Towards Predictions of Individual Benefits with Real Hearing Aids

* Presenting author
Day / Time: 21.03.2019, 15:00-15:20
Room: Saal 2
Typ: Regulärer Vortrag
Abstract: Accurate individual predictions of the benefit of hearing aids could be employed for a faster systematic development and better initial fitting of hearing aids. The Framework for Auditory Discrimination Experiments (FADE) was shown to accurately predict the speech reception thresholds (SRT) of hearing-impaired listeners and the benefit of several binaural noise reduction algorithms. However, several hours of processed signals were required for each predicted outcome which seem obstructive for predictions with physical hearing aids. Therefore, the framework was optimized to decrease required amount of realtime-data keeping the predictive power of FADE. Thus, an adaptive search paradigm across training and testing signal-to-noise ratios contrary to the original approach was implemented. As a result, this quick approach requires about 45 minutes of data for the prediction of the outcome of the matrix sentence test in an aided condition. The average difference in the predicted SRTs across a fluctuating masker was about 1 dB, i.e., within the empirical accuracy. The approach was evaluated with the open Master Hearing Aid with different hearing aid algorithms and different degrees of hearing loss. It showed SRTs similar to simulations with the original FADE approach. Ultimately, individual predictions for real (own) devices with recorded data are planned