Toward an Individual Binaural Loudness Model for Hearing Aid Fitting and Development.

Iko Pieper,Manfred Mauermann,Birger Kollmeier,Stephan D Ewert

doi:10.3389/fpsyg.2021.634943

Abstract

The individual loudness perception of a patient plays an important role in hearing aid satisfaction and use in daily life. Hearing aid fitting and development might benefit from individualized loudness models (ILMs), enabling better adaptation of the processing to individual needs. The central question is whether additional parameters are required for ILMs beyond non-linear cochlear gain loss and linear attenuation common to existing loudness models for the hearing impaired (HI). Here, loudness perception in eight normal hearing (NH) and eight HI listeners was measured in conditions ranging from monaural narrowband to binaural broadband, to systematically assess spectral and binaural loudness summation and their interdependence. A binaural summation stage was devised with empirical monaural loudness judgments serving as input. While NH showed binaural inhibition in line with the literature, binaural summation and its inter-subject variability were increased in HI, indicating the necessity for individualized binaural summation. Toward ILMs, a recent monaural loudness model was extended with the suggested binaural stage, and the number and type of additional parameters required to describe and to predict individual loudness were assessed. In addition to one parameter for the individual amount of binaural summation, a bandwidth-dependent monaural parameter was required to successfully account for individual spectral summation.

Highlights

Being “too loud” is the most frequent descriptor for fitting problems with hearing aids (Jenstad et al, 2003), and current hearing aid fitting procedures take loudness into consideration (e.g., Moore and Glasberg, 1998; Byrne et al, 2001; Keidser et al, 2012)
In this study binaural loudness summation is calculated from the ratios between binaural and monaural loudness in sones at a given level or by fitting a single model parameter that alters the modeled binaural inhibition to the empirical loudness data, van Beurden et al (2018) calculated the level differences between monaural and binaural loudness functions at given loudness categories, but if the loudness ratios are kept constant, the increase in the steepness of the loudness functions caused by the hearing impairment decreases the level differences between the functions (Moore et al, 2014)
Loudness perception of the normal hearing (NH) and hearing impaired (HI) listeners with sensorineural hearing loss was measured by categorical loudness scaling for narrowband and broadband stimuli, presented monaurally, and binaurally

Summary

Introduction

Being “too loud” is the most frequent descriptor for fitting problems with hearing aids (Jenstad et al, 2003), and current hearing aid fitting procedures take loudness into consideration (e.g., Moore and Glasberg, 1998; Byrne et al, 2001; Keidser et al, 2012). Loudness perception differs significantly across individuals with similar audiometric hearing loss (Moore, 2000) and, to some extent, for NH listeners (e.g., Pieper et al, 2018) This suggests that loudness models with parameters based on averaged data, and with individualization of Individual Binaural Loudness Model parameters for HI listeners inferred solely from their audiogram (e.g., Moore et al, 1999), might not be sufficient to predict individual loudness perception (Oetting et al, 2013; Pieper et al, 2018). In realistic environments, sounds are typically perceived binaurally in addition to showing broadband properties, as observed for, e.g., speech and environmental noise

Objectives

Methods

Results

Discussion

Conclusion