Offline rTMS over right IFG or pre-SMA:
effects on response capture and selective inhibition in the Simon task?
Master’s thesis report (1st internship) Master of Brain and Cognitive Science
Cognitive Science Center Amsterdam of the University of Amsterdam
date: 1/9/2011
author
R. Kunert (6317529)
research
A.D. van Campen R. Kunert
W.P.M. van den Wildenberg K.R. Ridderinkhof
internship supervision
supervisor: K.R. Ridderinkhof co-assessor: A.D. van Campen UvA representative: L. van Maanen
ABSTRACT
In order to perform in a goal directed way in a complex environment, one needs to suppress incorrect response tendencies triggered by irrelevant stimuli. Using the Simon task as a model for these situations, direct response capture of response selection by irrelevant stimuli and subsequent selective inhibition of this potentially false response were elicited. We probed for the involvement of two key brain areas in this process. 15 minutes of offline 1Hz repetitive transcranial magnetic stimulation (rTMS) was applied over the presupplementary motor area (pre-SMA) in order to amplify response capture and over the right inferior frontal gyrus (IFG) to reduce selective inhibition. We used distributional analyses of behavioural data to distinguish these processes. Even though both brain areas are widely reported to be involved in response conflict resolution across tasks as well as in the Simon task specifically, the results are not in line with predictions. PreSMA stimulation was without a measurable effect. Right IFG stimulation increased the reaction time effect of the irrelevant stimulus location. However, this main effect of stimulation does not interact with response time, i.e. it was not specific to late responses which were predicted to benefit more from selective inhibition than early responses.
INTRODUCTION
Response Conflict Resolution
The present study aimed to reveal the causal influence of two key brain regions involved in controlled behaviour – the presupplementary motor area (preSMA) and the right inferior frontal gyrus (rIFG). By applying repetitive transcranial magnetic stimulation (rTMS) and using reaction time (RT)-based distributional analyses, these brain areas’ different contributions to goal directed behaviour can be shown.
Goal pursuit can be impaired by unintentional response tendencies triggered by irrelevant stimuli or stimulus features. These prepotent, possibly false responses can influence early action selection. Thus, they need to be suppressed for the purpose of targeted behaviour, a mechanism referred to as response inhibition. Thus, by investigating these key processes of cognitive control we contribute to an understanding of how navigation in a variable and complex environment can be achieved.
Conflict tasks require both of these processes. Tasks such as the Stroop paradigm (Stroop, 1935), the Eriksen flanker task (Eriksen & Eriksen, 1974), and the Simon task (Simon & Rudell, 1967), require respondents to forego an automatic response triggered by an irrelevant stimulus aspect and instead carry out an appropriate, deliberate action which is selected based on a relevant stimulus aspect. Thus, they model response capture, i.e. the influence of irrelevant stimuli on early action selection, and selective inhibition, i.e. the progressive suppression of the response tendency triggered by irrelevant stimuli in order to carry out an alternative, correct response.
On congruent (CG) trials, the responses triggered by the relevant and irrelevant stimuli overlap and they tend to be faster and more accurate while on incongruent (IC) trials they are in opposition resulting in slower and less accurate responses. The magnitude of the behavioural difference between IC and CG trials, i.e. the interference effect of the irrelevant stimulus aspect, is typically taken as a proxy of the ability to resolve response conflict based on processes such as monitoring of response conflict, response selection and selective response inhibition.
Across tasks, response conflict resolution is associated with the medial frontal cortex and the bilateral inferior frontal cortex and insula as shown in meta analyses (Nee, Wager, & Jonides, 2007; Ridderinkhof, Ullsperger, Crone, & Nieuwenhuis, 2004) and multitask studies (Fan, Flombaum, McCandliss, Thomas, & Posner, 2003; Goghari & MacDonald, 2009; Liu, Banich, Jacobson, & Tanabe, 2004). Furthermore, in bilateral dorsolateral prefrontal cortex,
left premotor cortex, and parietal regions peak activation coordinates clustered across tasks (Nee et al., 2007).
Overall, two brain areas feature prominently across task: the rIFG and the preSMA. This picture is corroborated by non-fMRI data. PreSMA involvement is linked to response selection via individual differences in local gray matter density (van Gaal, Scholte, Lamme, Fahrenfort, & Ridderinkhof, 2011). In terms of the rIFG’s involvement in inhibition, lesions in this brain area are associated with stopping deficits (Aron, Robbins, & Poldrack, 2004). Furthermore, rIFG connectivity to posterior brain regions is linked to selective inhibition in the Simon task (Forstmann et al., 2008). Structual and functional connectivity analyses also showed the preSMA and the rIFG to be in a common inhibition-related network with the subthalamic nucleus (Aron, Behrens, Smith, Frank, & Poldrack, 2007; Duann, Ide, Luo, & Li, 2009; Jahfari et al., 2011) suggesting that the aforementioned pattern of finding both areas in conflict-related contrasts in fMRI studies has an anatomical substrate.
Despite the overlap in associated brain regions, differences between conflict tasks also mean different cognitive processes are needed. The Stroop task involves conflict at the semantic level and is consequently associated with more left lateralized activation (Nee et al., 2007). In the Eriksen flanker task the relevant and irrelevant stimulus dimensions overlap, i.e. stimulus representation codes may interfere besides response codes. Successful behaviour in the Simon task, on the other hand, in less reliant on processes other than response selection and selective motor inhibition and, thus, may offer a more direct focus on the cognitive control mechanisms of interest (Forstmann, van den Wildenberg, & Ridderinkhof, 2008; Hommel, 2011; Proctor, 2011).
The Simon task (Simon & Rudell, 1967) requires participants to overcome a fast response based on lateral stimulus location which automatically maps onto right or left button press responses, and instead they need to translate stimulus identity, e.g., colour or shape, into an appropriate response based on an instructed response mapping. Conflict resolution, often operationalised using the IC>CG BOLD contrast, is associated with medial frontal activity more anterior (BA9; Fruhholz, Godde, Finke, & Herrmann, 2011) or more ventral (BA32; Fan et al., 2003) than the preSMA, superior frontal areas including preSMA (BA6/8; Liu et al., 2004) and various other brain areas including bilateral insula (BA48; Liu et al., 2004) but not rIFG (Fan et al., 2003; Liu et al., 2004). Forstmann, van den Wildenberg et al. (2008) did not find any significant activation using the same contrast. Thus, from these results, it seems as if Simon task conflict resolution is based on somewhat different activation patterns than the aforementioned ones without much consistency in results.
However, as Forstmann, van den Wildenberg et al. (2008) argue, the irrelevant stimulus location’s ability to elicit response capture is also present on CG trials. Thus, both IC and CG trials may involve response selection and selective inhibition processes and as a consequence lead to a partial subtraction of activity of interest in the IC>CG contrast. The use of neutral trials with centrally presented stimuli and RT distribution analyses of interference effects is needed to overcome these shortcomings.
RT Distribution Analyses of Interference Effects
A more sophisticated way of data analysis – which we adopt in this study – involves splitting trials into equally sized RT-bins representing the fastest set of trials, the second fastest, and so on until the slowest set. This distributional analysis approach is motivated by dual-process models of interference effects in the Simon task (Ridderinkhof, Forstmann, Wylie, Burle, & van den Wildenberg, 2010) such as the activation-suppression hypothesis (Ridderinkhof, 2002). In short, they propose that a direct stimulus-response mapping facilitates (CG trials) or competes with (IC trials) a more deliberately chosen, rule-based response (Ridderinkhof et al., 2010).
Critically, the influence of these two response routes is dependent on response speed. Fast responses are more influenced by irrelevant stimuli than slower responses, e.g., fast responses can exhibit below chance accuracy on IC trials (Gratton, Coles, & Donchin, 1992). In terms of accuracy, this within-trial development is well captured by Conditional Accuracy Functions (CAFs, see Figure 1) which plot accuracy as a function of response speed. Given that fast responses tend to be dominated by the direct activation of the response system by the task-irrelevant stimulus position – low accuracy for IC trials, near asymptote accuracy for CG trials – the fast portion of CAFs can be seen as a parametric index of direct response capture by irrelevant stimulus aspects (Ridderinkhof, 2002; Ridderinkhof et al., 2010).
The conflict associated with multiple response activations in which some form of action selection module must work effectively is associated with the rostral cingulate zone generally (Ridderinkhof et al., 2004), an area which includes the preSMA. Furthermore, Forstmann, van den Wildenberg et al. (2008) found individual differences in the first CAF slope of IC trials to correlate with preSMA activation. More activation in this area indicated an increased speed of overcoming response capture.
In order to overcome response capture, the response activated by irrelevant stimulus features needs to be inhibited. However, selective inhibition requires time to exert its effect such that fast responses will be more affected by irrelevant stimulus features than slow ones
(Proctor, Miles, & Baroni, 2011; Ridderinkhof, 2002). As a consequence, slow responses will not show location-driven facilitation of correct responses in CG trials or a delay of correct responses in IC trials. Concerning RT, these within-trial dynamics are well represented by so called delta plots (see Figure 2) which plot the RT difference between corresponding RT-bins, i.e. the Simon effect, as a function of response speed (De Jong, Liang, & Lauber, 1994). Given the aforementioned reduction in the effect of irrelevant stimulus dimensions as selective inhibition builds up with time, one would predict a reduction in the Simon effect as a function of response time. Such a pattern is reliably seen in the standard visual Simon task (Proctor et al., 2011). Selective inhibition can then be operationalized as the slope in the late delta slope segments.
Given lesion evidence linking the rIFG with response inhibition (Aron et al., 2004), this area would be expected to correlate with late negative delta plot slopes. Indeed, Forstmann et al. (Forstmann et al., 2008; Forstmann, van den Wildenberg et al., 2008) found individual differences in the last delta plot slope to correlate with rIFG activation. More activation in this area indicated an increased speed of overcoming response capture. Furthermore, slow responses to IC trials – when answering correctly is critically dependent on selective inhibition – are associated with stronger rIFG activation than slow responses to CG trials (Fruhholz et al., 2011). Thus, introducing a more sophisticated behavioural analysis, brings brain areas associated with conflict resolution in the Simon task in line with those found in other conflict tasks.
rTMS over prefrontal areas and response conflict
However, apart from the lesion study review by Aron et al. (2004) all of the evidence reviewed above is correlational in nature. Determining whether a brain area is not only
associated with but also necessary for a given function requires brain intervention techniques,
the most popular of which is Transcranial Magnetic Stimulation (TMS). TMS transiently disrupts neural activity by discharging a magnetic field burst over the scalp, creating a so called ‘virtual lesion’. Neural processing can be disrupted during behavioural testing (online TMS) or – as applied here – before (offline repetitive TMS). This technique has been successfully employed to probe the involvement of prefrontal brain areas in conflict resolution.
Concerning the preSMA, TMS modulated complete and selective response inhibition as evidenced in mean RT and accuracy (Chen, Muggleton, Tzeng, Hung, & Juan, 2009; Taylor, Nobre, & Rushworth, 2007; but see Verbruggen, Aron, Stevens, & Chambers, 2010).
The effect of TMS over rIFG is similar even though task differences make comparisons difficult (Chambers et al., 2006; Chambers et al., 2007; Verbruggen et al., 2010). In summary, TMS over the prefrontal cortex largely agrees with the aforementioned fMRI findings. The preSMA and the rIFG appear to be causally involved in conflict resolution. Their precise contributions, however, remain unclear.
The present study
The experiment presented here goes beyond previously published studies by (1) using the Simon task whose conflict resolution requirements are thought to be less reliant on secondary processes besides action selection and selective response inhibition, (2) applying the more sensitive behavioural response analysis of CAFs and delta plots which proved valuable in Simon task fMRI studies, and (3) using rTMS over two brain regions which most consistently get associated with conflict resolution – the preSMA and the rIFG (Forstmann, van den Wildenberg et al., 2008). Thus, we aim to causally link two commonly reported brain areas to two key processes underlying action control in complex environments: response selection and selective inhibition. Distributional analyses allow us to distinguish these contributions to response conflict resolution.
We hypothesise that the preSMA is linked to action selection (Ridderinkhof et al., 2010) and is therefore susceptible to response capture by the irrelevant stimulus dimension as evidenced by the first CAF slope of IC trials (Forstmann, van den Wildenberg et al., 2008). Thus, preSMA stimulation should reduce the slope of the first CAF slope segment. The rIFG, on the other hand, is linked to response inhibition (Aron et al., 2004) and, therefore, to late delta plot segments (Forstmann et al., 2008; Forstmann, van den Wildenberg et al., 2008). Thus, rTMS over rIFG stimulation should increase late delta plot slopes.
METHODS
Participants
Thirteen healthy volunteers were recruited. Written consent was obtained prior to the first testing session and the experiment was approved by the ethics committee of the University of Amsterdam. Participants had normal or corrected to normal vision and were without a history of neurological, major medical, or psychiatric disorder. They were screened for counterindications to TMS according to guidelines suggested by Rossi, Hallett, Rossini, Pascual-Leone, & Safety of TMS Consensus Group (2009). One participant was removed from the experiment due to a resting motor threshold of above 64% of maximum output. The remaining twelve participants (seven women) were between 22 and 38 years of age (M=26,
SD=4.37), right handed as assessed by the Edinburgh Inventory (Oldfield, 1971), and had a
resting motor threshold of between 47% and 59% of maximum output (M=53%).
Behavioural Task
A modified visual Simon task was used (Simon & Rudell, 1967) in which a blue or green filled circle was associated either with a right or left hand response using the thumb, see Figure 3A. Across participants the response mapping was counterbalanced. Subjects were instructed to respond as fast and as accurately as possible and to ignore the circle’s location on the right or left side of a previous central fixation cross.
Each trial started with a fixation cross presented with equal likelihood for either 750ms, 1000ms or 1250ms. This was followed by the presentation of the blue or green stimulus which remained on the screen until a response was given or 1500ms elapsed. If the trial lasted longer than 800ms a reminder to respond faster appeared. Trials were pseudorandomised differently for each session with the constraint of a lower than chance exact repetition rate of on average 16.6% - between 15.5% and 17.7% across participants. This was intended to minimise the rejection of repeat trials when analysing sequential effects (analysis not presented here).
Each session consisted of five experimental blocks with 96 trials each. Participants received feedback on their performance after each block. Prior to experimental blocks 96 trials (first session) or 20 trials (all subsequent sessions) were presented to familiarise the subject with the task. An experimental block took approximately five minutes to complete. Each session lasted about 30 minutes.
TMS Procedure
Participants were invited three times with at least two days between each of the rTMS sessions, see Figure 3B. On the first day, after completing one set of training (n=96) and experimental trials (n=5×96) whose data will not enter present analyses, participants’ resting
At least one day break between days 1 and 2 At least two day break between days 2 and 3 Day 1 5 × 96 experimental trials 20 practice trials RMT determination 5 × 96 experimental trials 96 practice trials 20 practice trials 5 × 96 experimental trials 20 practice trials 5 × 96 experimental trials 20 practice trials 20 practice trials 20 practice trials 20 practice trials 20 practice trials 5 × 96 experimental trials 20 practice trials 5 × 96 experimental trials 20 practice trials 20 practice trials 5 × 96 experimental trials 20 practice trials Day 2 20 practice trials 5 × 96 experimental trials rTMS:15min,1Hz,90%RMT 20 practice trials Day 3 20 practice trials 5 × 96 experimental trials rTMS:15min,1Hz,90%RMT 20 practice trials 5 × 96 experimental trials
+
Jitter: 750-1250ms Congruent: position & response match+
Jitter: 750-1250ms
Incongruent: position & response do NOT match
presupplementary motor area (preSMA) right inferior frontal gyrus (rIFG) order counterbalanced across participants
B
A
C
Figure 3. Methods. A Schematic drawing of the different conditions in the Simon task. Note that only one response mapping is shown. They were counterbalanced across participants. Note also that responses were given with the thumb. B The order of experimental sessions and rTMS stimulation. C Stimulation sites.
motor thresholds (RMT) were determined using a biphasic stimulator. We defined an RMT as the minimal coil output necessary to elicit at least 5 motor evoked potentials (MEP’s) with an amplitude of at least 50µV out of 10 trials over the relaxed right abductor pollicis brevis muscle (Rossi et al., 2009).
During the second and third days, participants first completed one session of the behavioural task. Next, repetitive TMS (rTMS) was applied to either the rIFG or the preSMA at 1Hz for 15 minutes (900 pulses) within safety guidelines (Rossi et al., 2009). We used a biphasic stimulator in the form of a flat 70-mm figure of eight coil (Rapid 200, Magstim Rapid Machine, Whitland, Wales, UK) fixed in position using a holding clamp and tripod while the subject’s head rested in a chin rest. Stimulation sites were counterbalanced across participants, see Figure 3C. They were localised using MRI-guided neuronavigation based on previously generated structural MRI images. MRI scans were acquired using a 3T Philips Achieva Magnetic Resonance Imaging System with a voxel size of 1mm×1mm×1mm, 240mm FOV. Average brain coordinates derived from Forstmann, van den Wildenberg et al.'s (2008) fMRI study were translated into subject space and the coil placed above this point. The rIFG (BA 44) was localised at [x = 38, y = 20, z = 4]. The preSMA was localised at [x =4, y = 6, z = 52]. Coordinates are given in MNI space.
Given previous reports of a ‘comfort threshold’ for IFG stimulation of on average 92% of RMT (Chambers et al., 2006; Chambers et al., 2007) the stimulation intensity was set at 90% of RMT for both brain areas. Subjects wore foam earplugs during and after stimulation. Following the application of rTMS, subjects completed another session of the behavioural task.
Data Analysis
Outlier rejection was performed slightly differently for the accuracy data analysis and the RT analysis. Prior to analysing the accuracy data (mean accuracy and conditional accuracy function slopes), all responses faster than 150ms (fast guesses) were removed. Furthermore, all trials with RTs greater or smaller 3 SD off the mean were removed in a participant-, session-, block-, and congruity-specific approach iteratively until all remaining trials were within 3SD of the non-outliers’ mean. Overall, this led to the rejection of 3.33% of trials. Prior to RT-analyses (mean RT and delta slopes) the same outlier rejection with only correctly answered trials was performed. This led to the rejection of 3.27% of correct trials.
Behavioural data was first analysed in a nondistributional way, i.e. mean accuracies of all trials and reaction times (RTs) of correct trials were analysed with two 2(congruity) × 2(time: before or after stimulation) × 2(stimulation site) within-subjects ANOVAs.
Next, we performed distributional analyses. In order to investigate conditional accuracy functions (CAFs), for each participant and congruity level separately, all trials were split into six equally sized bins according to RT. For each RT-bin, we determined the accuracy rates. The slope values between bins were analysed in a 2(congruity) × 5(RT-segment) × 2(time: before or after stimulation) × 2(stimulation site) within-subjects ANOVA.
Similarly, correct trials were split into six equally sized bins based on RT – again, separately for each participant and congruity level – and the RT difference between equivalent congruent and incongruent bin means was determined. The resulting set of delta values track the development of the RT-Simon effect within a trial, as shown in delta plots. Slope values were analysed using a 5(RT-segment) × 2(before or after stimulation) × 2(stimulation site) within-subjects ANOVA. All reported p-values are two tailed.
RESULTS
Non-distributional Analyses
The ANOVA for accuracy values revealed a main effect of congruity, F(1,11)=24.51, p<.001,
pη2=.690, reflecting lower accuracy on incongruent trials (M=0.88, SD=0.05) than on
congruent trials (M=0.94, SD=0.03). The ANOVA for RT values similarly only revealed a main effect of congruity, F(1,11)=98.31, p<.001, pη2=.899, reflecting slower responses on
incongruent trials (M=334ms, SD=29.6) than on congruent trials (M=317ms, SD=30.5). Thus, the average RT-Simon effect was 18ms (SD=6.2) and was not an artefact of a speed-vs.-accuracy trade-off. Non-distributional analyses revealed no other main effects or interactions which approached significance (ps>.1).
Distributional Analyses
Direct Response Capture
The development of the different accuracies of congruent and incongruent trials with time within a trial is shown in Figure 1. The ANOVA for CAF slope values revealed a main effect of congruity, F(1,11)=83.47, p<.001, pη2=.884, RT-segment, F(4,8)=7.84, p<.01, pη2=.797 as
well as the interaction of these two factors, F(4,8)=17.56, p<.01, pη2=.898. Furthermore, the
interaction of RT-segment and time (pre or post stimulation) was significant, F(4,8)=11.55,
p<.01, pη2=.852.
In order to better understand the RT-segment × time interaction, separate ANOVAs for each stimulation site were carried out. While the main effects of congruity and RT-segment as well as their interaction could be found for both rTMS stimulation sites (ps<.05), the RT-segment × time interaction was significant for neither (ps>.1). Therefore, we are inclined to treat the RT-segment × time interaction as a result of training rather than of rTMS stimulation per se. No other main effect or interaction was found in the analysis of CAF slope values (ps>.05).
Selective Inhibition
The development of the RT-Simon effect with time within a trial is shown in Figure 2. The ANOVA for delta slope values revealed a main effect of RT-segment, F(4,8)=16.97, p<.01,
pη2=.895, and for time (pre or post rTMS stimulation), F(1,11)=7.25, p<.05, pη2=.397. The
latter referred to significantly greater slope values after stimulation (M=0.05, SD=0.10) compared to before (M= -0.01, SD=0.09).
Figure 1. Response Capture. A CAF’s regarding preSMA-stimulation separated into five segments. B CAF’s
regarding rIFG-stimulation separated into five segments. Error bar = SEM.
A
B
200 250 300 350 400 450 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 P ro p o rt io n C o rr e c t mean RT (ms) CG: pre preSMA IC: pre preSMA CG: post preSMA IC: post preSMA200 250 300 350 400 450 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 P ro p o rt io n C o rr e c t mean RT (ms) CG: pre rIFG IC: pre rIFG CG: post rIFG IC: post rIFG
In order to better understand the main effect of time, separate ANOVAs for each stimulation site were carried out. While the main effect for RT-segment was significant for both stimulation sites (ps<.05), the time main effect was only significant for the stimulation of the rIFG, F(1,11)=7.796, p<.05, pη2=.415. The latter referred to significantly more positive
slope values after stimulation (M=0.05, SD=0.15) compared to before (M= -0.04, SD=0.12). No other main effect or interaction approached significance in the analysis of delta slope values (ps>.1).
Figure 2. Selective Inhibition. A delta-plot regarding preSMA-stimulation separated into five segments. B delta-plot regarding rIFG-stimulation separated into five segments.
250 300 350 400 450 0 5 10 15 20 25 30 35 40 S im o n e ff e c t (m s ) mean RT (ms) pre rIFG post rIFG 240 260 280 300 320 340 360 380 400 420 440 0 5 10 15 20 25 30 35 40 S im o n e ff e c t (m s ) mean RT (ms) pre preSMA post preSMA
B
A
Still, given the rIFG stimulation effect and the prediction of a slope difference for late RT-segments, we also analysed the last slope segments’ differences in paired t-tests contrasting pre- and post-rIFG-stimulation values. While the third segment’s slopes were significantly different, t(11)= -2.32, p<.05, the fourth segment’s slopes were only marginally different, t(11)=-2.09, p=.06, while the last segment’s slope difference did not approach significance, t(11)=-0.57, p=.58. Mind that the two earliest RT-segments’ differences similarly did not approach significance (ps>.1).
DISCUSSION
Summary
In the present study we aimed to link the preSMA and the rIFG to two key processes underlying conflict resolution: action selection and selective inhibition. The former was thought to be influenced primarily by response capture by the irrelevant stimulus location in the Simon task and given previous fMRI findings based on the same behavioural task (Forstmann, van den Wildenberg et al., 2008) the preSMA was predicted to be involved. In agreement with predictions, CAF slopes increased for incongruent trials – indicating early response capture – but not for congruent trials. However, contrary to predictions, these dynamics were not modulated by rTMS to either the preSMA or the rIFG. Thus, no evidence was found linking the preSMA causally to response inhibition or any other conflict resolution process.
Selective inhibition was conceptualised as a time dependent reduction in the influence of the irrelevant stimulus location (Ridderinkhof et al., 2010). Previous fMRI findings acquired using the same behavioural task (Forstmann, van den Wildenberg et al., 2008) linked the rIFG to a measure of selective inhibition – the last slope segment on delta plots which plot the Simon effect as a function of response time. As predicted, the Simon effect was modulated by response time. However, contrary to predictions, the rIFG did not have a selective effect on late segments. Instead, it increased delta slopes generally, most clearly in the central RT-segment. PreSMA stimulation had no such effect.
Direct Response Capture
Concerning the preSMA null finding, the observed lack of a stimulation effect may be due to the TMS intensity chosen in combination with the precise location of the response selection hot spot in the brain. Previous successful TMS studies on the preSMA’s role in conflict tasks (Chen et al., 2009; Taylor et al., 2007) used higher TMS intensities than adopted here, while Verbruggen et al. (2010) used a lower intensity and also found no effect. Higher intensities also tend to penetrate the brain deeper. This may be crucial given that the precise location of the action selection hot spot inside the medial frontal cortex may lie deeper (see Ridderinkhof et al., 2004) or more to the left (see Forstmann, van den Wildenberg et al., 2008) than what we could reach (as also suggested by Verbruggen et al., 2010).
However, the null finding could also suggest a genuine lack of involvement of the preSMA in response selection. This view would be supported by an unsuccessful replication
attempt (Forstmann, personal communication) of the original preSMA finding reported by Forstmann, van den Wildenberg et al. (2008) and no causal involvement of the preSMA in selecting a nondominant response in a different task (Verbruggen et al., 2010). Ridderinkhof et al. (2010) suggest that the premotor cortex could also be the locus of the response capture effect while Forstmann, van den Wildenberg et al. (2008) could not rule out that their finding is due to a response conflict signalling role of the preSMA.
Selective Inhibition
Concerning the rIFG, TMS reduced the effectiveness of overcoming interference of the irrelevant stimulus dimension across RT-segments without affecting later responses more than earlier ones. This suggests that rIFG involvement in conflict resolution acts faster than previously thought, affecting even the earliest responses. One mechanism could be related to the anatomical and functional connectivity between the rIFG and the subthalamic nucleus (STN; Aron et al., 2007; Jahfari et al., 2011). Nambu, Tokuno, & Takada (2002) characterised this connection – which they call the hyperdirect pathway – as a fast general inhibition pathway of relevant and irrelevant motor programs. This general early inhibition is thought to be subsequently lifted for relevant actions only. This suggests that rIFG stimulation could impair the STN’s ability to inhibit irrelevant actions immediately. Thus, the irrelevant stimulus dimension may increase its influence on reaction times by simultaneously benefitting from a reduced early inhibition mediated by the STN as well as an impaired later inhibition mediated by the rIFG directly. If the latter impairment of the rIFG leads to a delay of selective inhibition rather than a general reduction of it, this may be the reason for the clearest impact of rIFG stimulation on the RT-interference effect of medium-long responses. These may constitute responses where the irrelevant response had time to develop in the absence of early inhibition while later inhibition does not yet have an effect.
In line with this hypothesis, it has been shown that following rTMS over the rIFG STN processing is modulated (Balaz, Srovnalova, Rektorova, & Rektor, 2010) and that direct STN stimulation modulates stop signal inhibition (van den Wildenberg et al., 2006), as well as Simon task response capture and selective inhibition of the irrelevant stimulus dimension (Wylie et al., 2010). However, this ad hoc explanation cannot account for the absent effect of rIFG rTMS on the conditional accuracy functions given that direct STN stimulation does have an effect on response capture evident in CAFs (Wylie et al., 2010).
Conclusion
Despite a great wealth of neuroimaging research linking the preSMA and the rIFG to conflict resolution across tasks, the present study failed to extend these findings by showing causal influences in the way we predicted. In sum, our results suggest that (1) preSMA stimulation parameters adopted here are without effect on the response selection area in the medial frontal cortex, (2) rIFG stimulation parameters may modulate both the cortical brain area itself as well as subcortical areas with which it is intimately connected, leading to increased interference by irrelevant stimulus aspects across reaction times, (3) that more research is needed to validate the causal role of brain areas associated with response conflict resolution.
ACKNOWLEDGEMENTS
This work was supported by grants to K.R.R and W.P.M.v.d.W by the Netherlands Science Association. R.K. was supported by the Studienstiftung des Deutschen Volkes. We would like to thank Birte Forstmann, Martijn Mulders and Wouter Boekel for providing structural MRI scans and stimulation site coordinates. Furthermore, Max Keuken helped with Simon task programming.
REFERENCES
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. The Journal of Neuroscience : The Official Journal of the Society
for Neuroscience, 27(14), 3743-3752. doi:10.1523/JNEUROSCI.0519-07.2007
Aron, A. R., Robbins, T. W., & Poldrack, R. A. (2004). Inhibition and the right inferior
frontal cortex. Trends in Cognitive Sciences, 8(4), 170-177.
doi:10.1016/j.tics.2004.02.010
Balaz, M., Srovnalova, H., Rektorova, I., & Rektor, I. (2010). The effect of cortical repetitive
transcranial magnetic stimulation on cognitive event-related potentials recorded in the
subthalamic nucleus. Experimental Brain Research.Experimentelle
Hirnforschung.Experimentation Cerebrale, 203(2), 317-327.
doi:10.1007/s00221-010-2232-4
Chambers, C. D., Bellgrove, M. A., Gould, I. C., English, T., Garavan, H., McNaught, E., . . .
Mattingley, J. B. (2007). Dissociable mechanisms of cognitive control in prefrontal and
premotor cortex. Journal of Neurophysiology, 98(6), 3638-3647.
doi:10.1152/jn.00685.2007
Chambers, C. D., Bellgrove, M. A., Stokes, M. G., Henderson, T. R., Garavan, H., Robertson,
I. H., . . . Mattingley, J. B. (2006). Executive "brake failure" following deactivation of
human frontal lobe. Journal of Cognitive Neuroscience, 18(3), 444-455.
Chen, C. Y., Muggleton, N. G., Tzeng, O. J., Hung, D. L., & Juan, C. H. (2009). Control of
prepotent responses by the superior medial frontal cortex. NeuroImage, 44(2), 537-545.
doi:10.1016/j.neuroimage.2008.09.005
De Jong, R., Liang, C. C., & Lauber, E. (1994). Conditional and unconditional automaticity:
A dual-process model of effects of spatial stimulus-response correspondence. Journal of
Experimental Psychology.Human Perception and Performance, 20(4), 731-750.
Duann, J. R., Ide, J. S., Luo, X., & Li, C. S. (2009). Functional connectivity delineates distinct
roles of the inferior frontal cortex and presupplementary motor area in stop signal
inhibition. The Journal of Neuroscience : The Official Journal of the Society for
Neuroscience, 29(32), 10171-10179. doi:10.1523/JNEUROSCI.1300-09.2009
Eriksen, B. A., & Eriksen, C. W. (1974). Effects of noise letters upon the identification of a
target letter in a nonsearch task. Perception & Psychophysics, 16(1), 143-149.
Fan, J., Flombaum, J. I., McCandliss, B. D., Thomas, K. M., & Posner, M. I. (2003).
Cognitive and brain consequences of conflict. NeuroImage, 18(1), 42-57.
Forstmann, B. U., Jahfari, S., Scholte, H. S., Wolfensteller, U., van den Wildenberg, W. P., &
Ridderinkhof, K. R. (2008). Function and structure of the right inferior frontal cortex
predict individual differences in response inhibition: A model-based approach. The
Journal of Neuroscience : The Official Journal of the Society for Neuroscience, 28(39),
9790-9796. doi:10.1523/JNEUROSCI.1465-08.2008
Forstmann, B. U., van den Wildenberg, W. P., & Ridderinkhof, K. R. (2008). Neural
mechanisms, temporal dynamics, and individual differences in interference control.
Fruhholz, S., Godde, B., Finke, M., & Herrmann, M. (2011). Spatio-temporal brain dynamics
in a combined stimulus-stimulus and stimulus-response conflict task. NeuroImage, 54(1),
622-634. doi:10.1016/j.neuroimage.2010.07.071
Goghari, V. M., & MacDonald, A. W.,3rd. (2009). The neural basis of cognitive control:
Response selection and inhibition. Brain and Cognition, 71(2), 72-83.
doi:10.1016/j.bandc.2009.04.004
Gratton, G., Coles, M. G., & Donchin, E. (1992). Optimizing the use of information: Strategic
control of activation of responses. Journal of Experimental Psychology.General, 121(4),
480-506.
Hommel, B. (2011). The simon effect as tool and heuristic. Acta Psychologica, 136(2),
189-202. doi:10.1016/j.actpsy.2010.04.011
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. The Journal of Neuroscience : The Official
Journal of the Society for Neuroscience, 31(18), 6891-6899.
doi:10.1523/JNEUROSCI.5253-10.2011
Liu, X., Banich, M. T., Jacobson, B. L., & Tanabe, J. L. (2004). Common and distinct neural
substrates of attentional control in an integrated simon and spatial stroop task as assessed
by event-related fMRI. NeuroImage, 22(3), 1097-1106.
doi:10.1016/j.neuroimage.2004.02.033
Nambu, A., Tokuno, H., & Takada, M. (2002). Functional significance of the
Nee, D. E., Wager, T. D., & Jonides, J. (2007). Interference resolution: Insights from a
meta-analysis of neuroimaging tasks. Cognitive, Affective & Behavioral Neuroscience, 7(1),
1-17.
Oldfield, R. C. (1971). The assessment and analysis of handedness: The edinburgh inventory.
Neuropsychologia, 9(1), 97-113.
Proctor, R. W. (2011). Playing the simon game: Use of the simon task for investigating
human information processing. Acta Psychologica, 136(2), 182-188.
doi:10.1016/j.actpsy.2010.06.010
Proctor, R. W., Miles, J. D., & Baroni, G. (2011). Reaction time distribution analysis of
spatial correspondence effects. Psychonomic Bulletin & Review, 18(2), 242-266.
doi:10.3758/s13423-011-0053-5
Ridderinkhof, K. R. (2002). Micro- and macro-adjustments of task set: Activation and
suppression in conflict tasks. Psychological Research, 66(4), 312-323.
doi:10.1007/s00426-002-0104-7
Ridderinkhof, K. R., Forstmann, B. U., Wylie, S. A., Burle, B., & van den Wildenberg, W. P.
M. (2010). Neurocognitive mechanisms of action control: Resisting the call of the sirens.
Wiley Interdisciplinary Reviews: Cognitive Science, 2(2) doi:DOI: 10.1002/wcs.99
Ridderinkhof, K. R., Ullsperger, M., Crone, E. A., & Nieuwenhuis, S. (2004). The role of the
medial frontal cortex in cognitive control. Science (New York, N.Y.), 306(5695), 443-447.
doi:10.1126/science.1100301
Rossi, S., Hallett, M., Rossini, P. M., Pascual-Leone, A., & Safety of TMS Consensus Group.
transcranial magnetic stimulation in clinical practice and research. Clinical
Neurophysiology : Official Journal of the International Federation of Clinical Neurophysiology, 120(12), 2008-2039. doi:10.1016/j.clinph.2009.08.016
Simon, J. R., & Rudell, A. P. (1967). Auditory S-R compatibility: The effect of an irrelevant
cue on information processing. The Journal of Applied Psychology, 51(3), 300-304.
Stroop, R. J. (1935). Studies of interference in serial verbal reactions. Journal of Experimental
Psychology, 18(6), 643-662.
Taylor, P. C., Nobre, A. C., & Rushworth, M. F. (2007). Subsecond changes in top down
control exerted by human medial frontal cortex during conflict and action selection: A
combined transcranial magnetic stimulation electroencephalography study. The Journal
of Neuroscience : The Official Journal of the Society for Neuroscience, 27(42),
11343-11353. doi:10.1523/JNEUROSCI.2877-07.2007
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(4), 626-636. doi:10.1162/jocn.2006.18.4.626
van Gaal, S., Scholte, H. S., Lamme, V. A., Fahrenfort, J. J., & Ridderinkhof, K. R. (2011).
Pre-SMA graymatter density predicts individual differences in action selection in the face
of conscious and unconscious response conflict. Journal of Cognitive Neuroscience,
23(2), 382-390. doi:10.1162/jocn.2010.21444
Verbruggen, F., Aron, A. R., Stevens, M. A., & Chambers, C. D. (2010). Theta burst
Proceedings of the National Academy of Sciences of the United States of America, 107(31), 13966-13971. doi:10.1073/pnas.1001957107
Wylie, S. A., Ridderinkhof, K. R., Elias, W. J., Frysinger, R. C., Bashore, T. R., Downs, K.
E., . . . van den Wildenberg, W. P. (2010). Subthalamic nucleus stimulation influences
expression and suppression of impulsive behaviour in parkinson's disease. Brain : A
