Do Children with Specific Language Impairment Have a Cognitive Profile Reminiscent of Autism? A Review of the Literature

Background: There is debate regarding the relationship between autism spectrum disorder (ASD) and specific language impairment (SLI), with some researchers proposing etiological overlap between the conditions and others maintaining their etiological distinction. The language phenotype of SLI is characterised by structural language difficulties, which contrast with the pragmatic language and broader developmental difficulties observed in children with ASD. On the basis of these differences, SLI and autism have traditionally been considered separate disorders, with distinct aetiologies. However, in recent years, evidence has emerged that the diagnostic boundary between autism and SLI may not be clear cut.

Objectives: Although there is an accumulating body of literature comparing the language phenotypes of ASD and SLI, no study has undertaken a systematic investigation of potential cognitive overlap between the two conditions. The overarching aim of the current review was to compare the cognitive phenotypes of ASD and SLI described in the published literature. The current review focused on the empirical literature that has investigated cognitive characteristics of ASD (namely impaired theory of mind and emotion recognition, executive dysfunction and weak central coherence) in children with SLI.

Methods: Published literature was examined for empirical articles that met the following criteria: (1) examined cognitive characteristics of ASD (Theory of Mind [ToM], emotion recognition, executive function and central coherence) in children with SLI, or (2) compared the cognitive phenotype of ASD and SLI with regard to the aforementioned cognitive characteristics of ASD. Findings were reviewed and synthesized.  

Results: Overall, findings were inconsistent and there is a lack of substantive evidence supporting overlapping cognitive phenotypes in autism and SLI. While several studies have found that children with SLI have intact ToM, other studies suggest that children with SLI have difficulty with ToM tasks. The mixed findings may relate to methodological limitations commonly observed across these studies. Observed ToM impairments in children with SLI may relate to the language difficulties of this population, rather than an underlying cognitive deficit. With regard to emotion recognition, while children with SLI are able to identify facial expressions, they have difficulty understanding and expressing auditory affective information. Results relating to executive function in children with SLI were also mixed. While some authors have reported planning and set-shifting deficits in children with SLI, there are a number of notable failures to replicate. There is also considerable debate as to whether language impairment is causal in the executive function deficits observed in children with SLI. Only one study has examined central coherence in children with SLI, finding that children with SLI did not manifest specific deficits in visuospatial processing (i.e., in either global or local processing).

Conclusions: There is little consistent evidence supporting the hypothesis that autism and SLI have overlapping cognitive phenotypes. Better powered and more rigorous experimental designs, as well as studies directly comparing the cognitive phenotypes for SLI and ASD will further elucidate the etiological relationship between these two conditions.