Timothy A. Wilson
Room No. 204A, Eng. Sci. Bldg., Dept. of Elec. Eng., Memphis State Univ., Memphis, TN 38152
William M. Siebert
MIT, Cambridge, MA 02139
The similarities and distinctions among linear, passive, cochlear models of one-, two-, and three-dimensional fluid motion---models popularized by (among others) Zwislocki, Ranke, and Steele, respectively---are confounded by fuzzy terminology (e.g., ``long-wave'' and ``short-wave''). Such models are frequently evaluated by comparing their place responses with measured frequency responses; their global impedance parameters are sometimes chosen solely to secure fit to local observations. Steele's WKB (phase-integral) approach is often treated as a technique for solving cochlear dynamical equations rather than as a conceptual framework yielding insight into cochlear phenomena. In this presentation, cochlear dynamical equations are developed for one-, two, and three-dimensional fluid motion in a box-cochlea model. The phase-integral approximate solution to these equations is described; the utility of this framework for explaining cochlear phenomena is discussed. Generalized representations for both cochlear-partition impedance and cochlear-gain response are developed highlighting the similarities and distinctions between the place response at a single frequency and the frequency response at a single place. The generalized representations clarify which aspects of partition impedance determine global features such as cochlear maps and which aspects determine local features such as magnitude-response peakiness and phase-response steepness.