Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Perception of amplitude modulation by hearing-impaired listeners: The audibility of component modulation and detection of phase change in three-component modulators Aleksander Seka兲 Institute of Acoustics, Adam Mickiewicz University, 85 Umultowska, 61-614 Poznan, Poland Brian C. J. Mooreb兲 Department of Experimental Psychology, University of Cambridge, Downing Street, Cambridge CB2 3EB, England 共Received 4 August 2005; revised 24 October 2005; accepted 25 October 2005兲 Two experiments were conducted to assess whether hearing-impaired listeners have a reduced ability to process suprathreshold complex patterns of modulation applied to a 4-kHz sinusoidal carrier. Experiment 1 examined the ability to “hear out” the modulation frequency of the central component of a three-component modulator, using the method described by Sek and Moore 关J. Acoust. Soc. Am. 113, 2801–2811 共2003兲兴. Scores were around 70–80% correct when the components in the three-component modulator were widely spaced and when the frequencies of the target and comparison different sufficiently, but decreased when the components in the modulator were closely spaced. Experiment 2 examined the ability to hear a change in the relative phase of the components in a three-component modulator with harmonically spaced components. The frequency of the central component, f c, was either 50 or 100 Hz. Scores were about 70% correct when the component spacing was 艋0.5f c, but decreased markedly for greater spacings. Performance was only slightly impaired by randomizing the overall modulation depth from one stimulus to the next. For both experiments, performance was only slightly worse than for normally hearing listeners, indicating that cochlear hearing loss does not markedly affect the ability to process suprathreshold complex patterns of modulation. © 2006 Acoustical Society of America. 关DOI: 10.1121/1.2139631兴 PACS number共s兲: 43.66.Mk, 43.66.Nm, 43.66.Sr 关AJO兴 I. INTRODUCTION The perception of amplitude modulation 共AM兲 plays an important role in many aspects of auditory perception, including speech perception 共Plomp, 1983; Shannon et al. 1995兲 and the ability to analyze mixtures of sounds arising from different sources 共Darwin and Carlyon, 1995兲. Hearingimpaired people often have difficulty in understanding speech in the presence of background sounds 共Plomp, 1978, 1994; Moore, 1998兲 and it is possible that part of this difficulty stems from abnormalities in the processing of amplitude modulation. This possibility is assessed in the present experiments. Previous studies of AM perception by hearing-impaired people have mostly focused on the detection of AM 共Formby, 1982; Bacon and Gleitman, 1992; Moore et al., 1992; Moore and Glasberg, 2001兲 or the discrimination of AM rate 共Formby, 1986; Grant, 1998兲. Overall, the results from these studies suggest that performance is not usually adversely affected by cochlear hearing loss, except when a portion of the carrier is inaudible. Here, we focus on the perception of complex patterns of modulation applied to a兲 Electronic mail: [email protected] Author to whom correspondence should be addressed. Electronic mail: [email protected] b兲 J. Acoust. Soc. Am. 119 共1兲, January 2006 Pages: 507–514 a single carrier for modulation depths well above threshold. We start with a review of previous related studies. Moore et al. 共1996兲 tested listeners with unilateral hearing loss. They were presented with an amplitude-modulated sinusoid to one ear, and were asked to adjust the AM depth of a sinusoid presented alternately to the other ear, so as to match the perceived modulation depth. The results indicated that the perceived amount of modulation in the impaired ear was “magnified” compared to that in the normal ear, and it was argued that this was a manifestation of loudness recruitment 共Steinberg and Gardner, 1937兲, probably resulting from the loss of fast-acting compression on the basilar membrane 共Rhode and Robles, 1974, Moore and Glasberg, 1997, 2004兲. Tandetnik et al. 共2001兲 measured second-order modulation detection thresholds in normally hearing and hearingimpaired listeners. Second-order modulation is modulation applied to the modulation depth of an AM signal, i.e., modulation of the modulator. The second-order modulation thresholds were measured as a function of the frequency of the modulation applied to the modulation depth 共referred to as fm ⬘ 兲, using values of f m⬘ from 1 to 11 Hz, and a fixed firstorder modulation frequency of 16 Hz. The audio-frequency carrier was broadband white noise. The thresholds for the hearing-impaired listeners were within the normal range for fm ⬘ = 3 , 5, and 11 Hz, and were higher 共poorer兲 than normal for f m ⬘ = 1 and 7 Hz. These results suggest that cochlear hearing loss may have some deleterious effect on the ability to 0001-4966/2006/119共1兲/507/8/$22.50 © 2006 Acoustical Society of America 507 process complex temporal envelopes, but the effect is not large. In a later study using a 2-kHz sinusoidal carrier, detection thresholds for second-order AM were found to be essentially the same for normally hearing and hearingimpaired listeners 共Füllgrabe et al., 2003兲. Turner et al. 共1995兲 used “noise-vocoder” processing to compare the ability of hearing-impaired and normally hearing listeners to identify speech on the basis of temporal envelope cues in a few spectral regions. They found that performance did not differ significantly for the two groups. However, recent unpublished data obtained in our laboratory indicate that hearing-impaired listeners often perform more poorly than normally hearing listeners when trying to understand “noise-vocoded” speech when background sounds are present. Although most of the studies described above suggest that the ability of hearing-impaired listeners to process amplitude modulation is nearly normal, there is at least one study suggesting that this may not be the case. Lorenzi et al. 共1997兲 measured the ability to detect 100-Hz sinusoidal AM applied to a white noise carrier, as a function of the frequency of a masking sinusoidal AM applied to the same noise carrier. For listeners with normal hearing, the modulation masking patterns had a broad bandpass characteristic with a peak at 100 Hz. One listener with a moderate hearing loss also showed a broadly tuned pattern with a peak at 100 Hz. However, the two other listeners with moderate or mild-to-severe hearing loss showed broader patterns with more of a lowpass characteristic. These results suggest that frequency selectivity in the modulation domain can be reduced in listeners with sensorineural hearing loss. In the present experiments, we examine the ability of hearing-impaired listeners to process complex patterns of suprathreshold modulation imposed on a single sinusoidal carrier, i.e., we examine within-channel modulation processing. The two tasks used 共Sek and Moore, 2003兲 were originally designed to test the concept that the envelopes of the outputs of the auditory filters are fed to a second array of overlapping bandpass filters tuned to different envelope modulation rates. This set of filters is usually called a “modulation filter bank” 共MFB兲 共Dau et al., 1997a, 1997b兲. The main purpose of the present experiments was not to test the concept of the MFB, but rather to use the two tasks as measures of the ability of hearing-impaired listeners to process complex suprathreshold patterns of AM imposed on a single sinusoidal carrier. This allowed us to examine modulation processing while avoiding possible confounding effects of reduced frequency selectivity in hearing-impaired listeners. It also allowed us to avoid confounding effects of variations in audibility with frequency, which can occur when a broadband carrier is used. II. EXPERIMENT 1: “HEARING OUT” COMPONENT MODULATION In our first experiment, we assessed the extent to which hearing-impaired listeners can “hear out” the component modulations in a complex modulator, in the same way that spectral components in a complex sound can be heard out in the audio-frequency domain 共Plomp, 1964; Plomp and Mimpen, 1968; Soderquist, 1970; Moore and Ohgushi, 508 J. Acoust. Soc. Am., Vol. 119, No. 1, January 2006 TABLE I. For each hearing-impaired listener, the table shows hearing loss at 4 kHz, test level, gender/age, and likely etiology of the hearing loss. Listener HL at 4 kHz 共dB兲 Test level 共dB SPL兲 Gender/Age 共years兲 S1 S2 S3 S4 S5 S6 S7 S8 60 40 50 50 65 65 80 75 85 83 88 85 90 80 96 93 M/62 M/55 F/68 M/47 M/53 F/69 F/73 M/71 Etiology Noise exposure Infection Antibiotics Head injury Antibiotics Unknown Presbycusis 共?兲 Noise exposure 1993兲. Performance was measured as a function of the frequency separation of the modulator components. The method was identical to that used by Sek and Moore 共2003兲 with normally hearing listeners, so it is described only briefly here. A. Listeners and test levels Eight hearing-impaired listeners were tested. All were paid for their services. The air-bone gap in the audiogram was always 艋10 dB, indicating that the hearing losses were sensorineural, presumably cochlear, in origin. Only one ear was tested for each listener. This was the ear that had a hearing loss closest to 50 dB at the test frequency of 4 kHz. The hearing loss at 4 kHz ranged from 40 to 80 dB. The level of the 4-kHz sinusoidal carrier was chosen to give a comfortable loudness for each listener. Initially, the carrier was presented with a level of 80– 85 dB SPL for listeners with hearing loss at 4 kHz up to 65 dB, and 85– 90 dB SPL for listeners with hearing loss at 4 kHz greater than 65 dB. Listeners were asked to indicate whether the carrier was at a comfortable level. If the carrier was judged too soft, the level was increased, while if it was judged too loud, the level was decreased. Generally, the required adjustments were small 共±2 to 4 dB兲. Table I shows, for each listener, the hearing loss at 4 kHz, the test level, the gender and age, and the probable etiology of the hearing loss. Listeners were trained both on the task for this experiment and the task for experiment 2 共described later兲, on alternate days. Training lasted 4 – 6 days with about 2.5 h of training per day. Following training, performance appeared to be stable for all listeners. B. Method and stimuli To maximize the likelihood that listeners could perform the task, we used a modulator containing only three components, with relatively wide frequency spacings between the components. We assessed whether listeners could hear out the central component of this complex modulator. The carrier was a 4-kHz sinusoid. To avoid any influence of resolvable spectral sidebands 共Sek and Moore, 1994; Kohlrausch et al., 2000; Moore and Glasberg, 2001兲, the highest modulation frequency used was 190 Hz, which is less than one-half of the “normal” bandwidth of the auditory filter, ERBN, at A. Sek and B. C. J. Moore: Modulation perception by hearing-impaired listeners FIG. 1. The left panel shows mean results of experiment 1 for the hearing-impaired listeners tested here. The right panel shows mean results for the same conditions for the normally hearing listeners tested by Sek and Moore 共2003兲. The percent correct identification of the target is plotted as a function of the ratio of f target and f comp, using as numerator whichever of the two was the larger. Each symbol denotes one pair of frequencies for the flanking modulator components, as indicated by the key. Error bars indicate ±1 standard deviation 共SD兲 across listeners. 4 kHz 共Glasberg and Moore, 1990兲. Since hearing-impaired listeners usually have auditory-filter bandwidths that are larger than normal 共Pick et al., 1977; Glasberg and Moore, 1986兲, it was very unlikely that the listeners would be able to “hear out” any spectral sidebands 共Moore and Glasberg, 2001兲. The task was designed so that listeners could not perform it by judging the mean rate of the complex modulator. On each trial, three modulated stimuli were presented. The modulator of the first stimulus contained three components. Within a run, the frequencies of the outer two components 共called “flanking” components兲 were fixed and the frequency of the central 共“target”兲 component, f target, was drawn randomly from one of five possible values: 80, 89.4, 100, 111.8, and 125 Hz. The modulators of the second and third stimuli contained just a single component. In one of these, selected at random, the modulation frequency was equal to that of the target. In the other, the 共“comparison”兲 modulation frequency, f comp, was randomly selected from one of the four other possible values of the target. Listeners were required to indicate whether the second or third stimulus contained a modulation component that was present in the first stimulus. Feedback was provided by lights on the response box indicating the correct interval. Listeners were seated in a doublewalled sound-attenuating chamber. The results were analyzed in terms of the ratio between f target and f comp, taking as the numerator whichever of these two was the larger. The smallest ratio was equal to 1.118. This was chosen as it is somewhat larger than the threshold for detection of a change in modulation frequency of a sinusoidal carrier 共Lee, 1994; Lemanska et al., 2002; Füllgrabe and Lorenzi, 2003兲. A run started with five trials using the easiest condition, with the largest ratio of f target and f comp. Scores for these initial five trials were discarded. Then each possible value of f target was paired with each possible value of f comp two times, giving a total of 40 scored trials per run. Data presented are the result of at least 17 runs per listener 共more usually, 20–23 runs兲. Stimuli were generated using a Tucker-Davis Technologies 共TDT兲 array processor 共TDT-AP2兲 in a host PC, and a 16-bit digital-to-analog converter 共TDT-DD1兲 operating at a 50-kHz sampling rate. The stimuli were attenuated 共TDT-PA4兲 and sent through an output amplifier 共TDT-HB6兲 to a Sennheiser HD580 earphone. Each modulator compoJ. Acoust. Soc. Am., Vol. 119, No. 1, January 2006 nent had a modulation index of 0.33. This was chosen so as to avoid overmodulation of the three-component modulator, while ensuring that the modulation was clearly audible. The starting phase of each modulator component was chosen randomly for each and every stimulus. On each trial, the carrier was presented in three bursts separated by silent intervals of 400 ms. Each burst had a 20-ms raised-cosine rise and fall, and an overall duration 共including rise/fall times兲 of 1000 ms. The modulation was applied during the whole of the carrier. The flanking component modulation frequencies were 10 and 190 Hz, 30 and 170 Hz, and 52.6 and 190 Hz. These values were also used by Sek and Moore 共2003兲. The first pair of flanking frequencies was chosen to be widely spaced, so that it would be possible to hear out the middle component even if selectivity in the modulation domain was very poor. The components were equally spaced from the central target component 共100 Hz兲 on a linear frequency scale. The second pair was chosen to be closer in modulation frequency to the target component, while keeping the components symmetrically placed 共on a linear frequency scale兲 around the frequency of the central target component. This was done to assess the ability to hear out the central modulation component under more difficult conditions. For the third pair, the components were equally spaced from the central target component on a logarithmic scale. C. Results and discussion The pattern of the results was similar across the eight listeners, and mean results across listeners are shown in the left panel of Fig. 1; the right panel shows the mean results for normally hearing listeners obtained by Sek and Moore 共2003兲 for the same stimuli, but using a carrier level of 70 dB SPL. Percent-correct scores are plotted as a function of the ratio f target / f comp, with the frequencies of the flanking components as parameter. For the smallest ratio f target / f comp, scores for the hearing-impaired listeners are close to the chance level of 50%. However, scores increase as the ratio increases, and for a ratio of 1.56 the mean scores are all above 50%. Performance was better for the widest spacing of the flanking modulator components 共10 and 190 Hz兲 than for the two smaller spacings. A within-subjects analysis of variance 共ANOVA兲 was A. Sek and B. C. J. Moore: Modulation perception by hearing-impaired listeners 509 conducted on the percent correct scores for the hearingimpaired listeners, with factors ratio f target / f comp and frequencies of the flanking components. There was a significant main effect of the ratio f target / f comp; F共3 , 21兲 = 28.55, p ⬍ 0.001. This suggests an ability to hear out the target modulation in the complex modulator when f target and f comp differ sufficiently. There was also a significant main effect of the frequencies of the flanking components; F共2 , 14兲 = 18.74, p ⬍ 0.001. Post hoc tests, based on the least-significant differences test, showed that performance was significantly better 共p ⬍ 0.01兲 for the 10- and 190-Hz flanking components than for either of the other two combinations of flanking components. There was no significant difference in scores between the 30- and 170-Hz combination and the 52- and 190-Hz combination. There was a significant interaction between the two main factors; F共6 , 42兲 = 2.42, p = 0.042. This reflects the fact that the improvement in performance with increasing value of the ratio f target / f comp was greater for the 10- and 190-Hz combination of flanking components than for either of the other two combinations. However, post hoc tests showed that the improvement in performance with increasing value of the ratio f target / f comp was significant 共p ⬍ 0.01兲 for all three pairs of flanking components. The pattern of results was very similar for the hearingimpaired listeners tested here and the normally hearing listeners tested by Sek and Moore 共2003兲. For the largest ratio f target / f comp, the normally hearing listeners performed slightly better than the hearing-impaired listeners when the flanking component frequencies were 10 and 190 Hz, and 30 and 170 Hz, but the reverse was true when the flanking component frequencies were 52.6 and 190 Hz. To assess whether the difference between the two groups was significant, an ANOVA was conducted on the mean percent correct scores for each group, with the following factors: ratio f target / f comp, frequencies of the flanking components, and group membership. The error variance was estimated from the three-way interaction term. The main effect of group membership was significant; F共1 , 6兲 = 6.39, p = 0.045. There was also a significant interaction between group membership and the frequencies of the flanking components; F共2 , 6兲 = 11.77, p = 0.008. This reflects the finding that performance was affected more by the frequencies of the flanking components for the normal than for the impaired listeners. In summary, the results indicate that the hearingimpaired listeners had some ability to hear out the target modulation in the complex modulator when the flanking components were widely spaced in frequency from the target, and when f target and f comp differed sufficiently. The ability is comparable to, but slightly worse than, that found for normally hearing listeners. Performance for both groups was never perfect even for the widest spacing of the modulator components, which is consistent with the relatively broad tuning of the modulation filters inferred from experiments on modulation masking 共Dau et al., 1997a, 1997b; Ewert and Dau, 2000兲 and on forward masking of AM 共Wojtczak and Viemeister, 2005兲. 510 J. Acoust. Soc. Am., Vol. 119, No. 1, January 2006 III. EXPERIMENT 2: THE DETECTION OF CHANGES IN MODULATOR COMPONENT PHASE A. Background In the audio-frequency domain, listeners are more sensitive to within-channel phase changes than to acrosschannel phase changes 共Zwicker, 1952; Patterson, 1987; Uppenkamp et al., 2001; Moore, 2003兲. Hence, changes in phase sensitivity with changes in the frequency spacing of the components probably reflect the influence of the frequency selectivity of the auditory system. Applying this rationale to the modulation domain, one would expect high sensitivity to component phase for closely spaced modulator components, and poorer sensitivity for widely spaced components. Such a pattern of results was observed by Sek and Moore 共2003兲 for normally hearing listeners in a task requiring detection of a change in relative modulator phase of a three-component modulator. One complicating factor when interpreting phase effects in the modulation domain is that the internal representation of the envelopes of sounds may be distorted by the presence of nonlinearities in the auditory system, such as basilarmembrane compression 共Rhode and Robles, 1974兲 and neural saturation and adaptation effects 共Shofner et al., 1996兲. These nonlinearities can result in the introduction of components in the effective modulation spectrum that were not present in the stimulus itself 共Moore et al., 1999; Verhey et al., 2003; Sek and Moore, 2004; Füllgrabe et al., 2005兲, and they can also result in differences in effective “internal” rootmean-square 共RMS兲 value of the modulator when the relative phases of modulator components are changed. Like Sek and Moore 共2003兲, we included a condition in which the modulation depth of all components was fixed, and a condition in which the effective modulation depth was made unreliable as a cue by randomizing the overall modulation depth from one stimulus to the next. Comparison of the results for the two conditions provides an indication of the extent to which effective modulation depth was used as a cue when there was no randomization of modulation depth. B. Listeners Eight hearing-impaired listeners were tested; they were the same as for experiment 1. All listeners were paid for their services. C. Stimuli The equipment and earphone were the same as for experiment 1. The carrier was again a 4-kHz sinusoid, and its level for each listener was the same as for experiment 1. Listeners were required to discriminate two three-component complex modulators differing only in the relative phases of their components. Each modulator component was of equal amplitude. The center component of the modulator had a frequency, f c, of 50 or 100 Hz. The spacing of the components was 5, 15, 25, 35, or 45 Hz for the 50-Hz f c, and 10, 30, 50, 70, or 90 Hz for the 100-Hz f c. For one modulator, the starting phase was 0° 共sine phase兲 for all components. We refer to this as 0-phase. The waveforms for the 0-phase A. Sek and B. C. J. Moore: Modulation perception by hearing-impaired listeners modulator had relatively high peak factors; see Sek and Moore 共2003兲 for details. For the other modulator, the lowest component was phase-shifted by 180° or radians 共the other two components starting at 0°兲. We refer to this as -phase. The waveforms for this modulator had lower peak factors than for the 0-phase modulator. On each trial, three stimuli were presented. Durations and levels were the same as for experiment 1. In one set of conditions, the modulation index m of each component was fixed at 0.2. In another set of conditions, each modulator component had the same value of m, but the value of m was randomly varied from one stimulus to the next, over a range of ±3 dB in terms of 20 log m 共on a uniform scale in terms of log m兲; the value of m varied between 0.1416 and 0.2825. D. Procedure A three-interval three-alternative forced-choice task was used. Two intervals contained the 0-phase stimulus. The other interval, selected at random from the three intervals, contained the -phase stimulus. The task of the listener was to select the interval that was different from the other two. Feedback was given via lights on the response box. A run started with five trials using the smallest frequency spacing of the modulator components; pilot trials indicated that performance was relatively good for this spacing. Then, in successive trials, stimuli with each frequency spacing were presented once, in ascending order. This sequence was repeated ten times to give a total of 55 trials per run. Results from the first five trials of each run were discarded. For each listener, each modulator center frequency, and each modulation-depth condition 共fixed or randomized兲, 20 runs were obtained. E. Results The pattern of results was reasonably similar across listeners, and mean results across listeners are shown in the top two panels of Fig. 2. Performance worsened with increasing frequency separation of the modulator components, for all listeners. However, even for the largest frequency separation, scores for all listeners remained above the score that would be achieved by guessing 共33.3%兲. For f c = 100 Hz, performance for listeners 1 and 2 was somewhat poorer when the modulation depth was randomized 共squares兲 than when it was fixed, suggesting that, when the modulation depth was fixed, these listeners may have made some use of a cue related to the change in “internal” modulation depth. For the other listeners, and for all listeners when f c = 50 Hz, there was no consistent effect of randomizing the modulation depth. The data for the hearing-impaired listeners were subjected to a within-subjects ANOVA with factors modulator center frequency 共50 or 100 Hz兲, spacing of the components relative to the center frequency 共five values兲, and randomization of modulation depth 共absent or present兲. The main effect of center frequency was significant; F共1 , 7兲 = 10.49, p = 0.014. Overall performance was a little better for f c = 100 Hz 共64%兲 than for f c = 50 Hz 共59%兲. The main effect of frequency spacing of the components was significant; F共4 , 28兲 = 85.88, p ⬍ 0.001. The main effect of randomization J. Acoust. Soc. Am., Vol. 119, No. 1, January 2006 FIG. 2. The top two panels show the mean results for the hearing-impaired listeners tested here, for modulator center frequencies of 50 Hz 共left兲 and 100 Hz 共right兲. The lower two panels show corresponding mean results for the same conditions for the normally hearing listeners tested by Sek and Moore 共2003兲. Error bars indicate ±1 SD across listeners. was also significant; F共1 , 7兲 = 13.21, p = 0.008. The mean score was 63.5% with no randomization and 59.5% with randomization. Finally, the interaction of randomization and frequency spacing of the components was significant, reflecting the fact that randomization of the modulation depth had a larger effect for small spacings than for large spacings; F共4 , 28兲 = 4.47, p = 0.006. No other interactions were significant. The lower two panels of Fig. 2 show the mean results for the normally hearing listeners tested by Sek and Moore 共2003兲 using a carrier level of 70 dB SPL. The results are similar for the two groups, except that the hearing-impaired listeners performed slightly more poorly overall than the normally hearing listeners. To assess whether the difference between the two groups was significant, an ANOVA was conducted on the mean percent correct scores for each group, with such factors as modulator center frequency 共50 or 100 Hz兲, spacing of the components relative to the center frequency 共five values兲, randomization of modulation depth 共absent or present兲, and group membership. The error variance was estimated from the three-way and four-way interaction terms. The main effect of group membership was significant; F共1 , 17兲 = 85.81, p ⬍ 0.001. There was also a significant interaction between group membership and spacing of the components; F共4 , 17兲 = 11.47, p ⬍ 0.001. This reflects the finding that performance changed more with spacing of the components for the normal than for the impaired listeners. F. Discussion Sek and Moore 共2003兲 considered various possible cues that might be used in the phase-discrimination task. Given A. Sek and B. C. J. Moore: Modulation perception by hearing-impaired listeners 511 the small size of the effect of randomization of modulation depth, they argued that listeners did not perform the task using a cue based on the effective internal depth of the modulation. Another possible cue introduced by nonlinearities in the auditory system is a distortion component at the envelope or “beat” rate of the modulation, which is determined by the spacing between modulator components 共Sek and Moore, 2004; Füllgrabe et al., 2005兲. The strength of this component might vary with the phase of the modulator components, and this could be used as a cue to discriminate the phase changes 共possibly mediated via analysis with an MFB兲. However, in several of the conditions of our experiment, the distortion component would have coincided in frequency with a component that was already present. Given that the distortion component appears to be weak in amplitude relative to the primaries, it is unlikely that the change in effective modulation depth of the distortion component would provide a useful cue in the phase discrimination task, especially in the condition where the modulation depth was randomized. Sek and Moore 共2003兲 also considered and rejected as possible cues the ratio of the maximum value to the minimum value of the modulator 共max-min兲 共Forrest and Green, 1987; Strickland and Viemeister, 1996兲 and the crest factor or the skewness of the modulator 共Lorenzi et al., 1999兲. It seems likely, as argued by Sek and Moore 共2003兲 and Ewert et al. 共2002兲, that listeners are sensitive to changes in the shape of the modulator waveform at the output of the modulation filter centered on 共or close to兲 the central component of the modulator. This shape would be influenced by the relative phase of the components. When the modulator components are widely spaced 共spacing greater than 0.5f c兲, they interact less at the output of the modulation filter, so the sensitivity to modulator phase is reduced. However, because of the broad tuning of the modulation filters, some interaction occurs even for wide spacings. This can explain why performance remained above chance for the largest spacing used. It was not possible to use much larger spacings, since if the spacing exceeds f c, one of the modulator components has a negative frequency. IV. GENERAL DISCUSSION In both experiments, the pattern of results obtained for the hearing-impaired listeners was similar to that obtained previously for normally hearing listeners. This suggests that the processing of suprathreshold complex patterns of modulation is not greatly affected by cochlear hearing loss. However, for both experiments, the hearing-impaired listeners tested here did perform slightly and significantly more poorly overall than the normally hearing listeners tested by Sek and Moore 共2003兲. Also, for both experiments, performance of the hearing-impaired listeners was affected less by the frequencies of the flanking components than performance of the normally hearing listeners. This is consistent with the idea that the hearing-impaired listeners had slightly reduced frequency selectivity in the modulation domain, consistent with the modulation-masking results of Lorenzi et al. 共1997兲. 512 J. Acoust. Soc. Am., Vol. 119, No. 1, January 2006 It is not obvious why this should be the case, since the MFB is usually assumed to occur relatively centrally in the auditory system 共Rees and Møller, 1983; Liegeois-Chauvel et al., 2004兲, and there is no obvious reason why the operation of central mechanisms should be affected by cochlear hearing loss 共Füllgrabe et al., 2003兲. While poorer selectivity in the modulation domain for hearing-impaired listeners might explain the reduced effect of the spacing of the flanking components in experiment 2, it does not account for the poorer overall performance; reduced selectivity would lead to more interaction of components, which would be expected to lead to better discrimination of the relative phase of the components. It is of interest to consider other factors that might have affected the performance of the impaired listeners tested here. As described in the Introduction, cochlear hearing loss usually leads to a loss of fast-acting compression on the basilar membrane 共Ruggero et al., 1997兲. For suprathreshold modulation depths, this can lead to a perceived amount of modulation that is greater than normal 共Moore et al., 1996兲. One might argue that in a person with cochlear hearing loss, the effective modulation depth at the input to central mechanisms is greater than normal. This might lead to an advantage in some modulation-processing tasks, and may account for why detection of AM is sometimes better than normal when the comparison is made at equal 共low兲 SLs 共Bacon and Gleitman, 1992; Moore et al., 1992兲. On the other hand, for normally hearing listeners, the high-frequency side of the excitation pattern is processed almost linearly, at least for high characteristic frequencies 共Robles and Ruggero, 2001兲. If a large effective modulation depth is beneficial in a specific task, a normally hearing listener can obtain that benefit by using information from the high-frequency side of the excitation pattern 共Zwicker, 1956兲, at least when the carrier is at a relatively high level. The normally hearing listeners of Sek and Moore 共2003兲 were tested using a carrier level of 70 dB SPL, which would have been sufficient to allow them to use information from the high-frequency side of the excitation pattern. One might therefore argue that differences in basilar-membrane compression between normal and hearingimpaired listeners should have little effect on performance. Another factor that may have affected the relative performance of the normally hearing listeners tested by Sek and Moore 共2003兲 and the hearing-impaired listeners tested here is the relatively low SL of the stimuli for the latter, which ranged from 15 to 43 dB SL 共mean= 27 dB SL兲. Performance on many tasks, including the detection of AM, worsens at very low SLs 共Kohlrausch et al., 2000兲. However, there is little information on the influence of SL on the processing of suprathreshold amounts of AM. The relatively low SL of the stimuli in our experiments might have led to slightly poorer-than-normal performance simply because relatively few neurons are excited at low SLs, so the neural code is relatively sparse. Put another way, neural “noise” may have a greater influence at low SLs than at high SLs. A final possible reason for the slightly poorer-thannormal performance of the hearing-impaired listeners is connected with the fact that they were older than the normally hearing listeners tested by Sek and Moore 共2003兲. It has been A. Sek and B. C. J. Moore: Modulation perception by hearing-impaired listeners suggested that age can have adverse effects on temporal processing 共Wingfield et al., 1985兲. However, Takahashi and Bacon 共1992兲 found that there were no significant effects of age on performance in a variety of modulation processing tasks once the effect of hearing loss was taken into account 共except for a very modest correlation between age and modulation detection sensitivity at low modulation frequencies兲. Furthermore, in our data there was no clear trend for the older listeners 共S7 and S8, aged 73 and 71, respectively兲 to perform more poorly than the younger listeners 共S4 and S5, aged 47 and 53, respectively兲. Thus it seems unlikely that age had a major influence on the results. Overall, the results of the present experiments indicate that hearing impairment has relatively little effect on the ability to process complex suprathreshold patterns of modulation, although the impaired listeners may have slightly reduced frequency selectivity in the modulation domain. V. CONCLUSIONS We have described two experiments assessing the ability of listeners with cochlear hearing loss to process suprathreshold amounts of AM applied to a 4-kHz sinusoidal carrier presented at a comfortable listening level. Experiment 1 examined the ability to “hear out” the modulation frequency of the central component of a three-component modulator. All listeners showed some ability to perform the task when the components in the modulator were widely spaced and when the frequencies of the target and comparison differed sufficiently. Scores were poorer when the two flanking components were closer to the central target component frequency of 100 Hz. This is consistent with the relatively broad tuning of the modulation filters inferred from experiments on modulation masking. The pattern of the results was very similar to that found earlier for normally hearing listeners, although performance overall was slightly poorer than for normally hearing listeners 共Sek and Moore, 2003兲. Experiment 2 examined the ability to hear a change in the relative phase of the components in a three-component modulator with harmonically spaced components. The frequency of the central component, f c, was either 50 or 100 Hz. Performance was good 共70–80% correct兲 when the component spacing was 艋0.5f c, but worsened markedly for frequency spacings greater than that. This is broadly the pattern of results predicted from the concept of the MFB. Performance was only slightly impaired by randomizing the overall modulation depth from one stimulus to the next. This suggests that listeners did not use the overall effective internal depth of the modulation as a cue. Again, the pattern of the results was very similar to that found earlier for normally hearing listeners, although performance overall was slightly poorer than for normally hearing listeners 共Sek and Moore, 2003兲. It may be concluded that the ability to process complex suprathreshold patterns of AM is affected only slightly by cochlear hearing loss. J. Acoust. Soc. Am., Vol. 119, No. 1, January 2006 ACKNOWLEDGMENTS This work was supported by the Wellcome Trust and by the MRC. We thank Christian Füllgrabe for helpful discussions and Brian Glasberg for help with statistical analyses. We also thank Associate Editor Andrew Oxenham, Christian Lorenzi, and an anonymous reviewer for helpful comments on an earlier version of this paper. Bacon, S. P., and Gleitman, R. M. 共1992兲. “Modulation detection in subjects with relatively flat hearing losses,” J. Speech Hear. Res. 35, 642–653. Darwin, C. J., and Carlyon, R. P. 共1995兲. “Auditory grouping,” In Hearing, edited by B. C. J. Moore 共Academic Press, San Diego兲. Dau, T., Kollmeier, B., and Kohlrausch, A. 共1997a兲. “Modeling auditory processing of amplitude modulation. I. Detection and masking with narrowband carriers,” J. Acoust. Soc. Am. 102, 2892–2905. Dau, T., Kollmeier, B., and Kohlrausch, A. 共1997b兲. “Modeling auditory processing of amplitude modulation. II. Spectral and temporal integration,” J. Acoust. Soc. Am. 102, 2906–2919. Ewert, S. D., and Dau, T. 共2000兲. “Characterizing frequency selectivity for envelope fluctuations,” J. Acoust. Soc. Am. 108, 1181–1196. Ewert, S. D., Verhey, J. L., and Dau, T. 共2002兲. “Spectro-temporal processing in the envelope-frequency domain,” J. Acoust. Soc. Am. 112, 2921– 2931. Formby, C. 共1982兲. “Differential sensitivity to tonal frequency and to the rate of amplitude modulation of broad-band noise by hearing-impaired listeners,” Ph.D. thesis, Washington University, St. Louis. Formby, C. 共1986兲. “Frequency and rate discrimination by Meniere patients,” Audiology 25, 10–18. Forrest, T. G., and Green, D. M. 共1987兲. “Detection of partially filled gaps in noise and the temporal modulation transfer function,” J. Acoust. Soc. Am. 82, 1933–1943. Füllgrabe, C., and Lorenzi, C. 共2003兲. “The role of envelope beat cues in the detection and discrimination of second-order amplitude modulation,” J. Acoust. Soc. Am. 113, 49–52. Füllgrabe, C., Meyer, B., and Lorenzi, C. 共2003兲. “Effect of cochlear damage on the detection of complex temporal envelopes,” Hear. Res. 178, 35–43. Füllgrabe, C., Moore, B. C. J., Demany, L., Ewert, S. D., Sheft, S., and Lorenzi, C. 共2005兲. “Modulation masking produced by second-order modulators,” J. Acoust. Soc. Am. 117, 2158–2168. Glasberg, B. R., and Moore, B. C. J. 共1986兲. “Auditory filter shapes in subjects with unilateral and bilateral cochlear impairments,” J. Acoust. Soc. Am. 79, 1020–1033. Glasberg, B. R., and Moore, B. C. J. 共1990兲. “Derivation of auditory filter shapes from notched-noise data,” Hear. Res. 47, 103–138. Grant, K. W. 共1998兲. “Modulation rate detection and discrimination by normal-hearing and hearing-impaired listeners,” J. Acoust. Soc. Am. 104, 1051–1060. Kohlrausch, A., Fassel, R., and Dau, T. 共2000兲. “The influence of carrier level and frequency on modulation and beat-detection thresholds for sinusoidal carriers,” J. Acoust. Soc. Am. 108, 723–734. Lee, J. 共1994兲. “Amplitude modulation rate discrimination with sinusoidal carriers,” J. Acoust. Soc. Am. 96, 2140–2147. Lemanska, J., Skrodzka, E., and Sek, A. 共2002兲. “Discrimination of the amplitude modulation rate,” Arch. Acoust. 27, 3–22. Liegeois-Chauvel, C., Lorenzi, C., Trebuchon, A., Regis, J., and Chauvel, P. 共2004兲. “Temporal envelope processing in the human left and right auditory cortices,” Cereb. Cortex 14, 731–740. Lorenzi, C., Berthommier, F., and Demany, L. 共1999兲. “Discrimination of amplitude-modulation phase spectrum,” J. Acoust. Soc. Am. 105, 2987– 2990. Lorenzi, C., Micheyl, C., Berthommier, F., and Portalier, S. 共1997兲. “Modulation masking in listeners with sensorineural hearing loss,” J. Speech Lang. Hear. Res. 40, 200–207. Moore, B. C. J. 共1998兲. Cochlear Hearing Loss 共Whurr, London兲. Moore, B. C. J. 共2003兲. An Introduction to the Psychology of Hearing, 5th ed. 共Academic Press, San Diego兲. Moore, B. C. J., and Glasberg, B. R. 共1997兲. “A model of loudness perception applied to cochlear hearing loss,” Aud. Neurosci. 3, 289–311. Moore, B. C. J., and Glasberg, B. R. 共2001兲. “Temporal modulation transfer functions obtained using sinusoidal carriers with normally hearing and hearing-impaired listeners,” J. Acoust. Soc. Am. 110, 1067–1073. A. Sek and B. C. J. Moore: Modulation perception by hearing-impaired listeners 513 Moore, B. C. J., and Glasberg, B. R. 共2004兲. “A revised model of loudness perception applied to cochlear hearing loss,” Hear. Res. 188, 70–88. Moore, B. C. J., and Ohgushi, K. 共1993兲. “Audibility of partials in inharmonic complex tones,” J. Acoust. Soc. Am. 93, 452–461. Moore, B. C. J., Sek, A., and Glasberg, B. R. 共1999兲. “Modulation masking produced by beating modulators,” J. Acoust. Soc. Am. 106, 908–918. Moore, B. C. J., Shailer, M. J., and Schooneveldt, G. P. 共1992兲. “Temporal modulation transfer functions for band-limited noise in subjects with cochlear hearing loss,” Br. J. Audiol. 26, 229–237. Moore, B. C. J., Wojtczak, M., and Vickers, D. A. 共1996兲. “Effect of loudness recruitment on the perception of amplitude modulation,” J. Acoust. Soc. Am. 100, 481–489. Patterson, R. D. 共1987兲. “A pulse ribbon model of monaural phase perception,” J. Acoust. Soc. Am. 82, 1560–1586. Pick, G., Evans, E. F., and Wilson, J. P. 共1977兲. “Frequency resolution in patients with hearing loss of cochlear origin,” in Psychophysics and Physiology of Hearing, edited by E. F. Evans and J. P. Wilson 共Academic Press, London兲. Plomp, R. 共1964兲. “The ear as a frequency analyzer,” J. Acoust. Soc. Am. 36, 1628–1636. Plomp, R. 共1978兲. “Auditory handicap of hearing impairment and the limited benefit of hearing aids,” J. Acoust. Soc. Am. 63, 533–549. Plomp, R. 共1983兲. “The role of modulation in hearing,” in Hearing— Physiological Bases and Psychophysics, edited by R. Klinke and R. Hartmann 共Springer, Berlin兲. Plomp, R. 共1994兲. “Noise, amplification, and compression: Considerations of three main issues in hearing aid design,” Ear Hear. 15, 2–12. Plomp, R., and Mimpen, A. M. 共1968兲. “The ear as a frequency analyzer II,” J. Acoust. Soc. Am. 43, 764–767. Rees, A., and Møller, A. R. 共1983兲. “Responses of neurons in the inferior colliculus of the rat to AM and FM tones,” Hear. Res. 10, 301–310. Rhode, W. S., and Robles, L. 共1974兲. “Evidence from Mössbauer experiments for non-linear vibration in the cochlea,” J. Acoust. Soc. Am. 55, 588–596. Robles, L., and Ruggero, M. A. 共2001兲. “Mechanics of the mammalian cochlea,” Physiol. Rev. 81, 1305–1352. Ruggero, M. A., Rich, N. C., Recio, A., Narayan, S. S., and Robles, L. 共1997兲. “Basilar-membrane responses to tones at the base of the chinchilla cochlea,” J. Acoust. Soc. Am. 101, 2151–2163. Sek, A., and Moore, B. C. J. 共1994兲. “The critical modulation frequency and its relationship to auditory filtering at low frequencies,” J. Acoust. Soc. Am. 95, 2606–2615. Sek, A., and Moore, B. C. J. 共2003兲. “Testing the concept of a modulation filter bank: The audibility of component modulation and detection of 514 J. Acoust. Soc. Am., Vol. 119, No. 1, January 2006 phase change in three-component modulators,” J. Acoust. Soc. Am. 113, 2801–2811. Sek, A., and Moore, B. C. J. 共2004兲. “Estimation of the level and phase of the simple distortion tone in the modulation domain,” J. Acoust. Soc. Am. 116, 3031–3037. Shannon, R. V., Zeng, F.-G., Kamath, V., Wygonski, J., and Ekelid, M. 共1995兲. “Speech recognition with primarily temporal cues,” Science 270, 303–304. Shofner, W. D., Sheft, S., and Guzman, S. J. 共1996兲. “Responses of ventral cochlear nucleus units in the chinchilla to amplitude modulation by lowfrequency, two-tone complexes,” J. Acoust. Soc. Am. 99, 3592–3605. Soderquist, D. R. 共1970兲. “Frequency analysis and the critical band,” Psychonomic Sci. 21, 117–119. Steinberg, J. C., and Gardner, M. B. 共1937兲. “The dependency of hearing impairment on sound intensity,” J. Acoust. Soc. Am. 9, 11–23. Strickland, E. A., and Viemeister, N. F. 共1996兲. “Cues for discrimination of envelopes,” J. Acoust. Soc. Am. 99, 3638–3646. Takahashi, G. A., and Bacon, S. P. 共1992兲. “Modulation detection, modulation masking, and speech understanding in noise in the elderly,” J. Speech Hear. Res. 35, 1410–1421. Tandetnik, S., Garnier, S., and Lorenzi, C. 共2001兲. “Measurement of firstand second-order modulation detection thresholds in listeners with cochlear hearing loss,” Br. J. Audiol. 35, 355–364. Turner, C. W., Souza, P. E., and Forget, L. N. 共1995兲. “Use of temporal envelope cues in speech recognition by normal and hearing-impaired listeners,” J. Acoust. Soc. Am. 97, 2568–2576. Uppenkamp, S., Fobel, S., and Patterson, R. D. 共2001兲. “The effects of temporal asymmetry on the detection and perception of short chirps,” Hear. Res. 158, 71–83. Verhey, J. L., Ewert, S. D., and Dau, T. 共2003兲. “Modulation masking produced by complex tone modulators,” J. Acoust. Soc. Am. 114, 2135–2146. Wingfield, A., Poon, L. W., Lombardi, L., and Lowe, D. 共1985兲. “Speed of processing in normal aging: Effects of speech rate, linguistic structure and processing time,” J. Gerontol. 40, 579–585. Wojtczak, M., and Viemeister, N. F. 共2005兲. “Forward masking of amplitude modulation: Basic characteristics,” J. Acoust. Soc. Am. 118, 3198–3210. Zwicker, E. 共1952兲. “Die Grenzen der Hörbarkeit der Amplitudenmodulation und der Frequenzmodulation eines Tones 共The limits of audibility of amplitude modulation and frequency modulation of a pure tone兲,” Acustica 2, 125–133. Zwicker, E. 共1956兲. “Die elementaren Grundlagen zur Bestimmung der Informationskapazität des Gehörs 共The foundations for determining the information capacity of the auditory system兲,” Acustica 6, 356–381. A. Sek and B. C. J. Moore: Modulation perception by hearing-impaired listeners