Suprathreshold auditory processing and speech perception in noise: Hearing-impaired and normal-hearing listeners

Van Summers, Matthew J. Makashay, Sarah Theodoroff, Marjorie R. Leek

Research output: Contribution to journalArticle

27 Citations (Scopus)

Abstract

Background: It is widely believed that suprathreshold distortions in auditory processing contribute to the speech recognition deficits experienced by hearing-impaired (HI) listeners in noise. Damage to outer hair cells and attendant reductions in peripheral compression and frequency selectivity may contribute to these deficits. In addition, reduced access to temporal fine structure (TFS) information in the speech waveform may play a role. Purpose: To examine how measures of peripheral compression, frequency selectivity, and TFS sensitivity relate to speech recognition performance by HI listeners. To determine whether distortions in processing reflected by these psychoacoustic measures are more closely associated with speech deficits in steady-state or modulated noise. Research Design: Normal-hearing (NH) and HI listeners were tested on tasks examining frequency selectivity (notched-noise task), peripheral compression (temporal masking curve task), and sensitivity to TFS information (frequency modulation [FM] detection task) in the presence of random amplitude modulation. Performance was tested at 500, 1000, 2000, and 4000 Hz at several presentation levels. The same listeners were tested on sentence recognition in steady-state and modulated noise at several signal-to-noise ratios. Study Sample: Ten NH and 18 HI listeners were tested. NH listeners ranged in age from 36 to 80 yr (M = 57.6). For HI listeners, ages ranged from 58 to 87 yr (M = 71.8). Results: Scores on the FM detection task at 1 and 2 kHz were significantly correlated with speech scores in both noise conditions. Frequency selectivity and compression measures were not as clearly associated with speech performance. Speech Intelligibility Index (SII) analyses indicated only small differences in speech audibility across subjects for each signal-to-noise ratio (SNR) condition that would predict differences in speech scores no greater than 10% at a given SNR. Actual speech scores varied by as much as 80% across subjects. Conclusions: The results suggest that distorted processing of audible speech cues was a primary factor accounting for differences in speech scores across subjects and that reduced ability to use TFS cues may be an important component of this distortion. The influence of TFS cues on speech scores was comparable in steady-state and modulated noise. Speech recognition was not related to audibility, represented by the SII, once high-frequency sensitivity differences across subjects (beginning at 5 kHz) were removed statistically. This might indicate that high-frequency hearing loss is associated with distortions in processing in lower-frequency regions.

Original languageEnglish (US)
Pages (from-to)274-292
Number of pages19
JournalJournal of the American Academy of Audiology
Volume24
Issue number4
DOIs
StatePublished - Apr 2013
Externally publishedYes

Fingerprint

Auditory Perception
Speech Perception
Hearing
Noise
Signal-To-Noise Ratio
Speech Intelligibility
Cues
Outer Auditory Hair Cells
Psychoacoustics
High-Frequency Hearing Loss
Aptitude

Keywords

  • Compression
  • Frequency selectivity
  • Hearing impaired
  • Hearing science
  • Sensorineural hearing loss
  • Speech perception
  • Temporal fine structure sensitivity

ASJC Scopus subject areas

  • Speech and Hearing

Cite this

Suprathreshold auditory processing and speech perception in noise : Hearing-impaired and normal-hearing listeners. / Summers, Van; Makashay, Matthew J.; Theodoroff, Sarah; Leek, Marjorie R.

In: Journal of the American Academy of Audiology, Vol. 24, No. 4, 04.2013, p. 274-292.

Research output: Contribution to journalArticle

@article{0aab634a615c44698c525873f8c887b9,
title = "Suprathreshold auditory processing and speech perception in noise: Hearing-impaired and normal-hearing listeners",
abstract = "Background: It is widely believed that suprathreshold distortions in auditory processing contribute to the speech recognition deficits experienced by hearing-impaired (HI) listeners in noise. Damage to outer hair cells and attendant reductions in peripheral compression and frequency selectivity may contribute to these deficits. In addition, reduced access to temporal fine structure (TFS) information in the speech waveform may play a role. Purpose: To examine how measures of peripheral compression, frequency selectivity, and TFS sensitivity relate to speech recognition performance by HI listeners. To determine whether distortions in processing reflected by these psychoacoustic measures are more closely associated with speech deficits in steady-state or modulated noise. Research Design: Normal-hearing (NH) and HI listeners were tested on tasks examining frequency selectivity (notched-noise task), peripheral compression (temporal masking curve task), and sensitivity to TFS information (frequency modulation [FM] detection task) in the presence of random amplitude modulation. Performance was tested at 500, 1000, 2000, and 4000 Hz at several presentation levels. The same listeners were tested on sentence recognition in steady-state and modulated noise at several signal-to-noise ratios. Study Sample: Ten NH and 18 HI listeners were tested. NH listeners ranged in age from 36 to 80 yr (M = 57.6). For HI listeners, ages ranged from 58 to 87 yr (M = 71.8). Results: Scores on the FM detection task at 1 and 2 kHz were significantly correlated with speech scores in both noise conditions. Frequency selectivity and compression measures were not as clearly associated with speech performance. Speech Intelligibility Index (SII) analyses indicated only small differences in speech audibility across subjects for each signal-to-noise ratio (SNR) condition that would predict differences in speech scores no greater than 10{\%} at a given SNR. Actual speech scores varied by as much as 80{\%} across subjects. Conclusions: The results suggest that distorted processing of audible speech cues was a primary factor accounting for differences in speech scores across subjects and that reduced ability to use TFS cues may be an important component of this distortion. The influence of TFS cues on speech scores was comparable in steady-state and modulated noise. Speech recognition was not related to audibility, represented by the SII, once high-frequency sensitivity differences across subjects (beginning at 5 kHz) were removed statistically. This might indicate that high-frequency hearing loss is associated with distortions in processing in lower-frequency regions.",
keywords = "Compression, Frequency selectivity, Hearing impaired, Hearing science, Sensorineural hearing loss, Speech perception, Temporal fine structure sensitivity",
author = "Van Summers and Makashay, {Matthew J.} and Sarah Theodoroff and Leek, {Marjorie R.}",
year = "2013",
month = "4",
doi = "10.3766/jaaa.24.4.4",
language = "English (US)",
volume = "24",
pages = "274--292",
journal = "Journal of the American Academy of Audiology",
issn = "1050-0545",
publisher = "American Academy of Audiology",
number = "4",

}

TY - JOUR

T1 - Suprathreshold auditory processing and speech perception in noise

T2 - Hearing-impaired and normal-hearing listeners

AU - Summers, Van

AU - Makashay, Matthew J.

AU - Theodoroff, Sarah

AU - Leek, Marjorie R.

PY - 2013/4

Y1 - 2013/4

N2 - Background: It is widely believed that suprathreshold distortions in auditory processing contribute to the speech recognition deficits experienced by hearing-impaired (HI) listeners in noise. Damage to outer hair cells and attendant reductions in peripheral compression and frequency selectivity may contribute to these deficits. In addition, reduced access to temporal fine structure (TFS) information in the speech waveform may play a role. Purpose: To examine how measures of peripheral compression, frequency selectivity, and TFS sensitivity relate to speech recognition performance by HI listeners. To determine whether distortions in processing reflected by these psychoacoustic measures are more closely associated with speech deficits in steady-state or modulated noise. Research Design: Normal-hearing (NH) and HI listeners were tested on tasks examining frequency selectivity (notched-noise task), peripheral compression (temporal masking curve task), and sensitivity to TFS information (frequency modulation [FM] detection task) in the presence of random amplitude modulation. Performance was tested at 500, 1000, 2000, and 4000 Hz at several presentation levels. The same listeners were tested on sentence recognition in steady-state and modulated noise at several signal-to-noise ratios. Study Sample: Ten NH and 18 HI listeners were tested. NH listeners ranged in age from 36 to 80 yr (M = 57.6). For HI listeners, ages ranged from 58 to 87 yr (M = 71.8). Results: Scores on the FM detection task at 1 and 2 kHz were significantly correlated with speech scores in both noise conditions. Frequency selectivity and compression measures were not as clearly associated with speech performance. Speech Intelligibility Index (SII) analyses indicated only small differences in speech audibility across subjects for each signal-to-noise ratio (SNR) condition that would predict differences in speech scores no greater than 10% at a given SNR. Actual speech scores varied by as much as 80% across subjects. Conclusions: The results suggest that distorted processing of audible speech cues was a primary factor accounting for differences in speech scores across subjects and that reduced ability to use TFS cues may be an important component of this distortion. The influence of TFS cues on speech scores was comparable in steady-state and modulated noise. Speech recognition was not related to audibility, represented by the SII, once high-frequency sensitivity differences across subjects (beginning at 5 kHz) were removed statistically. This might indicate that high-frequency hearing loss is associated with distortions in processing in lower-frequency regions.

AB - Background: It is widely believed that suprathreshold distortions in auditory processing contribute to the speech recognition deficits experienced by hearing-impaired (HI) listeners in noise. Damage to outer hair cells and attendant reductions in peripheral compression and frequency selectivity may contribute to these deficits. In addition, reduced access to temporal fine structure (TFS) information in the speech waveform may play a role. Purpose: To examine how measures of peripheral compression, frequency selectivity, and TFS sensitivity relate to speech recognition performance by HI listeners. To determine whether distortions in processing reflected by these psychoacoustic measures are more closely associated with speech deficits in steady-state or modulated noise. Research Design: Normal-hearing (NH) and HI listeners were tested on tasks examining frequency selectivity (notched-noise task), peripheral compression (temporal masking curve task), and sensitivity to TFS information (frequency modulation [FM] detection task) in the presence of random amplitude modulation. Performance was tested at 500, 1000, 2000, and 4000 Hz at several presentation levels. The same listeners were tested on sentence recognition in steady-state and modulated noise at several signal-to-noise ratios. Study Sample: Ten NH and 18 HI listeners were tested. NH listeners ranged in age from 36 to 80 yr (M = 57.6). For HI listeners, ages ranged from 58 to 87 yr (M = 71.8). Results: Scores on the FM detection task at 1 and 2 kHz were significantly correlated with speech scores in both noise conditions. Frequency selectivity and compression measures were not as clearly associated with speech performance. Speech Intelligibility Index (SII) analyses indicated only small differences in speech audibility across subjects for each signal-to-noise ratio (SNR) condition that would predict differences in speech scores no greater than 10% at a given SNR. Actual speech scores varied by as much as 80% across subjects. Conclusions: The results suggest that distorted processing of audible speech cues was a primary factor accounting for differences in speech scores across subjects and that reduced ability to use TFS cues may be an important component of this distortion. The influence of TFS cues on speech scores was comparable in steady-state and modulated noise. Speech recognition was not related to audibility, represented by the SII, once high-frequency sensitivity differences across subjects (beginning at 5 kHz) were removed statistically. This might indicate that high-frequency hearing loss is associated with distortions in processing in lower-frequency regions.

KW - Compression

KW - Frequency selectivity

KW - Hearing impaired

KW - Hearing science

KW - Sensorineural hearing loss

KW - Speech perception

KW - Temporal fine structure sensitivity

UR - http://www.scopus.com/inward/record.url?scp=84878309925&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84878309925&partnerID=8YFLogxK

U2 - 10.3766/jaaa.24.4.4

DO - 10.3766/jaaa.24.4.4

M3 - Article

C2 - 23636209

AN - SCOPUS:84878309925

VL - 24

SP - 274

EP - 292

JO - Journal of the American Academy of Audiology

JF - Journal of the American Academy of Audiology

SN - 1050-0545

IS - 4

ER -