role of the human medial frontal cortex in task … of the human medial frontal cortex in task...

Role of the Human Medial Frontal Cortex in Task Switching:A Combined fMRI and TMS Study

M.F.S. RUSHWORTH,1,2 K. A. HADLAND,1,✠ T. PAUS,3 AND P. K. SIPILA3

1Department of Experimental Psychology, University of Oxford, Oxford OX1 3UD; 2Oxford Centre for Functional MagneticResonance Imaging of the Brain, Department of Clinical Neurology, University of Oxford, John Radcliffe Hospital,Oxford OX3 9DU, United Kingdom; and 3Montreal Neurological Institute, Montreal, Quebec H3A 2B4, Canada

Received 3 October 2001; accepted in final form 8 January 2002

Rushworth, M.F.S., K. A. Hadland, T. Paus, and P. K. Sipila. Roleof the human medial frontal cortex in task switching: a combinedfMRI and TMS study. J Neurophysiol 87: 2577–2592, 2002;10.1152/jn.00812.2001. We used event-related functional magneticresonance imaging (fMRI) to measure brain activity when subjectswere performing identical tasks in the context of either a task-setswitch or a continuation of earlier performance. The context, i.e.,switching or staying with the current task, influenced medial frontalcortical activation; the medial frontal cortex is transiently activated atthe time that subjects switch from one way of performing a task toanother. Two types of task-set-switching paradigms were investi-gated. In the response-switching (RS) paradigm, subjects switchedbetween different rules for response selection and had to choosebetween competing responses. In the visual-switching (VS) paradigm,subjects switched between different rules for stimulus selection andhad to choose between competing visual stimuli. The type of conflict,sensory (VS) or motor (RS), involved in switching was critical indetermining medial frontal activation. Switching in the RS paradigmwas associated with clear blood-oxygenation-level-dependent signalincreases (“activations”) in three medial frontal areas: the rostralcingulate zone, the caudal cingulate zone, and the presupplementarymotor area (pre-SMA). Switching in the VS task was associated withdefinite activation in just one medial frontal area, a region on theborder between the pre-SMA and the SMA. Subsequent to the fMRIsession, we used MRI-guided frameless stereotaxic procedures andrepetitive transcranial magnetic stimulation (rTMS) to test the impor-tance of the medial frontal activations for task switching. ApplyingrTMS over the pre-SMA disrupted subsequent RS performance butonly when it was applied in the context of a switch. This result shows,first, that the pre-SMA is essential for task switching and second thatits essential role is transient and limited to just the time of behavioralswitching. The results are consistent with a role for the pre-SMA inselecting between response sets at a superordinate level rather than inselecting individual responses. The effect of the rTMS was not simplydue to the tactile and auditory artifacts associated with each pulse;rTMS over several control regions did not selectively disrupt switch-ing. Applying rTMS over the SMA/pre-SMA area activated in the VSparadigm did not disrupt switching. This result, first, confirms thelimited importance of the medial frontal cortex for sensory attentionalswitching. Second, the VS rTMS results suggest that just because anarea is activated in two paradigms does not mean that it plays the sameessential role in both cases.

I N T R O D U C T I O N

A number of imaging studies have identified activationwithin the human medial frontal cortex in tasks involvingresponse conflict and attention to action (Bench et al. 1993;Botvinick et al. 1999; Carter et al. 1995, 1998; Derbyshire et al.1998; Jueptner et al. 1997a,b; Leung et al. 2000; McDonald etal. 2000; Pardo et al. 1991; Paus et al. 1993; Taylor et al. 1994,1997). It has, however, proved difficult to interpret the func-tional role of these activations (Paus 2001) for a number ofreasons.

The behavioral tasks used in studies that have activated themedial frontal cortex are often complex, and it is difficult toknow exactly which aspect of the task was critical in producingmedial frontal activation (Bench et al. 1993; Botvinick et al.1999; Carter et al. 1995; Derbyshire et al. 1998; Leung et al.2000; McDonald et al. 2000; Pardo et al. 1991; Sohn et al.2000). There is longstanding ambiguity about whether conflictoccurs at the level of action or sensory selection in many of theparadigms, such as Stroop (1935) or flanker (Eriksen andEriksen 1974) paradigms, used in imaging studies of attention-to-action and response conflict (McLeod 1991).

There is uncertainty about which medial frontal areas haveattentional or task switching functions. Some studies have hadonly limited spatial resolution and others have been guided bya priori assumptions about the region of interest to be analyzed.There has been some inconsistency in the labeling of humanmedial frontal activations, and diverse medial frontal regionsare likely to have diverse functional roles. A number of studiesof response conflict have recorded activations in a relativelydorsal region of the medial frontal cortex defined by �10 �x � 10, 0 � y � 15, 45 � z � 55 (Talairach and Tournoux1988) in, around, and just posterior to the paracingulate sulcus(Paus et al. 1996a,b). When such activations have been re-corded in response conflict studies, they have sometimes beenassigned to cingulate cortex, but they may be in a humanequivalent of the presupplementary motor area (pre-SMA)(Crosson et al. 1999; Deiber et al. 1999; Luppino et al. 1991;Matsuzaka et al. 1992; Picard and Strick 1996; Sakai et al.1998–2000).

Although medial frontal activation has been prominent inimaging studies (Paus et al. 1998), there is a paucity of inac-

✠ Deceased 22 May 2001.Address for reprint requests: M.F.S. Rushworth, Dept. of Experimental

Psychology, University of Oxford, South Parks Rd., Oxford OX1 3UD, UK(E-mail: [email protected]).

The costs of publication of this article were defrayed in part by the paymentof page charges. The article must therefore be hereby marked ‘‘advertisement’’in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

J Neurophysiol87: 2577–2592, 2002; 10.1152/jn.00812.2001.

25770022-3077/02 $5.00 Copyright © 2002 The American Physiological Societywww.jn.org

tivation or lesion data to constrain its interpretation. Just be-cause an activation change is recorded in an area when twotasks are compared does not mean that the area carries out acognitive operation that is essential for one task but not theother. It is known from animal studies that the activity of singleneurons can be modulated during a task even when interferencestudies indicate that the neurons of the area are not essential forthe task’s performance. For example, the activity of single cellsin the dorsal and ventral premotor cortices is modulated whena monkey reaches, but removal or inactivation of the premotorcortices causes only minor disruption of reaching (Passingham1988; Rea et al. 1987; Wise et al. 1996). Instead interferencestudies suggest that the premotor cortex is important for as-pects of motor learning and the selection of learned movements(Kurata and Hoffman 1993; Kurata and Hoshi 1999; Passing-ham 1993; Petrides 1986; Schluter et al. 1998, 1999). More-over it is now becoming clear that the situation is even morecomplicated in the case of functional magnetic resonance im-aging (fMRI). A combined fMRI, single-unit and field-poten-tial recording study recently suggested that the fMRI signalwas more closely correlated with field potential recording thanwith action potentials recorded from single cells (Logothetis etal. 2001). If this is the case, then fMRI responses may be morereflective of the afferent input to a brain area than the area’soutput.

Finally, a related issue concerns when an activated brain areamakes its critical contribution to the performance of the task.Knowing an area’s critical time of operation may help eluci-date the nature of the cognitive process it performs in a task. Itis not clear if the temporal resolution of human neuroimagingtechniques, even event-related fMRI, is sufficient to disentan-gle the temporal order in which brain activity occurs in differ-ent areas. The interposition of large delays within a task(McDonald et al. 2000) may allow temporal separation of taskcomponents, but it may also drastically alter the nature of thetask by introducing a need for new cognitive processes such asworking memory.

To address these issues, we have used event-related fMRI todefine medial frontal activation in two simple task-set-switch-ing paradigms. In one paradigm, the response-switching para-digm (RS), switching and conflict only occurred with respect toresponse selection. Subjects selected between responses oneach trial, but the rules for response selection varied betweentrial blocks. In the second, the visual-switching (VS) paradigm,attentional switching and conflict occurred in relation to theselection of sensory attributes. Subjects selected between stim-uli according to their shapes or colors; the critical dimensionvaried between trial blocks. We have previously described theVS and RS paradigms as manipulating attentional and inten-tional sets, respectively (Rushworth et al. 2001c). Switching inboth paradigms is associated with similar increases in process-ing demands as indexed by reaction time (RT) increases afterswitching (M.F.S. Rushworth, R. E. Passingham, and A. C.Nobre, unpublished data). For both VS and RS paradigms, ouranalysis was performed by comparing increases in event-related blood-oxygenation-level-dependent (BOLD) signals(“activations”) that were time-locked either with the “switch”cue, which instructed subjects to switch from one way ofperforming the task to the other, or the control “stay” cue,which just instructed subjects to continue performing the par-adigm in the same way as before. We recorded prominent

dorsomedial frontal activation in both paradigms. The pre-SMA was activated in the RS task.

So that we could assess whether and when dorsomedialfrontal activity was essential for task-switching, we used frame-less stereotaxy (Paus 1999) to direct repetitive transcranialmagnetic stimulation (rTMS) at the activated region whilesubjects performed the two paradigms. TMS can be used todisrupt, reversibly and transiently, the normal activity of abrain area (Hallet 2000; Jahanshahi and Rothwell 2000; Pas-cual-Leone et al. 2000; Walsh and Rushworth 1999). Becauseit is an interference technique, TMS can be used to determinewhether a brain area is essential for task performance. Becauseits disruptive effect is transient, TMS can be used to determinewhen a brain area plays a critical role (Ashbridge et al. 1997;Schluter et al. 1998, 1999; Terao et al. 1998; Walsh and Cowey1998; Walsh et al. 1998a,b). When applied over the pre-SMA,TMS disrupted RS performance after a switch cue but not aftera stay cue. TMS over a midline control site 4-cm posterior tothe pre-SMA site did not disrupt performance in the same way.

To examine the specificity of the effect, we used severalprocedures during the course of additional TMS experiments(experiments 3 and 4). First we replicated the effect of rTMSin a different group of subjects, using higher frequency TMS(10 Hz rather than 5 Hz). Second, we examined the effect ofTMS at a different frontal control site. To control for theeffects of having inadvertently stimulated adjacent premotorareas, we tested the effect of TMS over the dorsal premotorcortex area in the vicinity of the superior branch of the superiorprecentral sulcus (Schluter et al. 1998, 1999).

Third, in addition to examining the spatial specificity of theTMS effect, we also examined its temporal specificity. Inaddition to applying TMS immediately after switch or staycues, we also tested its effect on the first trial of a task blockafter either a switch or stay cue. This is an important control forseveral reasons. Given the nature of the BOLD signal and itsmodeling, the results from the fMRI part of the experimentprobably reflect cognitive processes related to the performanceof the first trials of the block in addition to the preceding switchor stay cues. Moreover, it has been suggested that distinctcognitive processes occur during set switching; switching maybegin with a prospective process of set initiation and re-configuration after the switching instruction that can be distin-guished from subsequent performance of the new task (Meiran2000; Meiran et al. 2000; Monsell et al. 2000; Rogers andMonsell 1995; M.F.S. Rushworth, R. E. Passingham, and A. C.Nobre, unpublished results). In this way, we were able tocompare the role of the dorsomedial frontal cortex with that ofthe premotor cortex in prospective set re-configuration andsubsequent actual task performance.

The fourth set of control procedures involved testing theeffect of TMS on attentional switching in the VS paradigm.The effect of TMS over dorsomedial frontal cortex was com-pared during two different time periods, either immediatelyafter the switch or stay cues or at the time of the first trials ofswitch or stay blocks.

In combination, the fMRI experiments and the various TMSexperiments demonstrated that a dorsomedial frontal area,probably the pre-SMA, had a role in re-configuring intentionaltask set in the RS paradigm and that its role could be distin-guished from the role of the dorsal premotor cortex in selectingindividual task responses. Despite its activation the dorsome-

2578 RUSHWORTH, HADLAND, PAUS, AND SIPILA

J Neurophysiol • VOL 87 • MAY 2002 • www.jn.org

dial region did not appear to play the same role in the re-configuration of attentional set in the VS paradigm.

M E T H O D S

Experiments 1 and 2—fMRI and 5 Hz dorsomedialfrontal TMS

SUBJECTS. In total, 20 right-handed, healthy volunteers participatedin the fMRI recording study (ages 19–31 yr). The vision of allsubjects was normal or corrected to normal with MRI-compatibleglasses. Ten subjects performed the RS paradigm and 10 performedthe VS paradigm. The data from two subjects who performed the VStask was lost after mains power failures disrupted data acquisition andstorage. Eleven of the 20 subjects participated in the subsequent TMSstudy, 6 performed the RS paradigm, and 5 performed the VS para-digm. All subjects gave their informed written consent before partic-ipation. The procedures were approved by the Research Ethics Com-mittee of the Montreal Neurological Institute and Hospital.

BEHAVIORAL TASKS. Experiments were conducted both with sub-jects lying in the coil of the dimly illuminated MRI scanner room orin the dimly illuminated TMS laboratory. Stimulus presentation forfMRI and TMS tests was controlled by essentially identical computerprograms. Stimuli were presented on a computer monitor in front ofthe subjects in the TMS study. In the fMRI study, the stimuli wereprojected onto a screen using an LCD projector at the head of thescanner tube. Subjects performing the fMRI task used a mirror so thatthe stimulus appeared directly in front of them.

RS. Figure 1 (left) summarizes the RS paradigm. The RS paradigmconcerned intentional set switching and it targeted the mechanisms oftask-switching that depend on changing the rules for response selec-tion and response conflict. On each trial, subjects saw either a redtriangle (5.1° width, 2.7° high) or rectangle (3.7° width, 2.7°high).During the first set of trials, subjects made a right-hand response to therectangle and a left-hand response to the triangle. A small circle (0.9°diam, 70-ms duration) provided feedback to the subjects 100 ms afterthe response (yellow for correct responses and blue for incorrectresponses). An interval of 800 ms followed before the onset of thenext trial. The intervals between trial onsets varied according to thevariable reaction times, and averaged approximately 1,500 ms.

Each experimental session was broken down into blocks of 9–11trials. The rules by which responses were selected varied betweenblocks; on some blocks, subjects responded with a left-hand responseto rectangles and a right-hand response to triangles. Each block waspreceded by an instruction cue. Instruction cues were either a vertical(�) or a tilted (�) cross appearing in a white rectangular background(6° width, 5° high) presented for 200 ms. Cues indicated that thesubject should either switch rules for response selection or stay withthe current response selection rules. There was a 1,000-ms intervalbetween the onset of the instructive cue and the onset of the first pairof items. The meaning assignment (switch, stay) of each cue (�, �)was counterbalanced across subjects.

In the fMRI, study trials were presented in four sessions each of5-min duration. The event-related analysis was centered on a com-parison of BOLD signal after the switch and stay cues. In each of thefour sessions, there were approximately 10 switch and 10 stay cues(approximately 40 switch and 40 stay cues in total). Each cue wasseparated from the subsequent and preceding cues by a variableinterval of 11–14 s. The timing of both types of cue onset wasrecorded with respect to the onset of acquisition of each frame offMRI data (see following text).

The TMS study was conducted on a separate subsequent day. In theTMS study, each session consisted of 400 trials. As in the fMRIexperiment the switch and stay cues were the focus of the investiga-tion. Half of both types of cues were followed by a 5-Hz 4-pulserTMS train applied over the pre-SMA area activated in the fMRI

study (Fig. 2). It should be emphasized that the rTMS train wascompleted before the initiation of the first trial. The reaction times(RTs) for responses made on the first trial after stay or switch cues,either after rTMS or no rTMS, were recorded. Median RTs (averageof 8–10 trials) for each category of trial (stay with no rTMS, stay withrTMS, switch with no rTMS, switch with rTMS) for each subject werethen calculated. RTs from both correct and incorrect trials wereincluded; there was a slight, but nonsignificant, tendency for more

FIG. 1. Left: details of the response-switching (RS) task. Subjects werepresented with a series of task items. The items were always either rectanglesor triangles. Subjects alternated between 2 response selection rules eithertriangle-left hand and rectangle-right hand or triangle-right hand and rectangle-left hand. The figure shows an example where the subject started with the 1strule and later switches to the 2nd rule. Every 9–11 trials, a white cue shapeinstructed subjects to either stay with the current selection rule set or to switchto using the other rule set. Stay or switch cues were differentiated by a plus (�)or tilted cross (�) at their center. The meaning of the plus and tilted cross wascounterbalanced across subjects. In the example shown, the plus and the tiltedcross mean stay and switch, respectively. Right: examples of stimuli used inthe visual-switching (VS) task. The task was formally similar to the RS task.On each trial, subjects were presented with a pair of stimuli (items) either sideof a central fixation point. One of the stimuli was always red and the othergreen. One of the stimuli was always a rectangle and the other a triangle. Thesubject attended to just 1 of the 2 stimuli according to a rule based on eithercolor or shape. The design was fully counterbalanced so that some subjectsalternated between attending to red or rectangle stimuli, while others alternatedamong red or triangle, green or rectangle, or green or triangle. For example, thesubject might start (top) by attending to just the red stimulus item on everytrial, regardless of its shape. Every 9–11 trials a white cue shape instructedsubjects to either stay with the current selection rule or to switch to using aselection rule based on the other stimulus dimension. For example the subjectmight then attend to the rectangle stimulus regardless of color (7th panel).Again stay or switch cues were differentiated by a plus or tilted cross at theircenter. The meaning assignments of the plus and tilted cross were counterbal-anced across subjects as before. In the example shown, the plus and the tiltedcross mean stay and switch, respectively. The subject’s task was to detect arare target, “w” (4th panel) and respond with a key-press. The “w” only everappeared in the attended stimulus and only on 20% of trials. On other trials,only nontarget “v” were presented and no response was required. Both thetarget and nontarget, “w”and the “v,” were only present for the final 15 ms ofthe total 70 ms of stimulus presentation.

2579MEDIAL FRONTAL CORTEX AND TASK SWITCHING


errors to be made on trials preceded with rTMS. Two dorsomedialfrontal sites were tested: the pre-SMA (activated in the switch-staycomparison) and a control site, 4 cm posterior to the pre-SMA (aregion not activated in the switch-stay comparison). The RTs fromeach site were tested with two-way repeated-measures ANOVAs. Thefirst factor was switch, with two levels corresponding to stay andswitch trials. The second factor was TMS, with two levels TMS andnon-TMS control.

VS. Figure 1 (right) summarizes the VS task. The VS paradigmcomplemented the RS paradigm and shared most aspects of its formaldesign. It was designed to study the mechanisms involved in setswitching that depend on switching attention between different sen-sory dimensions of stimuli. Two visual stimulus items were presentedsimultaneously (70-ms duration) to either side (1.7° eccentricity) of awhite central fixation cross (1.3° width, 1.1° high) on a black back-ground on a PC monitor. The two items always consisted of onerectangle shape (1.7° width � 2° high) and one triangular shape (2.6°width [with base up], 2° high). One of the items was always green andone of the items was always red. Either shape could be combined witheither color. Subjects used either a particular shape (e.g., rectangle) orcolor (e.g., red) to direct their attention to the relevant item to detectoccasional embedded targets (see following text). There was a vari-able 1,200- to 1,500-ms interval between trials.

As in the RS paradigm, each experimental session was brokendown into shorter blocks of 9–11 trials. At the beginning of anexperimental block, during the first set of trials, subjects were told toattend to one particular stimulus feature (e.g., red color) and identifytargets that appeared within the relevant (red) item. Subsequently,instruction cues appeared before each set of 8–17 trials. Instructioncues were either a vertical (�) or a tilted (�) cross appearing in awhite rectangular background (6° width, 5° high) presented for 200ms. Cues indicated that the subject should switch the current visualrule for selection or stay with the current visual rule. There was a1,000-ms interval between the onset of the instructive cue and theonset of the first pair of items.

The visual selection rule was switched between particular pre-defined features in different dimensions (e.g., red and rectangle). Forexample, starting with the relevant feature “red,” the switch cue(e.g., �) would inform the subject that the relevant feature became“rectangle.” The next switch cue instructed the subject that the rele-vant feature returned to being red. The appearance of the stay cue (�)instructed subjects to continue selecting items based on their currentvisual rule. The meaning assignment (switch, stay) of each cue (�, �)was counterbalanced across subjects. The specific features for eachdimension relevant for selection (red/rectangle, red/triangle, green/rectangle, green/triangle) were also counterbalanced across subjects.

The counterbalancing of cue assignment and selection featuresensured that behavioral measures were un-confounded with artifactsdue to different physical appearances of the stimuli. Five levels ofmatched red and green luminosities were used randomly for itemcolors throughout the experiment. Differences in the physical intensityof stimuli therefore were unlikely to contribute systematically toattentional effects.

To ensure feature-guided sensory attention, subjects were asked todiscriminate small target stimuli embedded within the items. A small(0.7° long and 0.06° high) horizontal or angled line was presented inthe middle of each item. The embedded stimulus appeared only briefly(15 ms) at the end of each item presentation (55 ms after item onset)to maximize the advantage of orienting toward the relevant item. Onmost trials (80%), embedded nontargets were presented; the nontargetwas either a horizontal line or a line angled upward (approximating a“v”) to different degrees (0.06–2.9°). On rare (20%) target trials, theline was deviated downward (into a “w,” always by 2.9°). Subjectsresponded on the detection of the rare target (w) with a singlekey-press. Targets (w) only ever appeared in the relevant visualdimension to which subjects were attending.

As for RS, trials in the fMRI study of VS were presented in foursessions each of 5-min duration. The event related analysis wascentered on a comparison of BOLD signal after the switch and staycues. In the fMRI version of the paradigm, the low probability oftarget presentation (20%) meant that the event-related fMRI analysisof the switching of sensory attention would be largely uncontaminatedby response-related brain activity.

The TMS study version of the paradigm, RT on the trials afterswitch and stay cues was the measured behavioral index of attentionswitching. In the TMS version, there was a very high probability oftarget presentation (80%) on the first trial following both stay andswitch cues. This ensured sufficient data for analysis. Subjects werenot told of variations in target presentation probability.

The TMS study was conducted on a separate, subsequent day in asession of 600 trials. The switch and stay cues were the focus of theinvestigation. As in the RS paradigm, half of both types of cues werefollowed by a 5-Hz 4-pulse rTMS train applied over the dorsomedialfrontal cortex area activated in the fMRI study (Fig. 2). It should beemphasized that the rTMS train was completed before the initiation ofthe first trial. The RTs for responses made on the first trial after stayor switch cues, either after rTMS or no rTMS, were recorded. MedianRTs (average of 8–10 trials) for each category of trial (stay with norTMS, stay with rTMS, switch with no rTMS, switch with rTMS) foreach subject were then calculated. The RTs were tested with atwo-way repeated-measures ANOVA. The first factor was Switch,with two levels corresponding to stay and switch trials. The secondfactor was TMS, with two levels TMS and non-TMS control.

MRI ACQUISITION. Scanning was performed on a 1.5 T SiemensVision magnet. The scanning procedure began with the acquisition ofa T1 structural anatomical scan (80 slices at a thickness of 2 mm,256 � 256 matrix size, TR � 22 ms, TE � 10 ms, flip angle � 30°,voxel size � 1 � 1 � 2 mm3). This was immediately followed byacquisition of four series of 120 gradient-echo images (20 slices of5-mm thickness in the same orientation as the Sylvian fissure startingabove the most dorsal cortex, 64 � 64 matrix size, TR � 2.441 ms,TE � 50 ms, flip angle �90°, voxel size � 5 � 5 � 5 mm3) of BOLDsignal while subjects performed the behavioral tasks.

EVENT-RELATED FMRI DATA ANALYSIS. All images were trans-formed into standardized stereotaxic space. This was accomplished byusing an automatic image-registration method (Collins et al. 1994)based on multi-scale three-dimensional (3-D) cross-correlation withan average (n � 305) MR image aligned with the Talairach stereo-taxic space (Talairach and Tournoux 1988). The transformation islinear, yielding three scaling factors for the width (x axis), length (yaxis), and height (z axis) of the brain and effectively removinginter-individual differences in brain size. BOLD signal images were

FIG. 2. The RS and VS tasks were formally similar. Both tasks werecomposed of blocks of 9–11 items separated by the presentation of cues. Thecues were either switch or stay cues that instructed subjects to switch theselection rule or to continue with the current selection rule respectively. In thefunctional magnetic resonance imaging (fMRI) study (experiment 1), theblood-oxygenation-level-dependent (BOLD) signals following switch and staycues were compared. In the repetitive transcranial magnetic stimulation(rTMS) study in experiment 2, the interval between the presentation of a switchor stay cue and1st task item (the cue period) could either be unfilled or filledwith a train of rTMS (indicated by gray shading). The reaction times (RTs) ontrials preceded by rTMS were compared with control trials.



smoothed with a 3-D 6-mm (full-width half-maximum) Gaussiankernel, corrected for head motion artifact and transformed into thesame standard stereotaxic space. The statistical analysis was carriedout with adapted in-house software (Worsely et al. 2000) using amethod based on a linear model with correlated errors and a random-effects analysis. Task-related brain activity was measured by exam-ining the BOLD signal following the switch and stay cues in the VSand RS paradigms; the BOLD signal was convolved with a hemody-namic response function that was modeled as a gamma-density func-tion with a mean lag of 7 s and a SD of 3 s (Zarahn et al. 1997) timedto coincide with the onset of switch or stay cues. Drift was removedby adding third-order polynomial covariates in the volume acquisitiontimes in the design matrix (which were not convolved with thehemodynamic response function). Random effects T-statistical mapsof significant difference between cue related BOLD signals wereconstructed by using a spatially smoothed (150mm full width half-maximum Gaussian kernel) estimate of the random effects variance.The t-statistical maps were then thresholded (t � 4.75, P � 0.01; t �5.19, P � 0.001) in accordance with the Bonferonni correction formultiple comparisons (for the entire 20-slice brain-volume scanned)and nonisotropic random field theory (Worsely et al. 1996, 1999).

TMS. A Cadwell high-speed magnetic stimulator and a 5-cm-diamCadwell (Kennewick) figure-8 “cone” coil were used to administerrTMS. Each rTMS train was delivered as a 5-Hz sequences of fourpulses. Intensity of stimulation was set to be 5% above the thresholdfor eliciting a visible twitch of the foot when the TMS was appliedduring mild dorsiflexion of the ankle in all subjects (subjects wereinstructed to dorsiflect at 10% of full force). TMS intensity wastherefore set to be between 85 and 90% of the Cadwell stimulator’smaximum output. The rTMS trains used in the experiment began 200ms after the onset of the switch or stay cue.

Coil placement, in both VS and RS experiments, was guided by theposition of dorsomedial frontal activation in each individual subject.Because it was soon apparent that rTMS over the pre-SMA signifi-cantly disrupted RS task performance, we also tested the effects ofrTMS over a control site, 4 cm posterior to the pre-SMA. The controlsite stimulation was approximately over the site of the SMA (Fink etal. 1997). No significant switch-stay BOLD signal differences wererecorded at the control site.

The coordinates of the dorsomedial activation peak were deter-mined individually for each subject. This target position was markedon the subjects anatomical MRI scan using Brainsight (Rogue Re-search, Montreal, Canada) software. The subject’s anatomical MRIscan was then co-registered with the subject’s head using framelessstereotaxy (Paus 1999). The subject’s head position was assessed byusing the Polaris (Northern Digital, Waterloo, Canada) infra-redtracking system to measure the position of scalp land marks (nasion,nose-tip, intra-trageal notch of left and right ears) also visible on thesubject’s anatomical MRI. Once the subject’s head and MRI scanwere co-registered, infra-red tracking was used to monitor the positionof the TMS coil with respect to the subject’s brain. The TMS coil wasthen placed over the target brain area.

Experiment 3: 10-Hz dorsomedial frontal TMS at cue anditem periods

In this experiment, TMS was again applied over the dorsomedialfrontal cortex while subjects performed both VS and RS paradigms.Although the experiment resembled the preceding TMS experiment,there were three important differences.

First, a higher rate of TMS, 10 Hz, rather than 5 Hz, was used inthis experiment. It is possible that the failure of TMS trains to impairswitching in the VS paradigm was due to the relatively slow rate ofTMS pulse presentation in that experiment. In experiment 3, the rateof TMS pulse presentation was doubled. We (Hadland et al. 2001) andothers (Harmer et al. 2001) have shown that 10-Hz TMS, at or just

above the motor threshold for visible movement of the foot duringdorsiflexion, is sufficient to elicit behavioral effects when applied overthe dorsomedial frontal cortex.

Second, we employed longer testing sessions in experiment 3 sothat it was possible to gather more data for each subject performingeach condition. It was therefore possible to exclude incorrect re-sponses from the RT analysis of the TMS effects in experiment 3.

Third, the effects of TMS were tested at two different time periods.In experiment 2, TMS was applied after some switch and stay cues(Fig. 2). In experiment 3, we again tested the effect of applying TMSafter stay and switch cues, but we also tested the effect of applyingTMS on presentation of the first task item after either a switch or astay cue (Fig. 3). We have therefore referred to the two different timesof TMS application as the cue (Fig. 2) and item (Fig. 3) periods.

SUBJECTS. The effect of TMS during the cue period of the RSparadigm was studied in six subjects. The effect of the TMS duringthe item period of the RS paradigm was studied in six differentsubjects. The effect of TMS during the cue period of the VS paradigmwas studied in six subjects. The effect of TMS in during the itemperiod of the VS paradigm was studied in six different subjects. Allsubjects gave their informed consent before participation and theprocedures were approved by the Central Oxfordshire Research EthicsCommittee (reference No. C99.178).

BEHAVIORAL TASKS. The same RS and VS paradigms were used asin experiments 1 and 2. The sessions were longer than in experiment2; subjects usually performed about 800 trials, although, if too manymistakes were made or if the randomized delivery of TMS occurredon too few occasions, then the number of trials was sometimesincreased to 900 or 1,200 trials. The results for the RS and VSparadigms were analyzed separately using a between-subjectsANOVA approach similar to that used in experiment 1; within-subjectfactors of switch and TMS were used as before, and in addition, abetween subject factor of period (with 2 levels corresponding to thecue or item periods) was used.

TMS. A Magstim (Whitland, Wales, UK) rapid high-speed magneticstimulator and a Magstim double-cone coil were used to administerrTMS. The cue period rTMS consisted of a 1-s train at 10 Hz. Itemperiod rTMS consisted of 0.5-s train at 10 Hz (some subjects began torespond within 0.5 s of task item presentation). The rTMS trains usedin the experiment began with the onset of the switch or stay cue in thecue-period experiment or, in the case of the item-period experiment,30 ms prior to the onset of the first task item after the switch or staycue (in pilot experiments, we found that TMS trains that startedco-incidentally with the item period stimulus onset caused somesubjects to sometimes blink during presentation of the brief 15-mstarget used in the VS experiment). Intensity of stimulation was set to5% above the threshold for eliciting a visible twitch of the foot whenthe TMS was applied during mild dorsiflexion of the ankle in all

FIG. 3. Trains of rTMS pulses were delivered at 2 different times inexperiments 3–5. In some cases, rTMS trains were delivered during the cueperiod (Fig. 2). On other occasions, however, rTMS was delivered at adifferent time period referred to as the item period. The item period was theinterval after presentation of the first task item subsequent to either a switch orstay cue. This period could be either unfilled or filled with a train of rTMS(indicated by gray shading). The RTs on trials preceded by rTMS werecompared with control trials.



subjects (subjects were instructed to dorsiflect at 10% of full force).TMS intensity was therefore set to be between 45 and 70% of theMagstim stimulator’s maximum output.

Coil placement, in both VS and RS experiments, was determined asthe position 5 cm anterior to the maximally excitable leg representa-tion in the motor cortex (Hadland et al. 2001). We have previouslyshown that this leads to placement of the center of the coil over thepre-SMA region (Hadland et al. 2001), and this was confirmed inseven of the subjects using the MRI-guided frameless stereotaxicprocedures described in the preceding text. The frameless stereotaxicprocedure confirmed that the coil was placed similarly in both exper-iments.

Experiment 4: 10-Hz dorsal premotor TMS at cueand item periods

Experiment 4 was conducted in a similar way to experiment 3. Themain difference was that TMS was directed over the dorsal premotorcortex.

SUBJECTS. Five subjects were tested on two occasions in the RSparadigm. On each occasion TMS was either delivered in the cueperiod or in the item period. All subjects gave their informed consentbefore participation and the procedures were approved by the CentralOxfordshire Research Ethics Committee (reference No. C99.178).

BEHAVIORAL TASKS. The RS paradigm was used in the same way asin experiment 3. The analyses performed were the same as in exper-iment 3 except that the factor period was now a within-subject factorbecause the same subjects had been tested with both cue and itemperiod TMS.

TMS. A Mastim Rapid was used to apply 10-Hz TMS trains in eitherthe cue or item period as in experiment 3. Instead of a cone coil,however, a flat 70-mm Magstim figure-8 coil was used. Stimulationintensity was no longer set with respect to the threshold for stimulat-ing the leg area on the medial wall because now the stimulation sitewas on the lateral surface adjacent to the motor cortex representationof the hand area. Instead stimulation intensity was set at 5% above thethreshold for eliciting a visible thumb twitch when the coil was placedover the maximally excitable hand representation in the motor cortex.As before the intensity was set to be appropriate for each individualsubject (normally between 50 and 70% of stimulator maximum out-put). To place the coil over the dorsal premotor cortex, it was moved2 cm anterior and 2 cm medial to the motor cortex using proceduressimilar to those previously described and those that we and othershave shown leads to placement of the coil in the vicinity of thesuperior branch of the superior precentral sulcus (Praamstra et al.1999; Schluter et al. 1998, 1999), and this was confirmed in five of thesubjects using the MRI-guided frameless stereotaxic procedures de-scribed in the preceding text. The coil was held tangential to the skullwith the handle pointing backwards approximately parallel to themid-sagittal axis.

R E S U L T S

Experiment 1: fMRI

RS. In the scanner, switching was associated with a behav-ioral cost measurable in reaction time (RT). Nine of the 10 RSsubjects responded more slowly on the first trial of a switchblock (mean, 605 ms) than they had on the first trial of a stayblock (mean, 505 ms). The difference was significant (Wil-coxon T � 0, n � 10, P � 0.008). There were significantincreases in BOLD signal on switching (switch-stay compari-son) in four medial frontal regions (Table 1, Fig. 4). All fouractivations were in the left hemisphere. The most prominentactivation had a peak in or just posterior to the paracingulate

sulcus (x � �10, y � 9, z � 53) and extended dorsally to coverthe adjacent medial aspect of the superior frontal gyrus. Wehave therefore labeled this activation as pre-SMA. Two acti-vations had peaks in the cingulate sulcus, approximately 2 cmanterior and posterior the vertical plane at the anterior com-missure (VCA plane), and were labeled as rostral and caudalcingulate zones (RCZ and CCZ). The fourth medial frontalactivation was considerably more anterior, extended beyondthe medial surface and was labeled as frontal pole.

There were also areas of significant decrease in BOLDsignal on switching (stay-switch comparison) in the medialfrontal cortex (Table 2, Fig. 5). The peaks were situated in bothvery anterior (y � 42) and subcallosal cingulate cortex.

VS. In the scanner, switching was associated with a behav-ioral cost measurable in RT. All eight VS subjects respondedmore slowly on the first trial of a switch block (mean, 665 ms)than they had on the first trial of a stay block (mean, 579 ms).The difference was significant (Wilcoxon T � 0, n � 8, P �0.012). Fewer medial frontal increases in BOLD signal onswitching (switch-stay comparison) were recorded in the VSparadigm (Table 1, Fig. 4). There was a significant differenceat just two voxels in the cingulate sulcus (x � �2, y � 21, z �37) in the RCZ region. More prominent was a more dorsome-dial activation at the posterior end of the paracingulate sulcus(x � �8, y � 3, z � 60), adjacent to that labeled pre-SMA inthe RS experiment. The peak activation in the VS task wasmore than 9 mm (direct distance in 3-D space) from thatrecorded in the RS task. The more dorsal and caudal (justanterior to the VCA plane) position of the VS activation meantthat it was not clear if it should be ascribed to the SMA orpre-SMA. It was therefore described as “SMA/pre-SMA.”

There were also areas of significant decrease in BOLDsignal on switching (stay-switch comparison) in the medialfrontal cortex (Table 2, Fig. 5). As in the RS task, the peakswere situated in very anterior (y � 49) and subcallosal cingu-late cortex.

Experiment 2: 5-Hz dorsomedial frontal rTMS

RS. Figure 6 shows the target sites in the subjects taking partin the rTMS experiment. Figure 7 shows the co-registration ofthe TMS coil with the pre-SMA target area, using the framelessstereotaxic procedure in one subject.

There was a significant main effect of applying rTMS overthe pre-SMA site on subjects subsequent RTs (F � 13.253,df � 1, 4, P � 0.022). Subjects’ RTs on the first trials after

TABLE 1. Medial frontal areas of BOLD signal increase(switch-stay)

TalairachCoordinates t Values

RS taskFrontal pole �21 63 3 6.08Paracingulate sulcus region (pre-SMA) �10 9 53 7.69Cingulate sulcus (RCZ) �8 21 38 5.98Cingulate sulcus (CCZ) �5 �18 53 7.15

VS taskParacingulate sulcus region (SMA/pre-SMA) �8 3 60 7.74Cingulate �2 21 37 4.92*

* Only two suprathreshold voxels. BOLD, blood oxygenated level depen-dent.



switch cues were a mean of 265 ms slower when rTMS hadbeen given (Fig. 8A); this difference was significant (t � 2.668,df � 4, P � 0.028). The RTs of all five subjects were slowedby rTMS on switch trials. Subjects’ RTs on the first trials afterstay cues were a mean of 3 ms slower when rTMS had beengiven (Fig. 8A); this difference was not significant (t � 0.62,df � 4, P � 0.05). In summary, rTMS over pre-SMA disruptedRS performance but only on switch trials.

The effect of applying rTMS over the more posterior controlsite was quite distinct (Fig. 8B); none of the three subjects’ RTswere slowed on switch trials. None of the effects of applyingrTMS over the control site were significant.

VS. Figure 6 shows the targets sites for the subjects takingpart in the TMS experiment. The application of rTMS over theSMA/pre-SMA activation did not disrupt performance on ei-ther stay or switch trials (Fig. 8C). There was a general trendfor subjects to perform slightly faster after rTMS although theeffect did not approach significance.

Experiment 3: 10-Hz dorsomedial frontal TMS at cue anditem periods

Figure 9 shows the TMS target sites in a group of sevensubjects as measured with the frameless stereotaxic procedure.The average point of intersection of the coil trajectory withcortex was at �7, 8, 64) (Talairach and Tournoux 1988).

FIG. 4. T-statistical maps (3.5 � t �10;threshold: t � 4.75, P � 0.01) showing areasof greater activity on switching (switch-stay)from the fMRI experiment (experiment 1).The results for the RS and VS paradigms areshown in red and green, respectively, on axial(left) and sagittal (right) group-averaged ana-tomical MRI scans. The red line on the sagittalsections marks the vertical plane at the ante-rior commissure (VCA plane). The sagittalsections shows that, on the medial surface,switching in the RS paradigm is associatedwith BOLD signal changes in the presupple-mentary motor area (pre-SMA, b) and rostralcingulate zone (RCZ, d). The caudal cingulatezones (CCZ) region (c) is out of the plane ofsection in the sagittal view shown and only alimited area of BOLD signal can be seen.Instead the CCZ (c), together with the pre-SMA (b), can be seen clearly on the axialsection on the left. On the medial surface,switching in the VS task is only associatedwith BOLD signal change at the SMA/pre-SMA boundary (a) just anterior to the VCAplane. In addition, 2 voxels in the RCZ regionexceeded threshold. Note that the axial andsagittal sections showing the VS results areslightly more medial and dorsal than thoseshowing the RS results, reflecting the moredorsal and medial position of the VSactivation.

TABLE 2. Medial frontal areas of BOLD signal decrease(switch-stay)

Talairach Coordinates t Values

RS taskFrontal pole �15 54 21 �5.22Anterior paracingulate sulcus 7 53 20 �6.07Ventral subcallosal cingulate �8 48 2 �5.33

�18 43 1 �5.54�15 48 �10 �5.43

VS taskAnterior paracingulate sulcus 2 51 16 �7.08

�12 57 12 �6.79Ventral subcallosal cingulate �9 49 �8 �9.902Medial frontal cortex (area 9) �9 32 53 �17.381

FIG. 5. t-Statistical maps (�10 � t � �3.5; threshold: t � �4.75, P �0.01) showing areas of decreased activity on switching (switch-stay) from thefMRI experiment (experiment 1). The results for the RS and VS paradigms areshown in red and green, respectively, on sagittal group averaged anatomicalMRI scans. The red line marks the VCA plane. In both cases switching relateddecreases are seen on the very anterior ( y � 42) and ventral subcallosalcingulate cortex.



Figure 9 also shows the co-registration of the TMS coil withthe pre-SMA target area in one example subject.

RS. There was a significant main effect of applying TMS overthe pre-SMA site on subjects’ subsequent RTs (F � 6.025,df � 1, 10, P � 0.034) and a significant main effect of Switch(F � 21.364, df � 1, 10, P � 0.001). The effect of TMSdepended on whether it was applied after a switch or stay cue;there was a significant interaction between TMS and Switch

factors (F � 6.054, df � 1, 10, P � 0.034). In addition theeffect of TMS depended on whether it was applied in theearlier cue period or the later item period; there was a signif-icant three-way interaction among the factors of TMS, Switch,and period (F � 5.701, df � 1, 10, P � 0.038). From Fig. 10(A and B) it is clear that the statistical interactions were due toTMS having its most disruptive effect when it was applied inthe cue period after a switch cue (compare � and ▫ in Fig. 10A,right). Subjects’ RTs on the first trials after switch cues were amean of 295 ms slower when rTMS had been given; thisdifference was significant (1-tailed t � 2.499, df � 5, P �0.028). RTs were slowed for all six subjects.

VS. As in the RS paradigm, in the VS paradigm there was alsoa significant main effect of Switch (F � 8.431, df � 1, 10, P �0.016). There was, however, no significant main effect of TMSnor interaction between TMS and Switch factors. TMS andperiod factors did interact (F � 5.554, df � 1, 10, P � 0.040).From Fig. 10 (C and D) it is apparent that this is due to the factthat TMS tended to speed RTs when it was applied in the cueperiod (Fig. 10C), and it tended to slow RTs when it wasapplied in the item period (Fig. 10D), regardless of whether ornot subjects were switching between sets. The TMS inducedslowing in the item period was seen in half of individualsubjects’ data and was not significant. The TMS-induced fa-cilitation in the cue period, which was similar to that observedin experiment 2, again varied between subjects and did notreach significance.

Experiment 4: 10-Hz dorsal premotor TMS at cueand item periods

Figure 11 shows the TMS target sites in a group of fivesubjects as measured with the frameless stereotaxic procedure.The average point of intersection of the coil trajectory withcortex was at �36, 0, 64 (Talairach and Tournoux 1988).Figure 11 also shows the co-registration of the TMS coil withthe dorsal premotor target area in one example subject.

FIG. 6. Group-averaged MRI scans of subjects taking part in the VS (top)and the RS (bottom) paradigms. The circles indicate the targets for rTMS inindividual subjects (experiment 2). Two regions were targeted in the RS task:the pre-SMA (5 subjects) anterior to the VCA plane (shown as a red line) anda control region 4 cm posterior behind which no switch related activity hadbeen recorded (3 subjects). The control region is approximately over likelyposition of the SMA (Fink et al. 1997). The coil was always placed tangentialto the scalp, indicated by the yellow ‘T’ shapes. Just a single site was targetedin the VS task (5 subjects).

FIG. 7. Positioning the coil using the frame-less stereotaxic procedure (experiment 2). Oncethe subject’s head is co-registered with his orher MRI scan, the position of the TMS coil withrespect to the target area in the brain can bevisualized. On the left is a surface view of asingle subject’s MRI. The yellow and red crossindicates the main axes of the TMS coil. The redline projects above and into the coil center. Itcan be seen that the coil is positioned just to theleft of the midline (the activated midline areaswere always in the left hemisphere). At rightangles to the red line are smaller yellow linesrepresenting the coil’s anterior-posterior andmedial-lateral orientation. It can be seen that thecoil is positioned tangentially with respect to thebrain’s surface. On the right is a section takenthrough the same subject. The section is just outof the sagittal plane, instead its main axes aredefined by those of the coil’s orientation. A redx indicates the coil position. A yellow x markedcg indicates a medial frontal target for stimula-tion, just posterior to the paracingulate sulcus.



RS. As in the case of dorsomedial TMS during the RS task,TMS also had a significant effect when it was delivered overthe dorsal premotor cortex (F � 15.2044, df � 1, 4, P �0.018). In other respects, however, the results were different tothe dorsomedial TMS results in experiments 2 and 3. First,although there was a significant main effect of Switch (F �

9.599, df � 1, 4, P � 0.036), it clearly did not interact withTMS (F � 0.005, df � 1, 4, P � 0.945) nor was there anysuggestion of a three-way interaction of TMS, Switch, andperiod (F � 0.007, df � 1, 4, P � 0.935). On the other hand,again unlike dorsomedial TMS, there was a significant inter-action between TMS and period (F �16.090, df � 1, 4, P �0.016). From Fig. 12 (A and B) it is clear that dorsal premotorTMS slowed subjects performance when it was applied in theitem period (Fig. 12B), regardless of whether or not subjectshad just switched sets. All five subjects showed the samepattern of RT slowing when TMS was delivered in the itemperiod, and the slowing was significant, both in the context ofswitching set (t � 3.869, df � 4, P � 0.018) or staying withthe same set as previously (t � 4.002, df � 4, P � 0.0016).There was a slight and nonsignificant speeding of RT whenTMS was delivered during the cue period.

FIG. 9. Dorsomedial frontal coil position over the pre-SMA region inexperiment 3 in a group (top) and a representative single subject (bottom). Agroup-averaged MRI, in standard Talairach space, for 7 of the participatingsubjects is shown. The circles indicate the targets for rTMS in individualsubjects. The VCA plane is shown as a red line. Bottom: a section from a singlesubject is shown. The section is shifted away from the sagittal plane andinstead its main axes are defined by those of the coil’s orientation. A red xindicates the coil position. A yellow x indicates a medial frontal target forstimulation just superior to the caudal end of the paracingulate sulcus.

FIG. 8. Experiment 2 results. A: RTs, in the RS paradigm, on the 1st trialsafter stay and switch cues, preceded by pre-SMA rTMS or a control period.RTs on switch trials with rTMS are significantly slower than switch trialswithout rTMS. Pre-SMA rTMS disrupts switching from one way of perform-ing the task to another. B: RTs, in the RS paradigm, on the 1st trials after stayand switch cues, preceded by posterior control site rTMS or a control period.The RTs on switch trials with rTMS are no slower than switch trials withoutrTMS; in fact, the rTMS trials tend to be faster. This may be an artifact due tothe tactile and auditory sensations associated with each TMS. Such tactile andauditory sensations lead to inter-sensory facilitation of RT. C: RTs, in the VSparadigm, on the 1st trials after stay and switch cues, preceded by SMA/pre-SMA rTMS. The RTs on switch trials with rTMS are no slower than switchtrials without rTMS, in fact, the rTMS trials tend to be faster. This may be anartifact due to the tactile and auditory sensations associated with each TMS.Such tactile and auditory sensations lead to inter-sensory facilitation of RT.



D I S C U S S I O N

In the present experiments, we used event-related fMRI tomeasure brain activity when subjects were performing identi-cal tasks either in the context of a behavioral switch or in thecontext of a continuation of earlier performance. The context,switching-task set or staying with the current-task set, influ-enced medial frontal cortical activation; the medial frontalcortex is transiently activated at the time that subjects switchfrom one way of performing a task to another. The medialfrontal activation was more extensive in the RS paradigm,which required intentional set switching and involved changingthe rule for response selection and response conflict. One areaof activation was probably in the pre-SMA. The application ofrTMS over the pre-SMA disrupted RS performance but only onswitch trials. The effect was most clear when the TMS wasapplied during the cue period when subjects engage in aprospective process of set re-configuration prior to actual per-formance of the new task. TMS over adjacent premotor regionsdid not have the same effect. TMS over the dorsal premotorcortex disrupted the selection of individual task responses, butit did not affect wholesale task set reconfiguration. The resultssuggest a transient but essential role for the pre-SMA inintentional set switching that can be dissociated from the roleof the dorsal premotor cortex in selecting individual, specific

responses. An area at the SMA/pre-SMA border was the onlymedial frontal area activated in the VS paradigm, which en-tailed attentional set switching between rules for stimulusselection and stimulus conflict. The application of rTMS overthe SMA/pre-SMA border, at the same time (after the switchcue), did not disrupt performance of VS. There was someequivocal evidence for an effect of medial frontal TMS at thelater time period (the item period) when subjects were activelyperforming the task, but the effects were not specific to switch-ing trials and did not reach statistical significance. The medialfrontal cortex does not appear to play the same role in re-configuring attentional set in the VS task as it does in the RStask.

Location of switch-related activations in the medialfrontal cortex

Switching in the RS task was associated with activationchanges in several medial frontal regions (Fig. 4). The locations ofthe activations suggest that the attentional/task switching role ofthe medial frontal cortex is closely tied to its motor role; three ofthe activated regions, in the paracingulate and cingulate sulci,probably correspond to medial premotor areas (Deiber et al. 1999;Fink et al. 1997; Paus 2001; Paus et al. 1993; Picard and Strick

FIG. 10. Experiment 3 results: dorsomedial frontal pre-SMA region rTMS. A: cue period rTMS in the RS paradigm. RTs on the1st trials after stay and switch cues, preceded by cue period pre-SMA rTMS or a control period. RTs on switch trials with rTMSare significantly slower than switch trials without rTMS. B: item period rTMS in the RS paradigm. RTs on the 1st trials after stayand switch cues, preceded by item period pre-SMA rTMS or a control period. Pre-SMA item period rTMS does not significantlyslow performance in the RS paradigm. C: cue period rTMS in the VS paradigm. RTs on the 1st trials after stay and switch cues,preceded by item period pre-SMA rTMS or a control period. In the VS paradigm, cue period rTMS does not slow downperformance. D: item period rTMS in the VS paradigm. RTs on the 1st trials after stay and switch cues, preceded by item periodpre-SMA rTMS or a control period. Pre-SMA item period rTMS does not significantly slow performance in the VS paradigm.



1996). There were two activations in the cingulate sulcus, approx-imately 2 cm anterior and 2 cm caudal to the VCA plane. Theseactivations fall into premotor regions that have been described asRCZ and CCZ (Deiber et al. 1999; Picard and Strick 1996). Taskswitching was associated with BOLD signal decreases in both RSand VS in anterior and in ventral subcallosal cingulate areas (Fig.5). Such decreases are consistent with models of cingulate cortexthat emphasize its functional heterogeneity (Devinsky et al. 1995;Koski and Paus 2000; Paus et al. 1998). The proposed cognitivefunctions of the cingulate cortex seem closely related to its motorfunctions and depend on a relatively restricted supracallosal re-gion extending only a limited distance anterior to the VCA planein humans as is the case in monkeys (Rushworth et al. 2000). Themore anterior and ventral cingulate cortex in both species may bemore concerned with social and emotional processes (Bush et al.2000; Devinsky et al. 1995; Rushworth et al. 2000).

The more dorsal medial frontal activation recorded in RS

was in or just posterior to the paracingulate sulcus. Activationsin this region have been recorded in a number of responseswitching or response conflict paradigms (Pardo et al. 1991;Paus et al. 1993; Taylor et al. 1994) although they have notbeen labeled consistently. Crosson et al. (1999) have discussedthe difficulty of deciding whether activations in this region arein the pre-SMA or the cingulate cortex. In the present study,although its peak was in a sulcus, the activation appeared toextend dorsally into the medial aspect of the superior frontalgyrus. Because of this, and in accordance with previous studiesof this region (Crosson et al. 1999; Deiber et al. 1999; Picardand Strick 1996; Sakai et al. 1998, 1999a,b), the activation hasbeen labeled as pre-SMA. Disbrow et al. (2000) have com-pared the position of fMRI-recorded BOLD signals with mi-croelectrode recordings of activity in the same task and foundthat the BOLD signal may be biased toward the position oflocal blood vessels. In the same way, the current fMRI-basedestimate of the pre-SMA’s position may be biased ventrallytoward the vessels in and around the paracingulate sulcus.

VS was associated with definite activation in just one medialfrontal area in or just posterior and dorsal to the paracingulatesulcus (Fig. 4). The peak of this activation was 9 mm distantfrom the pre-SMA activation recorded in RS. We have referredto this peak as SMA/pre-SMA because of its position at theproposed boundary between the SMA and the pre-SMA (Pass-ingham 1995; Picard and Strick 1996). Stephan et al. (1995)suggested that even within the SMA proper there is a divisionor transition between a more anterior region, near the VCAplane, and a more posterior region. It is tempting to identify theVS activation with anterior SMA (see also following text). Thecertain conclusion that VS and RS activated distinct areas inthe anterior SMA and the pre-SMA respectively awaits anintra-individual direct comparison of the two paradigms with ahigher resolution scanning method. It should be emphasizedthat the lack of medial frontal activation in VS cannot beattributed to VS just being easier. Behavioral data gatheredduring scanning and in previous experiments showed a similarRT cost for switching in both VS and RS (Rushworth et al.2001; M.F.S. Rushworth, R. E. Passingham, and A. C. Nobre,unpublished data); moreover, the VS task clearly activatedother frontal areas, such as those in vicinity of superior pre-central sulcus (�35, �1, 69; 18, �13, 71; �26, �5, 61; 25,�2, 56; an example can be seen in the top left quadrant ofFig. 4).

Attention to action vs. sensory attention

All three medial premotor areas activated in RS, RCZ, CCZ,and pre-SMA, have previously been associated with responseconflict and attention to action and its consequences (Bench etal. 1993; Botvinick et al. 1999; Carter et al. 1995, 1999;Jueptner et al. 1997; Leung et al. 2000; MacDonald et al. 2000;Pardo et al. 1991; Passingham 1998; Paus et al. 1993, 1998;Posner and DiGirolamo 1998; Taylor et al. 1994, 1997; Turkenand Swick 1999). Because of the complex nature of the tasksused in some of these studies; however, it is not always clearthat conflict is occurring at the response level, as opposed to anearlier sensory level; there is ambiguity about the locus ofconflict in Stroop (Stroop 1935) and flanker (Eriksen andEriksen 1974) paradigms (MacLeod 1991). Confirmation ofthe importance of response rule switching and conflict for

FIG. 11. Coil position over the dorsal premotor region in experiment 4 in agroup (top) and a representative single subject (bottom). A group-averagedMRI, in standard Talairach space, for 5 of the participating subjects is shown.The circles indicate the targets for rTMS in individual subjects. Bottom: asection from a single subject is shown. The section is shifted out of the sagittalplane and instead its main axes are defined by those of the coil’s orientation.A red x indicates the coil position. A yellow x indicates the dorsal premotortarget for stimulation by the superior branch of the superior precentral sulcus.



activation of RCZ and CCZ came from the fMRI results in theVS paradigm; there was no activation in CCZ, and just twovoxels of significant activation in the RCZ region, associatedwith switching.

The importance of response conflict for RCZ and CCZactivation is consistent with a recent metanalysis of positronemission tomography (PET) studies recording cingulate acti-vation; Paus et al. (1998) found that blood flow changes in thisregion were associated with experiments involving fast manualresponding. Turken and Swick (1999) reported that a patientwith a restricted cingulate lesion was only impaired on theStroop task when responses were manual rather than vocal.The human RCZ and CCZ are thought to be homologous withrostral and caudal cingulate motor areas (CMAs) in the ma-caque monkey brain (Dum and Strick 1993; Shima et al. 1991).All the CMAs have direct connections with the spinal cord andthe motor cortex (Dum and Strick 1991, 1996; He et al. 1995;Lu et al. 1994; Luppino et al. 1991).

Exactly which aspect of the response demands of a task arecritical for activating the cingulate sulcal regions remains to beelucidated. It has been suggested that the cingulate may play acritical role in monitoring responses, perhaps for errors (Bushet al. 2000; Carter et al. 1998; Dehaene et al. 1994; Luu et al.2000; MacDonald et al. 2000). Single-cell, field recording, andmuscimol inactivation studies in the monkey have providedevidence consistent with this hypothesis (Gemba et al. 1986;Shima and Tanji 1998). It is possible that subjects in thepresent experiments might have monitored their own responsesfor errors in the context of a response set switch. It should benoted, however, that the cingulate BOLD increases cannotreflect the actual commission of errors; very few errors weremade in the fMRI scanner (only 1–4% of trials were error trialsacross all subjects), and these were not more frequent onswitch block trials.

The pre-SMA was also activated on switching in the RSparadigm. Identifying the pre-SMA with intentional switchingand attention to action, as opposed to sensory attention switch-ing, however, was more difficult; an adjacent area on theSMA/pre-SMA border was activated in VS. There are anatom-

ical reasons for thinking that the more caudal and dorsal VSactivation is a distinct region to the pre-SMA area activated inRS (see preceding text). The current rTMS results, however, doshow that the pre-SMA plays an essential role in intentional setswitching in RS but that the SMA/pre-SMA region, althoughactivated, is not essential for attentional set switching in VS.The delivery of TMS, in the cue period (Fig. 2) in the RS taskslowed subsequent performance when subjects were switchingset but not when they were staying with the current set (ex-periments 2 and 3, Figs. 8A and 10A). The delivery of TMS, inthe cue period in the VS task did not slow subsequent perfor-mance regardless of whether or not subjects were switching set(experiments 2 and 3, Figs. 8C and 10C). There was a slightslowing when TMS was delivered after the first item in a taskblock (Fig. 3), in both the VS and RS paradigm, in the contextof both switching and maintaining set (experiment 3, Fig. 10,B and D). These item period effects varied between subjectsand did not approach significance.

Essential and nonessential activations

To establish the causal importance of a brain area for acognitive function, it is necessary to use interference tech-niques in addition to techniques for measuring the activity ofsingle cells or populations of cells. For example, the activity ofpremotor neurons change when a monkey makes a reachingmovement, but lesion or inactivation of the premotor cortexhas a minimal effect on reaching movements under normalcircumstances (Passingham 1988; Rea et al. 1987; Wise et al.1996). Instead, interference studies, conducted with permanentlesions (Passingham 1993; Petrides 1986), temporary musci-mol inactivation (Hoshi and Kurata 1999; Kurata and Hoffman1993), and TMS (Schluter et al. 1998, 1999), suggest that thepremotor cortex is essential for aspects of motor learning andthe selection of learned movements. Although the importanceof fMRI imaging for the understanding of human brain func-tion cannot be overestimated, it can be instructive to combinethe technique with others. In the case of fMRI, is possible thatthe BOLD signal may reflect the afferent input to an area rather

FIG. 12. Experiment 4 results: dorsal premotor rTMS. A: cue period rTMS in the RS paradigm. RTs on the 1st trials after stayand switch cues, preceded by cue period dorsal premotor rTMS or a control period. No disruptive effect is seen when rTMS isdelivered over the dorsal premotor cortex rather than the pre-SMA (Figs. 8 and 10), during the cue period in the RS paradigm. B:item period rTMS in the RS paradigm. RTs on the 1st trials after stay and switch cues, preceded by item period dorsal premotorrTMS or a control period. A disruptive effect is seen when rTMS is delivered over the dorsal premotor cortex, rather than thepre-SMA (Fig. 10), during the item period in the RS paradigm. The disruptive effect of dorsal premotor rTMS is seen regardlessof whether or not subjects are switching set.



than the output of the area; a combined fMRI, single-unit andfield-potential recording study recently suggested that thefMRI signal was more closely correlated with field-potentialrecording than with action potentials recorded from single cells(Logothetis et al. 2001).

The pre-SMA region is critical for switching in RS. Theapplication of rTMS over this region, after a switch cue,disrupted subsequent performance (experiments 2 and 3, Figs.8A and 10A). The effect was unlike that seen when TMS wasdelivered over other areas. Dorsal premotor cortex TMS (ex-periment 4) did not disrupt task performance when it wasdelivered during the cue period (Fig. 2, Fig. 12A), although itdid interfere with the selection of individual task responseswhen it was delivered in conjunction with task items (Figs. 3and 12B).

The SMA/pre-SMA was activated in VS. We did not, how-ever, obtain evidence that the medial frontal cortex had thesame essential role in reconfiguring attentional set in the VSparadigm as was the case for the RS paradigm; rTMS, in thecue period (after the switch cue was presented and beforesubjects re-engaged in task performance) did not disrupt VSperformance, either on switch or nonswitch trials (Figs. 8C and10C). One explanation for the absence of a disruptive effect inVS might be that switching was not as difficult in VS as it wasin RS. This possibility seemed unlikely from our previousbehavioral experiments when longer RTs were recorded in theVS task than the RS task, suggesting that, if anything, the VStask was the more difficult (Rushworth et al. 2001; M.F.S.Rushworth, A. C. Nobre, and R. E. Passingham, unpublishedresults, Fig. 2). Unlike our previous experiments, however, thesubjects studied with rTMS in the present experiments dem-onstrated smaller RT costs on switching that the subjects takingpart in the RS rTMS experiment. Nevertheless there was clearevidence of an effect of the switching manipulation in both theexperiments that tested the effect of dorsomedial frontal TMSdirected over the SMA/pre-SMA region (experiments 2 and 3).All five subjects studied in the VS part of experiment 2 wereslower to respond after a switch cue than a stay cue, demon-strating that there was still a behavioral cost of switching. Notall subjects were slower on switching trials in the VS part ofexperiment 3, but the manipulation was still effective enoughto produce a statistically significant slowing effect on the groupas a whole. Moreover, it is not the case that the VS paradigmis somehow less susceptible to TMS; it is clear that the VSparadigm is susceptible to rTMS interference but only if it isapplied over brain areas other than the SMA/pre-SMA; spe-cifically posterior parietal rTMS disrupts VS performance(Hadland and Rushworth 2000). TMS over the posterior pari-etal cortex has been shown to affect several visual-attentiontasks (Ashbridge et al. 1997; Gobel et al. 2001; Rushworth etal. 2001; Walsh and Cowey 1998, 2000; Walsh et al. 1998,1999) and we have recently presented evidence that, not sur-prisingly, TMS at the same site affects VS performance (Had-land and Rushworth 2000).

Although there was no evidence that medial frontal TMSduring the cue period (immediately after the presentation of theswitching cue and prior to actual task performance) disruptedswitching, there was some equivocal evidence of a TMS effectwhen it was applied at the item period when subjects wereactively engaged in the task. This aspect of the results must beinterpreted with caution. First, it should be noted that because

these effects varied between subjects, they did not approachstatistical significance. Because these effects are not specific tojust switching trials and because a similar effect was notobserved when TMS was applied during the anticipatory in-terval of the cue period, it is difficult to associate them specif-ically with switching set rather than some other aspect ofcomponent task performance.

In summary, the interpretation of the SMA/pre-SMA acti-vation recorded during the VS paradigm is not clear. Althoughit may reflect a role for the medial frontal cortex in switchingattentional set in the VS paradigm, it is not clear that its rolehere is similar to its role in switching intentional set in the RSparadigm. An alternative hypothesis is that it might be due tomodulation of subjects’ sustained preparation to respond to therare targets. Cortex in the SMA/pre-SMA region has beenassociated with imagining or preparing responses (Lee et al.1999; Stephan et al. 1995). We have recorded a greater bereit-shafts potential (BP) after a switch as opposed to a stay cue inthe VS paradigm (M.F.S. Rushworth, A. C. Nobre, and R. E.Passingham, unpublished observations), and the BP has beenassociated with preparation (Ikeda et al. 1992). Such prepara-tory activity, however, is not confined to the pre-SMA and/orSMA but is also prominent in some parietal and premotor areas(Deiber et al. 1996; Krams et al. 1998; Rushworth et al. 2001;Wise 1985). If preparatory activity is so widespread, then thatseen in the pre-SMA may not be essential for response prep-aration.

Role of the pre-SMA in task switching

The present results demonstrate that the pre-SMA plays arole in task set switching. The role that it is played by thepre-SMA can now be distinguished from that of a number ofother medial and lateral premotor areas.

First, the role of the pre-SMA can be distinguished from thatof the cingulate cortices. Cingulate areas were also activatedduring intentional set switching in the RS task. A series ofrecent fMRI studies (Botvinick et al. 1999; Carter et al. 2000;MacDonald et al. 2000), however, have suggested that cingu-late activation is more closely tied to the subsequent periodwhen response conflict is greatest rather than the prior periodof top-down control of switching. We have also shown thatlesions in the macaque cingulate cortex do not affect thetop-down control of set switching (Rushworth et al. 2000). TheBOLD signal modeling in the present experiments did not justcapture set re-configuration related activity, but it is also likelyto have reflected activity associated with the first trials of thetask blocks when response conflict is greatest. The cingulatearea activation (experiment 1) may, therefore, reflect suchprocesses. The TMS results, however, suggest a very differentrole for the pre-SMA. The delivery of TMS over this region(experiments 2 and 3) was most effective during the cue period(Figs. 2 and 3) when subjects only had the opportunity tore-configure task set prospectively and before the subjects hadany opportunity to select specific responses to particular taskitems or before subsequent response conflict monitoring couldoccur.

The role of the pre-SMA can also be distinguished from thatof the dorsal premotor cortex. BOLD signal increases onswitching in the RS paradigm were also recorded in the vicinityof the superior precentral sulcus and adjacent gyrus (�30, �7,



73; �26, 14, 62; �16, 8, 62; �16, 8, 62; examples can be seenin the bottom left quadrant of Fig. 4). The timing of thedisruptive effects of dorsal premotor and pre-SMA TMS, how-ever, are quite distinct and suggest that the two regions havedistinct roles. There was an interaction between the maineffects of TMS and set switching, and a three-way interactionbetween the main effects of TMS, set switching and stimula-tion period (cue or item period) in the pre-SMA results (ex-periment 3, Fig. 10, A and B) that showed that the region wasmost important when subjects were switching set, not other-wise, and that the region was most important when set switch-ing was initiated rather than when subjects were selectingspecific responses. In the case of the dorsal premotor cortex(experiment 4), however, there was no interaction between theeffect of TMS and set switching, but there was an interactionbetween the main effects of TMS and stimulation period (Fig.12, A and B). This results suggests that the premotor cortex hasa role in selecting between specific responses rather than the inthe wholesale switching of response set. This is consistent withother single-unit recording, lesion, temporary inactivation,neuroimaging, and TMS data that confirms that the dorsalpremotor cortex (PMd) has a role in selecting individual re-sponses that are arbitrarily associated with stimuli (Kurata andHoffman 1994; Passingham 1993; Petrides 1986; Schluter etal. 1998; Toni et al. 1999; Wise et al. 1996). We suggest thatthe pre-SMA may have a role in selecting between sets of suchresponse selection rules. This hypothesis not only accommo-dates the current finding that the pre-SMA is involved in taskswitching in the RS paradigm but accommodates demonstra-tions of the pre-SMA’s involvement in motor sequences (Ger-loff et al. 1998; Hikosaka et al. 1996; Jueptner et al. 1997;Nakamura et al. 1998, 1999; Sakai et al. 1998, 1999; Shima etal. 1996). The pre-SMA concern with motor sequences may bejust a particular instance of its more general role in selectingbetween sets of responses.

Consistent with this view is the finding that pre-SMA activ-ity is most prominent during the learning of a new sequence(Hikosaka et al. 1996; Jueptner et al. 1997b; Nakamura et al.1998, 1999; Sakai et al. 1998; Shima et al. 1996). Hikosaka andcolleagues have developed the “2 � 5 task” (Hikosaka et al.1995) for testing the learning and retention of motor sequencesby monkeys. In this task, the monkey has to press two illumi-nated targets in the correct order. Five sets of two targets arepresented in a fixed order “hyperset.” Pre-SMA neurons areactive during the performance of the 2 � 5 task, but it is clearthat their activity is often restricted to just the first of the twomovements of a set or sometimes even just for the first move-ment of a hyperset (Nakamura et al. 1998). The cell recordingresults can be interpreted in the same way as the present fMRIand rTMS results; the pre-SMA is concerned with the selectionof a superordinate set of responses rather than each individualresponse.

Cell recording studies also confirm the anticipatory role ofthe pre-SMA in reconfiguring the response selection prior toexecution. Matsuzaka and Tanji (1996) taught monkeys tomake one of two movements on hearing a 1-kHz GO signal. On“stay” trials, the monkey simply made the same movement ason the previous trial. On “switch” trials, however, a 50-Hzsound indicated that the monkey should make the oppositemovement on hearing the next 1-kHz, imperative GO signal.Matsuzaka and Tanji found that pre-SMA neurons began firing

when the monkey heard the switch cue in advance of theimperative GO signal. In a similar vein, Gerloff et al. (1997)found that TMS over the pre-SMA disrupted movements madesome time later in a movement sequence; pre-SMA TMS-induced disruption occurred 1 s after motor cortex TMS-induced disruption.

In summary, the pre-SMA appears to have an anticipatory orprospective role in the wholesale re-configuring intentional orresponse set. This superordinate role in selecting between setsof responses can be distinguished from the subordinate role ofthe PMd in selecting between individual responses. The pre-SMA is more concerned with the switching of intentional setthan it is with the switching of attentional set.

We gratefully acknowledge the advice of K. Worsley, C. Liao, V. Petre, B.Pike, and H. Johansen-Berg and the assistance of P. Hobden.

This work was supported by the Royal Society, Medical Research Councilof Great Britain, the Canadian Institutes of Health Research, and the CanadianFoundation for Innovation.

REFERENCES

ALLPORT A AND WYLIE G. Task switching: positive and negative priming oftask set. In: Attention Space and Action, edited by Humphreys GW, DuncanJ, and Treisman A. New York: Oxford, 1999, p. 273–298.

ASHBRIDGE E, WALSH V, AND COWEY A. Temporal aspects of visual searchstudied by transcranial magnetic stimulation. Neuropsychologia 35: 1121–1131, 1997.

BENCH CJ, FRITH DC, GRASBY PM, FRISTON KJ,PAULESU E, FRACKOWIAK RSJ,AND DOLAN RSJ. Investigations of the functional anatomy of attention usingthe stroop test. Neuropsychologia 31: 907–922, 1993.

BOTVINICK M, NYSTROM LE, FISSELL K, CARTER CS, AND COHEN JD. Conflictmonitoring versus selection-for-action in anterior cingulate cortex. Nature402: 179–181, 1999.

BUSH G, LUU P, AND POSNER MI. Cognitive and emotional influences inanterior cingulate cortex. Trends Cogn Sci 4: 215–222, 2000.

BUSH G, WHALEN PJ, ROSEN BR, JENIKE MA, MCINERNEY SC, AND RAUCH SL.The counting stroop: an inteference task specialized for functional neuro-imaging—validation study with functional MRI. Hum Brain Map 6: 270–282, 1998.

CARTER CS, BRAVER RS, BARCH DM, BOTVINICK MM, NOLL D, AND COHEN

JD. Anterior cingulate cortex, error detection, and the online monitoring ofperformance. Science 280: 747–749, 1998.

CARTER CS, MACDONALD AM, BOTVINICK M, ROSS LL, STENGER VA, NOLL D,AND COHEN JD. Parsing executive processes: strategic vs. evaluative func-tions of the anterior cingulate cortex. Proc Natl Acad Sci USA 97: 1944–1948, 2000.

CARTER CS, MINTUM M, AND COHEN JD. Interference and facillitation effectsduring selective spatial attention: an H215O PET study of Stroop TaskPerformance. Neuroimage 2: 264–272, 1995.

COLLINS DL, NEELIN P, PETERS TM, AND EVANS AC. Automatic 3D intersub-ject registration of MR volumetric data in standardized Talairach space.J Comput Assist Tomogr 18: 192–205, 1994.

CROSSON B, SADEK JR, BOBHOLZ JA, GOKCAY D, MOHR CM, LEONARD CM,MARON L, AUERBACH EJ, BROWD SR, FREEMAN AJ, AND BRIGGS RW.Activity in the paracingulate and cingulate sulci during word generation: anfMRI study of functional anatomy. Cereb Cortex 9: 307–316, 1999.

DEHAENE S, POSNER ME, AND TUCKER DM. Localization of a neural system forerror detection and compensation. Psychol Sci 5: 303–305, 1994.

DEIBER M-P, HONDA M, IBANEZ V, SADATO N, AND HALLETT M. Mesial motorareas in self-initiated versus externally triggered movements with fMRI:effect of movement type and rate. J Neurophysiol 81: 3065–3077, 1999.

DEIBER M-P, IBANEZ V, SADATO N, AND HALLET M. Cerebral structuresparticipating in motor preparation in humans: a positron emission tomog-raphy study. J Neurophysiol 75: 233–247, 1996.

DERBYSHIRE SWG, VOGT BA, AND JONES AKP. Pain and stroop interferencetasks activate seperate processing modules in anterior cingulate cortex. ExpBrain Res 118: 52–60, 1998.

DEVINSKY O, MORRELL MJ, AND VOGT BA. Contributions of anterior cingulateto behaviour. Brain 118: 279–306, 1995.



DISBROW EA, SLUTSKY DA, ROBERTS TPL, AND KRUBITZER LA. FunctionalMRI at 1.5 tesla: a comparison of the blood oxygenation level-dependentsignal and electrophysiology. Proc Natl Acad Sci USA 97: 9718–9723,2000.

DUM RP AND STRICK PL. The origin of corticospinal projections from thepremotor areas in the frontal lobe. J Neurosci 11: 667–689, 1991.

DUM RP AND STRICK PL. Cingulate motor areas. In: Neurobiology of CingulateCortex and Limbic Thalamus, edited by Vogt BA and Gabriel M. Boston,MA: Birkhauser, 1993, p. 415–441.

DUM RP AND STRICK PL. Spinal cord terminations of the medial wall motorareas in macaque monkeys. J Neurosci 16: 6513–6525, 1996.

ERIKSEN BA AND ERIKSEN CW. Effects of noise letters upon the identificationof a target letter in nonsearch task. Percept Psychophys 16: 143–149, 1974.

EVANS AC, COLLINS DL, AND MILNER B. An MRI-based stereotaxic atlas from250 young normal subjects. Soc Neurosci Abstr 18: 408, 1992.

FINK GR, FRACKOWIAK RSJ, PIETRZYK U, AND PASSINGHAM RE. Multiplenonprimary motor areas in the human cortex. J Neurophysiol 77: 2164–2174, 1997.

GEMBA H, SASAKI K, AND BROOKS VB. “Error” potentials in limbic cortex(anterior cingulate area 24) of monkeys during motor learning. Neurosci Lett70: 223–227, 1986.

GERLOFF C, CORWELL B, CHEN R, HALLET M, AND COHEN LG. Stimulationover the human supplementary motor area interferes with the organization offuture elements in complex motor sequences. Brain 120: 1587–1602, 1997.

GOBEL S, RUSHWORTH MFS, AND WALSH V. The mental number line and thehuman angular gyrus. NeuroImage 14: 1278–1289, 2000.

HADLAND KA AND RUSHWORTH MFS. Anatomical localization of distinct timelimited processes in set switching. Soc Neurosci Abstr 30: 836.2, 2000.

HADLAND KA, RUSHWORTH MFS, PASSINGHAM RE, JAHANSHAHI M, AND ROTH-WELL J. An rTMS comparison of the dorsalateral prefrontal cortex and thepre-SMA in working memory and willed action. Soc Neurosci Abstr 29:753.5, 1999.

HADLAND KA, RUSHWORTH MFS, PASSINGHAM RE, JAHANSHAHI M, AND ROTH-WELL J. Interference with performance of a response selection task has noworking memory component: an rTMS comparison of the dorsolateralprefrontal and medial frontal cortex. J Cognit Neurosci 13: 1097–1108,2001.

HALLETT M. Transcranial magnetic stimulation and the human brain. Nature13: 147–150, 2000.

HARMER CJ, THILO KV, ROTHWELL JC, AND GOODWIN GM. Transcranialmagnetic stimulation of medial-frontal cortex impairs the processing ofangry facial expressions. Nat Neurosci 4: 17–18, 2001.

HE S-Q, DUM RP, AND STRICK PL. Topographic organization of corticospinalprojections from the frontal lobe: motor areas on the medial surface of thehemisphere. J Neurosci 15: 3284–3306, 1995.

HERSHENSON M. Reaction time as a measure of intersensory facilitation. J ExpPsychol 62: 289–293, 1962.

HIKOSAKA O, KATO RAND M, MIYACHI S, AND MIYASHITA K. Learning ofsequential movements in the monkey: process of learning and retention ofmemory. J Neurophysiol 74: 1652–1661, 1995.

HIKOSAKA O, SAKAI K, MIYAUCHI S, TAKINO R, SASAKI Y, AND PUTZ B.Activation of human presupplementary motor area in learning of sequentialprocedures: a functional MRI study. J Neurophysiol 76: 617–621, 1996.

IKEDA A, LUDERS HO, BURGESS RC, AND SHIBASAKI H. Movement-relatedpotentials recorded from supplementary motor area and primary motor area.Brain 115: 1017–1043, 1992.

JAHANSHAHI M AND ROTHWELL J. Transcranial magnetic stimulation studies ofcognition: an emerging field. Exp Brain Res 131: 1–9, 2000.

JUEPTNER M, FRITH DC, BROOKS DJ, FRACKOWIAK RSJ, AND PASSINGHAM RE.Anatomy of motor learning. II. Subcortical structures and learning by trial

and error. J Neurophysiol 77: 1325–1337, 1997a.JUEPTNER M, STEPHAN KM, FRITH CD, BROOKS DJ, FRACKOWIAK RSJ, AND

PASSINGHAM RE. Anatomy of motor learning. I. Frontal cortex and attentionto action. J Neurophysiol 77: 1313–1324, 1997b.

KOSKI L AND PAUS T. Functional connectivity of the anterior cingulate cortexwithin the human frontal lobe: a brain mapping meta-analysis. Exp BrainRes 133: 55–65, 2000.

KRAMS M, RUSHWORTH MFS, DEIBER MP, FRACKOWIAK RSJ, AND PASSINGHAM

RE. The preparation, suppression, and execution of copied movements in thehuman brain. Exp Brain Res 120: 386–398, 1998.

KURATA K AND HOFFMAN DS. Differential effects of muscimol microinjectioninto dorsal and ventral aspects of the premotor cortex of monkeys. J Neu-rophysiol 71: 1151–1164, 1994.

KURATA K AND HOSHI E. Reacquisition deficits in prism adaptation aftermuscimol microinjection into the ventral premotor cortex. J Neurophysiol81: 1927–1938, 1999.

LEE K-M, CHANG L-H, AND ROH J-K. Subregions within the supplementarymotor area activated at different stages of movement preparation and exe-cution. Neuroimage 9: 117–123, 1999.

LEUNG H-C, SKUDLARSKI P, GATENBY JC, PETERSEN BS, AND GORE JC. Anevent-related functional MR study of the Stroop Color Word InterferenceTask. Cereb Cortex 10: 552–560, 2000.

LOGOTHETIS NK, PAULS J, AUGATH M, TRINATH T, AND OELTERMANN A.Neurophysiological investigation of the basis of the fMRI signal. Nature412: 150–157, 2001.

LU M-T, PRESTON JB, AND STRICK PL. Interconnections between the prefrontalcortex and the premotor areas in the frontal lobe. J Comp Neurol 341:375–392, 1994.

LUPPINO G, MATELLI M, CAMARDA RM, GALLESE V, AND RIZZOLATTI G.Multiple representations of body movements in mesial area 6 and theadjacent cingulate cortex: an intracortical microstimulation study in themacaque monkey. J Comp Neurol 311: 463–482, 1991.

LUU P, FLAISCH T, AND TUCKER DM. Medial frontal cortex in action monitor-ing. J Neurosci 20: 464–469, 2000.

MACDONALD AW, COHEN JD, STENGER VA, AND CARTER CS. Dissociating therole of the dorsolateral prefrontal and anterior cingulate cingulate cortex incognitive control. Science 288: 1835–1838, 2000.

MACLEOD CM. Half a century of research on the Stroop effect: an interactivereview. Psychol Bull 109: 163–203, 1991.

MARZI CA, MINIUSSI C, MARAVITA A, BERTOLASI L, ZANETTE G, ROTHWELL

JC, AND SANES JN. Transcranial magnetic stimulation seletively impairsinterhemispheric transfer of visuo-motor information in humans. Exp BrainRes 118: 435–438, 1998.

MATSUZAKA Y, AIZAWA H, AND TANJI J. A motor area rostral to the supple-mentary motor area (presupplementary motor area) in the monkey: neuronalactivity during a learned motor task. J Neurophysiol 68: 653–662, 1992.

MATSUZAKA Y AND TANJI J. Changing directions of forthcoming arm move-ments: neuronal activity in the presupplementary and supplementary motorarea of monkey cerebral cortex. J Neurophysiol 76: 2327–2342, 1996.

MEIRAN N. Modelling cognitive control in task-switching. Psychol Res 63:234–249, 2000

MEIRAN N, CHOREV Z, AND SAPIR A. Component processes in task switching.Cognit Psychol 41: 211–53, 2000

MONSELL S, YEUNG N, AND AZUMA R. Reconfiguration of task-set: is it easierto switch to the weaker task? Psychol Res 63: 250–264, 2000

NAITO E, KINOMURA S, GEYER S, KAWASHIMA R, ROLAND RE, AND ZILLES K.Fast reaction to different sensory modalities activates common fields in themotor areas, but the anterior cingulate cortex is involved in the speed ofreaction. J Neurophysiol 83: 1701–1709, 2000.

NAKAMURA K, SAKAI K, AND HIKOSAKA O. Neuronal activity in the medialfrontal cortex during learning of sequential procedures. J Neurophysiol 80:2671–2687, 1998.

NAKAMURA K, SAKAI K, AND HIKOSAKA O. Effects of local inactivation ofmonkey medial frontal cortex in learning of sequential procedures. J Neu-rophysiol 82: 1063–1068, 1999.

PARDO JV, FOX PT, AND RAICHLE ME. Localization of a human system forsustained attention by positron emission tomography. Nature 349: 61–64,1991.

PASCUAL-LEONE A, BARTRES-FAZ D, AND KEENAN JP. Transcranial magneticstimulation: studying the brain-behaviour relationship by induction of “vir-tual lesions.” Philos Trans Roy Soc Lond B Biol Sci 354: 1229–1238. 1999.

PASCUAL-LEONE A, WALSH V, AND ROTHWELL V. Transcranial magnetic stim-ulation in cognitive neuroscience—virtual lesion, chronometry, and func-tional connectivity. Curr Opin Neurobiol 10: 232–237, 2000.

PASSINGHAM RE. Premotor cortex and preperation for movement. Exp BrainRes 70: 590–596, 1988.

PASSINGHAM RE. The Frontal Lobes and Voluntary Action. Oxford PsychologySeries 21. Oxford, UK: Oxford Univ. Press, 1993.

PASSINGHAM RE. The status of the premotor areas: evidence from PET scan-ning. In: Neural Control of Movement, edited by Ferrell WR and Proske U.New York: Plenum, 1995, p. 167–178.

PASSINGHAM RE. Attention to action. In: The Prefrontal Cortex: Executive andCognitive Functions, edited by Roberts AC, Robbins TW, and WeiskrantzL. Oxford, UK: Oxford Univ. Press, 1998, p. 131–143.

PAUS T. Imaging the brain before, during, and after transcranial magneticstimulation. Neuropsychologia 37: 219–224, 1999.



PAUS T. Primate anterior cingulate cortex: where motor control, drive andcognition interface. Nat Rev Neurosci 2: 417–24. 2001.

PAUS T, JECH R, THOMPSON CJ, COMEAU R, PETERS T, AND EVANS AC.Transcranial magnetic stimulation during positron emission tomography: anew method for studying connectivity of the human cerebral cortex. J Neu-rosci 17: 3178–3184, 1997.

PAUS T, KOSKI L, CARAMANOS Z, AND WESTBURY C. Regional differences inthe effects of task difficulty and motor output on blood flow response to inthe human anterior cingulate cortex: a review of 107 PET activation studies.Neuroreport 9: R37–R47, 1998.

PAUS T, OKAY N, CARAMANOS Z, MACDONALD D, ZIJBENDOS A, D’AVIRRO D,GUTMANS D, HOLMES C, TOMAIUOLO F, AND EVANS AC. In vivo mormom-etry of the intrasulcal gray matter in the human cingulate, paracingulate, andsuperior-rostral sulci: hemispheric asymetries, gender differences and prob-ability maps. J Comp Neurol 376: 664–673, 1996a.

PAUS T, PETRIDES MJ, EVANS AC, AND MEYER E. Role of the human anteriorcingulate cortex in the control of oculomotor, manual, and speech responses:a positron emission tomography study. J Neurophysiol 70: 453–469, 1993.

PAUS T, TOMAIUOLO F, OTAKY N, MACDONALD D, PETRIDES M, ATLAS J,MORRIS R, AND EVANS AC. Human cingulate and paracingulate sulci:pattern, variability, asymetry, and probabilistic map. Cereb Cortex 6: 207–214, 1996b.

PAUS T AND WOLFORTH M. Transcranial magnetic stimulation during PET:reaching and verifying the target site. Hum Brain Map 6: 399–402, 1998.

PETERS T, DAVEY B, MUNGER P, COMEAU R, EVANS A, AND OLIVIER A.Three-dimensional multimodal image guidance for neurosurgery. IEEETrans Med Imag 15: 121–128, 1996.

PETRIDES M. The effect of periarcuate lesions in the monkey on performanceof symmetrically and asymetrically visual and auditory go, no-go tasks.J Neurosci 6: 2054–2063, 1986.

PICARD N AND STRICK PL. Motor areas of the medial wall: a review of theirlocation and functional activation. Cereb Cortex 6: 342–353, 1996.

POSNER MI AND DIGIROLAMO GJ. Executive attention: conflict, target detec-tion, and cognitive control. In: The Attentive Brain, edited by ParasuramanR. Cambridge, MA: MIT Press, 1998, p. 401–423.

PRAAMSTRA P, KLEINE BU, AND SCHNITZLER A. Magnetic stimulation of thedorsal premotor cortex modulates the Simon effect. Neuroreport 10: 3671–3674, 1999.

REA GL, EBNER TJ, AND BLOEDEL JR. Evaluations of combined premotor andsupplementary motor cortex lesions on a visually guided arm movement.Brain Res 418: 58–67, 1987.

ROGERS RD AND MONSELL S. Costs of predictable switch between simplecognitive tasks. J Exp Psychol Gen 124: 207–231, 1995.

RUSHWORTH MFS, ELLISON A, AND WALSH V. Complementary localization andlateralization of orienting and motor attention. Nat Neurosci 4: 656–661,2001a.

RUSHWORTH MFS, HADLAND KA, GAFFAN D, AND PASSINGHAM RE. Thecingulate cortex in social behaviour and cognition. Soc Neurosci Abstr 26:205.1, 2000.

RUSHWORTH MFS, KRAMS M, AND PASSINGHAM RE. Complementary lateral-ization of attention and intention in the human brain. A PET study. J CognitNeurosci 13: 698–710, 2001b.

RUSHWORTH MFS, PAUS T, AND SIPILA PK. Attention systems and the organi-zation of the human parietal cortex. J Neurosci 21: 5262–5271, 2001c.

SAKAI K, HIKOSAKA O, MIYAUCHI S, SASAKI Y, FUJIMAKI N, AND PUTZ B.Presupplementary motor area activation during sequence learning reflectsvisuo-motor association. J Neurosci 19: RC1–RC6, 1999.

SAKAI K, HIKOSAKA O, MIYAUCHI S, TAKINO R, SASAKI Y, AND PUTZ B.Transition of brain activation from frontal to parietal areas in visuomotorsequence learning. J Neurosci 18: 1827–1840, 1998.

SAKAI K, HIKOSAKA O, TAKINO R, MIYAUCHI S, NIELSEN M, AND TAMADA T.What and when: parallel and convergent processing in motor control.

J Neurosci 20: 2691–2700, 2000.SCHLUTER ND, RUSHWORTH MFS, MILLS KR, AND PASSINGHAM RE. Signal,

set, and movement related activity in human lateral premotor cortex. A studyusing transcranial magnetic stimulation. Neuropsychologia 37: 233–243,1999.

SCHLUTER N, RUSHWORTH MFS, PASSINGHAM RE, AND MILLS K. Temporaryinteference in human lateral premotor cortex suggests dominance for theselection of movements: a study using transcranial magnetic stimulation.Brain 121: 785–799, 1998.

SHIMA K, AYA K, MUSHIAKE H, INASE M, AIZAWA H, AND TANJI J. Twomovement related foci in the primate cingualate cortex observed in signal-triggered and self-paced forelimb movements. J Neurophysiol 65: 188–202,1991.

SHIMA K, MUSHIAKE H, SAITO N, AND TANJI J. Role for cells in the presupple-mentary motor area in updating motor plans. Proc Natl Acad Sci USA 93:8694–8698, 1996.

SHIMA K AND TANJI J. Role for cingulate motor area cells in voluntarymovement selection based on reward. Science 282: 1335–1338, 1998.

SOHN MH, URSU S, ANDERSON JR, STENGER VA, AND CARTER CS. The role ofprefrontal cortex and posterior parietal cortex in task switching. Proc NatlAcad Sci USA 97: 13448–53, 2000.

STEPHAN KM, FINK GR, PASSINGHAM RE, SILBERSWEIG D, CEBALLOS-BAU-MANN AO, FRITH CD, AND FRACKOWIAK RSJ. Functional anatomy of themental representation of upper extremity movements in healthy subjects.J Neurophysiol 73: 373–386, 1995.

STROOP JR. Studies of interference in spatial verbal reactions. J Exp Psychol18: 643–662, 1935.

TALAIRACH J AND TOURNOUX P. Co-Planar Stereotaxic Atlas of the HumanBrain. Stuttgart, Germany: Thieme, 1988.

TAYLOR SF, KORNBLUM S, LAUBER EJ, MINOSHIMA S, AND KOEPPE RA.Isolation of specific inteference processing in the Stroop task: PET activa-tion studies. Neuroimage 6: 81–92, 1997.

TAYLOR SF, KORNBLUM S, MINOSHIMA S, OLIVERS LM, AND KOEPPE RA.Changes in medial cortical blood flow with a stimulus-response compati-bility task. Neuropsychologia 32: 249–255, 1994.

TERAO Y, FUKUDA H, UGAWA Y, HIKOSAKA O, HANAJIMA R, FURUBAYASHI T,SAKAI K, MIYAUCHI S, SASAKI Y, AND KANAZAWA I. Visualization of theinformation flow through human oculomotor cortical regions by transcranialmagnetic stimulation. J Neurophysiol 80: 936–946, 1998.

TONI I, SCHLUTER ND, JOSEPHS O, FRISTON K, AND PASSINGHAM RE. Signal-,set- and movement-related activity n the human brain: an event-relatedfMRI study. Cereb Cortex 9: 35–49, 1999.

TURKEN AU AND SWICK D. Response selection in the human anterior cingulatecortex. Nat Neurosci 2: 920–924, 1999.

WALSH V, ASHBRIDGE E, AND COWEY A. Cortical plasticity in perceptuallearning demonstrated by transcranial magnetic stimulation. Neuropsycho-logia 36: 45–49, 1998.

WALSH V AND COWEY A. Magnetic stimulation studies of visual cognition.Trends Cognit Sci 2: 103–110, 1998.

WALSH V AND COWEY A. Transcranial magnetic stimulation and cognitiveneuroscience. Nat Rev Neurosci 1: 73–79, 2000.

WALSH V, ELLISON A, ASHBRIDGE E, AND COWEY A. The role of the parietalcortex in visual attention—hemispheric asymmetries and the effects oflearning: a magnetic stimulation study. Neuropsychologia 37: 245–51, 1999.

WALSH V, ELLISON A, BATTELLI L, AND COWEY A. Task-specific impairmentsand enhancements induced by magnetic stimulation of human visual areaV5. Proc Roy Soc Lond B Biol Sci 265: 537–543, 1998.

WALSH V AND RUSHWORTH MFS. The use of transcranial magnetic stimulationin neuropsychological testing. Neuropsychologia 37: 125–135, 1999.

WISE SP. The primate premotor cortex: past, present, and preparatory. AnnuRev Neurosci 8: 1–20, 1985.

WISE SP, DI PELLIGRINO G, AND BOUSSAOUD D. The premotor cortex andnon-standard sensorimotor mapping. Can J Physiol Pharmacol 74: 469–482, 1996.

WISE SP AND MAURITZ K-H. Set-related neuronal activity in the premotorcortex of rhesus monkeys: effect of changes in motor set. Proc Roy Soc LondB Biol Sci 223: 331–354, 1985.

WORSLEY KJ, ANDERMANN M, KOULIS T, MACDONALD D, AND EVANS AC.Detecting changes in non-isotropic images. Hum Brain Map 8: 98–101,1999.

WORSELY KJ, LIAO CH, ASTON J, PETRE V, DUNCAN GH, MORALES F, AND

EVANS AC. A general statistical analysis for fMRI data. Neuroimage 15:1–15, 2002.

WORSLEY KJ, MARRETT S, NEELIN P, VANDAL AC, FRISTON KJ, AND EVANS

AC. A unified statistical approach for determing significant signals inimages of cerebral activation. Hum Brain Map 4: 58–73, 1996.

ZARAHN E, AGUIRRE G, AND D’ESPOSITO M. A trial-based experimental designfor fMRI. Neuroimage 6: 122–138, 1997.



role of the human medial frontal cortex in task … of the human medial frontal cortex in task...

Documents