Dissociations Within Semantic Processing in Young People with Down Syndrome: A Comparison with Children with Specific Language Impairment and Typically Developing Children

Down Syndrome (DS) is associated with patterns of strengths and weaknesses in language processing (Laws & Bishop, 2004), alongside marked cognitive impairment. A relative strength in lexical development has been characterised by receptive vocabulary as stronger than predicted i) from syntactic abilities, ii) from nonverbal MA (Glenn & Cunningham, 2005), and stronger compared to typically developing children (TD) matched for MLU. This study examines a possible differentiation within semantic processing system in DS, by making direct comparison with children with Specific Language Impairment, and TD children. Drawing on research with adult patients (Bozeat et al., 2000), conceptual matching tasks (e.g. matching camel to cactus), are sensitive to degraded semantic knowledge. Contemporary theories of semantic cognition (Lambon-Ralph 2011) imply a ‘semantic hub’ in the temporal cortex where links across different input and output modalities generate abstract conceptual knowledge. A key question was whether the marked cognitive impairment of individuals with DS further reflects degraded semantic knowledge and whether a semantic impairment in DS could fully characterise patterns of lexical development.

Young people with DS (N=16; 6 male, mean age 10y 2mo) were compared an SLI group (N=16; 9 male, mean age 7y 7mo), and a TD group (N=33; 17 male, mean age 6y 0mo). A novel conceptual matching test (Baby and Pram Test; 35 items) was developed and administered to all participants with standard tests of receptive vocabulary and phonological skills. A test of conceptual nonverbal ability (Leiter-R, Miller & Lucy, 1997) was administered to the DS and SLI groups only.

As shown in Table 1, all groups were matched on raw scores of the BPVS-II and were closely matched in verbal MA, but there was variable performance on the concept matching task. For Verbal MA, the DS group showed a relative strength compared to their Nonverbal MA (t (15) = -4.096, p < .0001), but the SLI showed a relative weakness of verbal MA, t (15) = -4.008, p < .0001. The key comparison was between the DS and SLI groups, matched for verbal MA, on two semantic tasks that used identical sets of target items i.e. conceptual matching (vocabulary depth) and word-picture matching (vocabulary breadth, Oullette, 1992). In a mixed ANOVA, main effects of task, p = .016, ηp2 = .09, and group, p < .0001 ηp2 = .28, and a significant interaction between task and group, F (2, 62) = 12.30, p < .0001 ηp2 = .28 were found.

As shown in Figure 1, the DS group were markedly impaired on the conceptual matching task (vocabulary depth), despite similar levels of performance on picture word matching (indexing vocabulary breadth) across groups. Findings imply that impaired conceptual development is a characterising feature of DS, despite their relative strength in vocabulary development, and will be discussed in relation to contemporary models of semantic cognition.