1. Neuroscience

Increased listening effort and cochlear neural degeneration underlie speech-in-noise deficits in normal-hearing middle-aged adults

  1. Maggie E Zink
  2. Leslie Zhen
  3. Jacie R McHaney
  4. Jennifer Klara
  5. Kimberly Yurasits
  6. Victoria E Cancel
  7. Olivia Flemm
  8. Claire Mitchell
  9. Jyotishka Datta
  10. Bharath Chandresekaran
  11. Aravindakshan Parthasarathy  Is a corresponding author
  1. Department of Communication Science and Disorders, School of Health and Rehabilitation Sciences, University of Pittsburgh, United States
  2. Department of Statistics, Virginia Polytechnic Institute and State University, United States
  3. Department of Bioengineering, Swanson School of Engineering, University of Pittsburgh, United States
  4. Department of Otolaryngology, School of Medicine, University of Pittsburgh, United States
https://doi.org/10.7554/eLife.102823.3
Share this article
Cite this article
  1. Maggie E Zink
  2. Leslie Zhen
  3. Jacie R McHaney
  4. Jennifer Klara
  5. Kimberly Yurasits
  6. Victoria E Cancel
  7. Olivia Flemm
  8. Claire Mitchell
  9. Jyotishka Datta
  10. Bharath Chandresekaran
  11. Aravindakshan Parthasarathy
(2025)
Increased listening effort and cochlear neural degeneration underlie speech-in-noise deficits in normal-hearing middle-aged adults
eLife 13:RP102823.
https://doi.org/10.7554/eLife.102823.3
  2. Download BibTeX
  3. Download .RIS
6 figures, 11 tables and 1 additional file

Figures

Figure 1
Download asset Open asset
Age-related cochlear neural degeneration (CND) occurs prior to overt changes in hearing thresholds and can be assessed noninvasively by measuring phase-locked neural envelope following responses.

(A) Thirty seven middle-aged (MA, 40–55 years, mean = 46.1 ± 4.6 years) and 35 young adults (YA, 18–25 years, mean = 21.17 ± 1.8 years) participated in this study. (B) All participants had clinically normal-hearing thresholds, with some evidence of threshold losses at extended high frequencies above 8 kHz typically not tested in the clinic. Hearing thresholds in dB HL are shown on the Y-axis and frequency in kHz is plotted on the X-axis. (C) Outer hair cell function assessed using distortion product otoacoustic emissions (DPOAEs) is comparable between YA and MA up to 4 kHz and showed age-related decreases at higher frequencies. Both cohorts show no evidence of self-reported tinnitus (D) or hyperacusis measured as loudness discomfort levels (LDLs) (E), have comparable self-reported noise exposure levels (F), and present with comparable working memory scores assessed using operation span task (OSPAN) (G). (H) Envelope following responses (EFRs) to modulation frequencies of 1024 Hz can be reliably recorded in YA and MA using ‘tiptrodes’. The panel shows grand-averaged fast Fourier transform (FFT) traces for YA and MA. (I) MA showed significant declines in EFR amplitudes at 1024 Hz amplitude modulation (AM), with putative neural generators in the auditory nerve. (J) Signal-to-noise ratios were 8 dB on average for YA and 4 dB for MA. (K) Statistically significant decreases in EFR amplitudes were selective for 1024 Hz AM, the modulation frequency with putative generators in the auditory nerve. All panels: Error bars and shading represent standard error of the mean (SEM). Asterisks represent p<0.05, analysis of variance (ANOVA).

Figure 2
Download asset Open asset
Cross-species experiments in a rodent model show that envelope following responses (EFRs) are a sensitive biomarker for histologically-confirmed cochlear neural degeneration (CND).

(A) Cross-species comparisons were made with young (22±0.86 weeks, n=14) and middle-aged (80±0.76 weeks, n=12) Mongolian gerbils, with identical stimuli, recording, and analysis parameters. (B) Middle-aged gerbils did not show any age-related decreases in hearing thresholds. (C) Age-related decreases in EFR amplitudes were isolated to the 1024 Hz modulation frequency, similar to middle-aged humans in Figure 1K. (D) CND was quantified using immunostained organ of Corti whole mounts, where afferent excitatory synapses were quantified using 3D reconstructed images. (E) Cochlear synapse counts at the 3 kHz cochlear region corresponding to the carrier frequency for the EFRs were significantly decreased in middle-aged gerbils, despite matched auditory thresholds. (F) EFR amplitudes at 1024 Hz amplitude modulation (AM) were significantly correlated with the number of remaining cochlear synapses, suggesting that these EFRs are a sensitive metric for CND with age. All panels: Error bars and shading represent standard error of the mean (SEM). Asterisks represent p<0.05, analysis of variance (ANOVA).

Figure 3
Download asset Open asset
Increased listening effort precedes behavioral deficits in speech-in-noise perception in middle-aged adults.

(A) Speech perception in noise was assessed using the Quick Speech-in-Noise (QuickSIN) test, which presents moderate context sentences in varying levels of multi-talker babble. Pupillary measures were analyzed in two time-windows: (1) during stimulus presentation and (2) after target sentence offset and prior to response initiation. (B) No significant age-related differences were observed in clinical QuickSIN scores presented as dB signal-to-noise ratio (SNR) loss. (C) QuickSIN performance is matched between middle-aged (MA) and younger adults (YA) until the most difficult noise condition (SNR 0). The x-axis shows the SNR condition that the target sentences were presented in, with 25 dB being the easiest noise condition, and 0 dB being the most difficult noise condition. The y-axis shows participant accuracy in repeating keywords from the target sentences as percent correct. (D) Grand-averaged pupillary responses, measured during task listening as an index of effort, exhibit modulation with task difficulty, with greater pupillary dilations observed in harder conditions for both groups. (E) Middle-aged adults show consistently higher pupillary responses during performance on the QuickSIN task and at SNR levels prior to when overt behavioral deficits are observed. (F) Grand-averaged pupillary responses measured after target sentence offset as an index of effort exhibit greater modulation with task difficulty, compared to changes in the listening window. (G) Trends seen in the listening window were amplified in this integration window, with middle-aged adults showing even greater effort, especially at moderate SNRs where behavior was matched.

Figure 4
Download asset Open asset
Listening effort and cochlear neural degeneration (CND) provide complementary contributions to speech-in-noise intelligibility.

(A) Behavioral performance at the most challenging signal-to-noise ratio (SNR) was significantly correlated with the envelope following response (EFR) measures of CND, with lower EFR amplitudes being associated with poorer behavioral performance. (B) Pupillary responses at 10 dB SNR from the integration window were significantly correlated with behavioral performance at 0 dB SNR. (B) These correlations between pupillary responses at 10 dB SNR and behavioral performance at 0 dB SNR were also found in the listening window, even though there were no group differences in age, further strengthening the link between listening effort at moderate SNRs and behavioral performance at challenging SNRs. (D) An elastic net regression model with 10-fold cross-validation (cv) was fit to the Quick Speech-in-Noise (QuickSIN) scores at 0 dB SNR. The tuning parameter Lambda controls the extent to which coefficients contributing least to predictive accuracy are suppressed. (E) A lollipop plot displaying the coefficients (β) contributing to explaining variance on QuickSIN performance suggests that CND, listening effort, and subclinical changes in hearing thresholds all contribute to QuickSIN performance. (F) QuickSIN scores predicted by the elastic net regression are correlated with actual participant QuickSIN scores.

Author response image 1
Download asset Open asset
Cochlear synapse counts per inner hair cell (IHC) in young and middle-aged gerbils as a function of cochlear frequency.
Author response image 2
Download asset Open asset
Harmonics analysis for the first four harmonics of envelope following responses elicited to the 110Hz AM stimulus.

Tables

Table 1
Comparison of air conduction thresholds using a three-way analysis of variance (ANOVA) (middle-aged adults [MA] =37, young adults [YA] =35).
EffectsDFnSum SqMean SqF-valuep-Value
Frequency6506784420.786<0.001***
Ear1330.0680.8
Group158315813143.521<0.001***
Freq:Ear685140.3490.9
Freq:Group69051513.712<0.001***
Ear:Group1770.1640.7
Freq:Ear:Group6234390.9610.5
Residuals8403412541

  1. *p<0.05; ***p<0.001.

Table 2
Comparison of extended high frequencies using three-way analysis of variance (ANOVA) (middle-aged adults [MA] =37, young adults [YA] =35).
EffectsDFnSum SqMean SqF-valuep-Value
Frequency2212091060574.523<0.001***
Ear1660.0390.8
Group13286832868230.978<0.001***
Freq:Ear21421420.4980.6
Freq:Group26016601621.137<0.001***
Ear:Group11521521.0690.3
Freq:Ear:Group238190.1340.9
Residuals35049805142

  1. *p<0.05; ***p<0.001.

Table 3
Comparison of right ear distortion product otoacoustic emissions using a two-way analysis of variance (ANOVA) (middle-aged adults [MA] =34, young adults [YA] =31).
EffectsDFnDFdF-valuep-Value
Group16325.85<0.001***
Freq9.55601.5658.786<0.001***
Group:Freq9.55601.567.341<0.001***
501:Group1630.7131.00
595:Group1631.9391.00
707:Group1630.7181.00
841:Group1630.2681.00
998:Group1634.380.84
1188:Group1630.7941.00
1414:Group1634.670.74
1681:Group1631.7241.00
2000:Group1630.871.00
2378:Group1630.0591.00
2828:Group1634.7550.69
3365:Group1632.0951.00
4001:Group16310.4630.04*
4757:Group16318.015<0.001***
5658:Group16329.947<0.001***
6727:Group16337.01<0.001***
8000:Group16328.94<0.001***
9514:Group16339.235<0.001***
11314:Group16326.847<0.001***
13454:Group1637.7710.147
160000:Group1630.4361.00

  1. *p < .05; ***p < .001.

Table 4
Comparisons using one-way analyses of variance (ANOVAs).
MeasureYA (n)MA (n)DFnDFdF-valuep-Value
THI33371680.8340.364
OSPAN34341663.5010.066
QuickSIN Clinical Score31341633.2140.078
NEQ32321660.83750.363

  1. Adjusted p-values are reported using Bonferroni correction.

Table 5
Comparison of envelope following responses (EFRs) using two-way analyses of variance (ANOVAs) (middle-aged adults [MA] =29, young adults [YA] =28).
EffectsDFnDFdF-valuep-Value
Group1540.2750.6
AM1.4779.49151.407<0.001***
Group:AM1.4779.490.1510.929
1024:Group15523.8<0.001***
512:Group1563.1710.083
110:Group1550.4910.487
40:Group1540.0270.870

  1. *p<0.05; ***p<0.001.

Table 6
Comparison of 22-week-old gerbil (n=14) and 80-week-old gerbil (n=12) envelope following responses (EFRs) using two-way analyses of variance (ANOVAs).
EffectsDFnDFdF-valuep-Value
Group1244.1250.053
AM2.6864.2874.636<0.001***
Group:AM2.6864.280.8750.449
16:Group1240.4560.506
40:Group1242.4610.130
110:Group1243.0560.093
256:Group1241.9590.174
724:Group1242.4830.128
1024:Group1245.1580.032*

  1. *p < .05; ***p < .001.

Table 7
Comparison of synapse counts at 3000 Hz in 22- and 80-week-old gerbils using one-way analysis of variance (ANOVA).
Measure22 weeks (n)80 weeks (n)DFnDFdF-valuep-Value
Synapse counts14121164.8770.042*

  1. *p<0.05; ***p<0.001.

Table 8
Comparison of Quick Speech-in-Noise (QuickSIN) performance using a mixed-design analysis of variance (ANOVA) (middle-age adults [MA] =34, young adults [YA] =31).
EffectsDFnDFdF-valuep-value
Group13723.790.0522
SNR5372541.81<0.001***
Group:SNR53726.57<0.001***
0:Group16210.5120.001*
5:Group1620.0020.956
10:Group1621.3360.252
15:Group1621.8340.180
20:Group1620.2270.634
25:Group1624.9800.0292*

  1. *p < .05; ***p < .001.

Table 9
Fixed-effect estimates for the model of pupillary responses from 0 s to 5.8 s time-locked to the babble masker onset to examine the effect of signal-to-noise ratio (SNR) and age group (observations =96,612, groups: participant × SNR =332, participant =63).
Fixed effectEstimateSE95% CItp
 Intercept7.261.26[4.79, 9.73]5.76<0.001***
 ot148.1410.62[27.33, 68.94]4.53<0.001***
 ot2–30.446.90[-43.96,–16.92]–4.41<0.001***
 SNR 0–0.431.80[–3.95, 3.09]–0.240.811
 SNR 5–3.451.80[–6.98, 0.07]–1.920.055
 SNR 10–3.531.80[−7.06,–0.01]–1.960.049*
 SNR 15–3.721.77[−7.19,–0.25]–2.100.035*
 SNR 20–2.321.80[–5.85, 1.20]–1.290.197
 ot1 × SNR 044.7714.58[16.19, 73.34]3.070.002*
 ot1 × SNR 57.2114.59[–21.39, 35.81]0.490.621
 ot1 × SNR 10–11.3014.59[–39.90, 17.30]–0.770.439
 ot1 × SNR 15–16.1914.34[–44.29, 11.92]–1.130.259
 ot1 × SNR 20–13.0814.58[–41.65, 15.50]–0.900.370
 ot2 × SNR 044.889.26[26.73, 63.02]4.85<0.001***
 ot2 × SNR 561.779.27[43.60, 79.94]6.66<0.001***
 ot2 × SNR 1026.069.27[7.88, 44.23]2.810.005*
 ot2 × SNR 1514.749.11[–3.11, 32.59]1.620.105
 ot2 × SNR 2020.079.26[1.92, 38.22]2.170.030*
 Group (MA vs. YA)1.031.85[–2.60, 4.65]0.550.579
 Group × SNR 01.532.64[–3.65, 6.70]0.580.564
 Group × SNR 51.872.63[–3.28, 7.02]0.710.476
 Group × SNR 10–4.82e–032.60[–5.10, 5.10]–1.85e–030.999
 Group × SNR 15–1.262.62[–6.39, 3.88]–0.480.632
 Group × SNR 20–1.262.64[–6.44, 3.91]–0.480.632
 ot1 × Group–4.3715.56[–34.87, 26.14]–0.280.779
 ot1 × Group × SNR 013.7121.46[–28.35, 55.77]0.640.523
 ot1 × Group × SNR 532.8921.35[–8.94, 74.73]1.540.123
 ot1 × Group × SNR 1024.2021.16[–17.26, 65.66]1.140.253
 ot1 × Group × SNR 1515.4621.28[–26.25, 57.18]0.730.468
 ot1 × Group × SNR 208.9721.45[–33.07, 51.00]0.420.676
 ot2 × Group–4.8910.11[–24.70, 14.92]–0.480.628
 ot2 × Group × SNR 03.1813.65[–23.57, 29.93]0.230.816
 ot2 × Group × SNR 5–11.6513.57[–38.26, 14.96]–0.860.391
 ot2 × Group × SNR 1016.7613.46[–9.62, 43.13]1.240.213
 ot2 × Group × SNR 1514.8213.53[–11.70, 41.35]1.100.273
 ot2 × Group × SNR 2012.1913.63[–14.54, 38.91]0.890.371

  1. Growth curve formula: lmer(Pupil ~ (ot1 + ot2)*Group*SNR + (0+ot1+ot2 | participant) + (ot1 +ot2 | participant:SNR), control = lmerControl(optimizer = ‘bobyqa’), REML = FALSE). Orthogonal polynomial terms: ot1=linear (slope); ot2=quadratic (curvature).

  2. *p<0.05; ***p<0.001.

Table 10
Fixed-effect estimates for model of pupillary responses from 0 s to 3 s time-locked to Quick Speech-in-Noise (QuickSIN) target sentence offset to examine the effect of signal-to-noise ratio (SNR) and age group (observations =63,184, groups: participant × SNR =359, participant =63).
Fixed effectEstimateSE95% CItp
Intercept–0.360.81[–1.95, 1.22]–0.450.652
ot1–10.336.06[–22.20, 1.54]–1.710.088
ot2–2.243.12[–8.35, 3.88]–0.720.474
SNR 07.401.00[5.45, 9.36]7.43<0.001***
SNR 56.931.00[4.97, 8.88]6.95<0.001***
SNR 101.861.00[–0.09, 3.82]1.870.062
SNR 150.841.01[–1.13, 2.81]0.830.404
SNR 20–0.551.00[–2.50, 1.41]–0.550.583
ot1 × SNR 060.927.15[46.91, 74.92]8.52<0.001***
ot1 × SNR 545.167.15[31.15, 59.16]6.32<0.001***
ot1 × SNR 1020.107.15[6.10, 34.11]2.810.005*
ot1 × SNR 1513.387.21[–0.76, 27.51]1.850.064
ot1 × SNR 2012.277.15[–1.74, 26.28]1.720.086
ot2 × SNR 0–3.414.19[–11.62, 4.81]–0.810.416
ot2 × SNR 5–14.974.19[-23.19,–6.75]–3.57<0.001***
ot2 × SNR 106.434.19[–1.78, 14.65]1.530.125
ot2 × SNR 158.834.23[0.54, 17.12]2.090.037*
ot2 × SNR 207.834.19[–0.39, 16.05]1.870.062
Group (MA vs YA)–0.301.16[–2.57, 1.97]–0.260.796
Group × SNR 01.641.44[–1.18, 4.46]1.140.254
Group × SNR 50.371.43[–2.43, 3.16]0.260.796
Group × SNR 103.161.43[0.36, 5.97]2.210.027*
Group × SNR 153.791.45[0.95, 6.63]2.620.009*
Group × SNR 202.631.45[–0.22, 5.47]1.810.071
ot1 × Group3.288.67[–13.72, 20.27]0.380.706
ot1 × Group × SNR 0–0.8910.33[–21.13, 19.36]–0.090.932
ot1 × Group × SNR 54.0510.23[–15.99, 24.10]0.400.692
ot1 × Group × SNR 1025.3310.26[5.21, 45.44]2.470.014*
ot1 × Group × SNR 1514.0110.40[–6.37, 34.39]1.350.178
ot1 × Group × SNR 206.2410.43[–14.20, 26.67]0.600.550
ot2 × Group5.504.48[–3.29, 14.29]1.230.220
ot2 × Group × SNR 0–11.676.04[–23.51, 0.18]–1.930.053
ot2 × Group × SNR 53.625.99[–8.11, 15.36]0.610.545
ot2 × Group × SNR 10–6.726.01[–18.50, 5.06]–1.120.264
ot2 × Group × SNR 15–18.836.09[-30.77,–6.90]–3.090.002*
ot2 × Group × SNR 20–17.106.10[-29.06,–5.15]–2.800.005*

  1. Growth curve formula: lmer(Pupil ~ (ot1 +ot2)*Group*SNR + (ot1 +ot2 | participant) + (ot1 + ot2 | participant:SNR), control = lmerControl(optimizer = 'bobyqa'), REML = FALSE). Orthogonal polynomial terms: ot1=linear (slope); ot2=quadratic (curvature).

  2. *p<0.05; ***p<0.001.

Appendix 1—key resources table
Reagent type (species) or resourceDesignationSource or referenceIdentifiersAdditional information
Genetic reagent (Meriones unguiculatus)Crl:MON(Tum)Charles River243
Antibodyms(1gG1) α CtBP2 (mouse monoclonal)BD Transduction LabsBDB612044
Antibodyrb α MyosinVIIa (rabbit polyclonal)Proteus Biosciences25-670
Antibodyms(1gG2a) α GluA2 (mouse monoclonal)MilliporeMAB397
Antibodygt α ms (IgG2a) AF 488 (goat polyclonal)Thermo FisherA-21131
Antibodygt α ms (IgG1) AF 568 (goat polyclonal)Thermo FisherA-21124
Antibodydk α rb AF 647 (donkey polyclonal)Thermo FisherA-31573
Chemical compound, drugIsofluraneCovetrus29405
Chemical compound, drugDexmedetomidineCovetrus60984
Software, algorithmLabviewNational Instrumentshttps://www.ni.com/en-us/shop/labview.html
Software, algorithmMATLABMathWorkshttps://www.mathworks.com/products/matlab.html
OtherEyelink 1000 PlusSR Researchhttps://www.sr-research.com/eyelink-1000-plus/
OtherBioSemi Active Two EEGBioSemihttps://www.biosemi.com/Products_ActiveTwo.htm
OtherInsert earphonesEtymoticER-3C
OtherMulti-I/O Processor- RZ6-A-P1TDThttps://www.tdt.com/product/rz6-multi-i-o-processor/
OtherABR Amplifier, Medusa 4ZTDThttps://www.tdt.com/product/medusa4z-amplifier/

Additional files

MDAR checklist
https://cdn.elifesciences.org/articles/102823/elife-102823-mdarchecklist1-v1.pdf

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Maggie E Zink
  2. Leslie Zhen
  3. Jacie R McHaney
  4. Jennifer Klara
  5. Kimberly Yurasits
  6. Victoria E Cancel
  7. Olivia Flemm
  8. Claire Mitchell
  9. Jyotishka Datta
  10. Bharath Chandresekaran
  11. Aravindakshan Parthasarathy
(2025)
Increased listening effort and cochlear neural degeneration underlie speech-in-noise deficits in normal-hearing middle-aged adults
eLife 13:RP102823.
https://doi.org/10.7554/eLife.102823.3