Cochlear synaptopathy, the loss of auditory nerve (AN) fiber synapses after a temporary threshold shift, has been demonstrated in several non-human mammals. By far, the mouse is the most used and best characterized species in connection with cochlear synaptopathy. Studies in human listeners are, however, inconclusive. The impossibility to directly assess the status of the AN in humans is a major challenge. Hence, similar studies carried out in humans and mice are of great interest. Previously, we proposed the use of envelope following responses (EFR) as a tool to investigate synaptopathy both in mice and humans. We reported similar patterns in synaptopathic mice and humans, which could be explained by a humanized model of the AN with simulated synaptopathy. In this study, the original animal version of the AN model developed using cat data was evaluated and modified to account for the EFRs recorded in mice.
The cat version of the AN model by Zilany et al. (2009, 2014) was used to simulate EFR level-growth functions using sinusoidally amplitude modulated tones with strong (m=85%) and shallow (m=25%) modulation depths in noise-exposed and control mice. The stimuli used in the mice recordings were adapted to the cat AN model by transposing them to the cat's relative cochlear place. A mouse model was proposed by modifying the middle-ear response, the AN tuning, and the sensitive frequency range of the model.
The cat AN model could account for the overall effect of synaptopathy on the recorded EFRs in noise-exposed mice. However, the shape of the recorded EFR level-growth functions both in the noise-exposed and the control mice did not fully agree with the simulated EFRs using the cat model. The mouse model also showed deviations from the data.
The analysis of the simulated results suggested that the inaccuracy of the model outcome may be caused by the differences in AN tuning between the model (cat) and the data (mouse). In addition, at supra-threshold levels, the total EFR is dominated by the off-frequency contributions (i.e., neuronal activity away from the characteristic place of the stimulus), which are limited in the high frequencies of the cat AN model. The proposed modifications of the model showed improvements which were not sufficient to entirely account for the data.
This work was supported by CHeSS at the Technical University of Denmark.
|Period||10 Feb 2018|
|Event title||41st Midwinter Meeting of the Association for Research in Otolaryngology : null|
|Location||San Diego, United States, California|
|Degree of Recognition||International|
- computational model
- mouse auditory nerve
- cochlear synaptopathy