Deficits in sound pattern sequencing in children with specific language impairment: A networks approach
Children with specific language impairment (SLI) demonstrate primary deficits in morphosyntax, which has served as the central theme in theoretical and clinical approaches. However, a striking number of children with SLI also exhibit speech sound deficits, characterized both by increased error patterns and by high levels of variability. These speech sound deficits have been under-studied and are not explicitly tied to accounts of SLI. In the present study, theoretical approaches drawn from dynamical systems and sequence learning are used to address speech production learning in children with SLI. Standard approaches to sound accuracy and variability and articulatory variability are integrated with novel applications of network science to assess sound learning trajectories over time. The purpose of the present study was to examine how measures of accuracy and variability are related when assessing nonword production over three sessions. A networks approach is proposed that highlights quantitative and qualitative relationships of sound sequences. Results demonstrate that children with SLI are less accurate and more variable, yet there is a dissociation between these two indices. Examination of movement trajectories reveals that group differences in performance cannot be accounted for solely by articulatory ability. There is a strong correlation between segmental variability and the networks measures, and the information provided by this novel methodology demonstrates gaps in classic approaches to error analysis. Results suggest that children with SLI have difficulty with sound sequencing, and that network science may capture error patterns that classic approaches do not.
Goffman, Purdue University.
Off-Campus Purdue Users:
To access this dissertation, please log in to our