Skip to main content
NCBI home page
The Journal of the Acoustical Society of America logoLink to The Journal of the Acoustical Society of America
letter
. 2008 Oct;124(4):1905–1908. doi: 10.1121/1.2968685

Factors contributing to comodulation masking release with dichotic maskers

Emily Buss 1,a), Joseph W Hall III 1
PMCID: PMC2600623  NIHMSID: NIHMS70239  PMID: 19062829

Abstract

Detection threshold for a pure tone signal centered in a narrow band of noise may be reduced by inclusion of additional flanking masker bands, provided that they share coherent amplitude modulation (AM) across frequency. This comodulation masking release (CMR) associated with coherent AM across frequency is often much smaller if the signal and on-signal masker are presented to one ear and the flanking masker band(s) are presented contralaterally. An experiment was carried out to explore the role of peripheral effects (e.g., suppression) and central effects (e.g., grouping) in this finding. As frequently reported, CMR was smaller when two or more flanking maskers were presented contralaterally to the signal than when presented ipsilaterally. An intermediate condition, where a subset of flanking maskers was presented to each ear, provided comparable benefit to presenting all flankers ipsilateral to the signal. This result suggests that central effects may play a significant role in the reduced dichotic CMR under some conditions.

INTRODUCTION

The detection threshold for a pure tone in a narrow band of noise is often lowered by the introduction of maskers at remote frequencies that have the same pattern of amplitude modulation as the on-signal masker, a finding described as comodulation masking release (CMR). This CMR effect can be demonstrated with a single wide band of noise or with multiple narrow band noise maskers distributed in frequency (Hall et al., 1984). While CMR is often described in terms of the across-channel cues associated with coherent amplitude modulation (AM) across frequency, coherent AM is also associated with cues at the signal frequency that could improve performance, such as the envelope modulation associated with beating between neighboring bands (Schooneveldt and Moore, 1987) or suppression (Oxenham and Plack, 1998; Ernst and Verhey, 2006). Identifying the contributions of within- and across-channel cues has been the subject of several studies (e.g., Schooneveldt and Moore, 1987; Carlyon et al., 1989) and a secondary goal of many more. It is widely believed that “true CMR,” based on across-channel effects, can be small relative to the total masking release observed for closely spaced masker bands.

Previous studies have attempted to discriminate within- from across-channel effects in CMR using modeling (Verhey et al., 1999), manipulation of stimulus features thought to disrupt across-channel processes (e.g., asynchronous gating; Dau et al., 2004), adjustment in the spectral proximity of masker bands (Schooneveldt and Moore, 1987), and dichotic stimulus presentation, wherein flanking masker bands are presented contralateral to the signal and on-signal band (Cohen and Schubert, 1987; Schooneveldt and Moore, 1987). Each of these approaches has both strengths and weaknesses. For example, recent work in our laboratory suggests that asynchronous gating can disrupt both within- and across-channel processes. Dichotic CMR results are often interpreted cautiously; some reports suggest that the cues underlying dichotic CMR may differ from those responsible for monaural CMR (e.g., Ernst and Verhey, 2008), though study of the combination of cues across ears has failed to uncover any substantive differences between monaural and dichotic CMR (Schooneveldt and Moore, 1989). This concern aside, the dichotic CMR paradigm is arguably the most straightforward means of eliminating within-channel cues while holding other stimulus features constant. For this reason, dichotic CMR was chosen for further study in the present investigation.

Dichotic CMR findings have been quite variable, ranging from no dichotic CMR to values comparable to those reported for monaural CMR. The wide variety of stimulus parameters used across studies could play a role in the range of results. Two studies that reported a dichotic CMR comparable to that found with monaural stimuli used two maskers which played continuously throughout a run (Cohen and Schubert, 1987; Schooneveldt and Moore, 1987). Using gated maskers, Hicks and Bacon (1995) reported no evidence of dichotic CMR, and Ernst and Verhey (2006) found no dichotic CMR for gated stimuli unless the flanking band level greatly exceeded the on-signal band level. Studies using monaural stimuli often report greater CMR for continuous than gated maskers (Fantini et al., 1993), an effect that may be inversely related to the number of flanking bands (Hatch et al., 1995). One goal of the present study was to test the hypothesis that dichotic CMR is similarly affected by masker gating, with a smaller dichotic CMR for gated as opposed to continuous stimuli.

Several reports of monaural CMR suggest that the outputs of auditory channels carrying information about the masker complex must be perceptually grouped in order to support the beneficial effects of coherent masker modulation. This conclusion rests primarily on the reduction or elimination of CMR under conditions of asynchronous onset of the on-signal and flanking maskers (Grose and Hall, 1993; Dau et al., 2004). While all of the above-cited gated dichotic CMR studies used synchronous onset across bands, contralateral masker presentation could itself introduce a strong segregation cue in that the signal and masker bands can be lateralized to opposite sides of the head. While some studies report a large dichotic CMR for pairs of continuous maskers (e.g., Cohen and Schubert, 1987), others using a family of flanking bands have reported a dichotic CMR about 50% of the analogous monaural CMR (Hall et al., 1990; Moore and Shailer, 1991). The finding of reduced dichotic CMR for multiple flanking maskers could reflect an increased tendency for stream segregation due to perceptual dissimilarity between the on-signal band and the family of contralateral flanking maskers. In the present study, it was hypothesized that distributing a family of flanking bands across ears would reduce the tendency to segregate the on-signal and flanking masker bands and thus increase the masking release produced by the contralateral flanking bands.

METHODS

Observers

Observers were ten normal hearing adults, four males and six females. All had pure tone thresholds of 20 dB HL or less at octave frequencies from 250 to 8000 Hz (ANSI, 1996) and no significant history of ear disease. Their ages ranged from 17 to 53, with a mean age of 29 years.

Stimuli

The signal was a 2 kHz pure tone, 400 ms in duration, including 50 ms raised-cosine ramps. The masker was one or more 20 Hz wide bands of Gaussian noise presented at 50 dB spectrum level. There was always an on-signal masker band centered on 2 kHz. When two flanking maskers were present they were centered on 1.6 and 2.4 kHz, the fourth and sixth harmonics of 400 Hz. When eight flankers were present the masker consisted of bands centered on the first to ninth harmonics of 400 Hz.

The signal and the on-signal masker band were always presented to the right ear. In the ipsilateral conditions, additional flanking bands were presented to the right ear, and in the contralateral conditions those flanking bands were presented to the left ear. The mixed condition included two flankers ipsilateral to the signal (the fourth and sixth harmonics of 400 Hz) and six flankers to the contralateral ear (first to third and seventh to ninth harmonics of 400 Hz). Maskers were gated on and off synchronously with the 400 ms signal or were presented continuously.

Maskers were generated in the frequency domain based on 218 points, which when played at 24.4 kHz resulted in a 10.7 s sample that repeated seamlessly. The on-signal band was generated based on random Gaussian draws defining the real and imaginary components associated with bins within the range ±10 Hz around 2 kHz. When flanking maskers were present, these same random draws were used to define the bands at different center frequencies.

Procedures

Stimuli were presented as a three-alternative forced choice, with the signal interval selected at random. Listening intervals were marked visually, with a 300 ms interstimulus interval. Correct answer feedback was provided after the observer response. The signal level was adjusted with a three-down one-up procedure (Levitt, 1971) estimating the 79% correct point. The initial step size was 4 dB prior to the second reversal and 2 dB thereafter. A track continued until eight reversals had been obtained, and the resulting threshold estimate was the average signal level at the last six reversals. Three such estimates were obtained in each condition, with a fourth collected in cases where the first three spanned a range of 3 dB or more. Thresholds were obtained in blocks by condition, and observers went through conditions in quasirandom order.

RESULTS AND DISCUSSION

Mean data are shown in Fig. 1. Thresholds are plotted as a function of the flanking masker condition, with symbols indicating ear of presentation (shading) and whether the masker was gated (squares) or continuous (circles). Error bars indicate 1 standard error (s.e.) of the mean. Thresholds in the on-signal condition (labeled “none” in Fig. 1) were quite similar in the gated and continuous conditions (t9=0.84, p=0.42), with means of 65.9 and 65.4 dB SPL, respectively.

Figure 1.

Figure 1

Open in a new tab

Mean thresholds are plotted as a function of flanking masker condition, with error bars indicating 1 s.e. of the mean. Symbols reflect the mode of presentation. Symbol shapes distinguish gated (squares) from continuous (circles) masker presentation. Symbol shadings distinguish ipsilateral (open), contralateral (solid), and mixed (gray-filled) flanker presentation.

Ipsilateral and contralateral conditions

Gated conditions will be considered first. For ipsilateral flanking bands, introducing two flanking masker bands decreased thresholds by only 1.4 dB, and eight flanking bands decreased thresholds by 6.6 dB. In contrast, there was little evidence of masking release for either number of bands presented contralaterally, with mean thresholds rising 1.6 dB for two bands and falling 0.9 dB for eight bands; thresholds in both conditions were within the 95% confidence interval around the on-signal masker threshold. For continuous ipsilateral masker presentation there was no effect of masker number, with masking release of 12.7 dB for both two and eight bands. Continuous contralateral maskers were slightly less effective when there were two bands as compared to eight bands, with 6.4 and 8.0 dB masking release, respectively.

A repeated-measures analysis of variance (ANOVA) was performed on estimates of masking release, calculated as the change in threshold relative to the associated on-signal threshold (either gated or continuous). There were two levels of GATING (gated, continuous), NUMBER (two bands, eight bands) and EAR (ipsilateral, contralateral). All three main effects were significant: GATING (F1,9=44.08, p<0.0001), NUMBER (F1,9=14.29, p<0.005), and EAR (F1,9=24.27, p<0.001). The interaction between GATING and NUMBER just failed to reach significance (F1,9=4.5, p=0.06), and the other two-way interactions were not significant (p>=0.45). The three-way interaction was significant (F1,9=7.20, p<0.05). These results are consistent with the interpretation that the GATING×NUMBER interaction demonstrated in previous studies with monaural stimuli is absent in the dichotic data. This result could be attributed to fundamental differences between monaural and dichotic CMR or to the fact that there are insufficient cues to promote perceptual integration of the on-signal and flanking bands in the dichotic gated condition.

Mixed conditions

Attention now turns to the mixed conditions, wherein two flanking masker bands were presented to the same ear as the signal and the remaining six bands were presented contralaterally. Thresholds in these conditions were quite low, with average masking release of 8.7 and 14.7 dB in the gated and continuous conditions, respectively. These values can be compared to masking release of 6.6 and 12.7 dB obtained in comparable eight-band ipsilateral masker conditions. Masking release in the mixed conditions was assessed with a repeated-measures ANOVA, with two levels of EAR (eight-band ipsilateral, mixed) and two levels of GATING (gated, continuous). There was a significant main effect of EAR (F1,9=59.02, p<0.0001) and of GATING (65.31, p<0.0001), but no interaction (F1,9=0.76, p=0.41). This outcome suggests that mixed presentation not only overcomes any reduction in CMR associated with contralateral flanking band presentation, but it may provide additional cues not present in the ipsilateral masker condition. It is not clear how this result comes about, but it could be related to previous findings of greater dichotic than monaural CMR in some conditions (Cohen and Schubert, 1987; Schooneveldt and Moore, 1989) or to previous reports of better diotic than monaural thresholds under conditions where performance is limited by internal noise (Langhans and Kohlrausch, 1992). It is possible that distributing flanking maskers across ears increases the number of independent auditory filters carrying information about the masker envelope; this possibility receives indirect support from the argument of Moore et al. (1993) that monaural CMR is reduced in hearing impaired listeners due to reduced frequency selectivity, but that this effect can be counteracted by presenting flanking bands contralateral to the signal and on-signal masker.

The most striking aspect of the mixed data is the 8.7 dB CMR obtained in the gated condition. If listeners were making use of cues present only in the signal ear, a modest CMR would be expected, similar to the 1.4 dB CMR for the two-band ipsilateral condition. The six contralateral flanking maskers in combination with these ipsilateral bands supported a robust CMR despite the fact that CMR was absent in the eight-band contralateral condition. This result is consistent with the hypothesis that CMR obtained with contralateral flanking bands may be strongly affected by auditory segregation. If bilaterally distributing masker bands improves performance by facilitating grouping, this result is consistent with the idea that little or no dichotic CMR is observed for gated stimuli due to a failure to group gated stimuli into a single auditory stream.

Central versus peripheral effects

Results presented here are consistent with the hypothesis that central effects related to auditory grouping can play a large role in the size of dichotic CMR. If grouping of comodulated bands is reduced by factors such as gating, a large disparity between numbers of bands presented to each ear, and differential lateralization, then the relative size of monaural and dichotic CMR should not be interpreted solely in terms of peripheral contributions to masking release. Results from the mixed conditions of this experiment are consistent with an interpretation that the poor CMR associated with the dichotic gated condition is related to a failure to process the on-signal and flanking bands as emanating from a single source. Gated masker bands in the nonsignal ear can be quite effective in contributing to making release when cues are provided that promote the fusion of bands across the ears.

SUMMARY

  • (1)

    With all flanking masker bands contralateral to the signal, a dichotic CMR was obtained for continuous but not gated presentation.

  • (2)

    Distributing flanking maskers across ears resulted in a relatively large CMR. This result was most striking for the gated masker presentation mode: Contralateral maskers in the mixed condition improved thresholds by 7 dB compared to the ipsilateral condition with just two flanking masker bands.

  • (3)

    These results are consistent with the hypothesis that the absence of a dichotic CMR for gated maskers may be due to perceptual segregation of the stimulus components presented to the two ears, and that the mixed dichotic presentation promotes perceptual grouping of those components.

  • (4)

    Differences between monaural and dichotic CMR cannot be attributed solely to peripheral effects.

ACKNOWLEDGMENTS

This work was supported by grants from the NIH NIDCD (Nos. R01 DC000418 and R01 DC007391). The authors thank Armin Kohlrausch, John Grose, Brian Moore, and two anonymous reviewers for contributing to this work.

References

  1. ANSI (1996). ANSI S3-1996, American National Standards Specification for Audiometers (American National Standards Institute, New York: ). [Google Scholar]
  2. Carlyon, R. P., Buus, S., and Florentine, M. (1989). “Comodulation masking release for three types of modulator as a function of modulation rate,” Hear. Res. 10.1016/0378-5955(89)90116-0 42, 37–45. [DOI] [PubMed] [Google Scholar]
  3. Cohen, M. F., and Schubert, E. D. (1987). “Influence of place synchrony on detection of a sinusoid,” J. Acoust. Soc. Am. 10.1121/1.394910 81, 452–458. [DOI] [PubMed] [Google Scholar]
  4. Dau, T., Ewert, S. D., and Oxenham, A. J. (2004). “Effects of concurrent and sequential streaming in comodulation masking release,” in Auditory Signal Processing: Physiology, Psychoacoustics and Models, edited by Pressnitzer D., de Cheveigne A., McAdams S., and Collet L. (Springer, Berlin: ), pp. 335–343. [Google Scholar]
  5. Ernst, S. M., and Verhey, J. L. (2006). “Role of suppression and retro-cochlear processes in comodulation masking release,” J. Acoust. Soc. Am. 10.1121/1.2361183 120, 3843–3852. [DOI] [PubMed] [Google Scholar]
  6. Ernst, S. M., and Verhey, J. L. (2008). “Peripheral and central aspects of auditory across-frequency processing,” Brain Res. 1220, 246–255 [DOI] [PubMed] [Google Scholar]
  7. Fantini, D. A., Moore, B. C. J., and Schooneveldt, G. P. (1993). “Comodulation masking release as a function of type of signal, gated or continuous masking, monaural or dichotic presentation of flanking bands, and center frequency,” J. Acoust. Soc. Am. 10.1121/1.406697 93, 2106–2115. [DOI] [PubMed] [Google Scholar]
  8. Grose, J. H., and Hall, J. W. (1993). “Comodulation masking release: Is comodulation sufficient?,” J. Acoust. Soc. Am. 10.1121/1.405809 93, 2896–2902. [DOI] [PubMed] [Google Scholar]
  9. Hall, J. W., Grose, J. H., and Haggard, M. P. (1990). “Effects of flanking band proximity, number, and modulation pattern on comodulation masking release,” J. Acoust. Soc. Am. 10.1121/1.399294 87, 269–283. [DOI] [PubMed] [Google Scholar]
  10. Hall, J. W., Haggard, M. P., and Fernandes, M. A. (1984). “Detection in noise by spectro-temporal pattern analysis,” J. Acoust. Soc. Am. 10.1121/1.391005 76, 50–56. [DOI] [PubMed] [Google Scholar]
  11. Hatch, D. R., Arne, B. C., and Hall, J. W. (1995). “Comodulation masking release (CMR): Effects of gating as a function of number of flanking bands and masker bandwidth,” J. Acoust. Soc. Am. 10.1121/1.412392 97, 3768–3774. [DOI] [PubMed] [Google Scholar]
  12. Hicks, M. L., and Bacon, S. P. (1995). “Some factors influencing comodulation masking release and across-channel masking,” J. Acoust. Soc. Am. 10.1121/1.413216 98, 2504–2514. [DOI] [PubMed] [Google Scholar]
  13. Langhans, A., and Kohlrausch, A. (1992). “Differences in auditory performance between monaural and dichotic conditions. I. Masking thresholds in frozen noise,” J. Acoust. Soc. Am. 10.1121/1.402834 91, 3456–3470. [DOI] [PubMed] [Google Scholar]
  14. Levitt, H. (1971). “Transformed up-down methods in psychoacoustics,” J. Acoust. Soc. Am. 10.1121/1.1912375 49, 467–477. [DOI] [PubMed] [Google Scholar]
  15. Moore, B. C. J., and Shailer, M. J. (1991). “Comodulation masking release as a function of level,” J. Acoust. Soc. Am. 10.1121/1.401950 90, 829–835. [DOI] [PubMed] [Google Scholar]
  16. Moore, B. C. J., Shailer, M. J., Hall, J. W., and Schooneveldt, G. P. (1993). “Comodulation masking release in subjects with unilateral and bilateral hearing impairment,” J. Acoust. Soc. Am. 93, 435–451. [DOI] [PubMed] [Google Scholar]
  17. Oxenham, A. J., and Plack, C. J. (1998). “Suppression and the upward spread of masking,” J. Acoust. Soc. Am. 10.1121/1.423933 104, 3500–3510. [DOI] [PubMed] [Google Scholar]
  18. Schooneveldt, G. P., and Moore, B. C. (1987). “Comodulation masking release (CMR): Effects of signal frequency, flanking-band frequency, masker bandwidth, flanking-band level, and monotic versus dichotic presentation of the flanking band,” J. Acoust. Soc. Am. 10.1121/1.395639 82, 1944–1956. [DOI] [PubMed] [Google Scholar]
  19. Schooneveldt, G. P., and Moore, B. C. J. (1989). “Comodulation masking release for various monaural and binaural combinations of the signal, on-frequency, and flanking bands,” J. Acoust. Soc. Am. 10.1121/1.397733 85, 262–272. [DOI] [PubMed] [Google Scholar]
  20. Verhey, J. L., Dau, T., and Kollmeier, B. (1999). “Within-channel cues in comodulation masking release (CMR): Experiments and model predictions using a modulation-filterbank model,” J. Acoust. Soc. Am. 10.1121/1.428101 106, 2733–2745. [DOI] [PubMed] [Google Scholar]

Articles from The Journal of the Acoustical Society of America are provided here courtesy of Acoustical Society of America

RESOURCES