA 15.2, MA 9.0) were matched to 12 adolescents with learning difficulties (CA 14.6, MA 8.6) on language ability assessed by the BPVS-II. In the Unlabelled condition, 2 ambiguous pictures were individually presented. For each trial, the participant was asked initially to copy the target picture, then probed for spontaneous or prompted reversal, and asked again to draw the target picture on a different sheet of paper. If reversal did not occur it was classed as ‘Refusal' and the alternative interpretation was demonstrated. The Labelled condition was similar except the target image was initially named before the individual made any drawings (e.g. can you draw this duck?). A control trial in each condition followed the same procedure but consisted of unambiguous line drawings, to ensure that responses to the test questions were not due to demand characteristics. Each participant received both conditions. Results: The pictures were scored by an independent, blind rater to determine similarity of the drawings of each target image before and after reversal. Drawings by individuals with ASD were more similar to each other than drawings by the control group in the Labelled condition (t =2.49, df =23; p =0.02), but not in the Unlabelled condition (t =0.41, df =23; p >0.05). Both groups showed mostly prompted reversals (44% ASD, 52% control), followed by spontaneous reversals (35% ASD, 33% control), then refusals (21% ASD, 15% control), a non-significant difference. Conclusions: These results suggest that adolescents with ASD are not influenced by contextual information (labels or experience of dual interpretations) when copying ambiguous drawings. That is, their pictures were more similar to each other and to the target picture. Consistent with prior research, individuals with ASD showed mostly prompted reversals, which did not differ from learning disabled peers. This research has implications for how individuals with ASD understand multiple representations and suggests that these individuals have a relative strength in accurate 2-D copying.