By comparing the pragmatic language abilities of children with idiopathic autism or FXS (both with and without autism), with children with DS and TD children, this study aimed to determine the extent to which pragmatic language impairment may overlap in autism and FXS, and may potentially be tied to underlying molecular genetic variation related to FMR1, the gene that causes FXS. Additionally, we explored theory of mind ability as a potential correlate of pragmatic language across groups. Prior studies have reported a link between impaired theory of mind and pragmatic language use in autism, but to our knowledge this question has not yet been addressed in FXS.
Results indicated that the ASD-O and FXS-ASD groups looked quite similar on direct-assessment of pragmatic language using the CASL, with both groups performing more poorly than the FXS-O, DS, and TD groups. Yet on teacher report findings were more divergent (e.g., Initiation, where the FXS-ASD group scored significantly lower than the ASD-O group). It could be the case that a global measure of pragmatic language ability such as the CASL obscures actual differences between these groups. Alternatively, informant-based methods such as the CCC-2 may introduce measurement error that complicates group comparisons (e.g., different teachers may have different thresholds for ratings, based on their prior experience, the composition of their classrooms, etc.). Further research comparing these groups using direct-assessment measures of specific types of pragmatic language ability will be valuable in addressing this question and determining the extent of overlap in pragmatic language impairment in autism and FXS.
Analyses of theory of mind ability revealed patterns of performance quite similar to those observed in the CASL test of pragmatic language – the ASD-O and FXS-ASD groups performed most poorly, and children with FXS-O did not differ significantly from controls. In this case, however, the DS group performed more like the ASD-O and FXS-ASD groups. We also found that theory of mind ability was associated with pragmatic language on the CASL for all groups, where better theory of mind scores were associated with more pragmatic language competence. Although we cannot draw definitive causal conclusions from the present data, these findings certainly support the hypothesis that the ability to understand and predict one’s own and others’ thoughts, feelings, intentions, and desires is a critical skill underpinning competent pragmatic language use (Sperber and Wilson, 2002
; Wilson and Sperber, 2004
). When theory of mind is impaired, as was the case for the ASD-O and FXS-ASD groups, children may be ill equipped to contend with the demands of social discourse, and less apt to glean information necessary for developing pragmatic language skills. Such a relationship has been demonstrated across a range of pragmatic language skills in autism (Loveland and Tunali, 1993
; Tager-Flusberg and Sullivan, 1995
; Surian et al., 1996
; Capps et al., 1998
; Tager-Flusberg, 2000
), and our findings suggest a similarly important role in the pragmatic language problems observed in a subgroup of children with FXS who show pragmatic language impairments as well. That significant associations were detected in all groups, even those who did not show significant pragmatic language impairment, may demonstrate the important role of theory of mind in supporting more fluent pragmatic language use as well. It is of course also possible that theory of mind tasks and pragmatic language are tapping some additional mediating (or moderating) abilities.
Patterns observed in the FXS-O and DS groups may also be informative, particularly with regard to defining syndrome-specific language and social cognitive profiles across these different groups. In particular, whereas social skills are generally considered to represent a relative strength in individuals with DS, the literature on pragmatic language in DS is actually quite mixed, with documented challenges compared with MA-matched TD children including initiation and elaboration of topics (Tannock, 1988
; Roberts et al., 2007a
), initiation of communicative repairs (Abbeduto et al., 2008
), and clarity of messages (Abbeduto et al., 2006
). Thus, our finding that boys with DS performed comparably to boys with FXS-ASD is not necessarily surprising. On the otherhand, we may have found significant differences between FXS-ASD and DS groups with a larger sample size or if we examined particular aspects of pragmatic language with direct-assessment measures (and it is important to note that the DS group did not differ significantly from the TD group, whereas the FXS-ASD group did perform significantly more poorly than the TD group). Thus, interpretation of these similarities with the present data is not straightforward.
In the FXS-O group, these data indicated that pragmatic language and theory of mind were relative strengths, and deficits in these areas may be restricted only to those with FXS-ASD, suggesting that pragmatic language deficit (or theory of mind) is not a core characteristic of FXS but rather autism in FXS. This is consistent with findings from Roberts et al. (2007a
), who found that boys with FXS-O did not produce more non-contingent language than TD boys, but that the FXS-ASD group produced more non-contingent language than both of these groups. However, it is important to note that the difference between the TD and FXS-O groups approached significance so may have revealed true differences with a larger sample.
Though not a primary focus of the current study, findings do have some important clinical implications. Given that boys with FXS-ASD showed more pragmatic language impairment than boys with FXS-O, performing comparably to boys with idiopathic autism on a direct-assessment measure, the diagnosis of ASD in boys with FXS should be considered during assessment and clinicians may consider interventions that have been studied in the context of ASD when tailoring intervention approaches for boys with FXS-ASD. Our divergent findings depending on assessment method also support the use of multiple assessments, including natural language samples, to fully characterize pragmatic language ability and identify specific targets for intervention which may differ across groups and individuals.
The group similarities in directly assessed pragmatic language ability and theory of mind in ASD-O and FXS-ASD may have important implications for furthering knowledge of the brain and gene basis of these complex skills. In particular, because much is known about the molecular and neurobiological basis of FXS, the considerable overlap observed with ASD-O may help to define specific phenotypes associated with known genetic variation, in this case variation in the FMR1. We observed correlations with molecular genetic variables that support this association – pragmatic language on the CASL and theory of mind were both associated with FMR1-related variation in the FXS group. Specifically, greater methylation was associated with lower theory of mind performance and more impaired pragmatic language ability. Higher CGG repeat numbers were also related to poorer pragmatic language skills. Genetic variables showed additional associations with general cognition and structural language, which is perhaps not surprising given that general cognitive and language functioning certainly contribute to pragmatic language and theory of mind abilities. By providing a link between genetic and phenotypic variation, these findings may offer a foothold for understanding gene-behavior relationships in atypical and typical development alike.
This study has some limitations. First, we determined autism status primarily with the ADOS, but future studies should utilize information from both the ADOS and ADI-R for all participants to confirm autism status. Second, we did not examine all potential underlying mechanisms of social communication, such as anxiety or various aspects of executive function. Third, we examined social communication and theory of mind at one time point and in boys only. Future studies should assess these skills longitudinally and in both boys and girls.
In sum, this study identified pragmatic language and theory of mind as important abilities that are impaired in autism, and in a subgroup of children with FXS who also meet criteria for autism. This considerable phenotypic overlap between autism and a known monogenic condition suggests that impairments in pragmatic language ability and theory of mind may be tied to a particular genetic variant – the FMR1. Further studies are needed to clarify those particular types of pragmatic language difficulties common to both conditions, given that results from the pragmatic language subscales on the informant-based CCC-2 were not as straightforward as those obtained from direct-assessment of pragmatic language ability, or theory of mind for that matter. An additional important area for further study concerns the brain basis of these abilities, and the extent to which impairments may stem from similar neural architectural differences. By integrating detailed phenotypic analysis with neuroimaging studies in autism and FXS, future research may provide important insights into the role of FMR1 in social-communicative phenotypes.