Limiting spectral resolution in speech for listeners with sensorineural hearing loss

Area of Science:

• Auditory perception
• Speech processing
• Sensorineural hearing loss

Background:

• Previous research indicated similar consonant recognition between normal-hearing and hearing-impaired listeners when speech had one spectral channel.
• Hearing-impaired listeners are hypothesized to perform similarly to normal-hearing listeners when spectral resolution is limited.

Purpose of the Study:

• To investigate consonant recognition in listeners with moderate sensorineural hearing loss compared to normal-hearing listeners.
• To examine the effect of varying spectral resolution on speech perception in both hearing groups.

Main Methods:

• Consonant recognition was measured using speech stimuli processed into 1, 2, 4, or 8 spectral channels.
• Participants included normally hearing listeners and listeners with moderate sensorineural hearing loss.

Main Results:

• Both groups performed similarly with 1-channel spectral resolution.
• Hearing-impaired listeners performed significantly worse than normal-hearing listeners with 2, 4, and 8 channels of spectral resolution.
• Deficient performance was observed in hearing-impaired listeners when combining information across multiple spectral channels.

Conclusions:

• The hypothesis that hearing-impaired listeners would perform similarly across limited spectral channels was not supported.
• Listeners with moderate sensorineural hearing loss demonstrate deficits in processing speech with multiple spectral channels.
• While temporal-envelope information in a single channel is utilized normally, multi-channel integration is impaired in hearing loss.