Elsevier

Hearing Research

Volume 298, April 2013, Pages 10-16
Hearing Research

Research paper
The perceptual enhancement of tones by frequency shifts

https://doi.org/10.1016/j.heares.2013.01.016Get rights and content

Abstract

In a chord of pure tones with a flat spectral profile, one tone can be perceptually enhanced relative to the other tones by the previous presentation of a slightly different chord. “Intensity enhancement” (IE) is obtained when the component tones of the two chords have the same frequencies, but in the first chord the target of enhancement is attenuated relative to the other tones. “Frequency enhancement” (FE) is obtained when both chords have a flat spectral profile, but the target of enhancement shifts in frequency from the first to the second chord. We report here an experiment in which IE and FE were measured using a task requiring the listener to indicate whether or not the second chord included a tone identical to a subsequent probe tone. The results showed that a global attenuation of the first chord relative to the second chord disrupted IE more than FE. This suggests that the mechanisms of IE and FE are not the same. In accordance with this suggestion, computations of the auditory excitation patterns produced by the chords indicate that the mechanism of IE is not sufficient to explain FE for small frequency shifts.

Highlights

► In a chord, a tone is perceptually enhanced by a shift in its frequency. ► This occurs even when the enhancing stimulus is softer than the subsequent stimulus. ► Frequency shifts of a few percent are sufficient to produce strong enhancement. ► The underlying mechanism seems to differ from that of other forms of enhancement.

Introduction

The auditory system tends to “enhance” acoustic changes consisting of the addition of power to a restricted part of the spectrum of a complex sound. Thus, for example, in a chord of pure tones with equal intensities, one of the tones (X) can be made to pop out perceptually by presenting, before this chord, a copy of it in which X is attenuated. Many variants of this enhancement effect have been described in the literature (Schouten, 1940; Wilson, 1970; Viemeister, 1980; Summerfield et al., 1987; Carlyon, 1989; Wright et al., 1993; Hartmann and Goupell, 2006; Serman et al., 2008; Cao and Richards, 2012; Cervantes Constantino et al., 2012). In everyday life, enhancement is presumably helpful for the perceptual segregation of new acoustic events in sound mixtures (Bregman, 1990).

A plausible explanation for enhancement is neural adaptation. In our example, the neural response to tone X in the second chord can be expected to be stronger than the neural responses to the other components of this chord because the latter responses should show more adaptation following the first chord. A related explanation is “adaptation of inhibition” (Viemeister and Bacon, 1982; Byrne et al., 2011): the first chord could reduce, in the second chord, the inhibition of neurons responding to X by neurons responding to other components of the chord. Nelson and Young (2010) have recently reported physiological data supporting the latter explanation. Remarkably, enhancement is still observable when the enhancing stimulus (hereafter called the “precursor” stimulus) and the subsequent stimulus (hereafter called the “test” stimulus) are separated by several seconds (Viemeister, 1980; Cao and Richards, 2012), or are presented to separate ears (Richards et al., 2004; Kidd et al., 2011; Erviti et al., 2011; Carcagno et al., 2012, in press), or are electrical signals directly exciting the auditory nerve through a cochlear implant (Goupell and Mostardi, 2012; Wang et al., 2012). These three facts indicate that enhancement cannot entirely originate from peripheral processes. Its main source could nevertheless be some form of stimulus-specific neural adaptation taking place at central levels of the auditory system (Ulanovsky et al., 2003; Malmierca et al., 2009; Antunes et al., 2010).

Erviti et al. (2011) observed that tones within chords can be strongly enhanced perceptually not only as a result of shifts in relative intensity – i.e., changes in “spectral profile” (Green, 1988) – but also as a result of shifts in frequency. In an array of frequency-selective neurons acting as bandpass filters and tuned to a wide range of frequencies, a shift in frequency will increase the excitation of some neurons, like an increase in intensity. In theory, therefore, it could be that the mechanism underlying the enhancement produced by frequency shifts (called “frequency enhancement” (FE) by Erviti et al. (2011) is the same as the mechanism underlying “classical” enhancement (called “intensity enhancement” (IE) by Erviti et al.). However, Erviti et al. provided one argument against that view. They found that presenting the precursor and test stimuli to opposite ears rather than to the same ear significantly diminishes IE but has little effect on FE.

In the present paper, we provide two additional pieces of evidence against the idea that FE can be fully explained in the same manner as IE. First, we report a psychophysical experiment comparing FE with IE regarding their sensitivity to a global attenuation of the precursor stimulus relative to the test stimulus. In most of the past experiments on enhancement, the background components of the stimuli (i.e., the stimulus components which were not the target of enhancement) had the same intensity in the precursor and the test stimulus. The effect of a global attenuation of the precursor on IE has been investigated by Carlyon (1989) and Viemeister et al. (in press). These two studies led to discrepant conclusions: while Carlyon (1989) found very little effect of precursor intensity in a 30-dB range, Viemeister et al. (in press) found that a 10-dB attenuation of the precursor background relative to the test background was sufficient to reduce IE substantially. The effect of such manipulations on FE has not been documented up to now.

We also report here the results of computations based on the excitation pattern model of Moore et al. (1997), revised by Glasberg and Moore (2006). Using this model, it is possible to estimate how any stationary sound stimulus excites a bank of filters consistent with the frequency selectivity of masking in humans. One can therefore compute the difference between the excitation patterns of two pure-tone chords, P and T, identical to each other except for one tone and such that the sequence P–T enhances this tone. For a given T chord and enhancement target, consider the case of two P chords, PF and PI, eliciting respectively FE and IE with the same strength. If FE and IE originate from the same process, namely some form of frequency-selective neural adaptation, the spectrally local increases of excitation produced by the sequences PF–T and PI–T should be similar in magnitude. Our computations were intended to test this prediction.

Section snippets

Overview

The experimental task was the same as that used by Erviti et al. (2011). On each trial, the listener was presented with three successive stimuli: a precursor chord, a test chord, and a probe tone. The test chord consisted of five synchronous pure tones with the same nominal sound pressure level (SPL); the frequencies of its components were renewed from trial to trial, according to rules specified below (Section 2.1.2). The following probe tone was, equiprobably, either identical in every

Computations of excitation patterns

We consider here the expected auditory excitation patterns of three prototypical samples from the stimulus set used in the experiment described above. One of these three stimuli, T, is the “mean” (or equivalently the “median”) of all the possibly selectable test chords in the SPL condition 70-70. The frequencies of its five component tones, separated by intervals of 8 semitones, are 317.5, 504.0, 800.0, 1269.9, and 2015.9 Hz. A second stimulus, PF, is the precursor chord that would have been

Discussion

We provide here two new arguments against the idea that FE can be completely explained in the same manner as IE, and more specifically as the consequence of frequency-specific adaptation in sets of neurons encoding only the spectral characteristics of sounds. One argument, stemming from the excitation-pattern computations, is that strong FE can be obtained with frequency shifts producing excitation increases that are not sufficient to result in strong IE. The other argument is based on our

Acknowledgment

This work was supported by a grant from the Agence Nationale de la Recherche (ANR-2010-BLAN-1906-02).

References (37)

  • F.M. Antunes et al.

    Stimulus-specific adaptation in the auditory thalamus of the anesthetized rat

    PloS One

    (2010)
  • A.S. Bregman

    Auditory Scene Analysis – The Perceptual Organization of Sound

    (1990)
  • A.J. Byrne et al.

    The enhancement effect: evidence for adaptation of inhibition using a binaural centering task

    J. Acoust. Soc. Am.

    (2011)
  • X. Cao et al.

    Enhancement in informational masking

    J. Speech Lang. Hear. Res.

    (2012)
  • S. Carcagno et al.

    Frequency-shift detectors bind binaural as well as monaural frequency representations

    J. Exp. Psychol. Hum. Percept. Perform.

    (2011)
  • S. Carcagno et al.

    Auditory enhancement of increments in spectral amplitude stems from more than one source

    J. Assoc. Res. Otolaryngol.

    (2012)
  • Carcagno, S., Semal, C., Demany, L. Enhancement of increments in spectral amplitude: further evidence for a mechanism...
  • R.P. Carlyon

    Changes in the masked thresholds of brief tones produced by prior bursts of noise

    Hear. Res.

    (1989)
  • F. Cervantes Constantino et al.

    Detection of appearing and disappearing objects in complex acoustic scenes

    PloS One

    (2012)
  • M. Cousineau et al.

    What makes a melody: the perceptual singularity of pitch sequences

    J. Acoust. Soc. Am.

    (2009)
  • L. Demany et al.

    Tuning properties of the auditory frequency-shift detectors

    J. Acoust. Soc. Am.

    (2009)
  • L. Demany et al.

    On the binding of successive sounds: perceiving shifts in nonperceived pitches

    J. Acoust. Soc. Am.

    (2005)
  • L. Demany et al.

    Fundamental differences in change detection between vision and audition

    Exp. Brain Res.

    (2010)
  • L. Demany et al.

    Implicit versus explicit frequency comparisons: two mechanisms of auditory change detection

    J. Exp. Psychol. Hum. Percept. Perform.

    (2011)
  • M. Erviti et al.

    Enhancing a tone by shifting its frequency or intensity

    J. Acoust. Soc. Am.

    (2011)
  • B.R. Glasberg et al.

    Prediction of absolute thresholds and equal-loudness contours using a modified loudness model

    J. Acoust. Soc. Am.

    (2006)
  • M.J. Goupell et al.

    Evidence of the enhancement effect in electrical stimulation via electrode matching

    J. Acoust. Soc. Am.

    (2012)
  • D.M. Green

    Profile Analysis – Auditory Intensity Discrimination

    (1988)
  • Cited by (11)

    • Automatic Frequency-Shift Detection in the Auditory System: A Review of Psychophysical Findings

      2018, Neuroscience
      Citation Excerpt :

      If the precursor chord is identical to the test chord except for a slight (e.g., 1-semitone) frequency shift in one tone, then this component of the test chord will also pop out perceptually when the test chord is heard. This form of enhancement has been called “frequency enhancement” (FE), and compared to the classical form, “intensity enhancement” (IE), in three studies (Erviti et al., 2011; Demany et al., 2013; Byrne et al., 2013). A priori, given the tonotopic organization of the auditory system, one might suppose that FE and IE are based on a common mechanism: in an array of frequency-tuned neurons acting as bandpass filters, a frequency shift of one tone will increase the excitation of some neurons, like an increase in intensity.

    • Detecting temporal changes in acoustic scenes: The variable benefit of selective attention

      2017, Hearing Research
      Citation Excerpt :

      However, given that we used rather low modulation frequencies, these spectral effects were negligible: according to the excitation-pattern model of Moore et al. (1997) revised by Glasberg and Moore (2006), there could be no auditory filter in which a modulation frequency change of 600 cents produced an excitation change exceeding 1 dB, and the amount of excitation change was most often well below 1 dB. The fact that the changes produced in each of our experiments could not be detected in the spectral domain may well be one of the reasons why change detection was so dependent on selective attention in these experiments: the detection of spectral changes is facilitated by automatic neural adaptation processes taking place sub-cortically (Palmer et al., 1995; Nelson and Young, 2010) as well as by other automatic processes (Carcagno et al., 2011; Demany and Ramos, 2005; Demany et al., 2009, 2010, 2011; Erviti et al., 2011; Demany et al., 2013; Moore et al., 2013). Experiments 1 and 2 suggested that a change in the modulation frequency of a tone does not draw attention to a larger extent when this change creates a discontinuity in the modulation domain than when that is not the case (the scene is always discontinuous, independently of the change to be detected): in the absence of pre-cues, performance was not significantly better in the Continuous conditions than in the Break conditions.

    • Effects of auditory enhancement on the loudness of masker and target components

      2016, Hearing Research
      Citation Excerpt :

      Nevertheless, any such bias should not affect the conclusions of the current study, as the bias should be eliminated by comparing the comparison level in the reference condition with that measured in the test conditions. In a typical auditory enhancement experiment, the loudness of the signal tone appears to be enhanced relative to the loudness of the flanking masker tones (e.g. Byrne et al., 2011; Demany et al., 2013). The results from experiment 1 show that neither the loudness of the signal in isolation (experiment 1A) nor the loudness of the flanking masking tones alone (experiment 1B) is affected by the presence of a precursor when presented at the same level as the signal and masker.

    • Neural auditory contrast enhancement in humans

      2021, Proceedings of the National Academy of Sciences of the United States of America
    • Effect of Context on the Contribution of Individual Harmonics to Residue Pitch

      2017, JARO - Journal of the Association for Research in Otolaryngology
    View all citing articles on Scopus
    View full text