A Link Between Grammar and Dyspraxia in Boys with Autism

Background:

Along with deficits in communication and language that define autism spectrum disorders (ASD), dyspraxia (motor skill deficits) is also commonly observed. It has been proposed that abnormalities of procedural memory may underlie impaired development of motor skill knowledge and grammatical (but not lexical) abilities. Examination of English past tense morphology offers a well-studied contrast between grammatical and lexical aspects of language. Dual-system models claim that regularized past-tenses (walked) depend on rule-based grammatical computation (stem + -ed), while irregularized past-tenses (slept) depend on memorized word-specific (lexical) knowledge. Thus the former but not the latter is predicted to share a common neurocognitive substrate with motor skill knowledge.

Objectives:

To examine the relationship between grammatical performance (past tense production) and dyspraxia in children with ASD.  

Methods:

Sixteen high-functioning boys with ASD (ages 8-15 years) participated in both a past-tense production task and a videotaped praxis examination. For the past tense production task we examined accuracy at producing the past tenses of real regular (walked; n=32) and real irregular (slept; n=32) verbs. For the praxis measure, we examined mean accuracy (percent correct responses) during performance of gestures-to-command, gestures-to-imitation, and gestures-with-tool-use. Pearson partial correlations between past tense accuracy and praxis accuracy were computed, with age (years) included as the partialled variable (given the potential for age effects in both the past-tense production and praxis measures). The pattern did not change when age was not partialled out of the correlations.

Results:

For the boys with ASD there was a significant positive correlation between mean regular verb and praxis accuracy (r(13)=0.58, p=0.02); whereas the correlation between mean irregular verb and praxis accuracy was not significant (r(13)=0.39, p=0.15). Significant negative correlations between regular verb performance and each of the praxis error measures (gesture-to-command, gesture-to-imitation, and gesture-with-tool-use) were also found (all p<0.05); whereas, again, none of these correlations reached significance for irregular verb performance (all p>0.12).  

Conclusions:

The relationship between regular (but not irregular) verb accuracy and praxis scores is consistent with a common neurocognitive substrate for grammar and motor skills, that may underlie impaired performance in both domains in boys with autism.  Specifically, the observed correlations suggest that both may be related to underlying difficulties with procedural memory necessary to acquire these skills.