ArticlePDF Available

Abstract and Figures

This magnetoencephalography (MEG) study investigated evoked ON and OFF responses to ramped and damped sounds in normal-hearing human adults. Two pairs of stimuli that differed in spectral complexity were used in a passive listening task; each pair contained identical acoustical properties except for the intensity envelope. Behavioral duration judgment was conducted in separate sessions, which replicated the perceptual bias in favour of the ramped sounds and the effect of spectral complexity on perceived duration asymmetry. MEG results showed similar cortical sites for the ON and OFF responses. There was a dominant ON response with stronger phase-locking factor (PLF) in the alpha (8–14 Hz) and theta (4–8 Hz) bands for the damped sounds. In contrast, the OFF response for sounds with rising intensity was associated with stronger PLF in the gamma band (30–70 Hz). Exploratory correlation analysis showed that the OFF response in the left auditory cortex was a good predictor of the perceived temporal asymmetry for the spectrally simpler pair. The results indicate distinct asymmetry in ON and OFF responses and neural oscillation patterns associated with the dynamic intensity changes, which provides important preliminary data for future studies to examine how the auditory system develops such an asymmetry as a function of age and learning experience and whether the absence of asymmetry or abnormal ON and OFF responses can be taken as a biomarker for certain neurological conditions associated with auditory processing deficits.
Content may be subject to copyright.
brain
sciences
Article
Perceptual Temporal Asymmetry Associated with
Distinct ON and OFF Responses to Time-Varying
Sounds with Rising versus Falling Intensity:
A Magnetoencephalography Study
Yang Zhang 1,2,3,*, Bing Cheng 4,*, Tess K. Koerner 1, Robert S. Schlauch 1, Keita Tanaka 5,
Masaki Kawakatsu 6, Iku Nemoto 6and Toshiaki Imada 7
1
Department of Speech-Language-Hearing Sciences, University of Minnesota, Minneapolis, MN 55455, USA;
koern030@umn.edu (T.K.K.); schla001@umn.edu (R.S.S.)
2Center for Neurobehavioral Development, University of Minnesota, Minneapolis, MN 55455, USA
3Speech-Language-Hearing Center, School of Foreign Languages, Shanghai Jiao Tong University,
Shanghai 200240, China
4English Department & Institute for Language, Cognition and Brain Sciences, School of Foreign Studies,
Xi’an Jiaotong University, Xi’an, Shaanxi 710049, China
5School of Science and Engineering, Tokyo Denki University, Ishizaka, Hatoyama, Hiki-gun,
Saitama 350-0394, Japan; ktanaka@mail.dendai.ac.jp
6School of Information Environment, Tokyo Denki University, 2-1200, Muzai-gakuendai, Inzai-shi,
Chiba 270-1382, Japan; kawakatu@asrl.dendai.ac.jp (M.K.); nemoto@sie.dendai.ac.jp (I.N.)
7Institute for Learning and Brain Sciences, University of Washington, Seattle, Washington 98195, USA;
imada@u.washington.edu
*Correspondences: zhanglab@umn.edu (Y.Z.); bch@mail.xjtu.edu.cn (B.C.);
Tel.: +1-612-624-7818 (Y.Z.); +86-29-8266-5237 (B.C.)
Academic Editor: Heather Bortfeld
Received: 5 May 2016; Accepted: 29 July 2016; Published: 5 August 2016
Abstract:
This magnetoencephalography (MEG) study investigated evoked ON and OFF responses
to ramped and damped sounds in normal-hearing human adults. Two pairs of stimuli that differed
in spectral complexity were used in a passive listening task; each pair contained identical acoustical
properties except for the intensity envelope. Behavioral duration judgment was conducted in separate
sessions, which replicated the perceptual bias in favour of the ramped sounds and the effect of spectral
complexity on perceived duration asymmetry. MEG results showed similar cortical sites for the ON
and OFF responses. There was a dominant ON response with stronger phase-locking factor (PLF) in
the alpha (8–14 Hz) and theta (4–8 Hz) bands for the damped sounds. In contrast, the OFF response
for sounds with rising intensity was associated with stronger PLF in the gamma band (30–70 Hz).
Exploratory correlation analysis showed that the OFF response in the left auditory cortex was a
good predictor of the perceived temporal asymmetry for the spectrally simpler pair. The results
indicate distinct asymmetry in ON and OFF responses and neural oscillation patterns associated
with the dynamic intensity changes, which provides important preliminary data for future studies to
examine how the auditory system develops such an asymmetry as a function of age and learning
experience and whether the absence of asymmetry or abnormal ON and OFF responses can be taken
as a biomarker for certain neurological conditions associated with auditory processing deficits.
Keywords:
MEG; auditory ON response; auditory OFF response; equivalent current dipole (ECD);
minimum norm estimation (MNE); phase locking factor (PLF); temporal asymmetry index (TAI)
Brain Sci. 2016,6, 27; doi:10.3390/brainsci6030027 www.mdpi.com/journal/brainsci
Brain Sci. 2016,6, 27 2 of 25
1. Introduction
One fundamental property of the human auditory system is to automatically detect and respond to
sound sources and changes. Rising and falling intensities are two basic dynamic patterns in the acoustic
environment. A damped sound is characterized by the abrupt occurrence of high-intensity onset and a
gradual fade-out whereas a ramped sound has the opposite pattern. Behavioral research has shown a
perceptual bias for rising-intensity (or ramped) sounds. Compared with the time-reversed stimuli,
listeners report hearing different timbre quality with stronger tonality for the ramped tone [
1
,
2
] and
overestimate the ramped sounds in subjective duration [
3
7
], overall loudness [
5
,
8
,
9
], and the amount
of perceived loudness change within the stimuli [
10
12
]. For instance, Schlauch et al. (2001) found
that ramped sounds ranging from 10 to 200 ms were perceived to be longer than the time-reversed
damped sounds, which had identical physical duration as well as long-term spectral power and
envelope spectra. This overestimation was also found for sounds longer than 200 ms [
6
]. In the
literature, the reported bias for ramped sounds is referred to as “auditory perceptual asymmetry” or
“perceptual looming”, which is affected by the spectral complexity of the carrier sound for the ramped
and damped intensity modulation. When the carrier is spectrally more complex than a sinusoidal tone,
the asymmetry decreases or may even disappear [1,3,10,13,14].
Presumably, the perceptual asymmetry phenomenon arises from distinct neural coding of the
two dynamic intensity patterns, which may (or may not) require attentional processing and learning
experience. One ecologically motivated explanation points to the attentional system of adaptive
alertness, which is associated with auditory motion perception of approaching versus receding sound
sources [
14
16
]. Two psychophysical accounts have also been proposed [
3
,
7
,
8
]. One is that listeners
can perceptually separate the attack and gradual release of a damped sound, ignoring a segment
of the decay portion as “echo” and thereby perceiving a shorter and softer sound. Another is that
the abrupt offset of ramped sounds may produce stronger and persistent activity, which results in
overestimation of duration and loudness. Digiovanni and Schlauch (2007) studied both of these
possible mechanisms [
7
]. They measured subjective durations in two groups of listeners. One group
was instructed to match the durations of the ramped and damped sounds without any special
instructions (null instructions) whereas the other group was instructed to “include all aspects of
the sounds”. The two different instruction sets produced very different results. When no special
instructions were offered, the ramped sounds were judged to be 50%–90% longer than the damped
sounds. By contrast, the group told to include all aspects of the sounds judged the ramped sound
to be between 10% and 20% longer than damped sounds. The larger perceptual asymmetry for the
null instruction set is believed to be a result of subjects ignoring the echo of the damped sounds.
The remaining perceptual asymmetry (10%–20%), when subjects were told to include all aspects of
the sounds, is attributed to persistence of excitation. Temporal masking patterns were measured and
the temporal extent of the excitation (9%–24% longer for ramped than for damped sounds), provided
further support for the persistence explanation. That is, temporal footprint representing the range of
durations over which a ramped or damped sound used as a masker could interfere with detection
of a signal was longer for ramped sounds than for damped ones, and that difference corresponds to
the perceptual judgments of duration when subjects are asked to attend to all aspects of the sounds.
Ries et al. (2008) reported a nearly identical finding for temporal masking patterns [5].
Although the behavioural temporal asymmetry is well established, the neural mechanisms
are not completely understood. Animal neurophysiology work has demonstrated asymmetric
neural representations for ramped and damped sounds at multiple processing sites in the auditory
pathway [
17
19
]. There is evidence that persistence of excitation after the stimulation offset is longer
for the ramped sound than that of the damped sound [
20
]. Stronger gamma (45–90 Hz) activities
were also found for the ramped auditory signals in comparison with damped sounds [
21
]. It has been
suggested that the ON-neurons in the auditory cortex may play an important role in detecting sound
source movement [
22
] whereas the OFF-neurons enable more precise temporal coding of the intensity
envelope [
17
,
23
]. Unlike the animal studies, human research has primarily relied on experimental
Brain Sci. 2016,6, 27 3 of 25
paradigms that require attentive tasks. In earlier work [
13
,
24
], a peripheral mechanism of input
suppression was proposed. However, the peripheral model was found to be not able to address the
auditory bias for sounds longer than 50 ms. There is a known asymmetric coding that differences
in temporal onsets are encoded more readily and accurately than differences in offsets for sounds of
various lengths [
25
]. A plausible account would require detailed knowledge about neural coding of
the ramped and damped sounds in the whole auditory pathway [
6
,
24
], especially at the cortical level.
Auditory stimulation in event-related paradigms typically produces a robust ON response
(also known as the auditory N1) in adult listeners irrespective of the listener’s attention. A similar
off-N1 response can be observed at the cessation of an auditory stimulus longer than 100 ms. Both the
ON and OFF responses have been considered to represent similar automatic cortical responses to
abrupt changes due to their similar properties in latency, topography, and source localization [
26
29
].
While the evoked ON and OFF responses are well documented in electroencephalography (EEG) and
MEG studies [
25
28
,
30
35
], there has been a lack of human neurophysiological data that systematically
examined the asymmetric ON and OFF responses for ramped and damped sounds. Findings and
interpretations are mixed. Several studies showed that perceived duration was not a simple reflection
of the differences between the ON and OFF responses [
3
,
36
,
37
]. Intracellular recordings from the
primary auditory cortex in animals [
38
] as well as from the left auditory cortex in humans [
39
] indicated
that the OFF response neurons played a more important role in sound duration perception than the ON
response neurons. In contrast, several MEG studies reported that the ON response was a good predictor
for perceived differences such as hissiness or pitch salience of ramped and damped stimuli [
40
42
].
However, these MEG studies used a ramped/damped envelope shorter than 50 ms and thus did not
provide OFF response measures in relation to the perceived temporal asymmetry.
The primary goal of the present study was to compare evoked ON and OFF responses to ramped
and damped auditory stimuli in terms of response amplitude, latency and source localization and
explore possible brain-behavior correlations for the temporal asymmetry phenomenon. To replicate
the perceptual temporal asymmetry, we conducted behavioural tests of duration comparison and
estimation. We hypothesized that the ON and OFF responses would show differences in amplitude,
latency measures and source localization. In particular, damped sounds would produce a larger and
earlier ON response but a smaller OFF response relative to the ramped sounds, which could provide a
good opportunity to verify the two explanations regarding persistence of excitation for ramped sounds
and ignoring the echo of the damped sounds. Given that the OFF response is also sensitive to the
intensity level and rise/fall-time of the auditory stimuli [
43
,
44
], we predicted that the ON and OFF
response amplitude measures might show good correspondence with the behavioral data [39].
The secondary goal was to examine how spectral complexity of the auditory stimuli might
influence perceived temporal asymmetry and the neural measures of ON and OFF responses. We used
two pairs of stimuli by manipulating the intensity envelope and spectral complexity, one pair based
on a sinusoidal tone (S stimuli) and the other based on a complex piano note (C stimuli). Based on
previous studies [
1
,
3
,
10
,
13
,
14
], we hypothesized that adding spectral complexity to the ramped and
damped sounds might reduce the perceptual bias effect due to the additional perceptual attributes and
thus weaken the brain-behavior correlational strength.
The third goal was to examine how the distinct intensity envelopes of the auditory stimuli
drove the ON and OFF responses via neural phase locking across trials in the different cortical
oscillatory frequency bands. Cortical oscillation rhythms reflect properties of large-scale neuronal
population excitability and discharge synchronization/desynchronization that subserve various
perceptual, attentional, and integrative functions [
45
47
]. The phase locking measure (or inter-trial
phase coherence) indicates the degree of consistency in temporal alignment of neural responses to the
stimulus/task characteristics. In the auditory domain, neural oscillatory activities are thought to reflect
different aspects of sound processing. Theta (4–8 Hz) activity has been associated with processing the
temporal and spectral attributes of spoken sentences, respectively [
48
,
49
]. Theta and alpha (8–14 Hz)
bands are also sensitive to the rise in time of the acoustic stimulus onset [
50
,
51
]. We hypothesized that
Brain Sci. 2016,6, 27 4 of 25
in comparison with ramped stimuli, the larger and earlier ON response for damped sounds would be
associated with enhanced phase locking in theta and alpha activities. We were also interested in testing
whether the ramped stimuli would induce stronger gamma activity in the OFF response, which was
shown in an earlier animal study [21].
2. Materials and Methods
2.1. Participants
Six right-handed male adults with normal hearing (thresholds < 25 dB HL for pure tones in the
range of 250–8000 Hz).participated in the study (24–38 in age). They were recruited after screening for
hearing, handedness, and auditory evoked responses. A steady 1 kHz tone (200 ms in duration) served
as a reference stimulus to test the robustness of the auditory ON and OFF responses in each subject.
For the current study, we only included normal-hearing subjects who participated in previous auditory
MEG experiments in the same lab and showed clear N1m response with bilateral dipole activity.
None of the subjects had medical history of speech, language, or hearing disorders. Informed consent was
obtained from each volunteer subject in accordance with approvals from the Institutional Review Boards
at the University of Minnesota and Tokyo Denki University (Ethic approval code: 0605M85808).
2.2. Stimuli
The ramped and damped stimuli consisted of two pairs of sounds (Figure 1). The two sets of
stimuli were 200 ms in physical duration. The first pair of ramped and damped sounds (S stimuli)
was based on a 1000 Hz sine wave tone. A linear fade-in envelope was applied to the simple tone
in making the ramped sound, and the damped sound was its time reversal. This process ensured
that the physical duration, intensity, and spectral contents were identical for the ramped and damped
sounds. The second pair of ramped and damped sounds (C stimuli) was based on a synthesized piano
note with a fundamental frequency of 440 Hz. The piano note was first synthesized in GuitarPro5
(Arobas Music, Lille, France). It had a nonlinear falling intensity envelope. Time reversal was applied
for the ramped counterpart. To verify the existence of ON and OFF responses within each subject,
we used a 200 ms long reference tone at 1000 Hz with a steady intensity envelope with a rise/fall time
of 10 ms. The same 10 ms rise time treatment was applied to the onset of the damped stimuli and the
10 ms fall time to the offset of the ramped stimuli to avoid the “click” percept associated with transient
distortion. All sounds were normalized to have the same RMS (root mean square) average intensity.
Brain Sci. 2016, 6, 27 4 of 24
2. Materials and Methods
2.1. Participants
Six right-handed male adults with normal hearing (thresholds < 25 dB HL for pure tones in the
range of 250–8000 Hz).participated in the study (24–38 in age). They were recruited after screening
for hearing, handedness, and auditory evoked responses. A steady 1 kHz tone (200 ms in duration)
served as a reference stimulus to test the robustness of the auditory ON and OFF responses in each
subject. For the current study, we only included normal-hearing subjects who participated in
previous auditory MEG experiments in the same lab and showed clear N1m response with bilateral
dipole activity. None of the subjects had medical history of speech, language, or hearing disorders.
Informed consent was obtained from each volunteer subject in accordance with approvals from the
Institutional Review Boards at the University of Minnesota and Tokyo Denki University (Ethic
approval code: 0605M85808).
2.2. Stimuli
The ramped and damped stimuli consisted of two pairs of sounds (Figure 1). The two sets of
stimuli were 200 ms in physical duration. The first pair of ramped and damped sounds (S stimuli)
was based on a 1000 Hz sine wave tone. A linear fade-in envelope was applied to the simple tone in
making the ramped sound, and the damped sound was its time reversal. This process ensured that
the physical duration, intensity, and spectral contents were identical for the ramped and damped
sounds. The second pair of ramped and damped sounds (C stimuli) was based on a synthesized piano
note with a fundamental frequency of 440 Hz. The piano note was first synthesized in GuitarPro5
(Arobas Music, Lille, France). It had a nonlinear falling intensity envelope. Time reversal was applied
for the ramped counterpart. To verify the existence of ON and OFF responses within each subject, we
used a 200 ms long reference tone at 1000 Hz with a steady intensity envelope with a rise/fall time of
10 ms. The same 10 ms rise time treatment was applied to the onset of the damped stimuli and the 10
ms fall time to the offset of the ramped stimuli to avoid the “click” percept associated with transient
distortion. All sounds were normalized to have the same RMS (root mean square) average intensity.
Figure 1. Acoustic representations of the S and C ramped and damped stimuli. (a) Sound waveforms;
(b) Power density spectra.
2.3. MEG Recording
The MEG experiment used a whole-scalp 122-channel neuromagnetometer system (Neuromag-
122, Neuromag Ltd., Helsinki, Finland) in a magnetically shielded room at the Research Center for
Advanced Technologies, Tokyo Denki University, Japan. The Neuromag-122 featured an inherent
device coordinate system with 122 sensors at 61 sites covering the whole head. The MEG recording
procedure was completely non-invasive. Prior to the MEG experiment, the subjects were taken to an
MRI facility in the same research center (Stratis II, a 1.5 T Superconductive Magnetic Resonance
Imaging System, Hitachi Co., Tokyo, Japan) for structural brain imaging. The MRI protocol
Figure 1.
Acoustic representations of the S and C ramped and damped stimuli. (
a
) Sound waveforms;
(b) Power density spectra.
Brain Sci. 2016,6, 27 5 of 25
2.3. MEG Recording
The MEG experiment used a whole-scalp 122-channel neuromagnetometer system (Neuromag-122,
Neuromag Ltd., Helsinki, Finland) in a magnetically shielded room at the Research Center for
Advanced Technologies, Tokyo Denki University, Japan. The Neuromag-122 featured an inherent
device coordinate system with 122 sensors at 61 sites covering the whole head. The MEG recording
procedure was completely non-invasive. Prior to the MEG experiment, the subjects were taken to
an MRI facility in the same research center (Stratis II, a 1.5 T Superconductive Magnetic Resonance
Imaging System, Hitachi Co., Tokyo, Japan) for structural brain imaging. The MRI protocol parameters
(TR = 36 ms, FA = 40 degrees, TE = 8.9 ms, NEX = 1.0) for the echo sequence were the same as in
previous MEG publications [
52
,
53
]. The number of slices were 193 with image slice thickness at 1 mm.
The MRIs allowed the construction of realistic head models for each individual subject to improve the
precision of source localization.
During MEG recording, the subjects were seated in a nonmagnetic chair inside the magnetically
shielded room. The subject’s head position in the MEG device was monitored using four head position
indicator (HPI) coils in reference to the spatial coordinate frame defined by the nasion, the left and
right preauricular points relative to the individual’s MRI head model. The positions of the HPI coils
with respect to the anatomical landmarks were first measured with a three-dimensional Polhemus
Isotrak digitizer outside the shielded room. When the subjects were seated under the MEG dewar,
HPIs were measured to ensure that the positioning accuracy was within 98%–100% for each coil.
Vertical electro-oculograms (EOGs) were recorded online with a pair of bipolar electrodes pasted at
the supraorbital and infraorbital ridge of the right eye. The impedances of the bipolar electrodes
were lower than 5 kOhm. The stimuli were binaurally delivered at a sensation level of 50 dB via
non-magnetic foam earplugs through a non-echoic plastic tube system. The sensation level was
individually calibrated for each subject using the steady 1 kHz reference tone with its mean RMS
(root mean square) level matched to that of the ramped and damped stimuli. Specifically, the binaural
hearing threshold for the 1 kHz steady tone was determined for each individual subject at the beginning
of the MEG session. Subjects were asked to verify that they could hear the acoustic stimuli clearly at
a comfortable level before proceeding to MEG recording. This stimulus presentation method based
on individually calibrated audiometric sensation level has previously been applied in our hearing
research studies [54,55] and other MEG studies [50,56].
A commonly used passive listening paradigm was adopted with a distraction task to help
minimize potential differences due to preferential listening [
57
]. Stimulus presentation used
an alternating short block design. There were five different stimuli, including the ramped and
damped sounds in simple and complex conditions and the reference tone. Each short block consisted
of 20 identical stimuli with an interstimulus interval randomized in the range of 1000–1100 ms.
The inter-block interval was 5 s. No two identical stimulus blocks were presented consecutively.
A similar alternating block design had been used in a previous study [
58
]. In the present study,
the subjects were instructed to watch a self-chosen movie projected to a white screen, which was
placed approximately 1.5 m in front of the subject. During the experiment, the subjects were asked to
concentrate on the muted movie with subtitles and ignore the auditory stimuli. The MEG signals were
bandpass-filtered from 0.03 to 100 Hz and digitized at 497 Hz. Epochs with amplitude greater
than 3000 fT/cm or EOG greater than 150
µ
V were rejected to exclude data with blinking and
movement-related artifacts or other noise contamination. For each subject, at least 80 good epochs
were averaged for each stimulus as in our previous MEG publications [52,53].
2.4. Global Field Power Analysis
To verify the experimental hypotheses, we conducted global field power analysis as well as
two source localization methods [
59
]. In the waveform analysis, the raw MEG data were digitally
low-pass filtered at 40 Hz and corrected with a pre-stimulus baseline of 100 ms. As our interest was
in the neural coding of the physical stimulus features instead of neural sensitivity in discriminatory
Brain Sci. 2016,6, 27 6 of 25
responses, the first stimulus in each block was omitted from averaging to avoid possible elicitation of
a mismatch field response to the previous block of stimuli. MEG waveform amplitude was defined
as the vector sum of amplitudes at two orthogonal channels in the same sensor location. To derive
a composite measure of differences in MEG activity for the averaged ON and OFF responses for the
ramped and damped stimuli, we analyzed the global field power (GFP) for each stimulus and each
subject. Similar to the calculation of GFP for EEG signals (e.g., [
54
]), the GFP is the root mean square of
magnetic fields across all the 61 recording sensor sites at each time sample. The GFP measure has been
shown to be an objective and reliable quantification method independent of sensor selection in EEG
and MEG studies [6062].
2.5. Source Localization Analysis
To localize the ON and OFF MEG components, we employed distributed source analysis using
minimum norm estimation (MNE). The MNE approach is thought to be more appropriate when the
distribution of source activity is poorly known [
63
]. The MNE analysis procedure here was performed
using the MNE-Suite by Dr. Matti Hämäläinen, including the following steps:
(1)
Head model preparation. The individual subject’s MRIs were converted and processed using the
Freesurfer software to derive the boundary element models.
(2)
MRI-MEG co-registration. After stimulus set operations were performed for filtering, artifact
rejection and averaging, the MEG waveform data were loaded for head position adjustment
relative to each subject’s Head Position Indicator (HPI) data and the 3-D locations of the nasion,
left and right preauricular points. We used four HPI coils in the experiment; two were positioned
in the forehead, and the other two were right behind the two ears.
(3)
Standardized forward and inverse solution. Forward and inverse solutions were derived
following the recommended settings in the MNE-Suite. Baseline noise covariance matrix was
calculated for each averaged data set. To perform MNE averaging across subjects and regions of
interest analysis, the individual MNE data were morphed to a standard brain model.
(4)
Regions of Interest (ROI) analysis. Based on the brain activation patterns, two ROIs in the
standard brain space were chosen, namely, the superior temporal (ST) and inferior parietal (IP).
The ROIs were anatomically defined and annotated with corresponding Talairach coordinates
in the MNE-Suite. The individual MNE waveforms for each ROI were exported and further
analyzed in Matlab.
(5)
MNE movie generation. For visualization purposes, the grand mean results for each stimulus
were exported as movie files with MNE results expressed in dSPM (dynamic statistical parametric
mapping) values integrated at every 10 ms [
64
,
65
]. The movie frames with peak activities were
selected in the two post-stimulus windows, 70–150 ms and 250–350 ms, and plotted to illustrate
the spatial localization of the ON and OFF responses.
To help determine whether the localization results were similar for the ON and OFF responses
in each direction of the x, y, and z head coordinates, we also applied the single moving equivalent
current dipole (ECD) model in each hemisphere [
63
]. Grand mean isofield contour map data for the
auditory ON and OFF responses were extracted with BESA (Version 6.0, MEGIS Software GmbH,
Gräfelfing, Germany). ECD modelling is often adopted in the analysis of auditory evoked responses,
which predominantly show dipolar field distributions. In the ECD approach, an optimal solution is
sought in the calculation of the dipole parameters in terms of location, amplitude, and direction of
the source current by minimizing the least-square error between the measured and predicted signals.
However, the point-like ECD source solution might be an oversimplification as the auditory N1 activity
has been shown to contain several subcomponents with contributions from temporal, parietal and
frontal regions in both hemispheres [
30
,
66
69
]. In contrast, the distributed source MNE model not only
shows the spread of activity but also allows visualization of simultaneous activities at multiple sites
as well as the strength of activity at different vertices within a specified region of interest. While the
Brain Sci. 2016,6, 27 7 of 25
ECD solutions are restrictive in the number of active point sources without any information about the
extent of activation, the MNE results tend to be biased towards superficial sources without properly
measuring the depth information [
70
]. In this regard, we combined both point source and distributed
source modelling in order to take advantage of their complimentary features to explain the cortical
source activities of the ON and OFF responses.
2.6. Phase-Locking Factor Analysis
Phase-locking factor (PLF) in gamma (30–70 Hz), alpha (8–14 Hz), and theta (4–8 Hz) bands
were analyzed for each subject. Phase consistency across trials was computed based on Morlet
wavelet analysis at the target frequency bands [
47
,
71
,
72
]. The absolute PLF values in the range of 0
(non-phase-locked activity) and 1 (strictly phase-locked activity) were calculated for the ramped and
damped sounds using PLF functions from the Matlab-based 4d Toolbox (version 1.2 developed by
Ole Jensen) [
72
]. The representation of the phase for trial j,
Φj
, at a given time (t) and frequency (f
0
),
was derived from the convolution of the complex wavelets defined as Formula (1–3)
wpt,f0q “ A exp t2{p2σt2qq exp p2iπf0tq(1)
where
σtm/2πf0(2)
A1{p2πσt2q1/2 (3)
iis the imaginary unit, mdefines the compromise between time resolution and frequency resolution,
and f
0
is the center frequency. The wavelet had a Gaussian shape in both the time domain and the
frequency domain around its central frequency f
0
. The width of Morlet wavelet (see Formula (4)) was
set at 7 [47].
pmf0{σfq(4)
The phase for trial jwas obtained for the signal sj(t),normalized by the amplitude, see Formula (5).
Φjpt,f0q “ wpt,f0qsjptq{|wpt,f0qsjptq|. (5)
The spectral amplitudes of the wavelets were normalized so that the total energy was 1.
The normalized complex time-varying energy of each single trial was averaged across trials for
each stimulus. The PLF value over N trials was calculated as Formula (6).
PLVpt,f0q “ 1{N|ÿj= 1NΦjpt,f0q| (6)
We specifically used a set of wavelets with the frequency (f0) ranging from 1 to 90 Hz in 1 Hz steps.
All the PLF computations were conducted on the individual subjects’ trial-by-trial raw data
after artifact rejection without applying the 40 Hz lowpass filter. As previous research has shown
that OFF response plays a more important role in duration coding [
38
], the gradiometer sensor that
showed the largest OFF responses in MEG waveform analysis was selected for each stimulus and each
subject. The mean absolute PLF values were averaged for theta band (4–8 Hz), alpha band (8–14 Hz),
and gamma band (30–70 Hz) in a window of 20 ms centering around the ON and OFF peaks for
statistical comparisons between ramped and damped sounds.
2.7. Behavioral Tests on Subjective Duration
Right after the passive listening session for MEG recording, two separate active listening tests
with 1-min breaks in between were administered using the same stimulus presentation level and
ear insert setting to verify the perceptual bias for auditory looming in terms of subjective duration.
In Test 1, the subjects were asked to press buttons to indicate whether the two sounds presented
Brain Sci. 2016,6, 27 8 of 25
in a stimulus-pair trial were equally long. If not, they were further instructed to indicate whether
the first or second stimulus sounded longer. The two sounds in a stimulus pair were separated by
a 250 ms silence interval, and the inter-trial interval was randomized in the range of 1500–2000 ms.
The stimulus presentation order was randomized for the stimulus-pair trials consisting of two stimulus
conditions (S and C stimuli), two trial orders (ramped-damped and damped-ramped), and two foil
orders (ramped-ramped and damped-damped). Each stimulus pair was presented 20 times. In Test 2,
the subjects were asked to perform a duration judgment task with a procedure modified after
DiGiovanni and Schlauch (2007) [
7
]. In this task, the subjects were instructed to include all aspects
of the auditory stimuli when making their judgments. In a given trial, the listeners were required to
adjust the length of a steady 1 kHz tone to match the duration of the target sound (one of the five
sounds for the MEG recordings; four S and C ramped/damped stimuli and the 200 ms steady 1 kHz
reference tone). Linear rise/decay of 10 ms was applied to the adjustable steady sound. The adjustable
duration range was set at
˘
80% of the target [
6
]. The inter-stimulus interval for each matching sound
pair (target sound followed by adjustable sound) was 500 ms, and the inter-trial interval was 1500 ms.
Trials were presented in random order with each target sound tested 20 times.
2.8. Statistical Analysis
For all behavioral and MEG measures, we conducted statistical tests for normality, variance
homogeneity, and outlier detection to ensure no existence of statistical outliers or other
variability/normality problems that would cause a violation of the assumptions of the statistical
tests. Results from behavioral Test 1 were calculated in terms of percentage of trials for which the
ramped sounds were judged to have a longer duration. Response sensitivity in terms of d-prime score
were then calculated using signal detection theory [
73
]. Given the categorical behavioral responses,
the Fisher’s Exact Test was applied to each individual subject’s data to verify the subjective duration
bias for auditory looming. Magnitude estimation results of behavioral Test 2 were averaged for
each target sound for each individual subject for an assessment of the relative strength of perceptual
asymmetry for the S and C stimuli. A repeated-measures ANOVA test was conducted for the perceived
duration data to verify the significance of the temporal asymmetry phenomenon with the two main
factors, stimulus intensity envelope (ramped vs. damped) and stimulus condition (S vs. C stimuli).
Repeated-measures ANOVA tests were performed on the peak ON and OFF responses for the GFP,
ECD, and MNE data, respectively. The main factors of interest included intensity envelope (ramped vs.
damped), response type (ON vs. OFF), stimulus condition (S vs. C stimuli), and hemisphere (left vs.
right). The hemisphere factor did not apply to the GFP data as the root mean square calculation used
for all 122 MEG sensors. To visualize the temporal evolution of significant differences between ramped
and damped stimuli, two-tailed point-to-point t-tests were conducted over the entire epoch [
58
,
74
].
In order for an interval to be considered significantly different between the two sounds, at least eight
consecutive points (approximately 16 ms) needed to reach the significance level of 0.01 [52,58,75].
To examine brain-behavior correlates, we adopted the temporal asymmetry index (TAI) formula
for both behavioral and brain measures (see Formula (7)) [24,40,76],
pQramped ´Qdampedq{pQramped `Qdampedq(7)
where Q stands for measures of the same type from an individual subject. In calculating the neural
TAI measures, we specifically looked at the ON and OFF responses in the left and right hemispheres
separately in terms of ECD amplitude and latency data. To confirm the correlation analysis with the
ECD data, TAI measures were also derived from MNE amplitude and latency data in the superior
temporal region. We were particularly interested in finding out which MEG response measure would
be a better predictor of perceived duration asymmetry and which source localization method provided
stronger correlation results. As the linear assumption in Pearson’s correlation is likely problematic for
the TAI measures for brain and behavioral responses, we adopted the Spearman rank correlation that
Brain Sci. 2016,6, 27 9 of 25
assumed a monotonic but not necessarily linear relationship [
77
]. Considering the sample size, we used
the Spearman test function in a new Matlab toolbox for robust correlation analysis, which included
a resampling bootstrap procedure to verify the significance of the correlation coefficient obtained
from the small sample [
78
]. We adopted false discovery rate (FDR) with the Benjamini-Hochberg
procedure [79] for correcting the p-values in multiple comparisons.
As our study was primarily interested in investigating the relationship between perceptual temporal
asymmetry and the ON and OFF responses of the ramped and damped stimuli, the comparison with
the steady reference tone was not included or discussed in the main text. More statistical results are
provided in the appendix to show the comparison of sounds with steady, rising and falling intensity
envelopes (Figures A1 and A2).
3. Results
3.1. Behavioral Data
Behavioral results replicated previous findings. The average d’ scores were above 3 for detecting
differences in subjective duration of the ramped and damped stimuli (3.7 for S stimuli and 3.6 for
C stimuli). Repeated measures ANOVA results showed a significant main effect of stimulus intensity
envelope (ramped vs. damped) (F(1, 5) = 94.23, p< 0.001), no significant effect of stimulus type
(S vs. C stimuli), and no significant interaction between intensity envelope and stimulus type.
Each subject judged the ramped sounds to be of longer duration than the damped sounds for both the
simple and C stimulus conditions (p< 0.00001; Fisher’s Exact Test). On average, the ramped sound was
perceived to be longer in 92.5% of the trials for the S stimuli and 91.7% of the trials for the C stimuli.
Duration matching results showed that the ramped sound was perceived to be 25.1% longer than
the damped sound for the S stimuli (219.2 ms vs. 175.2 ms). Consistent with previous behavioural
studies [
3
], a smaller perceptual bias effect was observed for the C stimuli with the ramped sound
perceived to be 14.7% longer than the damped sound (211.7 ms vs. 184.5 ms).
3.2. Global Field Power (GFP) Data
Repeated measures ANOVA results for GFP peak amplitude data showed a significant main effect
of intensity envelope (ramped vs. damped) (F(1, 5) = 70.1, p< 0.001) (Figure 2). There was also a main
effect of MEG response type (on vs. off)—the ON response was much larger than the OFF response
(F(1, 5) = 94.34, p< 0.001). A significant interaction was observed between intensity envelope and MEG
response type (F(1, 5) = 68.59, p< 0.001). Post-hoc two-tailed t-tests confirmed greater ON responses
for the damped stimuli than for the ramped stimuli in both S and C stimulus conditions (p< 0.01).
In the OFF response, the S stimulus condition showed dominance for the ramped sound whereas the
C stimulus condition did not (p< 0.05).
Peak latency data showed significantly later ON response for the ramped sounds than the damped
sounds (F(1, 5) = 96.8, p< 0.00001). Although the GFP data also showed a trend in the OFF response
latency in line with our hypothesis, there was no significant difference in the off latency responses
between the ramped and damped stimuli in either stimulus condition. These patterns were reflected
in the point-to-point comparisons in GFP data (Figure 1) as well as in the isofield contour maps at the
MEG sensor level (Figure 2).
Brain Sci. 2016,6, 27 10 of 25
1
Figure 2.
Neuromagnetic responses to the S and C ramped and damped stimuli. (
a
) Grand mean global
field power of the MEG responses. Significant differences between the ramped and damped sounds
are indicated on the horizontal axis (p< 0.01). Intervals with larger amplitudes for damped sounds
are marked by white bars on the x-axis, and intervals with larger amplitudes for ramped sounds are
marked by black bars. (b) Isofield contour maps of the grand mean ON and OFF responses.
3.3. Point Source Modelling: Equivalent Current Dipole (ECD) Data
The ECD locations were similar for the ramped and damped sounds in both hemispheres.
No significant differences in any of the x, y, and z dimensions were found for the ECD source location
between the ramped and damped sounds in either S or C stimulus condition. Repeated measures
ANOVA results of the ECD amplitude data showed a significant main effect of intensity envelope
(ramped vs. damped) (F(1, 5) = 72.3, p< 0.001). There was also a main effect for MEG response type
(on vs. off) (F(1, 5) = 116.73, p< 0.001).
The dominant ON responses were observed for the damped sounds relative to the ramped
sounds, which was observed in both hemispheres for both stimulus conditions (p< 0.01; post-hoc
two-tailed t-test). In the OFF response, there was a significant interaction between stimulus type
(ramped vs. damped), stimulus condition (simple vs. complex) and hemisphere (F(1, 5) = 11.32,
p< 0.05). The S stimulus condition showed auditory OFF response dominance in the right hemisphere
Brain Sci. 2016,6, 27 11 of 25
for the ramped sound in comparison with the damped sound (p< 0.05; post-hoc two-tailed t-test).
Like in the GFP data, the C stimulus condition did not show dominance in OFF response in favour of
the ramped sound over the damped sound in either hemisphere. Point-to-point comparisons of the
ECD source waveforms further confirmed these patterns (Figure 3).
Brain Sci. 2016, 6, 27 10 of 24
3.3. Point Source Modelling: Equivalent Current Dipole (ECD) Data
The ECD locations were similar for the ramped and damped sounds in both hemispheres. No
significant differences in any of the x, y, and z dimensions were found for the ECD source location
between the ramped and damped sounds in either S or C stimulus condition. Repeated measures
ANOVA results of the ECD amplitude data showed a significant main effect of intensity envelope
(ramped vs. damped) (F(1, 5) = 72.3, p < 0.001). There was also a main effect for MEG response type
(on vs. off) (F(1, 5) = 116.73, p < 0.001).
The dominant ON responses were observed for the damped sounds relative to the ramped
sounds, which was observed in both hemispheres for both stimulus conditions (p < 0.01; post-hoc
two-tailed t-test). In the OFF response, there was a significant interaction between stimulus type
(ramped vs. damped), stimulus condition (simple vs. complex) and hemisphere (F(1, 5) = 11.32, p <
0.05). The S stimulus condition showed auditory OFF response dominance in the right hemisphere
for the ramped sound in comparison with the damped sound (p < 0.05; post-hoc two-tailed t-test).
Like in the GFP data, the C stimulus condition did not show dominance in OFF response in favour
of the ramped sound over the damped sound in either hemisphere. Point-to-point comparisons of
the ECD source waveforms further confirmed these patterns (Figure 3).
Figure 3. Grand mean equivalent current dipole (ECD) waveform results in the left and right auditory
cortices for the ramped and damped sounds in the two stimulus conditions: (a) S stimuli; and (b) C
stimuli. Significant differences between the ramped and damped sounds are indicated on the
horizontal axis (p < 0.01). Intervals with larger amplitudes for damped sounds are marked by white
bars on the X-axis, and intervals with larger amplitudes for ramped sounds are marked by black bars.
3.4. Distributed Source Modelling: Minimum Norm Estimation (MNE) Data
In both S and C stimulus conditions, the MNE data showed two main regions of bilateral activity
(superior temporal and inferior parietal) for the ON response (Figure 4). In both superior temporal
(ST) and inferior parietal (IP) regions, there was a significant effect of response type (on vs. off) with
larger ON response than OFF response (F(1, 5) = 18,28, p < 0.01 for ST; F(1, 5) = 18,84, p < 0.01 for IP).
The ON responses were greater for the damped sounds than those for the ramped sounds in the ST
and IP regions, which was further confirmed in separate post-hoc t-tests for each hemisphere and
each stimulus condition (p < 0.05).
Figure 3.
Grand mean equivalent current dipole (ECD) waveform results in the left and right
auditory cortices for the ramped and damped sounds in the two stimulus conditions: (
a
) S stimuli;
and (
b
) C stimuli. Significant differences between the ramped and damped sounds are indicated on the
horizontal axis (p< 0.01). Intervals with larger amplitudes for damped sounds are marked by white
bars on the X-axis, and intervals with larger amplitudes for ramped sounds are marked by black bars.
3.4. Distributed Source Modelling: Minimum Norm Estimation (MNE) Data
In both S and C stimulus conditions, the MNE data showed two main regions of bilateral activity
(superior temporal and inferior parietal) for the ON response (Figure 4). In both superior temporal
(ST) and inferior parietal (IP) regions, there was a significant effect of response type (on vs. off) with
larger ON response than OFF response (F(1, 5) = 18,28, p< 0.01 for ST; F(1, 5) = 18,84, p< 0.01 for IP).
The ON responses were greater for the damped sounds than those for the ramped sounds in the ST
and IP regions, which was further confirmed in separate post-hoc t-tests for each hemisphere and each
stimulus condition (p< 0.05).
The OFF response showed dominance in favour of the ramped sounds only in the ST region
(F(1, 5) = 10.24, p< 0.05). There were significant interactions among stimulus condition (S vs. C) and
hemisphere (left vs. right) in both the ST (F(1, 5) = 23.31, p< 0.01) and IP (F(1, 5) = 6.65, p< 0.05)
regions, suggesting that stimulus complexity affected the involvement of the auditory areas in the two
hemispheres differently in the OFF response. Post-hoc t-tests showed OFF response dominance in
the right ST region for the ramped sounds in both stimulus conditions (p< 0.05). The MNE activity
patterns were confirmed in time-point-by-time-point t-test for the two regions of interest for each
stimulus condition (Figure 4). Consistent with the GFP and ECD results, the point-to-point MNE
comparison for the damped and ramped stimuli did not show evidence for our hypothesis that the
Brain Sci. 2016,6, 27 12 of 25
ramped stimuli would elicit stronger sustained activities following the OFF response in either the ST
or IP region in a passive listening condition.
Figure 4.
Grand mean MNE waveforms in the two regions of interest (ST and IP). (
a
) ST; superior
temporal region. (
b
) IP; inferior parietal region. Bars on x-axis indicate significant differences between
the ramped and damped stimuli using the same convention as in Figure 2.
3.5. Percentage Differences in ON and OFF Latencies
For a direct comparison with previously reported behavioral data, percentage differences in
ON and OFF response latencies between ramped and damped stimuli were calculated for all the
MEG analysis techniques we used. On average, the ON response latency for the S ramped sound
was delayed by 15.8% relative to the S damped sound, which was consistently observed in the GFP
Brain Sci. 2016,6, 27 13 of 25
peak measure. For the C stimuli, a similar ON response delay of 14.2% was observed. Consistent
percentages of delay in the ramped stimuli were found in the ECD analysis (for the S stimuli, 19.8% in
the left brain and 14.9% in the right; for the C stimuli, 16.4% in the left brain and 15.2% in the right) as
well as in the MNE analysis (for the S stimuli, 18.9% in the left brain and 16.3% in the right; for the
C stimuli, 20.7% in the left brain and 18.4% in the right). The grand mean off-minus-on latency value
showed a 17.0% longer duration for the S ramped sound than the S damped sound in the GFP peak
data. However, the C stimuli did not show such a pattern; in fact, the grand mean off-minus-on latency
value was 21.1% shorter in the C ramped sound relative to the C damped sound.
3.6. Phase-Locking Factor Data
The PLF results confirmed our hypothesis that different neural oscillation patterns mediated the
neural coding of rising vs. falling amplitude modulation (Figure 5). In particular, the stronger ON
responses for the damped sounds were coupled with stronger PLF at the MEG sensor level in delta
(4–7 Hz) (F(1, 5) = 38.10, p< 0.001) and alpha (8–14 Hz) (F(1, 5) = 26.36, p< 0.01) bands. In contrast,
the OFF responses for the ramped sounds were coupled with stronger PLF in gamma (30–70 Hz) in
comparison with damped sounds (F(1, 5) = 7.10. p< 0.05). Post-hoc tests confirmed these significant
differences in PLF in both the S and C stimulus conditions (p< 0.05).
Brain Sci. 2016, 6, 27 12 of 24
3.5. Percentage Differences in ON and OFF Latencies
For a direct comparison with previously reported behavioral data, percentage differences in ON
and OFF response latencies between ramped and damped stimuli were calculated for all the MEG
analysis techniques we used. On average, the ON response latency for the S ramped sound was
delayed by 15.8% relative to the S damped sound, which was consistently observed in the GFP peak
measure. For the C stimuli, a similar ON response delay of 14.2% was observed. Consistent
percentages of delay in the ramped stimuli were found in the ECD analysis (for the S stimuli, 19.8%
in the left brain and 14.9% in the right; for the C stimuli, 16.4% in the left brain and 15.2% in the right)
as well as in the MNE analysis (for the S stimuli, 18.9% in the left brain and 16.3% in the right; for the
C stimuli, 20.7% in the left brain and 18.4% in the right). The grand mean off-minus-on latency value
showed a 17.0% longer duration for the S ramped sound than the S damped sound in the GFP peak
data. However, the C stimuli did not show such a pattern; in fact, the grand mean off-minus-on
latency value was 21.1% shorter in the C ramped sound relative to the C damped sound.
3.6. Phase-Locking Factor Data
The PLF results confirmed our hypothesis that different neural oscillation patterns mediated the
neural coding of rising vs. falling amplitude modulation (Figure 5). In particular, the stronger ON
responses for the damped sounds were coupled with stronger PLF at the MEG sensor level in delta
(4–7 Hz) (F(1, 5) = 38.10, p < 0.001) and alpha (8–14 Hz) (F(1, 5) = 26.36, p < 0.01) bands. In contrast, the
OFF responses for the ramped sounds were coupled with stronger PLF in gamma (30–70 Hz) in
comparison with damped sounds (F(1, 5) = 7.10. p < 0.05). Post-hoc tests confirmed these significant
differences in PLF in both the S and C stimulus conditions (p < 0.05).
Figure 5. Grand mean phase locking factor results for the S stimulus condition (a) and C stimulus
condition (b). Enhanced gamma band activities associated with the OFF response were marked in
dotted circles with an arrow assign for the ramped sounds. Enhanced theta and alpha activities
associated with the ON response were indicated by the arrow sign for the damped sounds.
3.7. Brain-Behavior Correlates of Temporal Asymmetry
Spearman rank correlation analysis showed significant results (FDR-corrected) only in the OFF
response ECD amplitude of the left auditory cortex for the S stimuli (Table 1 and Figure 6). Even
though the ON responses showed a robust effect of ramped vs. damped differences, its temporal
asymmetry index (TAI) scores for the ECD data did not show any significant correlations with
behavioral data for either the S or the C stimuli. While we observed an overall consistency in
percentage differences in the off-minus-on latency for ramped vs. damped sounds with the
behavioral percentages of perceived duration, the MEG latency data for ON and OFF responses in
Figure 5.
Grand mean phase locking factor results for the S stimulus condition (
a
) and C stimulus
condition (
b
). Enhanced gamma band activities associated with the OFF response were marked in
dotted circles with an arrow assign for the ramped sounds. Enhanced theta and alpha activities
associated with the ON response were indicated by the arrow sign for the damped sounds.
3.7. Brain-Behavior Correlates of Temporal Asymmetry
Spearman rank correlation analysis showed significant results (FDR-corrected) only in the OFF
response ECD amplitude of the left auditory cortex for the S stimuli (Table 1and Figure 6). Even though
the ON responses showed a robust effect of ramped vs. damped differences, its temporal asymmetry
index (TAI) scores for the ECD data did not show any significant correlations with behavioral data for
either the S or the C stimuli. While we observed an overall consistency in percentage differences in the
off-minus-on latency for ramped vs. damped sounds with the behavioral percentages of perceived
duration, the MEG latency data for ON and OFF responses in the ECD data did not show any significant
brain-behavior correlations. The same brain-behavior correlation patterns were also found with the
MNE data (Table 1).
Brain Sci. 2016,6, 27 14 of 25
Brain Sci. 2016, 6, 27 13 of 24
the ECD data did not show any significant brain-behavior correlations. The same brain-behavior
correlation patterns were also found with the MNE data (Table 1).
Figure 6. Sample scatter plots of the temporal asymmetry index (TAI) measures with Spearman rank
correlation analysis. The OFF response dipole moment (amplitude) was used in calculating the neural
TAI scores in the left (LH) and right (RH) hemispheres for the S and C stimuli. Full details of the brain-
behavior correlate measures are reported in Table 1.
Table 1. Spearman correlation coefficient (ρ) results for brain-behavior correlations in temporal
asymmetry index scores, (Qramped Qdamped)/(Qramped + Qdamped), for the S and C stimuli (df = 5, ECD =
equivalent current dipole, MNE = minimum norm estimation, LH = left hemisphere, RH = right
hemisphere).
Brain Measures S Stimuli C STIMULI
ON Response OFF Response ON Response OFF Response
ECD
Amplitude (LH) ρ = 0.429 ρ = 0.943 * ρ = 0.257 ρ = 0.371
Latency (LH) ρ = 0.551 ρ = 0.086 ρ = 0.714 ρ = 0.086
Amplitude (RH) ρ = 0.143 ρ = 0.543 ρ = 0.371 ρ = 0.714
Latency (RH) ρ = 0.577 ρ = 0.086 ρ = 0.377 ρ = 0.029
MNE
Amplitude (LH) ρ = 0.257 ρ = 0.829 * ρ = 0.200 ρ = 0.086
Latency (LH) ρ = 0.200 ρ = 0.600 ρ = 0.489 ρ = 0.200
Amplitude (RH) ρ = 0.657 ρ = 0.714 ρ = 0.257 ρ = 0.486
Latency (RH) ρ = 0.029 ρ = 0.493 ρ = 0.086 ρ = 0.543
Significant correlations are indicated with a star sign (* stands for p < 0.05, FDR-corrected). The
significance of the correlational coefficients in bold italics obtained from our limited number of data
points was also confirmed in the resampling bootstrap procedure of up to 1000 samples using the
robust correlation toolbox [78].
Figure 6.
Sample scatter plots of the temporal asymmetry index (TAI) measures with Spearman rank
correlation analysis. The OFF response dipole moment (amplitude) was used in calculating the neural
TAI scores in the left (LH) and right (RH) hemispheres for the S and C stimuli. Full details of the
brain-behavior correlate measures are reported in Table 1.
Table 1.
Spearman correlation coefficient (
ρ
) results for brain-behavior correlations in temporal asymmetry
index scores, (Q
ramped ´
Q
damped
)/(Q
ramped
+ Q
damped
), for the S and C stimuli (df = 5, ECD = equivalent
current dipole, MNE = minimum norm estimation, LH = left hemisphere, RH = right hemisphere).
Brain Measures S Stimuli C STIMULI
ON Response OFF Response ON Response OFF Response
ECD
Amplitude (LH) ρ= 0.429 ρ=0.943 * ρ= 0.257 ρ= 0.371
Latency (LH) ρ=´0.551 ρ= 0.086 ρ=0.714 ρ=´0.086
Amplitude (RH) ρ=0.143 ρ=0.543 ρ=0.371 ρ=0.714
Latency (RH) ρ=´0.577 ρ=´0.086 ρ=0.377 ρ=´0.029
MNE
Amplitude (LH) ρ=0.257 ρ=0.829 * ρ=0.200 ρ=0.086
Latency (LH) ρ=0.200 ρ=0.600 ρ=0.489 ρ=0.200
Amplitude (RH) ρ=0.657 ρ=0.714 ρ=0.257 ρ=0.486
Latency (RH) ρ=0.029 ρ=0.493 ρ=0.086 ρ=0.543
Significant correlations are indicated with a star sign (* stands for p< 0.05, FDR-corrected). The significance of
the correlational coefficients in bold italics obtained from our limited number of data points was also confirmed
in the resampling bootstrap procedure of up to 1000 samples using the robust correlation toolbox [78].
4. Discussion
4.1. Dominant Auditory ON Response in Favour of Falling Intensity
Despite the limited number of subjects in our study, the analyses using GFP, ECD, MNE and PLF
data consistently showed dominant on-N1 response for the damped sounds relative to the ramped
sounds regardless of spectral complexity of the stimuli. The point-to-point comparisons for the MEG
data clearly demonstrated earlier and larger ON responses for the damped sounds relative to the
ramped sounds, and such temporal details were not available from previous functional Magnetic
Resonance Imaging (fMRI) data on auditory looming. The MEG source localization results indicated
that both left and right auditory cortices contributed to the dominance of the ON response in favour of
Brain Sci. 2016,6, 27 15 of 25
sounds with falling intensity. This pattern was attributable to the fact that damped sounds involved
an abrupt change from silence to the maximum intensity at the stimulus onset. In contrast, the ramped
sounds had a much slower intensity change at the onset, effectively reducing and delaying the on-N1
response [80].
4.2. Limited Evidence for Dominant Auditory OFF Response in Favor of Rising Intensity
The elicitation of auditory OFF responses was consistent with previous studies that used sounds
longer than 100 ms [
23
,
26
,
27
,
31
]. Consistent with previous fMRI data [
81
], our MNE results showed
right hemisphere dominance for encoding the OFF response to the rising intensity in the auditory
stimuli specifically in the superior temporal regionUnlike the auditory ON responses, there was not
a uniformly significant effect across all the MEG measures when comparing the OFF responses of
the ramped and damped stimuli. It was previously shown that cortical ON responses are encoded
more readily and accurately than OFF responses [
25
]. In our study, the GFP results showed stronger
OFF response in favour of the ramped sounds only in the S stimulus condition. This is consistent
with previous reports about a reduced asymmetry effect in spectral complex sounds, which is
possible from prior learning experience and stimulus familiarity [
1
,
3
,
10
,
13
,
14
]. Nevertheless, the MNE
and PLF data showed the asymmetry pattern to be biased towards ramped sounds in both S and
C stimulus conditions.
4.3. Source Localization for ON and OFF Responses
The source localization results suggest that the ON and OFF auditory evoked response may share
the same or overlapping cortical sites for coding abrupt acoustic change as previously suggested [
26
28
].
We did not find statistically different source locations for off and ON responses as has been noted in
two physiological studies where OFF responses were found to be either slightly more anterior [
35
]
or slightly more superior than the ON response [
32
]. As we tested only six subjects with large
intersubject variability in the source localization data, subtle differences and small effects might be
hard to verify with a small sample. While animal neurophysiology work suggest that different neuron
clusters or projection pathways may be responsible for coding ON and OFF response at the levels
of thalamus [
82
,
83
] and auditory cortex [
18
], the spatial resolution of our ECD and MNE analysis
methods for the human MEG data as implemented in the current study is rather limited and thus may
not be able to separate dipole sources for ON and OFF responses that are within the spatial radius
of 10 mm [52].
4.4. Distinct Neural Oscillations for Rising and Falling Intensities
The PLF data indicate the involvement of distinct neural oscillations for tracking the rising vs.
falling intensity modulation direction within the acoustic stimuli. In the present context with a passive
listening condition, the dominant PLF for the ON response to the damped sounds was mediated
by stronger alpha and theta activity, reflecting new information coding for the abrupt onset [
50
,
51
].
As the onset of damped sounds is prone to capture attention/arousal/alertful reaction, there could
be differences in involuntary attention to the arrival of the damped vs. ramped sounds mediated
by alpha activity, which is known to be influenced by attention [
84
]. The dominant PLF for the OFF
response to the ramped sounds was mediated by gamma activity, reflecting integrated temporal coding
of the rising intensity envelope. These data are consistent with animal neurophysiological findings.
Recent
in vivo
patch-clamp whole-cell recordings from the primary auditory cortex of anesthetized
rats indicate that the ON and OFF responses are driven by largely non-overlapping sets of synaptic
inputs in the auditory cortex [
85
]. Interestingly, the gamma activities associated with ramped auditory
stimuli in our study correspond nicely with MEG data for visual processing of moving vs. stationary
stimuli with moving objects eliciting higher gamma oscillations [86].
An alternative explanation for the distinct neural oscillation patterns is that phase synchrony may
reflect expectation or stimulus predictability [
87
,
88
]. One could argue that the offset for a damped
Brain Sci. 2016,6, 27 16 of 25
sound is more predictable than that for a ramped sound. In addition, our stimulus presentation
protocol used variable inter-stimulus intervals (ISIs) between trials, which would make the sound
offsets more predictable than sound onsets. Thus, endogenous anticipatory processes may have induced
or contributed to differences between the oscillatory responses elicited by the sound offsets and onsets.
While it is appealing to interpret the phase-locking factor as a measure of oscillatory activity on
a phase-resetting account, caution is necessary here as we cannot rule out the traditional additive
model for evoked responses [
51
]. Our PLF data as reported cannot provide conclusive evidence to
cleanly separate what might be due to phase-resetting of ongoing oscillations and what might be due to
additive evoked response (possibly non-oscillatory) for stimulus coding. In our view, the two models
are not necessarily exclusive of each other. Both the phase locking factor and the auditory evoked
response measures capture information about neural synchrony across trials in an event-related
experimental design.
4.5. Neural Correlates of the Perceptual Temporal Asymmetry
Spearman correlation analysis revealed significant results only in the amplitude measure of the
OFF response in the left auditory cortex for the S stimuli. This phenomenon is consistent with our
prediction based on the behavioral literature [
1
,
3
,
10
,
13
,
14
]. Behavioral data showed that the ramped
sound was 25.1% longer than the damped sound in the S stimulus condition. The subjective duration
difference was reduced to 14.7% for the C stimuli. Spectral complexity and sound familiarity of the
piano-sound quality of the C stimuli mostly likely reduced the monotonic relationship between the
temporal asymmetry index scores for the neural responses obtained in a passive listening condition
and behavioral results obtained in an active listening condition. The fact that the ECD and MNE
data showed consistent correlation results suggests that both source localization methods provide
good estimates of the ON and OFF activities in left and right auditory cortices. The fact that
significant brain-behavior correlation was only found in the left hemisphere is consistent with previous
reports [
38
,
39
], indicating that the OFF response might be more important for coding the perceived
duration asymmetry than the ON response. Due to the temporal order of neural responses, there could
be more influence from the more recent OFF response relative to the ON response in the internal
duration judgment. In this process, the left hemisphere might play a more important role for
detecting duration differences whereas the right hemisphere might be more important for frequency
discrimination as previous research suggested [89].
Previous studies have shown that there is no simple relationship between duration judgment and
differences between the ON and OFF responses [
36
,
37
]. Our data provided corroborating evidence that
OFF-ON latency differences in any of the MEG measures did not perfectly match perceived duration
differences for either the S and C stimuli. This could be partly due to the attentional factor in passive
(MEG) vs. active (behavior) listening conditions.
Compared with the ramped sounds, the ON and OFF responses for the damped sounds showed
greater differences in the amount of neural activation. This activation pattern would potentially
allow listeners to separate the onset and offset of damped sounds more easily. However, we did not
observe either a robust effect of earlier OFF response for the damped sounds or conclusive evidence
for stronger sustained neural activity associated with the ramped stimuli. The latency data did not
lend direct support to either a diminished perception of damped sounds or an augmented perception
of ramped sounds. If listeners typically ignore part of the decay portion for the damped sounds as
noted in attentive listening, one would expect the OFF response latency for damped sounds to be
earlier than that for the ramped sounds. Similarly, if the ramped offset generated a small amount
of persistent sustaining activity, one would expect later OFF response for ramped sounds relative to
damped sounds. Nevertheless, it is interesting to note that the percentage of ON response latency
delay for the ramped sounds in our two stimulus sets precisely fall within the previously reported
range of subjective duration differences for ramped vs. damped sounds when the listeners were
instructed to consider all aspects of the sounds [5,7].
Brain Sci. 2016,6, 27 17 of 25
4.6. Limitations and Future Directions
As the S stimuli in our experiment used a linear amplitude envelope and portions of the C stimuli
also appeared linear in the sound waveforms, there could be a confounding factor of amplitude
acceleration rate differences at the onset and offset of the stimuli when the intensity is expressed on
the dB scale. That is, although the waveform shape suggests a linear change in intensity over the
tone duration, the velocity of the intensity-change was not constant—there is a brief, high-velocity
change at the low-level part of the ramp, and a more gradual one at the high-level part. Previous
research has shown that many factors, including intensity level (or audibility), rise time, spectral
content, and stimulus duration, jointly influence auditory ON responses. In particular, Biermann and
Heil (2000) demonstrated that unlike intensity level and rise time which systematically modulated
the auditory ON response, varying acceleration rates of the stimulus envelope at the onset did not
affect the auditory ON response in human subjects [
43
]. While rise time directly affects auditory ON
response amplitude and latency [
90
,
91
], sound level exerts greater influences than the rising speed [
92
]
and the spectral content of the stimuli is also a very important determinant [
93
]. Previous studies
suggest that the amplitude of the ON response could get smaller with its latency delayed when the
amplitude acceleration rate at the sound onset is decreased [
94
]. Thus if our S stimuli had used a linear
envelope on the dB scale, the shallower acceleration rate at the onset of the ramped sound could have
resulted in later and smaller ON responses. This hypothetical scenario would then produce even larger
on/off differences between the ramped and damped stimuli than what we reported. Since much less
is known about how amplitude accelerate rate as well as the other acoustic factors affects the auditory
OFF responses [
95
], further research is necessary to investigate whether this predicted result truly
holds with a systematic control of the envelope shape, rise time, level, and duration of the stimuli.
Due to the limited number of participants and the lack of exact models to formulate hypotheses
on the brain-behavior correlates in our study, the statistical results as reported should be interpreted
as an exploratory analysis. In particular, one needs to be cautious about issues of low statistical
power—a significant pattern with low power from a small sample of six subjects may not necessarily
extend to the population level. Previous research on the auditory ON response showed that the adult
MEG data (both MEG field measure and dipole model) were highly reliable with a small sample of five
subjects in six repeated measures [
96
]. But it remains unclear whether the OFF response is also highly
replicable with a small subject sample. Future studies need to be conducted with a larger sample
size to verify the results and test the small or weak effects which a small subject sample may not be
able to reveal.
One potential confounding factor in comparing auditory evoked responses for time-varying sounds is
the lack of balance in selective attention to the different stimuli [
54
,
97
,
98
]. Previous behavioral and imaging
studies have used an active listening task to demonstrate the selective attentional bias for sounds
with rising intensity [
15
,
16
,
99
]. Some researchers have argued for the evolutionary preparation for the
perceptual priority and alertness of rising intensity as it is an intrinsic property of an approaching sound
source [
81
,
100
]. While rising and falling intensities are acoustic patterns associated with approaching
and receding sound sources, they are not necessarily reminiscent of auditory motion perception when
the ramped and damped stimuli are as short as 200 ms as in our design. In our experimental design
using a passive listening condition with a distraction task, the MEG data primarily reflect automatic
coding of the physical differences between the stimuli (with the listener presumably taking into account
of all aspects of the physical parameters of the ramped and damped stimuli). We were interested in
testing predictions related to the two psychophysical accounts by comparing neural coding of rising
and falling intensities in auditory stimulation independent of attentional bias. Our results could be
interpreted in favour of either theory. The MEG data showed shared cortical sites for the transient ON
and OFF responses as well as distinct neural oscillations for coding the dynamic intensity envelopes in
the absence of attentional efforts.
As the current study only used stimuli of 200 ms in a passive listening condition, it remains to be
tested whether the asymmetric response patterns are generalizable to stimuli that are shorter or longer
than 200 ms and how attention, which has been shown to affect the amount of perceptual asymmetry
Brain Sci. 2016,6, 27 18 of 25
in behavioural estimation [
7
], would modulate the neural responses. As our subject sample only
included male participants, it also remains to be tested whether the auditory ON and OFF responses
would faithfully reflect the sex difference found in behavioural data with females showing a larger
effect of perceptual asymmetry [101,102].
The preliminary findings of the current exploratory study have important implications for future
studies on developmental, cross-linguistic and pathological populations. Auditory ON and OFF
responses have been shown to be potential neural markers of brain immaturity in children [
32
] and
cortical dysfunction in adults to assess their ability to extract meaning from dynamic intensity changes
in music and spoken language [
103
]. In dyslexia research, a subgroup of children was found to be
linked with a potential deficit in neural discriminatory sensitivity to envelope changes in speech
and nonspeech sounds [
104
]. In a cross-language study, language experience has been shown to
play an important role in differentially coding speech sounds in onset and offset positions [
105
].
Future clinical work and cross-language neurophysiological studies can further test the reliability
and diagnostic utility of distinct ON and OFF responses to time-varying sounds with rising versus
falling intensity and how they are affected by linguistic experience or pathological conditions. Given
that similar looming biases for rising intensity also exist in the visual modality (e.g., [
106
]) and in
multisensory integration [
4
,
99
,
107
,
108
], future studies with the target populations can also investigate
the domain-general mechanisms for the asymmetric ON and OFF responses and potential cross-modal
interactions in multisensory integration.
5. Conclusions
In sum, the present study employed MEG techniques to explore and compare the cortical
responses to ramped and damped sounds with varying spectral complexity. First, the behavioral results
replicated previous studies that ramped sounds were perceived to be longer in subjective duration
compared to damped sounds. Second and critically, a robust difference in the ON response between
the ramped and damped sounds was observed in the superior temporal and inferior parietal regions,
showing a weaker and delayed pattern compared to damped sounds. Unlike the ON responses,
the OFF responses did not show a robust effect of dominance in favor of the ramped sounds in
either cortical regions. However, consistent effects in neuronal oscillations (alpha and theta activities
in the ON response and gamma activities in the OFF response) were observed in relation to the
differences between ramped and damped sounds. Finally, a significant correlation was found between
the OFF response amplitude in left auditory cortex and behavioral temporal asymmetry for the
spectrally simpler stimulus pair. The results indicate distinct asymmetry in ON and OFF responses and
trial-by-trial neural synchronization patterns for coding the dynamic intensity changes, which interact
with spectral complexity of the auditory stimuli to influence the perceptual bias in favour of rising
intensity. These preliminary data have implications for future studies to examine how the auditory
system develops such an asymmetry as a function of age and learning experience and whether the
absence of asymmetry or abnormal ON and OFF responses can be taken as a biomarker for certain
neurological conditions associated with auditory processing deficits.
Acknowledgments:
This work was supported by funding to the Research Center for Advanced Technologies at
Tokyo Denki University from the Ministry of Education, Culture, Sports, Science and Technology of Japan. Cheng
and Zhang received support from Humanity and Social Science Foundation of Ministry of Education of China for
Young Scholars (12YJCZH022), and National Social Science Foundation of China (15BYY005). Zhang additionally
received support from a Brain Imaging Research Project Award and the Grant-in-Aid of Research, Artistry and
Scholarship Program, University of Minnesota. Cheng received additional support from China Scholarship
Council for being a visiting professor at the University of Minnesota. We thank Lotus Jo-Fu Lin for assistance and
Matti Hämäläinen and Hui Zou for technical guidance respectively on MNE analysis and statistical techniques.
Author Contributions:
Y.Z., B.C., and R.S.S. conceived and designed the experiments; Y.Z., K.T., M.K., I.N.,
and T.I. performed the experiments and collected data; Y.Z., B.C., T.K.K., K.T. and M.K. analyzed the data; Y.Z.,
B.C., and R.S.S. wrote the paper.
Conflicts of Interest: The authors declare no conflict of interest.
Brain Sci. 2016,6, 27 19 of 25
Abbreviations
The following abbreviations are used in this manuscript:
MEG: Magnetoencephalography
EEG: Electroencephalography
EOG: Electro-oculogram
ECD: Equivalent current dipole
MNE: Minimum norm estimation
dSPM: Dynamic Statistical Parametric Mapping
PLF: Phase locking factor
LH: Left hemisphere
RH: Right hemisphere
ROI: Region of interest
ST: Superior temporal
IP: Inferior parietal
RMS: Root mean square
TAI: Temporal asymmetry index
MRI: Magnetic Resonance Imaging
fMRI: Functional Magnetic Resonance Imaging
TR: Repetition time
FA: Flip angle
TE: Echo time
NEX: Number of excitations
HPI: Head position indicator
ANOVA: Analysis of Variance
FDR: False discovery rate
ISI: Inter-stimulus interval
Appendix
Brain Sci. 2016, 6, 27 19 of 24
Appendix
Figure A1. MEG data with clear ON and OFF response patterns for the steady 1 kHz reference tone.
(a) 122 channel layout and grand mean MEG waveforms from selected sensors covering the left (LH)
and right (RH) hemispheres; (b) Grand mean equivalent current dipole waveforms in LH and RH; (c)
Grand mean phase locking factor plot as a function of time; (d) Representative isofield maps (derived
with MNE-suite from one individual subject) for the ON and OFF responses; (e) Grand mean MNE
activities projected to the standard brain for ON and OFF responses corresponding to the ON and
OFF peak latencies in GFP analysis. A 10 ms integration window was used in generating the MNE
activation maps.
Figure A2. MEG responses for the S steady tone in comparison with the S ramped and S damped
stimuli. (a) Grand mean global field power data; (b) Grand mean equivalent current dipole
waveforms showing differences in left and right auditory cortices depending ON response type (ON
vs. OFF) and stimulus type (ramped, damped, steady). Repeated measures ANOVA on ECD
amplitude data showed significant main effects of response type (F(1, 5) = 72.33, p < 0.001), stimulus
type (F(2, 10) = 8.92, p < 0.05, Greenhouse-Geisser corrected), and a significant three-way interaction
between response type, stimulus type and hemisphere (F(2, 10) = 24.75, p < 0.01, Greenhouse-Geisser
corrected). Repeated measures ANOVA on ECD latency data confirmed significant main effects of
response type (ON vs. OFF) (F(1, 5) = 1207.86, p < 0.00001), stimulus type (ramped, damped, steady)
(F(2, 10) = 13.22, p < 0.05, Greenhouse-Geisser corrected), and a three-way interaction between
response type, stimulus type and hemisphere (F(2, 10) = 18.54, p < 0.01, Greenhouse-Geisser corrected).
Figure A1.
MEG data with clear ON and OFF response patterns for the steady 1 kHz reference tone.
(
a
) 122 channel layout and grand mean MEG waveforms from selected sensors covering the left (LH)
and right (RH) hemispheres; (
b
) Grand mean equivalent current dipole waveforms in LH and RH;
(
c
) Grand mean phase locking factor plot as a function of time; (
d
) Representative isofield maps (derived
with MNE-suite from one individual subject) for the ON and OFF responses; (
e
) Grand mean MNE
activities projected to the standard brain for ON and OFF responses corresponding to the ON and
OFF peak latencies in GFP analysis. A 10 ms integration window was used in generating the MNE
activation maps.
Brain Sci. 2016,6, 27 20 of 25
Brain Sci. 2016, 6, 27 19 of 24
Appendix
Figure A1. MEG data with clear ON and OFF response patterns for the steady 1 kHz reference tone.
(a) 122 channel layout and grand mean MEG waveforms from selected sensors covering the left (LH)
and right (RH) hemispheres; (b) Grand mean equivalent current dipole waveforms in LH and RH; (c)
Grand mean phase locking factor plot as a function of time; (d) Representative isofield maps (derived
with MNE-suite from one individual subject) for the ON and OFF responses; (e) Grand mean MNE
activities projected to the standard brain for ON and OFF responses corresponding to the ON and
OFF peak latencies in GFP analysis. A 10 ms integration window was used in generating the MNE
activation maps.
Figure A2. MEG responses for the S steady tone in comparison with the S ramped and S damped
stimuli. (a) Grand mean global field power data; (b) Grand mean equivalent current dipole
waveforms showing differences in left and right auditory cortices depending ON response type (ON
vs. OFF) and stimulus type (ramped, damped, steady). Repeated measures ANOVA on ECD
amplitude data showed significant main effects of response type (F(1, 5) = 72.33, p < 0.001), stimulus
type (F(2, 10) = 8.92, p < 0.05, Greenhouse-Geisser corrected), and a significant three-way interaction
between response type, stimulus type and hemisphere (F(2, 10) = 24.75, p < 0.01, Greenhouse-Geisser
corrected). Repeated measures ANOVA on ECD latency data confirmed significant main effects of
response type (ON vs. OFF) (F(1, 5) = 1207.86, p < 0.00001), stimulus type (ramped, damped, steady)
(F(2, 10) = 13.22, p < 0.05, Greenhouse-Geisser corrected), and a three-way interaction between
response type, stimulus type and hemisphere (F(2, 10) = 18.54, p < 0.01, Greenhouse-Geisser corrected).
Figure A2.
MEG responses for the S steady tone in comparison with the S ramped and S damped
stimuli. (
a
) Grand mean global field power data; (
b
) Grand mean equivalent current dipole waveforms
showing differences in left and right auditory cortices depending ON response type (ON vs. OFF) and
stimulus type (ramped, damped, steady). Repeated measures ANOVA on ECD amplitude data showed
significant main effects of response type (F(1, 5) = 72.33, p< 0.001), stimulus type (F(2, 10) = 8.92,
p< 0.05, Greenhouse-Geisser corrected), and a significant three-way interaction between response type,
stimulus type and hemisphere (F(2, 10) = 24.75, p< 0.01, Greenhouse-Geisser corrected). Repeated
measures ANOVA on ECD latency data confirmed significant main effects of response type (ON vs.
OFF) (F(1, 5) = 1207.86, p< 0.00001), stimulus type (ramped, damped, steady) (F(2, 10) = 13.22, p< 0.05,
Greenhouse-Geisser corrected), and a three-way interaction between response type, stimulus type and
hemisphere (F(2, 10) = 18.54, p< 0.01, Greenhouse-Geisser corrected).
References
1.
Akeroyd, M.A.; Patterson, R.D. Discrimination of wideband noises modulated by a temporally asymmetric
function. J. Acoust. Soc. Am. 1995,98, 2466–2474. [CrossRef]
2.
Akeroyd, M.A.; Patterson, R.D. A comparison of detection and discrimination of temporal asymmetry in
amplitude modulation. J. Acoust. Soc. Am. 1997,101, 430–439. [CrossRef] [PubMed]
3.
Schlauch, R.S.; Ries, D.T.; DiGiovanni, J.J. Duration discrimination and subjective duration for ramped and
damped sounds. J. Acoust. Soc. Am. 2001,109, 2880–2887. [CrossRef] [PubMed]
4.
Van Wassenhove, V.; Buonomano, D.V.; Shimojo, S.; Shams, L. Distortions of subjective time perception
within and across senses. PLoS ONE 2008,3, e1437. [CrossRef] [PubMed]
5.
Ries, D.T.; Schlauch, R.S.; DiGiovanni, J.J. The role of temporal-masking patterns in the determination of
subjective duration and loudness for ramped and damped sounds. J. Acoust. Soc. Am.
2008
,124, 3772–3783.
[CrossRef] [PubMed]
6.
Grassi, M.; Darwin, C.J. The subjective duration of ramped and damped sounds. Percept. Psychophys.
2006
,
68, 1382–1392. [CrossRef] [PubMed]
7.
DiGiovanni, J.J.; Schlauch, R.S. Mechanisms responsible for differences in perceived duration for
rising-intensity and falling-intensity sounds. Ecol. Psychol. 2007,19, 239–264. [CrossRef]
8.
Stecker, G.C.; Hafter, E.R. An effect of temporal asymmetry on loudness. J. Acoust. Soc. Am.
2000
,107, 3358–3368.
[CrossRef] [PubMed]
9.
Canevet, G.; Scharf, B. The loudness of sounds that increase and decrease continuously in level.
J. Acoust. Soc. Am. 1990,88, 2136–2142. [CrossRef] [PubMed]
10. Neuhoff, J.G. Perceptual bias for rising tones. Nature 1998,395, 123–124. [CrossRef] [PubMed]
Brain Sci. 2016,6, 27 21 of 25
11.
Teghtsoonian, R.; Teghtsoonian, M.; Canevet, G. Sweep-induced acceleration in loudness change and the
“bias for rising intensities”. Percept. Psychophys. 2005,67, 699–712. [CrossRef] [PubMed]
12.
Olsen, K.N.; Stevens, C.J.; Tardieu, J. Loudness change in response to dynamic acoustic intensity.
J. Exp. Psychol. Hum. Percept. Perform. 2010,36, 1631–1644. [CrossRef] [PubMed]
13.
Irino, T.; Patterson, R.D. Temporal asymmetry in the auditory system. J. Acoust. Soc. Am.
1996
,99, 2316–2331.
[CrossRef] [PubMed]
14.
Neuhoff, J.G. An adaptive bias in the perception of looming auditory motion. Ecol. Psychol.
2001
,13, 87–110.
[CrossRef]
15.
Rosenblum, L.D.; Wuestefeld, A.P.; Saldana, H.M. Auditory looming perception: Influences on anticipatory
judgments. Perception 1993,22, 1467–1482. [CrossRef] [PubMed]
16.
Bach, D.R.; Schachinger, H.; Neuhoff, J.G.; Esposito, F.; Di Salle, F.; Lehmann, C.; Herdener, M.; Scheffler, K.;
Seifritz, E. Rising sound intensity: An intrinsic warning cue activating the amygdala. Cereb. Cortex
2008
,
18, 145–150. [CrossRef] [PubMed]
17.
Lu, T.; Liang, L.; Wang, X.Q. Neural representations of temporally asymmetric stimuli in the auditory cortex
of awake primates. J. Neurophysiol. 2001,85, 2364–2380. [PubMed]
18.
Wang, X.; Lu, T.; Snider, R.K.; Liang, L. Sustained firing in auditory cortex evoked by preferred stimuli.
Nature 2005,435, 341–346. [CrossRef] [PubMed]
19.
Qin, L.; Chimoto, S.; Sakai, M.; Wang, J.; Sato, Y. Comparison between offset and onset responses of primary
auditory cortex on-off neurons in awake aats. J. Neurophysiol. 2007,97, 3421–3431. [CrossRef] [PubMed]
20.
Wang, J.; Qin, L.; Chimoto, S.; Tazunoki, S.; Sato, Y. Response characteristics of primary auditory cortex
neurons underlying perceptual asymmetry of ramped and damped sounds. Neuroscience
2014
,256, 309–321.
[CrossRef] [PubMed]
21.
Maier, J.X.; Ghazanfar, A.A. Looming biases in monkey auditory cortex. J. Neurosci.
2007
,27, 4093–4100.
[CrossRef] [PubMed]
22.
Toronchuk, J.M.; Stumpf, E.; Cynader, M.S. Auditory cortex neurons sensitive to correlates of auditory
motion: Underlying mechanisms. Exp. Brain Res. 1992,88, 169–180. [CrossRef] [PubMed]
23.
Kuwada, S.; Batra, R. Coding of sound envelopes by inhibitory rebound in neurons of the superior olivary
complex in the unanesthetized rabbit. J. Neurosci. 1999,19, 2273–2287. [PubMed]
24.
Patterson, R.D.; Allerhand, M.H.; Giguere, C. Time-domain modeling of peripheral auditory processing:
A modular architecture and a software platform. J. Acoust. Soc. Am. 1995,98, 1890–1894. [CrossRef] [PubMed]
25.
Phillips, D.P.; Hall, S.E.; Boehnke, S.E. Central auditory onset responses, and temporal asymmetries in
auditory perception. Hear. Res. 2002,167, 192–205. [CrossRef]
26.
Pantev, C.; Eulitz, C.; Hampson, S.; Ross, B.; Roberts, L.E. The auditory evoked “off” response: Sources and
comparison with the “on” and the “sustained” responses. Ear Hear.
1996
,17, 255–265. [CrossRef] [PubMed]
27.
Hari, R.; Pelizzone, M.; Makela, J.P.; Hallstrom, J.; Leinonen, L.; Lounasmaa, O.V. Neuromagnetic responses
of the human auditory cortex to on- and offsets of noise bursts. Audiology
1987
,26, 31–43. [CrossRef]
[PubMed]
28.
Yamashiro, K.; Inui, K.; Otsuru, N.; Kida, T.; Kakigi, R. Automatic auditory off-response in humans: An meg
study. Eur. J. Neurosci. 2009,30, 125–131. [CrossRef] [PubMed]
29.
Nishihara, M.; Inui, K.; Motomura, E.; Otsuru, N.; Ushida, T.; Kakigi, R. Auditory N1 as a change-related
automatic response. Neurosci. Res. 2011,71, 145–148. [CrossRef] [PubMed]
30.
Näätänen, R.; Picton, T. The N1 wave of the human electric and magnetic response to sound—A review and
an analysis of the component structure. Psychophysiology 1987,24, 375–425. [CrossRef] [PubMed]
31.
Hillyard, S.A.; Picton, T.W. On and off components in the auditory evoked potential. Percept. Psychophys.
1978,24, 391–398. [CrossRef] [PubMed]
32.
Wakai, R.T.; Lutter, W.J.; Chen, M.; Maier, M.M. On and off magnetic auditory evoked responses in early
infancy: A possible marker of brain immaturity. Clin. Neurophysiol.
2007
,118, 1480–1487. [CrossRef]
[PubMed]
33.
Yamashiro, K.; Inui, K.; Otsuru, N.; Kakigi, R. Change-related responses in the human auditory cortex:
An meg study. Psychophysiology 2011,48, 23–30. [CrossRef] [PubMed]
34.
Hari, R.; Sams, M.; Järvilehto, T. Auditory evoked transient and sustained potentials in the human eeg:
I. Effects of expectation of stimuli. Psychiatry Res. 1979,1, 297–306. [CrossRef]
Brain Sci. 2016,6, 27 22 of 25
35.
Noda, K.; Tonoike, M.; Doi, K.; Koizuka, I.; Yamaguchi, M.; Seo, R.; Matsumoto, N.; Noiri, T.; Takeda, N.;
Kubo, T. Auditory evoked off-response: Its source distribution is different from that of on-response.
Neuroreport 1998,9, 2621–2625. [CrossRef] [PubMed]
36.
Bendixen, A.; Grimm, S.; Schroger, E. The relation between onset, offset, and duration perception as examined
by psychophysical data and event-related brain potentials. J. Psychophysiol. 2006,20, 40–51. [CrossRef]
37.
Sanders, L.D.; Astheimer, L.B. Temporally selective attention modulates early perceptual processing:
Event-related potential evidence. Percept. Psychophys. 2008,70, 732–742. [CrossRef] [PubMed]
38.
Qin, L.; Liu, Y.C.; Wang, J.Y.; Li, S.N.; Sato, Y. Neural and behavioral discrimination of sound duration
by cats. J. Neurosci. 2009,29, 15650–15659. [CrossRef] [PubMed]
39.
Liegeois-Chauvel, C.; de Graaf, J.B.; Laguitton, V.; Chauvel, P. Specialization of left auditory cortex for speech
perception in man depends on temporal coding. Cereb. Cortex 1999,9, 484–496. [CrossRef] [PubMed]
40.
Kult, A. An Integrated Magnetoencephalographic and Functional Magnetic Resonance Imaging Study on Temporal
Asymmetry Processing in the Human Auditory Cortex; The Ruperto-Carola University of Heidelberg: Heidelberg,
Germany, 2006.
41.
Tabas, A.; Balaguer-Ballester, E.; Pressnitzer, D.; Siebert, A.; Rupp, A. Hierarchical processing of temporal
asymmetry in human auditory cortex. Int. Workshop Pattern 2014, 1–4. [CrossRef]
42.
Tabas, A.; Siebert, A.; Supek, S.; Pressnitzer, D.; Balaguer-Ballester, E.; Rupp, A. Insights on the neuromagnetic
representation of temporal asymmetry in human auditory cortex. PLoS ONE
2016
,11, e0153947. [CrossRef]
[PubMed]
43.
Biermann, S.; Heil, P. Parallels between timing of onset responses of single neurons in cat and of evoked
magnetic fields in human auditory cortex. J. Neurophysiol. 2000,84, 2426–2439. [PubMed]
44.
Seither-Preisler, A.; Patterson, R.; Krumbholz, K.; Seither, S.; Lutkenhoner, B. Evidence of pitch processing in
the N100m component of the auditory evoked field. Hear. Res. 2006,213, 88–98. [CrossRef] [PubMed]
45.
Kaiser, J.; Ripper, B.; Birbaumer, N.; Lutzenberger, W. Dynamics of gamma-band activity in human
magnetoencephalogram during auditory pattern working memory. Neuroimage
2003
,20, 816–827. [CrossRef]
46.
Pantev, C.; Makeig, S.; Hoke, M.; Galambos, R.; Hampson, S.; Gallen, C. Human auditory evoked
gamma-band magnetic-fields. Proc. Natl. Acad. Sci. USA 1991,88, 8996–9000. [CrossRef] [PubMed]
47.
Jensen, O.; Gelfand, J.; Kounios, J.; Lisman, J. Oscillations in the alpha band (9–12 Hz) increase with memory
load during retention in a short-term memory task. Cereb. Cortex 2002,12, 877–882. [CrossRef] [PubMed]
48.
Ahissar, E.; Nagarajan, S.; Ahissar, M.; Protopapas, A.; Mahncke, H.; Merzenich, M.M. Speech comprehension
is correlated with temporal response patterns recorded from auditory cortex. Proc. Natl. Acad. Sci. USA
2001
,
98, 13367–13372. [CrossRef] [PubMed]
49.
Luo, H.; Poeppel, D. Phase patterns of neuronal responses reliably discriminate speech in human auditory
cortex. Neuron 2007,54, 1001–1010. [CrossRef] [PubMed]
50.
Shahin, A.J.; Trainor, L.J.; Roberts, L.E.; Backer, K.C.; Miller, L.M. Development of auditory phase-locked
activity for music sounds. J. Neurophysiol. 2010,103, 218–229. [CrossRef] [PubMed]
51.
Koerner, T.K.; Zhang, Y. Effects of background noise on inter-trial phase coherence and auditory N1–P2
responses to speech stimuli. Hear. Res. 2015,328, 113–119. [CrossRef] [PubMed]
52.
Zhang, Y.; Kuhl, P.K.; Imada, T.; Iverson, P.; Pruitt, J.; Stevens, E.; Kawakatsu, M.; Tohkura, Y.; Nemoto, I.
Neural signatures of phonetic learning in adulthood: A magnetoencephalography study. Neuroimage
2009
,
46, 226–240. [CrossRef] [PubMed]
53.
Zhang, Y.; Kuhl, P.K.; Imada, T.; Kotani, M.; Tohkura, Y. Effects of language experience: Neural commitment
to language-specific auditory patterns. Neuroimage 2005,26, 703–720. [CrossRef] [PubMed]
54.
Rao, A.; Zhang, Y.; Miller, S. Selective listening of concurrent auditory stimuli: An event-related potential
study. Hear. Res. 2010,268, 123–132. [CrossRef] [PubMed]
55.
Nie, Y.; Zhang, Y.; Nelson, P.B. Auditory stream segregation using bandpass noises: Evidence from
event-related potentials. Front. Neurosci. 2014,8, 277. [CrossRef] [PubMed]
56.
Obleser, J.; Scott, S.K.; Eulitz, C. Now you hear it, now you don’t: Transient traces of consonants and their
nonspeech analogues in the human brain. Cereb. Cortex 2006,16, 1069–1076. [CrossRef] [PubMed]
57.
Tervaniemi, M.; Rytkonen, M.; Schroger, E.; Ilmoniemi, R.J.; Näätänen, R. Superior formation of cortical
memory traces for melodic patterns in musicians. Learn. Mem. 2001,8, 295–300. [CrossRef] [PubMed]
58. Zhang, Y.; Koerner, T.; Miller, S.; Grice-Patil, Z.; Svec, A.; Akbari, D.; Tusler, L.; Carney, E. Neural coding of
formant-exaggerated speech in the infant brain. Dev. Sci. 2011,14, 566–581. [CrossRef] [PubMed]
Brain Sci. 2016,6, 27 23 of 25
59.
Zhang, Y.; Zhang, W.; Reynoso Alcántara, V.; Silva-Pereyra, J. Magnetoencephalography: Mapping the
spatiotemporal dynamics of neuronal activity. Suma Psicol. 2014,21, 45–53. [CrossRef]
60.
Lehmann, D.; Skrandies, W. Spatial analysis of evoked potentials in man—A review. Prog. Neurobiol.
1984
,
23, 227–250. [CrossRef]
61.
Yamasue, H.; Yamada, H.; Yumoto, M.; Kamio, S.; Kudo, N.; Uetsuki, M.; Abe, O.; Fukuda, R.; Aoki, S.;
Ohtomo, K.; et al. Abnormal association between reduced magnetic mismatch field to speech sounds and
smaller left planum temporale volume in schizophrenia. Neuroimage
2004
,22, 720–727. [CrossRef] [PubMed]
62.
Carbonell, F.; Galan, L.; Valdes, P.; Worsley, K.; Biscay, R.J.; Diaz-Comas, L.; Bobes, M.A.; Parra, M. Random
field-union intersection tests for EEG/MEG imaging. Neuroimage 2004,22, 268–276. [CrossRef] [PubMed]
63.
Hämäläinen, M.; Hari, R.; Ilmoniemi, R.J.; Knuutila, J.; Lounasmaa, O.V. Magnetoencephalography: Theory,
instrumentation, and applications to noninvasive studies of the working human brain. Rev. Mod. Phys.
1993
,
65, 413–497. [CrossRef]
64.
Lin, F.H.; Witzel, T.; Ahlfors, S.P.; Stufflebeam, S.M.; Belliveau, J.W.; Hämäläinen, M.S. Assessing and
improving the spatial accuracy in MEG source localization by depth-weighted minimum-norm estimates.
Neuroimage 2006,31, 160–171. [CrossRef] [PubMed]
65.
Dale, A.M.; Liu, A.K.; Fischl, B.R.; Buckner, R.L.; Belliveau, J.W.; Lewine, J.D.; Halgren, E. Dynamic statistical
parametric mapping: Combining fMRI and MEG for high-resolution imaging of cortical activity. Neuron
2000,26, 55–67. [CrossRef]
66.
Giard, M.H.; Perrin, F.; Echallier, J.F.; Thevenet, M.; Froment, J.C.; Pernier, J. Dissociation of temporal and
frontal components in the human auditory N1 wave—A scalp current-density and dipole model analysis.
Electroen. Clin. Neuro. 1994,92, 238–252. [CrossRef]
67.
Godey, B.; Schwartz, D.; de Graaf, J.B.; Chauvel, P.; Liegeois-Chauvel, C. Neuromagnetic source localization
of auditory evoked fields and intracerebral evoked potentials: A comparison of data in the same patients.
Clin. Neurophysiol. 2001,112, 1850–1859. [CrossRef]
68.
Zouridakis, G.; Simos, P.G.; Papanicolaou, A.C. Multiple bilaterally asymmetric cortical sources account for
the auditory N1m component. Brain Topogr. 1998,10, 183–189. [CrossRef] [PubMed]
69.
Sinai, A.; Crone, N.E.; Wied, H.M.; Franaszczuk, P.J.; Miglioretti, D.; Boatman-Reich, D. Intracranial mapping
of auditory perception: Event-related responses and electrocortical stimulation. Clin. Neurophysiol. Off. J. Int.
Fed. Clin. Neurophysiol. 2009,120, 140–149. [CrossRef] [PubMed]
70.
Jensen, O.; Hesse, C. Estimating distributed representation of evoked responses and oscillatory brain activity.
In MEG: An Introduction to Methods; Hansen, P., Kringelbach, M., Salmelin, R., Eds.; Oxford University Press:
New York, USA, 2010; pp. 156–185.
71.
Lachaux, J.P.; Rodriguez, E.; Martinerie, J.; Varela, F.J. Measuring phase synchrony in brain signals.
Hum. Brain Mapp. 1999,8, 194–208. [CrossRef]
72.
Tallon-Baudry, C.; Bertrand, O.; Delpuech, C.; Pernier, J. Stimulus specificity of phase-locked and
non-phase-locked 40 Hz visual responses in humans. J. Neurosci. 1996,16, 4240–4249. [PubMed]
73.
Macmillan, N.A.; Creelman, C.D. Detection Theory: A User’s Guide; Cambridge University Press: Cambridge,
UK, 1991.
74.
Stoppel, C.M.; Boehler, C.N.; Strumpf, H.; Krebs, R.M.; Heinze, H.J.; Hopf, J.M.; Schoenfeld, M.A.
Spatiotemporal dynamics of feature-based attention spread: Evidence from combined electroencephalographic
and magnetoencephalographic recordings. J. Neurosci. 2012,32, 9671–9676. [CrossRef] [PubMed]
75.
Guthrie, D.; Buchwald, J.S. Significance testing of difference potentials. Psychophysiology
1991
,28, 240–244.
[CrossRef] [PubMed]
76.
Seghier, M.L. Laterality index in functional MRI: Methodological issues. Magn. Reson. Imaging
2008
,
26, 594–601. [CrossRef] [PubMed]
77.
Rousselet, G.A.; Pernet, C.R. Improving standards in brain-behavior correlation analyses. Front. Hum. Neurosci.
2012,6, 119. [CrossRef] [PubMed]
78.
Pernet, C.R.; Wilcox, R.R.; Rousselet, G.A. Robust correlation analyses: False positive and power validation
using a new open source matlab toolbox. Front. Psychol. 2013,3, 606. [CrossRef] [PubMed]
79.
Benjamini, Y.; Hochberg, Y. Controlling the false discovery rate—A practical and powerful approach to
multiple testing. J. R. Stat. Soc. B 1995,57, 289–300.
80.
Tiitinen, H.; Makinen, V.T.; Kicic, D.; May, P.J.C. Averaged and single-trial brain responses in the assessment
of human sound detection. Neuroreport 2005,16, 545–548. [CrossRef] [PubMed]
Brain Sci. 2016,6, 27 24 of 25
81.
Seifritz, E.; Neuhoff, J.G.; Bilecen, D.; Scheffler, K.; Mustovic, H.; Schächinger, H.; Elefante, R.; Di Salle, F.
Neural processing of auditory looming in the human brain. Curr. Biol. 2002,12, 2147–2151. [CrossRef]
82.
He, J. On and off pathways segregated at the auditory thalamus of the guinea pig. J. Neurosci.
2001
,
21, 8672–8679. [PubMed]
83.
He, J. Corticofugal modulation on both on and off responses in the nonlemniscal auditory thalamus of the
guinea pig. J. Neurophysiol. 2003,89, 367–381. [CrossRef] [PubMed]
84.
Klimesch, W. Alpha-band oscillations, attention, and controlled access to stored information. Trends Cognit. Sci.
2012,16, 606–617. [CrossRef] [PubMed]
85.
Scholl, B.; Gao, X.; Wehr, M. Nonoverlapping sets of synapses drive on responses and off responses in
auditory cortex. Neuron 2010,65, 412–421. [CrossRef] [PubMed]
86.
Swettenham, J.B.; Muthukumaraswamy, S.D.; Singh, K.D. Spectral properties of induced and evoked gamma
oscillations in human early visual cortex to moving and stationary stimuli. J. Neurophysiol.
2009
,102,
1241–1253. [CrossRef] [PubMed]
87.
Fujioka, T.; Trainor, L.J.; Large, E.W.; Ross, B. Beta and gamma rhythms in human auditory cortex during
musical beat processing. Ann. N. Y. Acad. Sci. 2009,1169, 89–92. [CrossRef] [PubMed]
88.
Stefanics, G.; Hangya, B.; Hernadi, I.; Winkler, I.; Lakatos, P.; Ulbert, I. Phase entrainment of human delta
oscillations can mediate the effects of expectation on reaction speed. J. Neurosci.
2010
,30, 13578–13585.
[CrossRef] [PubMed]
89.
Molholm, S.; Martinez, A.; Ritter, W.; Javitt, D.C.; Foxe, J.J. The neural circuitry of pre-attentive auditory
change-detection: An fMRI study of pitch and duration mismatch negativity generators. Cereb. Cortex
2005
,
15, 545–551. [CrossRef] [PubMed]
90.
Kodera, K.; Hink, R.F.; Yamada, O.; Suzuki, J.I. Effects of rise time on simultaneously recorded
auditory-evoked potentials from the early, middle and late ranges. Audiology
1979
,18, 395–402. [CrossRef]
[PubMed]
91.
Thomson, J.M.; Goswami, U.; Badeweg, T. The ERP signature of sound rise time changes. Brain Res.
2009
,
1254, 74–83. [CrossRef] [PubMed]
92.
Mantani, M.; Nakagawa, S.; Kotani, K.; Horii, K. Effects of rise-time of sound stimulus on auditory evoked
magnetic fields. Trans. Jpn. Soc. Med. Biol. Eng. 2007,45, 114–120.
93.
Rupp, A.; Spachmann, A.; Dettlaff, A.; Patterson, R.D. Cortical activity associated with the perception of
temporal asymmetry in ramped and damped noises. In Basic Aspects of Hearing. Advances in Experimental
Medicine and Biology; Moore, B.C.J., Patterson, R.D., Winter, I.M., Carlyon, R.P., Gockel, H.E., Eds.; Springer:
New York, NY, USA, 2013; Volume 787, pp. 427–433.
94.
Escabi, M.A.; Miller, L.M.; Read, H.L.; Schreiner, C.E. Naturalistic auditory contrast improves spectrotemporal
coding in the cat inferior colliculus. J. Neurosci. 2003,23, 11489–11504. [PubMed]
95.
Baltzell, L.S.; Billings, C.J. Sensitivity of offset and onset cortical auditory evoked potentials to signals in
noise. Clin. Neurophysiol. Off. J. Int. Fed. Clin. Neurophysiol. 2014,125, 370–380. [CrossRef] [PubMed]
96.
Virtanen, J.; Ahveninen, J.; Ilmoniemi, R.J.; Naatanen, R.; Pekkonen, E. Replicability of MEG and EEG
measures of the auditory N1/N1m-response. Electroencephalogr. Clin. Neurophysiol.
1998
,108, 291–298.
[CrossRef]
97.
Näätänen, R. Selective attention and evoked potentials in humans—A critical review. Biol. Psychol.
1975
,
2, 237–307. [CrossRef]
98.
Alho, K. Selective attention in auditory processing as reflected by event-related brain potentials.
Psychophysiology 1992,29, 247–263. [CrossRef] [PubMed]
99.
Maier, J.X.; Chandrasekaran, C.; Ghazanfar, A.A. Integration of bimodal looming signals through neuronal
coherence in the temporal lobe. Curr. Biol. 2008,18, 963–968. [CrossRef] [PubMed]
100.
Ghazanfar, A.A.; Neuhoff, J.G.; Logothetis, N.K. Auditory looming perception in rhesus monkeys. Proc. Natl.
Acad. Sci. USA 2002,99, 15755–15757. [CrossRef] [PubMed]
101.
Grassi, M. Sex difference in subjective duration of looming and receding sounds. Perception
2010
,39,
1424–1426. [CrossRef] [PubMed]
102.
Neuhoff, J.G.; Planisek, R.; Seifritz, E. Adaptive sex differences in auditory motion perception: Looming
sounds are special. J. Exp. Psychol. Hum. 2009,35, 225–234. [CrossRef] [PubMed]
103.
Bach, D.R.; Buxtorf, K.; Strik, W.K.; Neuhoff, J.G.; Seifritz, E. Evidence for impaired sound intensity processing
in schizophrenia. Schizophr. Bull. 2011,37, 426–431. [CrossRef] [PubMed]
Brain Sci. 2016,6, 27 25 of 25
104.
Hämäläinen, J.A.; Leppanen, P.H.T.; Guttorm, T.K.; Lyytinen, H. Event-related potentials to pitch and rise
time change in children with reading disabilities and typically reading children. Clin. Neurophysiol.
2008
,
119, 100–115. [CrossRef] [PubMed]
105.
Cheng, B.; Zhang, Y. Syllable structure universals and native language interference in second language
perception and production: Positional asymmetry and perceptual links to accentedness. Front. Psychol.
2015
,
6, 1801. [CrossRef] [PubMed]
106.
Schiff, W.; Caviness, J.A.; Gibson, J.J. Persistent fear responses in rhesus monkeys to the optical stimulus of
“looming”. Science 1962,136, 982–983. [CrossRef] [PubMed]
107.
Cappe, C.; Thut, G.; Romei, V.; Murray, M.M. Selective integration of auditory-visual looming cues by
humans. Neuropsychologia 2009,47, 1045–1052. [CrossRef] [PubMed]
108.
Romei, V.; Murray, M.M.; Cappe, C.; Thut, G. Preperceptual and stimulus-selective enhancement of low-level
human visual cortex excitability by sounds. Curr. Biol. 2009,19, 1799–1805. [CrossRef] [PubMed]
©
2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).
... The model further postulates that the auditory cortex recognizes sound inputs by comparing it to the input of the previous instant, and thus acts as a detector of input changes. In this model, the auditory N1, a prominent electromagnetic response of the auditory cortex that is elicited ∼100 ms after the onset and offset of a discrete tone or after an alteration of a continuous tone (Zhang et al., 2016), serves as a marker of this change detection process within auditory cortex. Indeed, recent studies have revealed that the auditory N1 detects change by comparing the information of a preceding stimulus with that of a subsequent stimulus (Inui et al., 2010). ...
... Auditory N1 as a Change Detector As stated above, the auditory N1 is a prominent cortical electroencephalographic response to both the onset (Onresponse; On-N1) and offset (Off-response; Off-N1) of an auditory stimulus (Zhang et al., 2016). An auditory N1 can also be elicited by infrequent changes in pitch or timbre of a continuous complex tone (Vaz Pato and Jones, 1999;Change-N1). ...
Article
Full-text available
A growing body of research is focused on identifying and understanding the neurophysiological mechanisms that underlie tinnitus. Unfortunately, however, most current models cannot adequately explain the majority of tinnitus features. For instance, although tinnitus generally appears within minutes after entering a silent environment, most models postulate that tinnitus emerges over a much larger timescale (days). Similarly, there is a limited understanding of how the severity of tinnitus can differ in patients with a similar degree of hearing loss. To address this critical knowledge gap, we have formulated a novel explanatory model of tinnitus, the perception-update (PU) model, which rests on a theory of information processing and can explain several key characteristics of tinnitus onset. The PU model posits that the brain continuously updates the information received from the inner ear by comparing it to the received information immediately before. That is, the auditory system processes the relative change in sensory input, as opposed to the absolute value of the auditory input. This is analogous to the functioning of data compression technology used for music and images called differential pulse code modulation (differential PCM). The PU model proposes that the inner ear transmits sound change to the auditory cortex via an auditory N1 response, an event-related potential component that constitutes is a prime signaler of auditory input change. In cases of hearing impairment, the PU model posits that the auditory system finds itself in a state of uncertainty where perception has to be predicted based on previous stimulation parameters, which can lead to the emergence of tinnitus.
... This procedure was adapted from Schlauch et al. (2001). A similar magnitude estimation procedure was used in a recent study in our lab (Zhang et al., 2016b). Specifically, participants were presented with a sound and a bar on the computer screen for each trial, and they were instructed to adjust the length of the bar to match the duration of the sound. ...
... In the current study, both the rising and dipping tones included dynamic pitch contours, and the rising tone had a higher average F 0 than the dipping tone, which may partially contribute to the perceived duration difference between the two tones. Furthermore, there is a well-established perceptual looming bias with sounds that have rising pitch contour or intensity perceived as longer and louder than sounds with falling pitch contour or intensity (Ghazanfar and Maier, 2009;Zhang et al., 2016b). Our data indicate that the perceptual lengthening of rising tones does not appear to be influenced by language experience. ...
Article
A current topic in auditory neurophysiology is how brainstem sensory coding contributes to higher-level perceptual, linguistic and cognitive skills. This cross-language study was designed to compare frequency following responses (FFRs) for lexical tones in tonal (Mandarin Chinese) and non-tonal (English) language users and test the correlational strength between FFRs and behavior as a function of language experience. The behavioral measures were obtained in the Garner paradigm to assess how lexical tones might interfere with vowel category and duration judgement. The FFR results replicated previous findings about between-group differences, showing enhanced pitch tracking responses in the Chinese subjects. The behavioral data from the two subject groups showed that lexical tone variation in the vowel stimuli significantly interfered with vowel identification with a greater effect in the Chinese group. Moreover, the FFRs for lexical tone contours were significantly correlated with the behavioral interference only in the Chinese group. This pattern of language-specific association between speech perception and brainstem-level neural phase-locking of linguistic pitch information provides evidence for a possible native language neural commitment at the subcortical level, highlighting the role of experience-dependent brainstem tuning in influencing subsequent linguistic processing in the adult brain.
... Audition is known to mark more strongly the onset than the offset of auditory events (Philips, Hall, & Boehnke, 2002). In addition, the strength of the mark is a function of the rate of change in intensity of the onset and offset ramp (Phillips et al., 2002; see also Deneux et al., 2016;Lu, Liang, & Wang, 2001;Zhang, et al., 2016). We may observe a subjective counterpart of this difference in loudness perception: the beginning of a sound is the portion of the sound that contributes most to the overall loudness of the sound (Dittrich & Oberfeld, 2009). ...
Article
Hearing is the most accurate sense for perceiving duration. However, rarely it produces inaccurate estimates of duration, for example when it compares the subjective duration of tones that are increasing in intensity over time (i.e., ramped) with that of tones that are decreasing in intensity over time (i.e., damped). The literature reports that the damped tones are perceived as much being shorter than the ramped tones of the same length. The short subjective duration of damped tones may originate from a decay suppression mechanism that parses the source-informative part of many natural sounds (i.e., the beginning) from the less informative part of them (the decay): listeners may interpret the tail of damped tones like an echo or like the decay portion of an impact sound and exclude it from the account of the duration of the tone. In the natural soundscape, the tail of sounds produced in reverberant environments and the tail of impact sounds have a frequency content that is constant throughout the sound’s duration. Here, the carriers used for ramped and damped sounds were a tone constant in frequency and a tone modulated in frequency. The frequency modulation was introduced to prevent the listener from interpreting the tail of these tones as the result of reverberation or the decay portion of an impact sound. Frequency constant damped tones were largely underestimated in duration whereas frequency modulated ones were not (or were only slightly), demonstrating that the decay suppression mechanism is a worthy explanation for the short subjective duration of damped tones.
... While the two ERP studies reported here are clearly limited in scope and depth of analysis, the side-by-side comparisons clearly demonstrate the limitations and inappropriateness of the Pearson approach as well as its inflated correlation estimation results for the data sets. Given that multiple analysis techniques (for example, waveform analysis, source localization, time-frequency analysis) can be applied to the same neurophysiological data in cognitive neuroscience research (Koerner, Zhang, Nelson, Wang, & Zou, 2016;Koerner & Zhang, 2015;Zhang et al., 2011Zhang et al., , 2016, a cautionary note against the convenient use of the simple Pearson correlation test is necessary when selecting and applying statistical models to interpret brain-behavior correlations (e.g., biomarkers of various diseases and disorders) or correlations among the various brain measures with prior distributions and covariance structure for repeated measures. ...
... While the two ERP studies reported here are clearly limited in scope and depth of analysis, the side-by-side comparisons clearly demonstrate the limitations and inappropriateness of the Pearson approach as well as its inflated correlation estimation results for the data sets. Given that multiple analysis techniques (for example, waveform analysis, source localization, time-frequency analysis) can be applied to the same neurophysiological data in cognitive neuroscience research [54,[57][58][59], a cautionary note against the convenient use of the simple Pearson correlation test is necessary when selecting and applying statistical models to interpret brain-behavior correlations (e.g., biomarkers of various diseases and disorders) or correlations among the various brain measures with prior distributions and covariance structure for repeated measures. ...
Article
Full-text available
Neurophysiological studies are often designed to examine relationships between measures from different testing conditions, time points, or analysis techniques within the same group of participants. Appropriate statistical techniques that can take into account repeated measures and multivariate predictor variables are integral and essential to successful data analysis and interpretation. This work implements and compares conventional Pearson correlations and linear mixed-effects (LME) regression models using data from two recently published auditory electrophysiology studies. For the specific research questions in both studies, the Pearson correlation test is inappropriate for determining strengths between the behavioral responses for speech-in-noise recognition and the multiple neurophysiological measures as the neural responses across listening conditions were simply treated as independent measures. In contrast, the LME models allow a systematic approach to incorporate both fixed-effect and random-effect terms to deal with the categorical grouping factor of listening conditions, between-subject baseline differences in the multiple measures, and the correlational structure among the predictor variables. Together, the comparative data demonstrate the advantages as well as the necessity to apply mixed-effects models to properly account for the built-in relationships among the multiple predictor variables, which has important implications for proper statistical modeling and interpretation of human behavior in terms of neural correlates and biomarkers.
Article
GOUREVITCH, B., C. Martin, O. Postal, J.J. Eggermont. Oscillations in the auditory system, their possible role. NEUROSCI BIOBEHAV REV XXX XXX-XXX, 2020. - Neural oscillations are thought to have various roles in brain processing such as, attention modulation, neuronal communication, motor coordination, memory consolidation, decision-making, or feature binding. The role of oscillations in the auditory system is less clear, especially due to the large discrepancy between human, animal studies. Here we describe many methodological issues that confound the results of oscillation studies in the auditory field. Moreover, we discuss the relationship between neural entrainment and oscillations that remains unclear. Finally, we aim to identify which kind of oscillations could be specific or salient to the auditory areas and their processing. We suggest that the role of oscillations might dramatically differ between the primary auditory cortex and the more associative auditory areas. Despite the moderate presence of intrinsic low frequency oscillations in the primary auditory cortex, rhythmic components in the input seem crucial for auditory processing. This allows the phase entrainment between the oscillatory phase and rhythmic input, which is an integral part of stimulus selection within the auditory system.
Article
Full-text available
Communication sounds are typically asymmetric in time and human listeners are highly sensitive to this short-term temporal asymmetry. Nevertheless, causal neurophysiological correlates of auditory perceptual asymmetry remain largely elusive to our current analyses and models. Auditory modelling and animal electrophysiological recordings suggest that perceptual asymmetry results from the presence of multiple time scales of temporal integration, central to the auditory periphery. To test this hypothesis we recorded auditory evoked fields (AEF) elicited by asymmetric sounds in humans. We found a strong correlation between perceived tonal salience of ramped and damped sinusoids and the AEFs, as quantified by the amplitude of the N100m dynamics. The N100m amplitude increased with stimulus half-life time, showing a maximum difference between the ramped and damped stimulus for a modulation half-life time of 4 ms which is greatly reduced at 0.5 ms and 32 ms. This behaviour of the N100m closely parallels psychophysical data in a manner that: i) longer half-life times are associated with a stronger tonal percept, and ii) perceptual differences between damped and ramped are maximal at 4 ms half-life time. Interestingly, differences in evoked fields were significantly stronger in the right hemisphere, indicating some degree of hemispheric specialisation. Furthermore, the N100m magnitude was successfully explained by a pitch perception model using multiple scales of temporal integration of auditory nerve activity patterns. This striking correlation between AEFs, perception, and model predictions suggests that the physiological mechanisms involved in the processing of pitch evoked by temporal asymmetric sounds are reflected in the N100m.
Article
Full-text available
The present study investigated how syllable structure differences between the first Language (L1) and the second language (L2) affect L2 consonant perception and production at syllable-initial and syllable-final positions. The participants were Mandarin-speaking college students who studied English as a second language. Monosyllabic English words were used in the perception test. Production was recorded from each Chinese subject and rated for accentedness by two native speakers of English. Consistent with previous studies, significant positional asymmetry effects were found across speech sound categories in terms of voicing, places of articulation, and manner of articulation. Furthermore, significant correlations between perception and accentedness ratings were found at the syllable onset position but not for the coda. Many exceptions were also found, which could not be solely accounted for by differences in L1-L2 syllabic structures. The results show a strong effect of language experience at the syllable level, which joins force with acoustic, phonetic, and phonemic properties of individual consonants in influencing positional asymmetry in both domains of L2 segmental perception and production. The complexities and exceptions call for further systematic studies on the interactions between syllable structure universals and native-language interference and refined theoretical models to specify the links between perception and production in second language acquisition. (Open Access Web Link at: http://journal.frontiersin.org/article/10.3389/fpsyg.2015.01801/full)
Article
Full-text available
La magnetoencefalografía es una técnica de neuroimagen no invasiva que mide, con gran exactitud temporal, los campos magnéticos en la superficie de la cabeza producidos por corrientes neuronales en regiones cerebrales. Esta técnica es sumamente útil en la investigación básica y clínica, porque además permite ubicar el origen de la actividad neural en el cerebro. En esta revisión se abordan aspectos básicos de la biofísica del método y se discuten los hallazgos sobre procesos como la percepción del habla, la atención auditiva y la integración de la información visual y auditiva, que son importantes en la investigación. Igualmente, se ilustran sus ventajas, sus limitaciones y las nuevas tendencias en la investigación con magnetoencefalografía.
Article
Full-text available
The current study measured neural responses to investigate auditory stream segregation of noise stimuli with or without clear spectral contrast. Sequences of alternating A and B noise bursts were presented to elicit stream segregation in normal-hearing listeners. The successive B bursts in each sequence maintained an equal amount of temporal separation with manipulations introduced on the last stimulus. The last B burst was either delayed for 50% of the sequences or not delayed for the other 50%. The A bursts were jittered in between every two adjacent B bursts. To study the effects of spectral separation on streaming, the A and B bursts were further manipulated by using either bandpass-filtered noises widely spaced in center frequency or broadband noises. Event-related potentials (ERPs) to the last B bursts were analyzed to compare the neural responses to the delay vs. no-delay trials in both passive and attentive listening conditions. In the passive listening condition, a trend for a possible late mismatch negativity (MMN) or late discriminative negativity (LDN) response was observed only when the A and B bursts were spectrally separate, suggesting that spectral separation in the A and B burst sequences could be conducive to stream segregation at the pre-attentive level. In the attentive condition, a P300 response was consistently elicited regardless of whether there was spectral separation between the A and B bursts, indicating the facilitative role of voluntary attention in stream segregation. The results suggest that reliable ERP measures can be used as indirect indicators for auditory stream segregation in conditions of weak spectral contrast. These findings have important implications for cochlear implant (CI) studies – as spectral information available through a CI device or simulation is substantially degraded, it may require more attention to achieve stream segregation.
Conference Paper
Full-text available
Communication sounds are typically asymmetric in time and human listeners are highly sensitive to short-term temporal asymmetry. Nevertheless neurophysiological correlates of perceptual asymmetry remain largely elusive to current approaches. Physiological recordings suggest that perceptual asymmetry is based on multiple scales of temporal integration within the auditory processing hierarchy. To test this hypothesis, we used magneto-encephalographic recordings to perform a model-driven analysis of auditory evoked fields (AEF) elicited by asymmetric sounds characterised by rising or decreasing envelopes (ramped and damped, respectively), using a hierarchical model of pitch perception with top-down modulation. We found a strong correlation between the perceived salience of ramped and damped stimuli and the AEFs, as quantified by the amplitude of the N100m component. Furthermore, the N100m magnitude is closely mirrored by a hierarchical model with stimulus-driven temporal integration windows of auditory nerve activity patterns. This strong correlation of AEFs, perception and modelling suggests that temporal asymmetry is processed in a hierarchical manner where integration windows are top-down modulated.
Article
The common approach to the multiplicity problem calls for controlling the familywise error rate (FWER). This approach, though, has faults, and we point out a few. A different approach to problems of multiple significance testing is presented. It calls for controlling the expected proportion of falsely rejected hypotheses — the false discovery rate. This error rate is equivalent to the FWER when all hypotheses are true but is smaller otherwise. Therefore, in problems where the control of the false discovery rate rather than that of the FWER is desired, there is potential for a gain in power. A simple sequential Bonferronitype procedure is proved to control the false discovery rate for independent test statistics, and a simulation study shows that the gain in power is substantial. The use of the new procedure and the appropriateness of the criterion are illustrated with examples.
Article
Statistical analysis of natural sounds and speech reveals logarithmically distributed spectrotemporal modulations that can cover several orders of magnitude. By contrast, most artificial stimuli used to probe auditory function, including pure tones and white noise, have linearly distributed amplitude fluctuations with a limited average dynamic range. Here we explore whether the operating range of the auditory system is physically matched to the statistical structure of natural sounds. We recorded single-unit and multi-unit neuronal activity from the central nucleus of the cat inferior colliculus (ICC) in response to dynamic spectrotemporal sound sequences to determine whether ICC neurons respond preferentially to linear or logarithmic spectrotemporal amplitudes. We varied the intensity, dynamic range, and contrast statistics of these sounds to mimic those of natural and artificial stimuli. ICC neurons exhibited monotonic and nonmonotonic contrast dependencies with increasing dynamic range that were independent of the stimulus intensity. Midbrain neurons had higher firing rates and higher receptive field energies and showed a net improvement in spectrotemporal encoding ability for logarithmic stimuli, with an increase in the mutual information rate of similar to 50% over linear amplitude sounds. This efficient use of logarithmic spectrotemporal modulations by auditory midbrain neurons reflects a neural adaptation to structural regularities in natural sounds and likely underlies human perceptual abilities.
Article
Magnetoencephalography is a noninvasive imaging technique that measures the magnetic fields on the surface of the head --produced by neuronal currents in brain regions -- and provides highly accurate temporal information. Magnetoencephalography is extremely useful in basic and clinical research as it can also locate the sources of neural activity in the brain. This review chiefly approaches biophysics-related aspects of the method; findings are also discussed on issues such as speech perception, auditory attention and integration of visual-auditory information, which are quintessential in this type of research. Lastly, this review discusses the benefits and limitations of magnetoencephalography and outlines new trends in research with this technique.
Article
Cortical sources of neuromagnetic responses to noise bursts were compared in 7 healthy humans. The earliest response, P40m, peaking about 40 ms after the stimulus onset, was followed by a prominent deflection in the opposite direction at about 100 ms (N100m) and by another peak at 200 ms (P200m). A sustained field, seen near the end of the 400- to 550-ms stimuli, ended with an off-response of the same polarity as N100m. All deflections could be explained by cortical activity within the Sylvian fissure. The source of P200m was anterior to other sources of the on-response, except P40m. When noise bursts, or pauses of equal duration in a continuous noise were presented to the subject, striking similarities were found between the 100-ms deflections of the on- and off-responses: both peaked at about the same latency; their estimated sources were close to each other in the supratemporal plane and their amplitudes depended in a similar way on the interstimulus interval (1.1-8.8 s). However, only the on-response was preceded by P40m, suggesting that P40m and N100m are not causally linked. N100m seems to reflect cortical activity related to any abrupt change in the auditory environment. Les sources corticales des différentes déflections des réponses neuromagnétiques à des salves de bruit ont été comparées chez 7 sujets humains en bonne santé. Les réponses les plus précoces, P40m, présentant un pic aux environs de 40 ms après l'établissement du stimulus, étaient suivies par une déflection plus importante dans la direction oposée à environ 100 ms (N100m) puis par un autre pic à 200 ms (P200m). Un champ magnétique continu observé vers la fin des 400-550 ms du stimulus se terminait avec une réponse à l'extinction de la měme polarité que N100m. Toutes les déflections pouvaient s'expliquer par une activité corticale localisée dans la fissure de Sylvius. La source du P200m était significativement antérieure aux autres sources de la réponse à l'établissement sauf pour P40m. Quand des salves de bruit ou bien des pauses de durée égale dans un son continu étaient présentées au sujet, des similitudes frappantes étaient observées entre les déflections à 100 ms pour les réponses à l'établissement et à l'extinction: les deux présentaient un pic avec approximativement la meme latence, leurs sources étaient estimées voisines dans le plan supra-temporal et leurs amplitudes dépendaient de la měme manière de l'intervalle entre stimulus (1,1-8,8 s). Cependant, seule la réponse à l'établissement était précédée par P40m, suggérant par lá que P40m et N100m ne sont pas liées causalement. N100m semble refléter une activité corticale en rapport avec les changements abrupts dans l'environnement auditif.