Nonlinear interactions and the 'Kemp echo'

It would be convenient to assume one nonlinear interaction process as responsible for almost all cochlear nonlinearities including the 'Kemp echo'. In this paper various possible forms of such a process are discussed. In general the interaction assumes the form of a feedback system, its input being derived from physiological excitation and its output affecting the propagation of cochlear waves. Only first-harmonic effects are considered here. If the feedback derives from the averaged physiological activity, it effectively modifies cochlear mechanics. In the modeling exercises it proved impossible to construct a feedback system that would have the largest effects near the location of resonance, hence with this type of feedback system the desired goal cannot be attained. In another form the feedback signal is considered to be a variable that directly relates to physiological excitation, an intracellular potential, for instance. This type of feedback operates strictly locally. The functional forms of sharpened responses and (short-wave) cochlear mechanics are fairly well known, and the feedback system can be implemented rather easily. The results were somewhat unexpected: the resulting system has an innate tendency to be unstable. It seems that the types of feedback systems considered cannot be regarded as likely candidates for a nonlinear theory of the cochlea.