LAUSR

A variety of auditory-nerve models, as well as a vast amount of animal single-unit and population response data that can be used to set the parameters of such models, exists. However, it is hard to evaluate different model implementations from published data to decide whether the specific implementation is appropriate for your envisioned application. In this presentation, I will give an experience-based overview on the challenges faced with when evaluating the model parameters in auditory-nerve models. The adopted approach uses the available computer code of the different models under test, and compares their responses to the same input. This method is very efficient in testing the influence of changing one specific part of the model while leaving the rest unchanged, and can ultimately yield improved functional models of the auditory periphery.