Classical principles of cochlear operation have been challenged by recent discoveries that normal cochlear sound analysis is largely determined by motor responses of the outer hair cells (OHC), which in turn are under efferent control. It is now qualitatively clear that OHCs are responsible for the nonlinear and active response phenomena that have been studied extensively, including, two-tone suppression (2TS), combination tones (CT), otoacoustic emissions (OAE) and two-factor responses (2FR). Biophysical principles for these phenomena are unknown. An approach for developing the required understanding is proposed in which known cochlear structures are identified with signal processing operations that quantitatively represent extensive data on cochlear response phenomena. The research proposed builds upon the multiple-bandpass-nonlinearity (MBPNL) signal processing model of basilar membrane mechanical response, recently developed by the PI, and extends and develops it for CT and OAE data. The MBPNL model postulates two nonlinearly interacting filter systems corresponding to the """"""""tips"""""""" and """"""""tails"""""""" of cochlear frequency tuning curves. A working hypothesis is proposed that these two filtering systems are mediated by electrical and mechanical signals, respectively, which mix nonlinearly in the OHCS, and provide a synthesis of electromechanical and hydromechanical responses. Extension of the MBPNL model to CTs and OAEs adds new signal processing constraints on possible roles of cochlear structures. In contrast with 2TS, which is modeled as MBPNL processing localized along the cochlear partition, these phenomena suggest propagation of nonlinearly generated signals from one location to another, thereby providing opportunities for study of the roles of traveling wave theory in cochlear signal processing. Furthermore, CTs require differences in mechanism for physiological and psychophysical data, while OAEs demand structures for sustained and damped oscillations. The last requirement is met, while retaining previous ones, by reconfiguring the MBPNL model with two nested feedback loops (MFBPNL), in qualitative conformity with OHC bilateral and motility responses. Detailed study of these models is proposed to quantify the signal processing characteristics needed to represent the extensive existing data on cochlear nonlinear and active phenomena. The proposed development of quantitative cochlear signal processing models interactively with qualitative constraints from cochlear structure would contribute significantly to a modern biophysical theory of cochlear operation.
|Goldstein, J L (1995) Relations among compression, suppression, and combination tones in mechanical responses of the basilar membrane: data and MBPNL model. Hear Res 89:52-68|