Prevalence of excess binaural broadband loudness summation in the hearing-impaired population and implications for hearing aid gain targets

Abstract

Previous studies reported large individual differences in binaural broadband loudness summation in hearing-impaired listeners after narrowband loudness was normalized. These differences in loudness perception might require substantial fine-tuning for some hearing aid users to provide acceptable loudness in daily use. The present study aims at characterizing binaural broadband loudness summation for a hearing-impaired population, the prevalence of higher-than-normal values, and the potential implications for hearing aid target gains. For 180 hearing-impaired participants we measured standard audiological diagnostic parameters, binaural broadband loudness summation and computed gain targets according to NAL-NL2, DSLm[i/o] and trueLOUDNESS, a prescriptive procedure that includes individual loudness measurements. The observed binaural broadband loudness summation of the hearing-impaired participants was, on average, 13 dB higher and showed a higher variance relative to a normal-hearing reference group. In about 40% of all participants a binaural broadband loudness summation was beyond the normal-hearing range. The average excess loudness summation appears to be included in established prescriptive procedures, while relevant differences to trueLOUDNESS gain targets were observed in more than half of the participants. Elevated binaural broadband loudness summation appears to be a prevalent and individual trait in the hearing-impaired population and its assessment might be useful for individualized hearing aid fitting.

Fig 1. Air-conduction thresholds (black) and UCLs (red) for all participants, thick lines showing averages.

Solid lines show data for left ears, dashed lines for right ears.

Fig 2. Distribution of age and PTA thresholds for each participant group.

Horizontal lines indicate the median, boxes the interquartile range, and whiskers the full data range excluding outliers (crosses). Brackets with asteriscs above denote significant differences (p < 0.001).

Fig 3. Loudness functions for IFnoise with narrowband loudness normalization for hearing-impaired participants (gray lines, thick black line showing average) and normal-hearing reference function (thick green line).

Fig 4. Distribution of binaural broadband loudness summation, showing cumulative plots (bottom part) and boxplots (top part) for each group.

Diamonds in the boxplots denote the mean of the distributions.

Fig 5. Correlations between the individual excess binaural broadband loudness summation ΔL40 and other obtained audiological parameters (see x-axes).

Green shaded area denotes the 90% range for normal-hearing participants. Text in the top right corners shows coefficients of determination, stars indicate significant correlations.

Fig 6. Scatter plots of PTA gains against PTA hearing loss for each prescription rule and input level with given R2 values.

Boxplots show the distribution of PTA gains.

Fig 7. Boxplots of compression ratios at each frequencies across several level ranges in the regarded gain as specified in eq. (1) for level range given in the panel title and y-label.

Boxplots show distributions of compression ratios from gain targets for the given frequency determined in all ears of New and Experienced participants (320 ears).

Fig 8. Boxplots of the gain difference between trueLOUDNESS and NAL-NL2 (top row), and between trueLOUDNESS and DSLm[i/o] (bottom row) for each input level (see panel title) and frequency.

Black boxplots show distributions across all ears, colors indicate grouping by the excess binuaral broadband loudness summation (see legend). Percentages in the legend denote the percentage of participants falling into each category.

Fig 9. Cumulative occurrence of gain differences between trueLOUDNESS and other fitting rules as denoted by panel title based on the unsigned average of differences across 500, 1000, 2000 and 4000 Hz, input levels of 50, 65 and 80 dB, and both ears of a given participant.