Testing phenomenological auditory-nerve model predictions for selective inner- and outer-hair-cell dysfunction

Abstract

Sensorineural hearing loss (SNHL) effects on neural coding and perception have been largely associated with outer-hair-cell (OHC) dysfunction (e.g., reduced cochlear gain, reduced compression, broadened tuning). However, both inner-hair-cell (IHC) and OHC dysfunction occur in common hearing-loss etiologies, e.g., noise-induced and age-related (metabolic). Although IHC effects are largely ignored based on the insensitivity of audiograms to IHC loss up to 80%, the same selective IHC damage affects suprathreshold hearing (e.g., tone-in-noise detection, EFRs to periodic stimuli). Phenomenological auditory-nerve (AN) models are useful for studying neural-coding effects of SNHL, but their parametric control of OHC and IHC function (e.g., cohc and cihc parameters, Bruce et al., 2018) are designed based on data from animals with noise-induced hearing loss (OHC/IHC dysfunction combined). Although this approach has proven successful for OHC effects, testing of IHC-based effects has been largely indirect. Here, we compare model predictions to our previously collected AN-fiber responses to amplitude-modulated stimuli in animals with either carboplatin-induced selective IHC dysfunction or gentamicin-induced selective OHC damage to provide more direct testing of IHC and OHC effects. Model parameters were estimated from histology, ABR thresholds, AN-fiber rates, and spike-train statistical metrics (e.g., vector strength, discriminability, mutual information).