Behavioral Measures of Cochlear Gain Reduction Depend on Precursor Frequency, Bandwidth, and Level

Abstract

Sensory systems adjust to the environment to maintain sensitivity to change. In the auditory system, the medial olivocochlear reflex (MOCR) is a known physiological mechanism capable of such adjustment. The MOCR provides efferent feedback between the brainstem and cochlea, reducing cochlear gain in response to sound. The perceptual effects of the MOCR are not well understood, such as how gain reduction depends on elicitor characteristics in human listeners. Physiological and behavioral data suggest that ipsilateral MOCR tuning is only slightly broader than it is for afferent fibers, and that the fibers feed back to the frequency region of the cochlea that stimulated them. However, some otoacoustic emission (OAE) data suggest that noise is a more effective elicitor than would be consistent with sharp tuning, and that a broad region of the cochlea may be involved in elicitation. If the elicitor is processed in a cochlear channel centered at the signal frequency, the growth of gain reduction with elicitor level would be expected to depend on the frequency content of the elicitor. In the current study, the effects of the frequency content and level of a preceding sound (called a precursor) on signal threshold was examined. The results show that signal threshold increased with increasing precursor level at a shallower slope for a tonal precursor at the signal frequency than for a tonal precursor nearly an octave below the signal frequency. A broadband noise was only slightly more effective than a tone at the signal frequency, with a relatively shallow slope similar to that of the tonal precursor at the signal frequency. Overall, these results suggest that the excitation at the signal cochlear place, regardless of elicitor frequency, determines the magnitude of ipsilateral cochlear gain reduction, and that it increases with elicitor level.

Keywords: cochlear gain reduction; elicitor bandwidth; forward masking; frequency selectivity; medial olivocochlear reflex; psychoacoustics.

FIGURE 1

Schematic of the temporal masking paradigm used in this experiment, including a 50-ms precursor, 20-ms masker, and 6-ms signal. The precursor or masker is removed in some experiments, but the temporal relationships are not changed. The frequency content of the precursors and maskers also vary across experiments, but the signal is always presented at 4 kHz. The gray dotted line shows a schematic of the timecourse of forward masking due to neural excitation. The gray solid line shows a schematic of the timecourse of forward masking due to cochlear gain reduction with a precursor present.

FIGURE 2

Schematic of cochlear input-output functions and threshold predictions in Experiment 1. Signal threshold (S) occurs at a criterion signal-to-masker ratio (SMR), in this case 0 dB (first column) for an on-frequency (top row) and off-frequency (bottom row) masker (M). With the addition of a precursor (P), predictions differ for forward masking due to additivity of masking or gain reduction. With additivity of masking (second column), a similar shift in signal threshold is expected when the same precursor is presented with equally effective on- and off-frequency maskers (arrows in second column). With gain reduction (third column), a larger shift in signal threshold is expected in the off-frequency case, since the masker is not affected by gain reduction at the signal frequency place (arrow in third column). The input-output functions, S, and M from the first column are repeated in gray in the second and third columns to illustrate the predicted changes with the introduction of a precursor.

FIGURE 3

Individual GOM functions and masker-absent gain reduction estimates. Signal thresholds for the masker-alone condition are plotted as open circles and signal thresholds with the addition of a precursor are plotted as filled circles. Signal threshold without a preceding masker is plotted as an open triangle, and signal threshold with a precursor and 20-ms delay is plotted as a filled triangle. The difference between the triangles is the masker-absent gain reduction estimate. Arrows indicate the off-frequency masker levels used in the equally-effective-masker conditions. Error bars represent one standard deviation.

FIGURE 4

Bars indicate the group average increase in signal threshold with a precursor preceding equally effective off-frequency (2.4-kHz) and on-frequency (4-kHz) maskers. Signal threshold shift with a precursor was averaged for two matched signal levels for each participant (symbols). Error bars represent one standard deviation.

FIGURE 5

Individual estimated gain reduction with precursor level measured with 2.4-kHz (pink diamonds) and 4-kHz (green circles) precursors. Estimated gain reduction was calculated by subtracting quiet threshold for the signal from the signal threshold for each condition. The two data points measured for each precursor level and included in the analysis are plotted. Linear mixed-effects model fits (see Table 2 for model summary) are plotted as lines over the data.

FIGURE 6

Individual estimated gain reduction with precursor level measured with BBN (purple hourglasses), 2.4-kHz (pink diamonds, identical to data presented in Figure 5) and 4-kHz (green circles, identical to data presented in Figure 5) precursors. Precursor level is plotted in dB per equivalent rectangular bandwidth (ERB) for the noise precursors. Estimated gain reduction is calculated by subtracting quiet threshold for the signal from the signal threshold for each condition. LMM fits (see Table 3 for model summary) are plotted as lines over the data.