| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1Section of Biomagnetism, Department of Neurology, University of Heidelberg, Heidelberg, Germany; 2Chair for Philosophy, Eidgenössische Technische Hochschule Zürich, Zurich, Switzerland; and 3Centre for Applied Hearing Research, Department of Electrical Engineering, Technical University of Denmark, Lyngby, Denmark
Submitted 10 July 2007; accepted in final form 6 January 2008
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Glasberg and Moore 1990; Patterson 1976; Plomp 1964; Zwicker et al. 1957). Conversely, filter phase responses received much less attention, partly due to the popular belief that the ear is essentially "phase deaf" (Helmholtz 1954) and the observation that many perceptual properties can be explained by analyzing the power spectrum of a sound—i.e., by discarding the phase information. Although a number of studies have since shown that the ear is sensitive to changes in stimulus phase, both within an auditory filter (Goldstein 1967; Mathes and Miller 1947; Zwicker 1952) and, to a lesser extent, across filters (Patterson 1987), they have mostly assumed that any effect of the auditory filters themselves on the stimulus phase can be ignored. Indeed, there are only certain aspects of the phase response that have any psychophysically measurable influence. For instance, both the absolute phase and the group delay, proportional to d
/df, are generally meaningful only in the context of a fixed time reference and cannot be estimated psychophysically (Goldstein 1967; Kohlrausch and Sander 1995). When defining individual auditory filters, the first term that has any perceptual relevance is the phase curvature, d2/df2, which is the rate of change of group delay as a function of frequency.
Psychoacoustic experiments conducted by Smith et al. (1986) and Kohlrausch and Sander (1995) provided examples of instances where the phase response of the auditory filters cannot be ignored. Both studies used equal-amplitude harmonic-tone complexes with the phases set according to Schroeder (1970), which have come to be known as Schroeder-phase stimuli. Schroeder-phase stimuli have constant phase curvature, implying that the frequency sweep rate is constant (see Kohlrausch and Sander 1995). These stimuli are characterized by very flat temporal envelopes and can be thought of as repeating instantaneously linear rising (Schroeder negative, m–) or falling (Schroeder positive, m+) frequency sweeps. Examples of m– and m+ stimuli that have been used in Kohlrausch and Sander (1995) and also in the current study are illustrated in Fig. 1.
|
; Glasberg and Moore 1990; Patterson 1976), which predicts the same thresholds for m+ and m– maskers. Second, these complexes provide insight into the phase response of the auditory filters. If frequency-modulated (FM) stimuli, such as an m+ complex and the corresponding (temporally reversed) m– complex, are passed through a band-pass (auditory) filter, the FM is transferred into an amplitude modulation (AM). For a linear-phase filter, this would result in the same degree of modulation for the two stimuli after filtering. However, it has been proposed that auditory filtering at the level of the basilar membrane (BM) (Rhode and Recio 2000) alters the waveform such that m+ complexes result in a highly modulated or "peaky" envelope, whereas m– complexes produce a less modulated, flatter envelope. This in turn places constraints on the form of the auditory system's phase response (Kohlrausch and Sander 1995).
Certain semirealistic models of BM filtering (e.g., Giguère and Woodland 1994; Strube 1985) support this view. The output of these models is more highly modulated in response to an m+ complex than to an m– complex. One reason for the qualitative success of BM models in accounting for the perceptual difference between the m+ and m– complexes is that the phase response of these models has a negative curvature throughout most of the passband. In the case of the m+ complex, this negative curvature of the BM filter compensates the positive curvature of the stimulus, leading to a filtered waveform in which the starting phases of all the components come close to coinciding, producing a peaky envelope. Thus when the task is to detect a tone in the presence of the maskers, it appears that the much lower detection thresholds in the case of the m+ masker complex are due to the possibility of listening in the "dips" of the envelope at the output of the auditory filter (Gockel et al. 2002, 2003; Moore 2004).
Although these studies suggest a relationship between the dispersive basilar membrane transformation and results from perceptual masking, many details of the processing, such as the dependence of the phase response on stimulation level and amount of frequency selectivity, have yet to be elucidated in humans, and modeling has been only at a relatively qualitative level. In fact, information about human auditory signal processing can typically be derived only indirectly with noninvasive measurements. Furthermore, it is not clear how phase sensitivity of peripheral auditory processing is reflected at higher stages beyond cochlear processing. Specifically, it is unknown how the phase-sensitive masking effects as described earlier are represented in the central auditory system, up to the level of the human auditory cortex, and how the activation of these generators is related to perception. Here, we investigated the neural (objective) correlate of the perceptual findings obtained with Schroeder tone-complex maskers in the auditory cortex using whole head magnetoencephalography (MEG). The Pam and P1m responses of auditory evoked potentials and fields (AEPs and AEFs) play a major role in the investigation of the primary and secondary areas of the auditory cortex since there is converging evidence given by electro- and magnetoencephalographic as well as intracranial studies that these components originate, respectively, from the medial and lateral portion of Heschl's gyrus (Liégeois-Chauvel et al. 1994; Scherg et al. 1989, 1990). These responses are considered here to study the representation of auditory-filter phase dependence of the cochlea in the human auditory cortex.
In experiment 1, we investigated the AEFs evoked by long-duration tones presented in m+ and m– Schroeder-phase maskers, with parameters similar to those used in the psychoacoustic study of Kohlrausch and Sander (1995). In experiment 2, the neuromagnetic responses to brief tone pulses presented at different temporal positions within one period of the complex maskers were examined and compared with patterns from loudness experiments, using the same stimuli and derived from the same group of listeners. The results of both experiments were then correlated with simulations of neural activity patterns, using the auditory image model (AIM) of Patterson et al. (1995) that estimates the firing rate in the auditory nerve. These investigations were undertaken in an attempt to better understand 1) how the characteristics of peripheral signal processing are represented at the cortical level of processing in human subjects and 2) to determine the extent to which neuromagnetic responses can be interpreted as an objective correlate of perception.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
SUBJECTS. Fourteen subjects for the first experiment (eight female and six male, ages 18 to 43 yr, 13 right-handed) had normal audiometric thresholds and showed no history of peripheral or central hearing disorders. All subjects were familiar with MEG recording sessions and provided informed consent before participating in the experiment, approved by the local ethics committee. Eleven subjects (four female, seven male, ages 24 to 40 yr, ten right-handed) participated in the second experiment with nine overlapping in the first experiment.
STIMULI.
Like Kohlrausch and Sander (1995), we used maskers consisting of equal-amplitude harmonics (range 2 to 20) of a fundamental frequency set at f0 = 100 Hz; the starting phase of component n was assigned according to n = –
n(n – 1)/N, where N is the total number of components (in this case 19). Thus the complex is given by: 
n=220 A0 sin (2nf0t + n). The masker level was 70 dB SPL, corresponding to a level of the individual harmonics of about 57 dB SPL. The complexes derived using the "+" and "–" sign represent, respectively, the Schroeder positive (m+) and the Schroeder negative (m–) maskers. The resulting waveforms are shown in Fig. 1. The duration of the maskers was 100 ms and the onsets and offsets were gated using a 10-ms Hanning window.
The signal was a 1,000-Hz tone temporally centered in the masker, as indicated by the gray function in Fig. 1A. Its duration was 50 ms and included a 3-ms Hanning window for the onset and its offset. The signal was added in phase to the spectral component of the respective masker at the signal frequency. Two different signal levels for the MEG experiments were applied: 57 and 69 dB SPL, corresponding to 0 and 12 dB relative to the level of the individual masker components. These two levels were chosen in pilot listening experiments such that the signal was clearly detectable at both levels in the case of the m+ masker and clearly detectable only at the higher level in the case of the m– masker. Additionally, both maskers were presented without the signal to allow for the investigation of the signal-related responses given by the difference between masker-plus-signal response and masker-alone response.
Stimuli were generated digitally with a sampling frequency of 44.1 kHz. D/A-conversion was performed using an audio soundcard (RME Audio, Haimhausen, Germany) connected to a PC. Sounds were presented diotically with a custom-made sound list processor via ER-3 (Etymotic Research, Elk Grove Village, IL) earphones connected to 90-cm plastic tubes and foam earpieces. The interstimulus interval (ISI) was set to 270 ms, measured between the offset and onset of subsequently occurring maskers. The level was adjusted using a Brüel & Kjær sound-level meter type 2203 with an artificial ear (2 cc coupler) 4152. This equipment was attached to a digital oscilloscope (WaveSurver 424, LeCroy, Chestnut Ridge, NY) and used to investigate the phase and amplitude characteristics of the ER-3 devices. The acoustic waveform of the m+ masker is illustrated in Fig. 1B (gray) together with the digital waveform (black). The distortion due to the transmission through earphones and tubes is minimal.
DATA ACQUISITION. The gradients of the magnetic fields were recorded continuously with a Neuromag-122 whole head system (Elekta Neuromag Oy, Helsinki, Finland) inside a magnetically shielded room (IMEDCO, Hägendorf, Switzerland). Subjects sat in an upright position, viewed a silent movie of their choice, and listened passively to the stimuli. Four coils were attached to the scalp to determine the head position under the dewar during the recordings. Each MEG registration for both the m+ and m– Schroeder-phase masking conditions was about 30 min in duration.
DATA ANALYSIS.
Spatiotemporal dipole source analysis (Scherg 1990) of the auditory evoked fields was effected with the BESA (brain electrical source analysis) 5.1 software package (MEGIS Software, Munich, Germany). A spherical model was used for modeling and the position of the sphere was fitted to the individual surface of the head that had been digitized using a Fastrak system (Polhemus, Colchester, VT). Averaging with artifact monitoring was used to increase the signal-to-noise ratio. Single sweeps with MEG signal exceeding a peak level of 8,000 fT or with a gradient of 800 fT per sample were rejected. For each stimulus condition, about 1,200 single sweeps covering the range from 50 ms before to 300 ms after stimulus onset were averaged, thus providing a sufficient signal-to-noise ratio for a spatiotemporal source analysis of the early auditory evoked fields (AEFs).
Prior to source analysis, the averaged data were band-pass filtered off-line using a digital, zero-phase-shift Butterworth filter with a passband from 0.01 to 150 Hz (6 and 12 dB/octave). A BESA model with one equivalent dipole located near Heschl's gyrus in each hemisphere was used for fitting the P1m of the pooled conditions for the m– masker alone, m– masker plus signal at the higher signal level (69 dB), and m– masker plus signal at the lower signal level (57 dB). The fit interval covered the peak of the P1m ranging from the minimum between the Pam and P1m to the same amplitude value of the descending deflection between P1m and N1m. This masker condition was used for the fit because it generates a larger amount of neural synchrony across frequency and thus provided a large signal-to-noise ratio. No further constraints concerning dipole location and orientation or symmetry conditions were applied. The two-dipole model was held constant and used as a spatial filter for each stimulus condition.
Grand average source waveforms were separately computed for each stimulus condition and hemisphere. Difference waveforms were computed to derive the specific response elicited by the 69- and 57-dB signals such that the source waveform of the masker-alone condition without the signal was subtracted from the source waveform of the respective masker-plus-signal condition. To assess the significance of the resulting AEF complex, a permutation test for waveform differences, introduced by Blair and Karniski (1993), was applied. This distribution-free method produces P values without any assumption about a particular correlation structure among the waveforms. It can be computed for any number of variables and thus represents a robust alternative to the Hotelling T2 test when the number of time points exceeds the number of subjects tested (Picton et al. 2000). The statistic used herein was based on the whole waveform including the P1m maximum of the AEF complex. The output of this procedure is a single multivariate statistic denoted as tsum.
Magnetic resonance imaging (MRI) scans, available for 13 of 14 subjects, were obtained using a Siemens Symphony 1.5-T scanner. Three-dimensional reconstructions of the 176 (1-mm voxel) slices were computed using BrainVoyager software (version 4.4, Brain Innovation, Maastricht, The Netherlands). Dipole positions were coregistered onto the individual MRI and then transformed into the standard space of Talairach and Tournoux (1988) to illustrate the location of the generators.
Experiment 2: period patterns of excitation produced by tone pulses in complex maskers
STIMULI.
Kohlrausch and Sander (1995) measured psychoacoustic masking period patterns (the masked threshold of a brief tone pulse, the signal, in the presence of the m+ or m– masker) as a function of the temporal position of the signal within one period of the masker. In addition to the m+ and m– complexes, a sine-phase masker with zero starting phase for all masker components (200–2,000 Hz, f0 = 100 Hz) was utilized, referred to as the "m0" complex. The resulting stimulus resembled a pulse sequence with an interpulse interval of 10 ms, given by the inverse of the fundamental frequency. Portions of the waveforms of the stimuli used in this experiment are illustrated in Fig. 2. The masker level was 70 dB SPL as in experiment 1. The signal was a 5-ms 1,100-Hz tone pulse with 2.5-ms Hanning onset and offset ramps. In contrast to Kohlrausch and Sander, who varied the signal level during their experiment to obtain the corresponding masked threshold, we applied the signal at a fixed level to derive the signal's specific neuromagnetic response that was affected differently according to the temporal position of the signal in the masker. The level of the signal was fixed at 14 dB relative to the level of the individual masker components (i.e., at 71 dB SPL) such that it was detectable in all signal-masker conditions but close to the masked signal threshold for several positions of the signal in the m0 and the m– maskers (see Fig. 5 in Kohlrausch and Sander 1995). The signal was presented once every 200 ms with a delay of 0, 2, 4, 6, or 8 ms relative to the beginning of the respective masker period. The overall duration of the stimulus was 10 s and the ISI was 1.2 s.
|
|
|
|
). The partially masked loudness of the signal generally increases as the signal level rises above its masked threshold. Thus we expected that the partial loudness of the signal at the given suprathreshold signal level should be (inversely) related to the masked detection threshold (from Kohlrausch and Sander 1995); i.e., the lower the masked detection threshold for a particular signal-masker condition, the larger should be the corresponding partial loudness sensation. This was investigated for the three masker types (m+, m–, m0) as a function of the temporal position of the signal in the respective masker.
A two-alternative forced-choice task for paired comparisons of all masker-plus-signal combinations was applied. Subjects were instructed to indicate which signal was perceived as being louder. The Schroeder-phase maskers for the psychoacoustic task were 750 ms in duration and the signal occurred three times: 200, 400, and 600 ms after masker onset. Every combination of paired comparisons, involving three maskers (m+, m–, m0) and five temporal signal positions (with delays of 0, 2, 4, 6, and 8 ms relative to the onset of the masker period), was presented twice in a different order such that subjects judged 210 pairs overall. A scale for the relative loudness produced by the signal in the masker was derived from the paired comparisons for each subject, using the Bradley–Terry–Luce (BTL) method (David 1988), which assumes that perceived loudness can be ordered on a single scale.
MEG RECORDINGS. The effects of phase-sensitive cochlear processing in neural representations at the cortical level were studied using auditory evoked fields. These were recorded in response to the same stimuli as those in the psychoacoustic experiments. The patterns of excitation, produced by the signal added to the different maskers (m+, m–, m0) at different positions within a masker period, are referred to as neuromagnetic period patterns. Recording settings similar to those reported in experiment 1 were applied with the one exception that about 1,100 single sweeps, from –50 to +250 ms relative to the stimulus onset, were averaged for each stimulus condition. MEG registrations, obtained separately for each masker condition, were taken for about 30 min in each session.
DATA ANALYSIS. The fitting procedure for the source analysis was nearly the same as that in the first experiment, with the exception that the specific response evoked by a signal was fitted with a slightly different procedure. A BESA model with one equivalent dipole in each hemisphere was computed for each masker condition and fitted to the experimental condition that exhibited the largest response evoked by a signal (i.e., the 0-ms-delay condition for the m+ and m– maskers, and the 4-ms-delay condition for the m0 masker). This procedure resulted in three slightly different fits for each subject and masker condition, as shown in Fig. 3. A common reliable source model for each subject was derived by averaging the coordinates and orientations of the three independent source models. These mean parameters were used in the final model for all 15 experimental conditions according to masker type (m+, m–, m0) and signal position (0, 2, 4, 6, and 8 ms). The above-cited procedure, applied for each subject, assumed that the P1m response is evoked by the same generators within Heschl's gyrus; i.e., the location and orientation of the equivalent dipole remained constant for the signal presented at the different temporal positions.
An ANOVA (general linear model, SAS software, version 9) with repeated measurements including the factors masker and signal position was used to investigate the main and interaction effects of the P1m maxima elicited by the different masking types and relative delays of the signal.
Simulation of auditory-nerve activity patterns
A nonlinear model of cochlear processing was used to consider the effects of BM filtering and transformation into auditory-nerve (AN) activity for the stimuli of the present study. The auditory image model (AIM) of Patterson et al. (1995) was used to simulate the fine-grain spectrotemporal information in the auditory nerve. The first stage of this model consists of a one-dimensional transmission-line filterbank that accounts for cochlear hydrodynamics (Giguère and Woodland 1994). The output of this stage simulates basilar-membrane motion. The simulations in the present study were computed with a transmission line consisting of 500 sections covering a frequency range from 100 to 6,000 Hz. The tuning of the local segments of the basilar membrane was set to Qn = 8, as suggested by Carlyon and Datta (1997). They used the same modeling framework to account for the effects of the signal level, the number of components, and the phase of flanking components in psychoacoustic masking period patterns using the same type of maskers. The output of the transmission line was then converted into the neural activity pattern (NAP) using a hair-cell simulator with one hair cell per filterbank channel (Meddis 1988). A medium and a high spontaneous-rate fiber type were applied for the hair cell of each simulated channel. As in Berlin (1984), the excitation patterns of the medium and high spontaneous-rate fibers were weighted respectively by 0.35 for the medium spontaneous rate and 0.65 for the high spontaneous rate before adding the data for both fiber types.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
MEG DATA.
It was possible to consistently fit a stable spatiotemporal model with one equivalent dipole in each auditory cortex for all 14 subjects. The projection of the equivalent dipoles onto an averaged map of the sulcal borders of Heschl's gyrus (Schneider et al. 2005) showed dipole sources located bilaterally close to the border of the medial and lateral Heschl's gyrus (Fig. 3). Figure 4 displays individually the grand average source waveforms obtained with the m+ (Fig. 4A) and the m– conditions (Fig. 4B).
The results for the m+ condition (Fig. 4A) are as follows: The masker-alone condition (black curve, left) evoked a Pam followed by a prominent P1m and then a small N1m at about 130 ms. The addition of the more intense signal (69 dB SPL), indicated by the dark gray curve in the same panel, generated a clear additional response with a peak at about 60 ms relative to the signal onset. The signal-related response, derived by subtracting the waveform generated by the m+-masker alone from that produced by the m+-masker plus signal, is shown in the corresponding right panel. The waveform exhibited a significant positive plateau starting at about 60 ms after signal onset and lasting until about 150 ms (tsum = 227.2, P < 0.01). For the weaker signal (57 dB; indicated by the light gray curve), a smaller response was observed with a delayed peak where the largest positive response occurred near 100 ms. Permutation tests showed that this slow positive AEF represented a significant deflection (tsum = 455.2, P < 0.001).
For the m– condition (Fig. 4B), the source waveforms showed a steeper slope of the initial Pam activation than that for the m+ condition in response to the masker alone. The whole waveform exhibited a strong ripple with a periodicity of 10 ms corresponding to that of the masker waveform. The corresponding signal-related responses, i.e., the differences between the AEFs obtained with the m– masker plus signal and those with the masker alone, are shown in Fig. 4B (right). The more intense signal (69 dB) evoked a significant positivity (tsum = 227.9, P < 0.01). For the weaker signal, the specific response resulted in a difference source waveform with a much smaller magnitude than that obtained in the m+ condition, albeit with a significant deflection (tsum = 310.1, P < 0.01).
The direct comparison of the signal-related AEF (Fig. 4A, right vs. Fig. 4B, right) between the masker conditions did not reveal a significant difference for the more intense signal (tsum = 85.8, ns). However, a significant difference was found for the weaker signal such that the signal-related AEF was larger in the m+ condition than that in the m– condition (tsum = 241.4, P < 0.01).
Finally, we compared the responses obtained with the two maskers alone (Fig. 4C). The difference of the response waveforms is shown in the corresponding right panel. The permutation test showed a highly significantly larger response for the m– masker than that for the m+ masker in the interval from 30 to 150 ms that covers the whole positive deflection (tsum = 308.0, P < 0.0001).
SIMULATED AN ACTIVITY PATTERNS.
Figure 5A shows the internal representation of the four masker-plus-signal combinations after nonlinear basilar-membrane processing (Giguère and Woodland 1994) through a filter tuned to a signal frequency of 1,000 Hz. The gray functions show the filtered masker-alone responses and the black functions indicate the filtered representations of the sum of masker and signal. The filtered m+ and m– maskers have very different shapes, especially in terms of their "peakiness," as a consequence of the dispersive characteristics of the BM transformation. This can be seen most clearly in the first periods of the filtered maskers (in the absence of the signal). The m+ masker is modulated more strongly in amplitude than the m– masker. This is also reflected in the neural activity patterns (NAPs) shown in Fig. 5B. In the framework of the cochlea model, the addition of the signal produces a stronger change of neural activity in the case of the m+ masker (where the signal effectively "fills" the valleys of the filtered masker representation) than in the case of the m– masker.
Figure 5C shows the neural summary pattern at the AN level, integrated across the nominal frequency range between 200 and 2,000 Hz. The AN summary activity evoked by the m– masker alone (i.e., for the first and last 20 ms) shows a rippled structure. This result can be explained by the superposition of the neural activity across frequencies. The upward sweep (m– masker) partially compensates the traveling-wave delay across frequency, observed when the excitation by a (transient) stimulus progresses apically along the BM. This results in a relatively strong response amplitude for each period of the stimulus. In contrast, the summary NAP obtained with the m+ masker shows a less pronounced 10-ms periodicity. Instead, the m+ pattern exhibits much stronger fluctuations within a stimulation period due to the more peaky responses of the individual channels, as exemplified by the channel at the resonance frequency of 1,000 Hz in Fig. 5A. Since the peaks stemming from different peripheral channels occur at different times, their superposition results in a rippled fine structure of the summed waveform. At the high signal level, the signal-related change of excitation in the temporal portion where the signal is added to the masker is substantial in both masker conditions (m+ and m–). At the lower signal level, the responses are correspondingly weaker. However, the signal-related response appears to be stronger for the m+ masker than that for the m– masker. In the latter case, the signal contributes only within very short portions of activity at the beginning of each masker period.
Some of the characteristics in the summed NAPs after cochlear processing (Fig. 5C) are consistent with our measured neuromagnetic responses (from Fig. 4). The simulated periodic 10-ms modulation of the activity produced by the m– masker alone corresponds to the periodicity in the neuromagnetic response from Fig. 4B. Importantly, the signal-related responses in the NAP simulations are consistent with the corresponding signal-related neuromagnetic source waveforms, where the excitation due to the (lower-level) signal was larger in the presence of the m+ masker than that in the presence of the m– masker (Fig. 4, A and B, right panels).
Experiment 2: correlation of neuromagnetic period patterns, partial loudness patterns, and simulated auditory-nerve activity
MEG DATA. The grand average AEF source waveforms are shown in Fig. 6. The responses evoked by the maskers alone (thin black curves) indicate the baseline conditions for the m+ (Fig. 6, top), m– (Fig. 6, middle), and m0 masker (Fig. 6, bottom). As discussed in the first experiment, the responses to the m– masker show a pronounced ripple structure. The remaining curves in each panel display the responses to the different masker-plus-signal combinations. The addition of the transient 5-ms signal to the masker produced a large positive response that includes a Pam at about 30 to 40 ms in almost all conditions. This is followed by a large P1m peaking at about 80 ms, found for all three maskers and all temporal positions of the signal in the masker. The P1m peak value of the response waveform varies as a function of the temporal position of the signal in the masker.
|
4 ms. Here, the interaction masker x phase was significant for the m+ and m0 as well as the m– and m0 conditions.
|
|
), is shown in Fig. 7B. The general shape is similar to the neuromagnetic period pattern (Fig. 7A). The loudness produced by the signal presented in the m0 masker showed an almost inverted pattern relative to that obtained with the Schroeder-phase maskers. The highest loudness values for the m0 masker were found for signal delays at 4 to 8 ms, whereas a minimum was observed at 0 ms. SIMULATED AN ACTIVITY PATTERNS. The specific excitation produced by the signal was calculated by subtracting the spike rate evoked by the masker alone from that of the masker plus signal. This signal-related excitation is depicted in Fig. 8. For each masker condition, the additional activation was determined for the five different positions of the signal in the masker. The thin lines represent the activation produced by the maskers alone, summed across channels from 550 to 2,200 Hz. The shaded areas above the lines indicate additional activity from the increase in spike rate evoked by the signal under the various conditions. Integration of the shaded areas (i.e., the overall change in neural activity due to the addition of the signal) leads to the corresponding values shown in Fig. 7C. The resulting patterns for the different maskers and signal positions are similar to those obtained from the neuromagnetic recordings (Fig. 7A) and the psychoacoustic partial loudness data (Fig. 7B). The locations of the maxima and the minima of the patterns as well as the shift between them for the m+ and m– maskers are also similar.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
The projection of the equivalent dipoles that were in both experiments, based on a fit of the P1m onto a map of average sulcal borders (Schneider et al. 2005), revealed a position in the lateral portion of Heschl's gyrus close to the border of the core area. This is in line with the specific location of the P1m as reported by Schneider et al. (2005), which was based on the same transformation procedure and map. Furthermore, it is consistent with the intracranial recordings of Liégeois-Chauvel et al. (1994) who found the generators of the components with a latency of 60–75 ms to be localized in the lateral part of Heschl's gyrus.
The signal-related (difference) waveforms showed a prominent P1m from auditory cortex that started about 70 ms after signal onset in experiment 1 and after about 50 ms in experiment 2. This P1m was not followed by N1m deflection because a relatively short stimulus onset asynchrony between the signals was chosen (370 ms in experiment 1; 200 ms in experiment 2). Previous studies have shown that the N1m is strongly reduced for ISIs <0.5 s (Hari et al. 1982; Onitsuka et al. 2000) and not consistently observed for ISIs <300 ms (Carver et al. 2002), whereas the P1m is only slightly attenuated at an ISI of 200 ms and can still be observed for ISIs of 
100 ms (Gutschalk et al. 2004).
In both experiments of this study, the signal-related neuromagnetic responses were consistent with the perceptual data and with changes of excitation in the neural activity pattern at the output of cochlear processing. The difference waveforms of the responses in experiment 1 (Fig. 4, A and B) showed a plateau-like positive deflection for the more intense signal (69 dB SPL) for both masker types. At this level, the two signal-related responses did not differ in magnitude from each other. For the lower signal level (57 dB SPL), the responses were weaker and increased more gradually with time (and also decreased more gradually after signal offset). Here, the signal-related response was larger for the m+ than that for the m– condition. This observation was consistent with the simulations of the AN activity patterns (Fig. 5). These results are also consistent, at least qualitatively, with the perceptual masking results from Kohlrausch and Sander (1995) that demonstrated that the signal is masked more effectively by the m– complex than by the m+ masker.
A more detailed and direct investigation of the relation between AEF and perception was considered in our experiment 2. Regarding the correlation with the perceptual data, it is interesting to relate the loudness results from this experiment to the masked threshold data provided in Kohlrausch and Sander (1995). Figure 10 shows a replot of their masked detection data (their Fig. 5), using the same axes as in Fig. 7 of the present study. The masked detection threshold in Fig. 10 is shown as a function of the temporal position of the signal within (one period of) the masker. For the m0 masker, the peak in the masked period pattern occurs at the signal position close to 0 ms. The reason for this is that the envelope of this masker has a peak close to 0 ms (see Fig. 2) and thus makes it difficult to detect the signal. For the m+ and m– maskers, the maxima of the period patterns are at a signal position at which the instantaneous frequency of the masker coincides with the signal frequency. Since the signal frequency is in the spectral center of the complex, this time point occurs approximately in the temporal center of the period. Thus compared with the maxima in the period pattern for the m0 complex, the maxima for the Schroeder-phase complexes are temporally shifted by half a period (i.e., 5 ms), as can be seen in Fig. 10. The function for the m– masker, however, is shifted toward higher threshold values and is somewhat flatter than the pattern for the m+ masker. The differences between these two patterns have been attributed to the dispersive properties of the peripheral filter tuned to the signal frequency (Kohlrausch and Sander 1995; Smith et al. 1986).
|
).
A relationship between the P1m and perception has earlier been observed in a study of auditory stream segregation (Gutschalk et al. 2005), where the P1m was related to the strength of perceived segregation of one stream of tones from another one. Comparing stimuli of very different complexity, Lütkenhöner et al. (2006) provided evidence that activation strength in the auditory periphery correlates with the P1m magnitude. Moreover, Chait et al. (2004) demonstrated that the P1m was relatively independent of task demands and perturbation. Thus the P1m appeared to be well suited to study the relationship between peripheral filtering, auditory cortex function, and perception, as confirmed here by the high correlation found between summed AN activity and the P1m magnitude (Fig. 9C).
Another interesting observation was made in the present study with respect to the masker-alone responses. The masker-related response amplitude was larger for the m– masker than that for the m+ masker (Fig. 4C). This is consistent with the findings of Recio (2001) who recorded larger firing rates in AN fibers and neurons of the ventral cochlear nucleus (VCN) of chinchillas for their m– masker than for their m+ masker. Our results are also consistent with those of Rupp et al. (2002) who showed that middle-latency responses reflect the amount of synchronization along the basilar membrane. Their responses, obtained with upward sweeping chirps of a sweeping rate (Dau et al. 2000) different from that used here, also showed larger amplitudes than those obtained with a temporally reversed (downward sweeping) chirp. In Dau et al. (2000), Rupp et al. (2002), and related studies (e.g., Fobel and Dau 2004; Junius and Dau 2005; Stürzebecher et al. 2006), the approach was different from that in the present study. In their studies, the goal was to maximize neural synchronization across frequency that was achieved using a broadband stimulus that attempts to compensate for cochlear travel-time differences. Such a stimulus must be rising in instantaneous frequency with time since low-frequency components require more time to reach their place of maximum displacement (near the apex of the cochlea) than high-frequency components (close to the base). The upward-sweeping m– masker of the present study, even though much narrower in bandwidth and reflecting a linear sweeping rate [in contrast to the nonlinear sweeping rate in Dau et al. (2000)], produces a larger overall response amplitude than the m+ masker since it generates a larger amount of synchronization in the excited frequency region. This, in turn, is consistent with recent results by Mauermann and Hohmann (2007) on loudness perception where the listeners required a 6-dB higher stimulation level for m+ complexes to match the loudness of the m– complexes, a result that shows that phase-sensitive cochlear processing can affect loudness perception and should be incorporated in loudness models.
Physically, phase dispersion in the transfer function of individual cochlear filters in the form of an upward frequency "glide" (e.g., de Boer and Nuttall 1997) in the impulse responses is equivalent to (or a consequence of) the occurrence of the traveling wave along the basilar membrane. It is not possible to create a stimulus that compensates both for travel-time differences across frequency and for the phase curvature in the individual auditory filters. Thus although the m+ masker led to larger signal-related responses, which were in the focus of the present study, the m– stimulus produced the larger masker-alone response.
The results presented in this study suggest that neuromagnetic responses, and in particular the middle-latency AEFs, might be beneficially used as an objective correlate of the behavioral performance of human subjects in auditory masking conditions. They might also be helpful for elucidating models of cochlear processing in humans through comparison of neuromagnetic recordings with model predictions. The neuromagnetic responses considered here might also be interesting as an objective tool in hearing research involving patients that cannot participate actively in listening experiments such as newborns and small children. Recent results in psychoacoustic masking studies with hearing-impaired listeners, using Schroeder-type tone-complex maskers, showed much flatter masked period patterns compared with those of normal-hearing listeners (Oxenham and Dau 2004; Summers 2000; Summers and Leek 1998). The variation in these patterns has been suggested to reflect the state of hearing in human listeners.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: A. Rupp, Section of Biomagnetism, Department of Neurology, University of Heidelberg, Im Neuenheimer Feld 400, D-69120 Heidelberg, Germany (E-mail: andre.rupp{at}uni-heidelberg.de)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Blair RC, Karniski W. An alternative method for significance testing of waveform difference potentials. Psychophysiology 30: 518–524, 1993.[Web of Science][Medline]
Carlyon RP, Datta AJ. Masking period pattern of Schroeder-phase complexes: effects of level, number of components, and phase of flanking components. J Acoust Soc Am 101: 3648–3657, 1997.[CrossRef][Web of Science][Medline]
Carver FW, Fuchs A, Jantzen KJ, Kelso JAS. Spatiotemporal analysis of the neuromagnetic response to rhythmic auditory stimulation: rate dependence and transient to steady-state transition. Clin Neurophysiol 113: 1921–1931, 2002.[CrossRef][Web of Science][Medline]
Chait M, Simon JZ, Poeppel D. Auditory M50 and M100 responses to broadband noise: functional implications. Neuroreport 15: 2455–2458, 2004.[CrossRef][Web of Science][Medline]
Dau T, Wegner O, Mellert V, Kollmeier B. Auditory brainstem responses with optimized chirp signals compensating basilar-membrane dispersion. J Acoust Soc Am 103: 1530–1540, 2000.
David HA. The Method of Paired Comparisons. New York: Oxford Univ. Press, 1988.
de Boer E, Nuttall AL. The mechanical waveform of the basilar membrane. I. Frequency modulations ("glides") in impulse responses and cross-correlation functions. J Acoust Soc Am 101: 3583–3592, 1997.[CrossRef][Web of Science][Medline]
Fletcher H. Auditory patterns. Rev Mod Phys 12: 47–65, 1940.[CrossRef]
Fobel O, Dau T. Searching for the optimal stimulus eliciting auditory brainstem responses in humans. J Acoust Soc Am 116: 2213–2222, 2004.[CrossRef][Web of Science][Medline]
Giguère C, Woodland PC. A computational model of the auditory periphery for speech and hearing research. I. Ascending path. J Acoust Soc Am 95: 331–342, 1994.[CrossRef][Web of Science][Medline]
Glasberg BR, Moore BCJ. Derivation of auditory filter shapes from notched-noise data. Hear Res 47: 103–138, 1990.[CrossRef][Web of Science][Medline]
Gockel H, Moore BC, Patterson RD. Asymmetry of masking between complex tones and noises: the role of temporal structure and peripheral compression. J Acoust Soc Am 111: 2759–2770, 2002.[CrossRef][Web of Science][Medline]
Gockel H, Moore BC, Patterson RD. Asymmetry of masking between complex tones and noises: partial loudness. J Acoust Soc Am 114: 349–360, 2003.[CrossRef][Web of Science][Medline]
Goldstein JL. Auditory spectral filtering and monaural phase perception. J Acoust Soc Am 41: 458–479, 1967.[CrossRef][Web of Science][Medline]
Gutschalk A, Micheyl C, Melcher JR, Rupp A, Scherg M, Oxenham AJ. Neuromagnetic correlates of streaming in human auditory cortex. J Neurosci 25: 5382–5388, 2005.
Gutschalk A, Patterson RD, Uppenkamp S, Scherg M, Rupp A. Recovery and refractoriness of auditory evoked fields after gaps in click trains. Eur J Neurosci 20: 3141–3147, 2004.[CrossRef][Web of Science][Medline]
Hari R, Kaila K, Katila T, Tuomisto T, Varpula T. Interstimulus interval dependence of the auditory vertex response and its magnetic counterpart: implications for their neural generation. Electroencephalogr Clin Neurophysiol 54: 561–569, 1982.[CrossRef][Web of Science][Medline]
Helmholtz HLF. On the Sensations of Tone. New York: Dover, 1954.
Junius D, Dau T. Influence of cochlear traveling wave and neural adaptation on auditory brainstem responses. Hear Res 205: 53–67, 2005.[CrossRef][Web of Science][Medline]
Kohlrausch A, Sander A. Phase effects in masking related to dispersion in the inner ear. II. Masking period patterns of short targets. J Acoust Soc Am 97: 1817–1829, 1995.[CrossRef][Web of Science][Medline]
Liégeois-Chauvel C, Musolino A, Badier JM, Marquis P, Chauvel P. Evoked potentials recorded from the auditory cortex in man: evaluation and topography of the middle latency components. Electroencephalogr Clin Neurophysiol 92: 204–214, 1994.[CrossRef][Web of Science][Medline]
Lütkenhöner B, Seither-Preisler A, Seither S. Piano tones evoke stronger magnetic fields than pure tones or noise, both in musicians and non-musicians. Neuroimage 30: 927–937, 2006.[CrossRef][Web of Science][Medline]
Mathes RC, Miller RL. Phase effects in monaural perception. J Acoust Soc Am 19: 780–797, 1947.[CrossRef][Web of Science]
Mauermann M, Hohmann V. Differences in loudness of tone complexes with positive and negative Schroeder phase (Abstract). J Acoust Soc Am 117: 2597, 2007.
Meddis R. Simulation of auditory-neural transduction: further studies. J Acoust Soc Am 83: 1056–1063, 1988.[CrossRef][Web of Science][Medline]
Moore BCJ. An Introduction to the Psychology of Hearing. London: Elsevier Academic Press, 2004.
Onitsuka T, Ninomiya H, Sato E, Yamamoto T, Tashiro N. The effect of stimulus intervals and between-block rests on the auditory evoked potential and magnetic fields: is the auditory P50 in humans an overlapping potential? Clin Neurophysiol 111: 237–245, 2000.[CrossRef][Web of Science][Medline]
Oxenham AJ, Dau T. Towards a measure of auditory-filter phase response. J Acoust Soc Am 110: 3169–3178, 2001.[CrossRef][Web of Science][Medline]
Oxenham AJ, Dau T. Masker phase effects in normal-hearing and hearing-impaired listeners: evidence for peripheral compression at low signal frequencies. J Acoust Soc Am 116: 2248–2257, 2004.[CrossRef][Web of Science][Medline]
Patterson RD. Auditory filter shapes derived with noise stimuli. J Acoust Soc Am 59: 640–654, 1976.[CrossRef][Web of Science][Medline]
Patterson RD. A pulse ribbon model of monaural phase perception. J Acoust Soc Am 82: 1560–1586, 1987.[CrossRef][Web of Science][Medline]
Patterson RD, Allerhand MH, Giguère C. Time-domain modelling of peripheral auditory processing: a modular architecture and a software platform. J Acoust Soc Am 98: 1890–1894, 1995.[CrossRef][Web of Science][Medline]
Picton TW, Bentin S, Berg P, Donchin E, Hillyard SA, Johnson R, Miller GA, Ritter W, Ruchkin DS, Rugg MD, Taylor MJ. Guidelines for using human event-related potentials to study cognition: recording standards and publication criteria. Psychophysiology 37: 127–152, 2000.[CrossRef][Web of Science][Medline]
Plomp R. The ear as a frequency analyzer. J Acoust Soc Am 36: 1628–1636, 1964.[CrossRef][Web of Science]
Recio A. Representation of harmonic complex stimuli in the ventral cochlear nucleus of the chinchilla. J Acoust Soc Am 110: 2024–2033, 2001.[CrossRef][Web of Science][Medline]
Rhode WS, Recio A. Study of mechanical motions in the basal region of the chinchilla cochlea. J Acoust Soc Am 107: 3317–3332, 2000.[CrossRef][Web of Science][Medline]
Ruggero MA, Rich NC, Recio A, Narayan SS, Robles L. Basilar-membrane responses to tones at the base of the chinchilla cochlea. J Acoust Soc Am 101: 2151–2163, 1997.[CrossRef][Web of Science][Medline]
Rupp A, Uppenkamp S, Gutschalk A, Beucker R, Patterson RD, Dau T, Scherg M. The representation of peripheral neural activity in the middle-latency evoked field of primary auditory cortex in humans. Hear Res 2174: 19–31, 2002.
Scherg M. Fundamentals of dipole source analysis. In: Auditory Evoked Magnetic Fields and Electric Potentials: Advances in Audiology, edited by Grandori F, Hoke M, Romani GL. Basel: Karger, 1990, vol. 6, p. 40–69.
Scherg M, Hari R, Hämäläinen M. Frequency-specific sources of the auditory N19-P30-P50 response detected by a multiple source analysis of evoked magnetic fields and potentials. In: Advances in Biomagnetics, edited by Williamson SJ, Hoke M, Sroink G. New York: Plenum Press, 1989, p. 97–100.
Schneider P, Sluming V, Roberts N, Scherg M, Goebel R, Specht HJ, Dosch HG, Bleeck S, Stippich C, Rupp A. Structural and functional asymmetry of lateral Heschl's gyrus reflects pitch perception preference. Nat Neurosci 8: 1241–1247, 2005.[CrossRef][Web of Science][Medline]
Schroeder MR. Synthesis of low peak-factor signals and binary sequences with low autocorrelations. IEEE Trans Inf Theory 16: 85–89, 1970.[CrossRef]
Smith BK, Sieben UK, Kohlrausch A, Schroeder MR. Phase effects in masking related to dispersion in the inner ear. J Acoust Soc Am 80: 1631–1637, 1986.[CrossRef][Web of Science][Medline]
Strube HW. A computationally efficient basilar-membrane model. Acustica 58: 207–214, 1985.[Web of Science]
Stürzebecher E, Cebulla M, Elberling C, Berger T. New efficient stimuli for evoking frequency-specific auditory steady-state responses. J Am Acad Audiol 17: 448–461, 2006.[Web of Science][Medline]
Summers V. Effects of hearing impairment and presentation level on masking period patterns for Schroeder-phase harmonic complexes. J Acoust Soc Am 108: 2307–2317, 2000.[CrossRef][Web of Science][Medline]
Summers V, Leek MR. Masking of tones and speech by Schroeder-phase harmonic complexes in normally hearing and hearing impaired listeners. Hear Res 118: 139–150, 1998.[CrossRef][Web of Science][Medline]
Talairach P, Tournoux J. A Stereotactic Coplanar Atlas of the Human Brain. Stuttgart, Germany: Thieme, 1988.
Zwicker E. Die Grenzen der Hörbarkeit der Amplitudenmodulation und der Frequenzmodulation eines Tones. Acustica 2: 125–133, 1952.
Zwicker E, Fastl H. Psychoacoustics. Berlin: Springer-Verlag, 1999.
Zwicker E, Flottorp G, Stevens SS. Critical bandwidth in loudness summation. J Acoust Soc Am 29: 548–557, 1957.[CrossRef][Web of Science]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
