Modulation rate discrimination by normal-hearing and hearing-impaired listeners

Abstract

Sensorineural hearing loss has been consistently associated with deficits in frequency analysis in addition to the loss of sensitivity. However, temporal processing may be relatively normal in many hearing-impaired (HI) individuals. For these listeners it might be possible to improve speech recognition by coding certain dynamic spectral cues (e.g., variations in voice fundamental frequency, formant frequency transitions, etc.) as temporal modulations of simple carrier waves. For this signal-processing strategy to be successful, listeners must be able to discriminate changes in the rate of temporal modulations. In the present study, modulation detection and rate discrimination were examined at three different standard modulation rates (80, 160, and 320 Hz) and at various modulation depths for both normally hearing (NH) and HI subjects. Measures of pure-tone frequency discrimination were also obtained at these same test frequencies. For all subjects, modulation rate discrimination thresholds increased with increases in standard rate and decreases in depth of modulation. For HI subjects, modulation rate discrimination thresholds were approximately equal to pure-tone frequency discrimination thresholds at 80 Hz but were much worse at higher rates. These results suggest that for these subjects, transforming dynamic spectral speech cues into dynamic temporal cues would be of little benefit. [Work supported by NIH.]