Serial and Parallel Processing in
the Human Auditory Cortex: A
Koji Inui, Hidehiko Okamoto, Kensaku Miki, Atsuko Gunji and
Department of Integrative Physiology, National Institute for
Physiological Sciences, Okazaki 444-8585, Japan
Although anatomical, histochemical and electrophysiological find-
ings in both animals and humans have suggested a parallel and
serial mode of auditory processing, precise activation timings of
each cortical area are not well known, especially in humans. We
investigated the timing of arrival of signals to multiple cortical
areas using magnetoencephalography in humans. Following click
stimuli applied to the left ear, activations were found in six cortical
areas in the right hemisphere: the posteromedial part of Heschl’s
gyrus (HG) corresponding to the primary auditory cortex (PAC), the
anterolateral part of the HG region on or posterior to the transverse
sulcus, the posterior parietal cortex (PPC), posterior and anterior
parts of the superior temporal gyrus (STG), and the planum
temporale (PT). The mean onset latencies of each cortical activity
were 17.1, 21.2, 25.3, 26.2, 30.9 and 47.6 ms respectively. These
results suggested a serial model of auditory processing along the
medio-lateral axis of the supratemporal plane and, in addition,
implied the existence of several parallel streams running postero-
superiorly (from the PAC to the belt region and then to the posterior
STG, PPC or PT) and anteriorly (PAC--belt--anterior STG).
Keywords: auditory cortex, auditory evoked response,
In primates, auditory cortical areas are usually divided into three
regions — the core, belt and parabelt — based on their
cochleotopic organization, anatomical connections and archi-
tectonic features (for review, see Kaas et al., 1999). The core
field has histological features of the primary sensory cortex
(Galaburda and Pandya, 1983; Morel et al., 1993; Jones et al.,
1995; Hackett et al., 1998), and receives dense projections from
the ventral subdivision of the medial geniculate nucleus
complex (MGv) (Burton and Jones, 1976; Morel et al., 1993).
The belt region, which surrounds the core, does not exhibit the
histological features of the primary sensory cortex, and receives
main inputs from the core region with a minor contribution
from MGv (Galaburda and Pandya, 1983; Jones et al., 1995;
Hackett et al., 1998). The parabelt region, which surrounds the
belt, has strong connections with the belt region and minimal
connections with the core (Hackett et al., 1998). These findings
suggest that auditory information is processed serially through
the core, belt and parabelt. In addition, there are several
subdivisions in each of the three regions and a subregion tends
to receive main inputs from its adjacent area, for example, the
caudal parabelt receives main inputs from the caudal belt areas,
while the rostral parabelt receives inputs selectively from
the rostral belt fields (Hackett et al., 1998). These patterns
of connection indicate that there are several parallel streams
of auditory processing. In addition to these anatomical and
histochemical studies, there are some electrophysiological
studies that support such modes of auditory processing
(Rauschecker et al., 1997) or functional segregations of neurons
in each cortical zone (Kosaki et al., 1997). However, only a few
studies have compared the response latency of neurons in
different cortical fields of monkeys (Benson and Hienz, 1978;
Pfingst and O’Connor, 1981; Vaadia et al., 1982; Kajikawa et al.,
2005) and other mammals (Kowalski et al., 1995; Eggermont,
In humans, less is known about the processing mode of
auditory information because of limitations of experimental
methods. However, in general, data from humans suggest
a hierarchical and parallel auditory processing similar to that
found in monkeys. Cytoarchitectonic studies have distinguished
several human auditory areas (Galaburda and Sanides, 1980)
and, furthermore, have proposed a hierarchical and parallel
organization within them based on the systematic change of
histochemical features of each area (Mesulam and Geula, 1994;
Hutsler and Gazzaniga, 1996; Rivier and Clarke, 1997). The
distributions of response latencies in auditory evoked potential
(AEP) and magnetic field (AEF) studies are roughly consistent
with serial processing along the medio-lateral axis of the
supratemporal plane (Celesia, 1976; Lie ´ geois-Chauvel et al.,
1994; Gutschalk et al., 1999; Howard et al., 2000; Yvert et al.,
2001). However, precise temporal relationships among activi-
ties in multiple cortical areas are not well understood.
In a previous study using magnetoencephalography (MEG),
we showed that tactile signals are mainly processed in two
parallel pathways, each of which processes information in
a serial fashion: one through areas 3b, 1 and 5 (posterior parietal
cortex, PPC) and another through area 3b and/or area 1 and the
secondary somatosensory cortex (Inui et al., 2004), which is
consistent with anatomical and electrophysiological findings in
monkeys showing serial somatosensory processing along the
postcentral gyrus and through primary (SI) and secondary (SII)
somatosensory cortices (for review, see Iwamura, 1998). In the
present study, we sought to clarify the precise activation timing
of auditory cortical areas using similar methods and to find some
streams of auditory processing.
Materials and Methods
The experiment was performed on ten (two female and eight male)
healthy right-handed volunteers, aged 25--51 years (mean 33.5 ± 7.4).
Their hearing thresholds were below 15 nHL in the frequency range
from 250 to 8000 Hz as tested by means of pure tone audiometry. The
study was approved in advance by the Ethical Committee of the National
Institute for Physiological Sciences and written consent was obtained
from all the subjects.
The experiments were carried out in a quiet, magnetically shielded
room. The subjects lay in a right lateral position, and their heads were
? The Author 2005. Published by Oxford University Press. All rights reserved.
For permissions, please e-mail: email@example.com
Cerebral Cortex Advance Access published March 30, 2005
by guest on June 1, 2013
fixed to the device with adhesive tape to minimize head movements.
Clicks with a duration of 0.3 ms were created by a speaker placed
outsidetheroomand presented to thesubjects’left earthrougha plastic
tube 5 m in length and ear-pieces (E-A-Rtone 3A, Aero Company,
Indianapolis, IN). The intensity of the clicks was adjusted to 60 dB above
the threshold for each subject before the session. Clicks were delivered
at a random rate of between 2.7 and 3.3 Hz. While the subjects watched
silent movies, evoked magnetic fields were recorded with a 37-channel
biomagnetometer (Magnes, Biomagnetic Technologies, San Diego, CA)
as described previously (Inui et al., 2004). The magnetic fields were
recorded with a filter of 0.1--200 Hz at a sampling rate of 2083 Hz, and
then digitally filtered with a high-cut filter of 150 Hz. The window of
analysis was from 40 ms before to 120 ms after the stimulus onset, and
the pre-stimulus period was used as the DC baseline. In one trial, 1500
responses were collected and averaged. Two trials were obtained with
intervals of a few minutes. After the reproducibility had been confirmed,
the two trials were averaged and used for the analysis.
Source locations and time courses of source activities were de-
termined by a multiple source analysis, brain electric source analysis
(NeuroScan, Mclean, VA), as described previously (Inui et al., 2004).
Model adequacy was assessed by examining: (i) percent variance (Hari
et al., 1988); (ii) F-ratio (ratio of reduced chi-square values before and
after adding a new source) (Supek and Aine, 1993); and (iii) residual
waveforms (i.e. the difference between the recorded data and the
model). Percent variance measures the goodness-of-fit (GOF) of the
model comparing the recorded data and the model. The integral
probability of obtaining a F-ratio value equal to or greater than the
obtained value is calculated to evaluate whether a model with a larger
number of dipoles represents a statistically significant improvement of
<0.05,we consideredthe newdipole assignificant.Wecontinued toadd
a source to the model until the addition of a dipole did not significantly
improve the fit. The procedure used to assess the model’s accuracy was
basically the same as described elsewhere (Inui et al., 2004).
Magnetic resonance imaging (MRI) scans (Shimadzu, Kyoto, Japan,
150XT 1.5 T) were obtained from all subjects. T1-weighted coronal, axial
and sagittal image slices obtained every 1.5 mm were used for
superimposition of the MEG source locations. The same anatomical
landmarks used to create the MEG head-based three-dimensional (3D)
coordinate system (the bilateral pre-auricular points and nasion) were
visualized in the MR images by affixing to these points high-contrast cod
liver oil capsules (3 mm in diameter). The common MEG and MRI
anatomical landmarks allowed easy transformation of the head-based 3D
coordinate system used for MEG source analyses into the MRI co-
ordinate. The origin of the head-based coordinate system was the mid-
point between the pre-auricular points. The x-axis indicated the coronal
plane with a positive value toward the anterior direction, the y-axis
indicated the mid-sagittal plane with a positive value toward the left pre-
auricular point, and the z-axis indicated the transverse plane pre-
auricular to the x--y plane with a positive value toward the upper side.
To reduce the interindividual variability of the location and extension
of the auditory cortical area, we employed a common internal landmark.
The landmark was the crown of the posteriormost part of Heschl’s gyrus
(HG, point A), which was identified in the most medial sagittal slice
showing the HG clearly (Fig. 1) with the aid of the two other orthogonal
slices. When the HG was duplicated (Leonard et al., 1998), the most
anterior gyrus was taken. The coordinates of the landmark were
subtracted from those of the estimated sources, and the difference
coordinates were used to express the location of each dipole. To
examine the location of each dipole relative to the HG, we also
identified the anteriormost part of the HG (point B) in each subject
(Steinmetz et al., 1989). The HG usually flattens toward its lateral end on
the bulging rim of the superior temporal gyrus (STG). When a sagittal
slice moves laterally, the HG merges with the STG around its lateral end.
Point B was identified in the most lateral sagittal slice that still showed
the HG clearly. Usually we had to browse two other slices simulta-
neously to ensure a correct identification of point B. Individual head
system x,y,z coordinates for point A were 9.1 ± 7.0 (range, –5 to 16),
36.8 ± 1.5 (35--39) and 63.6 ± 6.0 (54--69) mm respectively. Respective
coordinates for point B were 26.1 ± 8.5 (6--35), 56.0 ± 2.0 (52--59) and
49.8 ± 4.2 (42--56) mm.
Waveforms and Topographies
In all subjects, two clear components around 30 and 50 ms after
the stimulation were identified (Fig. 2) by the time course of
the root mean square (RMS) and were termed 1M and 2M
respectively. To examine the reproducibility of the response
between two sessions, the peak latency, peak amplitude (peak
RMS) and field distribution patterns for 1M and 2M were
compared between the first and second session. The field
distribution was compared using a correlation coefficient, r.
A two-way analysis of variance (session 3 component) showed
no significant effect of the session on either peak latency (F =
0.48, P = 0.50) or peak amplitude (F = 0.022, P = 0.88). The mean
correlation coefficient r was 0.98 (0.95--0.99) and 0.97
(0.93--0.99) for 1M and 2M respectively. These results indicated
that the first and second 1500 click stimuli evoked magnetic
responses with a similar latency, amplitude and topography.
Therefore, trials of both sessions were averaged and the
averaged waveform was used for the analyses.
The peak latency (peak RMS) of 1M was 32.6 ms on average,
ranging from 29.3 to 35.0 ms. The peak latency of 2M was 53.0
ms on average, ranging from 49.4 to 57.1 ms. Both components
showed a dipolar pattern of field distributions consistent with
sources pointing antero-superiorly (Fig. 2). In six out of ten
subjects, the topography at ~40 ms showed a dipolar pattern
that was clearly different from that of 1M or 2M (see waveform
and topography at 39 ms in subject 2, Fig. 2). That is, the
isocontour map at this latency indicated a source in the parietal
region pointing posteriorly. After the 2M component, a topo-
graphy consistent with a source pointing superiorly with a more
superficial location than that for 1M or 2M was identified in
seven subjects at 55--85 ms. Similar topographies were some-
times identified at a latency of 40--50 ms. In five subjects,
a topography at 70--100 ms showed a dipolar pattern pointing
postero-inferiorly (see topography at 90 ms in subject 1, Fig. 2).
This topography indicated a source with a slightly posterior
location than that for 1M.
Figure 1. Landmarks A and B. The most medial sagittal MRI slice with a clear
Heschl’s gyrus (HG) for identification of point A (upper panels). The most lateral slice
with a clear HG for identification of point B. HS, Heschl sulcus; STG, superior temporal
gyrus; STS, superior temporal sulcus.
Page 2 of 13
Auditory Processing in Humans
Inui et al.
by guest on June 1, 2013
Two Sources during 1M
These topographies indicated that at least several distinct
sources were active during the period of analysis. Therefore,
we analyzed the data using a multiple source method to
differentiate each cortical activity. We started the analysis
with 1M. As shown in Figure 3A, the maximum influx and
outflux activities differed in latency, indicating that two or more
sources contributed to shape the 1M component. Figure 3B
shows that the waveforms recorded from 37 sensors did not
show a simple dipolar pattern, but showed peaks at different
latencies. For example, the outflux of the sensors 1, 7, 17 and 18
peaked slightly later than the peak of 1M, while that of sensors
20, 21 and 22 peaked slightly earlier than the 1M peak. When
a single dipole analysis was applied to 1M, the best GOF value
was obtained at 27 ms, 2 ms earlier than the peak of 1M in this
case (Fig. 3C). This finding was constant across subjects (Fig.
3D) and the best GOF value was obtained 3 ms earlier than the
peak of 1M on average. This finding indicated that one or more
sources were active at least at the peak of 1M in addition to the
first source. Therefore, we tried to separate these sources
during the period of 1M. The procedures and results of the
analysis in a representative case are shown in Figure 4. When
a single dipole analysis was applied to the peak of 1M (29 ms),
the dipole was estimated to be located in the deep area of the
Sylvian fissure (dipole A in Fig. 4Ca). Waveforms in Figure 4Ab
show the residual magnetic fields obtained by subtraction of the
theoretical magnetic fields due to this dipole from the original
data (Fig. 4Aa). The residual waveform shows weak but clear
activities in both the early and late parts of 1M (shown by
arrowheads in Fig. 4Ab). When the best dipole was calculated at
a latency point slightly earlier than the peak of 1M, the dipole
(dipole B) was located slightly more posterior, medial and
superior than dipole A. The residual waveform obtained by the
subtraction of magnetic fields due to dipole B from the original
record (Fig. 4Ac) shows that this dipole could explain the early
part of the 1M component well, while leaving substantial
activities during the late part of 1M unexplained. On the other
hand, when the best dipole was obtained at a latency point
slightly later than the peak of 1M, this dipole failed to explain
the early part of 1M (Fig. 4Ad). This dipole (dipole C) was
estimated to be located slightly more anterior and inferior than
dipole A. Therefore, the dipole location became more anterior,
lateral and inferior with an increase of the latency. These results
suggested that no single dipole could successfully explain the
overall 1M activities as expected. Therefore, at first, we tested
how a two-dipole model improved the fit during the latency
Figure 2. Evoked magnetic fields and topographies. Data from two subjects. Upper trace, superimposed waveform recorded from 37 channels. Lower trace, isocontour maps at
several latency points.
Cerebral Cortex Page 3 of 13
by guest on June 1, 2013
period of 1M. To obtain the best model, we started the analysis
with a tentative source (source 1) estimated at the latency point
where the highest GOF value was obtained (3 ms earlier than
the 1M peak on average). Then, a second source (source 2) was
calculated at the latency point where the residual activities
were largest in the presence of source 1. After the location and
orientation of the two dipoles were slightly adjusted to provide
the best fit during this period, we obtained the best two-dipole
model. Figure 4Ba shows the residual magnetic fields that could
not be explained by the two-dipole model (Fig. 4Cb). As
compared with each single dipole model (dipole A, B and C),
this two-dipole model significantly improved the fit. For
example, as compared with the dipole A model, the two-dipole
model increased the GOF value from 96.7 to 99.4% (F ratio =
4.63, P <0.0001) at 25 ms, and from 97.5 to 99.1% (F ratio = 2.45,
P = 0.01) at 32 ms. Figure 4Bb shows the time course of the two
source activities. Their time courses are very similar with a 5 ms
time delay for the second source. The activity of source 1
peaked at 27.4 ms and that of source 2 at 33.1 ms, supporting
the view that at least two source activities overlap at the peak
latency of 1M (29.3 ms). In this case, the activity of the second
source is relatively weak at the peak latency of the first source
(27.4 ms), which well explains why the largest GOF value was
obtained by a single-dipole model at 27.4 ms (Fig. 3C) not at
the peak latency of 1M. Since source 1 was located in the
postero-medial part of Heschl’s gyrus (HG) and source 2 in the
antero-lateral part of the HG region (Fig. 6), we refer to these
sources by the anatomical name, HG-M and HG-L, respectively
in the text.
Sources that Are Active Later than 1M
The location and orientation of the dipoles were fixed, and we
continued the analysis to include the best third source in the
model. Figure 5Ab shows residual magnetic fields obtained by
Figure 3. Analysis of the first MEG component, 1M, with a single dipole model. (A)
Superimposed waveform. (B) Waveforms recorded from each of the 37 channels. A
dotted line indicates the stimulus onset. A thin line indicates the peak latency (peak
root mean square, RMS) of 1M. (C, D) Relationship between the RMS and GOF value
obtained by a single dipole analysis in a single subject (C) and the group average
across all subjects (D). Vertical bars in D indicate ± SE. GOF, goodness of fit.
Figure 4. Two overlapping sources during 1M. (A) Recorded waveform (a) and
residual waveforms obtained by subtraction of theoretical waveforms due to dipole A
(b), dipole B (c) and dipole C (d) from the original data. Dipoles A, B and C were
obtained by a single dipole analysis at the peak of 1M, and 3 ms earlier and 3 ms later
than the 1M peak respectively (shown by vertical bars). The waveform in Ab shows
the residual magnetic fields that could not be explained by dipole A. There are weak
but clear magnetic fields in both early and late parts of 1M shown by arrowheads,
indicating that dipole A could not explain overall 1M activities. Likewise, both dipole B
and dipole C left some magnetic fields unexplained (Ac and Ad). (Ba) Residual
magnetic fields after subtractionof the theoreticalmagnetic fields due to the best two-
dipole model (source 1 and source 2). (Bb) Time course of source strength of source 1
and source 2. (C) Schematic presentation of locations of each dipole. (Ca) Dipole A, B
and C. (Cb) Source 1 and 2.
Page 4 of 13
Auditory Processing in Humans
Inui et al.
by guest on June 1, 2013
a subtraction of the magnetic fields due to the two-dipole model
from the original data. There are clear residual magnetic fields
around 40 ms. To explain the residual magnetic fields, the best
source was estimated to be located posterior to the postcentral
gyrus in the posterior parietal cortex (PPC). With the addition
of this source, the GOF value at 39.8 ms increased markedly
from 38.7 to 98.0%. Based on the residual magnetic fields that
could not be explained by the three-dipole model (Fig. 5Ac), the
best fourth source was estimated to be located around the
crown of the posterior part of the superior temporal gyrus
(STG-P). By adding this source, the GOF value at 51 ms
increased from 86.6 to 98.4% (F ratio = 8.46, P < 0.0001).
Finally, the fifth source to explain the residual magnetic fields
(Fig. 5Ad) was estimated to be located around the anterior part
of the STG (STG-A). The GOF value at 62.9 ms increased from
90.9to98.7% (Fratio = 3.78, P = 0.0064) with the addition of this
source. After the fitting of these five sources, the mean GOF
value for all the data (0--120 ms) was 98.7% and no additional
source significantly improved the fit. Figure 5B shows time
courses of each source strength and they were used for
the analysis of the latency of each activity. Figure 5C shows
the location and orientation of each source. Figure 6 shows the
location of each source superimposed on MR images.
Location of Each Source
Similar procedures were applied to data obtained for the
remaining nine subjects. By applying our criteria, two distinct
sources responsible for the 1M component were identified in
nine of the ten subjects. The mean x,y,z coordinates relative to
point A of each source are shown in Table 1. The mean location
of the HG-M source was 2.9 mm anterior, 6.4 mm lateral and
3.7 mm inferior to point A, which fell in the postero-medial part
of the HG. In the present study, five subjects had one HG (single
HG) and four subjects showed two HGs of a common stem type
(duplicated HG, Leonard et al., 1998). Among the five subjects
who had a single HG, the source location of the HG-M was
coextensive with the HG in four and was on the first Heschl
sulcus (HS1) in one. The HG-M source was located in the first
HG (H1) in all four subjects who had a duplicated HG. The
second source, HG-L, was located more anterior (3.9 mm),
lateral (8.9 mm) and inferior (5.2 mm) than the HG-M source.
Among the five subjects with a single HG, the HG-L source was
located in the HG in one, on the HS1 in two and posterior to the
HS1 in two. Among the four subjects with a duplicated HG, the
HG-L source was in the second HG (H2) in three and posterior
to the second Heschl sulcus (HS2) in one. Figure 7 shows the
location of the HG-M and HG-L sources of a representative case
with a duplicated HG.
The third source, PPC, was identified in seven subjects, and
was located on or slightly posterior to the postcentral sulcus.
The next two sources were located in the lateral aspect of the
STG, but were clearly separated by their anterior--posterior and
superior--inferior coordinates. That is, one (STG-P, six subjects)
was located in the postero-superior part of the STG, the other
(STG-A, six subjects) in the antero-inferior part of the STG. As
shown in Figure 8, the STG-P source was located around the
superior crown of the STG or, in some subjects, around the
inferior crown of the supramarginal gyrus. The STG-A source
tended to be located in the upper bank of the superior temporal
sulcus (STS). The latest source activity was identified in seven
subjects, which originated from an area slightly posterior to
point A (2.7 mm on average), corresponding to the planum
temporale (PT). Figure 9B shows schematic drawings of the
mean location and orientation of each source. Figure 9C shows
the locations of each source relative to points A and B of all
Figure 5. Multi-dipole analysis of magnetic fields later than 1M. (Note the different
time scales in Figures 4 and 5.) (A) Waveforms of recorded magnetic fields (a) and
residual magnetic fields obtained by subtraction of theoretical magnetic fields due
to the model in each step from the recorded magnetic fields (b--e). (B) Time course
of each source activity. (C) Schematic drawings of the location and orientation of
each source. HG, Heschl’s gyrus; PPC, posterior parietal cortex; STG, superior
Cerebral Cortex Page 5 of 13
by guest on June 1, 2013
Time Course of Each Cortical Activity
Figure 9A shows the time course of each source strength of all
subjects (blue lines) and their group-average (black lines). In
general, the activity in the HG-M, HG-L and PPC reversed its
polarity twice with an interval of ~10 ms, which resulted in
a triphasic waveform. On the other hand, the STG-A and PT
sources showed long-lasting (usually >50 ms) activities. STG-P
source activities were identified in six subjects. Three subjects
showed the triphasic waveform (Fig. 10), although the early two
components did not significantly increase the fit according to
our criteria in two subjects. The onset and peak latencies of
each source activity are shown in Table 1. The onset latency
of the HG-M source was the shortest (17.1 ms) followed by that
of the HG-L (21.2 ms), PPC (25.3 ms), STG-A (30.9 ms), STG-P
(39.8 ms) and PT (47.6 ms). Similarly, the peak latency became
longer in this order (28.0, 33.5, 38.2, 48.3, 55.3 and 76.2 ms).
If we accept the early component of the STG-P activity in
three subjects (Fig. 10), its mean onset latency was 26.2 ms,
which was 4.0 ms longer than that for the HG-L activity, and
its peak latency was 34.9 ms, which was longer by 2.9 ms than
that for the HG-L source in these subjects. The peak latency of
1M (32.6 ms) was between the peak latencies of the first
Figure 6. Locations of each cortical source in Figure 5 superimposed on the subject’s MR images.
Page 6 of 13
Auditory Processing in Humans
Inui et al.
by guest on June 1, 2013
component of the HG-M (28.0 ms) and HG-L (33.5 ms)
activities. The peak latency of 2M (53.0 ms) was between the
peak latencies of the third component of the HG-M (49.6 ms)
and HG-L (55.1 ms) activities.
The present MEG study revealed the precise timing of arrival of
signals to multiple auditory cortical areas and the time course of
each cortical activity. The shift of the response latency along the
medio-lateral axis of the supratemporal plane is consistent with
the hierarchical auditory processing in monkeys through the
core, belt and parabelt regions on the supratemporal plane (for
review, see Kaas et al., 1999). There are several possible
pathways other than cortico-cortical connections for activating
each cortical source such as thalamo-cortical, cortico-thalamo-
cortical and interhemispheric connections. Therefore, the
latency shift among cortical sources might be due to the
different conduction time of the pathway between the thalamus
and each cortical area or due to different kinds of pathways (for
discussions, see Brugge et al., 2003). However in animals, it has
been shown that the latency from the thalamus to a cortical cell
is remarkably constant across multiple cortical areas, irrespec-
tiveof the variability of traveling distance (Salami etal., 2003).In
addition, the present results that early cortical responses
exhibited a similar triphasic time course suggest that they
reflect a similar intra-laminar processing, probably ‘feedforward’
pattern processing, which supports that the early cortical
responses were driven mainly by successive cortico-cortical
Successive activities found in this study are summarized as
? 28 ms (peak latency of the first component, source HG-M):
activity originating from the postero-medial part of the HG.
? 33 ms (HG-L): activity in a region antero-lateral to the HG-M
source, just posterior to the first transverse sulcus of Heschl
(HS1) or intermediate sulcus.
? 38 ms (PPC): activity in the posterior parietal cortex.
? 35--50 ms (STG-P): activity in the posterior part of the STG
around its superior crown.
? 48 ms (STG-A): activity in the anterior part of the STG,
around the upper bank of the STS.
? 76 ms (PT): activity in a region posterior and slightly superior
to the HG-M source, in the planum temporale.
Activation In and Around the HG
We found two separate early activations in and near the HG,
HG-M and HG-L. The HG-M source was located in the dorso-
postero-medial part of the HG, which approximately corre-
sponds to the location of the primary auditory cortex (PAC)
identified by several techniques. First, this location appears to
correspond to cytoarchitectonically defined koniocortical
areas, KAm (Galaburda and Sanides, 1980), AI (Rivier and Clarke,
1997) and Te1.0 (Morosan et al., 2001) in humans, and KA
(Galaburda and Pandya, 1983) and AI (Morel et al., 1993) in
monkeys. In addition, the scatter plots of the location of the
HG-M source in Figure 9C are approximately coextensive to
a high probability area of the probabilistic map of the PAC
drawn by Rademacher et al. (2001). The present results that the
HG-M source was located within the H1 in most of the cases are
also consistent with findings in cytoarchitectonic studies that
the core region was usually identified in the most anterior HG
(Rademacher et al., 1993; Hackett et al., 2001). Second, this
region receives dense thalamic inputs in primates, as expected
for primary cortical fields (for reference, see Kaas and Hackett,
1998). Third, this location is consistent with the location of the
electrophysiologically defined primary area of the auditory
cortex. In intracerebral recording studies in humans, early
auditory evoked potentials were recorded in the posterior
part of the HG (Celesia, 1976; Lie ´ geois-Cauvel et al., 1991,
1994). In a unitary recording study in monkeys (Kosaki et al.,
1997), neurons in the core (AI) were sharply tuned, while
neurons in the surrounding fields (L) showed a broader
frequency tuning. In addition, clear tonotopic organizations
have been found in this region in both humans (Romani et al.,
1982; Pantev et al., 1989, 1995) and monkeys (Merzenich and
The mean latency and location of each cortical activity
Source Latency (ms)Location (mm)
Coordinates relative to point A (posteriormost point of Heschl’s gyrus). x, antero-posterior; y,
medio-lateral; z, superior-inferior. Latencies for the STG-P source activity were the mean value
among three subjects with the first component. In the remaining three subjects, the triphasic
structure was not clear and the mean onset latency of the activity was 36.3 ms.
Figure 7. Locations of the two sources in and around Heschl’s gyrus (HG) in a subject
with a duplicated HG. The HG is divided into the first (H1) and second (H2) HG by an
intermediate sulcus (SI). The HG-M source is located in H1, while the HG-L source is
located more laterally in H2.
Cerebral Cortex Page 7 of 13
by guest on June 1, 2013
Brugge, 1973; Pfingst and O’Connor, 1981; Morel et al., 1993).
Based on these observations, the earliest MEG activity of cortical
origin in this study is considered to arise from the PAC.
However, there remains a possibility that the present study
missed the earliest activity originating from the PAC. Previous
intracranial recordings (Celesia, 1976; Liegeois-Chauvel et al.,
1991) identified cortical activities originating from the deep H1
area with a latency earlier than that of the first activity in the
present study. The present MEG study might not have detected
this early activity dueto its deep origin. Previous studies showed
that it is difficult for MEG to detect PAC activities earlier than
1M (Yoshiura et al., 1995; Yvert et al., 2001).
The second source, HG-L, was located more anterior (4 mm)
and lateral (9 mm) than the HG-M source. This location is on the
HS1 or just posterior to it, and probably corresponds to
a parakoniocortical area, PaAi (Galaburda and Sanides, 1980)
or LA (Rivier and Clarke, 1997), rather than koniocortical areas.
The view that the HG-L source is located outside the PAC is
further supported by a histochemical study of the human
auditory cortex using acetylcholinesterase (AChE) staining
(Hutsler and Gazzaniga, 1996), in which the transition between
AChE-cell-sparse (PAC) and AChE-cell-dense (planum tempo-
rale) regions is at the HG, and therefore does not coincide with
the transverse sulcus. In an MEG study with deconvolution of 40
Hz steady-state magnetic fields, Gutschalk et al. (1999) con-
cluded that two adjacent sources are responsible for the middle
latency components, one in the medial part of the HG and the
other in a region 1 cm lateral to the former, which is very similar
to the locations of our HG-M and HG-L sources. In addition, the
time course of each activity is also similar to ours. Their
components P30 (28.8 ms) and N41 (40.6 ms) from the first
source and P36 (35.2 ms) and N46 (46.2 ms) from the second
source seem to correspond to our two successive peaks of the
HG-M (28.0 and 39.7 ms) and HG-L (33.5 and 44.2 ms) sources
respectively. Therefore, the present results confirmed that
these two sources are not specific for the steady state but also
are activated following a transient stimulation. The delay of both
the onset (4.1 ms) and peak (5.5 ms) latencies of the HG-L
activity relative to those of the HG-M activity suggests serial
activations through these areas. In support of this view,
Liegeois-Chauvel et al. (1991) recorded activities in the lateral
part of the HG in response to electrical stimulation of the medial
Figure 8. Locations of the STG-P and STG-A sources superimposed on six subjects’ coronal MR images.
Page 8 of 13
Auditory Processing in Humans
Inui et al.
by guest on June 1, 2013
part of the HG with a latency of 6--8 ms. In unitary recording
studies in monkeys, the response latency of neurons in the
lateral belt region was longer than that of neurons in the core
region (Benson and Hienz, 1978; Vaadia et al., 1982). Serial acti-
vations through these areas are also consistent with anatomical
findingsin monkeys that the beltregions are strongly connected
with adjacent core regions (Merzenich and Brugge, 1973;
Activity in the PPC
The third source was located in the posterior parietal cortex. In
functional MRI (fMRI) and positron emission tomography (PET)
studies, activation in this region is related to auditory localiza-
tion tasks (Griffiths et al., 1998; Bushara et al., 1999; Alain et al.,
2001; Zatorre et al., 2002). The involvement of this region in the
localization of sounds has been also demonstrated by a trans-
cranial magnetic stimulation study (Lewald et al., 2002).
Therefore, the PPC activity in this study may reflect an early
stage of localization processing of sounds. To localize a sensory
source and execute spatially accurate behaviors, transformation
of the spatial coordinates from head-centered to body-centered
or world-centered coordinates is required. The role of the
posterior parietal cortex in such a transformation is well
established (Andersen, 1995).
The onset latency of this source (25 ms) was clearly longer
than that of the HG-M (17 ms) or HG-L (21 ms) source, but
shorter than those of the STG and PT sources. Therefore, the
PPC source appears to depend on the HG-M or HG-L source for
activation. In monkeys, the PPC (area 7) receives projections
from a subdivision of the belt region, CL (Romanski et al.,
1999b). This may suggest that the PPC source is driven by the
HG-L source, forming a stream from the HG-M, HG-L and then
PPC, which may be involved in ‘where’ processing in the
Activity in the STG
We found two lateral activities in the STG, one in the anterior
and another in the posterior part. As for the STG-P source, its
location was similar to that reported in an intracerebral AEP
study (Howard et al., 2000). The reason why these source
activities could be identified in only six subjects in spite of their
more superficial location than the earlier sources in the HG
could be explained by their radial orientation. Since these
dipoles were located around the crown of the STG, they should
create mainly radially oriented components, which are difficult
to detect by MEG. Previous scalp EEG studies (Scherg and Von
Cramon, 1986; Cacace et al., 1990) have suggested a radial
component peaking at ~40 ms originating from the STG.
Another reason for the difficulty in detecting STG activities in
Figure 9. Time course of each source activity and the location of each source relative to the landmark. (A) Time course of each source activity of all subjects (blue line) and group-
averages (black line). (B) Schematic drawings of the mean locationand orientationof each source. (C) Dipole coordinates relative to point A plotted in three orthogonal planes. Point
A is the crown of the most posterior part of the HG. Point B is the most anterior part of the HG. Point B in this figure indicates the mean location of point B of all subjects.
Cerebral Cortex Page 9 of 13
by guest on June 1, 2013
the current study might be their sensitivity to the stimulation
rate. Howard et al. (2000) examined the effects of the
interstimulus interval on AEP activities originating from both
the HG and posterior STG (termed the PLST). The recovery rate
of the HG and STG activities at an interstimulus interval of 0.25 s
relative to that at 2 s was ~60% and ~25% respectively. The re-
sponse latency of the STG-P activity was consistent with values
reported in previous AEP (Scherg and Von Cramon, 1986: 39 ms;
Cacace et al., 1990: 41 ms), AEF (Yvert et al., 2001, 40--60 ms)
and intracranial recording (Celesia, 1976: 40 ms; Howard et al.,
2000: 45 ms) studies in humans, and an intracerebral AEP study
in monkeys (Arezzo et al., 1975: 27 ms). In the present study,
clear early activity from this area could be identified in three
subjects in whom the onset latency of the STG-P activity (26.2
ms) was longer than that of the HG-M and HG-L by 8.9 and 4.0
ms respectively. An electrophysiological study in humans
(Howard et al., 2000) provided evidence for the existence of
cortico-cortical projections from the HG to the posterior part of
the STG by showing that electrical stimulation of the HG elicits
responses in the STG with an onset latency of 3 ms and a peak
latency of 10--20 ms, which approximately corresponds to the
time delay between the HG-M and STG-P activities in this study.
Therefore, it seems possible that the STG-P was directly driven
by the HG-M or HG-L. Findings in a recent study using electrical
stimulation and intracranial recording methods in humans
(Brugge et al., 2003) favored a possibility that the STG-P source
is directly driven by the HG-M activity. They showed that
electrical stimulation of the medial HG elicited responses
consisting of several components in the posterior STG region
while stimulation of the lateral HG was ineffective in evoking
some of the components. Although the present results could
not determine the actual sequence of activations, successive
activations through the HG-M, HG-L and STG-P with a 4--5 ms
time delay implied a serial processing through them. The
sequential activation through the HG and STG-P is also
consistent with a histochemical study in humans using cyto-
chrome oxidase staining. Rivier and Clarke (1997) demon-
strated that the primary auditory area had a cytochrome oxidase
profile compatible with a primary sensory area, while STA,
which probably corresponds to our STG-P, had a very different
profile compatible with a high-order association area. The view
that the STG-P is located at a higher level in the auditory
processing stream than the HG is supported by recent PET and
fMRI studies, in which this region was usually more strongly
activated by complex acoustic (Thivard et al., 2000; Hall et al.,
2002), word (Petersen et al., 1988; Price et al., 1992) and
speech (Zatorre et al., 1992; Scott et al., 2000) stimuli than
Another source in the lateral aspect of the STG was located
around the superior bank of the STS. The mean onset latency of
this activity (30.9 ms) was clearly longer than that for the HG-M
(17.1 ms) or HG-L (21.2 ms) source. Therefore, the STS source
appears to depend for its activity on earlier activities in adjacent
auditory regions. However, the latency difference of 10 ms
between HG-L and STS sources left a possibility that the STS
source was driven by a feedback or lateral projection. Activation
in this area in neuroimaging studies was related to frequency-
modulated tone (Hall et al., 2002), speech (Binder et al., 2000),
intelligible speech (Scott et al., 2000) and voice (Belin et al.,
2000) stimuli, suggesting a role for this region in higher-level
acoustic processing. In monkeys, regions lateral to the core
contain neurons that are sensitive to complex sounds or
species-specific vocalizations (Rauschecker et al., 1995). There-
fore, the STS activity in the present study may represent
a beginning of the processing of sound patterns.
In monkeys, belt regions surrounding the core area receive
inputs from the core area, and the belt fields in turn project to
the surrounding parabelt fields and auditory-related fields on
the STG, caudal STG and rostral STG (Kaas and Hackett, 1999).
The anterior region of the belt projects to the anterior parabelt
and STGr, while the posterior region of the belt projects to the
posterior parabelt and STGc (Seltzer and Pandya, 1978; Hackett
et al., 1998). Then the subdivisions of the lateral belt and
parabelt/STG fields project to distinct regions of the prefrontal
cortex (Hackett et al., 1999; Romanski et al., 1999a,b). Anterior
auditory fields tend to project to non-spatial domains and
posterior fields tend to project to spatial domains, forming
separate auditory streams corresponding to ‘what’ and ‘where’
pathways. Therefore, the two separable source activities in the
STG region in this study may be the third stage of temporal lobe
processing of each stream, i.e. from the core or HG-M to the belt
or HG-L, the parabelt or STG and finally the prefrontal cortex.
Activation in the Planum Temporale
The last source was located in an area posterior to the HG in the
planum temporale and exhibited a long-duration activity peak-
ing at ~80 ms. Since both the orientation and timing of this
source activity were similar to those of a well-known auditory
evoked component around 100 ms, N1, we considered that the
activity from the PT source in this study contributes to the
formation of N1. The reduced strength of the activity in
the present study compared to other studies is due to our short
interstimulus interval, since it is reported that a short in-
terstimulus interval reduces the N1 amplitude (Schwent et al.,
1976; Hari et al., 1982). The view that the PT is one of the
source regions responsible for N1 is consistent with the result
of intracranial recording (Liegois-Chuvel et al., 1994; Godey
et al., 2001), AEP (Pantev et al., 1995) and AEF (Pelizzone et al.,
1987; Pantev et al., 1995; Lu ¨ tkenho ¨ ner and Steinstra ¨ ter, 1998)
studies. The slightly shorter latency of the PT source activity
Figure 10. Early STG-P source activities in three subjects. Thin lines indicate
waveforms of three subjects and thick lines indicate the averaged waveform. For
a comparison of latency, waveforms of the HG-L activity of these subjects are also
shown. Note the similar waveforms of the HG-L and STG-P activity with a few
milliseconds time delay for the latter.
Page 10 of 13
Auditory Processing in Humans
Inui et al.
by guest on June 1, 2013
than that of N1 (90--110 ms) in other reports may suggest that
this source is responsible for the early part of N1.
Comparison between Somatosensory and
In a previous MEG study, we showed the precise activation
timing of several somatosensory cortical areas following tactile
stimuli (Inui et al., 2004). Since we found some common
features between somatosensory and auditory processing, we
want to discuss this matter here briefly.
Serial Activation in a Stream in One Direction
As one might expect, signals in a cortical area are basically
conveyed to an adjacent area at least in the early stages of
processing and serial activations through several areas make
a stream in one direction. Following tactile stimulation, areas 3b,
1 (and 2) and 5 are serially activated, forming a stream running
posteriorly in the postcentral gyrus. In the present study, click
stimuli activated the HG-M, HG-L and STG sources sequentially,
forming a stream(s) running laterally in the supratemporal
plane. Such processing is quite natural in terms of speed.
Timing of Serial Activation
In somatosensory processing, the time delay between two
sequential activations is ~4 ms; 3.6 ms between areas 3b and
1, ~4.4 ms between areas 1 and 5, and ~3.7 ms between area 1
and SII. In the present study, the time delay was 4.1 ms between
HG-M and HG-L, 4.1 ms between HG-L and PPC, and 4.0 ms
between HG-L and STG-P (three subjects). Although the
physiological meaning of the 4 ms time delay is not clear, these
results suggest that somatosensory and auditory information is
processed in steps with similar timing.
Biphasic (Triphasic) Nature of Early Response
Early activities following tactile stimulation in area 3b, area 1,
the PPC and the SII region exhibit reversals of polarity after
10 ms once or in some cases twice, resulting in a biphasic or
triphasic time course. Similarly, the time course of activities of
the HG-M, HG-L, PPC and STG-P sources showed a triphasic
waveform. Such a structure was also found in the earliest
cortical activity in SI following noxious stimuli (Inui et al.,
2003). This phenomenon might correspond to the surface
positive--negative sequence of potentials, the so-called primary
evoked response, recorded from the brain surface of experi-
mental animals (for review, see Schlag, 1973). Although the
precise mechanisms responsible for the polarity reversals are
unclear in humans, these findings suggest that there is a similar
intra-laminar processing among these somatosensory and audi-
tory cortical areas. For dipoles with opposite orientations to be
produced in the same cortical field, different combinations of
a current sink and source in different layers must exist. In
monkeys, such an intra-laminar processing pattern has been
studied using a current source density (CSD) analysis. Accord-
ing to a series of studies by Schroeder’s group (Schroeder et al.,
1995, 1998, 2001; Schroeder and Foxe, 2002), CSD analysis can
clearly differentiate feedforward from feedback and lateral
input patterns. The laminar profiles of feedforward responses
triggered by ascending synaptic inputs show a characteristic
activation sequence composed of an initial depolarization of
input terminals in and near lamina 4 and a later excitation of
supra- and infragranular pyramidal cells, which is consistent
with the biphasic structure of the early responses in the present
study. In addition, their data indicated that such laminar profiles
of feedforward responses are common between auditory and
somatosensory processing. Based on our MEG and previous CSD
results, it seems natural to consider that the similar time course
of the response sequence of the early cortical activities in the
present (auditory) and previous (somatosensory) studies mainly
reflect the cortico-cortical feedforward pattern response.
In the present study, the second component of the triphasic
structure often did not exceed the baseline, especially the HG-L
source. This is probably due to the similar time course of the
HG-M and HG-L activities with a 4--5 ms time delay for the latter
and due to the similar orientation of these source activities.
Around the peak of the second component of the HG-M activity
oriented postero-inferiorly (Fig. 5B), the HG-L activity is still
directed antero-superiorly. Because of the proximal positions
and opposite orientations of the two sources, activities from the
two sources around this latency might cancel each other out to
some extent. This phenomenon may explain why auditory
evoked magnetic fields do not shape a clear component at
a latency corresponding to the second peak of these source
activities (~40 ms) like the well-known second component at
component at 50 ms following visual stimulation (K. Inui et al.,
Interestingly, the late activity from the SII region and late
activity from the PT donot have such clear polarity reversals and
are long-lasting. Therefore, these activities should differ in
function from early activities with the triphasic structure. We
postulate that a basic role of early cortical areas is to receive
inputs from the thalamus or convergent inputs from the
thalamus and/or adjacent multiple cortical areas and to send
their signals to the next point, while the later long-lasting
activities are related to recognition of the stimuli. Early activities
some 10 ms in duration or successive activations with a 4 ms
time delay seem not appropriate for activities relating to
recognition. Both the late SII activity (Hari et al., 1990) and
the N1 activity (Hillyard et al., 1973) are substantially influenced
by the subject’s attentional state.
Address correspondence to Koji Inui, Department of Integrative
Physiology, National Institute for Physiological Sciences, Okazaki 444-
8585, Japan. Email: firstname.lastname@example.org.
Alain C, Arnott SR, Hevenor S, Graham S, Grady CL (2001) ‘What’ and
‘where’ in the human auditory system. Proc Natl Acad Sci USA
Andersen RA (1995) Encoding of intention and spatial location in the
posterior parietal cortex. Cereb Cortex 5:457--469.
Arezzo J, Pickoff A, Vaughan HG Jr (1975) The sources and intracerebral
distribution of auditory evoked potentials in the alert rhesus
monkey. Brain Res 90:57--73.
Belin P, Zatorre RJ, Lafaille P, Ahad P, Pike B (2000) Voice-selective areas
in human auditory cortex. Nature 403:309--312.
Benson DA, Hienz RD (1978) Single-unit activity in the auditory cortex
of monkeys selectively attending left vs. right ear stimuli. Brain Res
Binder JR, Frost JA, Hammeke TA, Bellgrowan PSF, Springer JA,
Kaufman JN, Possing ET (2000) Human temporal lobe activation by
speech and nonspeech sounds. Cereb Cortex 10:512--528.
Brugge JF, Volkov IO, Garell PC, Reale RA, Howard MA III (2003)
Functional connections between auditory cortex on Heschl’s gyrus
Cerebral Cortex Page 11 of 13
by guest on June 1, 2013
and on the lateral superior temporal gyrus in humans. J Neurophysiol
Burton H, Jones EG (1976) The posterior thalamic region and its cortical
projection in new world and old world monkeys. J Comp Neurol
Bushara KO, Weeks RA, Ishii K, Catalan M-J, Tian B, Rauschecker JP,
Hallett M (1999) Modality-specific frontal and parietal areas for
auditory and visual spatial localization in humans. Nat Neurosci
Cacace AT, Satya-Murti S, Wolpaw JR (1990) Human middle-latency
auditory potentials: vertex and temporal components. Electroence-
phalogr Clin Neurophysiol 77:6--18.
Celesia GG (1976) Organization of auditory cortical areas in man. Brain
Eggermont JJ (1998) Representation of spectral and temporal sound
features in three cortical fields of the cat. Similarities outweigh
differences. J Neurophysiol 80:2743--2764.
Galaburda A, Sanides F (1980) Cytoarchitectonic organization of the
human auditory cortex. J Comp Neurol 190:597--610.
Galaburda AM, Pandya DN (1983) The intrinsic architectonic and
connectional organization of the superior temporal region of the
rhesus monkey. J Comp Neurol 221:169--184.
Godey B, Schwartz D, de Graaf JB, Chauvel P, Lie ´ geois-Chauvel C (2001)
Neuromagnetic source localization of auditory evoked fields and
intracerebral evoked potentials: a comparison of data in the same
patients. Clin Neurophysiol 112:1850--1859.
Griffiths TD, Rees G, Rees A, Green GGR, Witton C, Rowe D, Bu ¨ chel C,
Turner R, Frackowiak RSJ (1998) Right parietal cortex is involved in
the perception of sound movement in humans. Nat Neurosci
Gutschalk A, Mase R, Roth R, Ille N, Rupp A, Ha ¨ hnel S, Picton TW,
Scherg M (1999) Deconvolution of 40 Hz steady-state fields reveals
two overlapping source activities of the human auditory cortex.
Electroencephalogr Clin Neurophysiol 110:856--868.
Hackett TA, Stepniewska I, Kaas JH (1998) Subdivisions of auditory
cortex and ipsilateral cortical connections of the parabelt auditory
cortex in macaque monkeys. J Comp Neurol 394:475--495.
Hackett TA, Stepniewska I, Kaas JH (1999) Prefrontal connections of the
parabelt auditory cortex in macaque monkeys. Brain Res 817:45--58.
Hackett TA, Preuss TM, Kaas JH (2001) Architectonic identification of
the core region in auditory cortex of macaques, chimpanzees, and
humans. J Comp Neurol 441:197--222.
Hall DA, Johnsrude IS, Haggard MP, Palmer AR, Akeroyd MA,
Summerfield AQ (2002) Spectral and temporal processing in human
auditory cortex. Cereb Cortex 12:140--149.
Hari R, Kaila K, Katila T, Tuomisto T, Varpula T (1982) Interstimulus
interval dependence of the auditory vertex response and its
magnetic counterpart: implications for their neural generation.
Electroencephalogr Clin Neurophysiol 54:561--569.
Hari R, Joutsiniemi S-L, Sarvas J (1988) Spatial resolution of neuro-
magnetic records: theoretical calculations in a spherical model.
Electroencephalogr Clin Neurophysiol 71:64--72.
Hari R, Ha ¨ ma ¨ la ¨ inen H, Ha ¨ ma ¨ la ¨ inen M, Kekoni J, Sams M, Thiihonen J
(1990) Separate finger representations at the human second
somatosensory cortex. Neuroscience 37:245--249.
Hillyard SA, Hink RF, Schwent VL, Picton TW (1973) Electrical signs of
selective attention in the human brain. Science 182:177--180.
Howard MA, Volkov IO, Mirsky R, Garell PC, Noh MD, Granner M,
Damasio H, Steinschneider M, Reale RA, Hind JE, Brugge JF (2000)
Auditory cortex on the human posterior superior temporal gyrus.
J Comp Neurol 416:79--92.
Hutsler JJ, Gazzaniga MS (1996) Acetylcholinesterase staining in human
auditory and language cortices: regional variation of structural
features. Cereb Cortex 6:260--270.
Tran DT, Kakigi R (2003) Pain processing within the primary
somatosensory cortex in humans. Eur J Neurosci 18:2859--2866.
Inui K, Wang X, Tamura Y, Kaneoke Y, Kakigi R (2004) Serial processing
in the human somatosensory system. Cereb Cortex 14:851--857.
Iwamura Y (1998) Hierarchical somatosensory processing. Curr Opin
Jones EG, Dell’Anna ME, Molinari M, Rausell E, Hashikawa T (1995)
Subdivisions of macaque monkey auditory cortex revealed by
calcium-binding protein immunoreactivity. J Comp Neurol 362:
Kaas JH, Hackett TA (1998) Subdivisions of auditory cortex and levels
of processing in primates. Audiol Neurootol 3:73--85.
Kass JH, Hackett TA (1999) ‘What’ and ‘where’ processing in auditory
cortex. Nat Neurosci 2:1045--1047.
Kaas JH, Hackett TA, Tramo MJ (1999) Auditory processing in primate
cerebral cortex. Curr Opin Neurobiol 9:164--170.
Kajikawa Y, de la Mothe L, Blumell S, Hackett TA (2005) A comparison of
neuron response properties in areas A1 and CM of the marmoset
monkey auditory cortex: tones and broad band noise. J Neurophysiol
Kosaki H, Hashikawa T, He J, Jones EG (1997) Tonotopic organization of
auditory cortical fields delineated by parvalbumin immunoreactivity
in macaque monkeys. J Comp Neurol 386:304--316.
Kowalski N, Versnel H, Shamma A (1995) Comparison of responses in
the anterior and primary auditory fields of the ferret cortex.
J Neurophysiol 73:1513--1523.
Leonard CM, Puranik C, Kuldau JM, Lombardino LJ (1998) Normal
variation in the frequency and location of human auditory cortex
landmarks. Heschl’s gyrus: where is it? Cereb Cortex 8:397--406.
Lewald J, Foltys H, To ¨ pper R (2002) Role of the posterior parietal cortex
in spatial hearing. J Neurosci 22:207.
Liegeois-Chauvel C, Musolino A, Chauvel P (1991) Localization of the
primary auditory area in man. Brain 114:139--153.
Lie ´ geois-Chauvel C, Musolino A, Badier JM, Marquis P, Chauvel P (1994)
Evoked potentials recorded from the auditory cortex in man:
evaluation and topography of the middle latency components.
Electroencephalogr Clin Neurophysiol 92:204--214.
Lu ¨ tkenho ¨ ner B, Steinstra ¨ ter O (1998) High-precision neuromagnetic
study of the functional organization of the human auditory cortex.
Audiol Neurootol 3:191--213.
Mesulam M-M, Geula C (1994) Chemoarchitectonics of axonal and
perikaryal acetylcholinesterase along information processing sys-
tems of the human cerebral cortex. Exp Brain Res 33:137--153.
Merzenich MM, Brugge JF (1973) Representation of the cochlear
partition on the superior temporal plane of the macaque monkey.
Brain Res 50:275--296.
Morel A, Garraghty PE, Kaas JH (1993) Tonotopic organization,
architectonic fields, and connections of auditory cortex in macaque
monkeys. J Comp Neurol 335:437--459.
(2001) Human primary auditory cortex: cytoarchitectonic subdivi-
sions and mapping into a spatial reference system. Neuroimage
Pantev C, Hoke M, Lu ¨ tkenho ¨ ner B, Lehnertz K (1989) Tonotopic
organization of the auditory cortex: pitch versus frequency repre-
sentation. Science 246:486--488.
T (1995) Specific tonotopic organizations of different areas of the
human auditory cortex revealed by simultaneous magnetic and
Pelizzone M, Hari R, Ma ¨ kela ¨ JP, Huttunen J, Ahlfors S, Ha ¨ ma ¨ la ¨ inen M
(1987) Cortical origin of middle-latency auditory evoked responses
in man. Neurosci Lett 82:303--307.
Petersen SE, Fox PT, Posner MI, Mintun M, Raichle ME (1988) Positron
emission tomographic studies of the cortical anatomy of single-word
processing. Nature 331:585--589.
Pfingst BE, O’Connor TA (1981) Characteristics of neurons in auditory
cortex ofmonkeysperforminga simpleauditory task. J Neurophysiol
(1992) Regional response differences within the human auditory
cortex when listening to words. Neurosci Lett 146:179--182.
Rademacher J, Morosan P, Schormann T, Schleicher A, Werner C, Freund
H-J, Zilles K (2001) Probabilistic mapping and volume measurement
of human primary auditory cortex. Neuroimage 13:669--683.
Rademacher J, Caviness VS Jr, Steinmetz H, Galaburda AM (1993)
Topographical variation of the human primary cortices: implications
Page 12 of 13
Auditory Processing in Humans
Inui et al.
by guest on June 1, 2013
for neuroimaging, brain mapping, and neurobiology. Cereb Cortex Download full-text
Rauschecker JP, Tian B, Hauser M (1995) Processing of complex sounds
in the macaque nonprimary auditory cortex. Science 268:111--114.
Rauschecker JP, Tian B, Pons T, Mishkin M (1997) Serial and parallel
processing in rhesus monkey auditory cortex. J Comp Neurol
Rivier F, Clarke S (1997) Cytochrome oxidase, acetylcholinesterase,
and NADPH-d iaphorase staining in human supratemporal and
insular cortex: evidence for multiple auditory areas. Neuroimage
Romani GC, Williamson SJ, Kaufman L (1982) Tonotopic organization of
the human auditory cortex. Science 216:1339--1340.
Romanski LM, Bates JF, Goldman-Rakic PS (1999a) Auditory belt and
parabelt projections to the prefrontal cortex in the rhesus monkey.
J Comp Neurol 403:141--157.
Romanski LM, Tian B, Fritz J, Mishkin M, Goldman-Rakic PS,
Rauschecker JP (1999b) Dual streams of auditory afferents target
multiple domains in the primate prefrontal cortex. Nat Neurosci
Salami M, Itami C, Tsumoto T, Kimura F (2003) Changes of conduction
velocity by regional myelination yields constant latency irrespective
of distance between thalamus and cortex. Proc Natl Acad Sci USA
Scherg M, Von Cramon D (1986) Evoked dipole source potentials of the
human auditory cortex. Electroencephalogr Clin Neurophysiol
Schlag J (1973) Generation of brain evoked potentials. In: Bioelectric
recording techniques. Cellular processes and brain potentials
(Thompson RF, Patterson MM, eds), part A, pp. 273--316. New York:
Schroeder CE, Foxe JJ (2002) The timing and laminar profile of
converging inputs to multisensory areas of the macaque neocortex.
Brain Res Cogn Brain Res 14:187--198.
Schroeder CE, Seto S, Arezzo JC, Garraghty PE (1995) Electrophysiolog-
ical evidence for overlapping dominant and latent inputs to
somatosensory cortex in squirrel monkeys. J Neurophysiol 74:
Schroeder CE, Mehta AD, Givre SJ (1998) A spatiotemporal profile of
visual system activation revealed by current source density analysis
in the awake macaque. Cereb Cortex 8:575--592.
Schroeder CE, Lindsley RW, Specht C, Marcovici A, Smiley JF, Javitt DC
(2001) Somatosensory input to auditory association cortex in the
macaque monkey. J Neurophysiol 85:1322--1327.
Schwent VL, Hillyard SA, Galambos R (1976) Selective attention and the
auditory vertex potential. I. Effects of stimulus delivery rate. Electro-
encephalogr Clin Neurophysiol 40:604--614.
Scott SK, Blank CC, Rosen S, Wise RJS (2000) Identification of a pathway
Seltzer B, Pandya DN (1978) Afferent cortical connections and archi-
tectonics of the superior temporal sulcus and surrounding cortex in
the rhesus monkey. Brain Res 149:1--24.
Steinmetz H, Rademacher J, Huang Y, Hefter H, Zilles K, Thron A,
Freund H-J (1989) Cerebral asymmetry: MR planimetry of the human
planum temporale. J Comput Assist Tomogr 13:996-1005.
Supek S, Aine CJ (1993) Simulation studies of multiple dipole neuro-
magnetic source localization: model order and limits of source
resolution. IEEE Trans Biomed Eng 40:529--540.
Thivard L, Belin P, Zibovicius M, Poline J-B, Samson Y (2000) A cortical
region sensitive to auditory spectral motion. Neuroreport 11:
Vaadia E, Gottlieb Y, Abeles M (1982) Single-unit activity related to
sensorimotor association in auditory cortex of a monkey. J Neuro-
Yoshiura T, Ueno S, Iramina K, Masuda K (1995) Source localization of
middle latency auditory evoked magnetic fields. Brain Res 703:
Yvert B, Crouzeix A, Bertrand O, Seither-Preisler A, Pantev C (2001)
Multiple supratemporal sources of magnetic and electric auditory
evoked middle latency components in humans. Cereb Cortex 11:
Zatorre RJ, Evans AC, Meyer E, Gjedde A (1992) Lateralization of
phonetic and pitch discrimination in speech processing. Science
Zatorre RJ, Bouffard M, Ahad P, Belin P (2002) Where is ‘where’ in the
human auditory cortex? Nat Neurosci 5:905--909.
Cerebral Cortex Page 13 of 13
by guest on June 1, 2013