Peaks of ability as a subtyping tool for autism

Background: The autistic spectrum is subtyped in DSM-IV by the presence or absence of certain signs (e.g. speech delay). However, DSM-IV fails in operationalizing the autism vs Asperger syndrome distinction (Szatmari, 2000). Differences in cognitive profile between Asperger syndrome and autism are reported (Ghaziuddin & Mountain-Kimchi, 2004), but not included in diagnostic criteria. Despite clinical recognition of contrasted phenotypes, autism and Asperger syndrome are considered as arbitrary segments within an autistic spectrum continuum (Macintosh & Dyssanayake, 2004).

Objectives: The aim of the present study was to investigate how Wechsler intelligence scales subtest peaks can predict the presence or absence of DSM-IV phenotypic markers relevant for the autism vs. Asperger distinction.

Methods: Subjects were 250 participants ADI-R and ADOS-G positive for autism, from the Riviere-des-Prairies Hospital (Montreal) database. Participants without ADI-defined speech delay or atypicalities (echolalia, stereotyped language, pronoun reversal) were re-coded as Asperger. Participants with identified neurological or genetic syndromes (apart from idiopathic epilepsy) were excluded. Wechsler peaks were individually computed as a 1.5 SD difference between scaled scores and Wechsler full scale IQ baseline.

Results: Initial small sample results indicate the possibility of peaks which predict a "0" score for 5 ADI items (one word delay, two word delay, echolalia, stereotyped language, pronoun reversal) and peaks which predict scores higher than 1 at these ADI items. Results from a larger data set will be reported.

Conclusions: A combination of cognitive and behavioral dimensions may be a valid candidate for the subtyping of the autistic spectrum according to DSM-IV rules.