Measuring cognitive processes in children with language impairment

Research evidence indicates that the problems of children with language impairment whose language development is slower than that of typically developing peers can be accounted for not only by problems in areas of language, such as morphosyntax (word inflections and the grammatical rules governing them), semantics, and pragmatic competences,¹ but also by limitations in underlying cognitive processes. Relative to typically developing or language age-matched controls, children with language impairment are more likely to have slower response times across a wide range of tasks and problems in short-term and working memory. Theories of memory, such as that of Baddeley, distinguish between temporary, immediate short-term storage and recall of untransformed information – visual as well as verbal – which has not been subject to further processing (short-term memory) and the temporary storage and recall of processed verbal or visual material necessary for the completion of complex and demanding cognitive tasks (working memory).²

The distinction between separate processes for verbal and visual information is important as language impairment may be associated with ‘domain specific’ or ‘domain general’ impairments in working memory, which may have different implications for support and intervention. In the case of the former, the problems of children with language impairment would be more likely to be specific and confined to the verbal domain and in the latter, the problems would be more general, with deficits in both verbal and non-verbal domains.

Baird et al.³ used the data from school-aged participants recruited for a genetics study to shed further light on this issue by investigating whether the children’s performance on measures of visual and verbal memory, learning, and processing speed showed domain specificity. Children with current language impairment were compared with two control groups: a group of their siblings with resolved language impairment and a group of unaffected siblings with no history of language impairment. The findings reported by the authors provide support for the view that the cognitive processing problems of the children with language impairment in the study relative to non-affected siblings were specific to the verbal domain rather than global, and that the degree of impairment was correlated with the severity of the language impairment. Compared with unaffected children, the group of children with resolved language impairment also had problems in the verbal domain.

These findings are consistent with other published studies⁴ and help to clarify the theoretical relationship between memory functioning and language impairment. But uncertainty remains regarding the direction of causality in the light of the importance of linguistic knowledge for success in non-word repetition and sentence recall tasks. The results also have implications for compensatory educational support which capitalizes on the children’s relative areas of strength.

Yet the study also raises questions about the assessment of cognitive processes in children with language impairment. Studies of the task demands of some standardized measures of processing speed, short-term memory, and working memory reveal that the measures tap into a range of cognitive processes including, for example, attention and executive functions, or that the verbal demands of the tasks are problematic for the participants.⁵ Test sensitivity can also vary such that different test batteries may be more likely to yield between-group differences.

Taking speed of processing in children with language impairment as an example, many key research studies use direct measures of response times to verbal or non-verbal stimuli. In contrast, Baird et al. used standardized pencil and paper timed tasks. But although different in character from the simple and choice response time measures used by other researchers, these tasks produced a similar pattern of results, as the authors note.

Effects of heterogeneity seem to be more marked, however, in the case of measures of short-term memory and working memory. For example, a recent study carried out in Germany using a different battery of short-term memory measures from that used by Baird et al. reveals that children with receptive-expressive language impairment (i.e. with impairment in both language comprehension and production) had problems in both visual and verbal tasks in contrast to children with expressive language impairment who had difficulties in verbal tasks only.⁶ This suggest that children with receptive-expressive language impairment have domain-general problems in short-term memory. Further, a recent study investigating working memory in children with language impairment found that, relative to controls, they had less efficient inhibition, i.e. the ability to suppress a dominant but inappropriate response or irrelevant information, which was associated with working memory capacity.⁷

Care needs to be taken, therefore, to ensure that selection of test instruments for research in this area is informed by theoretical models and that batteries of reliable, valid, and sensitive measures of underlying cognitive processes of interest are administered.