Linking semantic priming effect in functional MRI and event-related potentials.
ABSTRACT The aim of this study is to examine the neural substrates involved in semantic priming using a combined event-related functional magnetic resonance imaging (fMRI) and event-related potentials (ERP) study. Twelve subjects were instructed to judge whether the presented target word was a real word or a nonword. Under the related condition, target words were preceded by a semantically related prime word. On the other hand, under the unrelated condition, prime words did not have semantic relatedness with the target word. The reaction time for reaching a judgment was longer under the unrelated condition than under the related condition, indicating that the recognition of target words is promoted by semantic priming under the related condition. In the fMRI results, we found reduced activity in the dorsal and ventral left inferior frontal gyrus, the anterior cingulate, and left superior temporal cortex for related versus unrelated conditions (i.e., the repetition suppression effect). ERP analysis revealed that the amplitude of the N400 component was reduced under the related condition compared with the unrelated condition (i.e., the N400 priming effect). Correlation analysis between the BOLD repetition suppression effect and the N400 priming effect decomposed by independent component analysis (ICA) across subjects showed significant correlation in the left superior temporal gyrus. This finding is consistent with the recent MEG data suggesting that the source of N400 is judged to be the bilateral superior temporal lobe. We discussed this finding herein in relation to the modulation of access to the phonological representation caused by semantic priming.
- SourceAvailable from: Markus Kiefer[Show abstract] [Hide abstract]
ABSTRACT: In classical theories of attention, unconscious automatic processes are thought to be independent of higher-level attentional influences. Here, we propose that unconscious processing depends on attentional enhancement of task-congruent processing pathways implemented by a dynamic modulation of the functional communication between brain regions. Using functional magnetic resonance imaging, we tested our model with a subliminally primed lexical decision task preceded by an induction task preparing either a semantic or a perceptual task set. Subliminal semantic priming was significantly greater after semantic compared to perceptual induction in ventral occipito-temporal (vOT) and inferior frontal cortex, brain areas known to be involved in semantic processing. The functional connectivity pattern of vOT varied depending on the induction task and successfully predicted the magnitude of behavioral and neural priming. Together, these findings support the proposal that dynamic establishment of functional networks by task sets is an important mechanism in the attentional control of unconscious processing. Hum Brain Mapp, 2014. © 2014 Wiley Periodicals, Inc.Human Brain Mapping 06/2014; · 6.92 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Mood is widely assessed in nutrition research, usually with rating scales. A core assumption is that positive mood reinforces ingestion, so it is important to measure mood well. Four relevant theoretical issues are reviewed: (i) the distinction between protracted and transient mood; (ii) the distinction between mood and emotion; (iii) the phenomenology of mood as an unstable tint to consciousness rather than a distinct state of consciousness; (iv) moods can be caused by social and cognitive processes as well as physiological ones. Consequently, mood is difficult to measure and mood rating is easily influenced by non-nutritive aspects of feeding, the psychological, social and physical environment where feeding occurs, and the nature of the rating system employed. Some of the difficulties are illustrated by reviewing experiments looking at the impact of food on mood. The mood-rating systems in common use in nutrition research are then reviewed, the requirements of a better mood-rating system are described, and guidelines are provided for a considered choice of mood-rating system including that assessment should: have two main dimensions; be brief; balance simplicity and comprehensiveness; be easy to use repeatedly. Also mood should be assessed only under conditions where cognitive biases have been considered and controlled.Nutrition Research Reviews 12/2014; · 3.86 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: This review examines existing cognitive experimental and brain imaging research related to cocaine addiction. In section 1, previous studies that have examined cognitive processes, such as implicit and explicit memory processes in cocaine users are reported. Next, in section 2, brain imaging studies are reported that have used chronic users of cocaine as study participants. In section 3, several conclusions are drawn. They are: (a) in cognitive experimental literature, no study has examined both implicit and explicit memory processes involving cocaine related visual information in the same cocaine user, (b) neural mechanisms underlying implicit and explicit memory processes for cocaine-related visual cues have not been directly investigated in cocaine users in the imaging literature, and (c) none of the previous imaging studies has examined connectivity between the memory system and craving system in the brain of chronic users of cocaine. Finally, future directions in the field of cocaine addiction are suggested.Journal of clinical toxicology. 07/2012; 2012(Suppl 7):003.
Linking semantic priming effect in functional MRI
and event-related potentials
Atsushi Matsumoto,a,*Tetsuya Iidaka,a,bKaoruko Haneda,a
Tomohisa Okada,cand Norihiro Sadatob,c
aGraduate School of Environmental Studies, Nagoya University, Nagoya, Aichi, 464-8601, Japan
bJST/RISTEX, Kawaguchi, Japan
cDepartment of Cerebral Research, National Institute for Physiological Sciences, Okazaki, Aichi, Japan
Received 3 February 2004; revised 20 June 2004; accepted 8 September 2004
Available online 14 November 2004
The aim of this study is to examine the neural substrates involved in
semantic priming using a combined event-related functional magnetic
resonance imaging (fMRI) and event-related potentials (ERP) study.
Twelve subjects were instructed to judge whether the presented target
word was a real word or a nonword. Under the related condition,
target words were preceded by a semantically related prime word. On
the other hand, under the unrelated condition, prime words did not
have semantic relatedness with the target word. The reaction time for
reaching a judgment was longer under the unrelated condition than
under the related condition, indicating that the recognition of target
words is promoted by semantic priming under the related condition. In
the fMRI results, we found reduced activity in the dorsal and ventral
left inferior frontal gyrus, the anterior cingulate, and left superior
temporal cortex for related versus unrelated conditions (i.e., the
repetition suppression effect). ERP analysis revealed that the amplitude
of the N400 component was reduced under the related condition
compared with the unrelated condition (i.e., the N400 priming effect).
Correlation analysis between the BOLD repetition suppression effect
and the N400 priming effect decomposed by independent component
analysis (ICA) across subjects showed significant correlation in the left
superior temporal gyrus. This finding is consistent with the recent
MEG data suggesting that the source of N400 is judged to be the
bilateral superior temporal lobe. We discussed this finding herein in
relation to the modulation of access to the phonological representation
caused by semantic priming.
D 2004 Elsevier Inc. All rights reserved.
Keywords: fMRI; Cortex; Gyrus
Semantic priming is a phenomenon in which a target word (e.g.,
bnurseQ) is recognized more rapidly when it is preceded by a
semantically related word (e.g., bdoctorQ) than when it is preceded
by an unrelated word (e.g., bbreadQ). A number of psychological
studies have investigated the cognitive basis of this phenomenon
and demonstrated that semantic priming consists of several
cognitive processes. Neely (1991) suggested that semantic priming
results from automatic and controlled processing mechanisms. In
the former, known as bautomatic spreading activationQ, the
presentation of prime words can activate the corresponding
semantic representations with subsequent spread to and automatic
activation of related nodes, thereby facilitating recognition when a
semantically related word is presented. In the latter, controlled
semantic processing involving prelexically developed expectancies
and postlexical semantic matching is under conscious control.
Subjects develop expectancies regarding the recognition of the
prime word, leading to attention directed to semantically related
words. Therefore, subsequent processing of a related word is
promoted, whereas processing of an unrelated (unexpected) word
is inhibited. Another study by Neely (1977) suggests that
automatic spreading activation is invoked when the interval
between presentation of prime and target words (i.e., stimulus
onset asynchrony; SOA) is short, while controlled processes are
engaged when the SOA is greater than 400 ms.
In recent years, several studies employing single cell recordings
and functional imaging techniques, such as fMRI and PET, have
been used to investigate the neural basis of priming and have
revealed that repetition priming inhibits activity in those brain
regions that process a presented stimulus (Badgaiyan et al., 1999;
Buckner et al., 2000; Desimone, 1996; Thiel et al., 2001). For
example, Desimone (1996) showed that repeated stimuli resulted
in decreases in neural firing in the infero-temporal neurons of
macaque monkeys and referred to this phenomenon as brepetition
suppressionQ (Henson, 2003). Desimone (1996) also proposed that
1053-8119/$ - see front matter D 2004 Elsevier Inc. All rights reserved.
* Corresponding author. Graduate School of Environmental Studies,
Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi, 464-8601,
E-mail address: firstname.lastname@example.org (A. Matsumoto).
Available online on ScienceDirect (www.sciencedirect.com).
NeuroImage 24 (2005) 624–634
a repeated stimulus produces a bsharpeningQ of its cortical
representation, whereby neurons that code features that are
unnecessary for processing the stimulus become progressively
less responsive. This results in a reduction in the firing rate of a
population of neurons and a decrease in the neuronal activity from
that region of the cortex.
Recent studies of semantic priming in event-related fMRI
showed clear repetition suppression in the left inferior frontal gyrus
and in the middle or superior temporal gyrus. Event-related design
reduces the effect of modes that are not sensitive to semantic
relations between word pairs per se (e.g. strategies or attention)
from the semantic priming effect. For instance, Kotz et al. (2002)
conducted an event-related fMRI experiment to investigate the
neural basis of auditory semantic priming and reported that the
activation of the left inferior frontal gyrus was lower in the related
condition than in the unrelated condition. Copland et al. (2003)
also characterized the repetition suppression effect in the left
inferior frontal cortex during a visual semantic priming task. The
left inferior frontal gyrus participates in the semantic processing of
words or objects, and a number of repetition priming studies have
shown that the activity of this region is attenuated by priming
(Buckner et al., 1995, 2000; Donaldson et al., 2001; Thiel et al.,
2001; Wagner et al., 2000). Kotz et al. (2002) suggested that the
repetition suppression effect in the left inferior frontal gyrus is due
to reduced demand for semantic memory retrieval or the result of
selection caused by the automatic and/or controlled semantic
priming effect. Furthermore, significant repetition suppression has
been observed in the left anterior medial temporal cortex (Rossell
et al., 2003) and superior temporal gyrus (Rissman et al., 2003)
during a semantic priming task, suggesting that the left medial and
superior temporal gyrus is associated with the preservation of the
word’s representation and that the repetition suppression in these
regions is due to the preceding activation of target word
representation caused by automatic spreading activation.
Semantic priming has also been investigated using neuro-
physiological methods, such as event-related potentials (ERPs).
ERP studies have shown that semantic priming affects the
amplitude of an ERP component called N400, which is a negative
deflection of the ERP signal between 300 and 500 ms, peaking at
about 400 ms after stimulus presentation (Kutas and Hillyard,
1980). N400 is sensitive to semantic deviation, with larger N400
amplitudes corresponding to semantically incongruent words at the
sentence and word level processing (Kutas and Hillyard, 1980). In
the semantic or repetition priming paradigm, N400 is reduced
when target words are preceded by semantically related primes or
by the same word (N400 priming effect) (Holcomb, 1993; Kiefer
and Spitzer, 2000; Van Petten, 1993).
There are two main hypotheses concerning the cognitive basis
of N400. The lexical hypothesis states that N400 reflects the
activation of the lexical and/or semantic representation of a word
(Fischler and Raney, 1989; Van Petten and Kutas, 1987), while the
postlexical hypothesis suggests that N400 reflects an associative
process that integrates word representation with ongoing context
(Holcomb, 1993; Neville et al., 1991; Rugg and Doyle, 1994).
Thus, the N400 priming effect may result from the ease of
integrating information into a context or from the access to word
representation caused by controlled or automatic semantic priming.
Several studies have suggested that N400 is generated from a
number of different regions (Guillem et al., 1999; Nobre and
McCarthy, 1995; Nobre et al., 1994). Intra-cranial depth recording
of ERP emphasized the importance of the medial temporal cortex,
including the amygdala and hippocampus, in the generation of
N400 (Guillem et al., 1999; McCarthy et al., 1995). McCarthy et
al. (1995) suggested that N400 is generated from the neocortex
near the collateral sulcus, including the anterior fusiform and
parahippocampal gyri. However, intra-cranial recordings from
other structures have shown that cortical areas along the superior
temporal sulcus are involved in the generation of N400 (Halgren et
al., 1994). Furthermore, a number of magnetoencephalography
(MEG) studies have suggested that N400 originates in the bilateral
superior temporal gyrus in the vicinity of the auditory cortex, with
the signal being stronger in the left hemisphere than in the right
hemisphere (Helenius et al., 1998, 2002; Laine et al., 2000;
Sekiguchi et al., 2001; Simos et al., 1997). Helenius et al. (1998)
found that semantic deviation in sentence reading elicited magnetic
fields at 350–450 ms after word onset (N400m) and suggested that
this process was mediated by the bilateral superior temporal gyrus.
Sekiguchi et al. (2001) reported that repetitive presentation of
visual words reduced the magnetic field associated with the left
superior temporal gyrus adjacent to the auditory cortex, suggesting
that the superior temporal gyrus mediates generation and priming
effect on N400m. Furthermore, Marinkovic et al. (2003) used a
distributed source modeling technique to demonstrate that the
priming effect on N400m is mediated via the left superior temporal
gyrus and the left inferior frontal gyrus. These results are consistent
with observations from fMRI studies, which reported that the
semantic priming caused a repetition suppression effect in the left
inferior frontal cortex and the left middle or superior temporal
cortex. However, the relationship between the repetition suppres-
sion effect in the BOLD signal and the N400 priming effect
Rossell et al. (2003) studied the neural distribution of semantic
priming using a combination of fMRI and ERP and a short and
long interval between prime and target items in order to manipulate
the degree of semantic expectancies. They showed a significant
repetition suppression in the anterior medial temporal lobe with the
N400 priming effect associated with the centro-parietal regions,
indicating the relationship between the activity of the anterior
medial temporal cortex and N400 generation. Furthermore, the
effect of the prime-target SOA difference was reflected in the
anterior cingulate cortex in their study. However, they recorded
ERP and fMRI data from different groups of subjects, making it
impossible to directly characterize the relationship between the
effect of semantic priming in BOLD activity and the ERP
In the present study, we employed ERP and event-related fMRI
recording in separate sessions and in the same subjects in order to
investigate the correlation between the magnitude of the N400
priming effect and the repetition suppression effect in the BOLD
signal caused by semantic priming. Since recent fMRI study has
suggested that similar BOLD responses were evoked when
repeating the same experiment (Wei et al., 2004), it is reasonable
that ERP and fMRI data measured in separated sessions are
integrated in our analysis. The correlation analysis between the
BOLD signal and an ERP component provide an alternative
method of examining the relationship between the neural activity
in a specific brain region and an ERP component (Horovitz et al.,
2002; Liebenthal et al., 2003; Opitz et al., 2001). Further, this
correlation analysis may identify the N400 priming effect in a
different manner than the equivalent current dipole (ECD) and
distributed source modeling methods used in MEG and EEG
studies. We conducted the correlation analysis in seven regions.
A. Matsumoto et al. / NeuroImage 24 (2005) 624–634
Four of these regions have been associated with N400 generation
in previous studies, and the other regions showed significant
repetition suppression effects in whole brain subtraction analysis.
Furthermore, we used a new ERP analysis technique, independent
component analysis (ICA), to eliminate the contamination of N400
by other ERP components. ICA enables us to conduct blind source
separation of a linear mixture of sources in an electroencephalo-
gram that is spatially fixed and temporally independent. Compo-
nents were determined using a neural network in order to train
unmixed weighted matrices that maximize the joint entropy
between nonlinearly transformed channel data (Makeig et al.,
1997). By using ICA, other components that overlap N400 based
on the scalp distribution were excluded, thereby resulting in a more
precise examination and correlation analysis.
Based on previous studies (Horovitz et al., 2002; Liebenthal et
al., 2003; Opitz et al., 2001) that demonstrated that ERP amplitude
correlated with BOLD activity across subject, we hypothesized that
the repetition suppression effect would be reflected in BOLD
activity. Further, the subjects with greater BOLD repetition
suppression effect in regions related to N400 generation would
show greater N400 priming effect, and the regions involved in the
semantic priming, such as the inferior frontal gyrus and the superior
or middle temporal gyrus of the left hemisphere, would show
significant repetition suppression effects in BOLD response, the
magnitude of which would correlate with the N400 priming effect.
Materials and methods
Electrophysiological and hemodynamic brain responses were
measured in 12 healthy right-handed volunteers (six males and six
females, mean age F SD, 20.8 F 1.6 years). Half of the subject
group participated in the fMRI experiment before the ERP
recording, and the other half of the subject group took part in
the fMRI experiment after the ERP experiment. Data from subject
was excluded from fMRI analysis because of an excess of head
motion artifacts. Before the experiment, written informed consent
was obtained from all subjects. This study was approved by the
ethics committee of the National Institute for Physiological
Stimuli consisted of 400 Japanese three-letter nouns and 70
pronounceable pseudo-words. All stimuli were presented visually
in Katakana script (a Japanese syllabic script) that were arranged
into 230 prime-target pairs; 80 pairs contained semantically related
words, another 80 pairs contained semantically unrelated words,
and the remaining 70 pairs had words followed by pseudo-words.
The stimuli were divided into two shorter lists (Lists 1 and 2), with
each list containing 40 semantically related pairs (related), 40
semantically unrelated pairs (unrelated), and 35 word–pseudo-
word pairs (pseudo). Assignment of the list to ERP and fMRI
recording was counterbalanced across subjects. Word frequency
(Amano and Kondo, 2000) was controlled for target words across
each condition and list. After the experiment, the strength of the
semantic relationship of each pair was evaluated in all subjects
using a 5-point scale. Table 1 summarizes the information
regarding stimulus material. Two-way ANOVA (List ? Condition)
revealed a significant main effect of Condition (F(1,9) = 311.86,
P b 0.0000001), suggesting that the strength of the semantic
relationship in related word pairs was significantly larger than that
in unrelated word pairs. No main effect of the List was observed.
During the task, 40 related, 40 unrelated, and 35 pseudo-events
were randomly presented along with 35 null events and fixation
points in two separate runs. After the presentation of fixation for
400 ms, a prime word was presented for 400 ms followed by an
inter-stimulus interval of 200 ms and a target word that was
presented for 1000 ms. The inter-trial interval was 3000 ms. Trial
scheme is illustrated in Fig. 1. Subjects were instructed to decide
whether the target word was a real word or a pseudo-word by
pressing unique buttons, and subject responses and reaction times
were recorded. An identical task procedure was used in for both the
ERP and the fMRI experiments.
The behavioral semantic priming effect was assessed by
repeated measures analysis of variance (ANOVAs), with compar-
ison of the condition (related, unrelated) and recording (ERP,
fMRI) with reaction time. Post hoc comparisons were conducted
using the Least Significant Difference (LSD) test.
The electroencephalograph was recorded from 14 international
10–20 system scalp locations (Fz, F3, F4, Cz, C3, C4, Pz, P3, P4,
T5, T6, Oz, O1, O2) that were referenced to the tip of the nose. Eye
movement was monitored by an electrode placed on the supra-
ridge of the left eye. Inter-electrode impedances were set below 5.
The EEG and EOG data were filtered using a bandpass of 0.5–
60 Hz. The data were digitized with an A–D conversion rate of
1000 Hz and sampled from a 300-ms preceding target word in a
trial until 1400 ms after the stimulus onset. Digital codes were sent
from the stimulus-presentation computer to mark the onset and
type of each target stimulus. EEG data were corrected to a 100-ms
baseline prior to the onset of the target word. Trials in which the
EEG or eye movements exceeded plus or minus 50 AV were
automatically rejected from the averaging process. ERPs were
analyzed using repeated measures ANOVAs that were conducted
on the mean amplitude of a 300–500 ms time window to compare
factors of condition (related, unrelated) and the electrode (Fz, F3,
F4, Cz, C3, C4, Pz, P3, P4, T5, T6, Oz, O1, O2). In the present
study, semantic priming seems to affect the latency of the late
positive component (LPC) as well as the amplitude of N400.
Therefore, we assessed the peak latencies and amplitudes of LPC at
Fz, Cz, and Pz. According to their appearance in the grand average
waveforms, the LPC peak was defined as the most positive voltage
within the interval of 450–700 ms. The analysis was performed to
compare factors of condition (related, unrelated) and the electrode
Mean and standard deviation of word familiarity of target words and
strength of prime-target semantic relatedness
A. Matsumoto et al. / NeuroImage 24 (2005) 624–634
(Fz, Cz, Pz). The Geisser–Greenhouse correction was always
applied when evaluating effects with more than one degree of
freedom were present in the numerator. Post hoc comparisons were
conducted using the Least Significant Difference (LSD) test.
Correlation analysis was conducted using a component
decomposed by ICA to diminish the effect of other component,
such as LPC. The difference wave of an unrelated minus-related
condition was computed for each subject in order to monitor the
N400 priming effect elicited by semantic priming. After the
decomposing original waveforms of 12 subjects were character-
ized, the amplitude of the N400 priming effect was correlated with
signal changes measured by fMRI. The LPC of the related
condition peaked earlier than that of unrelated condition; therefore,
the later part of the N400 priming effect on difference waves could
be contaminated by the effect of the LPC latency difference. In
order to diminish the effect of the LPC and/or other components
from the difference wave, we applied ICA to difference wave data.
The ICA was conducted using Psychophysiological Analysis
Software ver. 4.0, provided by the Computational Neuroscience
Laboratory of the Salk Institute, CA, USA, (http://www.cnl.
salk.edu/~tewon/ica_cnl.html) implemented in MATLAB version
6.1 (Mathworks). We conducted a single ICA analysis with
concatenated data from all subjects. This analysis enabled us to
obtain the N400 priming effect that has the same scalp distribution
for all subjects. After the ICA decomposition, a component that
was dominant at the centro-parietal region (IC1) was chosen from
14 decomposed components, and the peak amplitude within 350–
450 ms in the component was assessed in all subjects and used for
correlation analysis. Finally, correlation analysis was performed on
the differences of standard N400 amplitude (e.g., the mean
amplitude of 300–500 ms time window) to investigate whether a
correlation was present only when ICA was applied to the data.
fMRI data acquisition and analysis
Functional images of the whole brain were acquired in an axial
orientation usinga3-TSiemens AllegraMRI scannerequippedwith
matrix and 26 slices, voxel size = 3 ? 3 ? 4 mm) sensitive to BOLD
contrast (Ogawa et al., 1992). After discarding the first six images,
the next 166 successive images in each run were subjected to
analysis. An anatomical T1-weighted image was also acquired
(MPRAGE, TR = 3 s, TE = 4.6 ms, Flip Angle = 908, 256 ? 256
matrix and 26 slices, voxel size = 0.75 ? 0.75 ? 4 mm) for each
subject. The fMRI experiment was controlled using Presentation
software (Neurobehavioral Systems Inc. Albany, CA, USA).
Data were analyzed using SPM99 (Wellcome Department of
Imaging Neuroscience, London, UK). First, all volumes were
realigned spatially to the last volume, and the signal in each slice
was realigned temporally to that obtained in the middle slice using
a sinc interpolation. The resliced volumes were normalized to the
standard space of Talairach and Tournoux (1988) using a
transformation matrix obtained from the normalization process of
the T1-weighted anatomical image of each individual subject to the
T1 template image. The T1-weighted anatomical image was
coregistered to the mean EPI image in each subject. The
normalized images were spatially smoothed using an 8-mm
Gaussian kernel. Following pre-processing, statistical analysis of
each individual subject was conducted using the general linear
model. The hemodynamic response triggered by the target word in
each condition was modeled with two-basis function, that is, a
hemodynamic response function (HRF) and its temporal deriva-
tive. Low-pass and high-pass frequency filters were applied to the
time-series data. The images were scaled to a grand mean of 100
over all voxels and scans within a session. In the subtraction
analysis, four conditions (correct response for related target, correct
response for unrelated target, correct response for pseudoword
response, and incorrect response) were modeled separately.
Parameter estimates for each condition and for the difference
between the conditions were calculated from the least-mean-square
fit of the model to the time-series data. Images of parameter
estimates representing event-related activity at each voxel for each
condition and each subject were created.
At the second level, the results for each subject were entered
into the random effects model by applying t tests between the
contrast images to create a group statistical parametric map (SPM).
An SPM of voxels showing a significant response to stimulus
presentation versus the baseline and differences in the response
between the conditions was also created.
Fig. 1. (1) Task scheme: subjects were asked to judge whether or not the
target word is a real word. Under the related condition (the prime word
means bcheeseQ; the target word means bbutterQ), the prime word and the
target word semantically related, while with the unrelated condition (the
prime word means bgolf Q; the target word means bmouseQ), there was no
semantic relation between the prime word and the target word. Under the
nonword condition, a nonword was presented as a target word (prime word
means bballQ). (2) Behavioral results: mean RT and standard errors for the
related and unrelated target words in the ERP and fMRI sessions.
A. Matsumoto et al. / NeuroImage 24 (2005) 624–634
Whole brain analysis
The first analysis was performed by entering contrast images
containing parameter estimates of the difference in activity
between the related and unrelated conditions into a one-sample t
test to reveal the hemodynamic repetition suppression, that is, the
unrelated condition versus the related condition. Since we had
assumed that several brain regions (e.g. left inferior frontal cortex )
would show significant repetition suppression effect, the statistical
threshold was set at P = 0.001 without corrections for multiple
comparisons. The extent threshold was set at k = 10 voxels. Region
names, coordinates, and Z values are summarized in Table 2. The
distributions of regions that showed significant hemodynamic
repetition suppression are illustrated in Fig. 4.
Three regions that showed significant repetition suppression in
whole brain analysis were selected as regions of interest (ROIs)
[the dorsal left inferior frontal gyrus (?50, 28, 14), the ventral left
inferior frontal gyrus (?30, 22, ?10), and the anterior cingulate
cortex (0, 26, 44)]. For each location, a three-dimensional-ROI was
defined to include all voxels within a sphere of 8 mm radius by
using MarsBar software (http://www.mrc.cbu.cam.ac.uk/Imaging/
In addition, eight regions that were thought to be associated
with the generation of N400 in previous studies were selected as
ROIs. These regions included the amygdala–hippocampus (?23, 0,
?18), the anterior fusiform gyrus (?30, ?21, ?21), anterior
medial temporal cortex (?40, 14, ?34) and the superior temporal
gyrus (?57, ?12, ?1) of the left hemisphere and the correspond-
ing regions of the right hemisphere. In these regions, magnitude of
activation was extracted from an 8-mm spherical ROI centered on
each coordinate. These coordinates were derived from McCarthy et
al. (1995), Rissman et al. (2003), and Rossell et al. (2003).
First, for eight regions, t tests between conditions were
performed on the signals obtained in each ROI to investigate the
repetition suppression effect that was not characterized by whole
brain subtraction analysis. Second, we addressed the question of
whether the magnitude of the N400 priming effect measured by
ERP would correlate with repetition suppression of the BOLD
signal in these ROIs. The contrast images of 11 subjects pertaining
to the difference in activity between the unrelated and related
conditions and the amplitude of independent components derived
from the unrelated–related difference wave in each subject were
entered into a simple regression analysis.
Behavioral data from ERP and fMRI sessions are presented in
Fig. 1. There was a significant main effect on reaction time of
condition (F(1, 11) = 64.85, P b 0.00001). Post hoc analysis
revealed significant semantic priming, as measured by diminished
reaction time for related conditions compared with unrelated
conditions under both ERP and fMRI experiments (P b
0.000001). Neither the main effect due to recording nor
priming ? recording interaction was significant.
Grand-averaged waveforms elicited by targets under the related
and unrelated condition from all electrodes are shown in Fig. 2. The
amplitude of N400 was reduced under the related condition as
of the300–500mstimewindowrevealeda significantmaineffectof
condition (F(1, 10) = 23.10, P b 0.001), a main effect of the
electrode (F(2.6, 21.2) = 16.64, P b 0.000001), and condition ?
Brain areas demonstrating significant repetition suppression in whole brain
Region name (BA)L/R
x, y, zt value
Dorsal inferior frontal gyrus (45)
Ventral inferior frontal gyrus (47)
Anterior cingulate cortex (8/32)
?50, 28, 14
?30, 22, ?10
0, 26, 44
Fig. 2. Grand-averaged ERP under the related and unrelated condition at all sites. Positive polarity is plotted upward. Waveforms elicited by the target words
are depicted. The N400 component (black arrow) is attenuated under the related condition, and the LPC peak (gray arrow) is delayed under the unrelated
A. Matsumoto et al. / NeuroImage 24 (2005) 624–634
electrode interaction (F(13, 130) = 3.21, P b 0.0001). Post hoc
analysisrevealed that N400wassignificantlylargerfortheunrelated
condition than for the related condition at all sites (P b 0.000001).
The peak latency of LPC was slightly delayed under the unrelated
condition, but the difference did not reach statistical significance
(F(1, 11) = 4.72, P b 0.1). There was no significant effect in LPC
peak amplitude. Fig. 3 shows the ICA component derived from
concatenated data from all subjects and the scalp distribution.
Whole brain analysis
We hypothesized that regions showing greater activation under
the unrelated condition than under the related condition were
involved in repetition suppression induced by semantic priming.
As shown in the top panel of Fig. 4, when contrasted with the
unrelated condition, the related condition was associated with
decreased activity in the anterior cingulate cortex (BA 8/32), the
Fig. 3. (1) Difference waves between the unrelated and related conditions at the Fz, Cz, and Pz sites. Negative component peaking between 350 and 450 ms is
observed at all sites. Decomposed component and the scalp distribution. The component obtained from a single ICA with concatenated data from all subjects
and the time course of the activity from three representative subjects are shown. The component is dominant at the centro-parietal region and peaked between
350 and 450 ms.
Fig. 4. Brain areas demonstrating significant repetition suppression in whole brain analysis. Activations are displayed on a surface rendering of a typical brain
(top panel). The dorsal part of the left inferior frontal gyrus (A), the ventral part of the left inferior frontal gyrus (B), and the anterior cingulate cortex (C) are
more strongly activated under the unrelated condition than under the related condition. Lower graphs depict the mean percent signal change (column) and the
standard deviation (bar) in the peaks of these regions.
A. Matsumoto et al. / NeuroImage 24 (2005) 624–634
dorsal left inferior frontal cortex (BA 45), and the ventral left
inferior frontal cortex (BA47). In contrast, no region exhibited
greater activation under the related than under the unrelated
condition at this threshold.
The activation patterns in four priori ROIs and the regions that
showed significant repetition suppression effect were examined.
These included regions within the amygdala, the anterior fusiform
gyrus, the superior temporal gyrus, the anterior medial temporal
lobe, the dorsal left inferior frontal gyrus, the ventral left inferior
frontal gyrus, and the anterior cingulate cortex. t tests were
conducted separately for the priori ROIs with regard to activation.
For the left superior temporal gyrus, activation for the unrelated
condition was significantly larger when compared with that for the
related condition (t (10) = 2.49, P b 0.05), while the difference was
not significant in the right hemisphere (t (10) = 1.26, P N 0.2). The
result indicates that the repetition suppression is predominant in the
left hemisphere. There was no significant repetition suppression
effect in the amygdala–hippocampus, anterior fusiform gyrus, or
anterior medial temporal lobe, in either hemisphere.
The superior temporal gyrus in the left hemisphere exhibited
significant positive correlation between the magnitude of the N400
priming effect and the BOLD repetition suppression effect.
Subjects that showed a greater N400 priming effect exhibited a
greater BOLD repetition suppression effect in the left superior
temporal gyrus (r = 0.72, df = 9, P b 0.05). Although the dorsal
and ventral inferior frontal gyrus and the anterior cingulate gyrus
showed a significant BOLD repetition suppression effect, the
magnitude of BOLD response did not correlate with the amplitude
of N400 priming effect across subjects (see Fig. 5). There was no
significant BOLD-N400 correlation in the anterior fusiform gyrus,
anterior medial temporal lobe and amygdala. Furthermore, N400
priming effect assessed by standard N400 amplitude measures did
not show significant correlation with the repetition suppression
effect in all ROIs.
The goal of this study was to characterize the neural basis of the
N400 priming effect in semantic priming. We recorded ERP and
fMRI from the same group of subjects in separate sessions and
conducted correlation analysis between the N400 priming effect
decomposed by ICA and the BOLD repetition suppression effect.
In the whole brain analysis, we found that semantic priming caused
repetition suppression in the dorsal and ventral part of the left
inferior frontal gyrus and in the anterior cingulate cortex. Further,
ROI analysis showed significant repetition suppression in the left
superior temporal gyrus. Correlation analysis revealed that the
magnitude of the N400 priming effect was correlated with the
activity in the left superior temporal gyrus.
The behavioral results showed a clear semantic priming effect
during the lexical decision task. That is, the reaction time for the
target word preceded by a semantically related word was shorter
than that for the target word preceded by a semantically unrelated
word, suggesting that semantic priming facilitated the processing
of the target word.
The present study demonstrated that the amplitude of N400 was
attenuated by semantic, which is consistent with observations by
other investigators (Anderson and Holcomb, 1995; Bentin, 1987;
Bentin et al., 1985; Hill et al., 2002; Holcomb and Anderson,
1993). The ERP waveforms of the present study and the scalp
distribution resemble that reported by Rossell et al. (2003).
Although the amplitude of N400 was attenuated by semantic
priming, the present study also demonstrated that semantic priming
effect resulted in delayed peak latency of the late positive
component (LPC) under the unrelated condition. According to
Bentin et al. (1985), mechanisms that evoke the LPC component in
lexical decisions can be interpreted as P300, which is associated
with response decisions or the updating of working memory (Kok,
2001). Thus, it seems plausible that LPC latency is delayed under
the unrelated condition because a lexical decision requires a more
demanding process for decisions or for the evaluation of a target
word under the unrelated condition than under the related
condition. However, the delay of the LPC peak caused by semantic
priming could contaminate the N400 priming effect in the
unrelated–related difference wave and make it difficult to identify
the source of N400. Therefore, we used ICA to exclude
components that have inappropriate temporal or spatial distribution
in the difference wave for each subject, thereby avoiding
contamination of the other ERP components by the obtained
N400 priming effect and resulting in a more precise correlation
analysis for identification of the neural basis of the N400 priming
effect. The N400 priming effect assessed by the original N400
waveform did not have significant correlation with the repetition
suppression effect in all ROIs, suggesting that the original
waveforms might be contaminated with other components.
Whole brain fMRI analysis showed that there was significant
hemodynamic repetition suppression effect in the dorsal and
ventral left inferior frontal gyrus and in the anterior cingulate
cortex. The anterior medial temporal activity, which displayed
significant repetition suppression effect in Rossell et al. (2003), did
not differ between the related and unrelated condition. Whereas
Rossell’s study used a fixed-effects analysis, the present study used
a random-effects analysis. Friston et al. (1999) indicated that fixed-
effect analysis can yield results different from those obtained from
random-effects analysis, because the fixed-effects analysis is more
sensitive than the random-effects analysis. Thus, it is possible that
the difference of analysis technique (i.e. fixed-effects model vs.
random-effects model) contributed to the difference in results
between the present study and those reported by Rossell et al.
The left inferior frontal gyrus has been of particular interest to
researchers because of its strong association with the semantic
processing of words or objects. However, it is unlikely that this
region mediates the preservation of semantic memory representa-
tion, since lesions in this region do not cause deficits in semantic
knowledge (Gershberg, 1997; Swick, 1998). The left inferior
frontal gyrus mediates selection among competing alternatives
regulated by semantic knowledge, recent experience (Thompson-
Schill et al., 1997) or the central executive of retrieval of semantic
representation from semantic memory (Wagner et al., 2001). Thus,
the semantic priming effect (i.e., the repetition suppression effect)
of the left inferior frontal gyrus indicates that the selection or
retrieval of semantic representation for unrelated targets is more
demanding than that for related targets, possibly because the
processing of target words preceded by a semantically related word
was facilitated by conscious and unconscious priming effects, such
as automatic spreading activation or the expectancy effect.
Some researches have reported that the dorsal and ventral parts
of the left inferior frontal gyrus play different roles in semantic
processing (Bokde et al., 2001; Noppeney and Price, 2002).
Noppeney and Price (2002) suggested that the ventral part of left
A. Matsumoto et al. / NeuroImage 24 (2005) 624–634
anterior inferior frontal gyrus is involved in semantic retrieval or in
the semantic evaluation of stimulus, while the dorsal left inferior
frontal gyrus is associated with the general processing of semantic
or phonological information. Although we found significant
repetition suppression effect in the ventral part of the left inferior
frontal gyrus (BA47) and in the dorsal part of the left inferior
frontal gyrus (BA45), other visual semantic priming studies have
not found the repetition suppression effect in the dorsal part of the
inferior frontal gyrus (Copland et al., 2003; Rossell et al., 2003).
For example, Copland et al. (2003) found significant repetition
suppression effect only in the ventral left inferior frontal gyrus
(BA47). We speculate that the discrepancy between the results of
present study and those of Copland et al. (2003) can be attributed
to differences in procedural details; Copeland et al. focused on the
unconscious aspects of semantic priming and used short SOA
between the prime and the target word (150 ms), which can
attenuate the conscious semantic priming effect (e.g., semantic
matching or semantic integration processes). In contrast, the
present study used relatively long prime-target SOA (600 ms),
resulting in the semantic priming effect containing the effect of
automatic spreading activation and the effect of semantic matching
or semantic integration processes (Hill et al., 2002). Thus, it is
Fig. 5. Correlation between the magnitude of repetition suppression on BOLD and the magnitude of the N400 priming effect in the regions that show
significant repetition suppression. The horizontal axis represents the magnitude of the repetition suppression effect, and the vertical axis represents the
magnitude of the N400 priming effect. The correlation is significant only at the left superior temporal gyrus.
A. Matsumoto et al. / NeuroImage 24 (2005) 624–634
possible that the repetition suppression of the dorsal left inferior
frontal gyrus is associated with modulation in several types of
semantic processing, such as semantic matching or the semantic
integration process, while that of the ventral left inferior frontal
gyrus reflects the reduction of demand for retrieval or the selection
from semantic knowledge caused by automatic spreading activa-
tion. Further investigation to clarify the functional difference of
these two regions would be of benefit.
The anterior cingulate cortex also exhibited a significant
repetition suppression effect. It has been reported that the anterior
cingulate cortex plays an important role in the detection of errors
and in behavioral conflict, and is activated under the condition in
which response competition occurs (Gehring and Knight, 2000;
Kiehl et al., 2000). Thus, increased anterior cingulate activity under
the unrelated condition implies that a higher degree of response
conflict occurred under the unrelated condition than under the
related condition because conscious and unconscious semantic
priming can make it easier to make a decision regarding a target
word under the related condition.
In the ROI analysis, the left superior temporal gyrus showed
significant repetition suppression, and the magnitude of the effect
correlated with the magnitude of the N400 priming effect
decomposed by ICA across the subjects. Although N400 source
is assumed to be mediated by several brain regions (Guillem et al.,
1999; Nobre and McCarthy, 1995; Nobre et al., 1994), a number of
MEG studies estimated the source of N400 at the bilateral superior
temporal gyrus adjacent to the auditory cortex (Helenius et al.,
1998, 2002; Laine et al., 2000; Sekiguchi et al., 2001; Simos et al.,
1997). Furthermore, several studies have shown that the magnitude
of magnetic responses elicited by target words whose source is
estimated in the left peri-Sylvian area was reduced by semantic or
repetition priming for visual words (Koyama et al., 1999;
Sekiguchi et al., 2000, 2001). Thus, the results of present study
and those of previous studies suggest that the priming effect
observed in N400 is associated with the activity of the superior
Neuroimaging studies have reported that the left superior
temporal region is activated when subjects perform phonological
tasks for visually presented words or letters (Fujimaki et al., 1999;
Paulesu et al., 1993; Sergent et al., 1992). Bavelier et al. (1997)
showed that this area was more strongly activated by the visually
presented phonologically legal pronounceable word than by the
phonologically illegal consonant word. Herbster et al. (1997) also
reported that the left superior temporal cortex was related to the
processing of pronounceable words. Based on these observations,
Binder and Price (2001) suggested that the left superior temporal
cortex mediates the preservation of the phonological representation
of words. Notably, several studies have indicated the that N400 and
phonological processing are related. Nobre and McCarthy (1994)
reported that N400 was elicited by phonologically legal words but
not by phonologically illegal nonwords, indicating that N400 is
associated with access to the phonological representation of words.
Similar results were reported by Bentin et al. (1999) and Rugg and
Nagy (1987). The consistent result with regard to the involvement
of N400 response and superior temporal activity in word-related
phonological information supports the finding of the present study
that the N400 priming effect is correlated with the activity of the
superior temporal gyrus. If N400 reflects the activation of the
superior temporal cortex, then the N400 priming effect in the
present study suggests that semantic priming modifies phonolog-
Recent behavioral studies emphasize the importance of
phonological information in the recognition of visual words. For
example, Drieghe and Brysbaert (2002) showed that a target word
(e.g., bfrogQ) could be primed not only by an associated word (e.g.,
btoadQ) but also by a homophone (e.g., btowedQ) and pseudoho-
mophone (e.g., btodeQ) of the associated word, indicating that
semantic representations are coded as a word by a phonological
lexicon. According to this model, in a semantic priming task, a
prime word can activate the semantic representation of the word,
and the activation of semantic representation can activate the
phonological lexicon that is related to the semantic representation.
This would result in easier access to the phonological lexicon of a
target word when a subject makes a semantic judgment. If N400
reflects the access to the phonological lexicon of a word, the N400
priming effect in semantic or repetition priming may be mediated
by easier access to the phonological lexicon of the target word due
to its prior activation by the prime word.
Although the dorsal and ventral part of the left inferior frontal
gyrus and the anterior cingulate cortex showed significant
repetition suppression, the effect did not correlate with the
magnitude of the N400 priming effect from ICA component. This
would imply that the activities of these regions are independent of
the processing reflected in N400. As discussed above, these
regions may be related to retrieval or selection from semantic
memory or to response selection or verbal working memory.
Selection or working memory activity is associated with the
amplitude and latency of the LPC rather than the N400 (Kok,
2001). Thus, activity of the left and right inferior frontal gyrus and
anterior cingulate gyrus may be reflected in the LPC. Swick (1998)
reported that subjects with lesions in the left inferior frontal cortex
showed smaller repetition priming effect on the LPC when
compared to control subjects. This result supports the notion that
left inferior frontal gyrus activity is associated with the LPC.
However, it is difficult to investigate the correlation between the
latency or amplitude of LPC and BOLD responses because LPC is
a compound component consisting of working memory, episodic
memory retrieval, reallocation of cognitive resource, and response
selection. Therefore, we could not characterize the relationship
between the repetition suppression effect in the inferior frontal
gyrus or anterior cingulated cortex and the LPC in the present
study. Further investigation to examine the relationship between
the LPC and the inferior frontal gyrus or anterior cingulate cortex
would be of benefit.
In conclusion, this study provides new evidence that the
semantic priming effect on the N400 component is associated
with the activity of the left superior temporal cortex. The N400
priming effect decomposed by ICA was correlated with the
repetition suppression effect of the BOLD signal caused by
semantic priming at the left superior temporal gyrus but not with
the BOLD repetition suppression effect of the left inferior frontal
gyrus. This result is supported by MEG studies that suggest that the
source of N400m is the superior temporal cortex and by neuro-
imaging studies showing that the left superior temporal gyrus
mediates the processing of the phonological representation of
words. The present results indicate that the N400 priming effect
reflects the modification of the access to the phonological lexicon
of words and that the access to the phonological lexicon plays an
important role in the semantic priming task.
A. Matsumoto et al. / NeuroImage 24 (2005) 624–634
We wish to thank Dr. Hideki Ohira for his helpful comments on
this study. We also thank Prof. Kazuhiko Kakehi for his help in
Amano, S., Kondo, T., 2000. Lexical Properties of Japanese, NTT database.
Anderson, J.E., Holcomb, P.J., 1995. Auditory and visual semantic priming
using different stimulus onset asynchronies: an event-related brain
potential study. Psychophysiology 32, 177–190.
Badgaiyan, R.D., Schacter, D.L., Alpert, N.M., 1999. Auditory priming
within and across modalities: evidence from positron emission
tomography. J. Cogn. Neurosci. 11, 337–348.
Bavelier, D., Corina, D., Jezzard, P., Padmanabhan, S., Clark, V.P., Karni,
A., Prinster, A., Braun, A., Lalwani, A., Rauschecker, J.P., Turner, R.,
Neville, H., 1997. Sentence reading: a functional MRI study at 4 Tesla.
J. Cogn. Neurosci. 9, 664–686.
Bentin, S., 1987. Event-related potentials, semantic processes, and expect-
ancy factors in word recognition. Brain Lang. 31, 308–327.
Bentin, S., McCarthy, G., Wood, C.C., 1985. Event-related potentials,
lexical decision and semantic priming. Electroencephalogr. Clin.
Neurophysiol. 60, 343–355.
Bentin, S., Mouchetant-Rostaing, Y., Giard, M.H., Echallier, J.F., Pernier,
J., 1999. ERP manifestations of processing printed words at different
psycholinguistic levels: time course and scalp distribution. J. Cogn.
Neurosci. 11, 235–260.
Binder, J., Price, C.J., 2001. Functional Neuroimaging of Language. In:
Cabeza, R., Kingstone, A. (Eds.), Handbook of Functional Neuro-
imaging of Cognition. The MIT Press, Cambridge, pp. 157–252.
Bokde, A.L., Tagamets, M.A., Friedman, R.B., Horwitz, B., 2001.
Functional interactions of the inferior frontal cortex during the
processing of words and word-like stimuli. Neuron 30, 609–617.
Buckner, R.L., Petersen, S.E., Ojemann, J.G., Miezin, F.M., Squire, L.R.,
Raichle, M.E., 1995. Functional anatomical studies of explicit and
implicit memory retrieval tasks. J. Neurosci. 15, 12–29.
Buckner, R.L., Koutstaal, W., Schacter, D.L., Rosen, B.R., 2000. Func-
tional MRI evidence for a role of frontal and inferior temporal cortex in
amodal components of priming. Brain 123, 620–640.
Copland, D.A., de Zubicaray, G.I., McMahon, K., Wilson, S.J., Eastburn,
M., Chenery, H.J., 2003. Brain activity during automatic semantic
priming revealed by event-related functional magnetic resonance
imaging. NeuroImage 20, 302–310.
Desimone, R., 1996. Neural mechanisms for visual memory and their role
in attention. Proc. Natl. Acad. Sci. U. S. A. 93, 13494–13499.
Donaldson, D.I., Petersen, S.E., Buckner, R.L., 2001. Dissociating memory
retrieval processes using fMRI: evidence that priming does not support
recognition memory. Neuron 31, 1047–1059.
Drieghe, D., Brysbaert, M., 2002. Strategic effects in associative priming
with words, homophones, and pseudohomophones. J. Exper. Psychol.,
Learn., Mem., Cogn. 28, 951–961.
Fischler, I., Raney, G.E., 1989. Language by eye: behavioral, autonomic,
and cortical approaches to reading. In: Jennings, J.R., Coles, M.G.H.
(Eds.), Handbook of Cognitive Psychology: Central and Autonomic
Nervous System. Wiley, New York, pp. 511–574.
Friston, K.J., Holmes, A.P., Price, C.J., Buchel, C., Worsley, K.J., 1999.
Multisubject fMRI studies and conjunction analyses. NeuroImage 10,
Fujimaki, N., Miyauchi, S., Putz, B., 1999. Functional magnetic resonance
imaging of neural activity related to orthographic, phonological, and
lexico-semantic judgments of visually presented characters and words.
Hum. Brain Mapp. 8, 44–59.
Gehring, W.J., Knight, R.T., 2000. Prefrontal-cingulate interactions in
action monitoring. Nat. Neurosci. 3, 516–520.
Gershberg, F.B., 1997. Implicit and explicit conceptual memory following
frontal lobe damage. J. Cogn. Neurosci. 9, 105–116.
Guillem, F., Rougier, A., Claverie, B., 1999. Short- and long-delay
intracranial ERP repetition effects dissociate memory systems in the
human brain. J. Cogn. Neurosci. 11, 437–458.
Halgren, E., Baudena, P., Heit, G., Clarke, J.M., Marinkovic, K., Clarke,
M., 1994. Spatio-temporal stages in face and word processing: I. Depth-
recorded potentials in the human occipital, temporal and parietal lobes.
J. Physiol. (Paris) 88, 1–50.
Helenius, P., Salmelin, R., Service, E., Connolly, J.F., 1998. Distinct time
courses of word and context comprehension in the left temporal cortex.
Brain 121, 1133–1142.
H., 2002. Cortical activation during spoken-word segmentation in
nonreading-impaired and dyslexic adults. J. Neurosci. 22, 2936–2944.
Henson, R.N., 2003. Neuroimaging studies of priming. Prog. Neurobiol.
Herbster, A.N., Mintun, M.A., Nebes, R.D., Becker, J.T., 1997. Regional
cerebral blood flow during word and nonword reading. Hum. Brain
Mapp. 5, 84–92.
Hill, H., Strube, M., Roesch-Ely, D., Weisbrod, M., 2002. Automatic vs.
controlled processes in semantic priming—Differentiation by event-
related potentials. Int. J. Psychophysiol. 44, 197–218.
Holcomb, P.J., 1993. Semantic priming and stimulus degradation.
Implications for the role of the N400 in language processing.
Psychophysiology 30, 47–61.
Holcomb, P.J., Anderson, J.E., 1993. Cross-modal semantic priming: a
time-course analysis using event-related brain potentials. Lang. Cogn.
Processes 8, 379–411.
Horovitz, S.G., Skudlarski, P., Gore, J.C., 2002. Correlations and
dissociations between BOLD signal and P300 amplitude in an auditory
oddball task: a parametric approach to combining fMRI and ERP.
Magn. Reson. Imaging 20, 319–325.
Kiefer, M., Spitzer, M., 2000. Time course of conscious and unconscious
semantic brain activation. NeuroReport 11, 2401–2407.
Kiehl, K.A., Liddle, P.F., Hopfinger, J.B., 2000. Error processing and the
rostral anterior cingulate: an event-related fMRI study. Psychophysio-
logy 37, 216–223.
Kok, A., 2001. On the utility of P3 amplitude as a measure of processing
capacity. Psychophysiology 38, 557–577.
Kotz, S.A., Cappa, S.F., von Cramon, D.Y., Friederici, A.D., 2002.
Modulation of the lexical-semantic network by auditory semantic
priming: an event-related functional MRI study. NeuroImage 17,
Koyama, S., Naka, D., Kakigi, R., 1999. Evoked magnetic responses during
a word completion task. Electroencephalogr. Clin. Neurophysiol.,
Suppl. 49, 174–178.
Kutas, M., Hillyard, S.A., 1980. Reading senseless sentences: brain
potentials reflect semantic incongruity. Science 207, 203–205.
Laine, M., Salmelin, R., Helenius, P., Marttila, R., 2000. Brain activation
during reading in deep dyslexia: an MEG study. J. Cogn. Neurosci. 12,
Liebenthal, E., Ellingson, M.L., Spanaki, M.V., Prieto, T.E., Ropella, K.M.,
Binder, J.R., 2003. Simultaneous ERP and fMRI of the auditory cortex
in a passive oddball paradigm. NeuroImage 19, 1395–1404.
Makeig, S., Jung, T.P., Bell, A.J., Ghahremani, D., Sejnowski, T.J.,
1997. Blind separation of auditory event-related brain responses
into independent components. Proc. Natl. Acad. Sci. U. S. A. 94,
Marinkovic, K., Dhond, R.P., Dale, A.M., Glessner, M., Carr, V., Halgren,
E., 2003. Spatiotemporal dynamics of modality-specific and supra-
modal word processing. Neuron 38, 487–497.
McCarthy, G., Nobre, A.C., Bentin, S., Spencer, D.D., 1995. Language-
related field potentials in the anterior–medial temporal lobe: I.
Intracranial distribution and neural generators. J. Neurosci. 15,
Neely, J.H., 1977. Semantic priming and retrieval from lexical memory:
A. Matsumoto et al. / NeuroImage 24 (2005) 624–634
roles of inhibitionless spreading activation and limited-capacity
attention. J. Exp. Psychol. 106, 226–254.
Neely, J.H., 1991. Semantic priming effects in visual word recognition: a
selective review of current findings and theories. In: Besner, D.,
Humphreys, G.W. (Eds.), Basic Progress in Reading-Visual Word
Recognition. Erlbaum, Hillsdale, pp. 264–333.
Neville, H.J., Nicol, J., Brass, A., Forster, K.I., Garrett, M.F., 1991.
Syntactically based processing classes: evidence from event-related
potentials. J. Cogn. Neurosci. 3, 151–165.
Nobre, A.C., McCarthy, G., 1994. Language-related ERPs: scalp distribu-
tions and modulation by word type and semantic priming. J. Cogn.
Neurosci. 6, 233–255.
Nobre, A.C., McCarthy, G., 1995. Language-related field potentials in the
anterior–medial temporal lobe: II. Effects of word type and semantic
priming. J. Neurosci. 15, 1090–1098.
Nobre, A.C., Allison, T., McCarthy, G., 1994. Word recognition in the
human inferior temporal lobe. Nature 372, 260–263.
Noppeney, U., Price, C.J., 2002. A PETstudy of stimulus- and task-induced
semantic processing. NeuroImage 15, 927–935.
Ogawa, S., Tank, D.W., Menon, R., Ellermann, J.M., Kim, S.G., Merkle,
H., Ugurbil, K., 1992. Intrinsic signal changes accompanying sensory
stimulation: functional brain mapping with magnetic resonance imag-
ing. Proc. Natl. Acad. Sci. U. S. A. 89, 5951–5955.
Opitz, B., Rinne, T., Mecklinger, A., von Cramon, D.Y., Schroger, E., 2001.
Differential contribution of frontal and temporal cortices to auditory
change detection: fMRI and ERP results. NeuroImage 15, 167–174.
Paulesu, E., Frith, C.D., Frackowiak, R.S.J., 1993. The neural correlates of
the verbal component of working memory. Nature 362, 342–345.
Rissman, J., Eliassen, J.C., Blumstein, S.E., 2003. An event-related fMRI
investigation of implicit semantic priming. J. Cogn. Neurosci. 15,
Rossell, S.L., Price, C.J., Nobre, A.C., 2003. The anatomy and time course
of semantic priming investigated by fMRI and ERPs. Neuropsychologia
Rugg, M.D., Doyle, M.C., 1994. Event-related potentials and stimulus
repetition in direct and indirect tests of memory. In: Heinze, H., Munte,
T., Mangun, R. (Eds.), Cognitive Electrophysiology. Birkhauser,
Boston, pp. 124–148.
Rugg, M.D., Nagy, M.E., 1987. Lexical contribution to nonword-
repetition effects: evidence from event-related potentials. Mem. Cogn.
Sekiguchi, T., Koyama, S., Kakigi, R., 2000. The effect of word repetition
on evoked magnetic responses of the human brain. Jpn. Psychol. Res.
Sekiguchi, T., Koyama, S., Kakigi, R., 2001. The effect of stimulus
repetition on cortical magnetic responses evoked by words and
nonwords. NeuroImage 14, 118–128.
Sergent, J., Zuck, E., Levesque, M., MacDonald, B., 1992. Positron
emission tomography study of letter and object processing: empirical
findings and methodological considerations. Cereb. Cortex 2, 68–80.
N400 response in a sentence-reading paradigm using evoked magnetic
fields and magnetic resonance imaging. Brain Res. 762, 29–39.
Swick, D., 1998. Effects of prefrontal lesions on lexical processing and
repetition priming: an ERP study. Cogn. Brain Res. 7, 143–157.
Talairach, J., Tournoux, P., 1988. Co-Planar Stereotactic Atlas of the
Human Brain. Thieme, Stuttgart.
Thiel, C.M., Henson, R.N., Morris, J.S., Friston, K.J., Dolan, R.J., 2001.
Pharmacological modulation of behavioral and neuronal correlates of
repetition priming. J. Neurosci. 21, 6846–6852.
Thompson-Schill, S.L., D’Esposito, M., Aguirre, G.K., Farah, M.J., 1997.
Role of left inferior prefrontal cortex in retrieval of semantic knowl-
edge: a reevaluation. Proc. Natl. Acad. Sci. U. S. A. 94, 14792–14797.
Van Petten, C.A., 1993. A comparison of lexical and sentence-level context
effects in event-related potentials [Special issue: event-related brain
potentials in the study of language Lang]. Cogn. Processes 8, 485–531.
Van Petten, C.A., Kutas, M., 1987. Ambiguous words in context: an event-
related potential analysis of the time course of meaning activation.
J. Mem. Lang. 26, 188–208.
Wagner, A.D., Koutstaal, W., Maril, A., Schacter, D.L., Buckner, R.L.,
2000. Task-specific repetition priming in left inferior prefrontal cortex.
Cereb. Cortex 10, 1176–1184.
Wagner, A.D., Pare-Blagoev, E.J., Clark, J., Poldrack, R.A., 2001.
Recovering meaning: left prefrontal cortex guides controlled semantic
retrieval. Neuron 31, 329–338.
Wei, X., Yoo, S.S., Dickey, C.C., Zou, K.H., Guttmann, C.R., Panych, L.P.,
2004. Functional MRI of auditory verbal working memory: long-term
reproducibility analysis. NeuroImage 21, 1000–1008.
A. Matsumoto et al. / NeuroImage 24 (2005) 624–634