Abstract

Stopping an action requires suppression of the primary motor cortex (M1). Inhibitory control over M1 relies on a network including the right inferior frontal cortex (rIFC) and the supplementary motor complex (SMC), but how these regions interact to exert inhibitory control over M1 is unknown. Specifically, the hierarchical position of the rIFC and SMC with respect to each other, the routes by which these regions control M1, and the causal involvement of these regions in proactive and reactive inhibition remain unclear. We used off-line repetitive TMS to perturb neural activity in the rIFC and SMC followed by fMRI to examine effects on activation in the networks involved in proactive and reactive inhibition, as assessed with a modified stop-signal task. We found repetitive TMS effects on reactive inhibition only. rIFC and SMC stimulation shortened the stop-signal RT (SSRT) and a shorter SSRT was associated with increased M1 deactivation. Furthermore, rIFC and SMC stimulation increased right striatal activation, implicating frontostriatal pathways in reactive inhibition. Finally, rIFC stimulation altered SMC activation, but SMC stimulation did not alter rIFC activation, indicating that rIFC lies upstream from SMC. These findings extend our knowledge about the functional organization of inhibitory control, an important component of executive functioning, showing that rIFC exerts reactive control over M1 via SMC and right striatum.

INTRODUCTION

Stopping a manual response from being executed requires suppression of the primary motor cortex (M1; for a review, see Stinear, Coxon, & Byblow, 2009). Converging lines of evidence suggest that M1 is under control of a network in which the right inferior frontal cortex (rIFC) and the supplementary motor complex (SMC) have important roles (reviewed in Aron, 2011; Chambers, Garavan, & Bellgrove, 2009). Yet, how the regions in this network interact with each other to exert inhibitory control over M1 is intensely debated but remains unclear. First, little is known about the relative position of the rIFC and SMC within the inhibitory control network. Some studies report that the SMC exerts control over M1 via the rIFC (Neubert, Mars, Buch, Olivier, & Rushworth, 2010; Aron, Behrens, Smith, Frank, & Poldrack, 2007), whereas other studies provide support for the view that the rIFC exerts control over M1 via the SMC (Hwang, Velanova, & Luna, 2010; Duann, Ide, Luo, & Li, 2009). Second, it is unclear which of the multiple pathways between rIFC, SMC, and M1 are crucial to inhibitory control. The rIFC and SMC may exert control over M1 via cortico-BG pathways through the striatum and subthalamic nucleus (STN; Zandbelt & Vink, 2010; Aron et al., 2007; Aron & Poldrack, 2006), but also via direct cortico-cortical connections (Buch, Mars, Boorman, & Rushworth, 2010; Mars et al., 2009). Third, recent theories (Aron, 2011; Braver, Gray, & Burgess, 2007) propose that inhibitory control can be divided into proactive mechanisms (response slowing in anticipation of a stop-signal) and reactive mechanisms (outright stopping in response to a stop-signal) and signatures of proactive and reactive control have been observed in M1 (Cai, Oldenkamp, & Aron, 2011; van den Wildenberg et al., 2010; Coxon, Stinear, & Byblow, 2006). Nevertheless, to what extent and how rIFC and SMC contribute to proactive and reactive inhibition over M1 remains largely unknown. The SMC appears to be involved in both proactive and reactive inhibition, because intracortical stimulation in nonhuman primates increases proactive response slowing (Stuphorn & Schall, 2006) and transcranial brain stimulation in humans influences reactive stopping (Hsu et al., 2011; Chen, Muggleton, Tzeng, Hung, & Juan, 2009). In contrast, the rIFC seems to be crucial to reactive inhibition only. Although rIFC stimulation has robust effects on reactive stopping (Verbruggen, Aron, Stevens, & Chambers, 2010; Chambers et al., 2006, 2007), it does not influence response slowing (Verbruggen et al., 2010). Supporting this view, activation in the rIFC occurs rather late in the trial so that it unlikely contributes to processes underlying proactive inhibition (Zandbelt, Bloemendaal, Neggers, Kahn, & Vink, in press).

A typical task for studying inhibitory control is the stop-signal paradigm (reviewed in Verbruggen & Logan, 2008). In this paradigm, go-signals requiring a response are occasionally followed by a stop-signal, indicating that the planned response should be stopped. Reactive inhibition is measured in terms of the stop-signal RT (SSRT), the time it takes to stop the planned response, which can be estimated with a mathematical model of stop-signal task performance (Logan & Cowan, 1984). Investigation of proactive inhibition requires the task to be divided into two or more conditions that differ in the degree of preparatory control. Typically, this is achieved through manipulation of the probability that a stop-signal occurs (Zandbelt & Vink, 2010; Vink et al., 2005; Ramautar, Kok, & Ridderinkhof, 2004; Logan & Burkell, 1986); as stop-signal probability increases, participants proactively slow down responding in an attempt to increase inhibition accuracy in case a stop-signal occurs. The stop-signal anticipation task (Figure 1; Zandbelt & Vink, 2010) involves such stop-signal probability manipulation and can be used to examine both proactive and reactive inhibition.

Figure 1. 

Stop-signal anticipation task. Three horizontal lines formed the background displayed continuously during the task. (A) In each trial, a bar moved at constant speed from the bottom–up, reaching the middle line in 800 msec. The main task was to stop the bar as close to the middle line as possible by pressing a button with the right thumb (i.e., the target RT was 800 msec). These trials are referred to as go trials. (B) In a minority of trials, the bar stopped moving automatically before reaching the middle line (i.e., the stop-signal), indicating that a response had to be stopped. These trials are referred to as stop trials. Stop trials in which a response was successfully inhibited are referred to as successful stop trials, those in which inhibition failed are referred to as failed stop trials. After failed stop trials (as well as failed go trials), a red cross was presented (i.e., negative feedback). The stop-signal delay (SSD), the interval between trial onset and presentation of the stop-signal, was initially 550 msec and varied from one stop trial to the next according to a staircase procedure: If stopping was successful, then stopping was made more difficult on the next stop trial by increasing SSD by 25 msec. The process was reversed when stopping failed. (C) The probability that a stop-signal would occur was manipulated across trials and was indicated by the color of the target response line (i.e., cue). There were five stop-signal probability levels: 0% (green cue), 17% (yellow cue), 20% (amber cue), 25% (orange cue), and 33% (red cue). In total, 414 go trials (0%, n = 234; 17%, n = 30; 20%, n = 30; 25% n = 54; 33%, n = 48) and 60 stop trials (17%, n = 6; 20%, n = 12; 25%, n = 18; 33%, n = 24) were presented in a single run in pseudorandom order. For more details on the stop-signal anticipation task, see Zandbelt and Vink (2010).

Figure 1. 

Stop-signal anticipation task. Three horizontal lines formed the background displayed continuously during the task. (A) In each trial, a bar moved at constant speed from the bottom–up, reaching the middle line in 800 msec. The main task was to stop the bar as close to the middle line as possible by pressing a button with the right thumb (i.e., the target RT was 800 msec). These trials are referred to as go trials. (B) In a minority of trials, the bar stopped moving automatically before reaching the middle line (i.e., the stop-signal), indicating that a response had to be stopped. These trials are referred to as stop trials. Stop trials in which a response was successfully inhibited are referred to as successful stop trials, those in which inhibition failed are referred to as failed stop trials. After failed stop trials (as well as failed go trials), a red cross was presented (i.e., negative feedback). The stop-signal delay (SSD), the interval between trial onset and presentation of the stop-signal, was initially 550 msec and varied from one stop trial to the next according to a staircase procedure: If stopping was successful, then stopping was made more difficult on the next stop trial by increasing SSD by 25 msec. The process was reversed when stopping failed. (C) The probability that a stop-signal would occur was manipulated across trials and was indicated by the color of the target response line (i.e., cue). There were five stop-signal probability levels: 0% (green cue), 17% (yellow cue), 20% (amber cue), 25% (orange cue), and 33% (red cue). In total, 414 go trials (0%, n = 234; 17%, n = 30; 20%, n = 30; 25% n = 54; 33%, n = 48) and 60 stop trials (17%, n = 6; 20%, n = 12; 25%, n = 18; 33%, n = 24) were presented in a single run in pseudorandom order. For more details on the stop-signal anticipation task, see Zandbelt and Vink (2010).

Here, we used off-line repetitive TMS (rTMS) to interfere with neural activity in the rIFC and SMC and fMRI and the stop-signal anticipation task to examine effects on activation in the networks involved in proactive and reactive inhibition. Each participant underwent three rTMS-fMRI sessions; they received real stimulation over the rIFC, real stimulation over the SMC, and sham stimulation over the right superior parietal lobe in a counterbalanced crossover design. This experimental setup enabled us to visualize the interactions between the rIFC, the SMC, and the BG during proactive and reactive inhibition over M1.

On the basis of findings from studies that we discussed above, we formulated the following hypotheses. First, if the rIFC exerts proactive and reactive control over M1 via the SMC, then rIFC stimulation should alter activation in the rIFC, SMC, and M1, whereas SMC stimulation should only influence SMC and M1 activation, but not right IFC activation. Alternatively, if the SMC exerts proactive and reactive inhibitory control over M1 via the rIFC, then SMC stimulation should alter activation in the SMC, the rIFC, and M1, whereas rIFC stimulation should only alter rIFC and M1 activation, but not SMC activation. Second, if cortico-BG pathways are also crucial for proactive and reactive inhibitory control, then rIFC and SMC stimulation should not only alter activation in the stimulated region and M1 but also in BG regions, such as the striatum and the STN. In contrast, if the BG are not crucial for these functions and rIFC and SMC influence M1 directly, then we expect stimulation effects only in cortical regions, such as rIFC, SMC, and M1. Third, we expected that SMC stimulation would alter proactive as well as reactive inhibition, but that rIFC stimulation would affect reactive inhibition only.

METHODS

Participants

All 24 participants (mean age = 24.1 years, range = 20–38 years; 12 women) were right-handed (Oldfield, 1971), had normal or corrected-to-normal vision, had no signs of present or past neurological or psychiatric illness (Sheehan et al., 1998), had no contraindication to TMS (Keel, Smith, & Wassermann, 2001), and gave written informed consent. None of the participants reported adverse effects after rTMS, and none of them abandoned the study. Data from three participants had to be excluded from the fMRI analysis because of excessive head motion during data acquisition.

Experimental Design

All participants underwent three rTMS-fMRI sessions that were separated by at least 1 week: real stimulation over the rIFC, real stimulation over the SMC, and sham stimulation over the right superior parietal lobe. The order of these sessions was entirely counterbalanced across participants. Each session lasted about 75 min and consisted of the following steps: identification of the stimulation site (10 min), resting motor threshold (RMT) estimation (15 min), rTMS procedure (20 min), and fMRI procedure (30 min).

Identification of the Stimulation Site

All sessions started with identification of the stimulation site on the head of the participant. The Montreal Neurological Institute coordinates of the stimulation sites were x = 56, y = 16, z = −4 for the rIFC, x = 8, y = 0, z = 68 for the SMC, and x = 28, y = −56, z = 60 for sham stimulation over the right superior parietal lobe. These regions were activated both during proactive and reactive inhibition in previous fMRI studies using the same stop-signal task (Zandbelt, van Buuren, Kahn, & Vink, 2011; Zandbelt & Vink, 2010). These coordinates were marked on a normalized T1-weighted scan of each participant that was obtained before the first rTMS-fMRI session. The individual stimulation sites were derived by reversing the normalization procedure of the T1-weighted scan. Using this native T1-weighted scan and a frameless stereotactic neuronavigation system (The NeuralNavigator, BrainScienceTools B.V., Utrecht, The Netherlands, http://www.neuralnavigator.com, for details and validation see Neggers et al., 2004), we identified the site where the TMS coil had to be positioned to stimulate the rIFC and SMC. These sites were marked on a tight-fitting Lycra cap, which participants wore during the entire rTMS-fMRI session.

RMT Estimation

In each session and before rTMS, we determined the RMT according to a standardized procedure (Schutter & van Honk, 2006). RMT was defined as the lowest stimulator intensity at which 5 of 10 consecutive TMS pulses over the right primary motor cortex evoked a visually identifiable twitch in the left abductor pollicis brevis.

rTMS Procedure

We delivered rTMS with a Magstim Double 70-mm Air Film Coil (real rTMS) or a Magstim Double 70-mm Air Film Placebo Coil (sham rTMS), connected to a magnetic stimulator (Magstim Rapid2, Magstim, Welwyn Garden City, United Kingdom). The placebo coil is identical to the real coil in appearance, operation, and acoustic properties but did not produce the magnetic effects associated with TMS. The rTMS protocol (Iyer, Schleper, & Wassermann, 2003) consisted of 20 trains of 30 6-Hz pulses at 90% RMT with an intertrain interval of 25 sec, followed by 600 1-Hz pulses at 110% RMT. This protocol has previously been used to produce a long-lasting (i.e., >45 min) decrease of motor-evoked potentials (Iyer et al., 2003) and a long-lasting increase in resting state functional connectivity (Zandbelt et al., 2009), meaning that this protocol is appropriate for off-line mapping of stimulation effects on brain function (Siebner et al., 2009). During rTMS, participants wore earplugs to protect against TMS noise and were seated in a comfortable chair with their head stabilized by a forehead and chin rest. The stimulation site was marked on the cap with a vitamin E capsule for post hoc identification on the MRI scan.

fMRI Procedure

Immediately after rTMS, participants were placed in a 3.0-T MRI scanner (Philips Medical Systems, Best, the Netherlands). The fMRI scan started within 5 min after rTMS was completed. We collected 622 whole-brain T2*-weighted EPIs with BOLD contrast while participants performed the stop-signal anticipation task (see below) and a T1-weighted image thereafter, using scan parameters identical to those described before (Zandbelt & Vink, 2010).

Stop-signal Anticipation Task

Participants performed a modified version of the stop-signal paradigm (for a review, see Verbruggen & Logan, 2008), called the stop-signal anticipation task (Zandbelt et al., 2011; Zandbelt & Vink, 2010). The task is described in detail in Figure 1. In brief, the majority of trials were go trials, in which participants had to make a response. A minority of trials were stop trials, in which participants had to cancel their response. Stop-signal probability, indicated by a visual cue, was manipulated across trials and varied in five steps from 0% to 33%.

Data Analysis—Task Performance

For go trials, we computed mean RTs and omission error rates (i.e., the proportion of go trials in which no response was given), separately for each stop-signal probability level. Furthermore, we calculated the stop-signal probability slope (SSPS), a single-value index of proactive inhibition, defined as the linear effect of stop-signal probability on go trial RT. The SSPS was used only as a covariate of interest in the fMRI analysis. For stop trials, we computed mean RTs on failed stop trials, success rates for stop trials, and the SSRT (Logan & Cowan, 1984). The SSRT measures the time it takes to inhibit a response and is an index of reactive inhibition. SSRT was calculated according to the integration method (Verbruggen & Logan, 2009a). For each of these indices, the effect of stimulation was computed as the difference between real stimulation and sham stimulation (i.e., rIFC vs. sham and SMC vs. sham). We used these difference scores for statistical analyses (see below) because they reduce variability because of interindividual differences, thereby increasing statistical power (Overall & Tonidandel, 2010).

In keeping with previous studies (Zandbelt et al., 2011; Zandbelt & Vink, 2010; Verbruggen & Logan, 2009b; Vink, Ramsey, Raemaekers, & Kahn, 2006; Vink et al., 2005), proactive inhibition was measured as the effect of stop-signal probability on go trial RT. Reactive inhibition was studied in terms of the SSRT. To assess the validity of our SSRT estimates, we assessed the assumptions of the race model (see Verbruggen & Logan, 2009a).

We analyzed the effect of stimulation on go trial RT and go trial omission error rate using two repeated measures ANOVAs with “stop-signal probability” (0%, 17%, 20%, 25%, 33%) and “stimulation site” (rIFC, SMC) as within-subject factors. Note that, in such an analysis, the intercept reflects the difference between real stimulation and sham stimulation. A stronger positive (or negative) effect of stop-signal probability after real versus sham stimulation would indicate increased (or decreased) proactive inhibition. Similarly, a positive (or negative) SSPS after real versus sham stimulation would indicate increased (or decreased) of proactive inhibition. We analyzed the effect of stimulation on SSRT, stopping rate, and RTs on failed stop trials in three repeated measures ANOVAs, with “stimulation site” (rIFC, SMC) as within-subject factor. A shorter (or longer) SSRT after real versus sham stimulation would indicate improved (or declined) reactive inhibition. All analyses of task performance were conducted at the 5% level of statistical significance and Greenhouse–Geisser correction was used when the assumption of sphericity was violated.

Data Analysis—fMRI

Functional MRI data were analyzed with SPM5 (www.fil.ion.ucl.ac.uk/spm/software/spm5/). Preprocessing and first-level statistical analysis was performed as previously described (Zandbelt & Vink, 2010). Preprocessing involved correction for slice timing differences (resampling to the middle slice using Fourier interpolation), realignment for head motion correction (using fourth-degree B-spline interpolation), spatial normalization to the Montreal Neurological Institute template brain, and spatial smoothing (6-mm FWHM Gaussian kernel) to accommodate interindividual differences in neuroanatomy.

For each participant, we constructed a first-level general linear model containing the three rTMS-fMRI sessions. For each session, the following events of interest were included as regressors: successful stop trials, failed stop trials, and go trials with stop-signal probability > 0%. Go trials with stop-signal probability of 0% were not explicitly modeled and constituted an implicit baseline. For go trials with stop-signal probability > 0%, we included two parametric regressors modeling RT and stop-signal probability level. The regressor coding for stop-signal probability was used to investigate proactive inhibition (see below), whereas the regressor coding for RT was included because it has been suggested that stochastic variation in RTs may drive activation seen in inhibitory control contrasts (Zandbelt & Vink, 2010; Aron & Poldrack, 2006). Previously, we verified that parametric regressors showed sufficiently low correlation for reliable parameter estimation (Zandbelt & Vink, 2010). All regressors were modeled at response onset (or target RT for successful stop trials) by convolving delta functions with a canonical hemodynamic response function. The fMRI data were high-pass filtered (cutoff, 128 sec) to remove low-frequency drifts. A first-order autoregressive model was used to model the remaining serial correlations.

We generated contrast images for proactive inhibition (defined as the parametric effect of stop-signal probability on go trials) and reactive inhibition (defined as (1) the difference between successful stop trials and go trials in contexts in which stop-signals never occur (successful stop vs. go0%) and (2) the difference between successful stop trials and go trials in contexts in which stop-signals occur infrequently (successful stop vs. go17–33%), separately for three TMS conditions: sham, rIFC versus sham, and SMC versus sham. The sham contrast images were used to study activation during proactive and reactive inhibition under baseline conditions. The rIFC versus sham and SMC versus sham contrast images were used to examine how activation during proactive and reactive inhibition was influenced by rIFC and SMC stimulation. We analyzed the difference between real stimulation and sham stimulation to reduce variability because of interindividual differences, thereby increasing statistical power (Overall & Tonidandel, 2010). Contrast images were used for ROI analyses as well as a whole-brain voxel-wise analyses.

We started with the ROI analysis, because we had specific hypotheses about the influence of rIFC and SMC stimulation on activation in seven key ROIs of the inhibitory control network: rIFC, SMC, left striatum, right striatum, left STN, right STN, and left M1 (for details, see Figure 4). All these ROIs were defined as spheres with 6-mm radius around local maxima from a previous study (Zandbelt & Vink, 2010) in which an independent sample of healthy young adults performed the same stop-signal task, except for the left and right STN. STN ROIs were defined as 6-mm spheres around the coordinate [±12, −16, −4], in accordance with a human BG template (Prodoehl, Yu, Little, Abraham, & Vaillancourt, 2008) and a previous study investigating STN activation during inhibitory control (Aron & Poldrack, 2006). For each ROI, we extracted mean contrast estimates (i.e., β values) from the proactive inhibition and reactive inhibition contrast images of all rTMS conditions (sham, rIFC vs. sham, and SMC vs. sham). To verify whether ROIs were activated during baseline, we analyzed proactive and reactive contrast estimates of the sham rTMS condition, using one-sample t tests. To assess the effect of rIFC and SMC stimulation on activation during proactive and reactive inhibition, we analyzed proactive and reactive contrast estimates of the rIFC versus sham and SMC versus sham rTMS condition, using two-way ANOVAs with Stimulation Site (rIFC, SMC) and ROI (rIFC, SMC, left striatum, right striatum, left STN, right STN, left M1) as within-subject factors and Behavioral Effect (proactive: ΔSSPSreal-sham; reactive: ΔSSRTreal-sham) as between-subjects covariate of interest. Note that in such an analysis, the intercept reflects the difference between real stimulation and sham stimulation. All ROI analyses were performed at the 5% level of statistical significance and Greenhouse–Geisser correction was used when the assumption of sphericity was violated.

The ROI analyses were succeeded by whole-brain voxel-wise random effects analyses to explore activation during proactive and reactive inhibition outside the key ROIs of the inhibitory control network. To visualize whole-brain baseline activation during proactive and reactive inhibition, we analyzed proactive and reactive inhibition contrast images from the sham rTMS condition, using one-sample t tests. To explore effects of rIFC and SMC stimulation on whole-brain activation during proactive and reactive inhibition, we analyzed proactive and reactive inhibition contrast images from the rIFC versus sham and SMC versus sham rTMS conditions, using a factorial analysis with Stimulation Site (rIFC, SMC) as within-subject factor and Behavioral Effect (proactive: ΔSSPSrIFC-sham, ΔSSPSSMC-sham; reactive: ΔSSRTrIFC-sham, ΔSSRTSMC-sham) as between-subject covariate of interest. All group statistical parametric maps were tested for significance using cluster-level inference.

RESULTS

Behavior

Table 1 shows descriptive statistics of task performance on go and stop trials.

Table 1. 

Descriptive Statistics for Go and Stop Trials across rTMS Sessions

Trial Type
rTMS Session
Sham
rIFC
SMC
Go 
Omission error rate 0% (%) 3.7 ± 4.0 3.6 ± 4.2 2.1 ± 1.8 
Omission error rate 17–33% (%) 3.5 ± 1.7 2.6 ± 1.6 2.6 ± 1.7 
RT 0% (msec) 806 ± 14 805 ± 13 802 ± 15 
RT 17–33% (msec) 835 ± 26 832 ± 23 835 ± 28 
 
Stop 
Success rate (%) 42.8 ± 5.1 44.4 ± 3.6 44.4 ± 4.4 
SSD (msec) 516 ± 27 516 ± 22 517 ± 25 
SSRT (msec) 327 ± 19 318 ± 13 320 ± 16 
Failed RT (msec) 818 ± 25 819 ± 20 817 ± 26 
Trial Type
rTMS Session
Sham
rIFC
SMC
Go 
Omission error rate 0% (%) 3.7 ± 4.0 3.6 ± 4.2 2.1 ± 1.8 
Omission error rate 17–33% (%) 3.5 ± 1.7 2.6 ± 1.6 2.6 ± 1.7 
RT 0% (msec) 806 ± 14 805 ± 13 802 ± 15 
RT 17–33% (msec) 835 ± 26 832 ± 23 835 ± 28 
 
Stop 
Success rate (%) 42.8 ± 5.1 44.4 ± 3.6 44.4 ± 4.4 
SSD (msec) 516 ± 27 516 ± 22 517 ± 25 
SSRT (msec) 327 ± 19 318 ± 13 320 ± 16 
Failed RT (msec) 818 ± 25 819 ± 20 817 ± 26 

Proactive Inhibition: rIFC and SMC Stimulation Did Not Alter Response Slowing

Figure 2A shows mean go trial RT as a function of stop-signal probability at baseline (sham rTMS) and after rIFC and SMC stimulation. At baseline, go trial RTs increased as a function of stop-signal probability (linear contrast, F(1, 23) = 52.41, p < .001), indicating that participants slowed down proactively in anticipation of a stop-signal, replicating previous observations (Zandbelt et al., in press, 2011; Zandbelt & Vink, 2010; Verbruggen & Logan, 2009c; Ramautar, Slagter, Kok, & Ridderinkhof, 2006; Vink et al., 2005, 2006; Ramautar et al., 2004; Logan & Burkell, 1986). Go trial omission errors did not vary with stop-signal probability (linear contrast, F(1, 23) < 1, p = .46).

Figure 2. 

Effect of rIFC and SMC stimulation on task performance. (A) Proactive inhibition. Mean effect of rIFC stimulation (red) and SMC stimulation (blue), relative to sham stimulation (dashed line), on the effect of stop-signal probability on go trial RT. (B) Reactive inhibition. Mean effect of rIFC stimulation (red) and SMC stimulation (blue), relative to sham stimulation (dashed line) on the SSRT (reactive inhibition). Error bars indicate 95% confidence intervals. (C) Normalized inhibition functions, showing the probability of successful stopping against the Z-transformed relative finishing time (Logan & Cowan, 1984). The solid transparent graphs represent individual inhibition functions in the rIFC (red), SMC (blue), and sham (black) sessions. The dashed, smooth opaque graphs show a Weibull function fitted to the group averaged normalized inhibition function.

Figure 2. 

Effect of rIFC and SMC stimulation on task performance. (A) Proactive inhibition. Mean effect of rIFC stimulation (red) and SMC stimulation (blue), relative to sham stimulation (dashed line), on the effect of stop-signal probability on go trial RT. (B) Reactive inhibition. Mean effect of rIFC stimulation (red) and SMC stimulation (blue), relative to sham stimulation (dashed line) on the SSRT (reactive inhibition). Error bars indicate 95% confidence intervals. (C) Normalized inhibition functions, showing the probability of successful stopping against the Z-transformed relative finishing time (Logan & Cowan, 1984). The solid transparent graphs represent individual inhibition functions in the rIFC (red), SMC (blue), and sham (black) sessions. The dashed, smooth opaque graphs show a Weibull function fitted to the group averaged normalized inhibition function.

After frontal stimulation, the RT profile was almost identical (intercept, F(1, 23) < 1, p = .68) and did not differ between rIFC and SMC stimulation (Stimulation Site, F(1, 23) < 1, p = .55; Stimulation Site × Stop-Signal probability linear contrast, F(1, 23) < 1, p = .45), but participants tended to make less go omission errors overall (intercept, F(1, 23) = 4.43, p = .046; Stimulation Site, F(1, 23) < 1, p = .33; Stimulation Site × Stop-Signal probability linear contrast, F(1, 23) = 1.93, p = .18). These findings suggest that rIFC and SMC stimulation did not influence proactive inhibition.

Reactive Inhibition: rIFC and SMC Stimulation Shortened SSRT

Figure 2B depicts the SSRT at baseline and after rIFC and SMC stimulation. At baseline, the SSRT was on average 327 msec, similar to previous reports of the stop-signal anticipation task (Zandbelt et al., in press, 2011; Zandbelt & Vink, 2010).

After frontal stimulation, participants became significantly faster in reactive stopping (intercept, F(1, 23) = 6.66, p = .017), and this effect was similar for rIFC and SMC stimulation (Stimulation Site, F(1, 23) < 1, p = .63). There was a marginal effect of frontal stimulation on stop success rate (intercept, F(1, 23) = 3.46, p = .08; Stimulation Site, F(1, 23) < 1, p = .99), but no significant effect on RTs on stop failure trials (intercept, F(1, 23) < 1, p = .92; Stimulation Site, F(1, 23) < 1, p = .61). These results indicate that rIFC and SMC stimulation facilitated reactive inhibition.

In an additional analysis, we verified whether task performance was in line with predictions from the race model (Logan & Cowan, 1984), the theory that underlies computation of the SSRT. We analyzed inhibition functions to see if the proportion of successfully inhibited responses decreased with a later stop-signal onset, and we verified whether RTs were faster on stop failure trials than on go trials. As shown in Figure 2C, normalized inhibition functions confirmed that the probability of successfully stopping decreases with later stop-signal onset (i.e., lower ZRFT values) across participants and sessions. Moreover, RTs on failed stop trials were indeed faster than RTs on go trials (trial type, F(1, 23) = 185.2, p < .001), and this effect did not differ between stimulation sites (Stimulation Site, F(2, 46) < 1, p = .97; Trial Type × Stimulation Site interaction, F(2, 46) = 1.83, p = .17). Thus, under all rTMS conditions, stop-signal anticipation task performance was consistent with predictions from the race model.

Functional MRI

Proactive Inhibition: rIFC and SMC Stimulation Did Not Alter Activation Levels

Figure 3A shows the results of the ROI analysis, depicting mean activation levels during proactive inhibition in seven ROIs of the inhibitory control network at baseline and after rIFC and SMC stimulation. At baseline, there was a positive parametric effect in the rIFC and SMC (all ps < .05), indicating an increase in go trial activation as a function of stop-signal probability, in line with previous reports (Zandbelt et al., in press; Jahfari, Stinear, Claffey, Verbruggen, & Aron, 2010; Zandbelt & Vink, 2010; Chikazoe et al., 2009; Vink et al., 2005).

Figure 3. 

Effect of rIFC and SMC stimulation on activation during proactive inhibition in the seven ROIs. (A) Mean effect of rIFC stimulation (red) and SMC stimulation (blue), relative to sham stimulation (dashed line), on activation during proactive inhibition (i.e., the parametric effect of stop-signal probability on go trial activation). (B) Relationship between the neural effect of stimulation (vertical axis) and the behavioral effect of stimulation (horizontal axis) during proactive inhibition across participants. Positive (or negative) values on the vertical axis indicate increased (or decreased) activation or increased (or decreased) deactivation after real stimulation versus sham stimulation. Positive (or negative) values on the horizontal axis represent steeper (or flatter) SSPS after real stimulation versus sham stimulation. Error bars and bands represent 95% confidence intervals and bands.

Figure 3. 

Effect of rIFC and SMC stimulation on activation during proactive inhibition in the seven ROIs. (A) Mean effect of rIFC stimulation (red) and SMC stimulation (blue), relative to sham stimulation (dashed line), on activation during proactive inhibition (i.e., the parametric effect of stop-signal probability on go trial activation). (B) Relationship between the neural effect of stimulation (vertical axis) and the behavioral effect of stimulation (horizontal axis) during proactive inhibition across participants. Positive (or negative) values on the vertical axis indicate increased (or decreased) activation or increased (or decreased) deactivation after real stimulation versus sham stimulation. Positive (or negative) values on the horizontal axis represent steeper (or flatter) SSPS after real stimulation versus sham stimulation. Error bars and bands represent 95% confidence intervals and bands.

After frontal stimulation, mean activation in the inhibitory control network as a whole did not differ from baseline (intercept, F(1, 19) < 1, p = .52) and there was no difference between rIFC and SMC stimulation on overall network activation (site, F(1, 19) < 1, p = .81) or activation in any of the individual regions (ROI × Site, F(2.2, 41.2) < 1, p = .39). This may suggest that rIFC and SMC stimulation had no effect on activation during proactive inhibition. On the other hand, because the effect of low-frequency rTMS may vary substantially between individuals (Daskalakis et al., 2006; Romero, Anschel, Sparing, Gangitano, & Pascual-Leone, 2002; Sommer, Wu, Tergau, & Paulus, 2002; Maeda, Keenan, Tormos, Topka, & Pascual-Leone, 2000), this null finding might reflect that proactive slowing increased in some participants but decreased in others. Therefore, we also considered the neural effect of stimulation in relation to its behavioral effect. Figure 3B demonstrates for each ROI how the effect of rIFC and SMC stimulation on activation and proactive slowing varied across participants. Even when we studied the neural effect of stimulation in relation to its behavioral effect across ROIs, there was neither a common effect of rIFC and SMC stimulation (ΔSSPS × ROI, F(2.3, 44.4) < 1, p = .75) nor a differential effect (ΔSSSP × ROI × site, F(1.3, 20.5) < 1, p = .95). Thus, rIFC and SMC stimulation did not significantly alter activation during proactive inhibition within key regions of the inhibitory control network.

Finally, we examined rTMS effects on activation outside the key ROIs, using a whole-brain voxel-wise analysis. Figure 4 (yellow areas) shows brain regions in which activation on go trials increased as a function of stop-signal probability at baseline.1 A wide network of brain regions was found to be activated, including broad frontal and parietal lobe areas in both hemispheres, a cluster spanning the SMC and ACC, the posterior cingulate cortex, and a cluster spanning the right insula and right striatum. This corresponds to activation patterns previously reported in association with proactive inhibition (Zandbelt et al., in press, 2011; Jahfari et al., 2010; Zandbelt & Vink, 2010; Chikazoe et al., 2009; Vink et al., 2005). Even so, after frontal stimulation, there were no significant changes in whole-brain activation levels during proactive inhibition. Taken together, these findings suggest that rIFC and SMC stimulation did not affect activation during proactive inhibition.

Figure 4. 

Baseline (sham rTMS) activation pattern during proactive inhibition and reactive inhibition and location of ROIs in the inhibitory control network. Activation related to proactive inhibition (yellow) and reactive inhibition (red/magenta, activation; blue/cyan, deactivation) was significant (p < 05, family-wise error corrected) at the cluster level. ROIs (green) were the stimulated rIFC (top left; x = 56, y = 16, z = −4), the stimulated SMC (bottom left; x = 8, y = 0, z = 68), left striatum (top right; x = −20, y = 8, z = 0), right striatum (top right; x = 28, y = 8, z = −4), left STN (bottom right; x = −12, y = 16, z = −4), right STN (bottom right; x = 12, y = 16, z = −4), and left primary motor cortex (bottom right; x = −52, y = −16, z = 44). Each region was defined as a sphere with 6-mm radius. Numbers indicate the y coordinates of the coronal slices (in Montreal Neurological Institute space). The center image depicts a top view representation of a normalized brain, showing the location of coronal slices in green.

Figure 4. 

Baseline (sham rTMS) activation pattern during proactive inhibition and reactive inhibition and location of ROIs in the inhibitory control network. Activation related to proactive inhibition (yellow) and reactive inhibition (red/magenta, activation; blue/cyan, deactivation) was significant (p < 05, family-wise error corrected) at the cluster level. ROIs (green) were the stimulated rIFC (top left; x = 56, y = 16, z = −4), the stimulated SMC (bottom left; x = 8, y = 0, z = 68), left striatum (top right; x = −20, y = 8, z = 0), right striatum (top right; x = 28, y = 8, z = −4), left STN (bottom right; x = −12, y = 16, z = −4), right STN (bottom right; x = 12, y = 16, z = −4), and left primary motor cortex (bottom right; x = −52, y = −16, z = 44). Each region was defined as a sphere with 6-mm radius. Numbers indicate the y coordinates of the coronal slices (in Montreal Neurological Institute space). The center image depicts a top view representation of a normalized brain, showing the location of coronal slices in green.

Reactive Inhibition: rIFC and SMC Stimulation Caused Activation Changes in rIFC, SMC, Right Striatum, and Left M1

Figure 5A depicts mean activation during reactive inhibition, as measured with the stop versus go0% contrast, in the seven inhibitory control network ROIs at baseline and after rIFC and SMC stimulation. At baseline, all ROIs were activated, except for the left M1 that was deactivated (all ps < .05) and the right STN in which activation did not reach significance. These findings are broadly in line with activation patterns previously observed (Boehler, Appelbaum, Krebs, Hopf, & Woldorff, 2010; Sharp et al., 2010; Zandbelt & Vink, 2010; Leung & Cai, 2007; Aron & Poldrack, 2006; Li, Huang, Constable, & Sinha, 2006; Rubia, Smith, Brammer, & Taylor, 2003) and validates our ROI selection.

Figure 5. 

Effect of rIFC and SMC stimulation on activation during reactive inhibition (successful stop vs. go0%) in the seven ROIs. (A) Mean effect of rIFC stimulation (red) and SMC stimulation (blue), relative to sham stimulation (dashed line), on activation during reactive inhibition. (B) Relationship between the neural effect of stimulation (vertical axis) and the behavioral effect of stimulation (horizontal axis) during reactive inhibition across participants. Positive (or negative) values on the vertical axis indicate increased (or decreased) activation or increased (or decreased) deactivation after real stimulation versus sham stimulation. Positive (or negative) values on the horizontal axis represent longer (or shorter) SSRTs after real stimulation versus sham stimulation. The left datum above each panel indicates the p value of the simple effects analysis testing the common effect of rIFC and SMC stimulation (i.e., ΔSSRT) on activation, the right datum gives the p values of the simple effects analysis testing the differential effect of rIFC and SMC stimulation (i.e., Stimulation Site × ΔSSRT) on activation. Error bars and bands represent 95% confidence intervals and bands.

Figure 5. 

Effect of rIFC and SMC stimulation on activation during reactive inhibition (successful stop vs. go0%) in the seven ROIs. (A) Mean effect of rIFC stimulation (red) and SMC stimulation (blue), relative to sham stimulation (dashed line), on activation during reactive inhibition. (B) Relationship between the neural effect of stimulation (vertical axis) and the behavioral effect of stimulation (horizontal axis) during reactive inhibition across participants. Positive (or negative) values on the vertical axis indicate increased (or decreased) activation or increased (or decreased) deactivation after real stimulation versus sham stimulation. Positive (or negative) values on the horizontal axis represent longer (or shorter) SSRTs after real stimulation versus sham stimulation. The left datum above each panel indicates the p value of the simple effects analysis testing the common effect of rIFC and SMC stimulation (i.e., ΔSSRT) on activation, the right datum gives the p values of the simple effects analysis testing the differential effect of rIFC and SMC stimulation (i.e., Stimulation Site × ΔSSRT) on activation. Error bars and bands represent 95% confidence intervals and bands.

After frontal stimulation, mean activation in the inhibitory control network as a whole did not differ from baseline (intercept, F(1, 19) < 1, p = .75) and there was no difference between rIFC and SMC stimulation on overall network activation (Stimulation Site, F(1, 19) < 1, p = .70) or activation in any of the individual regions (ROI × Stimulation Site, F(3.1, 58.9) < 1, p = .46). In contrast, as shown in Figure 5B, there was a significant relation between the behavioral and neural effect of stimulation across participants (ΔSSRT × ROI, F(3.2, 60.3) = 4.45, p = .006), suggesting coupling of stimulation-induced changes in SSRT and activation levels across individuals in some of the ROIs. Post hoc tests revealed that participants who became faster in reactive inhibition after rIFC or SMC stimulation showed reduced SMC activation (p = .013), enhanced right striatum activation (p < .001), and enhanced left M1 deactivation (p = .041). Thus, frontal stimulation altered activation of cortical and subcortical brain regions and these neural effects were proportional to the stimulation-induced changes in SSRT. In addition, there were regions in which the relation between the behavioral and neural effect of stimulation across participants also differed between rIFC and SMC stimulation (ΔSSRT × ROI × Stimulation Site, F(3.1, 58.9) = 3.20, p = .028), providing information about the relative position of these regions in the inhibitory control network. Post hoc tests showed that rIFC stimulation had a significantly greater reduction of rIFC activation than SMC stimulation had in participants who became faster in stopping (p = .015, first panel in Figure 5B). Note that the regression line capturing the effect of SMC stimulation on rIFC activation was flat (p > .05) and that we already showed a common effect of rIFC and SMC stimulation on SMC activation.2 Thus, rIFC stimulation altered SMC activation, but SMC stimulation did not significantly alter rIFC activation, suggesting that the rIFC exerts control over M1 via the SMC.

Next, we investigated effects of rIFC and SMC stimulation on activation outside the predefined ROIs. Figure 4 (red and blue areas) shows that reactive inhibition elicited activation and deactivation in a widespread network of brain regions commonly reported in stop-signal tasks at baseline (Boehler et al., 2010; Zandbelt & Vink, 2010; Aron & Poldrack, 2006; Li et al., 2006), including activation in a mainly right-lateralized network of fronto-parietal and subcortical regions and deactivation in left sensorimotor and premotor areas (among other regions).1 After frontal stimulation (Figure 6; Table 2), participants who became faster in reactive inhibition after rIFC or SMC stimulation showed enhanced deactivation not only of the left M1 but also of the right M1. In addition, we observed reduced activation of the SMC and the left and right superior parietal cortex (among other regions) and a trend toward stronger activation of the right striatum. Furthermore, rIFC stimulation compared with SMC stimulation resulted in reduced left temporo-parietal cortex activation and slightly reduced rIFC activation in participants who became faster in stopping. In summary, rIFC and SMC stimulation modulated activation levels at the stimulation sites and in a variety of remote regions and these effects were proportional to the effects of stimulation on task performance.

Figure 6. 

Whole-brain voxel-wise analysis of the effect of rIFC and SMC stimulation on activation during reactive inhibition. Statistical parametric maps show regions in which the regression of the neural effect of rTMS on the behavioral effect of rTMS was significant. All activations are significant (p < .05, family-wise error corrected) at the cluster level, unless stated otherwise in Table 1. Clusters that were used for the ROI analysis are shown in green.

Figure 6. 

Whole-brain voxel-wise analysis of the effect of rIFC and SMC stimulation on activation during reactive inhibition. Statistical parametric maps show regions in which the regression of the neural effect of rTMS on the behavioral effect of rTMS was significant. All activations are significant (p < .05, family-wise error corrected) at the cluster level, unless stated otherwise in Table 1. Clusters that were used for the ROI analysis are shown in green.

Table 2. 

Brain Regions Showing an Effect of rIFC or SMC Stimulation on Activation Related to Reactive Inhibition


p (FWE corr.)
k (voxels)
x (mm)
y (mm)
z (mm)
Regression of Activation on SSRT: Real rTMS < Sham rTMS (Successful Stop vs. Go0%) 
Left M1 .002 75 −44 −4 28 
−48 −8 32 
−56 28 
−32 −20 32 
Left superior parietal cortex .001 77 −24 −44 64 
−20 −32 48 
−16 −40 72 
−12 −28 60 
Right superior and middle temporal cortex .015 51 60 −16 
60 
52 −32 
52 −24 
44 12 −8 
52 −4 
48 16 −24 
Left insula and IFC .017 50 −48 −8 
−40 
−44 16 20 
−36 12 12 
SMC < .001 100 −4 68 
68 
20 −16 64 
12 −24 40 
−16 68 
−24 52 
−4 −16 68 
32 −12 60 
−4 −4 56 
−16 60 
−28 56 
Right M1 .042 41 56 40 
56 −8 28 
48 −4 32 
36 32 
40 −12 40 
40 −8 48 
Brainstem, parahippocampal cortex, and cerebellum .009 56 −40 −8 
−4 −40 −8 
24 −36 
16 −48 −8 
−4 −48 
12 −36 
16 −40 −12 
−52 −8 
12 −44 
Right superior parietal cortex .023 47 12 −48 68 
16 −52 64 
28 −40 68 
40 −40 60 
16 −40 72 
24 −48 60 
20 −56 68 
 
Regression of Activation on SSRT: Real rTMS > Sham rTMS (Successful Stop vs. Go0%) 
Right striatum .139 30 24 −4 
 
Regression of Activation on SSRT: rTMS rIFC < rTMS SMC (Successful Stop vs. Go0%) 
Left TPJ .011 54 −52 −44 24 
−60 −32 24 
rIFC cortex .090 34 56 16 
48 24 
 
Regression of Activation on SSRT: rTMS rIFC < rTMS SMC (Successful Stop vs. Go17–33%) 
rIFC cortex .013 54 48 24 
56 12 −12 
56 16 −4 
Left IFC .078 36 −52 12 −12 
−60 

p (FWE corr.)
k (voxels)
x (mm)
y (mm)
z (mm)
Regression of Activation on SSRT: Real rTMS < Sham rTMS (Successful Stop vs. Go0%) 
Left M1 .002 75 −44 −4 28 
−48 −8 32 
−56 28 
−32 −20 32 
Left superior parietal cortex .001 77 −24 −44 64 
−20 −32 48 
−16 −40 72 
−12 −28 60 
Right superior and middle temporal cortex .015 51 60 −16 
60 
52 −32 
52 −24 
44 12 −8 
52 −4 
48 16 −24 
Left insula and IFC .017 50 −48 −8 
−40 
−44 16 20 
−36 12 12 
SMC < .001 100 −4 68 
68 
20 −16 64 
12 −24 40 
−16 68 
−24 52 
−4 −16 68 
32 −12 60 
−4 −4 56 
−16 60 
−28 56 
Right M1 .042 41 56 40 
56 −8 28 
48 −4 32 
36 32 
40 −12 40 
40 −8 48 
Brainstem, parahippocampal cortex, and cerebellum .009 56 −40 −8 
−4 −40 −8 
24 −36 
16 −48 −8 
−4 −48 
12 −36 
16 −40 −12 
−52 −8 
12 −44 
Right superior parietal cortex .023 47 12 −48 68 
16 −52 64 
28 −40 68 
40 −40 60 
16 −40 72 
24 −48 60 
20 −56 68 
 
Regression of Activation on SSRT: Real rTMS > Sham rTMS (Successful Stop vs. Go0%) 
Right striatum .139 30 24 −4 
 
Regression of Activation on SSRT: rTMS rIFC < rTMS SMC (Successful Stop vs. Go0%) 
Left TPJ .011 54 −52 −44 24 
−60 −32 24 
rIFC cortex .090 34 56 16 
48 24 
 
Regression of Activation on SSRT: rTMS rIFC < rTMS SMC (Successful Stop vs. Go17–33%) 
rIFC cortex .013 54 48 24 
56 12 −12 
56 16 −4 
Left IFC .078 36 −52 12 −12 
−60 

The table shows all local maxima per cluster > 6 mm apart. k = cluster extent in voxels.

Above, we contrasted activation on stop trials with activation on go trials in the 0% stop-signal probability condition (stop vs. go0%). Alternatively, one could compare stop trial activation with activation on go trials in which stop-signal probability was greater than 0% (stop vs. go17–33%), as is more common in fMRI studies of the stop-signal task. We repeated the ROI and whole-brain voxel-wise analyses with this alternative contrast and found almost identical results.

As shown in Figure 7, the ROI analysis revealed that after frontal stimulation, mean activation in the inhibitory control network as a whole did not differ from baseline (intercept, F(1, 19) < 1, p = .73) and that there was also no difference between rIFC and SMC stimulation on overall network activation (Stimulation Site, F(1, 19) < 1, p = .77) or activation in any of the individual regions (ROI × Stimulation Site, F(2.9, 54.4) < 1, p = .45). However, in some ROIs the behavioral and neural effect of stimulation were correlated across participants (ΔSSRT × ROI, F(2.9, 54.8) = 2.89, p = .046); post hoc tests revealed that participants who became faster in reactive inhibition after rIFC or SMC stimulation showed enhanced right striatum activation (p = .004) and slightly reduced activation in the SMC (p = .08). There were also regions in which the relation between the behavioral and neural effect of stimulation across participants differed between rIFC and SMC stimulation (ΔSSRT × ROI × Stimulation Site, F(2.9, 54.4) = 4.34, p = .009); post hoc tests showed that rIFC stimulation compared with SMC stimulation had a significantly greater reduction of rIFC activation (p = .003) in participants who became faster in stopping.

Figure 7. 

Effect of rIFC and SMC stimulation on activation during reactive inhibition (successful stop vs. go17–33%) in the seven ROIs. (A) Mean effect of rIFC stimulation (red) and SMC stimulation (blue), relative to sham stimulation (dashed line), on activation during reactive inhibition. (B) Relationship between the neural effect of stimulation (vertical axis) and the behavioral effect of stimulation (horizontal axis) during reactive inhibition across participants. Positive (or negative) values on the vertical axis indicate increased (or decreased) activation or increased (or decreased) deactivation after real stimulation versus sham stimulation. Positive (or negative) values on the horizontal axis represent longer (or shorter) SSRTs after real stimulation versus sham stimulation. The left datum above each panel indicates the p value of the simple effects analysis testing the common effect of rIFC and SMC stimulation (i.e., ΔSSRT) on activation; the right datum gives the p values of the simple effects analysis testing the differential effect of rIFC and SMC stimulation (i.e., stimulation site × ΔSSRT) on activation. Error bars and bands represent 95% confidence intervals and bands.

Figure 7. 

Effect of rIFC and SMC stimulation on activation during reactive inhibition (successful stop vs. go17–33%) in the seven ROIs. (A) Mean effect of rIFC stimulation (red) and SMC stimulation (blue), relative to sham stimulation (dashed line), on activation during reactive inhibition. (B) Relationship between the neural effect of stimulation (vertical axis) and the behavioral effect of stimulation (horizontal axis) during reactive inhibition across participants. Positive (or negative) values on the vertical axis indicate increased (or decreased) activation or increased (or decreased) deactivation after real stimulation versus sham stimulation. Positive (or negative) values on the horizontal axis represent longer (or shorter) SSRTs after real stimulation versus sham stimulation. The left datum above each panel indicates the p value of the simple effects analysis testing the common effect of rIFC and SMC stimulation (i.e., ΔSSRT) on activation; the right datum gives the p values of the simple effects analysis testing the differential effect of rIFC and SMC stimulation (i.e., stimulation site × ΔSSRT) on activation. Error bars and bands represent 95% confidence intervals and bands.

The whole-brain voxel-wise analysis showed that, at baseline, the activation pattern of the stop versus go17–33% contrast was less widespread but qualitatively similar to the stop versus go0% contrast (Figure 4, magenta and cyan areas). In addition, rIFC stimulation compared with SMC stimulation resulted in reduced rIFC activation and slightly reduced left IFC activation in participants who became faster in stopping (Figure 6; Table 2). Taken together, the findings from the stop versus go17–33% contrast echo the findings from the stop versus go0% contrast, showing that rIFC and SMC stimulation modulated activation levels at the stimulation sites and in a variety of remote regions.

DISCUSSION

The functional organization of the neural network of inhibitory control is intensely debated but remains unclear. This is the first study that combined TMS and fMRI to investigate the neural underpinnings of inhibitory control. Using a modified stop-signal task, we addressed three outstanding issues regarding the functional organization of this network: the relative position of the rIFC and the SMC with respect to each other, the routes by which these structures exert inhibitory control over the primary motor cortex (M1), and the degree to which these regions are involved in proactive and reactive inhibition. Our findings indicate that the rIFC lies upstream from the SMC, that rIFC and SMC modulate the primary motor cortex (M1) via cortico-BG pathways through the right striatum and that rIFC and SMC are crucial for reactive inhibition but that their involvement in proactive inhibition remains unclear.

Interestingly, stimulation of the rIFC and the SMC improved rather than impaired reactive inhibition, as evidenced by a shorter SSRT. Importantly, this improvement was associated with increased M1 deactivation. Because rTMS affected SSRT but not go RT and because reactive inhibition involves the suppression of activity in M1 (van den Wildenberg et al., 2010; Coxon, Stinear, & Byblow, 2007; Coxon et al., 2006), these findings suggest that rIFC and SMC stimulation improved reactive inhibition by suppressing activity in M1 more rapidly, resulting in greater M1 deactivation. These findings enable us to investigate the hierarchical position of the rIFC and SMC and the pathways from these regions to M1.

Hierarchical Position of rIFC and SMC in the Reactive Inhibition Network

rIFC and SMC stimulation had local effects on reactive inhibition-related activation: rIFC stimulation influenced rIFC activation and SMC stimulation influenced SMC activation. More importantly, we found that rIFC stimulation also altered SMC activation, whereas SMC stimulation did not significantly alter rIFC activation. These findings provide causal evidence for the hypothesis that the rIFC exerts reactive inhibition over M1 via the SMC and reject the hypothesis that SMC controls M1 via the rIFC. These findings concur with functional connectivity studies of inhibitory control that suggest an effect of rIFC on SMC rather than in the reverse direction (Hwang et al., 2010; Duann et al., 2009). More generally, the current findings are congruent with models of the functional organization of executive functioning in the frontal lobe, proposing that control is exerted in an anterior–posterior direction, rather than from posterior to anterior (Badre & D'Esposito, 2009; Koechlin & Summerfield, 2007).

Our findings seem at odds with results from studies investigating the interactions between rIFC, SMC, and M1 with paired-pulse TMS and a task switching paradigm (Neubert et al., 2010; Mars et al., 2009), a task that bears resemblance to the stop-signal task. These studies reported a facilitatory effect of SMC stimulation on M1 that occurred earlier (125 msec after the switch signal) than an inhibitory effect of rIFC stimulation on M1 (175 msec after switch signal), which has been taken as evidence for the view that the SMC exerts inhibitory control over M1 via the rIFC (Aron, 2011; Neubert et al., 2010). Although this interpretation is plausible for the task-switching paradigm, we argue that it may not explain findings obtained with the stop-signal task. First, the early SMC influence on M1 was facilitatory rather than inhibitory. In the task-switching paradigm, this probably reflects facilitation of the correct switch response. Yet, in the stop-signal task facilitation of the contralateral M1 does not make sense mechanistically, as responses have to be stopped, not switched. Second, the facilitatory effect of SMC stimulation in the task-switching paradigm occurred much earlier than the modulation of activity in SMC (and rIFC) in the stop-signal task that typically occurs between 175 and 250 msec after the stop-signal (Chen, Scangos, & Stuphorn, 2010; Swann et al., 2009). In fact, because Neubert et al. (2010) and Mars et al. (2009) measured effects of SMC and rIFC stimulation on M1 up to 175 msec after the switch cue, they could have failed to detect an inhibitory influence of SMC on M1 occurring later than 175 msec after the stop-signal. Third, Neubert et al. (2010) found that rTMS-induced inactivation of the SMC breaks down the interaction between the rIFC and M1. This result is in excellent agreement with the present findings and supports the view that rIFC exerts inhibitory control over M1 via the SMC. Thus, our data support the idea that, at least in the context of the stop-signal task, M1 activation is controlled by the rIFC via the SMC. Future paired-pulse TMS studies should further assess this idea, testing for an inhibitory influence of SMC on M1 occurring in the 175–250 msec interval following the switch- or stop-signal.

As an aside, we noticed differences between rIFC and SMC stimulation not only in the rIFC, but also in the left TPJ. Although we did not have a hypothesis about activation changes in this region, it is intriguing that improvements in inhibitory control induced by training are also associated with a reduced left TPJ response to no-go stimuli (Manuel, Grivel, Bernasconi, Murray, & Spierer, 2010). Because the rIFC and the left TPJ have both been implicated in detecting changes in sensory stimuli (Downar, Crawley, Mikulis, & Davis, 2000), we speculate that rIFC stimulation made stop-signal detection more efficient through interactions with left TPJ.

Pathways from rIFC and SMC to M1

In addition to providing insight into the relative position of rIFC and SMC with respect to each other, our findings shed light on the pathways through which the rIFC and the SMC exert inhibitory control over M1. We showed that rIFC and SMC stimulation not only increased deactivation of M1 but also increased activation of the right striatum. These findings directly support the hypothesis that reactive inhibition relies on cortico-BG pathways, in particular, pathways through the right striatum. Two aspects are worth noting here. First, we showed effects in the right rather than the left striatum. This may reflect that reactive inhibition involves processes that depend on the right hemisphere, such as detection of salient stimuli (Corbetta, Patel, & Shulman, 2008; Corbetta & Shulman, 2002) and updating of action plans (Mars, Piekema, Coles, Hulstijn, & Toni, 2007). The rIFC has a prominent role in these processes (Verbruggen et al., 2010), so the right striatum may mediate input from the right IFC, either directly or through connections with the SMC. Second, we found effects of stimulation in the striatum rather than in the STN. This is important, because the dominant view is that reactive inhibition acts via cortico-BG pathways through the STN (Neubert et al., 2010; Isoda & Hikosaka, 2008; Aron et al., 2007; Aron & Poldrack, 2006). However, it has now been recognized that pathways through the striatum are equally relevant to reactive inhibition (Jahfari et al., 2011; Zandbelt & Vink, 2010; Chambers et al., 2009). In particular, there is a tight correlation across trials between striatal activation and M1 deactivation during stopping (Zandbelt & Vink, 2010), individuals with short compared with long SSRTs have stronger striatal activation (Chao, Luo, Chang, & Li, 2009), and stimulation of striatal neurons improves inhibitory control (Watanabe & Munoz, 2010). We failed to detect an effect of rIFC and SMC stimulation on STN activation. These findings do not necessarily preclude a role for the STN in reactive inhibition, however. For example, deep brain stimulation of the STN in Parkinson's patients improves reactive inhibition by shortening the SSRT (Mirabella et al., 2011; van den Wildenberg et al., 2006). Future studies should investigate the relative contribution of the striatum and STN to reactive inhibition. In summary, rIFC and SMC exert reactive inhibitory control over M1 via cortico-BG pathways through the right striatum.

It is interesting that stimulation-induced improvement of reactive inhibition not only increased deactivation of the left (contralateral) M1, but also of the right (ipsilateral) M1, although only the right hand was used to perform the stop-signal anticipation task. Increased ipsilateral deactivation in association with improved reactive inhibition of the contralateral hand could reflect transcallosal interactions between left and right M1. However, transcallosal interactions are typically mutually inhibitory (Grefkes, Eickhoff, Nowak, Dafotakis, & Fink, 2008; Ferbert et al., 1992; Wassermann, Fuhr, Cohen, & Hallett, 1991) and therefore unlikely explain our finding. Alternatively, regions in the frontal BG network targeting M1 may provide common input to both left and right M1 in a nonselective, global manner. This appears in line with Aron's hypothesis that reactive inhibition has global effects on the motor system (Aron, 2011), except that Aron appears to suggest that global inhibition affects multiple effectors within the same hemisphere, whereas we found modulations of focal M1 regions in both hemispheres. Thus, rIFC and SMC appear to exert reactive inhibitory control over M1 in a nonselective manner. Future studies could test this hypothesis by assessing corticospinal excitability of the ipsilateral M1 during the stop-signal anticipation task.

Role of the rIFC and SMC in Proactive Inhibition

Despite profound effects on reactive inhibition, we did not detect significant effects of rIFC and SMC stimulation on proactive inhibition. We hypothesized that proactive inhibition would not be altered by rIFC stimulation, but we were surprised to find no effect of SMC stimulation either. First, previous studies have found that SMC stimulation alters proactive inhibition (Stuphorn & Schall, 2006), whereas rIFC stimulation does not (Verbruggen et al., 2010). Second, the SMC, but not the rIFC, activates while preparing for an upcoming stop-signal (Zandbelt et al., in press; Swann et al., 2012). Third, dorsal frontal regions, such as the SMC, have been implicated in goal-directed processes (e.g., proactive inhibition), whereas ventrolateral frontal regions, such as the rIFC, are thought to be involved in stimulus-driven processes (e.g., which reactive inhibition; Corbetta et al., 2008; Corbetta & Shulman, 2002). Then why did SMC stimulation influence reactive inhibition but not proactive inhibition? Because the efficacy of TMS decreases as a function of coil distance (Ruohonen & Ilmoniemi, 2002), we hypothesize that proactive and reactive inhibition depend on different subregions within the SMC that differ in position relative to the TMS coil. Specifically, reactive inhibition may rely on superior regions of the SMC (where stimulation was likely effective), whereas proactive inhibition may depend on areas deeper in the medial wall (where the effects of TMS, if any, were probably much weaker). This is supported by our baseline whole-brain voxel-wise imaging results demonstrating that proactive control recruits areas deeper in the medial wall than reactive control. It is also possible that the discrepancy between our findings and the results from Stuphorn and Schall (2006), who did find altered proactive inhibition after SMC stimulation, are because of differences in species (human vs. nonhuman), effector (hand vs. eye), and perturbation technique (rTMS vs. intracortical microstimulation).

Behavioral and Physiological Effects of Brain Stimulation

It is worth discussing the rTMS effects in terms of direction, consistency across individuals, and putative mechanism. First, we observed that off-line low-frequency stimulation improved task performance. This might initially seem counterintuitive, given the widely used rule of thumb that off-line low-frequency stimulation suppresses cortical activity, inducing “virtual lesions.” However, this principle is mainly derived from basic studies of motor cortex and does not necessarily apply to other brain regions and more complex cognitive functions (e.g., Snyder, 2009; Andoh et al., 2006, 2008; Hilgetag, Théoret, & Pascual-Leone, 2001). Moreover, the line of reasoning that a stimulation-induced decrease in cortical excitability causes a decline of task performance is not secure (Sandrini, Umiltà, & Rusconi, 2011). Indeed, off-line stimulation of the rIFC and the SMC in the context of the stop-signal task has resulted in impairments (Verbruggen et al., 2010; Chen et al., 2009; Chambers et al., 2006, 2007) but also improvements (Hsu et al., 2011; Jacobson, Javitt, & Lavidor, 2011). This variation in outcome demonstrates that, despite the acute perturbation of neuronal processing, brain stimulation can both inhibit and facilitate task performance, depending on the stimulated region, stimulation parameters, and the nature of the task in a manner that is not well understood (Siebner et al., 2009).

Second, the rTMS effect varied substantially between individuals. This interindividual variability in off-line rTMS has long been recognized (Daskalakis et al., 2006; Romero et al., 2002; Sommer et al., 2002; Maeda et al., 2000) and could arise from interindividual differences in neuroanatomy, such as functional organization of the brain, white matter characteristics (Kloppel et al., 2008), and scalp-to-brain distance (Stokes et al., 2005). On top of that, differences in task strategies could play a role, as task strategies may differ in their reliance on the stimulated region.

Third, stimulation-induced improvements in task performance were associated with a reduction in BOLD activation in the stimulated regions and several remote areas. This reduction in activation might seem contrary to what one would expect, but similar findings have been reported before (e.g., Ward et al., 2010). What does this rTMS-induced reduction in BOLD response reflect and how can it be reconciled with enhanced task performance? Two mechanisms might be at play. One is that rTMS amplifies background activity (i.e., introduces random neural noise; Walsh & Cowey, 2000), thereby decreasing the BOLD contrast between stop and go trials. Note that increased noise may not only interfere with but can also facilitate performance (Miniussi, Ruzzoli, & Walsh, 2010; Stein, Gossen, & Jones, 2005; Moss, Ward, & Sannita, 2004). Another is that off-line rTMS modulates plasticity (Hoogendam, Ramakers, & Di Lazzaro, 2010), rendering neural processing in the stimulated and connected regions more efficient. As a result, less activity may be needed to perform a cognitive process, effectively reducing the BOLD response measured. Future studies are needed to elucidate the effects of off-line rTMS on neuronal activity.

Stop-signal Anticipation Task

The present findings were obtained with the stop-signal anticipation task (Zandbelt & Vink, 2010), a paradigm based on the Slater–Hammel stop-signal task (Slater-Hammel, 1960). In this task, stop-signals occur occasionally before an anticipatory response is made. This differs somewhat from the standard stop-signal task (Verbruggen & Logan, 2008; Lappin & Eriksen, 1966), in which stop-signals occur infrequently after an imperative go-signal is presented. In other words, in the stop-signal anticipation task, participants stop the initiation of a prepared movement, whereas in the standard stop-signal task participants stop the initiation of a signaled response. This difference appears largely semantic, however, because commonalities between Slater–Hammel and standard stop-signal tasks appear to dominate: The race model applies to both types of paradigms (Zandbelt & Vink, 2010; Logan & Cowan, 1984), and striking similarities have been observed in terms of the degree of proactive slowing (Zandbelt & Vink, 2010; Vink et al., 2005), fMRI activation patterns (Zandbelt & Vink, 2010; Aron & Poldrack, 2006), and corticospinal excitability profiles (van den Wildenberg et al., 2010; Coxon et al., 2006). Accordingly, the present findings are probably not confined to the stop-signal anticipation task but extend to other inhibitory control paradigms, such as the standard stop-signal task.

fMRI Contrasts of Reactive Inhibition

There is an ongoing debate in the fMRI literature about which stop-signal task contrast is most suitable for assessing reactive inhibition. We therefore quantified activation during reactive inhibition by means of two contrasts: successful stop trials versus go trials in contexts in which stop-signals never occurred (successful stop vs. go0%) and successful stop trials versus go trials in contexts in which stop-signals did occur (successful stop vs. go17-33%). We demonstrated that the results from both contrasts are largely in agreement. Below we will describe the differences between them and discuss the specificity of these contrasts.

Successful stop versus go17–33% is the most commonly applied contrast in the fMRI literature of the stop-signal task and has the advantage that the same go trials are used for SSRT estimation and fMRI analysis. At first, it also seems better at isolating activation related to reactive inhibition than successful stop versus go0%, assuming that stop trials involve both proactive and reactive processes, go17–33% trials involve proactive processes only, and go0% trials involve neither of them. However, recently, we showed that this assumption is incorrect; activation on go17–33% trials reflects proactive inhibition only in some activated regions (striatum and the SMC), whereas in others (rIFC) it more likely represents expectancy violation (Zandbelt et al., in press). This makes interpretation of the successful stop versus go17–33% contrast problematic. Successful stop versus go0%, on the other hand, is easier to interpret because go0% trials provide a cleaner baseline in which proactive and reactive inhibition are not at play. Other advantages are that performance on go0% trials is more homogeneous across subjects than on go17–33% trials and that successful stop versus go0% has the best detection power. This is important given that off-line rTMS effects are subtle.

Although go0% trials provide adequate detection power and a clean baseline, activation in the successful stop versus go0% contrast could include effects of proactive inhibition and other processes that differ between successful stop and go0% trials. If the present rTMS effects on activation during reactive inhibition would reflect changes in proactive inhibition, then proactive inhibition should have been influenced by rTMS. Yet, neither at the behavioral level nor at the neural level did we detect such effects. It is also unlikely that the present rTMS effects on activation during reactive inhibition reflect changes in cognitive processes other than reactive inhibition, because all neural rTMS effects were proportional to the rTMS effects on SSRT. This suggests that the reported effects of rIFC and SMC stimulation were specific to reactive inhibition.

Conclusion

This is the first study combining TMS and fMRI to investigate inhibitory control in the stop-signal task. We show that rIFC lies upstream from the SMC in the network subserving reactive inhibition and that rIFC and SMC modulate the primary motor cortex (M1) via cortico-BG pathways through the right striatum.

Acknowledgments

This work was supported by the Netherlands Organization for Scientific Research (Veni grant to M. V.).

Reprint requests should be sent to Bram B. Zandbelt, Department of Psychology, Vanderbilt University, PMB 407817, 2301 Vanderbilt Place, Nashville, TN 37240-7817, United States, or via e-mail: bramzandbelt@gmail.com.

Notes

1. 

A table with local maxima of this contrast is available on-line at www.bramzandbelt.com.

2. 

A post hoc test revealed that rIFC stimulation also had a slightly stronger effect on SMC activation than SMC stimulation did (p = .053, second panel in Figure 5B).

REFERENCES

REFERENCES
Andoh
,
J.
,
Artiges
,
E.
,
Pallier
,
C.
,
Riviere
,
D.
,
Mangin
,
J. F.
,
Cachia
,
A.
,
et al
(
2006
).
Modulation of language areas with functional MR image-guided magnetic stimulation.
Neuroimage
,
29
,
619
627
.
Andoh
,
J.
,
Artiges
,
E.
,
Pallier
,
C.
,
Rivière
,
D.
,
Mangin
,
J.-F.
,
Paillère-Martinot
,
M.-L.
,
et al
(
2008
).
Priming frequencies of transcranial magnetic stimulation over Wernicke's area modulate word detection.
Cerebral Cortex
,
18
,
210
216
.
Aron
,
A. R.
(
2011
).
From reactive to proactive and selective control: Developing a richer model for stopping inappropriate responses.
Biological Psychiatry
,
69
,
e55
e68
.
Aron
,
A. R.
,
Behrens
,
T. E.
,
Smith
,
S.
,
Frank
,
M. J.
, &
Poldrack
,
R. A.
(
2007
).
Triangulating a cognitive control network using diffusion-weighted magnetic resonance imaging (MRI) and functional MRI.
Journal of Neuroscience
,
27
,
3743
3752
.
Aron
,
A. R.
, &
Poldrack
,
R. A.
(
2006
).
Cortical and subcortical contributions to stop signal response inhibition: Role of the subthalamic nucleus.
Journal of Neuroscience
,
26
,
2424
2433
.
Badre
,
D.
, &
D'Esposito
,
M.
(
2009
).
Is the rostro-caudal axis of the frontal lobe hierarchical?
Nature Reviews Neuroscience
,
10
,
659
669
.
Boehler
,
C. N.
,
Appelbaum
,
L. G.
,
Krebs
,
R. M.
,
Hopf
,
J. M.
, &
Woldorff
,
M. G.
(
2010
).
Pinning down response inhibition in the brain—Conjunction analyses of the stop-signal task.
Neuroimage
,
52
,
1621
1632
.
Braver
,
T. S.
,
Gray
,
J. R.
, &
Burgess
,
G. C.
(
2007
).
Explaining the many varieties of working memory variation: Dual mechanisms of cognitive control.
In A. Conway, C. Jarrold, M. Kane, A. Miyake, & K. Towse (Eds.)
,
Variation in working memory
(pp.
76
106
).
Oxford
:
Oxford University Press
.
Buch
,
E. R.
,
Mars
,
R. B.
,
Boorman
,
E. D.
, &
Rushworth
,
M. F. S.
(
2010
).
A network centered on ventral premotor cortex exerts both facilitatory and inhibitory control over primary motor cortex during action reprogramming.
Journal of Neuroscience
,
30
,
1395
1401
.
Cai
,
W.
,
Oldenkamp
,
C. L.
, &
Aron
,
A. R.
(
2011
).
A proactive mechanism for selective suppression of response tendencies.
Journal of Neuroscience
,
31
,
5965
5969
.
Chambers
,
C. D.
,
Bellgrove
,
M. A.
,
Gould
,
I. C.
,
English
,
T.
,
Garavan
,
H.
,
McNaught
,
E.
,
et al
(
2007
).
Dissociable mechanisms of cognitive control in prefrontal and premotor cortex.
Journal of Neurophysiology
,
98
,
3638
3647
.
Chambers
,
C. D.
,
Bellgrove
,
M. A.
,
Stokes
,
M. G.
,
Henderson
,
T. R.
,
Garavan
,
H.
,
Robertson
,
I. H.
,
et al
(
2006
).
Executive “brake failure” following deactivation of human frontal lobe.
Journal of Cognitive Neuroscience
,
18
,
444
455
.
Chambers
,
C. D.
,
Garavan
,
H.
, &
Bellgrove
,
M. A.
(
2009
).
Insights into the neural basis of response inhibition from cognitive and clinical neuroscience.
Neuroscience and Biobehavioral Reviews
,
33
,
631
646
.
Chao
,
H. H. A.
,
Luo
,
X.
,
Chang
,
J. L. K.
, &
Li
,
C.-S. R.
(
2009
).
Activation of the pre-supplementary motor area but not inferior prefrontal cortex in association with short stop signal reaction time—An intra-subject analysis.
BMC Neurosci
,
10
,
75
.
Chen
,
C.-Y.
,
Muggleton
,
N. G.
,
Tzeng
,
O. J. L.
,
Hung
,
D. L.
, &
Juan
,
C.-H.
(
2009
).
Control of prepotent responses by the superior medial frontal cortex.
Neuroimage
,
44
,
537
545
.
Chen
,
X.
,
Scangos
,
K.
, &
Stuphorn
,
V.
(
2010
).
Supplementary motor area exerts proactive and reactive control of arm movements.
Journal of Neuroscience
,
30
,
14657
14675
.
Chikazoe
,
J.
,
Jimura
,
K.
,
Hirose
,
S.
,
Yamashita
,
K.-I.
,
Miyashita
,
Y.
, &
Konishi
,
S.
(
2009
).
Preparation to inhibit a response complements response inhibition during performance of a stop-signal task.
Journal of Neuroscience
,
29
,
15870
15877
.
Corbetta
,
M.
,
Patel
,
G.
, &
Shulman
,
G. L.
(
2008
).
The reorienting system of the human brain: From environment to theory of mind.
Neuron
,
58
,
306
324
.
Corbetta
,
M.
, &
Shulman
,
G. L.
(
2002
).
Control of goal-directed and stimulus-driven attention in the brain.
Nature Reviews Neuroscience
,
3
,
201
215
.
Coxon
,
J. P.
,
Stinear
,
C. M.
, &
Byblow
,
W. D.
(
2006
).
Intracortical inhibition during volitional inhibition of prepared action.
Journal of Neurophysiology
,
95
,
3371
3383
.
Coxon
,
J. P.
,
Stinear
,
C. M.
, &
Byblow
,
W. D.
(
2007
).
Selective inhibition of movement.
Journal of Neurophysiology
,
97
,
2480
2489
.
Daskalakis
,
Z. J.
,
Moller
,
B.
,
Christensen
,
B. K.
,
Fitzgerald
,
P. B.
,
Gunraj
,
C.
, &
Chen
,
R.
(
2006
).
The effects of repetitive transcranial magnetic stimulation on cortical inhibition in healthy human subjects.
Experimental Brain Research
,
174
,
403
412
.
Downar
,
J.
,
Crawley
,
A. P.
,
Mikulis
,
D. J.
, &
Davis
,
K. D.
(
2000
).
A multimodal cortical network for the detection of changes in the sensory environment.
Nature Neuroscience
,
3
,
277
.
Duann
,
J.-R.
,
Ide
,
J. S.
,
Luo
,
X.
, &
Li
,
C.-S. R.
(
2009
).
Functional connectivity delineates distinct roles of the inferior frontal cortex and presupplementary motor area in stop signal inhibition.
Journal of Neuroscience
,
29
,
10171
10179
.
Ferbert
,
A.
,
Priori
,
A.
,
Rothwell
,
J. C.
,
Day
,
B. L.
,
Colebatch
,
J. G.
, &
Marsden
,
C. D.
(
1992
).
Interhemispheric inhibition of the human motor cortex.
Journal of Physiology
,
453
,
525
546
.
Grefkes
,
C.
,
Eickhoff
,
S. B.
,
Nowak
,
D. A.
,
Dafotakis
,
M.
, &
Fink
,
G. R.
(
2008
).
Dynamic intra- and interhemispheric interactions during unilateral and bilateral hand movements assessed with fMRI and DCM.
Neuroimage
,
41
,
1382
1394
.
Hilgetag
,
C. C.
,
Théoret
,
H.
, &
Pascual-Leone
,
A.
(
2001
).
Enhanced visual spatial attention ipsilateral to rTMS-induced “virtual lesions” of human parietal cortex.
Nature Neuroscience
,
4
,
953
957
.
Hoogendam
,
J. M.
,
Ramakers
,
G. M. J.
, &
Di Lazzaro
,
V.
(
2010
).
Physiology of repetitive transcranial magnetic stimulation of the human brain.
Brain Stimulation
,
3
,
95
118
.
Hsu
,
T. Y.
,
Tseng
,
L. Y.
,
Yu
,
J. X.
,
Kuo
,
W. J.
,
Hung
,
D. L.
,
Tzeng
,
O. J.
,
et al
(
2011
).
Modulating inhibitory control with direct current stimulation of the superior medial frontal cortex.
Neuroimage
,
56
,
2249
2257
.
Hwang
,
K.
,
Velanova
,
K.
, &
Luna
,
B.
(
2010
).
Strengthening of top–down frontal cognitive control networks underlying the development of inhibitory control: A functional magnetic resonance imaging effective connectivity study.
Journal of Neuroscience
,
30
,
15535
15545
.
Isoda
,
M.
, &
Hikosaka
,
O.
(
2008
).
Role for subthalamic nucleus neurons in switching from automatic to controlled eye movement.
Journal of Neuroscience
,
28
,
7209
7218
.
Iyer
,
M. B.
,
Schleper
,
N.
, &
Wassermann
,
E. M.
(
2003
).
Priming stimulation enhances the depressant effect of low-frequency repetitive transcranial magnetic stimulation.
Journal of Neuroscience
,
23
,
10867
10872
.
Jacobson
,
L.
,
Javitt
,
D. C.
, &
Lavidor
,
M.
(
2011
).
Activation of inhibition: Diminishing impulsive behavior by direct current stimulation over the inferior frontal gyrus.
Journal of Cognitive Neuroscience
,
23
,
3380
3387
.
Jahfari
,
S.
,
Stinear
,
C.
,
Claffey
,
M.
,
Verbruggen
,
F.
, &
Aron
,
A.
(
2010
).
Responding with restraint: What are the neurocognitive mechanisms?
Journal of Cognitive Neuroscience
,
22
,
1479
1492
.
Jahfari
,
S.
,
Waldorp
,
L.
,
van den Wildenberg
,
W. P.
,
Scholte
,
H. S.
,
Ridderinkhof
,
K. R.
, &
Forstmann
,
B. U.
(
2011
).
Effective connectivity reveals important roles for both the hyperdirect (fronto-subthalamic) and the indirect (fronto-striatal-pallidal) fronto-basal ganglia pathways during response inhibition.
Journal of Neuroscience
,
31
,
6891
6899
.
Keel
,
J.
,
Smith
,
M.
, &
Wassermann
,
E.
(
2001
).
A safety screening questionnaire for transcranial magnetic stimulation.
Clinical Neurophysiology
,
112
,
720
.
Kloppel
,
S.
,
Baumer
,
T.
,
Kroeger
,
J.
,
Koch
,
M. A.
,
Buchel
,
C.
,
Munchau
,
A.
,
et al
(
2008
).
The cortical motor threshold reflects microstructural properties of cerebral white matter.
Neuroimage
,
40
,
1782
1791
.
Koechlin
,
E.
, &
Summerfield
,
C.
(
2007
).
An information theoretical approach to prefrontal executive function.
Trends in Cognitive Sciences
,
11
,
229
235
.
Lappin
,
J. S.
, &
Eriksen
,
C. W.
(
1966
).
Use of a delayed signal to stop a visual reaction-time response.
Journal of Experimental Psychology
,
72
,
805
811
.
Leung
,
H.-C.
, &
Cai
,
W.
(
2007
).
Common and differential ventrolateral prefrontal activity during inhibition of hand and eye movements.
Journal of Neuroscience
,
27
,
9893
9900
.
Li
,
C.-S. R.
,
Huang
,
C.
,
Constable
,
R. T.
, &
Sinha
,
R.
(
2006
).
Imaging response inhibition in a stop-signal task: Neural correlates independent of signal monitoring and post-response processing.
Journal of Neuroscience
,
26
,
186
192
.
Logan
,
G. D.
, &
Burkell
,
J.
(
1986
).
Dependence and independence in responding to double stimulation. A comparison of stop, change, and dual-task paradigms.
Journal of Experimental Psychology: Human Perception and Performance
,
12
,
549
563
.
Logan
,
G. D.
, &
Cowan
,
W. B.
(
1984
).
On the ability to inhibit thought and action: A theory of an act of control.
Psychological Review
,
91
,
291
327
.
Maeda
,
F.
,
Keenan
,
J. P.
,
Tormos
,
J. M.
,
Topka
,
H.
, &
Pascual-Leone
,
A.
(
2000
).
Interindividual variability of the modulatory effects of repetitive transcranial magnetic stimulation on cortical excitability.
Experimental Brain Research
,
133
,
425
430
.
Manuel
,
A.
,
Grivel
,
J.
,
Bernasconi
,
F.
,
Murray
,
M.
, &
Spierer
,
L.
(
2010
).
Brain dynamics underlying training-induced improvement in suppressing inappropriate action.
Journal of Neuroscience
,
30
,
13670
.
Mars
,
R. B.
,
Klein
,
M. C.
,
Neubert
,
F.-X.
,
Olivier
,
E.
,
Buch
,
E. R.
,
Boorman
,
E. D.
,
et al
(
2009
).
Short-latency influence of medial frontal cortex on primary motor cortex during action selection under conflict.
Journal of Neuroscience
,
29
,
6926
6931
.
Mars
,
R. B.
,
Piekema
,
C.
,
Coles
,
M. G. H.
,
Hulstijn
,
W.
, &
Toni
,
I.
(
2007
).
On the programming and reprogramming of actions.
Cerebral Cortex
,
17
,
2972
2979
.
Miniussi
,
C.
,
Ruzzoli
,
M.
, &
Walsh
,
V.
(
2010
).
The mechanism of transcranial magnetic stimulation in cognition.
Cortex
,
46
,
128
130
.
Mirabella
,
G.
,
Iaconelli
,
S.
,
Romanelli
,
P.
,
Modugno
,
N.
,
Lena
,
F.
,
Manfredi
,
M.
,
et al
(
2011
).
Deep brain stimulation of subthalamic nuclei affects arm response inhibition in Parkinson's patients.
Cerebral Cortex
,
22
,
1124
1132
.
Moss
,
F.
,
Ward
,
L. M.
, &
Sannita
,
W. G.
(
2004
).
Stochastic resonance and sensory information processing: A tutorial and review of application.
Clinical Neurophysiology
,
115
,
267
281
.
Neggers
,
S. F.
,
Langerak
,
T. R.
,
Schutter
,
D. J.
,
Mandl
,
R. C.
,
Ramsey
,
N. F.
,
Lemmens
,
P. J.
,
et al
(
2004
).
A stereotactic method for image-guided transcranial magnetic stimulation validated with fMRI and motor-evoked potentials.
Neuroimage
,
21
,
1805
1817
.
Neubert
,
F. X.
,
Mars
,
R. B.
,
Buch
,
E. R.
,
Olivier
,
E.
, &
Rushworth
,
M. F.
(
2010
).
Cortical and subcortical interactions during action reprogramming and their related white matter pathways.
Proceedings of the National Academy of Sciences, U.S.A.
,
107
,
13240
13245
.
Oldfield
,
R. C.
(
1971
).
The assessment and analysis of handedness: The Edinburgh inventory.
Neuropsychologia
,
9
,
97
113
.
Overall
,
J. E.
, &
Tonidandel
,
S.
(
2010
).
The case for use of simple difference scores to test the significance of differences in mean rates of change in controlled repeated measurements designs.
Multivariate Behavioral Research
,
45
,
806
827
.
Prodoehl
,
J.
,
Yu
,
H.
,
Little
,
D. M.
,
Abraham
,
I.
, &
Vaillancourt
,
D. E.
(
2008
).
Region of interest template for the human basal ganglia: Comparing EPI and standardized space approaches.
Neuroimage
,
39
,
956
965
.
Ramautar
,
J. R.
,
Kok
,
A.
, &
Ridderinkhof
,
K. R.
(
2004
).
Effects of stop-signal probability in the stop-signal paradigm: The N2/P3 complex further validated.
Brain and Cognition
,
56
,
234
252
.
Ramautar
,
J. R.
,
Slagter
,
H. A.
,
Kok
,
A.
, &
Ridderinkhof
,
K. R.
(
2006
).
Probability effects in the stop-signal paradigm: The insula and the significance of failed inhibition.
Brain Research
,
1105
,
143
154
.
Romero
,
J. R.
,
Anschel
,
D.
,
Sparing
,
R.
,
Gangitano
,
M.
, &
Pascual-Leone
,
A.
(
2002
).
Subthreshold low frequency repetitive transcranial magnetic stimulation selectively decreases facilitation in the motor cortex.
Clinical Neurophysiology
,
113
,
101
107
.
Rubia
,
K.
,
Smith
,
A. B.
,
Brammer
,
M. J.
, &
Taylor
,
E.
(
2003
).
Right inferior prefrontal cortex mediates response inhibition while mesial prefrontal cortex is responsible for error detection.
Neuroimage
,
20
,
351
358
.
Ruohonen
,
J.
, &
Ilmoniemi
,
R. J.
(
2002
).
Physical principles of transcranial magnetic stimulation.
In A. Pascual-Leone, N. J. Davey, J. C. Rothwell, E. M. Wassermann, & B. K. Puri (Eds.)
,
Handbook of transcranial magnetic stimulation
(pp.
18 
29
).
London
:
Hodder Arnold Publishers
.
Sandrini
,
M.
,
Umiltà
,
C.
, &
Rusconi
,
E.
(
2011
).
The use of transcranial magnetic stimulation in cognitive neuroscience: A new synthesis of methodological issues.
Neuroscience and Biobehavioral Reviews
,
35
,
516
536
.
Schutter
,
D.
, &
van Honk
,
J.
(
2006
).
A standardized motor threshold estimation procedure for transcranial magnetic stimulation research.
The Journal of ECT
,
22
,
176
178
.
Sharp
,
D. J.
,
Bonnelle
,
V.
,
De Boissezon
,
X.
,
Beckmann
,
C. F.
,
James
,
S. G.
,
Patel
,
M. C.
,
et al
(
2010
).
Distinct frontal systems for response inhibition, attentional capture, and error processing.
Proceedings of the National Academy of Sciences, U.S.A.
,
107
,
6106
6111
.
Sheehan
,
D. V.
,
Lecrubier
,
Y.
,
Sheehan
,
K. H.
,
Amorim
,
P.
,
Janavs
,
J.
,
Weiller
,
E.
,
et al
(
1998
).
The mini-international neuropsychiatric interview (M.I.N.I.): The development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10.
Journal of Clinical Psychiatry
,
59(Suppl. 20)
,
22
33
.
Siebner
,
H. R.
,
Bergmann
,
T. O.
,
Bestmann
,
S.
,
Massimini
,
M.
,
Johansen-Berg
,
H.
,
Mochizuki
,
H.
,
et al
(
2009
).
Consensus paper: Combining transcranial stimulation with neuroimaging.
Brain Stimulation
,
2
,
58
80
.
Slater-Hammel
,
A. T.
(
1960
).
Reliability, accuracy, and refractoriness of a transit reaction.
Research Quarterly
,
31
,
217
228
.
Snyder
,
A.
(
2009
).
Explaining and inducing savant skills: Privileged access to lower level, less-processed information.
Philosophical Transactions of the Royal Society of London, Series B, Biological Sciences
,
364
,
1399
1405
.
Sommer
,
M.
,
Wu
,
T.
,
Tergau
,
F.
, &
Paulus
,
W.
(
2002
).
Intra- and interindividual variability of motor responses to repetitive transcranial magnetic stimulation.
Clinical Neurophysiology
,
113
,
265
269
.
Stein
,
R. B.
,
Gossen
,
E. R.
, &
Jones
,
K. E.
(
2005
).
Neuronal variability: Noise or part of the signal?
Nature Reviews Neuroscience
,
6
,
389
397
.
Stinear
,
C. M.
,
Coxon
,
J. P.
, &
Byblow
,
W. D.
(
2009
).
Primary motor cortex and movement prevention: Where stop meets go.
Neuroscience and Biobehavioral Reviews
,
33
,
662
673
.
Stokes
,
M. G.
,
Chambers
,
C. D.
,
Gould
,
I. C.
,
Henderson
,
T. R.
,
Janko
,
N. E.
,
Allen
,
N. B.
,
et al
(
2005
).
Simple metric for scaling motor threshold based on scalp-cortex distance: Application to studies using transcranial magnetic stimulation.
Journal of Neurophysiology
,
94
,
4520
4527
.
Stuphorn
,
V.
, &
Schall
,
J. D.
(
2006
).
Executive control of countermanding saccades by the supplementary eye field.
Nature Neuroscience
,
9
,
925
931
.
Swann
,
N. C.
,
Cai
,
W.
,
Conner
,
C. R.
,
Pieters
,
T. A.
,
Claffey
,
M. P.
,
George
,
J. S.
,
et al
(
2012
).
Roles for the pre-supplementary motor area and the right inferior frontal gyrus in stopping action: Electrophysiological responses and functional and structural connectivity.
Neuroimage
,
59
,
2860
2870
.
Swann
,
N.
,
Tandon
,
N.
,
Canolty
,
R.
,
Ellmore
,
T. M.
,
McEvoy
,
L. K.
,
Dreyer
,
S.
,
et al
(
2009
).
Intracranial EEG reveals a time- and frequency-specific role for the right inferior frontal gyrus and primary motor cortex in stopping initiated responses.
Journal of Neuroscience
,
29
,
12675
12685
.
van den Wildenberg
,
W. P.
,
Burle
,
B.
,
Vidal
,
F.
,
van der Molen
,
M. W.
,
Ridderinkhof
,
K. R.
, &
Hasbroucq
,
T.
(
2010
).
Mechanisms and dynamics of cortical motor inhibition in the stop-signal paradigm: A TMS study.
Journal of Cognitive Neuroscience
,
22
,
225
239
.
van den Wildenberg
,
W. P.
,
van Boxtel
,
G. J.
,
van der Molen
,
M. W.
,
Bosch
,
D. A.
,
Speelman
,
J. D.
, &
Brunia
,
C. H.
(
2006
).
Stimulation of the subthalamic region facilitates the selection and inhibition of motor responses in Parkinson's disease.
Journal of Cognitive Neuroscience
,
18
,
626
636
.
Verbruggen
,
F.
,
Aron
,
A. R.
,
Stevens
,
M. A.
, &
Chambers
,
C. D.
(
2010
).
Theta burst stimulation dissociates attention and action updating in human inferior frontal cortex.
Proceedings of the National Academy of Sciences, U.S.A.
,
107
,
13966
13971
.
Verbruggen
,
F.
, &
Logan
,
G. D.
(
2008
).
Response inhibition in the stop-signal paradigm.
Trends in Cognitive Sciences
,
12
,
418
424
.
Verbruggen
,
F.
, &
Logan
,
G. D.
(
2009a
).
Models of response inhibition in the stop-signal and stop-change paradigms.
Neuroscience and Biobehavioral Reviews
,
33
,
647
661
.
Verbruggen
,
F.
, &
Logan
,
G. D.
(
2009b
).
Proactive adjustments of response strategies in the stop-signal paradigm.
Journal of Experimental Psychology: Human Perception and Performance
,
35
,
835
854
.
Verbruggen
,
F.
, &
Logan
,
G. D.
(
2009c
).
Proactive adjustments of response strategies in the stop-signal paradigm.
Journal of Experimental Psychology: Human Perception and Performance
,
35
,
835
854
.
Vink
,
M.
,
Kahn
,
R. S.
,
Raemaekers
,
M.
,
van den Heuvel
,
M.
,
Boersma
,
M.
, &
Ramsey
,
N. F.
(
2005
).
Function of striatum beyond inhibition and execution of motor responses.
Human Brain Mapping
,
25
,
336
344
.
Vink
,
M.
,
Ramsey
,
N. F.
,
Raemaekers
,
M.
, &
Kahn
,
R. S.
(
2006
).
Striatal dysfunction in schizophrenia and unaffected relatives.
Biological Psychiatry
,
60
,
32
39
.
Walsh
,
V.
, &
Cowey
,
A.
(
2000
).
Transcranial magnetic stimulation and cognitive neuroscience.
Nature Reviews Neuroscience
,
1
,
73
79
.
Ward
,
N.
,
Bestmann
,
S.
,
Hartwigsen
,
G.
,
Weiss
,
M.
,
Christensen
,
L.
,
Frackowiak
,
R.
,
et al
(
2010
).
Low-frequency transcranial magnetic stimulation over left dorsal premotor cortex improves the dynamic control of visuospatially cued actions.
Journal of Neuroscience
,
30
,
9216
.
Wassermann
,
E. M.
,
Fuhr
,
P.
,
Cohen
,
L. G.
, &
Hallett
,
M.
(
1991
).
Effects of transcranial magnetic stimulation on ipsilateral muscles.
Neurology
,
41
,
1795
1799
.
Watanabe
,
M.
, &
Munoz
,
D.
(
2010
).
Saccade suppression by electrical microstimulation in monkey caudate nucleus.
Journal of Neuroscience
,
30
,
2700
.
Zandbelt
,
B. B.
,
Bloemendaal
,
M.
,
Neggers
,
S. F.
,
Kahn
,
R. S.
, &
Vink
,
M.
(
in press
).
Expectations and violations: Delineating the neural network of proactive inhibitory control.
Human Brain Mapping
.
Zandbelt
,
B. B.
,
Hoogendam
,
J. M.
,
Van Buuren
,
M.
,
Sack
,
A. T.
,
Kahn
,
R. S.
, &
Vink
,
M.
(
2009
).
Repetitive transcranial magnetic stimulation alters resting-state functional connectivity.
Neuroimage
,
47(Suppl. 1)
,
S74
.
Zandbelt
,
B. B.
,
van Buuren
,
M.
,
Kahn
,
R. S.
, &
Vink
,
M.
(
2011
).
Reduced proactive inhibition in schizophrenia is related to corticostriatal dysfunction and poor working memory.
Biological Psychiatry
,
70
,
1151
1158
.
Zandbelt
,
B. B.
, &
Vink
,
M.
(
2010
).
On the role of the striatum in response inhibition.
PLoS One
,
11
,
e13848
.