Gates, S.-S.; (2011) Speech processing in typical and atypical language development: using nonwords to map the way. Doctoral thesis , UCL (University College London).

Preview

PDF 1140259.pdf Download (2MB)

Abstract

Accurate differential diagnosis of specific language impairment (SLI) is essential to determine the optimum form and content of treatment. It is therefore important to address the cognitive processes underpinning SLI and to evaluate potential clinical markers. The experiments presented here were designed to investigate input and output phonological processing in typically developing (TD) children, children with SLI and children with SLI and a concomitant speech disorder (SLI+SSD). Participants carried out a battery of published assessments and three experimental tasks: nonword repetition (NWR), nonword discrimination (NWD) and categorical perception (CP). Each experimental task used the same nonword stimuli which had been created by manipulating the position and number of consonant clusters and reflected repetition errors previously observed in children with SLI and/or dyslexia (Marshall & van der Lely, 2009). The results showed that NWD and NWR were highly correlated in TD children, implicating the same phonological processes in performance accuracy. Furthermore, none of the tasks was related to published linguistic assessments, including measures of receptive vocabulary and phonological short-term memory. It seemed that the experimental tasks tapped phonological representations at an unspecified stage of sub-lexical speech processing. Subtle processing differences were found between the clinical groups. Children with SLI showed deficits in NWR and NWD but not CP. Children with SLI+SSD were as impaired as SLIs on NWD, but they showed a NWR deficit of greater magnitude and were additionally impaired on CP. It was proposed that the SLI deficit was related to the length and complexity of the nonword stimuli and that a bidirectional transfer of information between sub-lexical output and input phonological representations may explain SLI+SSD performance. Sentence recall was the most reliable marker of SLI. The clinical implications of the results were discussed and it was concluded that caution should be exercised when administering or interpreting NWR tasks.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as CSVJSONXML

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item