[Theory of mind in Asperger syndrome]

Theory of mind deficit can be used to explain social and communication impairments that define the autism spectrum disorder. Theory of mind is the ability to attribute mental states to self and others in order to understand and predict behavior. It involves the distinction between the real world and mental representations of the world. Several studies established that high functioning autistic individuals and individuals with Asperger syndrome (ASP) tend to be as proficient as controls in understanding first order false belief tasks. In contrast, they still lag behind their typical peers in understanding second order false belief tasks or more advanced tasks of theory of mind (e.g., Baron-Cohen, 1993). Most of these studies focus on the adult population and it seems particularly interesting to investigate whether children with ASP would present the same pattern of strengths and deficits as adults. In our research, children with ASP were tested in an advanced task of theory of mind based on a visual presentation of comic strips and in a more traditional assessment of false belief understanding: the Smarties test.

Two experimental groups participated in this study: a group of 16 high functioning children and adolescents with autism or with Asperger syndrome (ASP) and a group of 16 typically developing children matched on gender and age (CONT). The task was designed to assess the ability of children with ASP to infer the mental state of others. Stimuli were 26 different comic strips depicting a short story. Each comic strip was composed of three pictures and was shown on the upper half of the screen. Then three pictures numbered 1 to 3 showing possible outcomes of the scenario were superimposed on the bottom half of the screen. Only one of these three pictures represented a plausible conclusion to the scenario. This experiment contained two conditions: A Character intention (CI) condition and a Physical causality (PC) condition. The comic strips in the CI condition involved one character whose intentions had to be inferred by the subject in order to choose the correct picture. Comic strips in PC condition only required to understand physical causalities. Subjects were asked to watch the comic strip attentively and then they were required to make a choice between the three story endings by pressing the corresponding keyboard button. Both answers and response times were recorded. Additionally, all participants were enrolled in the classical false belief (Smarties) task.

Comic strips: An ANOVA [2 groups (CONT/ASP) x 2 conditions (CI/PC)] was performed on the number of correct responses. Neither the Group nor the Condition factor was significant (p>0.05). In contrast, the interaction Group x Condition reached significance level [F(1-30)=4.3, p<0.05)]. Further analysis revealed that performance of the ASP group was significantly higher in the CP (M=10.8, SD=2.5) than in the CI (M=9.8, SD=2) condition [t (1-14)=2.9, p<0.001)], whereas there was no condition difference in the control group (p>0.05). False belief: all ASP participants succeeded in the task.

Our data clearly demonstrated that children and adolescents with Asperger syndrome or high functioning autism exhibited an impairment in understanding the intention of others. In the comic strip task, children with ASP have more difficulties in the character's intention condition than in the physical causality condition. This impairment is not imputable to a deficit in taking into account the context (Weack Central Coherence theory) since they performed as well as controls in the physical causality condition which also required the processing of the whole scene. In contrast, all children with ASP succeeded at the false belief task. These contradictory findings suggest that, although testing theory of mind, the two tasks do not tap similar mechanisms. It is possible that the use of verbal material in the false belief task improved performance of the ASP children who are known to present particular strengths in this domain. Another possible explanation that needs further testing would be that the level of complexity differs between the two tasks.