*4.3. Group Differences across ADOS-2 Items*

We explored the presentation of ASD symptomatology in this sample by comparing groups (i.e., DS + ASD and DS-only) across algorithm items of the ADOS-2. Analyses were conducted at the module level because algorithm items differ between modules. Items on the SA subdomain of the ADOS-2 were more likely to differentiate individuals with DS + ASD from those with DS-only, with four common items emerging across modules: unusual eye contact, quality of social overtures, amount of reciprocal social communication and overall quality of rapport. Of note, all four significant algorithm items in the SA subdomain rely on pragmatic social communication skills, which have been found to be delayed among young children with DS and other forms of ID relative to normative age expectations. In previous studies, children with DS have also been found to have a relative weakness in pragmatics in comparison with their structural language abilities [37]. Additionally, the finding that unusual eye contact differentiated the groups is of interest given that atypical eye contact in DS has been detected as early as infancy [4] and the use of eye contact among children with DS is less clear than among their typically developing (TD) peers and peer with other developmental disabilities (DD) [38]. These results suggest that challenges in pragmatic social communication are characteristic of the DS phenotype and indicate the need to clarify boundaries between standard phenotype heterogeneity from comorbid ASD symptomatology.

At the same time, many commonalities were observed among SA items that did not differentiate the two DS groups. In Module 2, items measuring pointing, shared enjoyment and joint attention did not significantly differentiate individuals classified as DS-only from those classified as DS + ASD. Similarly, the following four SA items from Module 3 did not significantly differentiate the groups: conversation, descriptive gestures, facial expressions and shared enjoyment. Notably, shared enjoyment did not discriminate between groups in either module, a finding that is not entirely surprising given that, although individuals with DS may struggle with certain aspect of social communication, findings suggest relative strengths in social engagement and social orientation (i.e., sociability) [39]. Additionally, in Module 2, joint attention abilities did not differentiate between participants' group membership. This finding is consistent with literature finding that joint attention may be a relative strength among individuals with DS, who perform similarly to children with TD and at a higher level than children with ASD and other NDDs [40]. Although joint attention (JA) was not explicitly measured in Module 3, it could be argued that joint attention is

a developmental antecedent to the use of descriptive gestures, which is measured in Module 3 [41]. The finding that the item measuring the use descriptive gesture also did not differentiate groups may suggest that the use of JA may continue to be a strength among individuals with DS throughout childhood. These results suggest that some phenotypic characteristics of DS, such as relative strengths in social orientation and joint attention, may offset the presentation of these skills among individuals with DS + ASD. Further research is needed to better understand how phenotypic characteristics of DS affect ASD symptom presentation among individuals with DS + ASD.

Three items from the RRB subdomain of the ADOS-2 (stereotyped/idiosyncratic use of words and phrases, hand, finger and other complex mannerisms and unusually repetitive interests or stereotyped behaviors) significantly differentiated individuals with DS-only from those with DS + ASD in Module 2, and (after correcting for multiple comparisons) no item from the RRB subdomain differentiated the groups in Module 3. The finding that scores on the RRB subdomain differentiated groups in Module 2 but not Module 3 is of particular interest, as it mimics earlier reported findings that severity scores in the RRB subdomain did not differentiate individuals who were classified as having DS-only from those classified as having DS + ASD. As explained above, there are several possible explanations for this finding, although further research is needed to better understand this phenomenon.

The findings regarding differences in symptom profile and severity between the DS + ASD and DS-only groups have more general implications for the field. In particular, in examining the high cooccurrence of ASD and fragile X syndrome (FXS) [42], we have argued previously that the categorical diagnosis of ASD can hide mechanistically and clinically important differences among individuals with FXS, and between those with FXS+ASD and those with non-syndromic ASD. Moreover, we have provided empirical support for that claim in several studies providing in-depth analysis of both the ADOS-2 and the ADI-R and in multiple samples of different ages and degree of impairments [43,44]. In the present study, too, we have shown that an ASD diagnosis can "mask" different levels of severity and symptom profiles in individuals with DS as a function of the ADOS-2 module administered. This finding, we believe, has less to do with the specific characteristics of the module administered and is, instead, a reflection of that ways in which ASD symptoms are moderated by the phenotype of DS and within-syndrome variability in that phenotype. Simply focusing on whether an individual with DS meets or does not meet criteria for an ASD diagnosis, therefore, could have the consequences of a failure to understand the factors leading to those symptoms or the best approach to treatment to reduce those symptoms.
