Behavioural measurement of level-dependent shifts in the vibration pattern on the basilar membrane at 1 and 2 kHz
Introduction
Physiological measurements of the response of the basal end of the basilar membrane of cats and chinchillas show that the frequency of a sinusoid giving maximum response at a specific place decreases with increasing level of the sinusoid (Sellick et al., 1982, Ruggero, 1992, Ruggero et al., 1997). This is usually interpreted as indicating that the peak of the travelling wave on the basilar membrane evoked by a sinusoid shifts towards the base with increasing sound level. In a previous study (Moore et al., 2002) we presented psychoacoustical evidence supporting the idea that similar shifts occur on the human basilar membrane, for 4- and 6.5-kHz tones. Two experiments were performed, both using forward masking. The rationale for using forward masking is as follows. If masking patterns are measured in forward masking with a relatively long signal delay, the level of the signal at threshold is much lower than that of the masker (Zwislocki and Pirodda, 1952, Widin and Viemeister, 1979, Jesteadt et al., 1982, Moore and Glasberg, 1983). If level-dependent shifts occur, the peak of the travelling wave will occur at different places for the signal and masker. Therefore, the frequency at which the masker is most effective may be different from the signal frequency.
In the first experiment of Moore et al. (2002), masking patterns were measured in forward masking using a fixed 6.5-kHz masker tone presented at 65 or 85 dB sound pressure level (SPL). The threshold for detecting a brief sinusoidal signal was measured as a function of signal frequency for several time delays of the signal relative to the end of the masker. A background noise, with a fixed signal-to-noise ratio, was used to reduce ‘off-frequency listening’ (Johnson-Davies and Patterson, 1979, O’Loughlin and Moore, 1981a, O’Loughlin and Moore, 1981b, Oxenham and Plack, 1997). As the signal delay was increased, the signal level at the peaks of the masking patterns decreased and the signal frequency at the peaks of the patterns moved progressively towards higher frequencies. The pattern of results was consistent with the idea of a basalward shift of the peak of the travelling wave with increasing level of about 0.26 octave for a change in level from 45 to 85 dB SPL.
Although this experiment provided evidence for level-dependent shifts, the results showed high variability both within and across subjects. Moore et al. (2002) described two possible sources of this variability. Firstly, subjects reported that the background noise used to restrict off-frequency listening was distracting. The noise level was fixed relative to the signal level, and hence changed from trial to trial as the signal level was varied. Secondly, since the signal level was varied adaptively during a run, the changes in signal level from trial to trial meant that, assuming there are level-dependent shifts, the peak in the travelling wave pattern of the signal also varied from trial to trial.
The second experiment of Moore et al. (2002) was designed to overcome these problems. No background noise was used. To restrict off-frequency listening, the signal was presented at a low sensation level (SL), namely 10 dB SL. To prevent problems associated with trying to measure a ‘moving target’, the masker level was also fixed within a given run. For each masker level, the ‘gap threshold’ was determined. This is defined as the silent interval between the masker and the signal required to reach ‘threshold’, i.e. the gap leading to 79.4% correct detection of the signal. It was hypothesised that the gap threshold would be largest when the signal and the masker produced travelling waves with peaks at the same position on the basilar membrane. Finally, the signal frequency was fixed and the masker frequency was varied. This was done to reduce any influence on the results of microstructure in the absolute thresholds as a function of signal frequency (van den Brink, 1970, Long and Tubis, 1988). The masker level was always well above the absolute threshold, and at higher levels the influence of threshold microstructure appears to be minimal.
In the second experiment of Moore et al. (2002), the signal was a 4-kHz tone presented at 10 dB SL. For three fixed masker levels (65, 85 and 95 dB), the gap threshold was measured as a function of masker frequency. The gap-threshold patterns sometimes showed two peaks. One occurred just below the signal frequency and the frequency at the peak was hardly affected by masker level. The second peak fell at a lower frequency, and this frequency tended to decrease with increasing masker level. These results may reflect double-peaked frequency–response functions at the place on the basilar membrane responding to the signal frequency (Narayan and Ruggero, 2000, Ruggero et al., 2000). The gap-threshold patterns often had very shallow low-frequency sides for the highest masker level. The changes with level provided further evidence for a basalward spread of the basilar-membrane vibration pattern with increasing level.
Evidence for level-dependent shifts in the peak of the travelling wave at medium and low frequencies is somewhat conflicting. Data obtained from apical regions of the guinea pig and chinchilla basilar membranes show no shifts or only small shifts in best frequency with increasing level (Cooper and Rhode, 1997, Zinn et al., 2000). It is possible to make indirect inferences about changes in tuning with level from impulse responses derived from the auditory nerve using the reversed-correlation technique (de Boer and Kruidenier, 1990, Carney et al., 1999). It has been argued that the glides in instantaneous frequency observed in these responses (which are almost level-invariant in cats), combined with changes in the overall shapes of the impulse responses with level, are consistent with a basalward shift of the peak of the travelling wave for high frequencies, but no shift or an apical shift for low frequencies (Carney et al., 1999). Iso-intensity response functions obtained from auditory nerve fibres show that for fibres with high best frequencies, the frequency giving maximal response decreases with increasing level, while for fibres with low best frequencies the frequency giving maximal response increases with increasing level (Rose et al., 1971). The latter effect is consistent with an apical shift with increasing intensity.
Indirect psychophysical evidence for a basalward shift in humans comes from the finding that the maximum temporary threshold shift caused by exposure to an intense tone occurs for a test frequency somewhat above the frequency of the exposure tone, even when the exposure tone has a medium frequency (Hirsh and Ward, 1952, McFadden and Plattsmier, 1983, McFadden, 1986). However, evidence from studies of forward masking is less clear cut. Kidd and Feth (1981) measured psychophysical tuning curves (PTCs) for a 1-kHz signal at 10 dB SL, as a function of the masker–signal delay. With increasing delay, the masker level at the tip of the PTC increased. For three of the four subjects tested, the frequency at the tip (i.e. the frequency at which the masker was most effective) did not change systematically with increasing masker–signal delay, suggesting that there was no level-dependent shift in the peak of the travelling wave. However, for one subject the frequency at the tip decreased with increasing delay, suggesting a basalward shift with increasing level.
McFadden and Yama (1983) measured masking patterns in forward masking for a 2-kHz sinusoidal masker. They found that the peak of the masking pattern (quantified as the ‘centre of balance’) shifted towards higher frequencies with increasing masker level when they used a relatively long (50 ms) signal and long (50 ms) masker–signal interval. Similar findings had been reported earlier by Zwislocki and Pirodda (1952). However, shifts with level were absent or smaller when the signal was short (15 ms) and/or the interval was short (10 ms). No shift with masker level was found for a 750-Hz masker, although the peak of the masking pattern for both masker levels used (65 and 95 dB SPL) was somewhat above the masker frequency. In the experiments of Zwislocki and Pirodda (1952) and McFadden and Yama (1983), no precautions were taken to limit the extent of off-frequency listening, and this may have had an important influence on the results; we return to this point in Section 4. Overall, it remains unclear whether basalward shifts in the peak of the travelling wave with increasing level occur for medium and low frequencies in humans.
In the present paper, we used a forward-masking technique identical to that used in experiment 2 of Moore et al. (2002) to determine whether level-dependent shifts in the peak of the travelling wave occur for mid-range frequencies in humans.
Section snippets
Procedure
An adaptive three-interval two-alternative forced-choice procedure was used. The masker was presented in all intervals, and the signal followed the masker in one interval only, selected at random. Observation intervals were marked by lights on the response box. Subjects were required to indicate the interval containing the masker plus signal. Correct-answer feedback was provided by lights on the response box. The silent interval between the masker and signal was varied using a 3-down 1-up
Results
Absolute thresholds for detecting the brief 1-kHz signal were 14, 10.5, 21, 16 and 22 dB SPL for BG, CB, AD, LA and UK, respectively. Corresponding thresholds for the 2-kHz signal were 22, 16, 28, 21 and 22 dB. Thus, the signal levels used in the experiment (10 dB above the absolute threshold) were all at least 37 dB below the lowest masker level used. Note that level-dependent shifts in the basilar-membrane peak response are expected to be revealed as shifts in the gap-threshold patterns when
Discussion
In contrast to the results of our earlier study (Moore et al., 2002), which used signal frequencies of 4 and 6.5 kHz, the present results for signal frequencies of 1 and 2 kHz do not show consistent evidence for a basalward shift in the travelling wave with increasing level, at least for the range of levels from 10 dB above absolute threshold (the signal level) to 95 dB SPL (the highest masker level used). The results of a few subjects showed weak evidence for small shifts (about 5%), but these
Summary and conclusions
Physiological data suggest that the peak of the travelling wave on the basilar membrane evoked by a high-frequency sinusoid moves towards the base with increasing level, but it is not clear whether similar effects occur for mid-frequency sinusoids. We used a forward-masking technique to determine whether level-dependent shifts occur for mid-range frequencies. The signal was a 1-kHz or 2-kHz tone presented at 10 dB SL. This low level was chosen to minimise the possibility of subjects detecting
Acknowledgements
This work was supported by the MRC (UK). We thank Chris Bond, Lorna Arrol and Umaiya Kugathasan for gathering some of the data reported here. We also thank an anonymous reviewer for helpful comments.
References (38)
- et al.
Patterns of residual masking
Hear. Res.
(1981) - et al.
Formulae describing frequency selectivity as a function of frequency and level and their use in calculating excitation patterns
Hear. Res.
(1987) - et al.
Behavioural measurement of level-dependent shifts in the vibration pattern on the basilar membrane
Hear. Res.
(2002) - et al.
Improving psychoacoustical tuning curves
Hear. Res.
(1981) Responses to sound of the basilar membrane of the mammalian cochlea
Curr. Opin. Neurobiol.
(1992)- et al.
Evidence for active, nonlinear, negative feedback in the vibration response of the apical region of the in-vivo guinea-pig cochlea
Hear. Res.
(2000) - et al.
Frequency glides in the impulse responses of auditory-nerve fibers
J. Acoust. Soc. Am.
(1999) - et al.
Mechanical responses to two-tone distortion products in the apical and basal turns of the mammalian cochlea
J. Neurophysiol.
(1997) - et al.
On ringing limits of the auditory periphery
Biol. Cybern.
(1990) - et al.
Growth-of-masking functions for several types of maskers
J. Acoust. Soc. Am.
(1994)
Recovery of the auditory threshold after strong acoustic stimulation
J. Acoust. Soc. Am.
Forward masking as a function of frequency, masker level, and signal delay
J. Acoust. Soc. Am.
Psychophysical tuning curves: restricting the listening band to the signal region
J. Acoust. Soc. Am.
Modification of spontaneous and evoked otoacoustic emissions and associated psychoacoustic microstructure by aspirin consumption
J. Acoust. Soc. Am.
Frequency patterns of TTS for different exposure intensities
J. Acoust. Soc. Am.
Upward shifts in the masking pattern with increasing masker intensity
J. Acoust. Soc. Am.
Growth of forward masking for sinusoidal and noise maskers as a function of signal delay: implications for suppression in noise
J. Acoust. Soc. Am.
Masking patterns for sinusoidal and narrowband noise maskers
J. Acoust. Soc. Am.
Cited by (13)
Contributions of von Békésy to psychoacoustics
2012, Hearing ResearchCitation Excerpt :The shift is equivalent to a change in frequency up to one-half of an octave. Forward-masking experiments support the idea that such shifts occur in humans, at least for high frequencies (Moore et al., 2002; Moore and Glasberg, 2003). If pitch were really determined by the position of the peak excitation on the basilar membrane, large shifts in pitch with level should be heard.
A comparative study of eight human auditory models of monaural processing
2022, Acta AcusticaA clinically oriented introduction and review on finite element models of the human cochlea
2014, BioMed Research InternationalOn the controversy about the sharpness of human cochlear tuning
2013, JARO - Journal of the Association for Research in Otolaryngology