CLARIFYING THE SPEECH PERCEPTION DEFICIT IN DYSLEXIC CHILDREN Souhila Messaoud-Galusi, Valerie Hazan, Stuart Rosen Department of Phonetics and Linguistics, UCL, UK souhila@phon.ucl.ac.uk, v.hazan@ucl.ac.uk, stuart@phon.ucl.ac.uk ABSTRACT It has often been claimed that dyslexic children show deficits in various speech-perceptual tasks. In this study, dyslexic and chronological-age-matched control children were asked to identify words, and label monosyllables from a voiced/voiceless plosive continuum, in quiet and in noise. Correlations on these tasks with reading and reading-related skills were weak and about half of dyslexic children had categorization slopes within the normal range in quiet. Both reading groups performed similarly well for labeling in noise and when identifying words in noise. The identification of words in noise was found to be related neither to reading nor to the consistency of categorical labeling. This study confirms that only a subgroup of children with dyslexia appears to have speech- perceptual deficits. Keywords: Dyslexia, Categorical labeling, Speech Perception, Noise. 1. INTRODUCTION Dyslexia is defined as an unexpected difficulty in acquiring literacy despite average intelligence and appropriate learning opportunities. Given the large amount of evidence showing difficulties in processing phonological information [1], many investigators agree that poorly specified phonological representations are a key deficit in dyslexia [2]. Research on categorical perception has been conducted in dyslexic populations in order to determine if fine-grained speech perception abilities could be at the origin of the phonological deficit characterizing dyslexia. Various studies suggest that identification of items from an acoustic continuum representing a phonemic contrast (e.g. in voicing or place of articulation) differs between dyslexic and typically-developing listeners. Identification functions in dyslexic children were typically found to be shallower than those of chronological age controls [3, 4]; this indicates that phonemic categorization in dyslexic children is less consistent than average readers. Nevertheless, some studies failed to observe a systematic deficit in categorical labeling in dyslexic children [5, 6]. Adlard and Hazan [5] observed that despite significant group differences, only a subgroup of the dyslexic children were identifying the contrast less consistently than reading level matched controls controls. The present study evaluated categorical labeling in dyslexic and average reading children matched in age. A sizeable enough group of well- characterized dyslexics was recruited in order to identify the proportion of dyslexics performing outside the average reader range. The first aim of the study was to determine if categorical labeling deficits are restricted to a subgroup of dyslexic children, consistent with the findings of Adlard and Hazan [5]. Most communication situations take place with background noise, but little data are available on speech perception abilities of dyslexic listeners in noise. Therefore, we assessed categorical labeling and identification of words with noise in the background. If dyslexic children have poorly defined phonemic categories, they may be more affected by noise than average readers. 2. METHOD 2.1. Population 53 monolingual English native speakers aged between 7.11 and 13.06 were recruited. In order to be included in the study, participants were required to score within a standard deviation of the standardized mean in of receptive grammar [7], verbal [8] and non-verbal IQ [9]. 33 of the participants received a formal diagnosis of dyslexia by an educational psychologist. The 20 remaining participants (the controls) never experienced reading difficulty nor showed any reading delay in the tests we applied. ICPhS XVI ID 1457 Saarbrücken, 6-10 August 2007 www.icphs2007.de 697