Predicting speech intelligibility in normal-hearing and hearing-impaired listeners based on a physiologically inspired model of the auditory periphery

The present study extends and evaluates a speech intelligibility prediction model for normal-hearing (NH) and hearing-impaired (HI) listeners recently introduced by Scheidiger et al. [(2018). Acta Acustica united Ac. 104, 914-917]. The model combines a non-linear model of the auditory periphery with a decision process that is based on fluctuation profiles across characteristic frequency (CF) after modulation analysis in the range of the fundamental frequency of speech. Specifically, the short-term across-CF correlation between noisy speech and noise alone is assumed to be inversely related to speech intelligibility.The model has been shown to account for NH listeners’ speech intelligibility in conditions with stationary and fluctuating interferers. Furthermore, some of the main effects measured in HI listeners were predicted by adapting the peripheral model using a simplistic interpretation of the listeners’ hearing thresholds. Here, the effects of various impairment factors on the model performance were explored to better account for the individual HI listener, focusing particularly on the ratio of inner and outer hair cell loss and on the modulation-frequency tuning assumed in the model. The results are discussed regarding the validity of the individual best-matching model configurations and in relation to a further validation using data from auditory profiling.