Phonetic variation in dysarthric speech as a function of sampling task.

Abstract

Previous studies have demonstrated a number of ways in which normal speakers' phonetic performance varies across reading and spontaneous speech tasks. This study set out to investigate whether similar differences across speech sampling tasks were found in a mixed group of dysarthric subjects. A selection of segmental and prosodic parameters were investigated acoustically in the performance of five mild dysarthric speakers and five matched control subjects. The results demonstrated that breath–pause position, unstressed vowel duration and voice-onset time were subject to variation across sampling task in the speech produced by different types of dysarthric speaker. The results suggest that read material produced by dysarthric speakers may not be wholly representative of those speakers' spontaneous speech. Preliminary implications for clinical practice are discussed. The findings point to the need for further research to investigate the extent of such differences and their implications for dysarthria assessment, which up to the present has relied predominantly on read material