Longitudinal Look At Expressive, Receptive and Total Language Development in Individuals with Autism Spectrum Disorders

Background: Individuals with autism spectrum disorders are currently classified with impairments in both expressive and receptive language. Language impairments range from difficulty acquiring spoken language and in pragmatic language to never acquiring language. Longitudinal investigation of the development of specific types of language is needed from childhood into late stages of adulthood for individuals with autism. Doing so will help understand specific developmental trajectories and targets for new language interventions.

Objectives: Our goal was to understand longitudinal development of receptive, expressive, and total language in autism and controls by examining repeated measures of the CELF language assessment instrument.

Methods: Participants included 71 male controls (age range 3-39yrs at first test) and 106 males with autism (age range 3-45yrs at first test). Expressive, Receptive and Total language function was obtained from the Clinical Evaluation of Language Fundamentals (CELF) administered 1 to 4 times over 13 years (for a total of 324 assessments). A diagnosis of autism was obtained using the Autism Diagnostic Interview (ADI) and the Autism Diagnostic Observation Schedule (ADOS). In order to compare across multiple versions of the CELF (Preschool, CELF-3, CELF-4) we created z-scores for each test version based on the control sample. Linear Mixed models examined linear and nonlinear age effects as well as group by age interactions.

Results:  The z-scores for receptive, expressive and total language were significantly lower in the autism group than controls (all p<.001). For receptive language the autism group showed age related increase in z-score (on average 10% per year), which differed significantly from controls (on average 1.56% per year; group by linear age interaction (p<.001) with a trend toward group difference in nonlinear age related changes (p=.097). For expressive language we did not find age related changes in autism. For total language the autism group showed linear age related changes (p<.001) that did not differ significantly from controls. 

Conclusions: As expected the expressive, receptive and total language z-scores are decreased in high functioning individuals with autism compared to typical development. Receptive language performance in the autism sample increased over time, which was different than in typical development. However, this trend in improved CELF performance was not found in expressive language. Age related changes in total language were drive by receptive z-scores. To our knowledge, this is the first longitudinal study looking at raw age-related changes of language in autism. This may show continued increase in receptive language ability in individuals with autism and the need to create interventions that improve expressive language function.