ANALYSIS OF PERCEPTUAL CONFUSIONS BETWEEN 9 SETS OF CONSONANT-VOWEL SOUNDS IN NORMAL AND DYSLEXIC ADULTS

It is widely accepted that most developmental dyslexics perform poorly on tasks which assess phonological awareness. One reason for this ass ociation might be that the early or ''input'' phonological representat ions of speech sounds are distorted or noisy in some way. We have atte mpted to test this hypothesis directly. In Experiment 1, we measured t he confusions that adult dyslexics and controls made when they listene d to nine randomly presented consonant-vowel (CV) segments (/ba/, /da/ , /pa/, /t integral a/, /fa/, /integral a/, /la/, /wa/ & /ja/) under f our conditions of increasing white noise masking. Subjects could repla y stimuli and were under no obligation to respond quickly. Responses w ere selected with a computer mouse from a set of nine letter-strings, corresponding to the auditory stimuli, presented on a VDU. While the o verall pattern of confusions made by dyslexics and controls was very s imilar for this stimulus set, dyslexics confused /t integral a/ with / integral a/ and /pa/ with /fa/ significantly more than did controls. I n Experiment 2, subjects heard each stimulus once only and were forced to respond as quickly as possible. Under these timed conditions, the pattern of confusions made by dyslexics and controls was the same as b efore, but dyslexics took longer to respond than controls. The slower responses of dyslexics in Experiment 2 could have arisen because: (a) they were slower at processing the auditory stimuli than controls, (b) they had worse visual pattern memory for letter strings than controls , (c) they were slower than controls at using the computer mouse. In E xperiments 3, 4 and 5 subjects carried out control tasks which elimina ted each of these possibilities and confirmed that the results from th e auditory tasks genuinely reflected subjects' speech perception. We p ropose that the fine structure of dyslexics' input phonological repres entations should be further explored with this confusion paradigm by u sing other speech sounds containing VCs, CCVs and VCCs.