Characterising auditory filter nonlinearity

Test-retest evaluation of a notched-noise test using consumer-grade mobile audio equipment

OBJECTIVE

The aim of this study was to investigate whether consumer-grade mobile audio equipment can be reliably used as a platform for the notched-noise test, including when the test is conducted outside the laboratory.

DESIGN

Two studies were conducted: Study 1 was a notched-noise masking experiment with three different setups: in a psychoacoustic test booth with a standard laboratory PC; in a psychoacoustic test booth with a mobile device; and in a quiet office room with a mobile device. Study 2 employed the same task as Study 1, but compared circumaural headphones to insert earphones.

STUDY SAMPLE

Nine and ten young, normal-hearing participants completed studies 1 and 2, respectively.

RESULTS

The test-retest accuracy of the notched-noise test on the mobile implementation did not differ from that for the laboratory setup. A possible effect of the earphone design was identified in Study 1, which was corroborated by Study 2, where test-retest variability was smallest when comparing results from experiments conducted using identical acoustic transducers.

CONCLUSIONS

Results and test-retest repeatability comparable to standard laboratory settings for the notched-noise test can be obtained with mobile equipment outside the laboratory.

Comparison of Behavioral and Physiological Measures of the Status of the Cochlear Nonlinearity

While an audiogram is a useful method of characterizing hearing loss, it has been suggested that including a complementary, suprathreshold measure, for example, a measure of the status of the cochlear active mechanism, could lead to improved diagnostics and improved hearing-aid fitting in individual listeners. While several behavioral and physiological methods have been proposed to measure the cochlear-nonlinearity characteristics, evidence of a good correspondence between them is lacking, at least in the case of hearing-impaired listeners. If this lack of correspondence is due to, for example, limited reliability of one of such measures, it might be a reason for limited evidence of the benefit of measuring peripheral compression. The aim of this study was to investigate the relation between measures of the peripheral-nonlinearity status estimated using two psychoacoustical methods (based on the notched-noise and temporal-masking curve methods) and otoacoustic emissions, on a large sample of hearing-impaired listeners. While the relation between the estimates from the notched-noise and the otoacoustic emissions experiments was found to be stronger than predicted by the audiogram alone, the relations between the two measures and the temporal-masking based measure did not show the same pattern, that is, the variance shared by any of the two measures with the temporal-masking curve-based measure was also shared with the audiogram.

Toward Routine Assessments of Auditory Filter Shape

Purpose A Bayesian adaptive procedure, that is, the quick auditory filter (qAF) procedure, has been shown to improve the efficiency for estimating auditory filter shapes of listeners with normal hearing. The current study evaluates the accuracy and test-retest reliability of the qAF procedure for naïve listeners with a variety of ages and hearing status. Method Fifty listeners who were naïve to psychophysical experiments and exhibit wide ranges of age (19-70 years) and hearing threshold (-5 to 70 dB HL at 2 kHz) were recruited. Their auditory filter shapes were estimated for a 15-dB SL target tone at 2 kHz using both the qAF procedure and the traditional threshold-based procedure. The auditory filter model was defined using 3 parameters: (a) the sharpness of the tip portion of the auditory filter, p; (b) the prominence of the low-frequency tail of the filter, 10log( w); and (c) the listener's efficiency in detection, 10log( K). Results The estimated parameters of the auditory filter model were consistent between 2 qAF runs tested on 2 separate days. The parameter estimates from the 2 qAF runs also agreed well with those estimated using the traditional procedure despite being substantially faster. Across the 3 auditory filter estimates, the dependence of the auditory filter parameters on listener age and hearing threshold was consistent across procedures, as well as consistent with previously published estimates. Conclusions The qAF procedure demonstrates satisfactory test-retest reliability and good agreement to the traditional procedure for listeners with a wide range of ages and with hearing status ranging from normal hearing to moderate hearing impairment.

The role of spectral resolution, working memory, and audibility in explaining variance in susceptibility to temporal envelope distortion

BACKGROUND

Several studies have shown that hearing thresholds alone cannot adequately predict listeners' success with hearing-aid amplification. Furthermore, previous studies have shown marked differences in listeners' susceptibility to distortions introduced by certain nonlinear amplification parameters.

PURPOSE

The purpose of this study was to examine the role of spectral resolution, working memory, and audibility in explaining perceptual susceptibility to temporal envelope and other hearing-aid compression-induced distortions for listeners with mild to moderate and moderate to severe hearing loss.

RESEARCH DESIGN

A between-subjects repeated-measures design was used to compare speech recognition scores with linear versus compression amplification, for listeners with mild to moderate and moderate to severe hearing loss.

STUDY SAMPLE

The study included 15 adult listeners with mild to moderate hearing loss and 13 adults with moderate to severe hearing loss.

DATA COLLECTION/ANALYSIS

Speech recognition scores were measured for vowel-consonant-vowel syllables processed with linear, moderate compression, and extreme compression amplification. Perceptual susceptibility to compression-induced temporal envelope distortion was defined as the difference in scores between linear and compression amplification. Both overall scores and consonant feature scores (i.e., place, manner, and voicing) were analyzed. Narrowband spectral resolution was measured using individual measures of auditory filter bandwidth at 2000 Hz. Working memory was measured using the reading span test. Signal audibility was quantified using the Aided Audibility Index. Multiple linear regression was used to determine the predictive role of spectral resolution, working memory, and audibility benefit on listeners' susceptibility to compression-induced distortions.

RESULTS

For all listeners, spectral resolution, working memory, and audibility benefit were significant predictors of overall distortion scores. For listeners with moderate to severe hearing loss, spectral resolution and audibility benefit predicted distortion scores for consonant place and manner of articulation features, and audibility benefit predicted distortion scores for consonant voicing features. For listeners with mild to moderate hearing loss, the model did not predict distortion scores for overall or consonant feature scores.

CONCLUSIONS

The results from this study suggest that when audibility is adequately controlled, measures of spectral resolution may identify the listeners who are most susceptible to compression-induced distortions. Working memory appears to modulate the negative effect of these distortions for listeners with moderate to severe hearing loss.

The role of compression in the simultaneous masker phase effect

Peripheral compression is believed to play a major role in the masker phase effect (MPE). While compression is almost instantaneous, activation of the efferent system reduces compression in a temporally evolving manner. To study the role of efferent-controlled compression in the MPE, in experiment 1, simultaneous masking of a 30-ms 4-kHz tone by 40-ms Schroeder-phase harmonic complexes was measured with on- and off-frequency precursors as a function of masker phase curvature for two masker levels (60 and 90 dB sound pressure level). The MPE was quantified by the threshold range [min/max difference (MMD)] across the phase curvatures. For the 60-dB condition, the presence of on-frequency precursor decreased the MMD from 10 to 5 dB. Experiment 2 studied the role of the precursor on the auditory filter's bandwidth. The on-frequency precursor was found to increase the bandwidth, an effect incorporated in the subsequent modeling. A model of the auditory periphery including cochlear filtering and basilar membrane compression generally underestimated the MMDs. A model based on two-step compression, including compression of inner hair cells, accounted for the MMDs across precursor and level conditions. Overall, the observed precursor effects and the model predictions suggest an important role of compression in the simultaneous MPE.

Can place-specific cochlear dispersion be represented by auditory steady-state responses?

The present study investigated to what extent properties of local cochlear dispersion can be objectively assessed through auditory steady-state responses (ASSR). The hypothesis was that stimuli compensating for the phase response at a particular cochlear location generate a maximally modulated basilar membrane (BM) response at that BM position, due to the large "within-channel" synchrony of activity. This would lead, in turn, to a larger ASSR amplitude than other stimuli of corresponding intensity and bandwidth. Two stimulus types were chosen: 1] Harmonic tone complexes consisting of equal-amplitude tones with a starting phase following an algorithm developed by Schroeder [IEEE Trans. Inf. Theory 16, 85-89 (1970)] that have earlier been considered in behavioral studies to estimate human auditory filter phase responses; and 2] simulations of auditory-filter impulse responses (IR). In both cases, also the temporally reversed versions of the stimuli were considered. The ASSRs obtained with the Schroeder tone complexes were found to be dominated by "across-channel" synchrony and, thus, do not reflect local place-specific information. In the case of the more frequency-specific stimuli, no significant differences were found between the responses to the IR and its temporally reversed counterpart. Thus, whereas ASSRs to narrowband stimuli have been used as an objective indicator of frequency-specific hearing sensitivity, the method does not seem to be sensitive enough to reflect local cochlear dispersion.

Behavioural estimates of auditory filter widths in ferrets using notched-noise maskers

Frequency selectivity is a fundamental property of hearing which affects almost all aspects of auditory processing. Here auditory filter widths at 1, 3, 7, and 10 kHz were estimated from behavioural thresholds using the notched-noise method [Patterson, Nimmo-Smith, Weber, and Milroy, J. Acoust. Soc. Am. 72, 1788-1803 (1982)] in ferrets. The mean bandwidth was 21% of the signal frequency, excluding wider bandwidths at 1 kHz (65%). They were comparable although on average broader than equivalent measurements in other mammals (∼11%-20%), and wider than bandwidths measured from the auditory nerve in ferrets (∼18%). In non-human mammals there is considerable variation between individuals, species, and in the correspondence with auditory nerve tuning.

Parcellation of Human and Monkey Core Auditory Cortex with fMRI Pattern Classification and Objective Detection of Tonotopic Gradient Reversals

Auditory cortex (AC) contains several primary-like, or "core," fields, which receive thalamic input and project to non-primary "belt" fields. In humans, the organization and layout of core and belt auditory fields are still poorly understood, and most auditory neuroimaging studies rely on macroanatomical criteria, rather than functional localization of distinct fields. A myeloarchitectonic method has been suggested recently for distinguishing between core and belt fields in humans (Dick F, Tierney AT, Lutti A, Josephs O, Sereno MI, Weiskopf N. 2012. In vivo functional and myeloarchitectonic mapping of human primary auditory areas. J Neurosci. 32:16095-16105). We propose a marker for core AC based directly on functional magnetic resonance imaging (fMRI) data and pattern classification. We show that a portion of AC in Heschl's gyrus classifies sound frequency more accurately than other regions in AC. Using fMRI data from macaques, we validate that the region where frequency classification performance is significantly above chance overlaps core auditory fields, predominantly A1. Within this region, we measure tonotopic gradients and estimate the locations of the human homologues of the core auditory subfields A1 and R. Our results provide a functional rather than anatomical localizer for core AC. We posit that inter-individual variability in the layout of core AC might explain disagreements between results from previous neuroimaging and cytological studies.

Comparing auditory filter bandwidths, spectral ripple modulation detection, spectral ripple discrimination, and speech recognition: Normal and impaired hearing

Some listeners with hearing loss show poor speech recognition scores in spite of using amplification that optimizes audibility. Beyond audibility, studies have suggested that suprathreshold abilities such as spectral and temporal processing may explain differences in amplified speech recognition scores. A variety of different methods has been used to measure spectral processing. However, the relationship between spectral processing and speech recognition is still inconclusive. This study evaluated the relationship between spectral processing and speech recognition in listeners with normal hearing and with hearing loss. Narrowband spectral resolution was assessed using auditory filter bandwidths estimated from simultaneous notched-noise masking. Broadband spectral processing was measured using the spectral ripple discrimination (SRD) task and the spectral ripple depth detection (SMD) task. Three different measures were used to assess unamplified and amplified speech recognition in quiet and noise. Stepwise multiple linear regression revealed that SMD at 2.0 cycles per octave (cpo) significantly predicted speech scores for amplified and unamplified speech in quiet and noise. Commonality analyses revealed that SMD at 2.0 cpo combined with SRD and equivalent rectangular bandwidth measures to explain most of the variance captured by the regression model. Results suggest that SMD and SRD may be promising clinical tools for diagnostic evaluation and predicting amplification outcomes.

Frequency Selectivity of Voxel-by-Voxel Functional Connectivity in Human Auditory Cortex

While functional connectivity in the human cortex has been increasingly studied, its relationship to cortical representation of sensory features has not been documented as much. We used functional magnetic resonance imaging to demonstrate that voxel-by-voxel intrinsic functional connectivity (FC) is selective to frequency preference of voxels in the human auditory cortex. Thus, FC was significantly higher for voxels with similar frequency tuning than for voxels with dissimilar tuning functions. Frequency-selective FC, measured via the correlation of residual hemodynamic activity, was not explained by generic FC that is dependent on spatial distance over the cortex. This pattern remained even when FC was computed using residual activity taken from resting epochs. Further analysis showed that voxels in the core fields in the right hemisphere have a higher frequency selectivity in within-area FC than their counterpart in the left hemisphere, or than in the noncore-fields in the same hemisphere. Frequency-selective FC is consistent with previous findings of topographically organized FC in the human visual and motor cortices. The high degree of frequency selectivity in the right core area is in line with findings and theoretical proposals regarding the asymmetry of human auditory cortex for spectral processing.

Bayesian adaptive estimation of the auditory filter

A Bayesian adaptive procedure for estimating the auditory-filter shape was proposed and evaluated using young, normal-hearing listeners at moderate stimulus levels. The resulting quick-auditory-filter (qAF) procedure assumed the power spectrum model of masking with the auditory-filter shape being modeled using a spectrally symmetric, two-parameter rounded-exponential (roex) function. During data collection using the qAF procedure, listeners detected the presence of a pure-tone signal presented in the spectral notch of a noise masker. Dependent on the listener's response on each trial, the posterior probability distributions of the model parameters were updated, and the resulting parameter estimates were then used to optimize the choice of stimulus parameters for the subsequent trials. Results showed that the qAF procedure gave similar parameter estimates to the traditional threshold-based procedure in many cases and was able to reasonably predict the masked signal thresholds. Additional measurements suggested that occasional failures of the qAF procedure to reliably converge could be a consequence of incorrect responses early in a qAF track. The addition of a parameter describing lapses of attention reduced the likelihood of such failures.

Spectrotemporal modulation sensitivity as a predictor of speech intelligibility for hearing-impaired listeners

BACKGROUND

A model that can accurately predict speech intelligibility for a given hearing-impaired (HI) listener would be an important tool for hearing-aid fitting or hearing-aid algorithm development. Existing speech-intelligibility models do not incorporate variability in suprathreshold deficits that are not well predicted by classical audiometric measures. One possible approach to the incorporation of such deficits is to base intelligibility predictions on sensitivity to simultaneously spectrally and temporally modulated signals.

PURPOSE

The likelihood of success of this approach was evaluated by comparing estimates of spectrotemporal modulation (STM) sensitivity to speech intelligibility and to psychoacoustic estimates of frequency selectivity and temporal fine-structure (TFS) sensitivity across a group of HI listeners.

RESEARCH DESIGN

The minimum modulation depth required to detect STM applied to an 86 dB SPL four-octave noise carrier was measured for combinations of temporal modulation rate (4, 12, or 32 Hz) and spectral modulation density (0.5, 1, 2, or 4 cycles/octave). STM sensitivity estimates for individual HI listeners were compared to estimates of frequency selectivity (measured using the notched-noise method at 500, 1000, 2000, and 4000 Hz), TFS processing ability (2 Hz frequency-modulation detection thresholds for 500, 1000, 2000, and 4000 Hz carriers) and sentence intelligibility in noise (at a 0 dB signal-to-noise ratio) that were measured for the same listeners in a separate study.

STUDY SAMPLE

Eight normal-hearing (NH) listeners and 12 listeners with a diagnosis of bilateral sensorineural hearing loss participated.

DATA COLLECTION AND ANALYSIS

STM sensitivity was compared between NH and HI listener groups using a repeated-measures analysis of variance. A stepwise regression analysis compared STM sensitivity for individual HI listeners to audiometric thresholds, age, and measures of frequency selectivity and TFS processing ability. A second stepwise regression analysis compared speech intelligibility to STM sensitivity and the audiogram-based Speech Intelligibility Index.

RESULTS

STM detection thresholds were elevated for the HI listeners, but only for low rates and high densities. STM sensitivity for individual HI listeners was well predicted by a combination of estimates of frequency selectivity at 4000 Hz and TFS sensitivity at 500 Hz but was unrelated to audiometric thresholds. STM sensitivity accounted for an additional 40% of the variance in speech intelligibility beyond the 40% accounted for by the audibility-based Speech Intelligibility Index.

CONCLUSIONS

Impaired STM sensitivity likely results from a combination of a reduced ability to resolve spectral peaks and a reduced ability to use TFS information to follow spectral-peak movements. Combining STM sensitivity estimates with audiometric threshold measures for individual HI listeners provided a more accurate prediction of speech intelligibility than audiometric measures alone. These results suggest a significant likelihood of success for an STM-based model of speech intelligibility for HI listeners.

On the controversy about the sharpness of human cochlear tuning

In signal processing terms, the operation of the mammalian cochlea in the inner ear may be likened to a bank of filters. Based on otoacoustic emission evidence, it has been recently claimed that cochlear tuning is sharper for human than for other mammals. The claim was corroborated with a behavioral method that involves the masking of pure tones with forward notched noises (NN). Using this method, it has been further claimed that human cochlear tuning is sharper than suggested by earlier behavioral studies. These claims are controversial. Here, we contribute to the controversy by theoretically assessing the accuracy of the NN method at inferring the bandwidth (BW) of nonlinear cochlear filters. Behavioral forward masking was mimicked using a computer model of the squared basilar membrane response followed by a temporal integrator. Isoresponse and isolevel versions of the forward masking NN method were applied to infer the already known BW of the cochlear filter used in the model. We show that isolevel methods were overall more accurate than isoresponse methods. We also show that BWs for NNs and sinusoids equate only for isolevel methods and when the levels of the two stimuli are appropriately scaled. Lastly, we show that the inferred BW depends on the method version (isolevel BW was twice as broad as isoresponse BW at 40 dB SPL) and on the stimulus level (isoresponse and isolevel BW decreased and increased, respectively, with increasing level over the level range where cochlear responses went from linear to compressive). We suggest that the latter may contribute to explaining the reported differences in cochlear tuning across behavioral studies and species. We further suggest that given the well-established nonlinear nature of cochlear responses, even greater care must be exercised when using a single BW value to describe and compare cochlear tuning.

Basilar membrane responses to tones and tone complexes: nonlinear effects of stimulus intensity

The mammalian inner ear combines spectral analysis of sound with multiband dynamic compression. Cochlear mechanics has mainly been studied using single-tone and tone-pair stimulation. Most natural sounds, however, have wideband spectra. Because the cochlea is strongly nonlinear, wideband responses cannot be predicted by simply adding single-tone responses. We measured responses of the gerbil basilar membrane to single-tone and wideband stimuli and compared them, while focusing on nonlinear aspects of the response. In agreement with previous work, we found that frequency selectivity and its dependence on stimulus intensity were very similar between single-tone and wideband responses. The main difference was a constant shift in effective sound intensity, which was well predicted by a simple gain control scheme. We found expansive nonlinearities in low-frequency responses, which, with increasing frequency, gradually turned into the more familiar compressive nonlinearities. The overall power of distortion products was at least 13 dB below the overall power of the linear response, but in a limited band above the characteristic frequency, the power of distortion products often exceeded the linear response. Our results explain the partial success of a "quasilinear" description of wideband basilar membrane responses, but also indicate its limitations.

Auditory-filter characteristics for listeners with real and simulated hearing impairment

Functional simulation of sensorineural hearing impairment is an important research tool that can elucidate the nature of hearing impairments and suggest or eliminate compensatory signal-processing schemes. The objective of the current study was to evaluate the capability of an audibility-based functional simulation of hearing loss to reproduce the auditory-filter characteristics of listeners with sensorineural hearing loss. The hearing-loss simulation used either threshold-elevating noise alone or a combination of threshold-elevating noise and multiband expansion to reproduce the audibility-based characteristics of the loss (including detection thresholds, dynamic range, and loudness recruitment). The hearing losses of 10 listeners with bilateral, mild-to-severe hearing loss were simulated in 10 corresponding groups of 3 age-matched normal-hearing listeners. Frequency selectivity was measured using a notched-noise masking paradigm at five probe frequencies in the range of 250 to 4000 Hz with a fixed probe level of either 70 dB SPL or 8 dB SL (whichever was greater) and probe duration of 200 ms. The hearing-loss simulation reproduced the absolute thresholds of individual hearing-impaired listeners with an average root-mean-squared (RMS) difference of 2.2 dB and the notched-noise masked thresholds with an RMS difference of 5.6 dB. A rounded-exponential model of the notched-noise data was used to estimate equivalent rectangular bandwidths and slopes of the auditory filters. For some subjects and probe frequencies, the simulations were accurate in reproducing the auditory-filter characteristics of the hearing-impaired listeners. In other cases, however, the simulations underestimated the magnitude of the auditory bandwidths for the hearing-impaired listeners, which suggests the possibility of suprathreshold deficits.

Cascades of two-pole-two-zero asymmetric resonators are good models of peripheral auditory function

A cascade of two-pole-two-zero filter stages is a good model of the auditory periphery in two distinct ways. First, in the form of the pole-zero filter cascade, it acts as an auditory filter model that provides an excellent fit to data on human detection of tones in masking noise, with fewer fitting parameters than previously reported filter models such as the roex and gammachirp models. Second, when extended to the form of the cascade of asymmetric resonators with fast-acting compression, it serves as an efficient front-end filterbank for machine-hearing applications, including dynamic nonlinear effects such as fast wide-dynamic-range compression. In their underlying linear approximations, these filters are described by their poles and zeros, that is, by rational transfer functions, which makes them simple to implement in analog or digital domains. Other advantages in these models derive from the close connection of the filter-cascade architecture to wave propagation in the cochlea. These models also reflect the automatic-gain-control function of the auditory system and can maintain approximately constant impulse-response zero-crossing times as the level-dependent parameters change.

Vowel identification by listeners with hearing impairment in response to variation in formant frequencies

PURPOSE

This study examined the influence of presentation level and mild-to-moderate hearing loss on the identification of a set of vowel tokens systematically varying in the frequency locations of their second and third formants.

METHOD

Five listeners with normal hearing (NH listeners) and five listeners with hearing impairment (HI listeners) identified synthesized vowels that represented both highly identifiable and ambiguous examples of /i/, /[Please see symbol]/, and /[Please see symbol]/.

RESULTS

Response patterns of NH listeners showed significant changes, with an increase in presentation level from 75 dB SPL to 95 dB SPL, including increased category overlap. HI listeners, listening only at the higher level, showed greater category overlap than normal and overall identification patterns that differed significantly from those of NH listeners. Excitation patterns based on estimates of auditory filters suggested smoothing of the internal representations, resulting in impaired formant resolution.

CONCLUSIONS

Both increased presentation level for NH listeners and the presence of hearing loss produced a significant change in vowel identification for this stimulus set. Major differences were observed between NH listeners and HI listeners in vowel category overlap and in the sharpness of boundaries between vowel tokens. It is likely that these findings reflect imprecise internal spectral representations due to reduced frequency selectivity.

Auditory filter shapes and high-frequency hearing in adults who have impaired speech in noise performance despite clinically normal audiograms

Some individuals complain of hearing difficulties in the presence of background noise even in the absence of clinically significant hearing loss (obscure auditory dysfunction). Previous evidence suggests that these listeners have impaired frequency resolution, but there has been no thorough characterization of auditory filter shapes in this population. Here, the filter shapes of adults (n = 14) who self-reported speech recognition problems in noise and performed poorly on a sentence-in-noise perception test despite having clinically normal audiograms were compared to those of controls (n = 10). The filter shapes were evaluated using a 2-kHz probe with a fixed level of 30, 40, or 50 dB sound pressure level (SPL) and notched-noise simultaneous maskers that were varied in level to determine the masker level necessary to just mask the probe. The filters of the impaired group were significantly wider than those of controls at all probe levels owing to an unusual broadening of the upper slope of the filter. In addition, absolute thresholds were statistically indistinguishable between the groups at the standard audiometric frequencies, but were elevated in the impaired listeners at higher frequencies. These results strengthen the idea that this population has a variety of hearing deficits that go undetected by standard audiometry.

Isoresponse versus isoinput estimates of cochlear filter tuning

The tuning of a linear filter may be inferred from the filter's isoresponse (e.g., tuning curves) or isoinput (e.g., isolevel curves) characteristics. This paper provides a theoretical demonstration that for nonlinear filters with compressive response characteristics like those of the basilar membrane, isoresponse measures can suggest strikingly sharper tuning than isoinput measures. The practical significance of this phenomenon is demonstrated by inferring the 3-dB-down bandwidths (BW(3dB)) of human auditory filters at 500 and 4,000 Hz from behavioral isoresponse and isoinput measures obtained with sinusoidal and notched noise forward maskers. Inferred cochlear responses were compressive for the two types of maskers. Consistent with expectations, low-level BW(3dB) estimates obtained from isoresponse conditions were considerably narrower than those obtained from isolevel conditions: 69 vs. 174 Hz, respectively, at 500 Hz, and 280 vs. 464 Hz, respectively, at 4,000 Hz. Furthermore, isoresponse BW(3dB) decreased with increasing level while corresponding isolevel estimates remained approximately constant at 500 Hz or increased slightly at 4 kHz. It is suggested that comparisons between isoresponse supra-threshold human tuning and threshold animal neural tuning should be made with caution.

Diagnosing cochlear dead regions in children

OBJECTIVE

A dead region (DR) is defined as a region in the cochlea where inner hair cells and/or neurons are functioning so poorly that a tone producing peak vibration in this region is detected by off-frequency listening, i.e., via a place on the basilar membrane with a characteristic frequency different from that of the tone. The presence of a DR can have a significant effect on the perception of speech. People with and without DRs may differ in the benefit obtained from amplification and require different hearing aid settings. The Threshold Equalizing Noise (TEN) test and psychophysical tuning curves (PTCs) are two procedures used to identify a DR in adults. Because diagnosing a DR involves measuring masked thresholds, and there are reports in the literature that young children perform poorly compared with adults in background noise, it may be possible that the criteria used with adults may not be appropriate when testing children. Therefore, the aim of this study was to evaluate the consistency of the fast-PTC and TEN tests in diagnosing a DR in hearing-impaired children. In addition, the masked thresholds for normal-hearing children were measured with different TEN levels to assess whether any age-related effect in children compared with adults may occur.

DESIGN

Participants were divided into two groups: eight normal-hearing children (16 ears) and 12 hearing-impaired children (21 ears), aged 7 to 13 yr. TEN is based on measuring masked threshold in TEN. For normal-hearing participants, the masked thresholds were measured for five levels of noise (30, 40, 50, 60, and 70 dB per averaged equivalent rectangular bandwidth). For hearing-impaired participants, the level of the TEN was selected separately for each ear based on the highest acceptable level minus 5 dB. The TEN test results in hearing-impaired children were further validated by measuring fast-PTCs. The fast-PTC technique involves measuring the level of the narrowband noise masker needed to mask the signal. The center frequency of the masker sweeps across the required frequency range.

RESULTS

The masked thresholds in TEN measured for normal-hearing children were usually below and never higher than 5 dB above TEN level per averaged equivalent rectangular bandwidth. This suggests that no age-related effect on masked threshold in children compared with adults was observed. All hearing-impaired children were able to perform the TEN test and fast-PTCs. The results of the two tests were consistent in 17 of 21 ears (81%): eight ears did not show evidence of a DR and nine ears did. In three ears, the criteria for a DR were met on the TEN test, but there was no evidence of a DR on the fast-PTC test. In one ear, the TEN test did not show evidence of DRs at two frequencies, whereas fast-PTCs did.

CONCLUSIONS

The results of this study suggest that DRs can be detected in children using the fast-PTC technique and the TEN test interpreted with the adult criteria, which are the most appropriate in terms of specificity and sensitivity. However, in cases in which the masked threshold is 10 to 15 dB above the TEN level, it is recommended to confirm DR diagnosis with fast-PTC measurement.

Analysis and design of gammatone signal models

An established model for the signal analysis performed by the human cochlea is the overcomplete gammatone filterbank. The high correlation of this signal model with human speech and environmental sounds [E. Smith and M. Lewicki, Nature (London) 439, 978-982 (2006)], combined with the increased time-frequency resolution of sparse overcomplete signal models, makes the overcomplete gammatone signal model favorable for signal processing applications on natural sounds. In this paper a signal-theoretic analysis of overcomplete gammatone signal models using the theory of frames and performing bifrequency analyses is given. For the number of gammatone filters M> or =100 (2.4 filters per equivalent rectangular bandwidth), a near-perfect reconstruction can be achieved for the signal space of natural sounds. For signal processing applications like multi-rate coding, a signal-to-alias ratio can be used to derive decimation factors with minimal aliasing distortions.

On- and off-frequency forward masking by Schroeder-phase complexes

Forward masking by harmonic tone complexes was measured for on- and off-frequency maskers as a function of masker phase curvature for two masker durations (30 and 200 ms). For the lowest signal frequency (1 kHz), the results matched predictions based on the expected interactions between the phase curvature and amplitude compression of peripheral auditory filtering. For the higher signal frequencies (2 and 6 kHz), the data increasingly departed from predictions in two respects. First, the effects of the masker phase curvature became stronger with increasing masker duration, inconsistent with the expected effects of the fast-acting compression and time-invariant phase response of basilar membrane filtering. Second, significant effects of masker phase curvature were observed for the off-frequency masker using a 6-kHz signal, inconsistent with predictions based on linear processing of stimuli well below the signal frequency. New predictions were generated assuming an additional effect with a longer time constant, consistent with the influence of medial olivocochlear efferent activation on otoacoustic emissions in humans. Reasonable agreement between the predicted and the measured effects suggests that efferent activation is a potential candidate mechanism to explain certain spectro-temporal masking effects in human hearing.

Auditory responses in the barn owl's nucleus laminaris to clicks: impulse response and signal analysis of neurophonic potential

We used acoustic clicks to study the impulse response of the neurophonic potential in the barn owl's nucleus laminaris. Clicks evoked a complex oscillatory neural response with a component that reflected the best frequency measured with tonal stimuli. The envelope of this component was obtained from the analytic signal created using the Hilbert transform. The time courses of the envelope and carrier waveforms were characterized by fitting them with filters. The envelope was better fitted with a Gaussian than with the envelope of a gamma-tone function. The carrier was better fitted with a frequency glide than with a constant instantaneous frequency. The change of the instantaneous frequency with time was better fitted with a linear fit than with a saturating nonlinearity. Frequency glides had not been observed in the bird's auditory system before. The glides were similar to those observed in the mammalian auditory nerve. Response amplitude, group delay, frequency, and phase depended in a systematic way on click level. In most cases, response amplitude decreased linearly as stimulus level decreased, while group delay, phase, and frequency increased linearly as level decreased. Thus the impulse response of the neurophonic potential in the nucleus laminaris of barn owls reflects many characteristics also observed in responses of the basilar membrane and auditory nerve in mammals.

The relationship between frequency selectivity and pitch discrimination: sensorineural hearing loss

This study tested the relationship between frequency selectivity and the minimum spacing between harmonics necessary for accurate fo discrimination. Fundamental frequency difference limens (fo DLs) were measured for ten listeners with moderate sensorineural hearing loss (SNHL) and three normal-hearing listeners for sine- and random-phase harmonic complexes, bandpass filtered between 1500 and 3500 Hz, with fo's ranging from 75 to 500 Hz (or higher). All listeners showed a transition between small (good) fo DLs at high fo's and large (poor) fo DLs at low fo's, although the fo at which this transition occurred (fo,tr) varied across listeners. Three measures thought to reflect frequency selectivity were significantly correlated to both the fo,tr and the minimum fo DL achieved at high fo's: (1) the maximum fo for which fo DLs were phase dependent, (2) the maximum modulation frequency for which amplitude modulation and quasi-frequency modulation were discriminable, and (3) the equivalent rectangular bandwidth of the auditory filter, estimated using the notched-noise method. These results provide evidence of a relationship between fo discrimination performance and frequency selectivity in listeners with SNHL, supporting "spectral" and "spectro-temporal" theories of pitch perception that rely on sharp tuning in the auditory periphery to accurately extract fo information.

The relationship between frequency selectivity and pitch discrimination: effects of stimulus level

Three experiments tested the hypothesis that fundamental frequency (fo) discrimination depends on the resolvability of harmonics within a tone complex. Fundamental frequency difference limens (fo DLs) were measured for random-phase harmonic complexes with eight fo's between 75 and 400 Hz, bandpass filtered between 1.5 and 3.5 kHz, and presented at 12.5-dB/component average sensation level in threshold equalizing noise with levels of 10, 40, and 65 dB SPL per equivalent rectangular auditory filter bandwidth. With increasing level, the transition from large (poor) to small (good) fo DLs shifted to a higher fo. This shift corresponded to a decrease in harmonic resolvability, as estimated in the same listeners with excitation patterns derived from measures of auditory filter shape and with a more direct measure that involved hearing out individual harmonics. The results are consistent with the idea that resolved harmonics are necessary for good fo discrimination. Additionally, fo DLs for high fo's increased with stimulus level in the same way as pure-tone frequency DLs, suggesting that for this frequency range, the frequencies of harmonics are more poorly encoded at higher levels, even when harmonics are well resolved.

Comparison of the roex and gammachirp filters as representations of the auditory filter

Although the rounded-exponential (roex) filter has been successfully used to represent the magnitude response of the auditory filter, recent studies with the roex(p, w, t) filter reveal two serious problems: the fits to notched-noise masking data are somewhat unstable unless the filter is reduced to a physically unrealizable form, and there is no time-domain version of the roex(p, w, t) filter to support modeling of the perception of complex sounds. This paper describes a compressive gammachirp (cGC) filter with the same architecture as the roex(p, w, t) which can be implemented in the time domain. The gain and asymmetry of this parallel cGC filter are shown to be comparable to those of the roex(p, w, t) filter, but the fits to masking data are still somewhat unstable. The roex(p, w, t) and parallel cGC filters were also compared with the cascade cGC filter [Patterson et al., J. Acoust. Soc. Am. 114, 1529-1542 (2003)], which was found to provide an equivalent fit with 25% fewer coefficients. Moreover, the fits were stable. The advantage of the cascade cGC filter appears to derive from its parsimonious representation of the high-frequency side of the filter. It is concluded that cGC filters offer better prospects than roex filters for the representation of the auditory filter.

Level dependence of auditory filters in nonsimultaneous masking as a function of frequency

Auditory filter bandwidths were measured using nonsimultaneous masking, as a function of signal level between 10 and 35 dB SL for signal frequencies of 1, 2, 4, and 6 kHz. The brief sinusoidal signal was presented in a temporal gap within a spectrally notched noise. Two groups of normal-hearing subjects were tested, one using a fixed masker level and adaptively varying signal level, the other using a fixed signal level and adaptively varying masker level. In both cases, auditory filters were derived by assuming a constant filter shape for a given signal level. The filter parameters derived from the two paradigms were not significantly different. At 1 kHz, the equivalent rectangular bandwidth (ERB) decreased as the signal level increased from 10 to 20 dB SL, after which it remained roughly constant. In contrast, at 6 kHz, the ERB increased consistently with signal levels from 10 to 35 dB SL. The results at 2 and 4 kHz were intermediate, showing no consistent change in ERB with signal level. Overall, the results suggest changes in the level dependence of the auditory filters at frequencies above 1 kHz that are not currently incorporated in models of human auditory filter tuning.

Auditory filter nonlinearity across frequency using simultaneous notched-noise masking

Psychoacoustic masking experiments have been widely used to investigate cochlear function in human listeners. Here we use simultaneous notched-noise masking experiments in normal hearing listeners to characterize the changes in auditory filter shape with stimulus level over the frequency range 0.25-6 kHz. At each frequency a range of fixed signal levels (30-70 dB SPL) and fixed masker levels (20-50 dB SPL spectrum level) are used in order to obtain accurate descriptions of the filter shapes in individual listeners. The notched-noise data for individual listeners are fitted with two filter shape models: a rounded exponential (roex) shape in which the filter skirt changes as a linear function of probe-tone level and the other, in which the gain of the tip filter relative to the filter tail changes as a function of signal level [Glasberg and Moore, J. Acoust. Soc. Am. 108, 2318-2328 (2000)]. The parameters for these fitted models are then described with a simple set of equations that quantify the changes in auditory filter shape across level and frequency. Both these models fitted the data equally well and both demonstrated increasing tip-tail gain as frequency increased.

Psychophysical estimates of nonlinear cochlear processing in younger and older listeners

The primary goal of this project was to compare the performance of younger and older listeners on a number of psychophysical measures thought to be influenced by nonlinear cochlear processing. Younger (mean of 25.6 years) and older (mean of 63.8 years) listeners with normal hearing were matched (within 5 dB) according to their quiet thresholds at the two test frequencies of 1200 and 2400 Hz. They were similarly matched at the adjacent octave frequencies of 600 and 4800 Hz (within 5 dB at one and 9 dB at the other). Performance was compared on measures of auditory filter shape, psychophysical suppression, and growth of forward masking. There was no difference between the two age groups on these psychophysical estimates reflecting nonlinear processing, suggesting that aging per se does not affect the cochlear nonlinearity, at least for the ages sampled here. The results did, however, consistently demonstrate an age-related increase in the susceptibility to forward masking.

Perception of dissonance by people with normal hearing and sensorineural hearing loss

The purpose of this study was to determine whether the perceived sensory dissonance of pairs of pure tones (PT dyads) or pairs of harmonic complex tones (HC dyads) is altered due to sensorineural hearing loss. Four normal-hearing (NH) and four hearing-impaired (HI) listeners judged the sensory dissonance of PT dyads geometrically centered at 500 and 2000 Hz, and of HC dyads with fundamental frequencies geometrically centered at 500 Hz. The frequency separation of the members of the dyads varied from 0 Hz to just over an octave. In addition, frequency selectivity was assessed at 500 and 2000 Hz for each listener. Maximum dissonance was perceived at frequency separations smaller than the auditory filter bandwidth for both groups of listners, but maximum dissonance for HI listeners occurred at a greater proportion of their bandwidths at 500 Hz than at 2000 Hz. Further, their auditory filter bandwidths at 500 Hz were significantly wider than those of the NH listeners. For both the PT and HC dyads, curves displaying dissonance as a function of frequency separation were more compressed for the HI listeners, possibly reflecting less contrast between their perceptions of consonance and dissonance compared with the NH listeners.

A population study of the precedence effect

Data are reported from a population of untrained individuals under lag- and single-click conditions in a discrimination suppression precedence-effect task. The cue to be discriminated was an interaural level-difference (ILD). Each of 91 observers completed 10 runs in a two-interval forced-choice design under a lag-click condition and three runs under a single-click condition. Stimuli were 125-micros rectangular pulses and the interclick interval was 2 ms. Observers were randomly assigned to three groups of approximately 30. Each group was then tested at one stimulus intensity (43, 58, or 73 dB). Mean threshold within each group was greater than 15 dB for the lag-click condition and 6 dB for the single-click condition, although there was substantial interobserver variability. In contrast to [J. Acoust. Soc. Am. 114 (2003) 420] who reported a strong effect of intensity on lag-click ITD discrimination, no effect of intensity was observed on lag-click ILD thresholds. Analysis of over 50,000 near-threshold trials from 302 observers pooled across studies showed a spatial asymmetry in response patterns and a small, but statistically significant effect of gender. A model is proposed which shows that decay of sensory memory and increases in auditory filter bandwidths with intensity may predict the different findings for ILD versus ITD lag-click thresholds.

Extending the domain of center frequencies for the compressive gammachirp auditory filter

The gammatone filter was imported from auditory physiology to provide a time-domain version of the roex auditory filter and enable the development of a realistic auditory filterbank for models of auditory perception [Patterson et al., J. Acoust. Soc. Am. 98, 1890-1894 (1995)]. The gammachirp auditory filter was developed to extend the domain of the gammatone auditory filter and simulate the changes in filter shape that occur with changes in stimulus level. Initially, the gammachirp filter was limited to center frequencies in the 2.0-kHz region where there were sufficient "notched-noise" masking data to define its parameters accurately. Recently, however, the range of the masking data has been extended in two massive studies. This paper reports how a compressive version of the gammachirp auditory filter was fitted to these new data sets to define the filter parameters over the extended frequency range. The results show that the shape of the filter can be specified for the entire domain of the data using just six constants (center frequencies from 0.25 to 6.0 kHz and levels from 30 to 80 dB SPL). The compressive, gammachirp auditory filter also has the advantage of being consistent with physiological studies of cochlear filtering insofar as the compression of the filter is mainly limited to the passband and the form of the chirp in the impulse response is largely independent of level.

Estimates of human cochlear tuning at low levels using forward and simultaneous masking

Auditory filter shapes were derived from psychophysical measurements in eight normal-hearing listeners using a variant of the notched-noise method for brief signals in forward and simultaneous masking. Signal frequencies of 1, 2, 4, 6, and 8 kHz were tested. The signal level was fixed at 10 dB above absolute threshold in the forward-masking conditions and fixed at either 10 or 35 dB above absolute threshold in the simultaneous-masking conditions. The results show that filter equivalent rectangular bandwidths (ERBs) are substantially narrower in forward masking than has been found in previous studies using simultaneous masking. Furthermore, in contrast to earlier studies, the sharpness of tuning doubles over the range of frequencies tested, giving Q(ERB) values of about 10 and 20 at signal frequencies of 1 and 8 kHz, respectively. It is argued that the new estimates of auditory filter bandwidth provide a more accurate estimate of human cochlear tuning at low levels than earlier estimates using simultaneous masking at higher levels, and that they are therefore more suitable for comparison to cochlear tuning data from other species. The data may also prove helpful in defining the parameters for nonlinear models of human cochlear processing.

An account of monaural phase sensitivity

Listeners can detect phase differences between the envelopes of sounds occupying remote frequency regions, and between the fine structures of partials that interact within a single auditory filter. They are insensitive to phase differences between partials that differ sufficiently in frequency to preclude within-channel interactions. A new model is proposed that can account for all three of these findings, and which, unlike currently popular approaches, does not discard across-channel timing information. Sensitivity is predicted quantitatively by analyzing the output of a cochlear model using a spectro-temporal decomposition inspired by responses of neurons in the auditory cortex, and by computing a distance metric between the responses to two stimuli to be discriminated. Discriminations successfully modeled include phase differences between pairs of bandpass filtered harmonic complexes, and between pairs of sinusoidally amplitude modulated tones, discrimination between amplitude and frequency modulation, and discrimination of transient signals differing only in their phase spectra ("Huffman sequences").

Effects of signal delay on auditory filter shapes derived from psychophysical tuning curves and notched-noise data obtained in simultaneous masking

Psychophysical tuning curves (PTCs) measured in simultaneous masking usually sharpen as a short duration signal is moved from the onset to the temporal center of a longer duration masker. Filter shapes derived from notched-noise maskers have not consistently shown this effect. One possible explanation for this difference is that the signal level is fixed in the PTC paradigm, whereas the masker level is usually fixed in the notched-noise paradigm. In the present study, the signal level was fixed at 10 dB SL in both paradigms. The signal was 20 ms in duration, and presented at the onset or temporal center of the 400-ms masker. The masker was a pure tone presented in quiet (PTC) or in the presence of a pure-tone "restrictor" intended to limit off-frequency listening (PTCr), or it was a noise with a spectral notch placed symmetrically or asymmetrically about the 2-kHz signal frequency. Filter shapes were derived from the PTC, PTCr, and notched-noise data using the roex (p, w, t) model. The effects of signal delay and masking paradigm on filter bandwidth were analyzed with a two-factor repeated-measures ANOVA. There was a significant effect of signal delay (the filters sharpened with time) and masking paradigm (the filters derived from the notched-noise data were significantly wider than those derived from either of the PTC measurements, which did not differ from one another). Although the interaction between delay and paradigm was not significant, the filter derived from the notched-noise data sharpened more with time than did the other filters, and thus the bandwidth of the filters from the three paradigms were more similar at the longer delay than at the shorter delay. It is likely that the tuning-curve and notched-noise paradigms measure the same underlying filtering, but that various other factors contribute differentially to the derived filter shapes.

Asymmetry of masking between complex tones and noise: the role of temporal structure and peripheral compression

Thresholds for the detection of harmonic complex tones in noise were measured as a function of masker level. The rms level of the masker ranged from 40 to 70 dB SPL in 10-dB steps. The tones had a fundamental frequency (F0) of 62.5 or 250 Hz, and components were added in either cosine or random phase. The complex tones and the noise were bandpass filtered into the same frequency region, from the tenth harmonic up to 5 kHz. In a different condition, the roles of masker and signal were reversed, keeping all other parameters the same; subjects had to detect the noise in the presence of a harmonic tone masker. In both conditions, the masker was either gated synchronously with the 700-ms signal, or it started 400 ms before and stopped 200 ms after the signal. The results showed a large asymmetry in the effectiveness of masking between the tones and noise. Even though signal and masker had the same bandwidth, the noise was a more effective masker than the complex tone. The degree of asymmetry depended on F0, component phase, and the level of the masker. The maximum difference between masked thresholds for tone and noise was about 28 dB; this occurred when the F0 was 62.5 Hz, the components were in cosine phase, and the masker level was 70 dB SPL. In most conditions, the growth-of-masking functions had slopes close to 1 (on a dB versus dB scale). However, for the cosine-phase tone masker with an F0 of 62.5 Hz, a 10-dB increase in masker level led to an increase in masked threshold of the noise of only 3.7 dB, on average. We suggest that the results for this condition are strongly affected by the active mechanism in the cochlea.

Auditory filter nonlinearity in mild/moderate hearing impairment

Sensorineural hearing loss has frequently been shown to result in a loss of frequency selectivity. Less is known about its effects on the level dependence of selectivity that is so prominent a feature of normal hearing. The aim of the present study is to characterize such changes in nonlinearity as manifested in the auditory filter shapes of listeners with mild/moderate hearing impairment. Notched-noise masked thresholds at 2 kHz were measured over a range of stimulus levels in hearing-impaired listeners with losses of 20-50 dB. Growth-of-masking functions for different notch widths are more parallel for hearing-impaired than for normal-hearing listeners, indicating a more linear filter. Level-dependent filter shapes estimated from the data show relatively little change in shape across level. The loss of nonlinearity is also evident in the input/output functions derived from the fitted filter shapes. Reductions in nonlinearity are clearly evident even in a listener with only 20-dB hearing loss.

Quantifying the implications of nonlinear cochlear tuning for auditory-filter estimates

The relation between auditory filters estimated from psychophysical methods and peripheral tuning was evaluated using a computational auditory-nerve (AN) model that included many of the response properties associated with nonlinear cochlear tuning. The phenomenological AN model included the effects of dynamic level-dependent tuning, compression, and suppression on the responses of high-, medium-, and low-spontaneous-rate AN fibers. Signal detection theory was used to evaluate psychophysical performance limits imposed by the random nature of AN discharges and by random-noise stimuli. The power-spectrum model of masking was used to estimate psychophysical auditory filters from predicted AN-model detection thresholds for a tone signal in fixed-level notched-noise maskers. Results demonstrate that the role of suppression in broadening peripheral tuning in response to the noise masker has implications for the interpretation of psychophysical auditory-filter estimates. Specifically, the estimated psychophysical auditory-filter equivalent-rectangular bandwidths (ERBs) that were derived from the nonlinear AN model with suppression always overestimated the ERBs of the low-level peripheral model filters. Further, this effect was larger for an 8-kHz signal than for a 2-kHz signal, suggesting a potential characteristic-frequency (CF) dependent bias in psychophysical estimates of auditory filters due to the increase in strength of cochlear nonlinearity with increases in CF.

Asymmetry of masking between noise and iterated rippled noise: evidence for time-interval processing in the auditory system

This study describes the masking asymmetry between noise and iterated rippled noise (IRN) as a function of spectral region and the IRN delay. Masking asymmetry refers to the fact that noise masks IRN much more effectively than IRN masks noise, even when the stimuli occupy the same spectral region. Detection thresholds for IRN masked by noise and for noise masked by IRN were measured with an adaptive two-alternative, forced choice (2AFC) procedure with signal level as the adaptive parameter. Masker level was randomly varied within a 10-dB range in order to reduce the salience of loudness as a cue for detection. The stimuli were filtered into frequency bands, 2.2-kHz wide, with lower cutoff frequencies ranging from 0.8 to 6.4 kHz. IRN was generated with 16 iterations and with varying delays. The reciprocal of the delay was 16, 32, 64, or 128 Hz. When the reciprocal of the IRN delay was within the pitch range, i.e., above 30 Hz, there was a substantial masking asymmetry between IRN and noise for all filter cutoff frequencies; threshold for IRN masked by noise was about 10 dB larger than threshold for noise masked by IRN. For the 16-Hz IRN, the masking asymmetry decreased progressively with increasing filter cutoff frequency, from about 9 dB for the lowest cutoff frequency to less than 1 dB for the highest cutoff frequency. This suggests that masking asymmetry may be determined by different cues for delays within and below the pitch range. The fact that masking asymmetry exists for conditions that combine very long IRN delays with very high filter cutoff frequencies means that it is unlikely that models based on the excitation patterns of the stimuli would be successful in explaining the threshold data. A range of time-domain models of auditory processing that focus on the time intervals in phase-locked neural activity patterns is reviewed. Most of these models were successful in accounting for the basic masking asymmetry between IRN and noise for conditions within the pitch range, and one of the models produced an exceptionally good fit to the data.

Modeling temporal and compressive properties of the normal and impaired auditory system

Three modifications of a psychoacoustically and physiologically motivated processing model [Dau et al., J. Acoust. Soc. Am. 102 (1997a) 2892-2905] are presented and tested. The modifications aim at simulating sensorineural hearing loss and incorporate a level-dependent peripheral compression whose properties are affected by hearing impairment. Model 1 realizes this difference by introducing for impaired listeners an instantaneous level-dependent expansion prior to the adaptation stage of the model. Model 2 and Model 3 realize a level-dependent compression with time constants of 5 and 15 ms, respectively, for normal hearing and a reduced compression for impaired hearing. In Model 2, the compression occurs after the envelope extraction stage, while in Model 3, envelope extraction follows compression. All models account to a similar extent for the recruitment phenomenon measured with narrow-band stimuli and for forward-masking data of normal-hearing and hearing-impaired subjects using a 20-ms, 2-kHz tone signal and a 1-kHz-wide bandpass noise masker centered at 2 kHz. A clear difference between the different models occurs for the processing of temporally fluctuating stimuli. A modulation-rate-independent increase in modulation-response level for simulating impaired hearing is only predicted by Model 1 while the other two models realize a modulation-rate-dependent increase. Hence, the predictions of Model 2 and Model 3 are in conflict with the results of modulation-matching experiments reported in the literature. It is concluded that key properties of sensorineural hearing loss (altered loudness perception, reduced dynamic range, normal temporal properties but prolonged forward-masking effects) can effectively be modeled by incorporating a fast-acting expansion within the current processing model prior to the nonlinear adaptation stage. Based on these findings, a model of both normal and impaired hearing is proposed which incorporates a fast-acting compressive nonlinearity, representing the cochlear nonlinearity (which is reduced in impaired listeners), followed by an instantaneous expansion and the nonlinear adaptation stage which represent aspects of the retro-cochlear information processing in the auditory system.

Reconciling frequency selectivity and phase effects in masking

The effects of auditory frequency selectivity and phase response on masking were studied using harmonic tone complex maskers with a 100-Hz fundamental frequency. Positive and negative Schroeder-phase complexes (m+ and m-), were used as maskers and the signal was a long-duration sinusoid. In the first experiment, thresholds for signal frequencies of 1 and 4 kHz were measured as a function of masker bandwidth and number of components. A large difference in thresholds between the m+ and m- complexes was found only when masker components were presented ipsilateral to the signal over a frequency range wider than the traditional critical band, regardless of the absolute number of components. In the second experiment, frequency selectivity was measured in harmonic tone complexes with fixed or random phases as well as in noise, using a variant of the notched-noise method with a fixed masker level. The data showed that frequency selectivity is not affected by masker type, indicating that the wide listening bandwidth suggested by the first experiment cannot be ascribed to broader effective filters in complex-tone maskers than in noise maskers. The third experiment employed a novel method of measuring frequency selectivity, which has the advantage that the overall level at the input and the output of the auditory filter remains roughly constant across all conditions. The auditory filter bandwidth measured using this method was wider than that measured in the second experiment, but may still be an underestimate, due to the effects of off-frequency listening. The data were modeled using a single-channel model with various initial filters. The main findings from the simulations were: (1) the magnitude response of the Gammatone filter is too narrow to account for the phase effects observed in the data; (2) none of the other filters currently used in auditory models can account for both frequency selectivity and phase effects in masking; (3) the Gammachirp filter can be made to provide a good account of the data by altering its phase response. The final conclusion suggests that masker phase effects can be accounted for with a single-channel model, while still remaining consistent with measures of frequency selectivity: effects that appear to involve broadband processing do not necessarily require across-channel mechanisms.

A compressive gammachirp auditory filter for both physiological and psychophysical data

A gammachirp auditory filter was developed by Irino and Patterson [J. Acoust. Soc. Am. 101, 412-419 (1997)] to provide a level-dependent version of the linear, gammatone auditory filter, with which to explain the level-dependent changes in cochlear filtering observed in psychophysical masking experiments. In this 'analytical' gammachirp filter, the chirp varied with level and there was no explicit representation of the change in filter gain or compression with level. Subsequently, Carney et al. [J. Acoust. Soc. Am. 105, 2384-2391 (1999)] reviewed Carney and Yin's [J. Neurophysiol. 60, 1653-1677 (1988)] reverse-correlation (revcor) data and showed that the frequency glide of the chirp does not vary with level in their data. In this article, the architecture of the analytical gammachirp is reviewed with respect to cochlear physiology and a new form of gammachirp filter is described in which the magnitude response, the gain, and the compression vary with level but the chirp does not. This new 'compressive' gammachirp filter is used to fit the level-dependent revcor data reported by Carney et al. (1999) and the level-dependent masking data reported by Rosen and Baker [Hear. Res. 73, 231-243 (1994)].

Evaluation of maximum-likelihood threshold estimation with tone-in-noise masking

There has been much recent interest in the use of adaptive psychophysical procedures based on maximum-likelihood estimation (MLE) in order to minimize testing time. The speed and accuracy of MLE was compared to a standard transformed up-down algorithm in a two-interval forced-choice task. Thresholds for detecting a 2 kHz tone in either a broadband or a notched-noise were estimated in three normal-hearing listeners. The transformed up-down algorithm tracked 79% correct with either two, four, six or eight final turnarounds, whereas the MLE procedure tracked 70%, 80% or 90% correct. MLE was always quickest, but with a penalty in increased variability. Use of the MLE procedure to track 70% or 80% correct also resulted in a tendency to overestimate listeners' sensitivity. Reducing the number of turnarounds in the up-down procedure from eight to two reduced the number of trials required by nearly half and resulted in thresholds with similar magnitude and variability to those obtained using MLE to track 90% correct.

The effects of frequency region and level on the temporal modulation transfer function

Temporal modulation transfer functions (TMTFs) were measured using narrow-band AM and QFM noises with upper spectral edges from 0.6 to 4.8 kHz, and spectrum levels of 10 and 40 dB SPL. The cutoff frequency of the TMTF increases as the upper spectral edge is increased up to 4.8 kHz at low levels, and is constant at higher levels. Sensitivity increases with bandwidth if frequency region is constant. In a second experiment, these results were compared to predictions of a model incorporating peripheral and central limitations to modulation detection. To obtain an estimate of peripheral filtering, frequency selectivity was measured using the notched-noise method, with probe frequencies and levels chosen to parallel those in the first experiment. The TMTF data were then predicted using the model. Predicted cutoff frequencies as a function of the upper spectral edge of the test stimulus were lower than but parallel to those of the subjects at the lower stimulus level. The model predicted only a slight increase in cutoff frequency with level, and thus predicted an increase in cutoff frequency with frequency region at the higher level as well, in contrast to the measured data. These results suggest that there are peripheral and central limitations to temporal resolution, but the psychoacoustically derived auditory filter may be only an indirect measure of peripheral filtering, and/or a more complex model may be needed.

Different auditory filter bandwidth estimates based on profile analysis, notched noise, and hybrid tasks

Auditory filter bandwidths were estimated in three experiments. The first experiment was a profile-analysis experiment. The stimuli were composed of sinusoidal components ranging in frequency from 200 to 5000 Hz. The standard stimulus was the sum of equal-amplitude tones, and the signal stimulus had a power spectrum that varied up-down ... up-down. The number of components ranged from four to 60. Interval-by-interval level randomization prevented the change in level of a single component from reliably indicating the change from standard to signal. The second experiment was a notched-noise experiment in which the 1000-Hz tone to be detected was added to a noise with a notch arithmetically centered at 1000 Hz. Detection thresholds were estimated both in the presence of and in the absence of level randomization. In the third, hybrid, experiment a 1000-Hz tone was to be detected, and the masker was composed of equal-amplitude sinusoidal components ranging in frequency from 200 to 5000 Hz. For this experiment, thresholds were estimated both in the presence and absence of level variation. For both the notched-noise and hybrid experiments, only modest effects of level randomization were obtained. A variant of Durlach et al.'s channel model ["Towards a model for discrimination of broadband signals," J. Acoust. Soc. Am. 80, 63-72 (1986)] was used to estimate auditory filter bandwidths for all three experiments. When a two-parameter roex(p,r) filter weighting function was used to fit the data, bandwidth estimates were approximately two to three times as large for the two detection tasks than for the profile-analysis task.