Previous studies have found a subgroup of people with autism or Asperger Syndrome who pass second-order tests of theory of mind. However, such tests have a ceiling in developmental terms corresponding to a mental age of about 6 years. It is therefore impossible to say if such individuals are intact or impaired in their theory of mind skills. We report the performance of very high functioning adults with autism or Asperger Syndrome on an adult test of theory of mind ability. The task involved inferring the mental state of a person just from the information in photographs of a person's eyes. Relative to age-matched normal controls and a clinical control group (adults with Tourette Syndrome), the group with autism and Asperger Syndrome were significantly impaired on this task. The autism and Asperger Syndrome sample was also impaired on Happé's strange stories tasks. In contrast, they were unimpaired on two control tasks: recognising gender from the eye region of the face, and recognising basic emotions from the whole face. This provides evidence for subtle mindreading deficits in very high functioning individuals on the autistic continuum.

No abstract

Previous work suggests children with autism show superior performance (in relation to their general mental age) on the Embedded Figures Test (EFT). Frith interprets this as showing that they have "weak central coherence". In Experiment 1, using an adult level version of this task, we aimed to replicate and extend this finding, first, by collecting response time (RT) data; second, by testing adults with autism of normal intelligence; and third, by testing a group of adults with Asperger syndrome, in order to test for differences between autism and Asperger syndrome. Both clinical groups were significantly faster on the EFT. In Experiment 2, we investigated if this difference was due to a preference for local over global processing, using a novel drawing task based on the classical Rey Figure. The clinical groups did not differ significantly on this test, but there was a trend towards such a difference. Alternative explanations for the EFT superiority in autism and Asperger syndrome are considered.

No abstract

The results suggest that individuals on the autism spectrum have impaired global coherence. Arranging sentences and making global inferences correlated highly, suggesting that central coherence may be a unitary force in these different tasks. Of the two clinical groups, the autism group had the greater deficit. The effect that such a deficit would have on one's daily life is discussed, along with possible explanations for the clinical groups' greater difficulty, and suggestions for future research.

Two groups of individuals, one with high-functioning autism and the other with Asperger syndrome were tested using Happé's Strange Stories Test of a more advanced theory of mind (Happé, 1994). This assesses the ability to interpret a nonliteral statement. Relative to normal controls who were IQ and age-matched, individuals with autism or Asperger syndrome performed less well on the task, while performing normally on a non-mentalistic control task. Individuals with autism or Asperger syndrome could provide mental state answers, but had difficulty in providing contextually appropriate mental state answers. Rather, their answers tended to concentrate on the utterance in isolation. This replicates Happé's result. Although the majority of both clinical groups provided context-inappropriate interpretations, the autism group had the greater difficulty. Results are discussed in relation to both weak central coherence and theory of mind.

Weak central coherence was investigated by exploring the conceptual integration of objects. Normally intelligent adults with either autism or Asperger syndrome were given two novel experiments. Experiment 1, the Object Integration test, had sets of line drawings depicting objects and people. Each set had to be either visually integrated to make the most coherent scene, or compared for similarities. The clinical groups were significantly impaired in their ability to integrate objects, but they were not impaired in looking for similarities. Experiment 2, the Scenic test, presented black line drawings of scenes containing an item that was inappropriate for the context. Participants were required to describe the scenes, identify the type of scene and context-inappropriate object, and locate a name (incongruent) object as quickly as they could. The clinical groups' descriptions suggest that they did not spontaneously pay preferential treatment to local details, nor were they faster at locating a named incongruent object. Whereas only a few of their descriptions lacked coherence, there was a deficit in both their ability to spontaneously notice and identify incongruent objects, as well as to identify the scenes. These tests provide support for Frith's (1989) central coherence hypothesis. Conceptual or high-level processing seemed inferior, whereas perceptual or lowlevel processing seemed normal, but not superior. Poor performance on these tasks characterized the majority of clinical participants, but those with autism performed at a consistently poorer level than those with Asperger syndrome. Possible explanations for the clinical groups' difficulties are explored along with suggestions for future research.

Individuals with an autistic disorder are less able to integrate visual elements. Of the two clinical groups, the autism group had the greater deficit, and it applied to the majority of the group. Possible explanations for the clinical groups' weak central coherence are explored.