Estimating the form of the peripheral nonlinearity from slopes of psychometric functions in forward masking

Abstract

In previous work, we have shown that slopes of psychometric functions (PF) obtained for the detection of a signal in the presence of a forward masker are in general agreement with standard assumptions concerning the form of the peripheral nonlinearity. PFs are steep for conditions where signal levels are low and grow progressively shallower as the signal threshold increases in level. PFs become more parallel if signal levels are adjusted to reflect the effect of a nonlinearity of the form proposed by Yates et al. [G. K. Yates, I. M. Winter, and D. Robertson, Hear. Res. 45, 203–220 (1990)]. An alternative approach is to use PF data to estimate the form of the nonlinearity. The data suggest that the reciprocal of the slope of the PF is linearly related to the signal level. This would indicate that the nonlinearity is logarithmic. Use of a logarithmic nonlinearity as the first stage in a model of forward masking predicts thresholds as a function of the masker level and signal delay that are consistent with a formula proposed by Jesteadt et al. [W. Jesteadt, S. P. Bacon, and J. R. Lehman, J. Acoust. Soc. Am. 71, 950–962 (1982)] and predicts loudness growth rates consistent with the literature. [Work supported by NIDCD.]