Optimal L(1)-L(2) primary tone level separation remains independent of test frequency in humans

Previous studies described a systematic asymmetry of the level of the 2f(1)-f(2) distortion product otoacoustic emission (DP) in the space of the primary tones levels L(1) and L(2) in normal-hearing humans. Optimal primary tone level separations L(1)-L(2), which result in maximum DP levels, were close to L(1)=L(2) at high levels, but continuously increased with decreasing stimulus level towards L(1)>L(2) (Gaskill and Brown, 1990, J. Acoust. Soc. Am. 88, 821-839). At these optimal L(1)-L(2), however, not only DP levels in normal hearing were maximal, but also trauma-induced DP reductions. A linear equation that approximates optimal L(1)-L(2) level separations thus was suggested to be optimum for use in clinical applications (Whitehead et al., 1995, J. Acoust. Soc. Am. 97, 2359-2377). It was the aim of this study to extend the generality of optimal L(1)-L(2) separations to the typical human test frequency range for f(2) frequencies between 1 and 8 kHz. DPs were measured in 22 normal-hearing human ears at 61 primary tone level combinations, with L(2) between 5 and 65 dB SPL and L(1) between 30 and 70 dB SPL (f(2)/f(1)=1.2). It was found that the systematic dependence of the maximum DP level on the L(1)-L(2) separation is independent on frequency. Optimal L(1)-L(2) level separations may well be approximated by a linear equation L(1)=a L(2)+(1-a) b (after Whitehead et al., 1995) with parameters a=0.4 and b=70 dB SPL at f(2) frequencies between 1 and 8 kHz and L(2) levels between 20 and 65 dB SPL. Below L(2)=20 dB SPL, the optimal L(1) was found to be almost constant. Following previous notions (Gaskill and Brown, 1990), an analysis of basilar membrane response data in experimental animals (after Ruggero and Rich, 1991, Hear. Res. 51, 215-230) is further presented that relates optimal L(1)-L(2) separations to frequency-selective compression of the basilar membrane. Based on the assumption that optimal conditions for the DP generation are equal primary tone responses at the f(2) place, a linear increase of the optimal L(1)-L(2) level separation is graphically demonstrated, similar to our results in human ears.