A test of central coherence theory: Can adults with high-functioning autism or Asperger syndrome integrate objects in context?

Abstract

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.