TY - THES ID - 136964402 TI - THE ROLE OF STATISTICAL LEARNING IN 9-YEAR OLD CHILDREN WITH DYSLEXIA AU - Snip, Annoek AU - Ghesquière, Pol AU - Vandermosten, Maaike AU - KU Leuven. Faculteit Psychologie en Pedagogische Wetenschappen. Opleiding Master in de psychologie (Leuven) PY - 2018 PB - Leuven KU Leuven. Faculteit Psychologie en Pedagogische Wetenschappen DB - UniCat UR - https://www.unicat.be/uniCat?func=search&query=sysid:136964402 AB - When learning how to read, speech sounds need to be connected to letters, hence requiring good and well specified representations of speech sounds, and the ability to manipulate them. One of the predominant cognitive explanations of developmental dyslexia (DD) is that dyslexic readers display phonological problems that originate from underlying not-well specified phonological representations. This present study strives to partly decode to what extent statistical learning of speech sounds, i.e. the way we form phoneme categories by exposure to different distributions of phonetic units in the mother language, underlies the formation of phoneme representations in 19 dyslectic reading (DR) and 19 normal reading (NR) children. We examined if statistical learning of non-native phonemes took place, and if DR children displayed less of these learning derived developments after bimodal learning. Several reading related measures and a categorical perception task (CPT) of native sounds were assessed. In addition, to establish if children would learn to create phoneme categories based on statistical cues, they performed a CPT of non-native sounds before and after being exposed to a flow of non-native speech sounds according to a bimodal distribution. Findings revealed a group difference on the native CPT (with DR children performing worse than NR children) and DR children did not show any improvement in contrast to NR children on the post non-native CPT. These results may indicate that differential statistical learning underlies the impaired phonological representations that are observed in DR children. At the same time, the DR children did not perform worse than NR children on a phonemic awareness task, which was a somwhat unlikely finding that might indicate that our DR sample might not have been completely representative. Recommendations and implications for future studies are made. Keywords: developmental dyslexia, categorical perception, phoneme representations, statistical learning ER -