Interpretation of standard distortion product otoacoustic emission measurements in light of the complete parametric response.

Emission characteristics (at 2f1-f2) are measured in Mongolian gerbil as a function of the independent variation of all four stimulus parameters, the frequencies (f1 and f2) and the intensities (L1 and L2) of the two stimulus tones. The main five-dimensional display chosen is a logarithmic grid of frequencies, where for each frequency pair there is a contour map of the emission amplitude as a function of the two stimulus levels. The feature which leads to the greatest complexity in the proper interpretation of emission responses is the widespread presence of "notches" in these contour maps. Notches are lines of relative minima in the emission amplitude, and are found at either: (1) constant L1, but only in regions where L1 > L2; or (2) at constant L2, only where L2 > or = L1. Notches are not found at any other orientations, and are associated with emission phase shifts of about 180 degrees as the notch line is traversed. These notch characteristics are explained by phase cancellation in a simple cochlear amplifier model in which there is a change, as a function of the stimulus level alone, of relevant characteristics of the cochlear response to a single tone. Only one mechanism of emission generation is required to explain the observed patterns, i.e., there is no need to invoke different "active" and "passive" mechanisms. Unless properly accounted for, the presence of notches adversely affects all of the standard emission measurements, i.e., all methods which cover a restricted parameter set such as DPgrams, input-output or "growth" functions, and frequency ratio functions. Conversely, because the notch location appears approximately invariant in the cochlea, notches potentially make it possible to use certain emission growth functions to estimate forward and reverse middle-ear transfer functions.