Dochildrenwithautismfailtoprocessinformation in context? Beatriz Lo ´ pez 1 and Susan R. Leekam 2 1 School of Sports Science and Psychology, York St. John College, UK; 2 University of Durham, UK Background: This research investigated the proposal that children with autism are impaired in pro- cessing information in its context. To date, this proposal rests almost exclusively on evidence from verbal tasks. Given evidence of visuo-spatial proficiency in autism in other areas of functioning, it is possible that the ability to use context is spared in the visual domain but impaired in the verbal domain. Method: Fifteen children with autism and 16 age and IQ-matched typically developing chil- dren were tested on their ability to take account of visual context information (Experiment 1) and verbal context information (Experiment 2) using an adaptation of Palmer’s (1975) visual context task. They were also given an adaptation of Tager-Flusberg’s (1991) visual and verbal semantic memory task (Experiment 3) and Frith and Snowling’s (1983) homograph task (Experiment 4). Results: Experiment 1 showed that children with autism were facilitated by the provision of visual context information. Experiments 2 and 3 showed that the same children were also able to use both verbal context infor- mation when identifying words and semantic category information in a verbal task when naming and recalling words. However, in Experiment 4 these children had difficulties with a sentence-processing task when using sentence context to disambiguate homographs. Conclusions: These findings dem- onstrate that children with autism do not have a general difficulty in connecting context information and item information as predicted by weak central coherence theory. Instead the results suggest that there is specific difficulty with complex verbal stimuli and in particular with using sentence context to disam- biguate meaning. Keywords: Autism, cognition, context, central coherence. Abbreviations: AD: autistic disorder; TD: typically developing. Since Frith’s (1989) original claim that people with autism lack a natural drive towards central coher- ence, there has been increasing interest in this pro- posal as an explanation for the perceptual and cognitive difficulties of people with autism. A key contribution of the proposal is the claim that indi- viduals have a unique profile of perceptual and cognitive abilities in which superiority in processing local, featural information is contrasted with inferi- ority in processing global and context information. Evidence for this pattern has been examined with respect to low-level visual processing (Happe ´, 1996; Plaisted, Swettenham, & Rees, 1999; Ropar & Mitchell, 1999; O’Riordan & Plaisted, 2001; O’Rior- dan, Plaisted, Driver, & Baron-Cohen, 2001), high- level visuo-spatial processing (Shah & Frith, 1983, 1993; Brian & Bryson, 1996), semantic memory (Tager-Flusberg, 1991) and sentence processing (Hermelin & O’Connor, 1967; Happe ´ , 1997; Jolliffe & Baron-Cohen, 1999). The weak central coherence account of autism is wide ranging in its scope of application, yet the em- pirical evidence for this proposal is still not clearly established. On the one hand, there is considerable evidence to support the claim that children with autism have enhanced ability to discriminate fea- tures (Shah & Frith, 1983, 1993). On the other hand, there are contradictory findings as to whether this ability occurs alongside impairment in global pro- cessing. Some studies do show evidence of deficits in the ability to process global information (Hermelin & O’Connor, 1967; Frith & Snowling, 1983; Happe ´, 1997; Jolliffe & Baron-Cohen, 1999), whereas other studies show no differences between individuals with autism and non-autistic populations in global tasks (Ropar & Mitchell, 1999; Brian & Bryson, 1996; Plaisted et al., 1999; Pring & Hermelin, 1993; Ramondo & Milech, 1984). These contradictory findings may be best under- stood by examining the way that global processing has been conceptualised across different perceptual and cognitive tasks. The impairment in global pro- cessing has been interpreted both in terms of a conceptual semantic deficit as originally demon- strated in reading and memory tasks and in terms of a failure to extract holistic perceptual properties as demonstrated by studies of visuo-spatial processing (Happe ´, 1996). Studies that fail to show differences between autism and non-autism populations tend to be those using visuo-spatial tasks such as visual il- lusions or the Navon task (Ropar & Mitchell, 1999; Ozonoff, Strayer, McMahon, & Filloux, 1994). This lack of impairment in visuo-spatial tasks seems not to be solely confined to low-level visual processing. Evidence has also been found for intact higher-level semantic processing ability when information is presented visually (Brian & Bryson, 1996; Pring & Hermelin, 1993). The difficulty with global processing is often described as a difficulty with processing context. Journal of Child Psychology and Psychiatry 44:2 (2003), pp 285–300 Ó Association for Child Psychology and Psychiatry, 2003. Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA