This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
Author's personal copy
Local and global inhibition in bilingual word production: fMRI evidence from
⁎, Hongyan Liu
, Maya Misra
, Judith F. Kroll
State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University, China
Department of Communication Sciences and Disorders, The Pennsylvania State University, USA
Department of Psychology, The Pennsylvania State University, USA
Received 22 October 2010
Revised 4 March 2011
Accepted 17 March 2011
Available online 23 March 2011
The current study examined the neural correlates associated with local and global inhibitory processes used
by bilinguals to resolve interference between competing responses. Two groups of participants completed
both blocked and mixed picture naming tasks while undergoing functional magnetic resonance imaging
(fMRI). One group ﬁrst named a set of pictures in L1, and then named the same pictures in L2. The other group
ﬁrst named pictures in L2, and then in L1. After the blocked naming tasks, both groups performed a mixed
language naming task (i.e., naming pictures in either language according to a cue). The comparison between
the blocked and mixed naming tasks, collapsed across groups, was deﬁned as the local switching effect, while
the comparison between blocked naming in each language was deﬁned as the global switching effect. Distinct
patterns of neural activation were found for local inhibition as compared to global inhibition in bilingual
word production. Speciﬁcally, the results suggest that the dorsal anterior cingulate cortex (ACC) and the
supplementary motor area (SMA) play important roles in local inhibition, while the dorsal left frontal gyrus
and parietal cortex are important for global inhibition.
© 2011 Elsevier Inc. All rights reserved.
Bilinguals are able to speak two or even more languages. A crucial
feature of these bilinguals is that they typically make few mistakes
when switching between their two languages. Thus, an important
question is how bilinguals can select correct words in the correct
language. The mechanism of bilingual language selection has become
the focus of research interest in recent years. Overall, there are at least
two viewpoints on lexical selection in language production. While
there is agreement that words in both languages are activated in
parallel when a bilingual intends to speak (e.g., Costa, 2005; Kroll
et al., 2006), some researchers claim that only candidates in the target
language are considered for selection (e.g., Costa and Caramazza,
1999; Costa et al., 2000, 1999). In contrast, others argue that can-
didates in both languages are activated, but there is an inhibitory
mechanism to suppress the activation of the lexical representation of
the nontarget language (e.g., Green, 1998; Hermans et al., 1998). De
Groot and Christoffels (2006) further proposed that there are two
types of inhibitory control involved in suppressing the unwanted
language, i.e., global control and local control. Speciﬁcally, global
control refers to the activation and/or inhibition of the complete
language system, whereas local control refers to control exerted on a
restricted set of memory representations, such as speciﬁc lexical
One method used to examine whether bilingual lexical selection is
based on inhibition is the language switching paradigm or mixed
language naming task. In this paradigm, participants are instructed
to name digits pictures in either of their languages according to a cue.
For example, Meuter and Allport (1999) asked bilinguals to name
Arabic numbers either in their native language (L1) or in the second
language (L2) according to the background color. A switch trial was
deﬁned as a trial in which the response language differed from that of
the previous trial, while a nonswitch trial referred to a trial in which
the response was in the same language as the previous trial. They
found that naming times were slower on switch than on nonswitch
trials, a difference taken to indicate the cost of switching. Critically,
late bilinguals with a dominant L1 showed an asymmetry in switching
cost, such that larger switch costs were observed for the L1 than the
L2 (Meuter and Allport, 1999). These ﬁndings were interpreted as
reﬂecting inhibitory effects required to overcome the activation of
competitors from the nontarget language, and further suggested
that the more dominant L1 is inhibited to a larger degree than the less
In subsequent language switching studies, there has been con-
troversy regarding the interpretation of the asymmetric switch costs
observed for the bilingual's two languages. Some studies with highly
proﬁcient bilinguals have not observed the asymmetry and have
argued that highly proﬁcient bilinguals do not require an inhibitory
NeuroImage 56 (2011) 2300–2309
⁎Corresponding author at: State Key Laboratory of Cognitive Neuroscience and
Learning, Beijing Normal University, Beijing, 100875, PR China. Fax: +86 10 58806154.
E-mail address: firstname.lastname@example.org (T. Guo).
1053-8119/$ –see front matter © 2011 Elsevier Inc. All rights reserved.
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/ynimg
Author's personal copy
mechanism when producing words in one language alone (Costa
and Santesteban, 2004; Costa et al., 2006). A more recent study also
failed to observe the switch cost asymmetry for even less proﬁcient
bilinguals when the decision to switch between the two languages
was made spontaneously (Gollan and Ferreria, 2009). A careful
analysis of the Costa and Santesteban (2004) data also suggests that
even when the switch cost asymmetry is absent, under the mixed
language naming conditions in the switching paradigm, the L1
becomes slower to name than the L2, suggesting the presence of an
inhibitory process. The mixed ﬁndings from behavioral studies
suggest that the presence or absence of an asymmetrical switching
cost may not provide unequivocal support for the inhibitory control
hypothesis. In addition, the language switching task requires bi-
linguals to frequently switch between languages from trial to trial,
which may only provide evidence for local inhibition of speciﬁc
language representations and which may also have little bearing on
the typical experience of a bilingual who would be unlikely to switch
languages randomly and with such high frequency.
Past studies using the event-related potential (ERP) technique also
shed light on the hypothesis that inhibition of the nontarget language
is necessary to enable language selection during bilingual speech
planning. In these ERP studies (Jackson et al., 2001; Christoffels et al.,
2007; Verhoef et al., 2009), bilinguals performed the language
switching task, while their brain electrical potentials were recorded.
In these studies, switch trials produced signiﬁcant N2 ERP effects,
which are hypothesized to be related to inhibitory control (e.g.,
Falkenstein et al., 1999), and thus support the inhibitory control
model. However, the pattern of results has not been consistent across
studies. For example, Jackson et al. (2001) found a larger N2 ERP
component only for switch trials in L2 relative to nonswitch trials in
L2, but no such effect for L1, while Christoffels et al. (2007) observed a
larger N2 for nonswitch trials in L1 relative switch trials in L1, but
no such effect for L2. This raises the necessity of further studies to
disentangle the discrepancies across these studies. Likewise, these
language switching studies have primarily examined the issue of
whether there is local inhibition in bilingual language production,
operating over an immediate language switch. In a recent ERP study
(Misra et al., under review), Chinese–English bilinguals named a set of
pictures in L1 and then in L2 or in the reverse order, naming a set of
pictures in L2 and then in L1. A greater negativity was produced by
naming in the L1 after naming the same pictures in L2, even though
the pictures were repeated after an entire block. In contrast, the
expected priming, or facilitation effects, for repeated items were
observed when naming in L2 after L1. These results suggest that there
is persistent inhibition of the L1 when naming in the L2, which may
operate at a more global level and may be different from the inhibitory
processes revealed by the language switching paradigm.
Neuroimaging techniques such as functional magnetic resonance
imaging (fMRI) and Positron Emission Topography (PET) have also
recently been used to examine this issue of whether bilinguals require
inhibition in word production to attempt to resolve the current
controversies in the behavioral data (Abutalebi and Green, 2007). In
contrast to behavioral studies, where the presence of inhibition must
be inferred based on ﬁnding asymmetric switch costs or unexpected
slowing of the L1, in neuroimaging studies the activation of neural
areas observed to be required for other tasks requiring inhibition can
be sought during bilingual language production tasks. If these areas
are also activated in bilingual language production, it then provides
evidence for the inhibitory account. Previous neuroimaging studies
have revealed the neural mechanisms underlying cognitive control
by comparing results of switch trials with those of non-switch trials.
Most of these neuroimaging studies have provided evidence
for inhibition even in highly proﬁcient bilinguals, although results
are somewhat mixed.
Price et al. (1999) ﬁrst investigated the neural correlates of
translation and the mechanism required to switch between languages
with six proﬁcient German–English bilinguals using PET. Participants
were asked to translate or read visually presented words in German,
English, or alternating between the two languages. Switch conditions
relative to blocked word naming conditions increased activation in
the left inferior frontal region and bilateral supramarginal gyri.
Switching during translation also increased activity in bilateral ventral
cerebellum and the left medial fusiform.
However, an fMRI study by Hernandez et al. (2000), which com-
pared single and mixed-language picture naming with six early
Spanish–English bilinguals, found increased activation of the left
dorsolateral prefrontal cortex (DLPFC, BA 9/46) but not the supra-
marginal gyrus in language switching conditions relative to single-
language processing. In another study, Hernandez et al. (2001)
further explored whether similar neural correlates are involved in
within and between language switching. In the within-language
condition, six highly proﬁcient and balanced English–Spanish bi-
linguals named pictures as either the actions or the objects depicted
or switched between these two types of responses. In the between-
language switching condition, participants named pictures of objects
in English, in Spanish, or alternated between the two languages. The
right dorsolateral prefrontal cortex was signiﬁcantly more activated
for the mixed-language condition relative to the blocked-language
condition, while the comparison between the within-language mixed
condition and blocked condition yielded no signiﬁcant results even
with a lower threshold. Hernandez et al. claimed that switching
between languages involves increased general executive function
whereas within-language switching may not depend on executive
A criticism of the Hernandez et al. (2001) conclusion is that the
limited number of participants or the covert naming task used in that
study probably led to no increased activation in brain areas associated
with executive control in within-language switching. Using a similar
design, a more recent study (Abutalebi et al., 2008), examined overt
picture naming in a group of 12 German–French bilinguals. Abutalebi
et al. found activation of the left caudate and the anterior cingulate
cortex (ACC) for switching between languages as compared to
switching within a language, while activation in the left prefrontal
cortex was found for both types of task switching. They proposed that
the left prefrontal cortex is engaged in more general executive control,
but that the left caudate and the ACC are more speciﬁc for language
Most of the previous studies have examined the performance of
bilinguals whose two languages share the same script and might
therefore be hypothesized to be more likely to compete because of
their similarity. Wang et al. (2007) further investigated the neural
substrates of language switching among Chinese–English bilinguals
whose two languages use different scripts. During the experiment,
participants were asked to silently name pictures in each of their two
languages according to a visual cue. Several brain areas including
the left medial frontal gyrus and left ACC showed increased activa-
tion when switching into L2, however, no regions related to executive
control showed additional activation when switching into L1.
They claimed that language switching involved both general
executive regions and task-speciﬁc regions, but no regions dedicated
to language switching were observed. This is consistent with the
conclusions of previous studies (e.g., Hernandez et al., 2001) that the
language switching effect is task but not language speciﬁc.
Neural evidence for inhibition during bilingual production also
comes from neuroimaging studies using other tasks. In a study by
Rodriguez-Fornells et al. (2005) addressing the neural inhibition of
phonological interference from the non-target language, German–
Spanish bilinguals and German monolinguals were asked to respond
when the name of a picture started with a consonant but to withhold
responding for names starting with a vowel. The materials were
selected such that for half of the translation equivalents the German
and Spanish names both started with a vowel or consonant, requiring
2301T. Guo et al. / NeuroImage 56 (2011) 2300–2309
Author's personal copy
the same response (congruent), or started differently (incongruent).
Two regions, the DLPFC and the supplementary motor area (SMA),
were shown to be associated with the contrast between incongruent
and congruent trials in bilinguals when compared to monolinguals.
According to Rodriguez-Fornells et al. (2005), these results indicated
that the non-target language phonology was partly activated and
that bilinguals recruited executive control processing mechanisms
to negotiate the interference from the non-target name.
To summarize, a variety of regions, including the left inferior
frontal region (Price et al., 1999; Abutalebi et al., 2008), bilateral
supramarginal gyri (Price et al., 1999), the left dorsolateral prefrontal
cortex (Hernandez et al., 2000; Rodriguez-Fornells et al., 2005), the
right dorsolateral prefrontal cortex (Hernandez et al., 2001), the left
caudate (Abutalebi et al., 2008), the left anterior cingulate cortex
(Wang et al., 2007; Abutalebi et al., 2008), the left medial frontal gyrus
(Abutalebi and Green, 2007; Wang et al., 2007), and the supplemen-
tary motor area (Rodriguez-Fornells et al., 2005) have been observed
to be involved in inhibition of lexical competition between a
bilingual's two languages in order to select the correct language (for
reviews, see Abutalebi and Green, 2007; Rodriguez-Fornells et al.,
The abovementioned studies have highlighted the advantage
of using neuroimaging techniques to evaluate whether cognitive
control is necessary for bilingual language processing. Most studies
have found evidence that one or more areas believed to be involved
in inhibitory control processes are activated during bilingual lexical
selection, despite differences between tasks used for each study,
the level of language proﬁciency of the bilinguals tested, the type
of materials, and the similarity of the bilingual's languages. How-
ever, most of the past studies have typically sought evidence for
inhibition in situations in which a language switch occurred in a
The current study aimed to examine lexical selection and
inhibitory processes in conditions which might be expected to invoke
local versus global inhibitory processes to enable a richer under-
standing of the generality of the proposed inhibitory mechanisms.
By using a design similar to an ERP study we previously conducted
(Misra et al., under review), we attempted to provide neural evidence
for local and global inhibitory processes used by bilinguals to resolve
interference between competing responses. We assume that bi-
linguals may use different levels of executive control, relying on
distinct neural processes, to achieve a given goal in a given situation.
Comparisons were made between local switches (i.e., switching
between languages from trial to trial) and more global switches
(i.e., switching between languages on successive blocks of trials). As
in Wang et al. (2007) we were speciﬁcally interested in Chinese–
English bilinguals whose languages use different scripts. However,
Wang and colleagues used a silent naming task, which might
have underestimated the neural activity related to production (e.g.,
Palmer et al., 2001). Movement artifacts, including head movements,
can be problematic for studies using fMRI to evaluate overt speech
production, but a carefully time-locked event-related (ER) design was
used in the current study to obtain artifact-free images (Huang et al.,
2001). Furthermore, while the Wang et al. study only evaluated
neural correlates of local switching effects, the current study also
aimed to examine the neural mechanism of global inhibition.
In the present experiment, two groups of Chinese–English
bilinguals completed both blocked and mixed picture naming tasks.
One group ﬁrst named a set of pictures in L1, and then named the
same pictures in L2. The other group ﬁrst named pictures in L2,
and then in L1. Following the blocked naming tasks, both groups
performed a mixed language naming task (i.e., naming pictures
in either language according to a cue). The comparison between
the blocked and the mixed naming tasks, collapsed across groups,
was operationalized as the local switching effect, with the switching
effects calculated separately for each language. In addition, the
comparison between blocked naming in each language was oper-
ationalized as the global switching effect.
The logic of our study is as follows:
1) In the mixed naming task, participants cannot select a response
language until they see a cue, so both languages need to be kept
active throughout the task. Therefore, as in previous studies, the
results should reveal the neural mechanism of local inhibition in
the switching effects. Relative to blocked naming, an increased
activation in neural areas associated with cognitive control should
be observed in the mixed naming condition.
2) In contrast, in the blocked naming task, participants are able to
select a response language in advance. However, if both languages
are always active, then even during blocked naming a global
inhibitory process may be required to attenuate activation of
the other language in order to complete naming pictures in the
required language. To boost the probability that the non-target
language label for each picture would be activated, our task
involved repetition of the same pictures from one block to the
other. The results of the blocked switching manipulation thus
should reveal the neural mechanism of the global inhibition.
Speciﬁcally, based on the results of our previous ERP and behavioral
results (Misra et al., under review), we expected that in a typical
block, naming pictures in L2 would activate the dominant L1, while
naming pictures in L1 would not activate L2 to the same degree.
Therefore, naming in L1 after L2 might lead to increased activation
of brain areas related to executive control because of a need to
overcome the L1 inhibition from the previous block. However,
naming pictures in L2 after L1 might be predicted to show a
different picture. Since L1 should be active regardless of the
block conﬁguration, additional inhibition should not be required to
name in L2 after naming in L1.
Materials and methods
Twenty four Chinese–English bilinguals participated in the pres-
ent experiment. All of the participants had normal or corrected-to-
normal vision and were free of neurological diseases. Participants were
paid a small amount of money for their participation. They were
randomly divided into two groups, 12 (6 male and 6 female) for each
group. Group A ﬁrst named a set of pictures in L1, and then named the
same set of pictures in L2. Group B ﬁrst named pictures in L2 and
then in L1. After the blocked naming tasks, both groups performed the
mixed naming task described below. All participants began to learn
English at approximately age 12, and had no experience of studying
Participants in the two groups were closely matched in age,
language proﬁciency level (measured by self assessments provided in
a language history questionnaire and by performance on the Waters
and Caplan (1996) Reading Span task conducted in English), and
executive control (measured by performance of a Simon task). An
independent-samples ttest showed that there was no signiﬁcant
difference between the ages of the participants in each group (mean:
21.4 vs. 22.8 years), t(22) = −1.45, pN0.1. According to the outcome
of a three-way ANOVA on the self-rating scores (mean scores on a 10
point scale —Group A: L1 reading 8.8, writing 8.2, speaking 8.6,
listening 8.8; L2 reading 6.8, writing 6.1, speaking 5.8, listening 6.3;
Group B: L1 reading 7.8, writing 7.7, speaking 7.8, listening 8.1; L2
reading 5.9, writing 5.4, speaking 4.9, listening 5.8), both groups
were more proﬁcient in L1 than L2, F(1, 22) = 76.63, pb0.001; but
there was no signiﬁcant difference in language proﬁciency between
groups, F(1, 22)= 1.87, pN0.1. In the reading span task, performance
on sentence processing was measured along with sentence ﬁnal word
recall. There was no signiﬁcant difference between how many words
2302 T. Guo et al. / NeuroImage 56 (2011) 2300–2309
Author's personal copy
participants in each group recalled correctly (14.5% vs. 14.1%) or the
correct judgment of sentence plausibility (36.8% vs. 33.3%), t(22) b1.
In the Simon task, participants were instructed to judge whether a
colored square was red or blue by pressing a button. The button
for “blue”responses was on the left side of the response pad, while
the button for “red”responses was on the right side. However, squares
could appear on either side of the screen, leading to congruent
trials (where the square of a given color appeared on the same side
as the correct response button) and incongruent trials (where the
square appeared on the side opposite to the appropriate response
button). The difference between the incongruent trials and the
congruent trials is termed the Simon Effect. There was no signiﬁcant
difference between the Simon effect for both groups in reaction
time (RT: 40 ms vs. 43 ms), F(1, 22) b1, or error rates (ER: 2.2%
vs. 3.2%), F(1, 22) = 2.76, pN0.1. These comparisons indicate that
the two groups were well matched on both measures of language
proﬁciency and cognitive control. All of the participants were L1
Seventy line drawings sampled from a wide range of semantic
categories were selected as experimental stimuli. Another 10 pictures
were used as practice items. A plus sign “+”presented on a white
screen was used as the baseline. The ratio of pictures to baselines
was: 5:4. The trial order was randomized, and item condition was
counterbalanced across participants. Pictures were presented on
either a red or blue background (see below).
The experiment was carried out in an isolated room, free from
external noises other than those associated with the fMRI scanning
procedures. There were four runs: the ﬁrst two runs were blocked
naming, and the last two runs were mixed naming. Each run included
the same set of pictures. In the blocked naming task, participants
named pictures in L1 or L2 in separate runs. The order of the language
presented ﬁrst was counterbalanced such that Group A named
pictures ﬁrst in L1, while Group B named pictures ﬁrst in L2. In the
mixed naming runs, participants named successive pictures either in
L1 or in L2 randomly according to the background color of the picture.
The background color–language mapping was explained to partici-
pants verbally, and practice trials were used to verify that they
understood the mapping approach. This mapping was counter-
balanced across participants such that red could indicate either
Chinese or English, depending on the participant. In the blocked
naming condition, the background colors were kept the same as in the
mixed naming condition. Thus, in the blocked naming condition the
background color also alternated, but participants did not need to
attend to the switches in color.
For each run, the same 70 pictures were used as the experimental
stimuli, and another 56 plus signs ‘+’were used as the baseline
stimuli. Each picture or plus sign was presented for 1000 ms and then
replaced by a blank screen for 2000 ms. Participants were instructed
to name pictures as quickly and accurately as possible in a soft voice,
but to remain silent when seeing a plus sign (i.e., baseline).
Participants were instructed to say NO or BUZHIDAO depending on
the language of output if they did not know a name. They were also
instructed to minimize head, jaw and tongue movement when
naming in the scanner. As in our previous studies (e.g., Misra et al.,
under review), participants were not pre-trained with the picture
names to reduce repetition effects. Past studies using pre-training
have assumed that the procedure will simply enhance performance
but not differentially affect L1 and L2. Because it is well known that
the less dominant language is likely to beneﬁt more from repetition
(e.g., Hernandez and Reyes, 2002), and because the goal of the present
study was to examine language selection, we decided that it was
better to tolerate some trials on which bilinguals would not know
the picture's name in L2 rather than to prime L2 in a way that would
create bias. A short break was provided between runs.
Although participants made verbal responses to the pictures in the
scanner, technical limitations did not allow these responses to be
recorded. Therefore, behavioral data were collected again, outside of
the scanner, about 1 month later, in a quiet, isolated room. After the
behavioral picture naming experiment, participants completed the
language history questionnaire, the Simon task, and the Reading
All images were acquired using a 3 T Siemens Sonata whole-body
MRI scanner. Participants' heads were secured to minimize movement.
Functional scans were obtained using a single shot T2*-weighted
gradient echo planar imaging (EPI) sequence. The following scan
parameters were used: TR (repetition time) =3000 ms, TE (echo
time) =30 ms, ﬂip angle = 90°, ﬁeld-of-view (FOV)= 200× 200 mm
matrix size= 64 × 64, and slice thickness/gap =4 mm/0.8 mm.
Thirty three contiguous axial slices were acquired to cover the whole
brain at 128 time points. High-resolution, T1-weighted 3D images
were also obtained. The following scan parameters were used:
TR= 2530 ms, TE= 3.39 ms, ﬂip angle =7°, FOV=256 × 256 mm
matrix size=256 × 256, slice thickness =1.33 mm, and number of
Imaging data analysis
Image processing and statistical analyses were performed using
SPM2 (Wellcome Department of Cognitive Neurology, London, UK) in
conjunction with the ArtRepair software (Mazaika et al., 2007)
implemented in MATLAB (The Mathworks Inc., Natick, Mass., USA).
For each participant, the ﬁrst two volumes in each run were discarded
to allow magnetization to reach the equilibrium state. For the
remaining functional images, slice timing correction was performed
ﬁrst to minimize differences in acquisition time between slices. The
images were realigned to the ﬁrst volume in the time series to correct
for head motion. They were then normalized to the EPI template
in SPM2, based on the Montreal Neurological Institute (MNI) stereo-
tactic space, and then re-sampled with 2× 2 × 2 mm spatial resolu-
tion. The images were spatially smoothed with a cubic Gaussian ﬁlter
(8-mm full width at half-maximum).
To quantify the effect of head movement on the quality of the data,
the data were inspected by using the ArtRepair toolbox for SPM2
(Mazaika et al., 2007), and the realignment parameters provided by
the SPM2 motion correction procedure were examined before the
general linear model (GLM) estimation. Of particular interest is the
scan-to-scan (incremental) motion during the task (e.g., Barch et al.,
1999), i.e., the change in position between two successive images. The
movement parameters for extreme movements were also inspected
by considering absolute movements, i.e., the displacement of a scan
with respect to the ﬁrst image in the time series. The criteria for
inclusion were that a participant did not show absolute motion
greater than the voxel size and incremental motion greater than
1 mm. All participants met the absolute motion inclusion criteria, but
one participant did not meet the incremental motion inclusion
criteria. As there were only 4 volumes with incremental motion
greater than 1 mm, the data of this participant were kept in the
following analyses after being corrected with ArtRepair. Individual
Note that these very low accuracy scores are a consequence of completing this
challenging task in one's L2. While reading span is generally considered a measure of
working memory (see e.g., Waters and Caplan, 1996), in this context it is more
accurately described as an L2 proﬁciency measure.
2303T. Guo et al. / NeuroImage 56 (2011) 2300–2309
Author's personal copy
volumes showing rapid inter-TR movements of greater than 0.5 mm
were excluded via interpolation of the two nearest non-repaired
volumes. Consequently, approximately 1.64% of total volumes for
the 24 participants were corrected. Interpolated volumes were
then partially de-weighted when ﬁrst-level models were calculated
on the repaired images (Mazaika et al., 2007).
Statistical analyses were performed by modeling different condi-
tions (i.e., blocked naming in Chinese, mixed naming in Chinese,
blocked naming in English, mixed naming in English) as explanatory
variables within the context of the general linear model on a voxel-by-
voxel basis. The analyses were performed individually for each
participant. Using group analyses based on a random effects model,
we ﬁrst identiﬁed brain regions that showed a signiﬁcant response to
(a) mixed vs. blocked naming in Chinese and (b) mixed vs. blocked
naming in English for all twenty-four participants.
Only the clusters
larger than 10 voxels (2 × 2 × 2 mm) activated above the height
threshold of pb0.05 (FWE corrected for multiple comparisons) were
considered as signiﬁcant. To detect the interaction effect between
Language (L1/L2) and task (mixed naming/blocked naming), the
mixed naming effect in the two languages was compared, i.e.,
[Chinese (mixed–blocked) vs. English (mixed–blocked)] and [English
(mixed–blocked) vs. Chinese (mixed–blocked)]. To avoid false
alarms from deactivations, the direct comparisons between languages
were inclusively masked by activations in the ﬁrst effect using a
lenient uncorrected pb0.001 threshold. Furthermore, to identify the
global switching effect, we directly compared the activations across
the entire brain of the participants in the different groups that
named pictures in Chinese or English in different orders, using two-
sample t-tests. Due to the reduced statistic power associated with the
between-participant comparison, a more lenient threshold (pb0.001
without further correction for clusters containing at least ten con-
tiguous voxels) was used.
Participants' responses during the behavioral data collection
session were coded as correct using criteria that took into account
the difﬁculty of producing an accurate name for items in both L1
and L2 without pre-training. Therefore, they were given credit for a
correct response if they named the item correctly, if they used
an appropriate category label for the item (e.g., naming a “coat”as
“clothes”), or if they used a correct label with an L2 pronunciation
error. Repetitions were also counted as correct as long as the name
used fell into one of the previous categories. Response accuracy data
for the 24 participants were analyzed with the assumption that
similar accuracies would have been obtained in the original task.
Reaction time data for ﬁve participants were not available due to
equipment malfunction, but the timing data from those participants
for whom it was available were also analyzed. The local and global
switching effects were analyzed separately.
For the local switching effect, the 2 (task: mixed/blocked) 2
(language: L1/L2) two-way ANOVAs on error rates found that the
main effect of taskwas not signiﬁcant, F(1, 23) b1, but the main effect of
language was signiﬁcant, F(1, 23)= 28.99, pb0.001. The interaction
between task and language was also signiﬁcant, F(1, 23) = 11.28,
pb0.005. Further paired ttests showed that mixed naming in L1 elicited
more errors than blocked naming in L1 (5.24% vs. 2.5%), t(23)= 4.05,
pb0.005; however, there was no signiﬁcant difference between blocked
and mixed naming in L2 (14.76% vs. 13.15%), t(23)=1.31, pN0.2. The
two-way ANOVAs on reaction times found that the main effect of
task was signiﬁcant, F(1, 18)= 10.25, pb0.01, but the main effect of
language was not signiﬁcant, F(1, 18)=2.88, pN0.1. The interaction
between the task and language was also signiﬁcant, F(1, 18) = 122.47,
pb0.001. Further paired ttests showed that mixed naming in L1
was signiﬁcantly slower than blocked naming in L1 (1145 ms vs.
929 ms, t(18)=10.15, pb0.001, however, mixed naming in L2 was
faster than blocked naming in L2 (1034 ms vs. 1100 ms, t(18)= −2.12,
pb0.05. This result suggested that an asymmetric local switching
effect was present.
The 2 (task order: ﬁrst/second) 2 (language: L1/L2) two-way
ANOVAs on the global switching effects (Group A: L1 2.98% vs. L2
18.33%; Group B: L1 2.02% vs. L2 11.19%) found a signiﬁcant main
effect of language, F(1, 22)= 46.58, pb0.001, suggesting that naming
pictures in L2 elicited more errors. The main effect of task order
trended towards signiﬁcance, F(1, 22)= 3.33, p= 0.08. The interac-
tion between language and task order was also marginally sig-
niﬁcant, F(1, 22)= 2.97, p= 0.099. The two-way ANOVAs on reaction
times (Group A: L1 908 ms vs. L2 1083 ms; Group B: L1 952 ms vs.
L2 1118 ms) only revealed a signiﬁcant main effect of language,
F(1, 17) = 33.81, pb0.001, indicating that naming pictures in L1 was
faster than naming pictures in L2.
Neuroimaging results: local switching effect
The neuroimaging results for the local switching effect are
summarized in Table 1 and Figs. 1 and 2. Relative to blocked naming
in Chinese (L1), mixed naming in L1 signiﬁcantly activated frontal
cortical areas including bilateral dorsal anterior cingulate gyri, and
supplementary motor area (SMA), as well as the left precentral gyrus.
Signiﬁcant activations were also noted in posterior areas including
bilateral cerebellum and several regions of the occipital lobe including
bilateral superior occipital gyri, middle occipital gyri, inferior occipital
gyri, fusiform gyri, lingual gyri, cuneus, and left precuneus. Additional
areas of enhanced activation were noted in the parietal lobe in the
left superior parietal gyrus and bilaterally in the postcentral gyri.
English (L2) mixed naming resulted in increased activation over L2
blocked naming in similar areas as observed for mixed versus blocked
naming in L1. Speciﬁcally, enhanced activations for mixing naming
were observed in frontal cortical areas including bilateral dorsal
anterior cingulate gyri, and SMA, as well as in left precentral gyrus and
superior medial frontal gyrus. Posterior activations were also again
noted in bilateral cerebellum and bilaterally in occipital lobe regions
including the cuneus, the superior occipital gyri, the middle occipital
gyri, the inferior occipital gyri, the fusiform gyri, and the lingual gyri.
Mixed vs. blocked naming in L2 also resulted in signiﬁcant activation
of left postcentral gyrus.
For the interaction between language and task, the two contrasts,
i.e., [Chinese (mixed–blocked) vs. English (mixed–blocked)] and
[English (mixed–blocked) vs. Chinese (mixed–blocked)], revealed no
signiﬁcant differences in any brain areas, suggesting that no sig-
niﬁcant interaction was found.
Neuroimaging results: global switching effect
The neuroimaging results for the global switching effect are shown
in Table 2 and Fig. 3. Compared with naming pictures ﬁrst in L1,
naming pictures in L1 after naming pictures in L2 led to activation of
the right postcentral gyrus and a series of left hemisphere areas
including the middle frontal gyrus, the middle temporal gyrus, the
precuneus, the inferior parietal gyrus, and the angular gyrus. In
contrast, relative to naming pictures ﬁrst in L2, naming pictures in L2
To identify whether the order in which a given language was named in the blocked
sessions affects the results of local switching effect, we performed two-sample t-tests
between the two groups (i.e., Group A (mixed vs. blocked naming)–Group B (mixed
vs. blocked naming) for each language. The results revealed that no areas showed
signiﬁcant activation to the order effect in either Chinese or English naming, at the
height threshold of pb0.05 (corrected). Thus, the local switching effect was measured
by comparing the blocked naming and mixed naming tasks collapsed across the two
2304 T. Guo et al. / NeuroImage 56 (2011) 2300–2309
Author's personal copy
after naming pictures in L1 merely increased the activity of the right
Bilinguals commonly use only one of their two languages to
communicate with others at a given time. However, previous
behavioral and neuroimaging studies have provided converging
evidence that both languages are activated during bilingual language
production and that bilinguals need to inhibit the interference from
activation of the nontarget language, especially when they speak their
non-dominant language (for a review, see Kroll et al., 2008). Recent
fMRI studies have attempted to identify the neural mechanism for
bilingual inhibitory process. However, these studies have mainly
evaluated inhibitory mechanisms using the language switching
paradigm. While there are circumstances, such as when translating
between languages, in which bilinguals must ﬂuidly switch back and
forth between languages, most interactions do not require this type of
switching. It is not yet clear whether bilinguals allocate different
levels of neural inhibition when their languages are used in distinct
conversational contexts. It is therefore important to identify whether
different types of neural inhibition are necessary for bilinguals to
reach different goals.
The current study was designed to distinguish between the neural
networks involved in local and global inhibition in bilingual word
production. Speciﬁcally, local inhibition was investigated with the
local switching effect (i.e., switching languages in the context of a
mixed naming condition), whereas global inhibition was examined
with the global switching effect (i.e., switching languages across
naming blocks). As summarized and discussed below, distinct net-
works were observed to be engaged in these two contexts.
The neural basis of local inhibition in bilingual word production
In the mixed naming task, a visual cue instructed participants to
name pictures in either L1 or L2. In the blocked naming task,
participants named pictures in one of two languages in the entire
block, although the same pictures were presented across blocks.
Consistent with previous studies (e.g., Meuter and Allport, 1999), the
behavioral results showed a signiﬁcant asymmetric switching
cost with a larger switching cost in the dominant L1, which further
suggests the dominant L1 is inhibited to a larger degree when less
proﬁcient bilinguals produce words in their L2.
More critically, the fMRI results demonstrated that switching into
L1 or L2 engages a largely similar neural network. Speciﬁcally,
switching into either language in the mixed naming condition, as
Brain activation in (1) mixed vs. blocked naming in Chinese and (2) mixed vs. blocked naming in English (pb0.05, kN10, FWE corrected).
Comparison Areas BA MNI coordinates (x, y, z) T value
Mixed vs. blocked naming in Chinese L_SMA 6/32 −4 2 66 9.84
R_SMA 6 2 2 64 7.45
L_Anterior Cingulate Gyrus 24/32 −4 10 42 8.85
R_Anterior Cingulate Gyrus 32 4 16 44 7.08
L_Precentral Gyrus 6 −50 −2 44 8.39
L_Superior Occipital Gyrus 17 −8−98 5 12.43
R_Superior Occipital Gyrus 17/18 22 −99 7 9.90
L_Middle Occipital Gyrus 17/18/19 −16 −100 0 12.96
R_Middle Occipital Gyrus 18/19 32 −90 15 8.55
L_Inferior Occipital Gyrus 19/37 −42 −80 −9 9.11
R_Inferior Occipital Gyrus 18/19 31 −80 −9 8.68
L_Fusiform Gyrus 18/19/37 −24 −76 −12 12.61
R_Fusiform Gyrus 18/19/37 36 −55 −19 10.21
L_Lingual Gyrus 17/18/19 −22 −76 −13 11.69
R_Lingual Gyrus 17/18 10 −82 0 15.63
L_Superior Parietal Gyrus 7 −16 −72 52 7.95
L_Precuneus 7 −6−76 44 8.53
L_Cuneus 19 −8−79 42 7.02
R_Cuneus 17 19 −98 8 8.92
L_Postcentral Gyrus 4/6/43/48 −50 −7 41 7.82
R_Postcentral Gyrus 4 52 −8 32 7.51
L_Cerebellum −6−98 2 14.34
R_Cerebellum 12 −91 2 12.80
Mixed vs. blocked naming in English L_Superior Medial Frontal Gyrus 32 −4 18 42 7.55
L_SMA 6/32 0 12 50 11.93
R_SMA 6/32 2 1 66 9.64
L_Anterior Cingulate Gyrus 24/32 −4 14 44 8.96
R_Anterior Cingulate Gyrus 32 7 18 41 7.50
L_Precentral Gyrus 6 −47 −7 45 7.65
L_Superior Occipital Gyrus 17 −8−100 8 9.79
R_Superior Occipital Gyrus 17/18 23 −96 10 7.65
L_Middle Occipital Gyrus 17/18 −9−95 1 8.28
R_Middle Occipital Gyrus 17/18 24 −96 10 7.80
L_Inferior Occipital Gyrus 18/19 −25 −81 −11 9.50
R_Inferior Occipital Gyrus 19 32 −80 −15 8.28
L_Fusiform Gyrus 18/19/37 −30 −72 −16 11.26
R_Fusiform Gyrus 18/19/37 37 −64 −15 9.94
L_Lingual Gyrus 17/18/19 −26 −80 −12 9.30
R_Lingual Gyrus 17/18 13 −77 −13 7.70
L_Cuneus 17 −6−99 13 8.38
R_Cuneus 17/18 12 −100 8 8.33
L_Postcentral Gyrus 3/4/6/43/48 −48 −8 44 8.54
L_Cerebellum −32 −70 −20 10.03
R_Cerebellum 26 −62 −26 12.37
2305T. Guo et al. / NeuroImage 56 (2011) 2300–2309
Author's personal copy
opposed to blocked naming, was associated with increased activation
of the bilateral dorsal anterior cingulate (ACC) gyri, the bilateral
supplementary motor area (SMA), left precentral gyrus, left post-
central gyri, bilateral cerebellum, bilateral superior occipital gyri,
bilateral middle occipital gyri, bilateral inferior occipital gyri, bilateral
fusiform gyri, bilateral cuneus, and bilateral lingual gyri. Switching
into naming in L1 was also associated with increased activation of the
left superior parietal gyrus, left precuneus, and right postcentral
gyrus, while switching into naming in L2 engaged increased activation
of the left medial superior frontal gyrus. However, direct contrasts
between the two directions of switching revealed no signiﬁcant
differences in their neural substrates.
Increased activity of posterior regions such as occipital gyri, and
cuneus may be due to the fact that participants needed to pay more
visual attention to the background colors in the mixed picture naming
condition, as the background color was a visual cue for the language
of naming. However, in the blocked naming condition, although
pictures were also presented against different background colors, the
cues were redundant with the language of the blocked task so that
participants did not need to consciously process the background
colors. In addition, increased activation in lingual gyri, and fusiform
gyri may be due to the fact that lexical candidates in both languages
are activated for selection in the mixed naming condition, while
enhanced activation during the mixing condition in the bilateral
cerebellum (e.g., Ackermann et al., 1998) and postcentral gyri (Hillis
et al., 2004) may have been related to articulatory processing. These
results suggest that the switch trials may have required more effort to
generate the picture names in the target language. These are not
the main foci of the current study, so we will not discuss these results
Importantly, there was a signiﬁcant difference between mixed and
blocked naming in several regions related to attentional control in the
frontal and parietal lobes. Speciﬁcally, the left medial superior frontal
gyrus, bilateral SMAs, left precentral gyrus, the dorsal anterior
cingulate gyri, the left precuneus, and the left superior parietal
gyrus were involved in mixed naming relative to blocked naming.
Activation of the SMA and precentral gyrus has been reported to be
related to motor control and is reliably found in the go/nogo task
(Talati and Hirsch, 2005). One possible reason for enhanced activation
in these areas for switch trials in the current study, regardless of
language, is that words in two languages are activated to a larger
extent in the mixed naming task compared with the blocked naming
task, and that bilinguals need to select the correct words in the target
language by inhibiting the non-target verbal response in the mixed
naming task. However, it is unclear why switching into L2 (English)
engaged more anterior areas of the attentional network such as the
left superior medial frontal gyrus, while switching into L1 (Chinese)
involved in more activation in the posterior areas of the attentional
network such as the left precuneus and left superior parietal gyrus.
Fig. 1. Group-averaged t-maps for mixed vs. blocked naming in Chinese (L1) and in English (L2).
Fig. 2. Activation in SMA and ACC for mixed vs. blocked naming in Chinese (L1) and
activation in SMA, ACC and superior MFC for mixed vs. blocked naming in English (L2).
Brain activation for direct comparison of two groups: (1) Chinese (L1) second vs.
Chinese ﬁrst and (2) English (L2) second vs. English ﬁrst (pb0.001, kN10, uncorrected).
Comparison Areas BA MNI coordinates
(x, y, z)
vs. Chinese ﬁrst
−36 46 8 5.53
37 −58 −58 14 3.66
L_Precuneus 7/19 −12 −70 36 3.67
R_Postcentral Gyrus 3/4 38 −34 66 4.27
7/39/40 −42 −60 52 5.41
L_Angular Gyrus 39/40 −42 −54 46 4.28
vs. English ﬁrst
R_Cuneus/Precuneus 23 18 −66 24 4.45
2306 T. Guo et al. / NeuroImage 56 (2011) 2300–2309
Author's personal copy
The precuneus, in the posterior region of the medial parietal cortex,
may play an important role in higher cognitive functions such as
episodic memory retrieval (for a review, see Cavanna and Trimble,
2006). Activation of the left superior parietal gyrus was previously
found when participants switched between two manual responses
(Rushworth et al., 2001) and when participants switched between
verbal ﬂuency tasks (Gurd et al., 2002). These results suggest that
language proﬁciency (L2 vs. L1) or language similarity (English vs.
Chinese) might modulate engagement of the frontal and parietal
attention networks in bilingual language production although direct
contrasts between the two directions of switching effects in the
present study revealed no signiﬁcant differences. Further studies may
determine which factors play a role.
Enhanced activation in the dorsal ACC was also observed for
switching versus blocked naming in both languages. There is some
controversy about whether the ACC is related to conﬂict detection or
error detection, or whether these two functions might be processed
separately in distinct subdivisions in ACC (for reviews, see Bush et al.,
2000; Rushworth et al., 2004). Typically, activation of a more dorsal
level in the ACC has been found in cognitive tasks that involve
response selection or conﬂict monitoring. For instance, in a study
using the counting Stroop paradigm, participants were asked to report
the number of words on the screen while ignoring word meaning
(Bush et al., 1998). Compared with neutral trials that contained
common animal words, the interference trials consisting of number
words that were incongruent with the correct response (e.g., “two”
written three times) increased the activation of the dorsal ACC. In the
current experiment, the mixed naming task may involve more
competition for response selection, leading to increased activation
in bilateral dorsal ACC in mixed compared with blocked naming.
The abovementioned areas such as the dorsal ACC and the SMA
have also been reported to be active in previous studies on lexical
selection in bilingual language production (Abutalebi et al., 2008;
Rodriguez-Fornells et al., 2005; Wang et al., 2007), and further
suggest that cognitive control is necessary for bilinguals to select
the correct language in contexts in which they must make a decision
between languages frequently. However, we did not observe acti-
vation of the left caudate and the supramarginal gyrus, which have
been found in some other studies of language switching (e.g.,
Abutalebi et al., 2008). One possibility is that different types of
switching effects were compared in different studies. Most previous
studies (e.g., Price et al., 1999) have only compared neural activity
for switching trials and nonswitching trials collapsed across
languages, rather than separating the two languages for analysis.
Wang et al. (2007) evaluated the two directions of switching effects
separately, and only found activation of the right supramarginal
gyrus in switching for L1, but no activation of the left caudate.
Another possible reason why we did not ﬁnd activation in the left
caudate in our study is that the left caudate may be more likely to be
activated in bilinguals who speak two similar languages. In other
words, it may be easier to switch between two languages of dif-
ferent scripts such as Chinese and English, as opposed to languages
such as Spanish and English, resulting in no signiﬁcant activation of
these brain areas. In addition, most previous studies used a covert
naming task, rather than overt naming, and presented a visual cue
prior to or after the picture, rather than simultaneously with picture
onset. These task differences may have also impacted the observed
results. Further studies will be required to investigate these issues
The neural basis of global inhibition in bilingual word production
In the blocked naming task, participants named pictures in one of
their two languages in the same block, but the order of the language
of naming differed. One group of participants ﬁrst named pictures in
L1 and then in L2, while the other group ﬁrst named pictures in L2
and then in L1. Behaviorally, there was no signiﬁcant interaction
between task order and language, but our ﬁndings revealed different
patterns of brain activation when switching into L1 after a block of
naming in L2 as compared to switching into L2 after an L1 naming
Compared with naming pictures in L1 ﬁrst, naming pictures in
L1 after naming pictures in L2 resulted in activation of a network of
brain regions including left middle frontal gyrus, the left temporal
gyrus, the left precuneus, the left inferior parietal gyrus, the left
angular gyrus, and the right postcentral gyrus. Many prior studies
have provided evidence for the role of these brain regions in cognitive
The DLPFC, which encompasses the middle frontal gyrus, has been
associated with interference suppression in tasks such as the ﬂanker
task (Bunge et al., 2002), and is recruited during maintenance of task-
relevant information (for review, see Miller and Cohen, 2001). The
activation of this area when naming in L1 after a block of naming in L2
suggests its role in overcoming inhibition of L1 that is required in
the previous L2 naming block to maintain a goal through the entire
experimental session (e.g., naming pictures in L1 only).
Activation of the temporal and parietal cortices is often associated
with visual attention. For example, in a task where participants were
required to switch attention between local or global features of
hierarchically organized letters (e.g., a large letter h consisting of
small letters s), the temporal–parietal areas were activated (Fink et al.,
1997). Left inferior parietal gyrus has also been observed to be
activated in tasks requiring interference suppression, such as the
ﬂanker task (Bunge et al., 2002), and in cases where visual attentional
set shifts are required, such as in the attention-switching paradigm
(Rushworth et al., 2001). In the current blocked naming tasks,
Fig. 3. Direct comparison between groups: activation network for Chinese (L1) second vs. Chinese ﬁrst and right cuneus/precuneus activation for English (L2) second vs. English ﬁrst.
2307T. Guo et al. / NeuroImage 56 (2011) 2300–2309
Author's personal copy
switching languages between blocks was endogenously driven. The
activation of left middle temporal gyrus, left inferior parietal gyrus,
and left precuneus might therefore be related to selecting the
appropriate language to name the pictures.
Increased activation in the right postcentral gyrus may be due
to recruitment of additional resources for articulatory processes
when naming in L1 after L2. However, the role of the left angular
gyrus during this task is less clear. Previously, this region has been
hypothesized to be related to mapping visual inputs onto phonolog-
ical representations (Horwitz et al., 1998), but its role in language
production has not been well addressed. The present ﬁnding that this
area was activated in response to the block order effect in L1 might
simply suggest that naming pictures in L1 after naming pictures in L2
recruits more processing resources, but further studies are needed to
clarify the role of this region in language production.
The above results from the global switching effect for blocked
naming in L1 suggest that when bilinguals speak the L2 over an
extended block, there is inhibition of the L1 that is persistent to the
next block, which lead to activation of the neural correlated associated
with cognitive control. In contrast, naming a block of pictures in L2
after naming the same pictures in L1 enhanced only the activation
of the right cuneus/precuneus relative to naming the same set of
pictures ﬁrst in L2. This result may be due to participants' awareness
of picture repetitions, as this area is related to visual processing.
The dorsal ACC was not differentially activated for either
comparison in our blocked naming analysis. This result is consistent
with previous ﬁndings that the ACC is activated only when par-
ticipants need to make a decision between changing response sets or
in situations with response conﬂicts (for a review, see Rushworth
et al., 2004). In contrast, during the mixed naming task, in which
participants needed to frequently select which language to produce,
dorsal ACC activations were observed.
The general pattern of differential inhibition for the L1 relative to
the L2 converges closely with the results obtained by Misra et al.
(under review) when examining similar conditions using ERPs. That
is, during blocked language naming, participants need to inhibit the
other language in order to maintain the same goal through the
experiment. For naming in the L2, the presence of a previous block of
trials in L1 may have posed only minimal additional interference over
the general situation of naming in one's non-dominant language.
However, if completing a block of trials in which the L2 name had
been retrieved required inhibition of the L1, this may have lead to
apparent interference when retrieving the L1 label in a subsequent
In summary, the present study attempted to investigate whether
unbalanced bilinguals exhibited a neural dissociation in cognitive
control in situations requiring rapid, frequent language switches, as
opposed to situations which require sustained processing in one
language versus the other. To our knowledge, the present ﬁndings
are the ﬁrst to demonstrate the neural correlates of a functional
dissociation between two levels of inhibition in bilingual speakers.
We found distinct patterns of neural activation for local inhibition
as compared to global inhibition in bilingual word production.
Speciﬁcally, our results suggest that the dorsal ACC and the SMA
play important roles in local inhibition, while the dorsal left frontal
gyrus and parietal cortex are important for global inhibition.
However, it is still unknown whether these effects were speciﬁc for
bilinguals or induced by the task requirements. In future studies it will
be important to recruit monolinguals as a control group to exclude
the latter possibility.
This work was supported by the National Natural Science
Foundation of China (Grant Number 30600179) and the Fundamental
Research Funds for the Central Universities to Taomei Guo. The
writing of this article was supported in part by NIH Grant R01-
HD053146 to Judith F. Kroll, Taomei Guo, and Maya Misra, and by
NSF Grant OISE-0968369 from the Partnerships for International
Research and Education (PIRE) Program to Judith F. Kroll. We thank
Wenping You, Jingjing Guo, Min Chen, and Xiujun Li for their help
with collecting data. We are also grateful to the two anonymous
reviewers for their helpful comments to the previous versions of the
Abutalebi, J., Green, D., 2007. Bilingual language production: the neurocognition
of language representation and control. J. Neurolinguistics 20, 242–275.
Abutalebi, J., Annoni, J.M., Zimine, I., Pegna, A.J., Seghier, M.L., Lee-Jahnke, H., Lazeyras,
F., Cappa, S.F., Khateb, A., 2008. Language control and lexical competition in
bilinguals: an event-related fMRI study. Cereb. Cortex 18, 1496–1505.
Ackermann, H., Wildgruber, D., Daumc, I., Grodd, W., 1998. Does the cerebellum
contribute to cognitive aspects of speech production? A functional magnetic
resonance imaging (fMRI) study in humans. Neurosci. Lett. 247, 187–190.
Barch, D.M., Sabb, F.W., Carter, C.S., Braver, T.S., Noll, D.C., Cohen, J.D., 1999. Overt verbal
responding during fMRI scanning: empirical investigations of problems and
potential solutions. Neuroimage 10, 642–657.
Bunge, S.A., Dudukovic, N.M., Thomason, M.E., Vaidya, C.J., Gabrieli, J.D.E., 2002.
Immature frontal lobe contributions to cognitive control in children: evidence from
fMRI. Neuron 33, 301–311.
Bush, G., Whalen, P.J., Rosen, B.R., Jenike, M.A., McInerney, S.C., Rauch, S.L., 1998. The
counting Stroop: an interference task specialized for functional neuroimaging-
validation study with functional MRI. Hum. Brain Mapp. 6, 270–282.
Bush, G., Luu, P., Posner, M.I., 2000. Cognitive and emotional inﬂuences in anterior
cingulate cortex. Trends Cogn. Sci. 4, 215–222.
Cavanna, A.E., Trimble, M.R., 2006. The precuneus: a review of its functional anatomy
and behavioural correlates. Brain 129, 564–583.
Christoffels, I.K., Firk, C., Schiller, N.O., 2007. Bilingual language control: an event-
related brain potential study. Brain Res. 1147, 192–208.
Costa, A., Caramazza, A., 1999. Is lexical selection in bilingual speech production
language speciﬁc? Further evidence from Spanish–English and English–Spanish
bilinguals. Bilingualism Lang. Cogn. 2, 231–244.
Costa, A., Miozzo, M., Caramazza, A., 1999. Lexical selection in bilinguals: do words in
the bilingual's two lexicons compete for selection? J. Mem. Lang. 41, 365–397.
Costa, A., Caramazza, A., Sebastián-Gallés, N., 2000. The cognate facilitation effect:
implications for the models of lexical access. J. Exp. Psychol. Learn. Mem. Cogn. 26,
Costa, A., Santesteban, M., 2004. Lexical access in bilingual speech production: Evidence
from language switching in highly proﬁcient bilinguals and L2 learners. J. Mem.
Lang. 50, 491–511.
Costa, A., 2005. Lexical access in bilingual production. In: Kroll, J.F., De Groot, A.M.B.
(Eds.), Handbook of Bilingualism: Psycholinguistic Approaches. Oxford University
Press, NY, pp. 308–325.
Costa, A., Santesteban, M., Ivanova, I., 2006. How do highly proﬁcient bilinguals control
their lexicalization process? Inhibitory and language-speciﬁc selection mecha-
nisms are both functional. J. Exp. Psychol. Learn. Mem. Cogn. 32, 1057–1074.
De Groot, A.M.B., Christoffels, I.K., 2006. Language control in bilinguals: monolingual
tasks and simultaneous interpreting. Bilingualism Lang. Cogn. 9, 189–201.
Falkenstein, M., Hoormann, J., Hohnsbein, J., 1999. ERP components in Go/Nogo tasks
and their relation to inhibition. Acta Psychol. 101, 267–291.
Fink, G.R., Halligan, P.W., Marshall, J.C., Frith, C.D., Frackowiak, R.S.J., Dolan, R.J., 1997.
Neural mechanisms involved in the processing of global and local aspects of
hierarchically organized visual stimuli. Brain 120, 1779–1791.
Gollan, T.H., Ferreria, V.S., 2009. Should I stay or should I switch? A cost–beneﬁt analysis
of voluntary language switching in young and aging bilinguals. J. Exp. Psychol.
Learn. Mem. Cogn. 35, 640–665.
Green, D.W., 1998. Mental control of the bilingual lexico-semantic system. Bilingualism
Lang. Cogn. 1, 67–81.
Gurd, J.M., Amunts, K., Weiss, P.H., Zaﬁris, O., Marshall, J.C., Fink, G.R., 2002. Posterior
parietal cortex is implicated in continuous switching between verbal ﬂuency tasks:
an fMRI study with clinical implications. Brain 125, 1024–1038.
Hernandez, A.E., Martinez, A., Kohnert, K., 2000. In search of the language switch: an
fMRI study of picture naming in Spanish–English bilinguals.Brain Lang. 73, 421–431.
Hernandez, A.E., Dapretto, M., Mazziotta, J., Bookheimer, S., 2001. Language switching
and language representation in Sp anish–English bilinguals: an fMR I study.
Neuroimage 14, 510–520.
Hernandez, A.E., Reyes, I., 2002. Within- and betwe en-language priming differ:
evidence from repetition of pictures in Spanish–English bilinguals. J. Exp. Psychol.
Learn. Mem. Cogn. 28, 726–734.
Hermans, D., Bongaerts, T., de Bot, K., Schreuder, R., 1998. Producing words in a foreign
language: can speakers prevent interference from their ﬁrst language? Bilingualism
Lang. Cogn. 1, 213–229.
Hillis, A.E., Work, M., Barker, P.B., Jacobs, M.A., Breese, E.L., Maurer, K., 2004. Re-
examining the brain regions crucial for orchestrating speech articulation. Brain 127,
Horwitz, B., Rumsey, J.M., Donohue, B.C., 1998. Functional connectivity of the angular
gyrus in normal reading and dyslexia. Proc. Natl. Acad. Sci. 95, 8939–8944.
2308 T. Guo et al. / NeuroImage 56 (2011) 2300–2309
Author's personal copy
Huang, J., Carr, T.H., Cao, Y., 2001. Comparing cortical activations for silent and overt
speech using event-related fMRI. Hum. Brain Mapp. 15, 39–53.
Jackson, G.M., Swainson, R., Cunnington, R., Jackson, S.R., 2001. ERP correlates of
executive control during repeated language switching. Bilingualism Lang. Cogn. 4,
Kroll, J.F., Bobb, S.C., Misra, M., Guo, T.M., 2008. Language selection in bilingual speech:
evidence for inhibitory processes. Acta Psychol. 128, 416–430.
Kroll, J.F., Bobb, S., Wodniecka, Z., 2006. Language selectivity is the exception, not the
rule: arguments against a ﬁxed locus of language selection in bilingual speech.
Bilingualism Lang. Cogn 9, 119–135.
Mazaika, P., Whitﬁeld-Gabrieli, S., Reiss, A., Glover, G., 2007. Artifact repair for fMRI data
from high motion clinical participants. Poster Presented at 13th Annual Meeting of
the Organization for Human Brain Mapping, Chicago, IL.
Meuter, R.F.I., Allport, A., 1999. Bilingual language switching in naming: asymmetrical
costs of language selection. J. Mem. Lang. 40, 25–40.
Miller, E.K., Cohen, J.D., 2001. An integrative theory of prefrontal cortex function. Annu.
Rev. Neurosci. 24, 167–202.
Misra, M., Guo, T., Bobb, S.C., Kroll, J.F., under review. Electrophysiological evidence for
global inhibition in bilingual word production.
Palmer, E.D., Rosen, H.J., Ojemann, J.G., Buckner, R.L., Kelley, W.M., Petersen, S.E., 2001.
An event-related fMRI study of overt and covert word stem completion. Neuro-
Image 14, 182–193.
Price, C.J., Green, D., von Studnitz, R.A., 1999. Functional imaging study of translation
and language switching. Brain 122, 2221–2236.
Rodriguez-Fornells, A., Van der Lugt, A., Rotte, M., Britti, B., Heinze, H.J., Munte, T.F.,
2005. Second language interferes with word production in ﬂuent bilinguals: brain
potential and functional imaging evidence. J. Cogn. Neurosci. 17, 422–433.
Rodriguez-Fornells, A., De Diego Balaguer, R., Munte, T.F., 2006. Executive control in
bilingual language processing. Lang. Learn. 56, 133–190.
Rushworth, M.F.S., Paus, T., Sipila, P.K., 2001. Attention systems and the organization
of the human parietal cortex. J. Neurosci. 21, 5262–5271.
Rushworth, M.F.S., Walton, M.E., Kennerley, S.W., Bannerman, D.M., 2004. Action sets
and decisions in the medial frontal cortex. Trends Cogn. Sci. 8, 410–417.
Talati, A., Hirsch, J., 2005. Functional specialization within the medial frontal gyrus
for perceptual go/no-go decisions based on “what”,“when”, and “where”related
information: an fMRI study. J. Cogn. Neurosci. 17, 981–993.
Verhoef, K., Roelofs, A., Chwilla, D.J., 2009. Role of inhibition in language switching:
evidence from event-related brain potentials in overt picture naming. Cognition
Wang, Y., Xue, G., Chen, C., Xue, F., Dong, Q., 2007. Neural bases of asymmetric language
switching in second-languag e learners: an ER- fMRI study. Neuroimage 35,
Waters, G.S., Caplan, D., 1996. The measurement of verbal working memory capacity
and its relation to reading comprehension. Q. J. Exp. Psychol. 49A, 51–79.
2309T. Guo et al. / NeuroImage 56 (2011) 2300–2309