Speech Perception/Psychoacoustics

NIDCD funded R03 project: Spectral Cue Use in Older Listeners with Poor Speech Perception

Background:

Among those older listeners who have a hearing loss, most have good speech perception once the issue of audibility is addressed via amplification. However, there is a subgroup of older listeners with hearing loss who have an inordinate amount of difficulty understanding even amplified speech.

My previous research indicated that these listeners have significantly more difficulty with spectral resolution if the signal is complex. Performance on the NST suggested that these listeners might have difficulty with voiced and initial consonants in addition to the typical unvoiced and final speech sounds.

Experiment I:

• Hypothesis: Older listeners with poor speech perception abilities not only make more of the same kinds of errors as older listeners with good speech perception, they make different kinds of errors.
• Methods: 12 older listeners (aged 60-80 years) in three groups (normal hearing, hearing-impaired with good Word Recognition Scores, and hearing-impaired with poor WRS) identified the naturally spoken CV in an 8-choice task. The 8 consonants used were /b/, /d/, /g/, /m/, /n/, /δ/, /v/, and /z/. Three vowel contexts were used: /?/, /o/, and /u/. The CVs were recorded from a male and female speaker.
• Phillips, S.L., Richter, S.J. & McPherson, D. (2009). Voiced initial consonant perception deficits in older listeners with hearing loss and good and poor word recognition. Journal of Speech, Language & Hearing Research, 52(1), 118-129.