Listener profiles for dysarthric and accented speech

Abstract

In order to determine which aspects of speech contribute most to speech intelligibility, listeners rated several speech features for sentences produced by speakers with dysarthria and for passages read aloud by speakers of English as a second language. For the dysarthric speech, speech components included rate, stress, intonation, articulation, voice quality, and nasality in addition to speech intelligibility. For the accented English passages, the speech components were rate, stress, intonation, articulation, and other aspects of speech as well as overall intelligibility, accentedness, and language competence. For both types of speech, articulation was the best predictor of intelligibility ratings. Voice quality, rate, and nasality made minor contributions to intelligibility ratings for dysarthric speech. Rate also contributed slightly to intelligibility ratings for the accented English passages. Potential differences in the salience of speech components for listeners will be discussed. Use of the perceptual profile approach for planning treatment and assessing treatment efficacy for dysarthric and accented speech will also be addressed.