Expressive Language Profiles of Children with Idiopathic Autism and Fragile X

Background: Fragile X syndrome (FXS) is associated with significant genetic risk for autism, with 67-74% of males with FXS meeting criteria for ASD (Harris et al., 2008). The study of autism in FXS (a single gene disorder) provides a paradigm for examining phenotypic variation within a simplified genetic context, potentially informing genetic and neural mechanisms in autism. Both children with ASD and FXS show impaired performance on standardized expressive language assessments (Kjelgaard & Tager-Flusberg, 2001; Roberts et al., 2001). This study aimed to define potentially overlapping language phenotypes in ASD and FXS, and to build on previous findings by examining expressive language in spontaneous, naturalistic language samples, which may compliment findings from standardized assessments (Evans, 1996).

Objectives: To compare intelligibility, fluency, rate, and pragmatic skills within a semi-naturalistic conversational sample among boys with ASD, FXS with ASD (FXS-ASD), FXS-only (FXS-O) and typical development (TD), in order to define syndrome-specific expressive language profiles.

Methods: Participants included boys with ASD (n=21), FXS-ASD (n=34), FXS-O (n=21) and TD (n=19), aged 3-12. Groups were similar according to nonverbal mental age (Leiter-R) and receptive/expressive language (EVT-2 and PPVT-4 composite). The ADOS was administered to determine ASD status, and served to elicit the semi-naturalistic conversational language. The first 150 conversational turns were transcribed using SALT conventions (Miller & Chapman, 2000). Overall intelligibility was measured as the percent of intelligible utterances. Fluency was assessed via the number of reformulations that included word repetitions, and the number that included word revisions. Rate was measured as the number of words spoken by the child or examiner divided by the total minutes elapsed. Pragmatic variables included the number of exact repetitions of the examiner’s utterance (echolalia), and the percent of questions to which the child responded. MANCOVA was conducted to test for group performance on the six variables, co-varying for mean length of utterance in morphemes.  

Results: Results indicated a significant overall effect for group. Bonferoni-corrected group comparisons revealed that ASD and TD groups responded to significantly fewer questions than FXS-O (M=-.095; M=-.107, respectively) and FXS-ASD (M=-.092; M=-.104). Children with FXS-ASD were less intelligible than controls (M=-.051) and produced more reformulations with repetitions than children with ASD (M=7.10). Both FXS groups used more words per minute than ASD (M=-19.12; M=-19.02) and TD (M= 14.49; M=14.38) groups. Groups did not differ on imitations or reformulations with revisions.

Conclusions: Children with FXS, regardless of autism status, were more responsive to questions and showed faster rates of speech than children with ASD and TD. The FXS-ASD and FXS-O groups did not differ on any of the variables measured. However, children with FXS-ASD were less intelligible than those with TD and produced more repetitious reformulations than children with ASD, while these differences did not reach significance in FXS-O. Findings provide evidence of a FXS-specific expressive language profile that presents regardless of ASD status, and may differ from that seen in ASD.