Second Biennial Hearing Aid Research and Development Conference
September 22-24, 1997
National Institutes of Health
Effects of Flat and Frequency-shaped Amplification on Neural Measures of Vowel Discriminability in Hearing-impaired Cats
B.M. Calhoun, E.D. Young, Center for Hearing Sciences, The Johns Hopkins School of Medicine, Baltimore, Maryland
We have studied the representation of vowels in cat auditory-nerve fibers after inducing sensorineural hearing loss in the 2 kHz region, in order to investigate the neurophysiological basis of diminished speech recognition in human listeners with hearing loss. Here, we describe the discriminability of four /eh/-like steady-state vowels, created with a Klatt synthesizer in which the second formant was varied in frequency (F2 = 1.4, 1.5, 1.7, 2.0 kHz). Placement of the first and third formant was left unchanged (F1 = 0.5 kHz, F3 = 2.5 kHz), making representation of F2 the dominant feature for discriminating between these stimuli. Normal fibers show a strong phase-locked response to F2 (synchrony capture) and a clear rate-place dif-ference for these vowels, when presented at 50 or 70 dB SPL (JASA 98, 3223-34).
Hearing loss was induced by a 4 hour exposure to a 50 Hz noiseband presented at 103 or 108 dB SPL, centered at 2 kHz, and allowing a minimum recovery period of 54 days. Fibers in the F2 region had broad tuning curves and showed a minimum threshold shift of about 50 dB, consistent with OHC damage. Audibility, as indicated by an increase in response rate, was restored to the damaged fibers by simple amplification, i.e. presenting the vowel at levels above 90 dB SPL. However, a normal temporal pattern of response was not present in the damaged cochlea, in that synchrony capture by F2 was not present. The broader tuning of damaged fibers increased sensitivity to spectral features below BF and allowed multiple harmonics, including F1, to drive fibers which normally encode F2 features. As a result of this impaired response pattern, hearing-impaired animals do not show a rate-place difference between these vowels.
We also tested contrast-enhancing frequency-shaped (CEFS) amplification which has a gain profile governed by the frequency profile of the hearing loss and each vowel. Conventional frequency-shaped amplification, in which the frequency receiving greater amplification is based upon hearing loss alone, adds power to the spectral trough between F1 and F2 and reduces spectral contrast in these vowels. As a result, fibers with BFs between F1 and F2 synchronize to their own BF, reducing the spectral contrast between F1 and F2 in the temporal-place code. In CEFS, the F2-F3 portion of each vowel received 30dB more gain than the F1 portion and the energy trough between F1 and F2 was preserved. This paradigm provided the highest degree of phase-locking to F2, as well some restoration of the rate-place difference between the vowels, in the impaired animals. These results demonstrate the merit of hearing aid processing schemes which increase the spectral contrast of the speech spectrum.
[Supported by NIDCD Grant]
Return to Conference Abstracts List