Doctoral Dissertations

Date of Award

8-1998

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Speech and Hearing Science

Major Professor

A. K. Ananthanarayan, Mark Hedrick

Committee Members

Jim Thelin, Samuel Burchfield, Mary Sue Younger

Abstract

The effects of varying the onset frequency of the second formant and the overall intensity level of synthetic speech on the identification of stop consonant place of articulation and the neural representation of second formant transitions were compared in normal and hearing-impaired listeners. Comparisons of the phonetic boundary and slope of group identification functions for each stop consonant were made at each intensity level. Phonetic boundaries for normal and hearing-impaired listeners were relatively similar for each stop consonant at each intensity level. Slopes of the /d/ and /g/ identification functions for normal and hearing-impaired listeners were similar at each intensity level, however, slopes of the identification of /b/ functions were significantly steeper for the normal group at each intensity level except 72 dB SPL.

Comparisons of stop consonant identification at each intensity level were also made for the stimuli which corresponded to the highest percentage identification of each stop consonant. Results indicated significantly lower identification of /b/ and /g/ for hearing-impaired listeners at each intensity level and significantly lower identification of /d/ for hearing-impaired listeners at higher intensity levels.

The neural representation of the stimuli which corresponded to the highest percentage identification of each stop consonant was evaluated for each subject using the frequency-following response. Results demonstrated that the frequency-following response was sufficiently dynamic to encode time-varying formant transitions in normal listeners and in some listeners with hearing-impairment. Furthermore, comparisons of identification and neural representation results indicated that the identification of stop consonant place of articulation was directly related to the neural encoding of the time variant second formant transition reflected in the human frequency-following response.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS