Special Data Dialogue: Peculiar properties of traveling waves in the cochlea and an iterative linearization technique for modeling
The cochlea is the sound-retrieving organ of mammals. Its main feature is a tonotopic mapping that allows us to distinguish sounds of different frequencies. Further, when waves propagate along its basilar membrane (BM), they are amplified by outer hair cells (OHCs) via an electromotile feedback. This feedback allows us to hear sounds at a rather soft level while the response to loud sounds are compressed because of the saturation of OHC transduction current. The overall input (sound level)-output (BM displacement) relation becomes nonlinear, and it is well-known that the response to one tone can be suppressed by another tone. In this talk, I will review an iterative technique that stems from the intuition of Kanis and de Boer (1993) to represent the nonlinear cochlea by equivalent linear models. Using this technique, the mechanical structure of the model is preserved, so linear concepts can still be used locally such as feedback, wave impedance, and reflectance. Based on this framework, we have carried out details for linearization and compute cochlear response to broadband noise (Liu, 2014) and backward wave propagation that escapes the cochlea to become emissions (Liu and Liu, 2016). I will show empirical comparison to brute-force, time-domain finite difference simulation, and attempt to discuss the limitations of this method.