||Stabilité dans le temps des déficits en et hors lecture chez des adolescents dyslexiques (données longitudinales)
The goals of the present study are twofold: (1) to assess dyslexics’ (DYS) reading and reading-related skills compared to Chronological Age Controls (CAC); and (2) to examine the predictors of their reading level from skills assessed at age 8 and 17. At age 17, reading skills have been assessed through high frequency word and pseudoword reading, the items in the 2nd category being supposed to be accurately processed through the phonological reading route, assumed to be deficient in DYS. In addition, some phonological reading-related skills, also assumed to be deficient in DYS, were assessed (phonemic segmentation, phonological short-term memory [STM] and Rapid Automatic Naming [RAN]), as well as some visual skills (visual STM and visual attention). Except for STM, accuracy and speed have been taken into account. As expected, the DYS were found to have lower scores than the CAC in reading, both for accuracy and latency (the delay between the display of a word on the computer and the beginning of the vocal response), their deficit being more noticeable on pseudowords than on words. The results observed for latency cannot be due to the articulatory component of the task, given that there was no difference between the groups for vocal response duration. The DYS were also found to have lower scores on tests involving phonological reading-related skills, but not on tests involving visual skills. In addition, phonological reading-related skills explained 63% of the variance in reading skills at age 17 but only processing speed in the RAN and in the phonemic segmentation tasks added a unique and significant part of the variance in reading skills. The skills assessed at age 8 (reading level, verbal and non-verbal IQ, syllabic and phonemic segmentation, phonological STM) explained 60% of the variance in reading skills at age 17 but only the reading level added a unique and significant part of the explained variance (27%). These results, as those of other longitudinal studies (6, 11, 19, 28, 29, 31), indicate (1) the persistence over time of reading deficits; (2) that phonological skills largely explain reading success or failure; (3) the importance of processing time, included for the assessment of phonological reading-related skills.