A systematic review and meta-analysis of morphosyntactic skills in Williams syndrome

Williams syndrome (WS) is a rare genetic neurodevelopmental disorder, with a cognitive phenotype characterised by relative peaks and valleys. Findings on language skills remain controversial and sometimes contradictory. While, traditionally, some authors advocated for fully preserved language skills in WS, more recent studies point to delayed language skills, or even language impairment. To gain more clarity in this area, we performed a systematic search and meta-analysis on morpho-syntactic skills in WS. We conducted a systematic search covering six relevant databases that yielded a total of 383 articles, of which 38 met the inclusion criteria: (a) peer review articles; (b) comparisons between WS and control groups; (c) effect sizes were provided or could be determined; and (d) tasks assessed morpho-syntactic abilities. In total, 1878 participants and 92 effect sizes were included in the meta-analysis. The calculation of the cumulative effect size and publication bias was assessed using Robust Bayesian Meta-Analysis. Results showed that people with WS have worse morpho-syntactic abilities compared to typically developing (TD) people, both when matched by chronological or mental age. However, individuals with WS perform better compared to people with other disabilities. According to task type, while results revealed a worse performance in WS compared to controls in production, receptive and metalinguistic knowledge tasks, no significant differences were found in repetition tasks between WS and controls. These results support the neuroconstructivist approach, in which language would follow a distinct and unique development in WS, due to their genetic setup, which can be modeled by using connectionist methods.