Indicators of Linguistic Processing Constraints in the Narratives of Individuals with High-Functioning Autism

Background: Few standardized measures are available to assess the language skills of adolescents and adults with HFA. Furthermore, current measures fail to capture the “idiosyncratic” and “stereotyped” speech productions that are characteristic of HFA. The overuse of formulaic language is one of the behaviors that is suggestive of autism on the ADOS, however, methods to measure the use of this type of language have not been formalized. Wray (2000) proposed that the overuse of formulaic language or prefabricated patterns may be evidence of linguistic processing constraints. The measurement of formulaic language would be important with respect to information processing models of autism. Systems for the measurement of formulaic language have been developed and used with adults with brain injuries (Van Lancker Sidtis & Postman, 2006) and second language learners (Wray, 2000), populations that are reported to overuse formulaic language. These analyses may be useful for characterizing the language of individuals with HFA.

Objectives: To analyze spoken language samples of individuals with HFA as compared to age- and IQ- matched controls to determine if differences occur in measures of formulaic expressions (FE) under two different language formulation conditions.

Methods: Language samples for analysis were taken from 15- to 35-year old males with HFA (n=20) and neurotypicals matched for age and IQ all with verbal IQs greater than 85. Autism diagnosis was established for the affected group with the ADOS and ADI-R, and confirmed by expert clinical impression. Narrative language samples for both groups were collected from the ADOS “Telling a Story from a Book” and “Create a Story” tasks. Language samples were transcribed using Hunt's (1970) utterance boundary coding procedure. Transcript reliability was established (.99) with another graduate student. The Systematic Analysis of Language Transcripts (SALT) software (Miller & Chapman, 2000) was used for standard language analyses (e.g., type token ratio and mean length of utterance). Analysis of the transcripts for evidence of FE is ongoing using a classification system developed by Van Lancker Sidtis and Postman (2006).

Results: Initial analyses suggest a wide range of verbal fluency among the individuals with HFA. Four participants with HFA had clinically significant rates of speech disruptions (11.49, 11.51, 13.24, and 18.09). Classification and measurement of FE is ongoing for both groups. Further analyses will be reported and comparisons will be made to the performance of age- and IQ-matched controls. The relation between the rate of occurrence of FE between the two tasks that differ in linguistic formulation demands (one with a given story with picture support and the other requiring creation of a novel story with generic props) will also be discussed.

Conclusions: High-functioning adolescents and adults with autism are reported to produce “stereotyped” and “idiosyncratic” utterances at higher rates than neurotypical individuals with similar language abilities. However, objective methods of measuring this type of language use are not generally available. Measurements of FE during spontaneous speech productions may be useful for characterizing the language production of individuals with HFA and may provide evidence related to the use of linguistic processing resources.