Drawing Corners: Using a Drawing Reproduction Task to Test Theories of Local-Global Processing in Children with Autism

Background: Much research attention has centred on strengths and weaknesses in local-global processing in individuals with autism. On some accounts, individuals with autism excel at local processing at the cost of global processing. On other accounts, autistic perception is characterised by enhanced local processing without any impairment in integrative processing. Although drawing is usually examined in relation to autistic savant abilities, it can also greatly inform our understanding of autistic perception. Studies of drawing reproduction in autism, however, have yielded mixed results and, to our knowledge, have not investigated local and global processing in relation to perceptuomotor coding of spatial relations.

Objectives: The current study investigated local and global processing in children with and without autism by examining their perceptuomotor representations of object relations in a drawing task. Specifically, we examined children’s between-object coding skills (local processing) and their within-object coding skills (global processing) by assessing drawing accuracy as they copied simple hierarchical geometric figures. Perceptual grouping cues were manipulated by altering elements of the figures such that some figures were more likely to group into a coherent global form than others.

Methods: Participants were 24 children with autism and 24 typically developing children matched on age, gender and ability. They completed a copying task comprising 24 trials presented in a randomised order. On each trial, participants copied a set of four corners arranged to form a 6cm x 6cm square. In order to examine the relationship between perceptual grouping cues and drawing accuracy we manipulated the size (0.5cm, 1.5cm or 2.5cm), orientation (rotated clockwise or anticlockwise by 10°), and number of rotations (2 or 4) of the corners in each trial. We coded each participant’s drawings following procedures outlined in a previous experiment with constructional apraxia patients.

Results: There were two indices of global processing: (1) a measure of global shape accuracy (based on the internal angles of the copy produced by participants) and (2) the extent of deviation from a 1:1 height/width ratio (i.e., to form a square as opposed to a rectangle). Children with autism made more errors in their reproductions of the global shape than typical children. For children with autism, we also found that changes in orientation of the corners disproportionately affected the global reproduction of figures. There were no significant group differences in the reproduction of the individual (local) elements.

Conclusions: The current results suggest that children with autism show difficulties reproducing the spatial relations between the elements when the individual corners do not form a strong perceptual whole – consistent with accounts that postulate global processing difficulties in autism. This may suggest that children with autism rely more on grouping cues than what has been previously reported in the literature.