New publication: Mutualistic coupling of vocabulary and non-verbal reasoning in children with and without language disorder
One of the key findings from the first phase of SCALES was that children with language disorder also frequently have lower the an average non-verbal reasoning skills. This finding has been used to challenge the use of non-verbal ability cut-offs for diagnosis and for accessing speech and language services. In our new paper, recently published in Developmental Science, we explore why difficulties in non-verbal reasoning seem to go hand in hand with language disorder.
People sometimes think of general intelligence as a stable trait which determines other skills, such as understanding and producing spoken language. Many speech and language therapy services do not offer therapy to children with low non-verbal reasoning skills, because they believe non-verbal reasoning places a cap on children’s ability to learn language making the therapy ineffective. However a recent theory of cognitive development called “mutualism” has challenged the idea that all abilities depends on a stable general intelligence trait, instead suggesting that abilities in different areas (such as language and non-verbal reasoning) can support each other’s development.
Mutualism is a term borrowed from ecology. In ecology it describes the relationship between two species, like aphids and ants, which live together because they each benefit from the others existence. In cognitive development, language and non-verbal reasoning are the aphids and ants. Mutualism theory suggests that children with good language should develop reasoning skills more quickly than children with poor language, because their language is helping them gain new reasoning skills. And likewise, children with more general cognitive deficits may find language learning more challenging than peers. This would explain why children with language disorder also often have difficulties with non-verbal reasoning, and why children with intellectual disability also often have poor language.
Previous studies have focused on children who do not have developmental disorders and have demonstrated that changes in language are ‘coupled’ with changes in non-verbal reasoning. Importantly, these relationships go in both directions – better language drive changes in non-verbal reasoning just as much as good non-verbal reasoning leads to greater improvement in language. In typical development then, language and non-verbal reasoning are mutually influential. However, it was an open question as to whether these mutualistic relationships were intact in developmental disorders. We teamed up with the lead author from these previous studies, Rogier Kievit, to explore whether mutualism between verbal and non-verbal skills was also present in children from SCALES that met the criteria for language disorder.
We analysed SCALES data from Year 3, Year 6 and Year 8 (collected prior to pandemic onset). To measure non-verbal reasoning, we used scores from a task called Matrix Reasoning, in which children had to select a geometric shape to complete a puzzle. To measure language, we used scores on a receptive vocabulary test. The advantage of using this language measure, out of all the other ones we have in SCALES, is that we used the same version of the test at all three time-points, whereas other language measures had to change as the children got older and the tasks became too easy.
We first replicated the findings that language and non-verbal reasoning are mutually influential in the entire cohort, such that those with better language skills show greater growth in non-verbal reasoning and vice versa. We then tested whether there was a difference in strength of mutualistic relationships between language and reasoning in children with and without language disorder and found no evidence for a difference.
We think these findings have some important implications for practice. First, they are another nail in the coffin for Specific Language Impairment (SLI) as a diagnostic category. Selective impairments in language are pretty rare and it is clear that language is intimately related to other aspects of cognitive and behaviour development. Second, our findings suggest that targeting the language of children with low IQ has the potential to help not just their language but also their non-verbal reasoning skills.
Therefore low non-verbal reasoning IQ scores should not be a reason for exclusion from speech and language therapy. However, it may be that children with broader cognitive challenges (who are likely to have more severe language difficulties) require more language input and/or input that is structured in a way that makes language easier to learn. Finally, our findings raise interesting questions about how to optimise interventions for language disorder. For example, it might be that focusing not just on children’s language skills but also on general cognitive skills could be a more efficient way to support language development. We definitely need intervention trials to include children with a range of cognitive profiles to test these predictions!