Detection of frequency modulation by hearing-impaired listeners ...

Detection of frequency modulation by hearing-impaired listeners: Effects of carrier frequency, modulation rate, and added amplitude modulation Brian C. J. Moorea) Department of Experimental Psychology, University of Cambridge, Downing Street, Cambridge CB2 3EB, England

Ewa Skrodzkab) Institute of Acoustics, Adam Mickiewicz University, 85 Umultowska, 61-614 Poznan, Poland

共Received 6 July 2001; revised 5 September 2001; accepted 8 October 2001兲 It has been proposed that the detection of frequency modulation 共FM兲 of sinusoidal carriers can be mediated by two mechanisms: a place mechanism based on FM-induced amplitude modulation 共AM兲 in the excitation pattern, and a temporal mechanism based on phase-locking in the auditory nerve. The temporal mechanism appears to be ‘‘sluggish’’ and does not play a role for FM rates above about 10 Hz. It also does not play a role for high carrier frequencies 共above about 5 kHz兲. This experiment examined FM detection in three young subjects with normal hearing and four elderly subjects with cochlear hearing loss. Carrier frequencies were 0.25, 0.5, 1, 2, 4, and 6 kHz and modulation rates were 2, 5, 10, and 20 Hz. FM detection thresholds were measured both in the absence of AM, and with AM of a fixed depth (m⫽0.33) added in both intervals of a forced-choice trial. The added AM was intended to disrupt cues based on FM-induced AM in the excitation pattern. Generally, the hearing-impaired subjects performed markedly more poorly than the normal-hearing subjects. For the normal-hearing subjects, the disruptive effect of the AM tended to increase with increasing modulation rate, for carrier frequencies below 6 kHz, as found previously by Moore and Sek 关J. Acoust. Soc. Am. 100, 2320–2331 共1996兲兴. For the hearing-impaired subjects, the disruptive effective of the AM was generally larger than for the normal-hearing subjects, and the magnitude of the disruption did not consistently increase with increasing modulation rate. The results suggest that cochlear hearing impairment adversely affects both temporal and excitation pattern mechanisms of FM detection. © 2002 Acoustical Society of America. 关DOI: 10.1121/1.1424871兴 PACS numbers: 43.66.Fe, 43.66.Hg, 43.66.Ba 关MRL兴

I. INTRODUCTION

It has been proposed that the frequency discrimination of steady pulsed tones by normal-hearing listeners is largely based on temporal information 共cues derived from phase locking兲 for frequencies up to 4 –5 kHz 共Moore, 1973a, b, 1974; Goldstein and Srulovicz, 1977; Sek and Moore, 1995; Moore, 1997; Micheyl et al., 1998兲. Above 4 –5 kHz, frequency discrimination is thought to depend on place mechanisms, based on changes in the excitation pattern 共Moore, 1973b; Sek and Moore, 1995兲. The mechanisms underlying the detection of frequency modulation 共FM兲 of sinusoidal carriers are thought to depend on the modulation rate. For sinusoidal modulation with rates above about 10 Hz, detection is probably largely based on excitation-pattern cues 共Zwicker, 1956; Zwicker and Fastl, 1990; Moore and Sek, 1994, 1995; Saberi and Hafter, 1995; Sek and Moore, 1995兲. FM results in modulation of the excitation level at each place on the pattern, so the FM is effectively transformed into amplitude modulation 共AM兲. Thus the FM can be detected as AM, either by using information from the single point on the excitation pattern where the AM is greatest 共Zwicker, 1956; a兲

Electronic mail: [email protected] Electronic mail: [email protected]

b兲

J. Acoust. Soc. Am. 111 (1), Pt. 1, Jan. 2002

Zwicker and Fastl, 1990兲 or by combining information from different parts of the excitation pattern 共Moore and Sek, 1994兲. For very low FM rates 共around 2 Hz兲, temporal information may also play a role 共Moore and Sek, 1995, 1996; Plack and Carlyon, 1995; Sek and Moore, 1995兲; the short-term pattern of phase locking can be used to estimate the momentary frequency, and changes in phase locking over time indicate that FM is present. A similar temporal mechanism probably plays a role in the detection of FM of the fundamental frequency (F0) of harmonic complex tones, when those tones are bandpass filtered so as to contain only unresolved harmonics 共Plack and Carlyon, 1994, 1995; Shackleton and Carlyon, 1994; Carlyon et al., 2000兲. Indeed, for such tones, place information is not available at all, so subjects are forced to rely on temporal information. The temporal mechanism may become less effective for modulation rates above about 5 Hz because it is ‘‘sluggish,’’ and cannot follow rapid changes in frequency. Consistent with this idea, thresholds for detecting FM of the F0 of harmonic complex tones containing only unresolved harmonics increase with increasing modulation rate over the range 1 to 20 Hz, reaching 20% 共defined as the peak deviation in F0 divided by the mean F0兲 for a modulation rate of

0001-4966/2002/111(1)/327/9/$19.00

© 2002 Acoustical Society of America

327

20 Hz 共Carlyon et al., 2000兲. In the case of sinusoidal carriers, performance does not change much with increasing modulation rate 共Zwicker and Fastl, 1990; Moore and Sek, 1995, 1996; Sek and Moore, 1995兲, presumably because the place mechanism ‘‘takes over’’ from the temporal mechanism for modulation rates above 5–10 Hz. Hearing-impaired people generally have higher FM detection thresholds 共called hereafter FMDLs兲 than normalhearing subjects. This might occur because their frequency selectivity is poorer than normal, disrupting the excitationpattern mechanism 共Pick et al., 1977; Glasberg and Moore, 1986; Moore, 1998兲. Alternatively, or in addition, the use of cues based on phase locking may be impaired, either because phase locking itself is weaker than normal 共Woolf et al., 1981兲, or because temporal information is decoded via spatial cross correlation 共Loeb et al., 1983兲 or coincidence detection 共Shamma and Klein, 2000兲, and these mechanisms are disrupted by changes in frequency-to-place mapping or traveling wave velocity associated with cochlear hearing loss 共Ruggero, 1994兲. Zurek and Formby 共1981兲 measured FMDLs in ten subjects with sensorineural hearing loss 共assumed to be mainly of cochlear origin兲 using a 3-Hz modulation rate and carrier frequencies between 125 and 4000 Hz. Subjects were tested at a sensation level 共SL兲 of 25 dB, a level above which performance was found 共in pilot studies兲 to be roughly independent of level. The FMDLs tended to increase with increasing hearing loss at a given frequency. For a given degree of hearing loss, the worsening of performance with increasing hearing loss was greater at low frequencies than at high frequencies. Zurek and Formby suggested two possible explanations for the greater effect at low frequencies. The first is based on the assumption that two mechanisms are involved in coding frequency, a temporal mechanism at low frequencies and a place mechanism at high frequencies. The temporal mechanism may be more disrupted by hearing loss than the place mechanism. An alternative possibility is that, at low frequencies, absolute thresholds do not provide an accurate indicator of the extent of cochlear damage, since these thresholds may be determined by neurons with characteristic frequencies above the test frequency. For example, if the inner hair cells at a region of the basilar membrane tuned to low frequencies are damaged, then low frequencies may be detected via the ‘‘tails’’ of the tuning curves of neurons with higher characteristic frequencies. In extreme cases, there may be a dead region at low frequencies, with no functioning inner hair cells and/or neurons 共Thornton and Abbas, 1980; Florentine and Houtsma, 1983; Turner et al., 1983; Moore et al., 2000; Moore, 2001兲. Moore and Glasberg 共1986兲 measured both FMDLs and thresholds for detecting AM 共called hereafter AMDLs兲, using a 4-Hz modulation rate. They used subjects with moderate unilateral and bilateral cochlear impairments. Stimuli were presented at a fixed level of 80 dB SPL, which was at least 10 dB above the absolute threshold. The FMDLs were larger for the impaired than for the normal ears, by an average factor of 3.8 for a frequency of 500 Hz and 1.5 for a frequency of 2000 Hz, although the average hearing loss was 328

J. Acoust. Soc. Am., Vol. 111, No. 1, Pt. 1, Jan. 2002

similar for these two frequencies. The greater effect at low frequencies is consistent with the results of Zurek and Formby 共1981兲. The AMDLs were not very different for the normal and impaired ears. The AMDLs provide an estimate of the smallest detectable change in excitation level. Moore and Glasberg 共1986兲 also used the notched-noise method 共Patterson, 1976; Glasberg and Moore, 1990兲 to estimate the slopes of the auditory filters, at each test frequency. The slopes, together with the AMDLs, were used to predict the FMDLs on the basis of Zwicker’s excitation-pattern model. The obtained FMDLs were reasonably close to the predicted values. In other words, the results were consistent with the excitation-pattern model. This contrasts with results from normal-hearing subjects, for whom FMDLs for a 4-Hz modulation rate are not well predicted by an excitationpattern model 共Moore and Glasberg, 1989兲. Grant 共1987兲 measured FMDLs for three normal-hearing subjects and three subjects with profound hearing losses. The sinusoidal carrier was modulated in frequency by a 3-Hz triangle function. Stimuli were presented at 30 dB SL for the normal subjects and at a ‘‘comfortable listening level’’ 共110– 135 dB SPL兲 for the impaired subjects. For all carrier frequencies 共100 to 1000 Hz兲, FMDLs were larger, by an average factor of 9.5, for the hearing-impaired than for the normal-hearing subjects. Grant also measured FMDLs when the stimuli were simultaneously amplitude modulated by a noise that was low-pass filtered at 3 Hz. The slow random amplitude fluctuations produced by this amplitude modulation would be expected to impair the use of cues for FM detection based on changes in excitation level. Consistent with the predictions of the excitation-pattern model, the random AM led to increased FMDLs. Interestingly, the increase was much greater for the hearing-impaired than for the normal-hearing subjects. When the random AM was present, thresholds for the hearing-impaired subjects were about 16 times those for the normal-hearing subjects. As described earlier, it is likely that, for low modulation rates and for carriers below about 5 kHz, normal-hearing subjects can extract information about FM both from changes in excitation level and from phase locking 共Moore and Sek, 1995; Sek and Moore, 1995兲. The random AM disrupts the use of changes in excitation level, but does not markedly affect the use of phase locking cues. The profoundly hearing-impaired subjects of Grant appear to have been relying mainly or exclusively on changes in excitation level. Hence, the random AM had severe adverse effects on the FMDLs. Lacher-Fouge`re and Demany 共1998兲 measured FMDLs for a 500-Hz carrier, using modulation rates of 2 and 10 Hz. They used five normal-hearing subjects and seven subjects with cochlear hearing loss ranging from 30 to 75 dB at 500 Hz. Stimuli were presented at a ‘‘comfortable’’ loudness level. The subjects with losses up to 45 dB had thresholds that were about a factor of 2 larger than for the normalhearing subjects. The subjects with larger losses had thresholds that were as much as ten times larger than normal. The effect of the hearing loss was similar for the two modulation rates. Lacher-Fouge`re and Demany suggested that cochlear hearing loss disrupts excitation-pattern 共place兲 cues and B. C. J. Moore and E. Skrodzka: FM detection by impaired listeners

phase-locking cues to a roughly equal extent. The present experiment was intended to provide further insight into the mechanisms underlying FM detection in subjects with moderate cochlear hearing loss. FMDLs were measured for a wide range of carrier frequencies 共0.25– 6 kHz兲 for modulation rates, f m , of 2, 5, 10, and 20 Hz. Thresholds were determined when FM only was present and when the carriers in both intervals of a forced-choice trial were amplitude modulated at the same rate as the FM with a modulation index of 0.333. The added AM was intended to disrupt cues based on FM-induced AM in the excitation pattern. In a similar experiment conducted using normal-hearing subjects, Moore and Sek 共1996兲 found that, for carrier frequencies up to 4 kHz, the deleterious effect of the added AM increased with increasing f m . For the 6-kHz carrier, the deleterious effect was independent of f m . This pattern of results was interpreted as being consistent with the hypothesis that both temporal and place mechanisms are involved in FM detection. The temporal mechanism is assumed to dominate for carriers below about 4 kHz, and for very low modulation rates; in these conditions the added AM had only a small disruptive effect. The place mechanism is assumed to dominate for high carrier frequencies, and for lower carrier frequencies when stimuli are frequency modulated at high rates; in these conditions the added AM had a larger disruptive effect. We reasoned that if hearing-impaired subjects are able to use phase-locking information for low modulation rates 共and for carriers below about 5 kHz兲, the added AM should have a relatively small effect for these rates, similar to the small effect found for normal-hearing listeners. However, if the extraction of phase-locking cues is disrupted, hearingimpaired subjects may rely on excitation-pattern changes for all modulation rates. In this case, the added AM should have a marked disruptive effect for all modulation rates. II. METHOD A. Procedure

FMDLs were measured using a two-interval twoalternative forced-choice task. On each trial, two successive stimuli were presented, one with sinusoidal FM and one without FM. The order of the two stimuli in each pair was random. Subjects were required to identify the interval with the FM by pressing the appropriate button on the response box. In some conditions, AM with a modulation depth, m, of 0.33 共corresponding to a peak-to-valley ratio of 6 dB兲 was added to both stimuli in a trial. This relatively small modulation depth was chosen to minimize AM-induced pitch fluctuations 共Burns and Turner, 1986兲. A two-down one-up adaptive procedure was used 共Levitt, 1971兲. The amount of FM was changed by a factor of 1.5 until four reversals had occurred, and by a factor of 1.25 for a further eight reversals. The threshold was estimated as the geometric mean of the amounts of FM at the last eight reversal points. Each threshold reported is the geometric mean of at least three 共usually four兲 estimates. Subjects were tested in a double-walled sound-attenuating chamber. Correct-answer feedback was provided by lights on the response box. J. Acoust. Soc. Am., Vol. 111, No. 1, Pt. 1, Jan. 2002

B. Stimuli

The carrier frequency was 0.25, 0.5, 1, 2, 4, or 6 kHz. The overall level was 85 dB SPL for the hearing-impaired subjects and 70 dB SPL for the normal-hearing subjects; the level is specified as the estimated level at the eardrum. For the hearing-impaired subjects, the carrier was at least 10 dB above the absolute threshold for all frequencies tested, and in most cases was more than 20 dB above the absolute threshold 共see Sec. II C兲. The modulation frequency, f m , was 2, 5, 10, or 20 Hz. The starting phase of the FM for each stimulus was chosen randomly from one of four values: 0, ␲/2, ␲, and 3␲/2. Each stimulus had an overall duration of 1000 ms, including 20-ms raised-cosine rise/fall times. When AM was present, it was added with a phase of 180° relative to the FM. This differs somewhat from earlier experiments, where the relative phase varied randomly from one stimulus to the next 共Moore and Glasberg, 1989; Moore and Sek, 1996兲. Pilot experiments using normal-hearing subjects indicated that the disruptive effect of the AM was almost unaffected by the phase of the AM relative to the FM. Stimuli were generated using a Tucker-Davis 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. They were attenuated 共TDT-PA4兲 and sent through an output amplifier 共TDT-HB6兲 to a Sennheiser HD580 earphone. The headphone was chosen for its relatively smooth frequency response. This was important, since if the frequency response is irregular, FM can be transformed into a combination of FM and AM. Measurements using a probe microphone close to the eardrum 共not using the subjects of the present experiment兲 and using a KEMAR manikin 共Burkhard and Sachs, 1975兲 showed that the response was very flat 共⫾0.6 dB兲 from 0.25 up to 1 kHz. The response showed a peak around 3 kHz, but with a smooth variation with frequency up to about 5 kHz. However, around 6 kHz, the response showed some irregularities, which varied from one ear to another. These irregularities might have led to detectable amounts of FMinduced AM when the FMDLs were relatively large. Stimuli were presented to one ear only. For the hearingimpaired subjects, the ear with better hearing 共averaged over the frequencies from 0.5 to 4 kHz兲 was tested. For the normal-hearing subjects, the test ear was the preferred ear of each subject. C. Subjects

Three subjects with normal hearing and four subjects with mild to moderate hearing loss were tested. One normalhearing subject was author ES. All other subjects were paid for their services. The normal-hearing subjects had absolute thresholds less than 20 dB HL at all audiometric frequencies and had no history of hearing disorders. Their ages ranged from 27 to 37 years. The audiograms for the test ears of the hearing-impaired subjects are shown in Fig. 1. The ages of the hearing-impaired subjects were 80, 61, 83, and 72 years, for TT, MP, GW, and AR, respectively. All losses were diagnosed as being of cochlear origin as indicated by lack of an air-bone gap; signs of loudness recruitment; normal tympanograms; and normal acoustic reflex thresholds. All hearing-

B. C. J. Moore and E. Skrodzka: FM detection by impaired listeners

329

performance appeared to be relatively stable; practice effects for FM detection appear to be relatively small 共Moore, 1976兲. III. RESULTS

FIG. 1. Audiograms of the four hearing-impaired subjects.

impaired subjects were given the test for diagnosis of dead regions described by Moore et al. 共2000兲. No evidence for dead regions was found. All subjects, including the hearingimpaired subjects, had extensive previous experience in psychoacoustic tasks, but not specifically in the detection of FM. All subjects received at least 2 h of practice on a selection of conditions from the present experiment, after which their

The pattern of results was similar for the three normalhearing subjects, and the mean results are shown in Figs. 2–5 by the short-dashed lines 共detection of FM alone兲 and the long-dashed lines 共detection of FM with added AM兲. The FMDLs are expressed as peak-to-peak deviation divided by the carrier frequency. The overall pattern of the results is similar to that found by Moore and Sek 共1996兲. For the lower carrier frequencies, the disruptive effect of the added AM tends to increase with increasing modulation rate, but for the 6-kHz carrier there is little effect of modulation rate. The results differ somewhat from earlier results using similar stimuli 共Moore and Glasberg, 1989; Moore and Sek, 1996兲 in that thresholds are not higher for the 6-kHz carrier than for the lower carrier frequencies. This may simply reflect individual differences, as frequency discrimination varies more across individuals at high frequencies than at medium frequencies 共Nelson et al., 1983兲. It is possible that the subjects in the present experiment used FM-induced AM 共introduced by irregularities in earphone response兲 as a detection cue at 6 kHz. However, this seems unlikely, as the disruptive effect of the added AM was not greater at 6 kHz than at lower carrier frequencies. To assess the statistical significance of the effects described above, a within-subjects analysis of variance 共ANOVA兲 was conducted with factors carrier frequency, modulation frequency, and presence or absence of AM. The analysis was conducted on the logarithms of the FMDLs 共expressed as a proportion of carrier frequency兲, as the vari-

FIG. 2. FMDLs plotted as a function of modulation frequency. Each panel shows results for one carrier frequency. The circles indicate FMDLs for hearingimpaired subject TT, without 共open circles兲 and with 共filled circles兲 added AM. The short- and long-dashed lines show corresponding mean results for the normalhearing subjects. Error bars indicate ⫾ one standard deviation. They are omitted when they would span a range less than 0.1 log units 共corresponding to a ratio of 1.26兲

330

J. Acoust. Soc. Am., Vol. 111, No. 1, Pt. 1, Jan. 2002

B. C. J. Moore and E. Skrodzka: FM detection by impaired listeners

FIG. 3. As in Fig. 2, but for subject MP.

ability was more uniform across conditions with this transformation. The main effect of carrier frequency was significant; F(5,10)⫽26.21, p⬍0.001: the mean threshold was lowest at 2 kHz and highest at 0.25 kHz. The main effect of modulation rate was significant; F(3,6)⫽46.87, p⬍0.001: the mean threshold was highest for the 20-Hz rate. The main effect of presence or absence of AM was significant; F(1,2)⫽41.9, p⫽0.023: the added AM led to larger FMDLs. The interaction of the presence or absence of AM with

modulation rate was significant; F(3,6)⫽4.84, p⫽0.048: the effect of the AM increased with increasing modulation rate. The interaction of carrier frequency and modulation rate was also significant; F(15,30)⫽2.54, p⫽0.014: this reflects the fact that for the 6-kHz carrier FMDLs did not vary markedly with modulation rate, while for the lower carrier frequencies FMDLs tended to increase with increasing modulation rate. However, the three-way interaction failed to reach significance at the 0.05 level.

FIG. 4. As in Fig. 2, but for subject GW.

J. Acoust. Soc. Am., Vol. 111, No. 1, Pt. 1, Jan. 2002

B. C. J. Moore and E. Skrodzka: FM detection by impaired listeners

331

FIG. 5. As in Fig. 2, but for subject AR.

The individual results for the hearing-impaired subjects are shown in Figs. 2–5 by open symbols 共FM alone兲 and filled symbols 共with added AM兲. Generally, the FMDLs were markedly larger than for the normal-hearing subjects, and the disruptive effect of the AM was greater. Although the effect of the AM increased with increasing modulation rate in some cases 共e.g., subject TT for carriers from 0.5 to 4 kHz, subject GW for carriers from 2 to 6 kHz, and subject AR for the

4-kHz carrier兲, this effect was not consistently observed and GW even showed a trend opposite to this for the 0.25-, 0.5-, and 1.0-kHz carriers. FMDLs did not vary markedly with carrier frequency, except for subject MP, who showed poorer performance for the 4-kHz carrier than for the other carriers. Figure 6 shows the mean data for the normal-hearing 共open symbols兲 and hearing-impaired 共filled symbols兲 subjects. For the latter, the overall magnitude of the disruptive

FIG. 6. Mean results for the normal-hearing subjects 共open symbols兲 and the hearing-impaired subjects 共filled symbols兲. Otherwise, as in Fig. 2.

332

J. Acoust. Soc. Am., Vol. 111, No. 1, Pt. 1, Jan. 2002

B. C. J. Moore and E. Skrodzka: FM detection by impaired listeners

modulation rate, regardless of whether or not AM was present. For the three higher carrier frequencies, FMDLs tended to decrease with increasing modulation rate when AM was absent. Post hoc tests based on the least-significant differences test showed that the FMDLs in the absence of AM were significantly lower for the 20-Hz rate than for the 2-Hz rate for the carriers of 4 and 6 kHz 共p⬍0.002 and p⬍0.01, respectively兲. When AM was present, there was a slight trend for the FMDLs to increase with increasing modulation rate for the higher carrier frequencies, but this effect was significant only for the 2-kHz carrier (p⬍0.05). IV. DISCUSSION

FIG. 7. FMDLs for the individual hearing-impaired subjects, plotted as a function of carrier frequency, with modulation rate as parameter. Open and filled symbols indicate FMDLs without and with added AM, respectively.

effect of the AM is similar across carrier frequencies. The effect is roughly independent of modulation rate for the three lowest carrier frequencies, but there is a trend for the effect to increase with increasing modulation rate for the three higher carrier frequencies. Figure 7 shows the individual data for the hearingimpaired subjects plotted as a function of carrier frequency with modulation rate as a parameter. Open symbols indicate conditions with FM only, and filled circles indicate conditions with added AM. This figure shows more clearly the local increase in FMDLs at 4 kHz for subject MP. The reason why thresholds increased at 4 kHz and decreased at 6 kHz is not clear. Possibly, at 6 kHz he was able to use FM-induced AM produced by irregularity in the response of the earphone; without such a cue, his performance at 6 kHz might have been worse, as would be expected from his increasing hearing loss with increasing frequency. At 4 kHz, he was probably not able to use such a cue as the response of the earphone was smoother around 4 kHz. Consistent with this idea, for MP the added AM had a very large effect at 6 kHz, but a relatively small effect at 4 kHz. For the three highest carrier frequencies, FMDLs in the condition without AM tend to decrease with increasing modulation frequency 共in Fig. 7, the open circles are above the open diamonds for all subjects兲. In the condition with added AM, the FMDLs for high carrier frequencies do not show a consistent trend with changes in modulation frequency. To assess the statistical significance of the effects described above, a within-subjects ANOVA was conducted with factors carrier frequency, modulation frequency, and presence or absence of AM. As before, the analysis was conducted on the logarithms of the FMDLs. The main effects of carrier frequency and modulation rate were not significant. The main effect of presence or absence of AM was highly significant; F(1,3)⫽144.29, p⬍0.001, performance being worse with the added AM. There was also a significant threeway interaction; F(15,45)⫽3.22, p⬍0.001. For the three lower carrier frequencies, the FMDLs hardly varied with J. Acoust. Soc. Am., Vol. 111, No. 1, Pt. 1, Jan. 2002

The relatively poor performance of the hearing-impaired subjects is consistent with the findings of earlier studies on FM detection, as described in the Introduction 共Zurek and Formby, 1981; Moore and Glasberg, 1986; Grant, 1987; Lacher-Fouge`re and Demany, 1998兲. For example, LacherFouge`re and Demany 共1998兲 found that FMDLs were about a factor of 2 larger than normal for hearing-impaired subjects with mild losses, and as much as ten times larger than normal for subjects with larger losses. Our subjects had moderate losses, and had FMDLs that were three to eight times as large as normal. The added AM produced a strong worsening in performance for the hearing-impaired subjects, even for low carrier frequencies and low modulation rates. For the 2-Hz modulation rate, the FMDL averaged across the four lowest carrier frequencies was a factor of 2.5 larger when AM was present than when it was absent. In contrast, the corresponding ratio for the normal-hearing subjects was only 1.45. Grant 共1987兲 found an even larger disruptive effect of AM on FM detection for his profoundly hearing-impaired subjects; thresholds were increased by an average factor of 16. His greater effect is probably connected with the greater hearing loss of his subjects, who may have been relying exclusively on excitation-pattern cues, but for whom the excitation patterns would have been much broader than normal. It has been argued in the past that the relatively small disruptive effect of AM at low modulation rates for normalhearing subjects reflects the use of temporal information 共Moore and Sek, 1996兲. The larger effect found for the hearing-impaired subjects suggests that they were not using temporal information effectively. Rather, the FMDLs were probably based largely on excitation-pattern cues 共FMinduced AM in the excitation pattern兲, and these cues were strongly disrupted by the added AM. An alternative explanation for the disruptive effect of the AM is that the AM led to level-dependent pitch shifts. For hearing-impaired subjects, sound level can have an abnormally large effect on the pitch of a pure tone, at both low and high frequencies 共Burns and Turner, 1986兲. The AMinduced pitch fluctuations could make it more difficult to detect the FM of the stimuli. However, the AM depth used in our experiment produced only small fluctuations in level; the peak-to-valley ratio was 6 dB. The data of Burns and Turner 共1986兲 suggest that a 6-dB change in level would typically lead to pitch shifts of less than 1%. This is small compared with the thresholds that we measured for our hearing-

B. C. J. Moore and E. Skrodzka: FM detection by impaired listeners

333

impaired subjects in the presence of the AM, which were typically around 10% of the carrier frequency. Thus, we believe that AM-induced pitch fluctuations probably made only a minor contribution to the disruptive effect of the AM on FM detection. Another possible explanation for the disruptive effect of the AM is that, in the absence of AM, the detection of FM by the hearing-impaired listeners may have been based on the detection of loudness fluctuations; such fluctuations might be especially salient for subjects with sloping audiograms. The added AM would introduce prominent loudness fluctuations in both intervals of a 2-AFC trial, making the detection of FM-induced loudness fluctuations more difficult. This explanation is weakened by the finding that the added AM had a large disruptive effect even for frequencies where the audiogram was relatively flat, e.g., subject TT at 0.25 and 0.5 kHz, and subject AR at all frequencies. Also, even for frequencies where the audiogram was sloping, the FM-induced loudness fluctuations would have been small. The slopes of the audiograms of our subjects did not usually exceed 10 dB/oct. The FM detection thresholds in the absence of AM were typically around 2% 共peak-to-peak兲, which means that the sensation level 共SL兲 of the carrier would have fluctuated by less than 0.3 dB as a result of the FM. This would have produced only very small loudness fluctuations. We conclude that the hearing-impaired subjects probably did not detect the FM alone using FM-induced loudness fluctuations. Therefore, the effect of the AM on FM detection probably cannot be explained by disruption of these loudness cues. Another factor that may have contributed to the large disruptive effect of the AM for the hearing-impaired subjects is that the effective ‘‘internal’’ modulation depth produced by the AM was larger for the hearing-impaired than for the normal-hearing subjects, because of the loss of cochlear compression in the former 共Moore et al., 1996兲. However, this does not provide an explanation for why the disruptive effect of the AM increased with increasing modulation rate for the normal-hearing subjects 共for carriers below 6 kHz兲, but did not change consistently with modulation rate for the hearing-impaired subjects. For the three higher carrier frequencies, FMDLs for the hearing-impaired subjects decreased with increasing modulation frequency up to 10 Hz when no AM was present; this effect was statistically significant at 4 and 6 kHz. A similar trend has been observed for normal-hearing subjects for high carrier frequencies in some earlier studies 共Sek and Moore, 1995; Moore and Sek, 1996兲, although it was not clearly apparent for the normal-hearing subjects of the present study. The trend for improving performance with increasing modulation frequency up to 10 Hz is similar to that observed in studies of AM detection when the carrier is gated 共Yost and Sheft, 1997兲; for a review, see Kohlrausch et al. 共2000兲. It is consistent with the idea that the FM was detected via the AM that it introduced in the auditory system. However, for the three lower carrier frequencies, FMDLs in the absence of added AM did not change consistently with modulation frequency. Perhaps the lack of effect for these low frequencies reflects some residual ability to use temporal information for low modulation rates. 334

J. Acoust. Soc. Am., Vol. 111, No. 1, Pt. 1, Jan. 2002

When AM was present, FMDLs did not change consistently with modulation rate for the hearing-impaired subjects; the effect of modulation rate was just significant at 2 kHz, but not for any other carrier frequency. The lack of effect of modulation rate when AM was present may have occurred because performance was limited by the imposed AM; essentially, subjects had to discriminate the change in AM depth produced by adding the FM to the AM. The task thus resembles an AM depth discrimination task, for which thresholds are roughly independent of modulation rate for rates up to 32 Hz and for moderate modulation depths 共Ozimek and Sek, 1988兲. Overall, the conclusions from this study are consistent with those of Lacher-Fouge`re and Demany 共1998兲. The results suggest that both temporal and place mechanisms are adversely affected by cochlear hearing loss. The disruption of place mechanisms is not surprising, given the ample evidence that frequency selectivity is poorer than normal in subjects with cochlear hearing loss; for reviews, see Tyler 共1986兲 and Moore 共1998兲. The origin of the disruption of temporal mechanisms is somewhat uncertain. Phase locking to sinusoids was found to be abnormal in animal models of cochlear hearing loss in one study 共Woolf et al., 1981兲 but not in two others 共Harrison and Evans, 1979; Miller et al., 1997兲. Possibly, the decoding of temporal information depends on having a ‘‘normal’’ traveling were pattern on the basilar membrane 共Loeb et al., 1983; Shamma and Klein, 2000兲, and cochlear hearing loss changes the traveling wave sufficiently 共Ruggero, 1994兲 to severely impair the extraction of temporal information. It should be noted, however, that our hearing-impaired subjects were all elderly. It is possible that part of the reduced ability to make use of temporal information was related to the age of these subjects rather than to the hearing loss per se. Consistent with this idea, it has been shown that elderly subjects with near-normal audiometric thresholds have larger-than-normal thresholds for discriminating interaural time differences and smaller-than-normal binaural masking level differences 共Strouse et al., 1998兲. V. SUMMARY AND CONCLUSIONS

In this study, we measured the ability of both normalhearing and elderly hearing-impaired subjects to detect FM, using a wide range of carrier and modulation frequencies. FMDLs were measured both in the absence and presence of AM 共in both intervals of a forced-choice trial兲. The AM was intended to disrupt cues for detection of FM based on FMinduced AM in the excitation pattern. The results showed that the hearing-impaired subjects performed markedly more poorly than the normal-hearing subjects. For the normalhearing subjects, the disruptive effect of the AM tended to increase with increasing modulation rate, for carrier frequencies below 6 kHz. For the hearing-impaired subjects, the disruptive effective of the AM was generally larger than for the normally hearing subjects, and the magnitude of the disruption did not consistently increase with increasing modulation rate. The results suggest that cochlear hearing impairment adversely affects both temporal and excitation pattern mechanisms of FM detection. B. C. J. Moore and E. Skrodzka: FM detection by impaired listeners

ACKNOWLEDGMENTS

This work was supported by the Medical Research Council 共UK兲. E.S. was partly supported by the State Committee for Scientific Research 共Poland兲, Grant No. 8T11E01717. The authors thank Aleksander Sek and Brian Glasberg for their assistance and Kathy Arehart for helpful discussions and comments on an earlier version of this article. They also thank Marjorie Leek, Laurent Demany, and an anonymous reviewer for helpful comments. Burkhard, M. D., and Sachs, R. M. 共1975兲. ‘‘Anthropometric manikin for acoustic research,’’ J. Acoust. Soc. Am. 58, 214 –222. Burns, E. M., and Turner, C. 共1986兲. ‘‘Pure-tone pitch anomalies. II. Pitchintensity effects and diplacusis in impaired ears,’’ J. Acoust. Soc. Am. 79, 1530–1540. Carlyon, R. P., Moore, B. C. J., and Micheyl, C. 共2000兲. ‘‘The effect of modulation rate on the detection of frequency modulation and mistuning of complex tones,’’ J. Acoust. Soc. Am. 108, 304 –315. Florentine, M., and Houtsma, A. J. M. 共1983兲. ‘‘Tuning curves and pitch matches in a listener with a unilateral, low-frequency hearing loss,’’ J. Acoust. Soc. Am. 73, 961–965. 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. Goldstein, J. L., and Srulovicz, P. 共1977兲. ‘‘Auditory-nerve spike intervals as an adequate basis for aural frequency measurement,’’ in Psychophysics and Physiology of Hearing, edited by E. F. Evans and J. P. Wilson 共Academic, London兲. Grant, K. W. 共1987兲. ‘‘Frequency modulation detection by normally hearing and profoundly hearing-impaired listeners,’’ J. Speech Hear. Res. 30, 558 – 563. Harrison, R. V., and Evans, E. F. 共1979兲. ‘‘Some aspects of temporal coding by single cochlear fibres from regions of cochlear hair cell degeneration in the guinea pig,’’ Arch. Otolaryngol. 224, 71–78. 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. Lacher-Fouge`re, S., and Demany, L. 共1998兲. ‘‘Modulation detection by normal and hearing-impaired listeners,’’ Audiology 37, 109–121. Levitt, H. 共1971兲. ‘‘Transformed up–down methods in psychoacoustics,’’ J. Acoust. Soc. Am. 49, 467– 477. Loeb, G. E., White, M. W., and Merzenich, M. M. 共1983兲. ‘‘Spatial cross correlation: A proposed mechanism for acoustic pitch perception,’’ Biol. Cybern. 47, 149–163. Micheyl, C., Moore, B. C. J., and Carlyon, R. P. 共1998兲. ‘‘The role of excitation-pattern cues and temporal cues in the frequency and modulation-rate discrimination of amplitude-modulated tones,’’ J. Acoust. Soc. Am. 104, 1039–1050. Miller, R. L., Schilling, J. R., Franck, K. R., and Young, E. D. 共1997兲. ‘‘Effects of acoustic trauma on the representation of the vowel /}/ in cat auditory nerve fibers,’’ J. Acoust. Soc. Am. 101, 3602–3616. Moore, B. C. J. 共1973a兲. ‘‘Frequency difference limens for narrow bands of noise,’’ J. Acoust. Soc. Am. 54, 888 – 896. Moore, B. C. J. 共1973b兲. ‘‘Frequency difference limens for short-duration tones,’’ J. Acoust. Soc. Am. 54, 610– 619. Moore, B. C. J. 共1974兲. ‘‘Relation between the critical bandwidth and the frequency-difference limen,’’ J. Acoust. Soc. Am. 55, 359. Moore, B. C. J. 共1976兲. ‘‘Comparison of frequency DL’s for pulsed tones and modulated tones,’’ Br. J. Audiol. 10, 17–20. Moore, B. C. J. 共1997兲. An Introduction to the Psychology of Hearing, 4th ed. 共Academic, San Diego兲. Moore, B. C. J. 共1998兲. Cochlear Hearing Loss 共Whurr, London兲. Moore, B. C. J. 共2001兲. ‘‘Dead regions in the cochlea: Diagnosis, perceptual consequences, and implications for the fitting of hearing aids,’’ Trends Amplif. 5, 1–34. Moore, B. C. J., and Glasberg, B. R. 共1986兲. ‘‘The relationship between frequency selectivity and frequency discrimination for subjects with unilateral and bilateral cochlear impairments,’’ in Auditory Frequency Selectivity, edited by B. C. J. Moore and R. D. Patterson 共Plenum, New York兲.

J. Acoust. Soc. Am., Vol. 111, No. 1, Pt. 1, Jan. 2002

Moore, B. C. J., and Glasberg, B. R. 共1989兲. ‘‘Mechanisms underlying the frequency discrimination of pulsed tones and the detection of frequency modulation,’’ J. Acoust. Soc. Am. 86, 1722–1732. Moore, B. C. J., and Sek, A. 共1994兲. ‘‘Effects of carrier frequency and background noise on the detection of mixed modulation,’’ J. Acoust. Soc. Am. 96, 741–751. Moore, B. C. J., and Sek, A. 共1995兲. ‘‘Effects of carrier frequency, modulation rate and modulation waveform on the detection of modulation and the discrimination of modulation type 共AM vs FM兲,’’ J. Acoust. Soc. Am. 97, 2468 –2478. Moore, B. C. J., and Sek, A. 共1996兲. ‘‘Detection of frequency modulation at low modulation rates: Evidence for a mechanism based on phase locking,’’ J. Acoust. Soc. Am. 100, 2320–2331. 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. Moore, B. C. J., Huss, M., Vickers, D. A., Glasberg, B. R., and Alca´ntara, J. I. 共2000兲. ‘‘A test for the diagnosis of dead regions in the cochlea,’’ Br. J. Audiol. 34, 205–224. Nelson, D. A., Stanton, M. E., and Freyman, R. L. 共1983兲. ‘‘A general equation describing frequency discrimination as a function of frequency and sensation level,’’ J. Acoust. Soc. Am. 73, 2117–2123. Ozimek, E., and Sek, A. 共1988兲. ‘‘AM difference limens for noise bands,’’ Acustica 66, 153–160. Patterson, R. D. 共1976兲. ‘‘Auditory filter shapes derived with noise stimuli,’’ J. Acoust. Soc. Am. 59, 640– 654. 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, London兲. Plack, C. J., and Carlyon, R. P. 共1994兲. ‘‘The detection of differences in the depth of frequency modulation,’’ J. Acoust. Soc. Am. 96, 115–125. Plack, C. J., and Carlyon, R. P. 共1995兲. ‘‘Differences in frequency modulation detection and fundamental frequency discrimination between complex tones consisting of resolved and unresolved harmonics,’’ J. Acoust. Soc. Am. 98, 1355–1364. Ruggero, M. A. 共1994兲. ‘‘Cochlear delays and traveling waves: Comments on ‘Experimental look at cochlear mechanics,’ ’’ Audiology 33, 131–142. Saberi, K., and Hafter, E. R. 共1995兲. ‘‘A common neural code for frequencyand amplitude-modulated sounds,’’ Nature 共London兲 374, 537–539. Sek, A., and Moore, B. C. J. 共1995兲. ‘‘Frequency discrimination as a function of frequency, measured in several ways,’’ J. Acoust. Soc. Am. 97, 2479–2486. Shackleton, T. M., and Carlyon, R. P. 共1994兲. ‘‘The role of resolved and unresolved harmonics in pitch perception and frequency modulation discrimination,’’ J. Acoust. Soc. Am. 95, 3529–3540. Shamma, S., and Klein, D. 共2000兲. ‘‘The case of the missing pitch templates: how harmonic templates emerge in the early auditory system,’’ J. Acoust. Soc. Am. 107, 2631–2644. Strouse, A., Ashmead, D. H., Ohde, R. N., and Grantham, D. W. 共1998兲. ‘‘Temporal processing in the aging auditory system,’’ J. Acoust. Soc. Am. 104, 2385–2399. Thornton, A. R., and Abbas, P. J. 共1980兲. ‘‘Low-frequency hearing loss: Perception of filtered speech, psychophysical tuning curves, and masking,’’ J. Acoust. Soc. Am. 67, 638 – 643. Turner, C. W., Burns, E. M., and Nelson, D. A. 共1983兲. ‘‘Pure tone pitch perception and low-frequency hearing loss,’’ J. Acoust. Soc. Am. 73, 966 – 975. Tyler, R. S. 共1986兲. ‘‘Frequency resolution in hearing-impaired listeners,’’ in Frequency Selectivity in Hearing, edited by B. C. J. Moore 共Academic, London兲. Woolf, N. K., Ryan, A. F., and Bone, R. C. 共1981兲. ‘‘Neural phase-locking properties in the absence of outer hair cells,’’ Hear. Res. 4, 335–346. Yost, W. A., and Sheft, S. 共1997兲. ‘‘Temporal modulation transfer functions for tonal stimuli: Gated versus continuous conditions,’’ Aud. Neurosci. 3, 401– 414. Zurek, P. M., and Formby, C. 共1981兲. ‘‘Frequency-discrimination ability of hearing-impaired listeners,’’ J. Speech Hear. Res. 24, 108 –112. Zwicker, E. 共1956兲. ‘‘Die elementaren Grundlagen zur Bestimmung der Informationskapazita¨t des Geho¨rs,’’ Acustica 6, 356 –381. Zwicker, E., and Fastl, H. 共1990兲. Psychoacoustics—Facts and Models 共Springer-Verlag, Berlin兲.

B. C. J. Moore and E. Skrodzka: FM detection by impaired listeners

335