Control of prepotent responses by the superior medial frontal cortex
Chiao-Yun Chena,b,c, Neil G. Muggletond, Ovid J.L. Tzenga,b,e,f, Daisy L. Hunga,b,e, Chi-Hung Juana,b,⁎
aInstitute of Cognitive Neuroscience, National Central University, Jhongli 320, Taiwan
bLaboratories for Cognitive Neuroscience, National Yang-Ming University, Taipei 112, Taiwan
cSocial Science Research Center, National Science Council, Taipei 115, Taiwan
dInstitute of Cognitive Neuroscience and Department of Psychology, University College London, London WC1N 3AR, UK
eInstitute of Neuroscience, National Yang-Ming University, Taipei 112, Taiwan
fInstitute of Linguistics, Academia Sinica, Taipei 115, Taiwan
a b s t r a c t a r t i c l ei n f o
Received 15 April 2008
Revised 10 September 2008
Accepted 11 September 2008
Available online 24 September 2008
The inhibitory control of prepotent action is vital for appropriate behaviour. An example of the importance of
such control can be seen in the inhibition of aggressive behavior, deficits in which may have broader
consequences for society. Many studies have related lesions or the under-development of the prefrontal
cortex to inefficiency of inhibitory control. Here we used transcranial magnetic stimulation and a stop-signal
task, which occasionally requires the inhibition of a prepotent motor response, to investigate the role of pre-
supplementary motor area (Pre-SMA) in inhibitory control. While no effects were seen on the ability to
generate responses, TMS delivered over the Pre-SMA disrupted the ability to respond to a stop signal. These
results are the first to establish a casual link between Pre-SMA and inhibitory control in normal subjects. The
understanding of the underlying mechanisms of inhibitory control may lead to clearer understanding of the
neural basis of inappropriate behaviour.
© 2008 Elsevier Inc. All rights reserved.
The control of voluntary action involves not only choosing from a
range of possible actions but also the inhibition of responses as
circumstances demand. The ability to inhibit prepotent responses is
important to prevent execution of a behaviour in circumstances
where to do so may be detrimental. Oft mentioned examples of such
behaviour include the withholdingof responses by batsmen in sports
such as cricket and baseball, when there is a very brief period in
whichthechoicemust bemadewhethertomakeamotor responseor
This behaviour can be investigated experimentally using stop-
signal tasks. These involve the presentation of a target to which
subjects have to respond unless a (relatively rare) stop signal is
presented. Performance of such tasks is usually described in terms of a
race between the go response and the stop response, with whichever
reaches its threshold first governing the response made (Logan and
Cowan,1984; Logan et al.,1984; Boucher et al., 2007). Alteration in the
time of onset of the stop signal allows for variation in the probability
of responding in the trials with stop signals. The distribution of go
reaction times and the probability of responding in the trials with stop
signals can then be used to estimate the time required to inhibit the
planned response, namely the stop-signal reaction time (SSRT).
The stop-signal task can be used to reliably estimate the response
time of an internally generated act of control. This task therefore has
been applied to investigate the underlying causes of some clinical
syndromes related to impulsivity control. Deficits in performance on
this task have been seen in attention-deficit hyperactivity disorder
and conduct disorder (Schachar and Logan,1990; Schachar et al.,1993,
1995; Armstrong and Munoz, 2003), Tourette's syndrome (Li et al.,
2006a) and cocaine-dependent men (Li et al., 2008). This task has also
been used to gauge the impulsivity of certain groups of subjects. For
instance: Logan et al. (1997) found that highly impulsive subjects
showed longer stop signal reaction times on this task. A similar line of
study has shown that impulsive-violent offenders have longer SSRTs,
requiring a longer time to inhibit their actions, compared to matched
controls(Chenetal.,2008).Thesestudies have demonstratedthewide
utility of the task and its robust effects.
Recently, many studies have been focused on the neural correlates
of the inhibitory processes recruited in the task. Evidence from both
electrophysiological studies in non-human primates and human
neuroimaging data have indicated that several frontal cortical regions
are involved in the cognitive processes required for successful
performance of the stop-signal task. These include frontal eye fields
(FEF, Hanes and Schall,1996), supplementary eye fields (SEF, Stuphorn
et al., 2000; Stuphorn and Schall, 2006), anterior cingulated cortex
(ACC, Ito et al., 2003; Chevrier et al., 2007), inferior frontal gyrus (IFG,
area (Pre-SMA, Li et al., 2006b). Several comprehensive papers have
further illustrated the neural mechanisms of the inhibitory processes
NeuroImage 44 (2009) 537–545
⁎ Corresponding author. Fax: +886 3 426 3502.
E-mail address: firstname.lastname@example.org (C.-H. Juan).
1053-8119/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/ynimg
involved in theparadigm in regards of both awake-behavingmonkeys'
data (e.g. Schall et al., 2002; Schall and Boucher, 2007; Isoda and
Hikosaka, 2007) and evidence from human neuroimaging studies (e.g.
Aron et al., 2007a). With neurodisruption techniques, it is now
established that both SEF (Stuphorn and Schall, 2006) and IFG
(Chambers et al., 2006, 2007) are critical for inhibitory control.
Stuphorn and Schall (2006) showed that such SEF stimulation
increases the probability of countermanding a saccade, but there was
no effect when no stop signal occurred. Chambers et al. (2006, 2007)
first demonstrated that IFG is critical when manual responses are
necessary for inhibition and the SMA only critical for motor initiation.
In comparison with the findings of IFG and SEF involvement in
response inhibition, whetherthe Pre-SMA has a critical functional role
in this process is relatively less understood. The role of the Pre-SMA in
action control has been investigated in several studies. For example,
the Pre-SMA is involved when subjects must switch between making
manual or saccade responses or when inhibiting such responses
(Rushworth et al., 2007; Taylor et al., 2007). Imaging studies have
implicated this area in the initiation of voluntary responses (Lau et al.,
2004a,b). It is also involved in switching between tasks (Rushworth
et al., 2002a,b) and between action sequences (Kennerleyet al., 2004).
have been implicated in self initiated or goal driven actions and it has
automatically triggered actions (Sumner et al., 2007). In recent years,
several studies have also investigated the functional role of the Pre-
SMA in response inhibition. Li et al. (2006b) used the duration of the
SSRT in a stop-signal task to index the efficiency of inhibitory control
and to investigate the functional role of the Pre-SMA in this control by
use of fMRI. They compared groups of subjects with short and long
SSRTs and found that higher activation in the Pre-SMA was associated
with shorter SSRTs (i.e. more efficient performance). This was argued
to be consistent with Pre-SMA having a role in mediating the motor
2006; Aron et al., 2007b). In addition, human patient studies revealed
that damage to right superior frontal regions (including Pre-SMA and
SMA) elevated subjects' SSRT in the stop-signal paradigm (Floden and
al. (2007) also observed that patients with damage to the left superior
regions of Brodmann area 6 (i.e. in the vicinity of the left Pre-SMA)
in a go–nogo task. The pattern of the results suggests that damage in
the vicinity of the left Pre-SMA may impair inhibitory control in
patients. However, it is not yet clear whether the Pre-SMA plays a
critical role in inhibitory control in healthy human subjects. This is
especially important since the findings of humanpatients studies may
also involve an unknown degree of reorganization in the damaged
brain which may occur rapidly. In particular, the Pre-SMA has direct
connections to both the rIFG and the subthalamic nucleus, which are
considered to be the main neural substrates for inhibitory control
(Aron etal.,2007b). Thefindings from thehumanpatientsstudies may
the inhibitory network after the Pre-SMA lesion but not the actual
with several cognitive functions related to the control of action, its
critical role in response inhibition has not been tested directly with
shed some lights on the mechanisms underlying the operation of the
inhibitory neural network.
While neuroimaging studies provide evidence of areas in the brain
in which activity correlates with task performance, this approach
alone cannot establish a causal link showing that an area is essential
for task performance (Li et al., 2006b) and the human patients studies
cannot avoid the potential confounds of neural plasticity or reorga-
nization which may occur following the lesion. We therefore used
temporally precise transcranial magnetic stimulation (TMS), which
allows investigationof whetheran area is requiredfor performance, to
evaluate the contribution of Pre-SMA to stop signal performance.
Rushworth and colleagues (Rushworth et al., 2002a; Kennerley et al.,
2004) have demonstrated that high frequency repetitive TMS (rTMS,
subjects' performance on the switch of motor sequences. In this study,
we employed 10 Hz rTMS over the Pre-SMA but the duration of rTMS
was 100 ms (i.e.: 2 pulses). This rTMS protocol has beenwidely used to
investigate the functional roles of the primary visual cortex (V1), the
frontal eye fields (FEF) and the posterior parietal cortex (PPC) in visual
search tasks. Juan and Walsh (2003) used this rTMS protocol over V1
and found that rTMS impaired visual search function by decreasing
subjects' d′ score in two discrete time windows (for review see Juan
et al., 2004; Chambers and Mattingley, 2005). A similar approach has
been used to investigate the critical time windows of FEF involvement
in visual search (O'Shea et al., 2004), to elucidate the different
temporal involvements between FEFand PPC (Kalla et al., 2008) and to
probe temporal dissociation between visual selection and saccade
preparation in FEF (Juan et al., 2008). Both O'Shea et al. (2004) and
Kalla et al. (2008) found that rTMS FEF and rTMS PPC decrease
subjects' d' score and Juan et al. (2008) found that subjects' saccade
latencies were prolonged by rTMS FEF in two separated time windows
which represented the stages of visual selection and saccade
preparation, respectively. Because Rushworth et al demonstrated
that Pre-SMA rTMS interfered the function motor switching and the
abovementioned findings of neurodisruptive effects of the rTMS
protocol, we therefore hypothesized that the rTMS delivered over Pre-
SMA would disrupt stop signal performance and such disruption
would be manifested as modulation of the ability to respond to the
stop signal (i.e. inhibitory control would be affected) rather than
affecting action execution. The present study represents the first
attempt to use rTMS to test the functional necessity of Pre-SMA in the
inhibition of the prepotent responses with the stop-signal paradigm.
Aron et al. (2007a,b) suggested that the Pre-SMA, IFG and STN are
parts of a neural network for the inhibitory control in stop-signal
paradigm. Recently, Chambers and colleagues (2006, 2007) have used
TMS over IFG and found the rate of noncancelled responses following a
stop signal was increased by TMS (i.e.: inhibitory control was affected).
Their findings have successfully established a causal role for IFG in the
inhibitory control. The TMS protocol used in their studies was so called
1 Hz repetitive TMS stimulation. rTMS pulses were delivered for
10∼15 min at a 1 Hz rate. The merit of this protocol is that it produces
rTMS effects which last beyond the period of stimulation (for around
15 min)sosubjects canbe stimulatedfirstthentested.However,theTMS
effects of such a protocol are less event-related. In this study, because
subjects received rTMS while they were performing the task, any rTMS
by the task. The current study therefore not only probes the critical
involvements of Pre-SMA in the inhibitory control of the stop-signal task
but also the event-related rTMS effects over the stop-signal task.
Materials and methods
Nine volunteer college students (aged 21 to 35 years, mean 25.7, 7
male, 2 female, all right handed) took part in the experiment. All gave
informed consent prior to participation. The experiment was
approved by Institutional Review Board of the Veterans General
Testing took place in a sound attenuated room. Stimuli were
presented on a 19-inch CRT screen using video resolution of
C.-Y. Chen et al. / NeuroImage 44 (2009) 537–545
800×600 pixels and a vertical refresh rate of 100 Hz. The subjects sat
75 cm in front of the screen which was at eye level. The task was
programmed using E-prime running on a Pentium IV PC which
controlled the presentation of the stimuli as well as recording
In the stop-signal task, the stop signal delay (SSD) is the most
critical independent variable and it is manipulated by adjusting the
time between the onset of the go stimulus and that of the stop signal
(Logan,1994).The outcomeof theracebetween thego process andthe
stop process is reflected by the inhibition function. This describes the
probability of responding given a stop signal delay in accordance with
the race model of Logan and Cowan (1984). The stop signal reaction
time (SSRT) represents the latency of the stop process and it is the
most important dependent variable in the task. The SSRT can be
estimated from theobserved distributionofRTs in no-stop signaltrials
in combination with the inhibition function (Logan, 1994). According
to Band and colleagues' comprehensive review of the stop-signal
paradigm, several methods areavailable forestimatingthe SSRT (Band
et al., 2003; see also Logan,1994). In the current study, SSRTs for each
SSD were estimated using the integration method and one summary
SSRT was calculated by averaging the three SSRTs acquired in three
SSDs in our experiments (Logan, 1994; Band et al., 2003).
In the current study, each trial of the stop-signal task began with a
central fixation dot which appeared for 500 ms. Following offset of
this dot, a white target dot was presented to the left or right of the
fixation at 9° of eccentricity on the horizontal meridian (see Fig.1). On
75% of trials (go trials) subjects were required to make a key-press
response on a response box with the left index finger when the dot
was presented on the left, or with the right index finger when the dot
was presented on the right. On 25% of the trials (stop trials), the
central fixation dot reappeared and acted as an instruction to
withhold responses to the peripheral target.
In order to reduce the number of trials in task and thereby the
number of TMS pulses in the formal experiments, each subject's Mean
RT for go (no-stop-signal) trials and critical SSD were acquired with
three pre-TMS sessions (see Fig. 2). Every subject started with a
session of the choice RT task (50 trials). Subjects were asked to
respond to a target which appeared in the left or the right visual field
with their corresponding index fingers. They were encouraged to
make the responses correctly and as quickly as possible. The purpose
of this session was to obtain each subject's mean go RT and standard
deviation in the absence of stop signals. Each subject's mean go RTs
plus two standard deviations was set as his/her time limit for go RT
trialsin thesubsequent sessions.If thesubject did not respondquicker
than his/her time restriction in a go trial, the trial was counted as a
non-responding error and a warning beep was delivered. It has been
demonstrated that this procedure can effectively limit the strategy of
slow responses to avoid errors, which subjects may use to reduce the
rate of non-cancelled errors (Chen et al., 2008). Floden and Stuss
(2006) also used a similar procedure to evaluate baseline go RTand to
restrain strategic slowing. Moreover, Garavan et al's study (2002)
suggested that the pre-SMA was more involved in inhibitory control
when the time pressure was present in the go-nogotask. We therefore
used this procedure to reduce strategic slowing and to increase the
involvement of the pre-SMA in the task.
A practice session which was composed 24 go trials and 8 stop
trials followed the choice go RT session. The SSD was fixed at 170 ms
in the stop trials in this session, but the experimental sequence of the
trials in this session was otherwise identical to the abovementioned
sequence and the subsequent formal TMS sessions. After the subjects
performed the go time restricted session and the practice session,
they were required to carry out a critical SSD session (see Fig. 2). The
Fig.1. Task procedure. The stop-signal task consisted of go and stoptrials.Alltrials began centralfixation.Followingoffsetof thecentral fixation,awhite peripheral dot was presentedto
the left or rightof the fixation. Subjects were required to make a keypress response on a response box with the left index finger when the dot was presented on the leftor with the right
C.-Y. Chen et al. / NeuroImage 44 (2009) 537–545
purpose of this session was to estimate every subject's SSD at which
their non-cancelled rate would be around 50%. This session also
helps to decrease the number of trials in the formal TMS sessions
since the critical SSD does not then have to be obtained during the
TMS trials and there is then a better chance to observe the effects of
TMS (see also Chambers et al., 2007). The tracking procedure was
used for acquiring the critical SSD. According to the results of our
pilot experiments, the initial SSD was set at 170 ms. The SSD of each
subject was adjusted until the subject's accuracy on stop-trials
reached 50%. The program monitored subjects' performance block by
block. If the subject's non-cancelled rate was lower than 37.5%, the
SSD was increased by 40 ms. Conversely, if the non-cancelled rate
was higher than 62.5%, the SSD was decreased by 40 ms. A critical
SSD could be computed that represented the time delay required for
the subject to succeed in withholding a response in the stop trials
half of the time. Subject's critical SSD was determined when their
non-cancelled rate was within 37.5%∼62.5% for two consecutive
blocks. It usually took subjects less than 500 trials to obtain their
The main body of the experiment consisted of three conditions;
two of these involved TMS (over the left Pre-SMA and the vertex) and
one with no TMS. Three SSDs were presented to each subject based
on their individual critical SSDs; critical SSD, 40 ms less than critical
SSD, and 40 ms more than critical SSD. For example, if a subject's
critical SSD was 210 ms (acquired in the Critical SSD session), the
other two conditions were 170 ms and 250 ms. During blocks with
TMS, two pulses with an inter-pulse interval of 100 ms were
delivered over the relevant stimulation site concurrent with the
onset of the go signal.
Each experimental block included 48 trials and lasted approxi-
mately 4 min; the occurrence and order of the three stop signal
presentation conditions was randomized within each block. For each
TMS site, subjects received 20 blocks of trials, of which 10 blocks were
with TMS and the other 10 blocks were without TMS, serving as
control trials. An ABBAdesignwas used to control for sequence effects.
EachTMSsessionwas divided intotwosub-sessions. Subjects received
5 TMS and 5 control blocks totalling 10 blocks in the first sub-session.
After a break of 10 min, subjects received the second sub-session of 10
blocks. Subjects only received one TMS session in a day and received
the second TMS session a week after the first. Subjects were randomly
assigned to receive either the Pre-SMA TMS or the vertex session as
their first TMS session (see Fig. 2).
TMS parameters and site localization
A Magstim Super-Rapid Stimulator was used to deliver TMS at 60%
of maximum machine output (approx.1.2 Tesla, duration of one pulse:
less than 1 ms) over Pre-SMA and the vertex. A fixed stimulation level
was used because it has proven successful and replicable in many
studies and over a wide range of tasks (e.g.: Ashbridge et al., 1997;
Rushworth et al., 2002a,b; Muggleton et al., 2003; Hung et al., 2005;
Ellison and Cowey, 2007, Juan et al., 2008) and because motor cortex
excitability does not provide a good guide to TMS thresholds in other
cortical areas (Stewart et al., 2001). Stimulation was delivered via a
70 mm figure of eight coil held clamped in position with the handle
parallel to the sagittal midline (with the direction of the current over
the stimulation site travelling in the same lateral to medial direction).
The stimulation site for the Pre-SMA was localized in each subject
using a magnetic resonance image (MRI)-guided frameless stereotaxy
system (Brainsight, Rogue Research, Montreal, Canada). Briefly,
identification of the Pre-SMA site on the structural scans was achieved
by the following procedure. Individual MRIs were normalized against
a standard template using the FSL software package (FMRIB, Oxford).
This produced a matrix describing the transformation applied to the
structural scan to result in the normalised brain. This was then
reverse-applied to the coordinates for Pre-SMA (−4, 32, 51, Li et al.,
2006b see Fig. 3) to obtain the location of the site in the original
structural scan for each subject. The location was then marked on the
MRI scan in the Brainsight system.
After the Pre-SMA location had been identified in a subject's
structural MRI scan, a Polaris infra-red tracking system (Northern
Digital, Waterloo, Canada) was then used to co-register the positions
of anatomical landmarks on each subject's head which were also
visible on each MRI scan (bridge of nose, nose tip, left and right
intra-trageal notches). Another infra-red tracker was placed over the
TMS coil and was utilized to identify the scalp point over the
identified Pre-SMA site. For vertex stimulation the coil was held
anterior to the handle which was oriented parallel to the sagittal
midline. This site was localized by marking the point midway
between the intertragal notches and midway between the inion and
nasion. The scalp positions of the Pre-SMA and the vertex were
marked on a cloth swimming cap which was worn throughout the
Fig. 2. The procedures of the experimental sessions. Three pre-TMS sessions were carried out to establish each subject's baseline RTand to acquire his/her critical SSD. Subjects then
received two TMS sessions. The order of the two TMS sessions were counterbalanced across subjects.
C.-Y. Chen et al. / NeuroImage 44 (2009) 537–545
The go RTs were filtered by removing non-response trials, trials
with responses to the wrong target and trials with latencies below
200 ms. In addition, trials with latencies more than 2 standard
deviations away from each subject's mean Go RT in the baseline GO
session were also excluded from further analysis. When TMS was
applied over pre-SMA, the error rates for non-response trials, trials
with responses to the wrong target, the proportions of trials with
latencies below 200 ms and GO RT above the mean plus 2 SD were
0.002, 0.003, 0.017 and 0.046, respectively. The mean GO RT plus 2 SD
across subjects was 421 ms (SD of 82 ms).
The same procedures were also applied to the Vertex TMS trials.
When TMS was applied over vertex, the error rates for non-response
trials, trials with responses to the wrong target, trials with latencies
below 200 ms and GO RTabove the meanplus 2 SD were 0.003, 0.002,
0.008 and 0.03, respectively. The mean of GO RT plus 2 SD across
subjects was 425 ms (SD was 53 ms). These trials were excluded from
For the no TMS condition, the same exclusion criteria were again
applied and the error rates were 0.004, 0.001, 0.014 and 0.048,
respectively. The mean GO RT plus 2 SD across subjects was 413 ms
(SD was 35 ms).
Individual mean reaction times for the correct trials were
analyzed after removal of the abovementioned trials. The estimation
of the internal response time to the stop signal (SSRT) was
calculated using the distribution of go signal reaction times and
the probability of responding given a stop signal delay (the
inhibition function) in accordance with the race model of Logan
and Cowan (1984). The inhibition function described the probability
of responding (noncancelled rate) as a function of SSD. The latency
of the stop process can be estimated from the RTs on no-stop-signal
trials (the observed go RT distribution) in combination with the
inhibition function. In this study, SSRTs for each stop signal delay
were estimated using the integration method (Hanes et al., 1998;
Hanes and Carpenter, 1999) then averaged to obtain a summary
SSRT (i.e.: SSRTaverage in Band et al's terminology (2003)). We
followed the method introduced in Logan (1994; see also Logan and
Cowan, 1984; Band et al., 2003) to calculate the SSRT in each SSD.
Briefly, if the non-cancelled rate=x, at a given SSD, the stop
processes must have finished at point x of the observed go RT
distribution. The value of the x point minus SSD yields the SSRT. For
example, if SSD=130 ms, non-cancelled rate=0.4, and the 40th
percentile RT of the observed go RT distribution=330 ms, the
Fig. 4. Stop signal reaction time calculation. The figure shows the relationship between stop signal delay, the stop signal reaction time, and the distribution of go reaction times. The
distribution of go reaction times is integrated from the time of go signal presentation. For each stop-signal delay, a probability of responding is obtained. If the stop-signal delay of
50 ms resulted in an error rate=0.20, this means that the end of the stop process should be at a point equal to 20% of the go RT distribution. If the point of 20% of the go RT distribution
was 252 ms, so the observed SSRT would be 252−50=202 ms. The rest of the SSRTs were calculated with the same procedure. A summary SSRT was acquired by averaging the
observed three SSRTs that corresponded to 0.15bp (respond)b0.85 (Band et al., 2003).
Fig. 3. Site localization. The Pre-SMA magnetic stimulation sites were localized using
the Brainsight TMS-MRI co-registration system (Rogue Research, Montreal, Canada).
The Tailarach coordinates of −4, 32, 51 reported by Li et al. (2006b) were used and lay in
as the superior medial frontal cortex. The vertex was defined as a point midway
between the inion and the nasion and equidistant from the left and right intertrachial
C.-Y. Chen et al. / NeuroImage 44 (2009) 537–545
observed SSRT will be 330–130=200 ms for this SSD. Although this
method assumes that the SSRTs are constant across trials, violation
of this assumption does not significantly alter the outcome of the
analysis (Logan and Cowan, 1984; Hanes et al., 1998). Notwithstand-
ing this, Band et al. (2003) found observed SSRTs changed with
stop-signal delays. Consequently, we calculated the average SSRT for
stop signal delays where the proportion of trials in which subjects
failed to inhibit responding lay between 0.15 and 0.85 (Band et al.,
2003). This range of the inhibition function was selected because
the slope of this function can be approximated as a straight line.
Outside this range, ceiling and floor effects may result in a shallower
slope of the inhibition function (Band et al., 2003). Fig. 4 shows an
example how a summary SSRT was acquired in one subject.
Repeated measures analysis of variance was carried out for go RT
(for correct and noncancelled trials), SSRTand accuracy with factors of
TMS site (Pre-SMA, vertex and no TMS), and response hand. As there
was no significant effect of response hand (F(1,8)=1.559, P=0.247) (see
Supplementary Information A for a comparison of the performance of
the two hands). The data were collapsed for this factor and
the analysis repeated with it omitted. The descriptive data are
summarized in Table 1.
Go RTs (correct responses)
Fig. 5a shows the mean go RTs. There were no significant effects of
TMS condition on go RTs (pre SMA 285.7±24.9, vertex 284.2±21.6, no
TMS 285.5±29.8 ms, F(2,16)=0.113, P=0.894).
Go RTs (noncancelled responses)
Fig. 5b shows the mean go reaction times when responses were
not inhibited appropriately. Again there was no significant effect of
TMS on go RTs for these responses (pre SMA 269.5±20.7, vertex
269.6.2±18.4, no TMS 268.9±21.3 ms, F(2,16)=0.029, P=0.971).
Mean error rates
Fig. 5c shows the noncancelled rates. Significant differences were
observed among the noncancelled rates for the three TMS condi-
tions (F(2,16)=4.071, P=0.037). The mean noncancelled rate when
TMS was applied to Pre-SMA (0.65±0.14) was significantly higher
than when applied to vertex (0.56±0.12, t(8)=2.396, P=0.043) and
also higher than in the no TMS condition (0.57±0.11, t(8)=2.565,
Fig. 5d shows the inhibition function for the different TMS
conditions. The noncancelled rate was significantly increased with
the incrementof SSDs. (F(2,16)=131.8, P=0.000). The main effect of TMS
conditionwas significant (F(2,16)=6.367, P=0.009). The post hoc results
showed that the Pre-SMA TMS significantly increased the noncan-
celled rate in comparison to Vertex TMS (t(8)=2.396, P=0.044) and
NO TMS (t(8)=2.565, P=0.033) conditions. The interaction between
the SSD factor and the TMS conditions factor was not significant
Stop signal reaction times
Fig. 5e shows the mean SSRTs for the different TMS conditions.
There were significant differences across the different conditions
(F.(2,16)=9.866, P=0.002). The SSRT for applying TMS to Pre-SMA
(210.5±15.7 ms) was significantly longer than applying TMS to vertex
(193.7±15.1, t(8)=3.505, P=0.008) and no TMS condition (196.9±11.2,
In the study of the role of IFG in inhibitory control by Chambers
et al. (2006) they found a significant effect of session on the
modulation of performance by TMS. For this reason we analysed the
data described above with an additional factor of sub-session order.
The go reaction times of first-TMS and second-TMS
Fig. 6a shows the mean go reaction times of 1st TMS sub-session
and 2nd TMS sub-session for three stimulation site conditions.
Analysis of go RTs used ANOVA with sub-sessions and stimulation
sites as factors. There was no interaction between the go RTs of the
two sub-sessions across the different stimulation site conditions
(F(2,16)=0.759, P=0.484). Therewas no main effect for stimulation sites
(F(1,8)=0.175, P=0.841). There was also no main effect for sub-sessions
The mean reaction times of noncancelled trials of 1st TMS and 2nd TMS
There was no interaction between the mean reaction times of
noncancelled trials of the two sub-sessions across the different
stimulation site conditions (F(2,16)=0.428, P=0.659). There was no
main effect for stimulation sites (F(2,16)=0.423, P=0.662). There
was also no main effect for sub-sessions (F(1,8)=0.428, P=0.659)
The mean noncancelled rates of first-TMS and second-TMS
A significant interaction was observed between the mean
noncancelled rates of the two sub-sessions across different stimula-
tion site conditions (F(2,16)=11.022, P=0.001) (Fig. 6c).
In the 1st TMS sub-session, there was a simple main effect for
stimulation sites (F(2,16)=9.025, P=0.002). The mean noncancelled
rate for applying TMS to Pre-SMA (0.69±0.12) was significantly
higher than applying TMS to vertex (0.55±0.13, t(8)=3.924,
p=0.004) and no TMS condition (0.58±0.11, t(8)=3.892, p=0.005).
The noncancelled rate of 1st TMS sub-session (0.69±0.12) was
significantly higher than that of 2nd TMS sub-session (0.6±0.15)
while TMS was applied to Pre-SMA (F(1,8)=14.37, P=0.005). There
was no significant difference between the 1st TMS and 2nd TMS
sub-sessions while TMS was applied to vertex (F(1,8)=1.933,
P=0.202). There was also no significant difference between the
1st and 2nd sub-sessions in the no TMS condition (F(1,8)=0.452,
Go RT (ms)% Go% NoncancelledSSRT (ms)
% Go=percentage of successful go trials.
% Noncancelled=percentage of noncancelled trials.
SSRT=stop-signal reaction time.
All numbers are mean±standard deviation.
C.-Y. Chen et al. / NeuroImage 44 (2009) 537–545
Stop signal reaction times of first-TMS and second-TMS
In the case of the SSRT variable, a significant interaction effect
was seen between the SSRTs of two sub-sessions across three
stimulation sites (F(2,16)=32.719, P=0.000) (Fig. 6d). In 1st TMS sub-
session, there was a simple main effect for stimulation sites
(F(2,16)=30.274, P=0.000). The mean SSRT for applying TMS to Pre-
SMA (219.5±14.6 ms) was significantly longer than applying TMS to
vertex (191.3±13 ms, t(8)=6.189, P=0.000) and no TMS condition
(198.1±6.9 ms, t(8)=5.563, P=0.001). In 2nd TMS sub-sessions, there
was no main effect for stimulation sites (F(1,8)=2.065, P=0.159).
The SSRT of 1st TMS sub-sessions (219.5±14.6) was significantly
longer than that of 2nd TMS sub-sessions (202.8±15.2) while TMS
was applied to Pre-SMA (F(1,8)=11.928, P=0.009). The SSRT of 1st
TMS sub-sessions (191.3±13) was significantly shorter than that of
2nd TMS sub-sessions (200.1±15.8) while TMS was applied to vertex
(F(1,8)=6.496, P=0.034). There was also no significant difference
between the pre- and post-sub-sessions in the no TMS condition
In summary, our results indicated that TMS applied over the Pre-
SMA significantly affected subjects' inhibitory control. The Pre-SMA
TMS effects were reflected in the increase of noncancelled rates
(Fig. 5c), the alteration of the inhibition function (Fig. 5d) and the
increase of SSRTs (Fig. 5e). The Pre-SMA TMS effects seem more
substantial in sub-session I (Figs. 6c, d).
The ability to appropriately inhibit prepotent responses is necessary
to prevent certain behaviours in circumstances where to do so may be
associated with adverse consequences. Several areas in the prefrontal
cortex have been associated with the mechanisms underlying such
inhibitory control, with a network including left and right IFG, right
MFG, ACC, Pre-SMA, FEFand right inferior parietal lobule all implicated
2007; Hanes and Schall,1996; Ito et al., 2003; Li et al., 2006b; Stuphorn
et al., 2000; Stuphorn and Schall, 2006). It has previously been shown,
using TMS, that the IFG plays a causal role in inhibitory control
(Chambersetal., 2006, 2007). We investigatedtheroleof anotherof the
Fig. 5. Panel (a) to (e) show the mean go reaction times, mean reaction times of noncancelled trials, mean noncancelled rates, inhibition function of the three SSDs and stop-signal
reaction times across three different conditions (TMS, vertex and no TMS). Error bars show a 95% confidence interval (CI). CI was computed on a within-subjects basis in accordance
with Loftus and Masson (1994).
C.-Y. Chen et al. / NeuroImage 44 (2009) 537–545
implicated areas, the Pre-SMA, in inhibitory control. We tested the
hypothesis that this area performs functions related to inhibition of
responses, by use of a stop-signal task which allows measurement of
both impulsivity and inhibition (Logan et al.,1997).
TMS delivered overleft Pre-SMA resulted in effects consistent with
this hypothesis, producing both elevated SSRTs and increased error
rates compared to control stimulation. Both of these measures are
indices of inhibitory control, and so the results are related to a pattern
of disruption consistent with the findings of Li et al. (2006b). This
study found that activation of the Pre-SMA was correlated with
shorter SSRTs, which suggested that greater activity in Pre-SMA is
indicative of more effective inhibitory control. Our data showed that
TMS delivered over the left Pre-SMA resulted in deteriorated
inhibitory control for both hands. This pattern of results may seem
surprising on initial consideration. However, Isoda and Hikosaka
(2007) recently applied microstimulation to monkeys' Pre-SMA and
found that the Pre-SMA played a critical role in suppressing an
unwanted action and facilitating a desired one. Intriguingly, they also
found that a certain proportion of the recorded neurons responded to
ipsi-lateral saccadic responses or even responded to bilateral saccadic
responses. Similar findings were reported in Stuphorn and Schall's SEF
microstimulation study (2006). Furthermore, subjects were asked to
respond within a limited time period in the current study, and this
may increase the involvement of Pre-SMA (Garavan et al., 2002) with
Pre-SMA TMS therefore affecting motor selection and inhibition in
both hands. It is also worth mentioning that the effect here may be a
consequence of the medial location of the coil such that it might also
activate the right Pre-SMA to some extent then cause the left hand
effects, although this is considered unlikely.
The role of Pre-SMA in behavioural performance has been posited
to be one of error monitoring (e.g. Klein et al., 2007). The findings here
are inconsistent with this. The most obvious reason for conflict
between our data and an error monitoring account is that TMS
delivery occurs prior to a response. As such, the data are incompatible
with TMS modulation of activity related to the accuracy of
performance. TMS has high temporal specificity, indeed this is one
ofit's oft mentioned advantages, andconsequentlydelivery duringthe
trial is most likely to have effects, if any, on performing the trial
effectively, rather than on the much later, in psychophysical terms,
process of monitoring the outcome of the trial.
Even if TMS at this time point were to affect error monitoring, the
pattern of effects seen are not what would be predicted. The
disruption of feedback related to responses would be more inclined
to lead to less conservative responding due tothe reduced influence of
making an error. This would be manifested in the results as a reduced
go RT, in addition to increased error rates, and failure to inhibit
responses. The data obtained here are consistent with the findings of
Taylor et al. (2007). They found that Pre-SMA plays a direct and casual
role in conflict resolution of an action selection task using a
combination of TMS and electroencephalographic recording. Aron
and colleagues (2007b) conducted a fMRI study with a conditional
stop-signal paradigm and suggested similar functional role of the Pre-
SMA in conflict resolution.It hasalso been suggested thatthe Pre-SMA
is in charge of both response inhibition and response selection (e.g.
Mostofsky and Simmonds, 2008). It would also be crucial to test
whether the Pre-SMA becomes more important for response execu-
tion under conditions of increased response conflict.
Furthermore, our finding confirms the critical role of Pre-SMAwithin
the neural network of the Pre-SMA, IFG and STN, which are fundamen-
tally involved in inhibitory control in the stop-signal paradigm (Aron et
al., 2007b). We found that the effects of TMS over Pre-SMA also were
dependant upon familiarity with the task. Modulation of performance
for other tasks, notably for conjunction visual search performance and
TMS delivered over parietal cortex. Walsh et al. (1998) showed that TMS
related disruption of conjunction search no longer occurred when
subjects were well practiced at the task following learning over a period
of days. Data obtainedheremight reflectPre-SMAbeingmoreimportant
in response inhibition during the formation of new (inhibitory)
associations. Even if this is the case, it is noteworthy that there was no
indication of a learning effect across sessions. Another possible cause of
the session effects observed in the current study is that the functional
weight of the Pre-SMA in inhibitory control may change across time. In
Fig. 6. Panel (a to d) show the mean go reaction times, mean reaction times of
noncancelled trials, mean noncancelled rates and stop-signal reaction times of first-
TMS and second-TMS blocks separately for three different conditions (TMS, vertex and
no TMS). Error bars show a 95% confidence interval (CI).
C.-Y. Chen et al. / NeuroImage 44 (2009) 537–545
otherwords,otherneuralnodesintheinhibitorynetworksuchasIFGand Download full-text
STN may be more important once the inhibitory associations are more
familiar to the subject. Future work might fruitfully test these ideas
further (see also Chambers et al., 2006). In conclusion, the Pre-SMA has
been implicated in inhibitory control by a number of studies, both in
monkeys and in humans. TMS delivered over this area showed, using a
We are grateful to the reviewers' constructive comments and
with the preparation of the manuscript. This work was sponsored by the
National Science Council, Taiwan and the Academia Sinica, Taiwan. CYC,
OJLT, DLH and CHJ were supported by the National Science Council,
008-005-MY3) and Academia Sinica, Taiwan (AS-94-TP-C06). NGM was
supported by the Medical Research Council, UK.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.neuroimage.2008.09.005.
Armstrong, I.T., Munoz, D.P., 2003. Inhibitory control of eye movements during
oculomotorcountermanding in adults with attention-deficit hyperactivity disorder.
Exp. Brain Res. 152, 444–452.
Aron, A.R., Poldrack, R.A., 2006. Cortical and subcortical contributions to Stop signal
response inhibition: role of the subthalamic nucleus. J. Neurosci. 26, 2424–2433.
Aron, A.R., Fletcher, P.C., Bullmore, E.T., Sahakian, B.J., Robbins, T.W., 2003. Stop-signal
inhibition disrupted by damage to right inferior frontal gyrus in humans. Nat.
Neurosci. 6, 115–116.
Aron, A.R., Durston, S., Eagle, D.M., Logan, G.D., Stinear, C.M., Stuphorn, V., 2007a.
Converging evidence for a fronto-basal-ganglia network for inhibitory control of
action and cognition. J. Neurosci. 27, 11860–11864.
Aron, A.R., Behrens, T.E., Smith, S., Frank, M.J., Poldrack, R.A., 2007b. Triangulating a
cognitive control network using diffusion-weighted magnetic resonance imaging
(MRI) and functional MRI. J. Neurosci. 27, 3743–3752.
Ashbridge, E., Walsh, V., Cowey, A., 1997. Temporal aspects of visual search studied by
transcranial magnetic stimulation. Neuropsychologia 35, 1121–1131.
Band, G.P., van der Molen, M.W., Logan, G.D., 2003. Horse-race model simulations of the
stop-signal procedure. Acta. Psychol. (Amst) 112, 105–142.
an interactive race model of countermanding saccades. Psychol. Rev.114, 376–397.
Chambers, C.D., Mattingley, J.B., 2005. Neurodisruption of selective attention: insights
and implications. Trends Cogn. Sci. 9 (11), 542–550.
Chambers, C.D., Bellgrove, M.A., Stokes, M.G., Henderson, T.R., Garavan, H., Robertson,
I.H., Morris, A.P., Mattingley, J.B., 2006. Executive qbrake failureq following
deactivation of human frontal lobe. J. Cogn. Neurosci. 18, 444–455.
Chambers, C.D., Bellgrove, M.A., Gould, I.C., English, T., Garavan, H., McNaught, E.,
Kamke, M., Mattingley, J.B., 2007. Dissociable mechanisms of cognitive control in
prefrontal and premotor cortex. J. Neurophysiol. 98 (6), 3638–3647.
Chen, C.Y., Muggleton, N.G., Juan, C.H., Tzeng, O.J., Hung, D.L., 2008. Time pressure leads to
Chevrier, A.D., Noseworthy, M.D., Schachar, R., 2007. Dissociation of response inhibition
and performance monitoring inthe stop-signal task using event-relatedfMRI. Hum.
Brain Mapp. 28, 1347–1358.
Ellison, A., Cowey, A., 2007. Time course of the involvement of the ventral and dorsal
visual processing streams in a visuospatial task. Neuropsychologia 45, 3335–3339.
Floden, D., Stuss, D.T., 2006. Inhibitory control is slowed in patients with right superior
medial frontal damage. J. Cogn. Neurosci. 18, 1843–1849.
Garavan, H., Ross, T.J., Murphy, K., Roche, R.A., Stein, E.A., 2002. Dissociable executive
functions in the dynamic control of behavior: inhibition, error detection, and
correction. Neuroimage 17, 1820–1829.
Hanes, D.P., Schall, J.D., 1996. Neural control of voluntary movement initiation. Science
Hanes, D.P., Carpenter, R.H.,1999. Countermanding saccades in humans. Vision Res. 39,
saccades: visual, movement, and fixation activity. J. Neurophysiol. 79, 817–834.
Hung, J., Driver, J., Walsh, V., 2005. Visual selection and posterior parietal cortex: Effects
of repetitive transcranial magnetic stimulation on partial report analyzed by
Bundesen's theory of visual attention. J. Neurosci. 19, 9602–9612.
Isoda, M., Hikosaka, O., 2007. Switching from automatic to controlled action by monkey
medial frontal cortex. Nat. Neurosci. 10, 240–248.
Ito, S., Stuphorn, V., Brown, J.W., Schall, J.D., 2003. Performance monitoring by the
anterior cingulate cortex during saccade countermanding. Science 302, 120–122.
Juan, C-H., Walsh, V., 2003. Feedback to V1: A Reverse Hierarchy in Vision. Exp. Brain
Res. 150, 259–263.
Juan, C-H., Campana, G., Walsh, V., 2004. Cortical interactions in vision and awareness:
hierarchies in reverse. Prog. Brain Res. 144, 117–130.
Juan, C-H., Muggleton, N.G., Tzeng, O.J.L., Hung, D., Cowey, A., Walsh, V., 2008.
Segregation of visual selection and saccades in human frontal eye fields. Cereb.
Cortex 18 (10), 2410–2415.
Kalla, R., Muggleton, N.G., Juan, C.-H., Cowey, A., Walsh, V., 2008. The timing of the
involvement of the frontal eye fields and posterior parietal cortex in visual search.
Neuroreport 19 (10), 1067–1071.
Kennerley, S.W., Sakai, K., Rushworth, M.F., 2004. Organization of action sequences and
the role of the pre-SMA. J. Neurophysiol. 91, 978–993.
Klein, T.A., Endrass, T., Kathmann, N., Neumann, J., von Cramon, D.Y., Ullsperger, M.,
2007. Neural correlates of error awareness. Neuroimage 34 (4), 1774–1781.
Lau, H.C., Rogers, R.D., Haggard, P., Passingham, R.E., 2004a. Attention to intention.
Science 303, 1208–1210.
Lau, H.C., Rogers, R.D., Ramnani, N., Passingham, R.E., 2004b. Willed action and
attention to the selection of action. Neuroimage 21, 1407–1415.
Leung, H.C., Cai, W., 2007. Common and differential ventrolateral prefrontal activity
during inhibition of hand and eye movements. J. Neurosci. 27, 9893–9900.
Li, C.S., Chang, H.L., Hsu, Y.P., Wang, H.S., Ko, N.C., 2006a. Motor response inhibition in
children with Tourette's disorder. J. Neuropsychiatry Clin. Neurosci. 18, 417–419.
Li, C.S., Huang, C., Constable, R.T., Sinha, R., 2006b. Imaging response inhibition in a
stop-signal task: neural correlates independent of signal monitoring and post-
response processing. J Neurosci 26, 186–192.
Li, C.S., Huang, C., Yan, P., Bhagwagar,Z., Milivojevic, V., Sinha, R.,2008. Neural Correlates
of Impulse Control During Stop Signal Inhibition in Cocaine-Dependent Men.
Neuropsychopharmacology 33 (8), 1798–1806.
Loftus, G.R., Masson, M.E.J., 1994. Using confidence intervals in within-subject designs.
Psychon. Bull Rev. 1, 476–490.
Logan, G.D.,1994. On the ability to inhibit thought and action: a users' guide to the stop
signal paradigm. In: Dagenbach, D., Carr, T.H. (Eds.), Inhibitory processes in
attention, memory, and language. Academic Press, San Diego, pp. 189–239.
Logan, G.D., Cowan, W.B.,1984. On the ability to inhibit thought and action: a theory of
ac act of control. Psychol. Rev. 295–327.
Logan, G.D., Cowan, W.B., Davis, K.A., 1984. On the ability to inhibit simple and choice
reaction time responses: a model and a method. J. Exp. Psychol. Hum. Percept.
Perform. 10, 276–291.
Logan, G.D., Schachar, R.J., Tannock, R.,1997. Impulsivity and inhibitory control. Psychol.
Sci. 8, 60–64.
Mostofsky, S.H., Simmonds, D.J., 2008. Response inhibition and response selection: two
sides of the same coin. J. Cogn. Neurosci. 20 (5), 751–761.
Muggleton, N.G., Juan, C-H., Cowey, A., Walsh, V., 2003. Human frontal eye fields and
visual search. J. Neurophysiol. 89, 3340–3343.
Nachev, P., Rees, G., Parton, A., Kennard, C., Husain, M., 2005. Volition and conflict in
human medial frontal cortex. Curr. Biol. 15, 122–128.
O'Shea, J., Muggleton, N.G., Cowey, A., Walsh, V., 2004. Timing of target discrimination
in human frontal eye fields. J. Cogn. Neurosci. 16 (6), 1060–1067.
Picton, T.W., Stuss, D.T., Alexander, M.P., Shallice, T., Binns, M.A., Gillingham, S., 2007.
Effects of focal frontal lesions on response inhibition. Cereb. Cortex 17, 826–838.
Rushworth, M.F., Hadland, K.A., Paus, T., Sipila, P.K., 2002a. Role of the human medial
frontal cortex in task switching: a combined fMRI and TMS study. J. Neurophysiol.
Rushworth, M.F., Passingham, R.E., Nobre, A.C., 2002b. Components of switching
intentional set. J. Cogn. Neurosci. 14, 1139–1150.
organization of the medial frontal cortex. Curr. Opin. Neurobiol,.17, 220–227.
Schachar, R., Logan, G.D., 1990. Impulsivity and inhibitory control in normal
development and childhood psychopathology. Dev. Psychol. 26, 710–720.
Schachar, R., Tannock, R., Logan, G.D., 1993. Inhibitory control, impulsiveness, and
attention deficit hyperactivity disorder. Clin. Psychol. Rev. 13, 721–739.
Schachar, R., Tannock, R., Marriott, M., Logan, G., 1995. Deficient inhibitory control in
attention deficit hyperactivity disorder. J. Abnorm. Child. Psychol. 23, 411–437.
Schall, J.D., Boucher, L., 2007. Executive control of gaze by the frontal lobes. Cogn. Affect.
Behav. Neurosci. 7, 396–412.
Schall, J.D., Stuphorn, V., Brown, J.W., 2002. Monitoring and control of action by the
frontal lobes. Neuron 36, 309–322.
Stewart, L.M., Walsh, V., Rothwell, J.C., 2001. Motor and phosphene thresholds:
a transcranial magnetic stimulation correlation study. Neuropsychologia 39,
Stuphorn, V., Schall, J.D., 2006. Executive control of countermanding saccades by the
supplementary eye field. Nat. Neurosci. 9, 925–931.
Stuphorn, V., Taylor, T.L., Schall, J.D., 2000. Performance monitoring by the supple-
mentary eye field. Nature 408, 857–860.
Sumner,P., Nachev,P., Morris, P., Peters,A.M.,Jackson, S.R., Kennard,C., Husain, M.,2007.
Human medial frontal cortex mediates unconscious inhibition of voluntary action.
Neuron 54, 697–711.
Taylor, P.C., Nobre, A.C., Rushworth, M.F., 2007. Subsecond changes in top down control
exerted by human medial frontal cortex during conflict and action selection:
a combined transcranial magnetic stimulation electroencephalography study.
J. Neurosci. 27, 11343–11353.
Walsh, V., Ashbridge, E., Cowey, A., 1998. Cortical plasticity in perceptual learning
demonstrated by transcranial magnetic stimulation. Neuropsychologia 36, 45–49.
C.-Y. Chen et al. / NeuroImage 44 (2009) 537–545