Hearing Research 335 (2016) 25e32

Contents lists available at ScienceDirect

Hearing Research
journal homepage: www.elsevier.com/locate/heares

Research paper

Functional magnetic resonance imaging conrms forward suppression

for rapidly alternating sounds in human auditory cortex but not in the
inferior colliculus
Christian Harm Uhlig*, Andrew R. Dykstra, Alexander Gutschalk**
Department of Neurology, University of Heidelberg, INF 400, 69120 Heidelberg, Germany

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 19 November 2015
Received in revised form
8 February 2016
Accepted 15 February 2016
Available online 17 February 2016

Forward suppression at the level of the auditory cortex has been suggested to subserve auditory stream
segregation. Recent results in non-streaming stimulation contexts have indicated that forward suppression
can also be observed in the inferior colliculus; whether this holds for streaming-related contexts remains
unclear. Here, we used cardiac-gated fMRI to examine forward suppression in the inferior colliculus (and the
rest of the human auditory pathway) in response to canonical streaming stimuli (rapid tone sequences
comprised of either one repetitive tone or two alternating tones). The rst stimulus is typically perceived as a
single stream, the second as two interleaved streams. In different experiments using either pure tones
differing in frequency or bandpass-ltered noise differing in inter-aural time differences, we observed
stronger auditory cortex activation in response to alternating vs. repetitive stimulation, consistent with the
presence of forward suppression. In contrast, activity in the inferior colliculus and other subcortical nuclei did
not signicantly differ between alternating and monotonic stimuli. This nding could be explained by active
amplication of forward suppression in auditory cortex, by a low rate (or absence) of cells showing forward
suppression in inferior colliculus, or both.
2016 Elsevier B.V. All rights reserved.

Keywords:
Forward suppression
Selective adaptation
Stimulus-specic adaptation
Inferior colliculus
Stream segregation
Interaural time differences
Functional magnetic resonance imaging

1. Introduction
Separating sound sources in complex environments is a critical
function of the auditory system that allows humans and other
animals to hear out and selectively attend sources of interest. A
widely used paradigm in the study of such source segregation is
auditory 'stream-segregation', or 'streaming' (Bregman, 1990;
Miller and Heise, 1950; van Noorden, 1975), in which a repetitive
sequence of alternating tones (A and B) can be perceived either as
one integrated stream or two distinct, segregated streams. Most
previous studies have utilized frequency differences (DF) between
the A and B tones to examine streaming, but differences along other
dimensions e e.g. pitch (Vliegen et al., 1999), amplitude modulation

Abbreviations: DF, Frequency differences; DITD, Differences of Interaual Time

Differences; AC, Auditory Cortex; CN, Cochlear Nucleus; fMRI, functional magnetic
resonance imaging; HG, Heschl's gyrus; ITD, Interaural Time Difference; MGB,
Medial geniculate nucleus; NLL, Nuclei of the lateral lemniscus; PT, Planum Temporale; ROI, Region of Interest; S.D., Standard Deviation; S.E.M, Standard Error of
the Mean; SOC, Superior Olivary Complex
* Corresponding author.
** Corresponding author.
E-mail addresses: christian@uhlig.email (C.H. Uhlig), alexander.gutschalk@med.
uni-heidelberg.de (A. Gutschalk).
http://dx.doi.org/10.1016/j.heares.2016.02.010
0378-5955/ 2016 Elsevier B.V. All rights reserved.

(Grimault et al., 2002) or spatial lateralization (Boehnke and

Phillips, 2005) e can also produce streaming.
Microelectrode studies in animal models suggest that streaming
based on DF could be subserved by frequency-specic forward suppression in auditory cortex (AC) (Bee and Klump, 2004; Brosch and
Schreiner, 1997; Fishman et al., 2004, 2001; Scholes et al., 2015,
2011). Similarly, the spatial tuning of AC neurons is known to be
sharpened in streaming contexts that can be well explained by a
stimulus-specic forward suppression model (Middlebrooks and
Bremen, 2013). On a macroscopic scale, analogous results have
been obtained in human listeners with magnetoencephalography
(Gutschalk et al., 2005), electroencephalography (Snyder et al., 2006),
and functional magnetic resonance imaging (fMRI) (Gutschalk et al.,
2007; Schadwinkel and Gutschalk, 2010; Wilson et al., 2007). However, one question that remains is whether the streaming-related
forward suppression that has been observed in AC emerges there
or instead reects the output of subcortical processes.1
Forward suppression has been observed in the inferior colliculus

1
Long-term adaptation, which has been suggested to subserve stream segregation independent of forward suppression (Micheyl et al., 2005), has in fact been
observed in the cochlear nucleus (CN) of anaesthetized guinea pigs (Pressnitzer
et al., 2008).

26

C.H. Uhlig et al. / Hearing Research 335 (2016) 25e32

(IC) of awake marmosets using a two-tone paradigm (Nelson et al.,

rez-Gonz

2009). Furthermore, stimulus-specic adaptation (Pe
alez
and Malmierca, 2014), a phenomenon akin to forward suppression, has been observed in the rodent IC and medial geniculate
nucleus (MGB) within the context of the auditory oddball paradigm, where rare deviant sounds are presented amidst frequent
standards (Anderson et al., 2009; Malmierca et al., 2009; Patel et al.,

rez-Gonza

lez et al., 2005; Zhao et al., 2011). Whether for2012; Pe
ward suppression and stimulus-specic adaptation reect a common or different underlying mechanisms remains unclear. Both are
simply dened by reduction of neural responses when a stimulus is
repeated, which is stronger for identical stimuli and weaker for
more dissimilar stimuli. Two of us have previously used another
synonym e selective adaptation e in the context of streaming
(Gutschalk et al., 2005; Gutschalk and Dykstra, 2014) and consider
the three terms interchangeable. In the present paper we use forward suppression in the context of streaming and stimulus-specic
adaptation in the context of the oddball paradigm, as that is how
previous, disparate lines of research have used them. This also allows for the possibility that different mechanisms or anatomical
centres may be recruited by the two paradigms.
Using sparse-sampled (Edmister et al., 1999; Hall et al., 1999),
cardiac-gated (Guimaraes et al., 1998) fMRI, we examined whether
streaming-related forward suppression observed in AC is inherited
from subcortical structures, particularly the IC. We hypothesized
that stimulus sequences with alternating frequencies (DF, experiment 1) or interaural time differences (DITD, experiment 2) would
produce larger blood-oxygenation level-dependent (BOLD) activity
than monotone control sequences, reecting the presence of forward suppression in both the IC and AC.
2. Materials and methods
2.1. Participants
30 healthy listeners participated in the study, 15 per experiment.
The mean age was 25 3 years (standard deviation: S.D.) in
experiment 1 and 23 2 years (S.D.) in experiment 2. Ten participants of experiment 1 and twelve participants of experiment 2
were female. An additional ve participants (three from experiment 1 and two from experiment 2) were excluded from analysis
due to excessive movement in the scanner. All participants had
clinically normal pure-tone audiograms with threshold less than
15 dB HL between 0.125 and 12.5 kHz. Each participant provided
written informed consent prior to their participation in the experiments. All experiments were approved by the ethics committee
of Heidelberg University Medical School and conform with the in
accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki).
2.2. Stimuli and procedures
All stimuli were generated with MATLAB (The Mathworks,
Natick, MA, USA) and stored as wave les with a sample-rate of
48 kHz. The wave les were presented via a D/A converter and
headphone amplier (MR Confon; MR confon GmbH, Magdeburg,
Germany) with Sensimetrics S14 in-ear headphones (Sensimetrics
Corporation, Malden, MA, USA). The non-linear transfer function of
the in-ear headphones was corrected using software provided by
the manufacturer.
2.2.1. Experiment 1: frequency separation e DF
The stimuli consisted of sequences of pure tones labelled A and
B, with frequencies of 600 and 849 Hz, respectively (i.e., the B-tone
was 6 semitones higher than the A tone), at 74 dB sound-pressure

level (SPL). The three conditions were: CONST1 (AAAA ), CONST2

(BBBB ), and ALT (ABAB ). The two CONST conditions were used
in order to equalize the long-term power spectra between the ALT
and CONST conditions so as to avoid such a confound in the ALTversus-CONST contrast. The length of each tone was 100 ms with
15 ms on- and off-raised-cosine ramps. There were no silent interstimulus intervals between subsequent tones. Each tone sequence
comprised 320 tone repetitions, amounting to a duration of 32 s.
The ALT condition was repeated 32 times and the CONST1 and
CONST2 conditions were presented 16 times each. The sequences
were presented in pseudo-random order with a variable interstimulus interval of 24e32 s. This was done in order to stagger
the BOLD volume acquisitions with respect to stimulus onset to
allow for reconstruction of the BOLD signal timecourse (see Section
2.3.). The presentation of each 32-s tone sequence was started with
a delay of 0, 2, 4, or 6 s relative to the trigger sent at the onset of the
preceding scanner acquisition. Because of the cardiac gating procedure (see below), there is additional variability of the interstimulus interval in the sub-second range.
2.2.2. Experiment 2: spatial separation e DITD
In experiment 2, stimuli were composed of band-limited
(0e2 kHz) noise bursts, lateralized to the left (L) or right (R) with
an ITD of 500 ms. Sound level was 75 dB SPL. In analogy to
experiment 1, three conditions were used: CONST1 (RRRR ),
CONST2 (LLLL ), and ALT (RLRL ). Parameters and procedures
were otherwise similar to experiment 1. Conditions CONST1 and 2
were presented 12 times each, and ALT was presented 24 times.
2.2.3. Procedures for both experiments
The task was explained one day before the fMRI session, including
one or two training runs with circumaural headphones connected to
a desktop computer. The training ended after the experimenter was
condent that the participants understood the task, although some
participants nonetheless made a small number of erroneous button
presses. To determine if the participants perceived the stimuli as one
or two separate streams in the scanner, they were instructed to
evaluate the stimuli by pressing one of two buttons after the end of
each 32 s sequence. One button indicated that they had heard one
sequence most of the time, while the other button indicated that they
had heard two separate streams most of the time.
2.3. Imaging
All MRI data were acquired in a Siemens 3T Magnetom Tim Trio
scanner (Siemens, Erlangen, Germany) with a 32-channel, phasedarray head coil. Two T1-weighted magnetization-prepaired rapid
gradient echo sequences (MPRAGE) with a dimension of
256
256
192 voxel, an isovoxel resolution of 1 mm3, a TR of
1570 ms, a TE of 2.63 ms, a TI of 900 ms, and a Flip Angle of 9 in one
frame for each participant was collected. These scans were used to
place the functional volume. To cover the complete ascending
auditory pathway, the functional volume was placed in a nearcoronal orientation orthogonal to the Sylvian ssure, such that AC
and the brainstem were inside the volume. The functional volume
comprised 21 slices (2.1 mm thickness, distance 10%) with a eld of
view of 204
204 mm (120
120 voxel, resolution 1.7
1.7 mm)
and included the brain-stem as well as most of AC. The parameters
for BOLD imaging were echo time (TE) 42 ms, inversion time
(TI) 1 ms, Flip Angle 90 the phase coding was chosen from feet
to head. In-house software was used for stimulus presentation and
collection of participants' responses with a LUMItouch optical
response keypad (Photon Control, Burnaby, BC, Canada). A modied
sparse-imaging (Edmister et al., 1999; Hall et al., 1999) and cardiacgated (Guimaraes et al., 1998) paradigm with an average TR of 8.0 s

C.H. Uhlig et al. / Hearing Research 335 (2016) 25e32

was used. Each acquisition was started after a minimum delay of

7.5 s plus the delay to the next peak of the pulse curve, measured by
using a pulse-oxymeter at the nger tip. Thus, with pulse rates
around 60/min, the TR varies in a range of 7.5e8.5 s, and is typically
close to 8 s on average.
2.4. Data analysis
2.4.1. Activation maps
The structural data were analysed with Freesurfer Version 5.3.0
(http://surfer.nmr.mgh.harvard.edu/) using the surface-based
stream (Dale et al., 1999; Fischl et al., 1999) for the cortical anatomy and the volume-based stream for the sub-cortical anatomy
(Fischl et al., 2002).
The functional data were motion corrected, slice timing corrected, smoothed, and aligned to a template brain using Freesurfer
on an Ubuntu 12.04 LTS 64 bit operating system with an Intel Core
i5-2500 CPU @ 3.3 GHz
4 processor. Because the freesurfer architecture is based on a regular TR, we performed the analysis under
the assumption of a regular TR of 8 s and individually corrected the
offset-times in the paradigm les for the auditory stimulation to
match the actual acquisition time of the directly following acquisition. The acquisition time of the scanner was recorded by the
stimulation software and matched with the scanner timing stored
in the DICOM les to provide exact timing information. Individual
activation maps were calculated for the sound-versus-silence and
ALT-versus-CONST contrasts. These contrasts were used to perform
a second-level, mixed-effects group analysis with a cluster-wise
correction for multiple comparisons using the algorithm
described in (Hagler et al., 2006) as implemented in Freesurfer.
Synthesized z-score maps are created using Gaussian random noise,
smoothed, and thresholded to consider only those vertices or voxels
with z-score >1.3 (corresponding to a vertex-/voxel-wise p-value of
0.05); only positive clusters were considered. Then, the size of the
largest cluster is recorded, and this process is repeated N times
(here, N 10,000) in order to generate a distribution of maximum
cluster sizes under the null hypothesis of no difference between
conditions. Cluster-wise p-values from the actual contrast (after the
same vertex- or voxel-wise thresholding) are then determined by
where these cluster sizes fall within the distribution of cluster sizes
created by the simulation. The threshold for signicant clusters was
set at p < 0.05 for the cortical and the subcortical analysis. Note that
our setup, and consequently our analysis, was limited to the auditory pathways and that other areas were not consistently covered
across participants. Accordingly, signicant activation in areas
outside of the auditory pathway will be briey reported but will not
be evaluated in detail.
2.4.2. Regions-of-interest analysis
Regions of interest (ROIs) were dened as intersection of voxels
(vertices for cortex) within an anatomical structure of interest that
were signicantly activated by the sound-versus-silence contrast
at the group level (all sound conditions vs. baseline). The
anatomical constraints of cortical ROIs was taken as the combination of Freesurfer labels (Destrieux et al., 2010) comprising the
superior temporal plane. The anatomical constraints of brain-stem
ROIs were dened manually based on the atlas of Paxinos (Paxinos
et al., 2012) in analogy to a previous study from our laboratory
(Gutschalk and Steinmann, 2015). For each ROI, a small sphere was
dened with the following centre coordinates and diameters in
Talairach coordinates: CN 12, 41.5, 34 mm (diameter 6 mm);
superior olivary complex (SOC) 6, 38, 32 mm (diameter
6 mm); nucleus of the lateral lemniscus (NLL) 10, 35, 17 mm
(diameter 6 mm); IC 5, 35, 7 mm (diameter 7 mm); MGB 16,
26, 3 mm (diameter 7 mm).

27

The mean beta values within each ROI for each participant were
entered into a two-way repeated-measurements analysis of variance (ANOVA) using R Software (R Core Team, 2015) with the factors condition (ALT and CONST) and hemisphere. To evaluate
differences between AC and the IC, ROI (AC, IC) was added as third
factor for an additional analysis.
Finally, to examine whether classically dened sub-regions of
AC might show different results, we performed an additional ROI
analysis using hand-drawn anatomical labels that encompassed (i)
medial HG, (ii) lateral HG, and (iii) PT. However, because this
analysis yielded results that were qualitatively similar to that from
the analysis in which these sub-ROIs were grouped (i.e. main effects
of condition in each ROI, in each hemisphere, for both experiments), we report only the grouped analysis in detail.
3. Results
3.1. Experiment 1: frequency separation e DF
Stimuli were 32-s-long sound sequences of 10-Hz repetitive
pure tones with tone frequencies of A 600 Hz and B 849 Hz,
corresponding to a frequency difference (DF) of 6 semitones for the
alternating tone conditions (ALT). In the constant-tone condition
(CONST), both tone sequences (either AAAA or BBBB ) were
used with equal probability.
3.1.1. Behavioural data
On average, the participants indicated that they perceived two
streams in 90.0 5.4% (S.E.M.: Standard Error of the Mean) of the
alternating condition and one stream in 93.8 1.6% (S.E.M.) of the
constant condition. Although this seems to suggest that some
participants on some trials may have perceived two streams for the
monotone sequences and one stream for the alternating sequences,
given the above percentages, these are likely to reect simple
erroneous button presses over the approximately 80min experimental session.
3.1.2. Functional MRI data
The acquisition volume was restricted to cover the auditory
cortex and the brain-stem, to reduce the overall scanner sound
level caused by gradient pulses during the experiment. Within this
region, the sound-versus-silence contrast revealed activations of
Heschl's gyrus (HG) in both hemispheres and parts of the Planum
temporale (PT) in the right hemisphere, as well as subcortical activations in the brain-stem and thalamus. The activation includes
the whole human auditory pathway (Fig. 1A).
The ALT-versus-CONST contrast shows AC activation in both
hemispheres. This contrast reveals signicantly stronger activation
for ALT compared to CONST in HG and PT (Fig. 1B). In the subcortical
analysis, signicant clusters of activation were only observed in
medial temporal cortex on both sides for the ALT-versus-CONST
contrast. Because we had no prior hypothesis for this area we do
not report this results in detail. Part of a left activation cluster
overlapped with our anatomical constraint of the MGB region of
interest. However, since there was no MGB activity in the soundversus-silence contrast on the left and the cluster lay predominantly outside of the MGB, we did not consider this activity to
clearly indicate enhanced MGB activity. Otherwise, the subcortical
analysis revealed no signicant activation for the ALT-versusCONST contrast.
For a more detailed comparison of activity evoked by the ALT
and CONST conditions in the auditory pathway, an ROI analysis was
performed using all voxels that (i) were inside anatomically dened
areas and (ii) showed signicant activation in the sound-versussilence contrast. Fig. 1C shows the mean beta-values of the ROI

28

C.H. Uhlig et al. / Hearing Research 335 (2016) 25e32

Fig. 1. Activation maps and ROI Analysis of the DF experiment (N 15). (A) Activation map for the sound-versus-silence contrast in cortex (left) and the subcortical auditory
pathway (right). Cluster-based correction for multiple comparisons (p < 0.05). (B) The ALT-versus-CONST contrast reveals highly signicant activation in AC. (C) ROI analysis based
on voxels active in the sound-versus-silence task within each of the referenced anatomical sites: CN: cochlear nucleus; SOC: supperior olivary complex; NLL: nucleus of the lateral
lemniscus; IC: inferior colliculus; MGB: medial geniculate nucleus; AC: auditory cortex. Signicant differences between conditions are indicated by stars (*: p < 0.05, **: p < 0.01; ***:
p < 0.001). Refer to Table 1 for the details of the statistical comparison.

analysis. The ROI analysis conrmed that there was no signicant

difference between CONST and ALT in the CN, SOC, NLL, IC and MGB
(Table 1). Conversely, as could already be expected from the vertexwise analysis, ALT evoked signicantly higher activity than CONST
in AC. An ANOVA with the additional factor ROI was performed to
further examine the null hypothesis that response enhancement
for ALT over CONST was similar in the IC and AC. This analysis
yielded a signicant interaction (F(1,14) 17.57, p 0.000905) of
ROI
condition (ALT versus CONST), conrming that the response
pattern in AC was signicantly different from the pattern observed
in the IC.

3.2. Experiment 2: spatial separation e DITD

In experiment 2, stimuli were 10-Hz noise bursts, lateralized to
the left (L) or right (R) with an ITD of 500 ms. Other parameters
were similar to experiment 1.

3.2.1. Behavioural data

The participants indicated that they perceived 80.6 5.6%
(S.E.M.) of the alternating condition as two streams and 82.8 3.9%
(S.E.M.) of the constant condition as one stream. The higher error
rate for the one-stream condition is probably related to ITD-based
streaming being less distinct than frequency-based streaming
(Boehnke and Phillips, 2005).

3.2.2. Functional MRI data

Fig. 2A shows the sound-versus-silence contrast, which revealed
activation of HG in both hemispheres and parts of the PT in the
right hemisphere, as well as in the complete ascending auditory
pathway. Fig. 2B shows the ALT-versus-CONST contrast for AC. A
signicant difference between the ALT and CONST conditions was
observed in both left and right AC, though the effect in left AC was
conned to posterolateral HG.
In the subcortical analysis, the ALT-versus-CONST contrast
revealed an activation cluster that overlapped with parts of the
anatomically dened ROI of the right SOC, but which was otherwise
not related to the auditory pathway and which was therefore not
evaluated in more detail. Another signicant cluster was observed
in the right medial temporal lobe, similar to that observed in
experiment 1. However, this cluster did not overlap with any
anatomical ROI of the auditory pathway, and was therefore outside
the scope of the present study.
Fig. 2C shows the results of the ROI analysis (c.f. Table 2 for the
statistical analysis). Note the higher activation levels for noise
bursts in comparison to pure tones in experiment 1 (Fig. 1), which
are likely related to stimulus bandwidths (Hawley et al., 2005;
Overath et al., 2012). The ROI analysis conrmed that there were
no signicant differences between CONST and ALT in the CN, SOC,
NLL, IC and MGB (Table 2). As expected, the ALT condition evoked
signicantly higher activity than CONST in AC. The comparison of
the effects in AC and the IC revealed no signicant results for an
ROI
condition interaction (F(1,14) 0.231, p 0.638). This is

C.H. Uhlig et al. / Hearing Research 335 (2016) 25e32

29

Table 1
Results of a two-way ANOVA for repeated measurements of the DF experiment (N 15). The entered factors are hemisphere (left and right) and condition (CONST and ALT).
Signicant differences between conditions are indicated by stars (***: p < 0.001).
ROI

Hemisphere

Condition

Hemisphere

condition

Condition (in left

hemisphere)

Condition (in right

hemisphere)

Hemisphere (for
CONST condition)

Hemisphere (for
ALT condition)

CN

F(1,14) 0.301, p 0.592

F(1,14) 0.083,
p 0.778
F(1,14) 0.596,
p 0.453
F(1,14) 0.43,
p 0.523
F(1,14) 0.259,
p 0.618

F(1,14) 0.028,
p 0.970
F(1,14) 0.173,
p 0.684
F(1,14) 1.266,
p 0.279
F(1,14) 1.101,
p 0.312

F(1,14) 0.068,
p 0.798
F(1,14) 0.056,
p 0.817
F(1,14) 0, p 0.993

F(1,14) 0.047,
p 0.831
F(1,14) 0.948,
p 0.347
F(1,14) 1.278,
p 0.277
F(1,14) 0.805,
p 0.385

F(1,14) 0.161,
p 0.694
F(1,14) 0.683,
p 0.422
F(1,14) 2.327,
p 0.149
F(1,14) 0, p 0.991

F(1,14) 0.028,
p 0.869
F(1,14) 0.035,
p 0.855
F(1,14) 0.042,
p 0.841
F(1,14) 1.569,
p 0.231

F(1,14) 1.373,
p 0.261

F(1,14) 1.634,
p 0.222

SOC F(1,14) 0.462, p 0.508

NLL F(1,14) 0.983, p 0.338
IC

F(1,14) 0.425,
p 0.525F(1,14) 0.161,
p 0.694

F(1,14) 0.017,
p 0.897

MGB
AC

F(1,14) 1.497, p 0.241

F(1,14) 90.94,
p 1.68e07 ***

F(1,14) 0.719, F(1,14) 78.25,

p 0.411
p 4.17e07 ***

probably because the overall enhancement in auditory cortex is

smaller for ITD compared to the frequency-based streaming
experiment. Numerically, activity in IC is on average smaller for the
alternating compared to the monotone stimulation.
Independent of our research question, activity was signicantly
stronger in the right AC, IC, and CN. While right dominance in the
AC and IC would t together well, it appears a bit odd at the CN level
where most excitatory bres have not yet crossed sides.
The left- and right-lateralized condition of CONST (CONST1:
right; CONST2: left) were also evaluated to probe for the general

F(1,14) 1,293,
p 0.275
F(1,14) 71.69,
p 7.03e07 ***

effects of ITD. The ANOVA revealed a signicant interaction of

hemisphere
condition (F(1,14) 5.828, p 0.03, which was
driven by signicantly stronger activity in right AC for left-than for
right-lateralized sounds (F(1,14) 9.033, p 0.00945). No signicant ITD dependence was observed in left AC or in the subcortical
pathway.
4. Discussion
The results of this study cannot conrm forward suppression in

Fig. 2. Activation maps and ROI Analysis of the DITD experiment (N 15). (A) The sound-versus-silence contras in cortex (left) and the subcortical auditory pathway (right).
Cluster-based correction for multiple comparisons (p < 0.05). (B) The ALT-versus-CONST contrast reveals differences in right AC. (C) ROI analysis based on voxels active in the soundversus-silence task within each of the referenced anatomical sites: CN: cochlear nucleus; SOC: supperior olivary complex; NLL: nucleus of the lateral lemniscus; IC: inferior colliculus; MGB: medial geniculate nucleus; AC: auditory cortex. Signicant differences between conditions are indicated by stars (*: p < 0.05, **: p < 0.01; ***: p < 0.001). Refer to
Table 2 for the details of the statistical comparison. not be part of the PDF.

30

C.H. Uhlig et al. / Hearing Research 335 (2016) 25e32

Table 2
Results of a two-way ANOVA for repeated measurements of the DITD experiment (N 15). The entered factors are hemisphere (left and right) and condition (CONST and
ALT). Signicant differences between conditions are indicated by stars (*: p < 0.05, **: p < 0.01).
ROI

Hemisphere

F(1,14) 5.142,
p 0.0397 *
SOC F(1,14) 0.416,
p 0.530
NLL F(1,14) 0.256,
p 0.621
IC
F(1,14) 5.411,
p 0.036 *
MGB F(1,14) 0.589,
p 0.456
AC F(1,14) 10.76
p 0.00547 **
CN

Condition

Hemisphere

condition

Condition (in left

hemisphere)

Condition (in right

hemisphere)

Hemisphere (for CONST Hemisphere (for ALT

condition)
condition)

F(1,14) 0.038,
p 0.848
F(1,14) 0.602,
p 0.451
F(1,14) 0.454,
p 0.511
F(1,14) 1.306,
p 0.272
F(1,14) 0.168,
p 0.688
F(1,14) 15.43,
p 0.00152 **

F(1,14) 0.325,
p 0.578
F(1,14) 1.004,
p 0.333
F(1,14) 1.257,
p 0.281
F(1,14) 0.309,
p 0.587
F(1,14) 0.665,
p 0.428
F(1,14) 0.39,
p 0.542

F(1,14) 0.044,
p 0.838
F(1,14) 0, p 1

F(1,14) 0.122,
p 0.732
F(1,14) 1.047,
p 0.324
F(1,14) 1.137,
p 0.304
F(1,14) 0.764,
p 0.397
F(1,14) 0.045,
p 0.836
F(1,14) 12.8,
p 0.00303 **

F(1,14) 3.05, p 0.103 F(1,14) 4.018,

p 0.0648.
F(1,14) 1.042,
F(1,14) 0.004,
p 0.325
p 0.953
F(1,14) 0.063,
F(1,14) 0.835,
p 0.806
p 0.376
F(1,14) 4.372,
F(1,14) 4.448,
p 0.0553
p 0.0534
F(1,14) 1.699,
F(1,14) 0.052,
p 0.213
p 0.823
F(1,14) 8.1, p 0.0129 F(1,14) 13.97,
*
p 0.00221 **

F(1,14) 0.017,
p 0.899
F(1,14) 1.73,
p 0.21
F(1,14) 0.649,
p 0.434
F(1,14) 11.29,
p 0.00467 **

the IC in the context of streaming stimuli when using fMRI. In

contrast, activity in AC was lower for constant compared to alternating stimulus sequences, consistent with previous fMRI data
(Gutschalk et al., 2007; Schadwinkel and Gutschalk, 2011; Wilson
et al., 2007). The latter effect has been linked to forward suppression demonstrated in animal models for both two-tone paradigms
(Brosch and Schreiner, 1997) and typical streaming stimuli (Bee and
Klump, 2004; Fishman et al., 2004, 2001; Middlebrooks and
Bremen, 2013). The most straight-forward potential explanation
of our data would therefore be that streaming-related forward
suppression is mostly a cortical phenomenon, and not inherited (at
least not entirely) from the IC or other sub-cortical auditory centers.
However, several microelectrode studies in animal models
indicate that frequency-specic response suppression, termed
either forward suppression in the context of forward-masking
(Nelson et al., 2009) or stimulus-specic adaptation in the
context of oddball paradigms (Duque et al., 2012; Malmierca et al.,

rez-Gonz

2009; Malone and Semple, 2001; Pe
alez et al., 2005) are
already present at the level of the IC. Moreover, fMRI studies using
auditory oddball paradigms in both rat (Gao et al., 2014) and man
(Cacciaglia et al., 2015) are consistent with the presence of
stimulus-specic adaptation in the IC. The latter study is of
particular interest, because it suggests that the relative response
enhancement in the IC was in the same range as in AC, which was
clearly not the case for alternating compared to monotone sequences in the present study. The comparison between the present
study and the study by Cacciaglia et al. (Cacciaglia et al., 2015).
Therefore suggests that the mechanisms for BOLD signal
enhancement may be different in the case of the streaming and
oddball paradigms. Therefore, even if it is not possible to rule out
that forward suppression is present in the IC in the context of
streaming based on the present fMRI data, it would at least appear
that such an effect is enhanced in AC (either by amplication or by a
larger proportion of neurons showing the effect, or both e see
below), if not emergent there.
For the comparison of streaming and oddball paradigms, it is
important to consider the extent to which forward suppression and
stimulus-specic adaptation reect distinct underlying mechanisms, a question that is not yet settled. While some authors have
used forward suppression and adaptation interchangeably and in a
phenomenological sense (Gutschalk and Dykstra, 2014; Scholes
et al., 2011), others have argued that they reect distinct mechanisms, one e stimulus-specic adaptation e that produces longerlasting effects, and the other e forward suppression e limited to a
few hundred milliseconds (Ulanovsky et al., 2003). The effect
observed for streaming stimuli in AC (Bee and Klump, 2004;
Fishman et al., 2001) is usually interpreted as forward suppression, as in forward-masking paradigms (Brosch et al., 1999). Based

on in-vivo patch-clamp recordings, it has been suggested that

GABAergic inhibition can not account for forward suppression
beyond a time range of about 100 ms (Wehr and Zador, 2005). The
authors suggested instead that other cortical mechanisms e most
likely synaptic depletion e was the cause of forward suppression
for longer time scales, and that this emerged in AC. A relationship
between streaming perception and longer-term adaptation has also
been discussed, rst in AC (Bee et al., 2010; Micheyl et al., 2007,
2005), but subsequently also in the CN (Pressnitzer et al., 2008).
These effects are distinct from the effects attributed to forward
suppression, but their relationship to stimulus-specic adaptation
with longer time constants in the context of oddball paradigms
(Ulanovsky et al., 2003) is also unclear.
It has recently been discussed that the stimulus-specic adaptation observed in the IC in the oddball paradigm may not reect a
bottom-up process, but might instead be controlled by top-down
mechanisms originating in AC. This possibility has been explored
for the MGB by two studies that inactivated ipsilateral AC using
either cortical cooling (Antunes and Malmierca, 2011) or the
uerle et al., 2011). The results of
inhibitory agonist muscimol (Ba
these studies are similar in that they nd a modulating effect of AC
uerle et al., 2011)
on the MGB. However, while one study (Ba
concluded that stimulus-specic adaptation in the MGB is more or
less absent when AC is inactivated, another study (Antunes and
Malmierca, 2011) suggested that it is only modulated by AC activity [see (Yu et al., 2009), for an alternative explanation involving the
thalamic reticular nucleus].
In the IC, cooling of AC has produced similar effects (Anderson
and Malmierca, 2012), i.e. modulation but not complete suppression of stimulus-specic adaptation by ipsilateral AC cooling. More
recently, comparison of the anatomical connectivity of IC neurons
showing stimulus-specic adaptation with those that do not suggest a role for AC: While neurons showing no stimulus-specic
adaptation received their input primarily from brain-stem nuclei,
those that showed stimulus-specic adaptation received strong
input from AC and comparatively sparse input from brain-stem
nuclei (Ayala et al., 2015). It is conceivable that the role of topdown signals is different for stimulus-specic adaptation as
recorded with typical oddball paradigms and for forward suppression in the context of streaming stimuli, and that the two
constitute different mechanisms, but this remains speculative.
Along the same lines, it has been suggested that the response
pattern observed during reversals of bistable streaming stimuli is
similar to that observed for deviants in an oddball paradigm (Kondo
and Kashino, 2009). Thus, IC activation during streaming reversals
(Schadwinkel and Gutschalk, 2011) may also be explained by topdown projections from auditory cortex, similar to the response
enhancement observed for rare deviants in an oddball paradigm.

C.H. Uhlig et al. / Hearing Research 335 (2016) 25e32

The absence of a signicant BOLD effect in the IC during

continuous streaming stimuli, where tones alternate at a much
faster rate, could then be attributed to a general difference between
forward suppression and stimulus-specic adaptation. Alternatively, it could be that the fast and continuous streaming stimuli
evoke such strong activity in neurons not showing stimulusspecic adaptation (which tend to be located in the central nucleus of the IC) (Ayala et al., 2015) that these neurons dominate the
IC BOLD signal, such that the enhancement in other parts of the IC
(in particular in lateral and rostral regions) (Ayala et al., 2015) is so
small in comparison that it is missed in the average IC signal. It is
also unclear if forward suppression in the context of classical
forward-masking is restricted to a subset of IC neurons or represents a more general phenomenon (Nelson et al., 2009). In fMRI,
potential correlates of adaptation or forward suppression have
been demonstrated only indirectly via rate dependence, and
selectivity of the adaptation has not been demonstrated. For
example, it has been shown that activity in AC increases with pulserepetition only up to rates of about 10 Hz. For higher rates or
shorter inter-stimulus intervals, the sustained BOLD response is in
fact lower than at 10 Hz (Harms et al., 2005; Harms and Melcher,
2002). In contrast, activity in the IC increases up to at least 20 Hz
and remains elevated for higher rates and even continuous noise
(Hawley et al., 2005; Steinmann and Gutschalk, 2012). Another
observation is that addition of sinusoidal amplitude modulation on
a continuous carrier enhances activity in the IC, MGB and primary
AC for relatively fast modulation rates of 40 Hz (Steinmann and
Gutschalk, 2012), whereas amplitude enhancement was conned
to the auditory belt cortex and not observed subcortically for lower
rates (8 Hz) (Gutschalk and Steinmann, 2015). While another study
found no rate effects in the IC when normalizing for average sound
intensity (Overath et al., 2012), the same study conrmed that
auditory cortex produces higher activity at 3 Hz than at rates of
30 Hz and more.
Together these results from BOLD fMRI are consistent with the
notion that sensitivity to high rates decreases along the ascending
auditory pathway (Creutzfeldt et al., 1980), or that forward suppression is stronger in cortex (Wehr and Zador, 2005). We therefore
expected that forward suppression in the IC would be more likely to
be observed with fast stimulation rates or short inter-stimulus intervals (Fishman et al., 2004); however, despite the parameters
chosen e 10 Hz repetition rate and no silent interval except for
falling/rising ramps e no signicant evidence for forward suppression in the IC was found. Note that this result is limited to a
macroscopic view, and that some IC neurons may very well show
forward suppression for the stimuli used here. However, our ndings suggest that it is unlikely to be as strong or prevalent as in AC.
In our data, activity in MGB was similar to that in IC in that it was
not signicantly higher for alternating vs. monotone sequences.
Note, however, that BOLD activity in the MGB is generally weaker
than in IC (Sigalovsky and Melcher, 2006), and that the results are
therefore more variable here. Even if forward suppression was
observed only in AC, there might still be a role for MGB neurons;
considering that BOLD is more sensitive to synaptic activity
(Logothetis et al., 2001), fMRI may not see spiking activity from the
MGB that projects to AC. Thus, with the spatial resolution used in
this study, we cannot dissociate whether the response suppression
we see for monotone sequences in AC reects thalamo-cortical or
cortico-cortical mechanisms.
Potential explanations for the nding of response enhancement
for alternating sequences in AC but not IC or MGB include (i) that
forward suppression is actively amplied in auditory cortex or (ii)
that a larger proportion of AC vs. IC neurons show forward suppression. Very recent data from an animal model support our results and suggest that forward suppression e at least in the context

31

of streaming e emerges at thalamo-cortical synapses (Yao et al.,

2015a). These authors evaluated streaming based on spatial separation with unit recordings in the IC, MGB, and AC of anesthetized
rats. While evidence of stream segregation was found in the nucleus of the brachium of the IC and in about 2/3 of MGB neurons
[conrming previous ndings of spatial specicity in this species
(Yao et al., 2015b)], relevant suppression of responses by intervening streams, even when their lateralization was identical, was
limited to A1. In accordance with previous ndings (Wehr and
Zador, 2005), this reduction could not be explained by intracortical, GABAergic inhibition, which is why the authors suggest
that the thalamocortical synapse is its most likely origin (Yao et al.,
2015a). Insofar as ndings in anesthetized rats can be translated to
awake humans, these data support the rst explanation of our data
given above, that forward suppression in the context of streaming is
strongly enhanced in auditory cortex. Our data extend these ndings in that they suggest the same pattern for spatial and frequency
cues.
Contributions
C.H.U. and A.G. conceived the experiments. C.H.U. conducted the
experiments and analyzed the data. C.H.U., A.R.D., A.G. discussed
the data and wrote the manuscript.
Acknowledgements
Research supported by Deutsche Forschungsgemeinschaft (DFG,
grant GU593/3-2). All authors declare no conict of interest. We are
grateful to Sabine Heiland, Dorothea Willich, and Andreas Bartsch
for support with MRI acquisition. The authors declare no competing
nancial interests.
References
Anderson, L.A., Christianson, G.B., Linden, J.F., 2009. Stimulus-specic adaptation
occurs in the auditory thalamus. J. Neurosci. Off. J. Soc. Neurosci. 29,
7359e7363. http://dx.doi.org/10.1523/JNEUROSCI.0793-09.2009.
Anderson, L.A., Malmierca, M.S., 2012. The effect of auditory cortex deactivation on
stimulus-specic adaptation in the inferior colliculus of the rat. Eur. J. Neurosci.
http://dx.doi.org/10.1111/ejn.12018.
Antunes, F.M., Malmierca, M.S., 2011. Effect of auditory cortex deactivation on
stimulus-specic adaptation in the medial geniculate body. J. Neurosci. Off. J.
Soc. Neurosci. 31, 17306e17316. http://dx.doi.org/10.1523/JNEUROSCI.191511.2011.
Ayala, Y.A., Udeh, A., Dutta, K., Bishop, D., Malmierca, M.S., Oliver, D.L., 2015. Differences in the strength of cortical and brainstem inputs to SSA and non-SSA
neurons in the inferior colliculus. Sci. Rep. 5 http://dx.doi.org/10.1038/
srep10383.
uerle, P., von der Behrens, W., Ko
ssl, M., Gaese, B.H., 2011. Stimulus-specic
Ba
adaptation in the gerbil primary auditory thalamus is the result of a fast
frequency-specic habituation and is regulated by the corticofugal system.
J. Neurosci. Off. J. Soc. Neurosci. 31, 9708e9722. http://dx.doi.org/10.1523/
JNEUROSCI.5814-10.2011.
Bee, M.A., Klump, G.M., 2004. Primitive auditory stream segregation: a neurophysiological study in the songbird forebrain. J. Neurophysiol. 92, 1088e1104.
http://dx.doi.org/10.1152/jn.00884.2003.
Bee, M.A., Micheyl, C., Oxenham, A.J., Klump, G.M., 2010. Neural adaptation to tone
sequences in the songbird forebrain: patterns, determinants, and relation to the
build-up of auditory streaming. J. Comp. Physiol. A Neuroethol. Sens. Neural.
Behav. Physiol. 196, 543e557. http://dx.doi.org/10.1007/s00359-010-0542-4.
Boehnke, S.E., Phillips, D.P., 2005. The relation between auditory temporal interval
processing and sequential stream segregation examined with stimulus laterality differences. Percept. Psychophys. 67, 1088e1101.
Bregman, A.S., 1990. Auditory Scene Analysis: the Perceptual Organization of Sound.
The MIT Press, Massachusetts.
Brosch, M., Schreiner, C.E., 1997. Time course of forward masking tuning curves in
cat primary auditory cortex. J. Neurophysiol. 77, 923e943.
Brosch, M., Schulz, A., Scheich, H., 1999. Processing of sound sequences in Macaque
auditory cortex: response enhancement. J. Neurophysiol. 82, 1542e1559.
Cacciaglia, R., Escera, C., Slabu, L., Grimm, S., Sanju

an, A., Ventura-Campos, N.,


Avila,
C., 2015. Involvement of the human midbrain and thalamus in auditory
deviance detection. Neuropsychologia 68, 51e58. http://dx.doi.org/10.1016/
j.neuropsychologia.2015.01.001.

32

C.H. Uhlig et al. / Hearing Research 335 (2016) 25e32

Creutzfeldt, O., Hellweg, F.C., Schreiner, C., 1980. Thalamocortical transformation of

responses to complex auditory stimuli. Exp. Brain Res. 39, 87e104.
Dale, A.M., Fischl, B., Sereno, M.I., 1999. Cortical surface-based analysis. I. Segmentation and surface reconstruction. NeuroImage 9, 179e194. http://
dx.doi.org/10.1006/nimg.1998.0395.
Destrieux, C., Fischl, B., Dale, A., Halgren, E., 2010. Automatic parcellation of human
cortical gyri and sulci using standard anatomical nomenclature. NeuroImage 53,
1e15. http://dx.doi.org/10.1016/j.neuroimage.2010.06.010.

rez-Gonza

lez, D., Ayala, Y.A., Palmer, A.R., Malmierca, M.S., 2012.
Duque, D., Pe
Topographic distribution, frequency, and intensity dependence of stimulusspecic adaptation in the inferior colliculus of the rat. J. Neurosci. 32,
17762e17774. http://dx.doi.org/10.1523/JNEUROSCI.3190-12.2012.
Edmister, W.B., Talavage, T.M., Ledden, P.J., Weisskoff, R.M., 1999. Improved auditory
cortex imaging using clustered volume acquisitions. Hum. Brain Mapp. 7,
89e97, 10.1002/(SICI)1097-0193(1999)7:2<89::AID-HBM2>3.0.CO;2-N.
Fischl, B., Salat, D.H., Busa, E., Albert, M., Dieterich, M., Haselgrove, C., van der
Kouwe, A., Killiany, R., Kennedy, D., Klaveness, S., Montillo, A., Makris, N.,
Rosen, B., Dale, A.M., 2002. Whole brain segmentation: automated labeling of
neuroanatomical structures in the human brain. Neuron 33, 341e355.
Fischl, B., Sereno, M.I., Dale, A.M., 1999. Cortical surface-based analysis. II: ination,
attening, and a surface-based coordinate system. NeuroImage 9, 195e207.
http://dx.doi.org/10.1006/nimg.1998.0396.
Fishman, Y.I., Arezzo, J.C., Steinschneider, M., 2004. Auditory stream segregation in
monkey auditory cortex: effects of frequency separation, presentation rate, and
tone duration. J. Acoust. Soc. Am. 116, 1656e1670. http://dx.doi.org/10.1121/
1.1778903.
Fishman, Y.I., Reser, D.H., Arezzo, J.C., Steinschneider, M., 2001. Neural correlates of
auditory stream segregation in primary auditory cortex of the awake monkey.
Hear. Res. 151, 167e187. http://dx.doi.org/10.1016/S0378-5955(00)00224-0.
Gao, P.P., Zhang, J.W., Cheng, J.S., Zhou, I.Y., Wu, E.X., 2014. The inferior colliculus is
involved in deviant sound detection as revealed by BOLD fMRI. NeuroImage 91,
220e227. http://dx.doi.org/10.1016/j.neuroimage.2014.01.043.
Grimault, N., Bacon, S.P., Micheyl, C., 2002. Auditory stream segregation on the basis
of amplitude-modulation rate. J. Acoust. Soc. Am. 111, 1340e1348.
Guimaraes, A.R., Melcher, J.R., Talavage, T.M., Baker, J.R., Ledden, P., Rosen, B.R.,
Kiang, N.Y.S., Fullerton, B.C., Weisskoff, R.M., 1998. Imaging subcortical auditory
activity in humans. Hum. Brain Mapp. 6, 33e41.
Gutschalk, A., Dykstra, A.R., 2014. Functional imaging of auditory scene analysis.
Hear. Res. 307, 98e110. http://dx.doi.org/10.1016/j.heares.2013.08.003.
Gutschalk, A., Micheyl, C., Melcher, J.R., Rupp, A., Scherg, M., Oxenham, A.J., 2005.
Neuromagnetic correlates of streaming in human auditory cortex. J. Neurosci.
25, 5382e5388. http://dx.doi.org/10.1523/JNEUROSCI.0347-05.2005.
Gutschalk, A., Oxenham, A.J., Micheyl, C., Wilson, E.C., Melcher, J.R., 2007. Human
cortical activity during streaming without spectral cues suggests a general
neural substrate for auditory stream segregation. J. Neurosci. Off. J. Soc. Neurosci. 27, 13074e13081. http://dx.doi.org/10.1523/JNEUROSCI.2299-07.2007.
Gutschalk, A., Steinmann, I., 2015. Stimulus dependence of contralateral dominance
in human auditory cortex. Hum. Brain Mapp. 36, 883e896. http://dx.doi.org/
10.1002/hbm.22673.
Hagler Jr., D.J., Saygin, A.P., Sereno, M.I., 2006. Smoothing and cluster thresholding
for cortical surface-based group analysis of fMRI data. NeuroImage 33,
1093e1103. http://dx.doi.org/10.1016/j.neuroimage.2006.07.036.
Hall, D.A., Haggard, M.P., Akeroyd, M.A., Palmer, A.R., Summereld, A.Q., Elliott, M.R.,
Gurney, E.M., Bowtell, R.W., 1999. Sparse temporal sampling in auditory fMRI.
Hum. Brain Mapp. 7, 213e223.
Harms, M.P., Guinan, J.J., Sigalovsky, I.S., Melcher, J.R., 2005. Short-term sound
temporal envelope characteristics determine multisecond time patterns of activity in human auditory cortex as shown by fMRI. J. Neurophysiol. 93, 210e222.
http://dx.doi.org/10.1152/jn.00712.2004.
Harms, M.P., Melcher, J.R., 2002. Sound repetition rate in the human auditory
pathway: representations in the waveshape and amplitude of fMRI activation.
J. Neurophysiol. 88, 1433e1450.
Hawley, M.L., Melcher, J.R., Fullerton, B.C., 2005. Effects of sound bandwidth on fMRI
activation in human auditory brainstem nuclei. Hear. Res. 204, 101e110. http://
dx.doi.org/10.1016/j.heares.2005.01.005.
Kondo, H.M., Kashino, M., 2009. Involvement of the thalamocortical loop in the
spontaneous switching of percepts in auditory streaming. J. Neurosci. Off. J. Soc.
Neurosci. 29, 12695e12701. http://dx.doi.org/10.1523/JNEUROSCI.1549-09.2009.
Logothetis, N.K., Pauls, J., Augath, M., Trinath, T., Oeltermann, A., 2001. Neurophysiological investigation of the basis of the fMRI signal. Nature 412, 150e157.
http://dx.doi.org/10.1038/35084005.

rez-Gonza

lez, D., Covey, E., 2009. Stimulus-specic
Malmierca, M.S., Cristaudo, S., Pe
adaptation in the inferior colliculus of the anesthetized rat. J. Neurosci. Off. J.
Soc. Neurosci. 29, 5483e5493. http://dx.doi.org/10.1523/JNEUROSCI.415308.2009.
Malone, B.J., Semple, M.N., 2001. Effects of auditory stimulus context on the representation of frequency in the gerbil inferior colliculus. J. Neurophysiol. 86,
1113e1130.
Micheyl, C., Carlyon, R.P., Gutschalk, A., Melcher, J.R., Oxenham, A.J.,
Rauschecker, J.P., Tian, B., Courtenay Wilson, E., 2007. The role of auditory cortex
in the formation of auditory streams. Hear. Res. 229, 116e131. http://dx.doi.org/
10.1016/j.heares.2007.01.007.

Micheyl, C., Tian, B., Carlyon, R.P., Rauschecker, J.P., 2005. Perceptual organization of
tone sequences in the auditory cortex of awake macaques. Neuron 48, 139e148.
http://dx.doi.org/10.1016/j.neuron.2005.08.039.
Middlebrooks, J.C., Bremen, P., 2013. Spatial stream segregation by auditory cortical
neurons. J. Neurosci. Off. J. Soc. Neurosci. 33, 10986e11001. http://dx.doi.org/
10.1523/JNEUROSCI.1065-13.2013.
Miller, G.A., Heise, G.A., 1950. The trill threshold. J. Acoust. Soc. Am. 22, 637e638.
http://dx.doi.org/10.1121/1.1906663.
Nelson, P.C., Smith, Z.M., Young, E.D., 2009. Wide-dynamic-range forward suppression in marmoset inferior colliculus neurons is generated centrally and
accounts for perceptual masking. J. Neurosci. Off. J. Soc. Neurosci. 29,
2553e2562. http://dx.doi.org/10.1523/JNEUROSCI.5359-08.2009.
Overath, T., Zhang, Y., Sanes, D.H., Poeppel, D., 2012. Sensitivity to temporal modulation rate and spectral bandwidth in the human auditory system: fMRI evidence.
J.
Neurophysiol. 107,
2042e2056.
http://dx.doi.org/10.1152/
jn.00308.2011.
Patel, C.R., Redhead, C., Cervi, A.L., Zhang, H., 2012. Neural sensitivity to novel
sounds in the rat's dorsal cortex of the inferior colliculus as revealed by evoked
local eld potentials. Hear. Res. 286, 41e54. http://dx.doi.org/10.1016/
j.heares.2012.02.007.
Paxinos, G., Huang, X.-F., Sengul, G., Watson, C., 2012. Organization of brainstem
nuclei. In: The Human Nervous System. Elsevier Academic, Amsterdam,
pp. 260e327.

rez-Gonza

lez, D., Malmierca, M.S., 2014. Adaptation in the auditory system: an
Pe
overview. Front. Integr. Neurosci. 8, 19. http://dx.doi.org/10.3389/
fnint.2014.00019.

rez-Gonza

lez, D., Malmierca, M.S., Covey, E., 2005. Novelty detector neurons in
Pe
the mammalian auditory midbrain. Eur. J. Neurosci. 22, 2879e2885. http://
dx.doi.org/10.1111/j.1460-9568.2005.04472.x.
Pressnitzer, D., Sayles, M., Micheyl, C., Winter, I.M., 2008. Perceptual organization of
sound begins in the auditory periphery. Curr. Biol. CB 18, 1124e1128. http://
dx.doi.org/10.1016/j.cub.2008.06.053.
R Core Team, 2015. R: a Language and Environment for Statistical Computing. R
Foundation for Statistical Computing, Vienna, Austria.
Schadwinkel, S., Gutschalk, A., 2011. Transient bold activity locked to perceptual
reversals of auditory streaming in human auditory cortex and inferior colliculus. J. Neurophysiol. 105, 1977e1983. http://dx.doi.org/10.1152/jn.00461.2010.
Schadwinkel, S., Gutschalk, A., 2010. Functional dissociation of transient and sustained fMRI BOLD components in human auditory cortex revealed with a
streaming paradigm based on interaural time differences. Eur. J. Neurosci. 32,
1970e1978. http://dx.doi.org/10.1111/j.1460-9568.2010.07459.x.
Scholes, C., Palmer, A.R., Sumner, C.J., 2015. Stream segregation in the anesthetized
auditory cortex. Hear. Res. 328, 48e58. http://dx.doi.org/10.1016/
j.heares.2015.07.004.
Scholes, C., Palmer, A.R., Sumner, C.J., 2011. Forward suppression in the auditory
cortex is frequency-specic. Eur. J. Neurosci. 33, 1240e1251. http://dx.doi.org/
10.1111/j.1460-9568.2010.07568.x.
Sigalovsky, I.S., Melcher, J.R., 2006. Effects of sound level on fMRI activation in
human brainstem, thalamic and cortical centers. Hear. Res. 215, 67e76. http://
dx.doi.org/10.1016/j.heares.2006.03.002.
Snyder, J.S., Alain, C., Picton, T.W., 2006. Effects of attention on neuroelectric correlates of auditory stream segregation. J. Cogn. Neurosci. 18, 1e13. http://
dx.doi.org/10.1162/089892906775250021.
Steinmann, I., Gutschalk, A., 2012. Sustained BOLD and theta activity in auditory
cortex are related to slow stimulus uctuations rather than to pitch.
J. Neurophysiol. 107, 3458e3467. http://dx.doi.org/10.1152/jn.01105.2011.
Ulanovsky, N., Las, L., Nelken, I., 2003. Processing of low-probability sounds by
cortical neurons. Nat. Neurosci. 6, 391e398. http://dx.doi.org/10.1038/nn1032.
van Noorden, L.P.A.S., 1975. Temporal Coherence in the Perception of Tone Sequences. Technical University Eindhoven, The Netherlands. The Netherlands,
Technical University Eindhoven.
Vliegen, J., Moore, B.C., Oxenham, A.J., 1999. The role of spectral and periodicity cues
in auditory stream segregation, measured using a temporal discrimination task.
J. Acoust. Soc. Am. 106, 938e945.
Wehr, M., Zador, A.M., 2005. Synaptic mechanisms of forward suppression in rat
auditory
cortex.
Neuron
47,
437e445.
http://dx.doi.org/10.1016/
j.neuron.2005.06.009.
Wilson, E.C., Melcher, J.R., Micheyl, C., Gutschalk, A., Oxenham, A.J., 2007. Cortical
FMRI activation to sequences of tones alternating in frequency: relationship to
perceived rate and streaming. J. Neurophysiol. 97, 2230e2238. http://
dx.doi.org/10.1152/jn.00788.2006.
Yao, J.D., Bremen, P., Middlebrooks, J.C., 2015a. Emergence of spatial stream
segregation in the ascending auditory pathway. J. Neurosci. Off. J. Soc. Neurosci.
35, 16199e16212. http://dx.doi.org/10.1523/JNEUROSCI.3116-15.2015.
Yao, J.D., Bremen, P., Middlebrooks, J.C., 2015b. Transformation of spatial sensitivity
along the ascending auditory pathway. J. Neurophysiol. 113, 3098e3111. http://
dx.doi.org/10.1152/jn.01029.2014.
Yu, X.-J., Xu, X.-X., He, S., He, J., 2009. Change detection by thalamic reticular
neurons. Nat. Neurosci. 12, 1165e1170. http://dx.doi.org/10.1038/nn.2373.
Zhao, L., Liu, Y., Shen, L., Feng, L., Hong, B., 2011. Stimulus-specic adaptation and its
dynamics in the inferior colliculus of rat. Neuroscience 181, 163e174. http://
dx.doi.org/10.1016/j.neuroscience.2011.01.060.