Early Cognitive Indicators of Dyslexia in Preliterate Children at Genetic Risk

Abstract

The aim of this thesis was to assess a sample of 81 preliterate children (3 to 5 years old) who were from a Family At Risk of dyslexia (FAR) or Not Easily Associated with Risk (NEAR) in respect to a range of working memory and standardised language tasks (children's measures) in an attempt to identify cognitive features that might be implicated in dyslexia. Because dyslexia appears to have hereditary components, parent's performance on a range of reading related tasks was also correlated with children's performance to assess which tasks were most predictive of risk status. Utilising information gained by previous longitudinal studies, standardized tasks provided a continuous classification of the children's likely risk status based on their own performance. Three factors were gleaned from a factorial analysis, component manipulation ability, rapid automatised naming, and letter knowledge, and this accords with current research (Bishop, 2003) suggesting that these three factors are good predictors of future reading success. In terms of the children's non-standardised measures, the ability to pair visual stimuli with articulated and synthesised sounds was investigated. Measures of central executive functions (cognitive inhibition, susceptibility to interference, various span measures, speed, and controlled attention) were also examined. Tasks were adapted, or developed, specifically to meet the attentional, knowledge and skill levels of these young children. Children were categorised in terms of risk status using two methods based upon their parent's literacy performance. Firstly, several continuous predictors were identified. Of these, the worst parent's comprehension score was the best predictor of children's performance and supports the idea that the hereditary component of dyslexia appears more likely to involve central executive functions, like speed, and possibly central coherence, which in turn impact on phonological processing. The second method produced the discrete categories of family at risk (FAR) and not easily associated with risk (NEAR). This method proved to be most informative when categorisation was based on self-report by the parents concerning their past competence in reading.