BOLD MRI responses to repetitive TMS over human dorsal premotor cortex

www.elsevier.com/locate/ynimg

NeuroImage 28 (2005) 22 – 29

BOLD MRI responses to repetitive TMS over human

dorsal premotor cortex

Sven Bestmann,a,b,* Jurgen Baudewig,a Hartwig R. Siebner,c

John C. Rothwell,b and Jens Frahma

aBiomedizinische NMR Forschungs GmbH am Max-Planck-Institut fur Biophysikalische Chemie, 37070 Gottingen, GermanybSobell Department of Motor Neuroscience and Movement Disorders, Institute of Neurology, University College of London, UKcInstitute of Neurology, University of Kiel, Germany

Received 20 June 2004; revised 20 April 2005; accepted 20 May 2005

Available online 5 July 2005

Functional magnetic resonance imaging (fMRI) studies in humans have

hitherto failed to demonstrate activity changes in the direct vicinity of

transcranialmagnetic stimulation (TMS) that cannot be attributed to re-

afferent somatosensory feedback or a spread of excitation. In order to

investigate the underlying activity changes at the site of stimulation as

well as in remote connected regions, we applied short trains of high-

intensity (110% of resting motor threshold) and low-intensity (90% of

activemotor threshold) repetitive TMS (rTMS; 3 Hz, 10 s duration) over

the presumed location of the left dorsal premotor cortex (PMd) during

fMRI. Signal increases in the direct vicinity of the stimulated PMd were

observed during rTMS at 110% RMT. However, positive BOLD MRI

responses were observed with rTMS at both 90% and 110% RMT in

connected brain regions such as right PMd, bilateral PMv, supplemen-

tary motor area, somatosensory cortex, cingulate motor area, left

posterior temporal lobe, cerebellum, and caudate nucleus. Responses

were generally smaller during low-intensity rTMS. The results indicate

that short trains of TMS can modify local hemodynamics in the absence

of overt motor responses. In addition, premotor rTMS cannot only

effectively stimulate cortico-cortical but also cortico-subcortical con-

nections even at low stimulation intensities.

D 2005 Elsevier Inc. All rights reserved.

Keywords: fMRI; EMG; SMA; BOLD; Echo-planar imaging; Motor cortex

Introduction

A number of recent combined transcranial magnetic stimulation

(TMS) and functional magnetic resonance imaging (fMRI) studies

have confirmed that TMS leads to changes in neural activity in

structures beyond the site of stimulation (Bohning et al., 1998,

1053-8119/$ - see front matter D 2005 Elsevier Inc. All rights reserved.

doi:10.1016/j.neuroimage.2005.05.027

* Corresponding author. Biomedizinische NMR Forschungs GmbH am

Max-Planck-Institut fur Biophysikalische Chemie, 37070 Gottingen,

Germany. Fax: +49 551 201 1307.

E-mail address: [email protected] (S. Bestmann).

Available online on ScienceDirect (www.sciencedirect.com).

1999, 2000b; Bestmann et al., 2004). Although similar effects have

been observed with positron emission tomography (PET), the

ability to perform similar experiments in fMRI allows us safely to

make repeated studies on the same individuals at higher temporal

and spatial resolution than can be obtained in PET (Bohning et al.,

1998, 1999, 2000b; Shastri et al., 1999; Baudewig et al., 2001;

Nahas et al., 2001; Bestmann et al., 2004). However, despite their

advantages, combined TMS-fMRI studies have produced results

that have led to one puzzling question: what happens at the site of

stimulation itself? A problem has been that most studies have used

TMS over motor cortex, and stimulation here can lead to

contralateral movements and sensory feedback that additionally

contribute to BOLD MRI signal changes. In fact, stimulation of

motor cortex at sub-movement threshold intensities fails to change

BOLD activity under the TMS coil, even though changes can be

seen at other distant connected sites (Bohning et al., 1998, 1999,

2000b; Baudewig et al., 2001; Kemna and Gembris, 2003;

Bestmann et al., 2004). Interestingly, no significant changes in

activity have been reported at sites of stimulation in the parietal and

premotor cortex in a recent repetitive TMS (rTMS) study even

when using intensities well above the threshold needed to activate

the primary motor cortex (Kemna and Gembris, 2003). One

possible explanation is that the threshold for motor cortex

stimulation is not the same as the threshold in non-motor regions

(Stewart et al., 2001). Thus, the absence of significant local activity

changes in these experiments may simply reflect the fact that the

intensity or duration was not sufficient to cause significant BOLD

MRI response changes.

Here, we performed TMS-fMRI over the presumed location of

the dorsal premotor cortex in order to investigate the occurrence of

activity changes in the direct vicinity of the stimulation site as well as

remote brain regions. This stimulation site has an advantage over

motor cortex stimulation in that it does not provoke movements of

contralateral muscles and therefore avoids the confounding BOLD

signals that result from reafferent feedback. Although the precise site

of stimulation with the TMS coil is not well defined, the effects of

S. Bestmann et al. / NeuroImage 28 (2005) 22–29 23

TMS over this site differ from those seen after stimulation over the

motor cortex in both their physiological effects on corticocortical

connectivity (Civardi et al., 2001; Mochizuki et al., 2004; Rizzo et

al., 2004) as well as their metabolic consequences as recorded with

PET (Chouinard et al., 2003; Siebner et al., 2003). For these reasons,

we have referred to the site as dorsal premotor cortex since it is

anterior and slightly medial to the TMS ‘‘hot spot’’ for provoking

EMG activity in contralateral hand muscle.

Using the premotor site as a model, we asked (a) whether TMS-

fMRI is capable of visualizing activity changes in the direct

vicinity of the TMS probe and (b) whether secondary motor and

non-motor regions that are connected to PMd show activity

changes during TMS-fMRI. We show that rTMS over premotor

cortex activates a large network of remote interconnected cortical

and subcortical motor regions. The results furthermore demonstrate

for the first time that short TMS trains modulate local hemody-

namic responses in the absence of overt muscle movements.

Materials and methods

Subjects

We studied nine right-handed healthy subjects without any

previous personal or family neuropsychiatric or neurological

history (mean age 29 years, range 25–42 years; six female) after

receiving written informed consent. Examinations conformed to

the standards laid down by the Declaration of Helsinki and local

ethics board approval.

Experimental procedures

In each of three experimental conditions, eight stimulation

epochs (9.96 s) alternated with resting periods (23.24 s). Subjects

were instructed to keep their eyes closed and to relax their hands

while repetitive TMS was applied at 3 Hz using two different

intensities: 110% of individual resting motor threshold (RMT) and

90% of individual active motor threshold (AMT) (mean: 92%

(73%) of the maximal output of the stimulator, range: 82–100%

(65–81%). The high stimulation intensity values are caused by the

introduction of the long connecting cable to the TMS coil allowing

it to be used within the scanner. This increases the resistance and

inductance of the circuit so that higher intensities of stimulator

output are required to generate the same maximum current flow.

Using the same protocol timings, participants dorsiflexed the right

index finger in order to locate areas of increased BOLD MRI signal

related to focal hand movement. The time of each movement was

cued by applying (ineffective) rTMS at 15% of stimulator output

(mean: 21% of AMT, range: 19–23%). The order of experimental

conditions was pseudo-randomized in order to avoid the same

TMS condition occurring in successive scans (five subjects started

with suprathreshold TMS). The three experimental runs were

separated by 5 min to reduce carry-over effects of stimulation. The

number and intensity of TMS pulses conformed to presently

available safety guidelines for TMS (Wassermann, 1998).

Magnetic resonance imaging

MRI was performed at 2.9 T (Siemens Trio, Erlangen,

Germany) using a standard transmit-receive head-coil. Anatomical

images were acquired using a short-echo time 3D FLASH

sequence (TR/TE = 11/4.92 ms, flip angle 15-, 1 mm isotropic

resolution) covering the whole head. Functional MRI was

conducted using a T2*-weighted single-shot gradient-echo EPI

sequence (frequency-selective fat suppression, TR/TE = 3320/36

ms, flip angle 70-, 2 � 2 mm2 resolution, 4 mm section thickness).

Twenty oblique brain sections covered a region from the primary

and secondary motor cortex down to the basal ganglia and the

thalamus. Each fMRI run lasted 4 min 59 s, corresponding to 90

volumes including periods for signal equilibration.

Transcranial magnetic stimulation and electromyographic

recordings

TMS was conducted using a non-ferromagnetic figure-of-eight

coil (two windings of ten turns each; inner winding diameter 53

mm; distance between outer coil surface and windings: 2–3 mm

(variation due to manufacturing tolerance); coil inductance includ-

ing cable: 20 AH; maximal current at 100% stimulator output:

approximately 5 kA). The coil was connected to a Magstim Rapid

stimulator (The Magstim Company, Wales, UK) outside the

radiofrequency-shielded cabin via an 8 m cable inserted through

a filter tube. This ensured that no other material other than the

MRI-compatible TMS coil and the connecting cable could be

introduced into the scanner room. The TMS coil was mechanically

strengthened for the requirements at 3 T by inclusion of an

appropriately shaped 8 mm plastic former on the reverse side of the

coil which was attached to the coil by high strength fiber tape (The

Magstim Company). Each subject wore earplugs and headphones

to reduce acoustic noise from the discharging TMS coil.

Biphasic electrical pulses (approximate rise time based on full

coil inductance including connecting cable and coil windings:

approximately 90 As with a duration of 250 As) were applied with

an initial anteroposterior direction of current flow. TMS-fMRI was

synchronized using Presentation 0.51 software (Neurobehavioral

Systems, Inc., San Francisco, USA).

Before scanning, the optimal position of the coil for evoking

movements of the contralateral hand (referred to as motor hot spot)

was determined. Individual motor thresholds were then obtained

with single TMS pulses. RMT was defined as the minimum

intensity that evoked a muscle twitch in 5 out of 10 trials that was

clearly observable to two of the experimenters (S.B., J.B.). The

minimum intensity that caused an observable muscle twitch during

10% of maximum voluntary contraction in 5 out of 10 trials was

defined as AMT.

The site for premotor stimulation was defined with reference to

the motor hot spot according to previously published procedures

(Schluter et al., 1998, 1999; Johansen-Berg et al., 2002). The coil

was moved 2 cm anterior and 1 cm medial and fixed using a

custom-made adjustable coil holder (Polyetheretherketon, PEEK)

mounted onto the MRI head coil. In two subjects, suprathreshold

TMS at the target site evoked electromyographic (EMG) responses

in the relaxed right first dorsal interosseous (FDI) muscle, and the

coil was moved 0.5 and 1 cm anterior to the target site,

respectively. This was motivated by the fact that the presumed

area of stimulation along the junction of the TMS coil wings spans

a region of about 4 cm (Barker, 1999) and thus the motor hot spot

was still within the target region of the PMd. Moreover, previous

studies have shown a good correspondence of scalp positions

between 2 and 3 cm anterior to the motor hot spot and the

underlying dorsal premotor complex (Johansen-Berg et al., 2002;

Chen et al., 2003; Siebner et al., 2003; Oxley et al., 2004).

S. Bestmann et al. / NeuroImage 28 (2005) 22–2924

Consequently, we aimed to minimize spread of excitation and overt

muscle movements while at the same time retaining the TMS probe

over the presumed dorsal premotor region.

During fMRI, EMG responses from the right FDI muscle were

continuously monitored to assess a possible spread of excitation

into the adjacent M1. Surface EMG was recorded with non-

ferromagnetic sintered Ag/AgCl surface electrodes (9 mm diam-

eter) using a belly-tendon montage. The electrodes were connected

to a BrainAmp MRI-compatible EEG amplifier (Brain Products,

Munich, Germany). EMG data were recorded at a sampling rate of

2000 Hz, rectified, and low-pass filtered at 50 Hz using Brain

Vision 1.0 (Brain Products, Munich, Germany). Following the

removal of MRI gradient artefacts, the residual EMG allowed for a

qualitative detection of electromyographic activity. Recordings

from each experimental run were normalized to the mean EMG

amplitude and paired t tests with Bonferroni correction for multiple

comparisons were used to compare EMG levels during TMS and

resting epochs.

Combined TMS-fMRI

Combined TMS-fMRI was accomplished as described previ-

ously (Bohning et al., 1998; Shastri et al., 1999; Bestmann et al.,

2003a, 2004). Functional images were acquired every 166 ms,

while each image acquisition lasted for 91 ms and TMS pulses

were applied immediately afterwards. The chosen stimulation

frequency of 3 Hz (every 332 ms) provided waiting periods

between TMS pulses and subsequent image acquisitions of 75 ms

which allowed for unperturbed fMRI as confirmed in pilot

experiments.

Data analysis

Image processing and statistical analyses were carried out using

BrainVoyager 2000 (Brain Innovation, Maastricht, The Nether-

lands). Realignment, intensity normalization, spatial smoothing

and linear drift removal was performed prior to statistical analysis

following procedures described elsewhere (Goebel et al., 1998).

2D functional slice-time data were co-registered with anatomical

T1-weighted images from the same session. Functional images

were transformed into 3 mm isotropic resolution. Individual pre-

processed volume time courses were analyzed using the general

linear model (GLM) with stimulation or movement epochs as the

effects of interest. Multi-subject analysis was conducted using a

fixed effects model to test for significant changes in BOLD MRI

signal during each experimental condition at a group level (P <

0.01 adjusted for multiple comparisons). Coordinates of the center

of gravity of activation clusters were determined with reference to

the MNI template (Montreal Neurological Institute) in stereotaxic

space (Talairach and Tournoux, 1988). EMG recordings from each

session were normalized to the mean EMG amplitude and mean

normalized EMG levels during TMS and resting epochs were

compared using paired samples t tests.

Fig. 1. Electromyographic (EMG) recordings of (a) voluntary finger tapping

(right hand) and (b) suprathreshold rTMS of the left PMd from a

representative subject recorded during simultaneous fMRI. Stimulation

epochs (boxes) were time-locked averaged to the onset of stimulation.

Absence of EMG activity in the contralateral first dorsal interosseous (FDI)

muscle during suprathreshold rTMS indicated that stimulation at 110%

resting motor threshold did not spread into the primary motor hand area.

Results

None of the subjects reported any side effects from the

experimental procedure when asked immediately after the experi-

ment. No additional interrogation was conducted subsequently.

Compared with resting epochs, no significant EMG activity was

observed during suprathreshold rTMS in any of the subjects

(paired samples t test, P = 0.24, df = 8), whereas voluntary finger

movement epochs evoked clear EMG responses (P < 0.01, df = 8).

Fig. 1 presents EMG recordings of a representative subject.

Repetitive TMS at 110% RMT produced a localized BOLD

MRI signal increase in the left PMd (mean coordinates of activated

cluster: x = �40, y = �11, z = 54; see Fig. 2 and Table 1).

Additional activity increases were found in the homologous PMd,

the bilateral ventral premotor region, the supplementary motor area

(SMA) in the medial aspect of the superior frontal gyrus, and the

putative cingulate motor area. In contrast, no significant changes

were found in the left primary sensorimotor cortex. Repetitive

TMS at 110% RMT was also associated with increases in BOLD

MRI signal in the left posterior middle temporal gyrus, and large

parts of the putative bilateral auditory cortex including the superior

temporal plane, superior temporal gyrus, and planum temporale,

extending dorsally into putative SII (Fig. 2, Table 1). Subcortically,

bilateral activity was found in the middle part of the caudate

nucleus, as well as in the thalamus of the left hemisphere and the

bilateral inferior colliculi. In addition, focal BOLD MRI responses

were found in the medio-dorsal cerebellum, with a right-hemi-

spheric preponderance (Table 1).

As shown in Fig. 3 and Table 2, there was a close spatial

correspondence between brain regions activated during voluntary

finger movement and rTMS of the left PMd. This was confirmed

by a conjunction analysis which tested for brain regions showing

an increase in BOLD MRI signal during both voluntary finger

movement and rTMS at 110% RMT (Table 2). The results

demonstrate that activations common to both experimental con-

ditions comprised large parts of motor system known to be

involved in manual motor control except for the left M1 which was

only found to be activated during voluntary movement.

The brain regions that revealed an increase in BOLD MRI

signal during subthreshold rTMS are listed in Table 3. In contrast

to suprathreshold stimulation, no significant response was

observed in the targeted presumed PMd (Fig. 5a). BOLD MRI

signal increases in distant brain regions to low-intensity rTMS

were highly co-localized with changes during stimulation at 110%

RMT. However, the extent of activation clusters as well as peak

Table 1

Activity evoked by suprathreshold rTMS over the left PMd

Anatomical/functional location xa ya za t valueb Volumec

L premotor (dorsal) �40 �11 54 8.24 438

L premotor (ventral) �49 2 17 10.29 1339

L auditory cortexd �45 �25 15 16.42 7836

L middle temporal gyrus �50 �52 3 9.26 1299

L thalamus �14 �17 8 6.04 150

L caudate nucleus �11 �4 18 6.72 68

L putamen �27 �11 3 5.82 276

L cerebellar hemisphere �6 �59 �20 6.99 355

L inferior colliculus �3 �20 �2 7.00 56

BL medial superior frontal

gyrus (SMA)

3 �5 55 8.73 898

BL cingulate gyrus 3 4 44 9.26 908

R premotor (dorsal) 43 �3 49 8.92 367

R premotor (dorsal) 42 7 26 7.64 238

R premotor (ventral) 54 0 13 9.14 1127

R auditory cortexd 50 �22 15 15.72 7311

R thalamus 12 �14 8 6.97 172

R caudate nucleus 12 0 14 7.42 630

R inferior colliculus 6 �25 �6 7.80 93

R cerebellar hemisphere 5 55 �15 8.38 701

a Coordinates correspond to center of gravity of respective activation

clusters.b Peak activation within cluster with P < 0.01, corrected.c Rescaled to voxel size 1 � 1 � 1 mm.d Including superior temporal gyrus, ventral part of parietal operculum

and planum temporale. L: left, R: right, BL: bilateral.Fig. 2. BOLD MRI responses to suprathreshold rTMS of left premotor

cortex (group analysis, n = 9, P < 0.01, corrected). (a) Sagittal (x = �40),

coronal ( y = �11), and transverse (z = 55) view of activity in the left PMd.

(b) Six transverse sections showing activity changes in the cingulate gyrus,

PMv, auditory cortex, caudate nucleus, left posterior temporal lobe, medial

geniculate nucleus, and cerebellum with coordinates indicated. Activation

maps are projected onto a template brain (Montreal Neurological Institute,

MNI). L: left, R: right.

S. Bestmann et al. / NeuroImage 28 (2005) 22–29 25

activity within activation clusters were markedly reduced during

low-intensity rTMS (Tables 1, 3). A direct comparison revealed

that stimulation at 110% RMT was associated with a stronger

BOLD MRI signal at the site of stimulation and in distant brain

regions than low-intensity rTMS, including lateral and mesial

premotor regions, left posterior temporal gyrus, and the thalamus

bilaterally (Fig. 4, Table 4). Neither high- nor low-intensity rTMS

evoked significant BOLD MRI signal changes in the caudally

adjacent M1 hand area (Fig. 5b).

Fig. 3. Brain regions with BOLD MRI responses to both suprathreshold

rTMS of left premotor cortex and voluntary finger movement of the right

hand (group analysis, n = 9; P < 0.01, corrected). The data are projected

onto a left-hemispheric 3D surface reconstruction (MNI) and reveal

enhanced activity in the stimulated left PMd (red circle). Additional

activity in the left hemisphere was evoked in PMv, SMA, cingulate gyrus,

auditory cortices, and left posterior middle temporal gyrus.

Discussion

The present study provides evidence that suprathreshold rTMS

over the left premotor cortex is able to elicit BOLD MRI signal

increases in the direct vicinity of the targeted region, as well as in a

range of cortical and subcortical distant brain regions. In contrast,

stimulation at 90% AMT increased activity in connected cortical

and subcortical areas but failed to induce significant BOLD MRI

changes in the directly targeted premotor region. Further ROI

analysis confirmed that no consistent change in average BOLD

MRI signal occurred under the probe in the low-intensity

condition. While the spatial pattern of distant activity changes

was comparable for both intensities, these were stronger during

rTMS at 110% RMT as compared to subthreshold stimulation.

Local effects of premotor rTMS

Previous studies reported a local increase in BOLD MRI signal

at the site of stimulation only when rTMS was applied to M1 at

intensities suprathreshold for evoking a contralateral muscle twitch

(Bohning et al., 1998, 1999, 2000a,b; Baudewig et al., 2001;

Kemna and Gembris, 2003; Bestmann et al., 2003b, 2004).

However, these effects most likely had confounding contributions

from somatosensory feedback caused by the contralateral muscle

Fig. 4. Comparison of BOLD MRI responses to supra- and subthreshold

rTMS over left premotor cortex (group analysis, n = 9; P < 0.01, corrected)

in a sagittal (a), coronal (b), and two transverse (c, d) sections of a reference

brain (MNI) with Talairach coordinates indicated. Suprathreshold rTMS

evoked significantly stronger activity changes than subthreshold rTMS at

the stimulated left PMd (red circle), the right PMd, SMA, cingulate gyrus,

right cerebellum, bilateral superior temporal plane (putative auditory cortex

and SII), bilateral caudate nucleus, and left posterior middle temporal gyrus.

Table 4

Differences in activity between supra- and subthreshold rTMS over left

Table 2

Conjunction analysis finger tapping-suprathreshold rTMS

Anatomical/functional location xa ya za t valueb Volumec

L premotor (ventral) �50 0 12 9.21 1047

L auditory cortexd �45 �27 16 13.93 6370

L middle temporal gyrus �50 �53 3 9.26 1014

L thalamus �12 �17 5 5.89 122

L putamen �26 �11 3 6.12 180

L inferior colliculus �5 �19 �3 6.47 119

BL medial superior frontal

gyrus (SMA)

2 �5 53 8.72 727

BL cingulate gyrus 3 5 44 8.39 679

R premotor (ventral) 53 1 14 9.14 620

R auditory cortexd 53 �22 15 14.11 5308

R caudate nucleus 13 �2 15 6.86 200

R cerebellar hemisphere 5 �55 �15 8.38 588

a Coordinates correspond to center of gravity of respective activation

clusters.b Peak activation within cluster with P < 0.01, corrected.c Rescaled to voxel size 1 � 1 � 1 mm.d Including superior temporal gyrus, ventral part of parietal operculum

and planum temporale. L: left, R: right, BL: bilateral.

S. Bestmann et al. / NeuroImage 28 (2005) 22–2926

movements. In the present study, short trains of rTMS at 110%

RMT induced consistent BOLD MRI signal increases in the

stimulated left PMd. This shows for the first time that rTMS

evokes a local hemodynamic response that can now be visualized

by concurrent fMRI. Our simultaneous EMG recordings indicate

that movement-evoked sensory feedback was not the source of

these effects.

Using suprathreshold rTMS, previous prefrontal rTMS-fMRI

studies have failed to provoke significant local BOLD MRI signal

changes with pulse trains of 1 s duration at 10 Hz (Baudewig et al.,

2001) or 4 Hz (Kemna and Gembris, 2003), respectively. This

suggests that not only the high intensity but also the total number

of pulses (i.e., the length of the rTMS train) is critical for inducing

significant activity changes (as indexed by the BOLD MRI signal).

A similar conclusion was reached in several previous PET studies

(Paus et al., 1998; Speer et al., 2003a,b). Dose-dependent effects of

TMS have also been found in the global mean field amplitude,

taken as an index of the overall brain response due to TMS

(Komssi et al., 2004).

Table 3

Activity evoked by subthreshold rTMS over the left PMd

Anatomical/functional location xa ya za t valueb Volumec

L premotor (ventral) �52 �3 15 7.30 811

L auditory cortexd �47 �23 14 12.86 5619

L middle temporal gyrus �48 �53 4 7.63 251

L thalamus �19 �15 4 6.03 60

BL medial superior frontal

gyrus (SMA)

�1 �6 52 6.02 75

BL cingulate gyrus 3 6 43 6.23 231

R premotor (ventral) 53 �1 11 7.79 575

R auditory cortexd 49 �26 16 12.29 5709

R thalamus 12 �16 9 6.63 103

a Coordinates correspond to center of gravity of respective activation

clusters.b Peak activation within cluster with P < 0.01, corrected.c Rescaled to voxel size 1 � 1 � 1 mm.d Including superior temporal gyrus, ventral part of parietal operculum

and planum temporale. L: left, R: right, BL: bilateral.

Evidence from electrophysiological studies of stimulation over

M1 suggests that there is a progressive increase in the

excitability of local circuits during rTMS especially during

stimulation at frequencies above 1 Hz (Pascual-Leone et al.,

1994; Wu et al., 2000). This may cause a progressive increase of

neural activity and subsequent positive BOLD MRI changes.

PMd

Anatomical/functional location xa ya za t valueb Volumec

L premotor (dorsal) �37 �10 53 5.08 157

L premotor (ventral) �51 5 12 5.52 530

L posterior temporal lobe �50 �51 3 4.96 485

L caudate nucleus �10 �5 19 4.14 55

L auditory cortexd �48 �24 16 5.30 532

BL medial superior frontal

gyrus (SMA)

3 �5 55 6.53 297

BL cingulate gyrus 2 5 47 5.40 319

R premotor (dorsal) 43 �8 52 6.79 130

R premotor (ventral) 41 7 26 5.80 130

R caudate nucleus 13 0 15 4.63 414

R auditory cortexd 51 �20 14 5.72 411

R cerebellar hemisphere 5 �55 �14 5.61 336

a Coordinates correspond to center of gravity of respective activation

clusters.b Peak activation within cluster with P < 0.01, corrected.c Rescaled to voxel size 1 � 1 � 1 mm.d Including superior temporal gyrus, ventral part of parietal operculum

and planum temporale. L: left, R: right, BL: bilateral.

Fig. 5. BOLD MRI signal intensity time courses (mean T SEM, n = 9) in (a) the left PMd and (b) the left M1 as determined from a priori anatomically defined

regions-of-interest (Bestmann et al., 2004). Voluntary finger movement (solid line) and suprathreshold rTMS (dashed line) evoked responses in the stimulated

left premotor cortex, while no significant signal changes were found during subthreshold rTMS (dotted line). In the left M1 caudal to the site of rTMS, only

finger movements resulted in significant activity changes.

S. Bestmann et al. / NeuroImage 28 (2005) 22–29 27

Another possibility is that with increasing numbers of TMS

pulses or higher intensities of individual pulses, there is a gradual

accumulation of oxygen debt that eventually triggers subsequent

blood flow changes. If this were the case, a certain amount of

TMS-evoked activity within a certain time window would be

required to trigger BOLD MRI signal changes. Both views

would be consistent with the lack of significant local BOLD

MRI response changes during short subthreshold TMS applica-

tions (Bohning et al., 2000b; Baudewig et al., 2001). The

absence of significant BOLD MRI changes during subthreshold

rTMS, however, does not exclude the possibility that stimulation

led to changes in local intracortical activity (Sanger et al., 2001).

However, this would mean that combined TMS-fMRI may be

blind to subtle and short-lasting activity changes within intra-

cortical circuits that can be readily detected using surface EMG

or cervical epidural recordings.

Distant effects of premotor rTMS

Remote cortical and subcortical activity changes occurred in

secondary motor areas including the contralateral PMd, the ventral

premotor cortex, cingulate, and SMA bilaterally, and subcortically,

the thalamus, caudate nucleus, and putamen. The nature of the

remote effect of TMS is not well understood. The presumed net

facilitatory effect on overall (inhibitory and excitatory) neural

activity in remote regions may be produced by transsynaptic or

direct (orthodromic or antidromic) activation of cortico-cortical, or

cortico-subcortical neurons. While a combination of these is likely

to account for remote activity changes at sites with relatively direct

connections, antidromic activation is a less likely candidate for a

polysynaptic propagation of activity to regions such as the

contralateral thalamus.

Activity increases in the homologous right PMd during

suprathreshold rTMS could be caused by direct activation of

callosal connections (Marconi et al., 2003) by each TMS pulse.

However, direct activation of transcallosal fibers during subthres-

hold stimulation is less likely given that these fibers have a smaller

diameter than the corticospinal fibers from M1 that are activated at

suprathreshold stimulation intensities. It seems more likely that

subthreshold rTMS changed the ongoing pattern of activity in

connections from the PMd to remote target regions by altering the

activity in local circuits intrinsic to PMd itself.

Notably, no significant activity changes were found in the

adjacent left M1/S1. This contrasts with recent studies that have

shown physiological effects on motor cortex excitability after

single pulse TMS (Civardi et al., 2001) and metabolic as well as

physiological effects on M1/S1 after long periods of premotor

rTMS (Chouinard et al., 2003; Siebner et al., 2003). The different

results in the present experiment may be due to a combination of

the fact that we used a different frequency of rTMS than had been

used in previous studies as well as a smaller number of pulses.

Since the pattern of excitatory and inhibitory activity produced by

rTMS is critically dependent on these parameters, it may have been

that we evoked activation of M1/S1 that resulted in no net change

in metabolic activity.

As TMS generates auditory and vibrotactile stimulation on the

scalp, this raises the question whether activity changes are

mediated by these factors rather than direct effective cortical

stimulation and its propagation into remote regions. In the present

experiments, the comparison of the two effective TMS intensities

revealed a number of secondary motor regions, including

contralateral PMd, PMv, SMA, cingulate motor region, and the

caudate nucleus, that were activated more strongly during supra-

threshold rTMS. In contrast, auditory activation was only margin-

ally larger during suprathreshold rTMS. The finding that there was

extensive overlap in auditory cortices even during finger tapping

(at which TMS was applied at 15% of stimulator output) suggests

that these activations are not due to the cortical stimulation itself,

but rather due to the auditory stimulation.

Notwithstanding the secondary activity changes related to

somatosensory and auditory stimulation, TMS applications to the

motor system target distinct cortical and subcortical circuits that

reflect the known anatomical relationship between such areas.

This means that it is possible to test the effective connectivity of

a variety of brain regions by using TMS to provide a controlled

input into a cortical circuit(s). Although the importance of

stimulus intensity on local and distal BOLD signal changes

needs to be explored further across a wide range of stimulus

intensities, the present results are proof of principle that combined

TMS-fMRI over non-primary motor regions can be used to

monitor functional connectivity in cortical circuits. In the future,

we hope to be able to apply the method to explore how these

systems react to acute and chronic perturbation of function in

healthy and disease.

S. Bestmann et al. / NeuroImage 28 (2005) 22–2928

Methodological considerations

Several factors suggest that the PMd was the target of our

TMS pulses. First, the lack of EMG activity makes a contribution

from the caudally adjoining M1 unlikely. Second, stimulation of

the nearby frontal eye field (Paus et al., 1997) or the adjacent

anterior dorsolateral prefrontal cortex (Nahas et al., 2001) has not

yielded prefrontal motor regions in a comparable way. Third, the

activation center in left caudal PMd during suprathreshold rTMS

(x = �40, y = �11, z = 54) was 12 mm anterior to the activation

center in M1/S1 during finger tapping (x = 33, y = �23, z = 55).

This corresponds well to the average distance between the PMd

and M1 (Picard and Strick, 2001). This pattern was preserved

when the two subjects with a more anterior stimulation site were

excluded from the analysis. Previous studies have shown that

when the TMS coil is over the motor hot spot, the center of the

junction region lies 5–10 mm from the site of maximum BOLD

signal change during finger movement (Herwig et al., 2002; Lotze

et al., 2003). Thus, movement of the TMS probe 2–2.5 cm

anterior to the motor hot-spot would be compatible the idea that

the BOLD MRI signal change 12 mm anterior to the M1/S1

activation center was caused by TMS pulses over PMd. However,

it is important to emphasize that we cannot exclude the possibility

of additional stimulation of more prefrontal regions. Nevertheless,

given that micro- and macroanatomical differences between brain

regions make a determination of the exact site of the maximal

induced current and most effective stimulation difficult, activity

changes in the direct vicinity of stimulation may provide an

indication of the most effectively targeted cortical region. We

conclude that in the present study movement of the TMS coil 2–

2.5 cm anterior to the motor hotspot effectively influenced activity

in the dorsal premotor cortex, rather than the anterior dorsolateral

prefrontal cortex.

Due to technical restrictions we used a stimulation frequency

between the commonly applied frequencies of 1 Hz and 5 Hz. It is

therefore difficult to interpret the results directly with regard to the

predominant inhibitory (1 Hz) and excitatory (5 Hz) effects of

rTMS. However, our results clearly demonstrate that rTMS at a

relatively low frequency can already exert significant influence on

both local and remote brain regions. In the future, the direct

comparison of different stimulation frequencies may provide more

detailed information regarding the contribution of inhibitory and

excitatory circuits to such activity changes.

Acknowledgments

SB (DFG GK-GRK 632/1-00) and HRS (DFG SI 738/1) are

grateful for financial support from the Deutsche Forschungsge-

meinschaft. The authors would like to thank John R Hernshey and

Anders A Baumann for assistance and are indebted to Anthony

Thomas and Stefan Cohrs for technical support.

References

Barker, A.T., 1999. The history and basic principles of magnetic nerve

stimulation. Electroencephalogr. Clin. Neurophysiol., Suppl. 51, 3–21.

Baudewig, J., Siebner, H.R., Bestmann, S., Tergau, F., Tings, T., Paulus, W.,

Frahm, J., 2001. Functional MRI of cortical activations induced by

transcranial magnetic stimulation (TMS). NeuroReport 12, 3543–3548.

Bestmann, S., Baudewig, J., Frahm, J., 2003a. On the synchronization of

transcranial magnetic stimulation and functional echo-planar imaging.

J. Magn. Reson. Imaging 17, 309–316.

Bestmann, S., Baudewig, J., Siebner, H.R., Rothwell, J.C., Frahm, J.,

2003b. Subthreshold high-frequency TMS of human primary motor

cortex modulates interconnected frontal motor areas as detected by

interleaved fMRI-TMS. NeuroImage 20, 1685–1696.

Bestmann, S., Baudewig, J., Siebner, H.R., Rothwell, J.C., Frahm, J., 2004.

Functional MRI of the immediate impact of transcranial magnetic

stimulation on cortical and subcortical motor circuits. Eur. J. Neurosci.

19, 1950–1962.

Bohning, D.E., Shastri, A., Nahas, Z., Lorberbaum, J.P., Andersen, S.W.,

Dannels, W.R., Haxthausen, E.U., Vincent, D.J., George, M.S., 1998.

Echoplanar BOLD fMRI of brain activation induced by concurrent

transcranial magnetic stimulation. Invest. Radiol. 33, 336–340.

Bohning, D.E., Shastri, A., McConnell, K.A., Nahas, Z., Lorberbaum, J.P.,

Roberts, D.R., Teneback, C., Vincent, D.J., George, M.S., 1999. A

combined TMS/fMRI study of intensity-dependent TMS over motor

cortex. Biol. Psychiatry 45, 385–394.

Bohning, D.E., Shastri, A., McGavin, L., McConnell, K.A., Nahas, Z.,

Lorberbaum, J.P., Roberts, D.R., George, M.S., 2000a. Motor cortex

brain activity induced by 1-Hz transcranial magnetic stimulation is

similar in location and level to that for volitional movement. Invest.

Radiol. 11, 676–683.

Bohning, D.E., Shastri, A., Wassermann, E.M., Ziemann, U., Lorber-

baum, J.P., Nahas, Z., Lomarev, M.P., George, M.S., 2000b. BOLD-

fMRI response to single-pulse transcranial magnetic stimulation

(TMS). J. Magn. Reson. Imaging 11, 569–574.

Chen, W.H., Mima, T., Siebner, H.R., Oga, T., Hara, H., Satow, T., Begum,

T., Nagamine, T., Shibasaki, H., 2003. Low-frequency rTMS over

lateral premotor cortex induces lasting changes in regional activation

and functional coupling of cortical motor areas. Clin. Neurophysiol.

114, 1628–1637.

Chouinard, P.A., Van Der Werf, Y.D., Leonard, G., Paus, T., 2003.

Modulating neural networks with transcranial magnetic stimulation

applied over the dorsal premotor and primary motor cortices. J. Neuro-

physiol. 90, 1071–1083.

Civardi, C., Cantello, R., Asselman, P., Rothwell, J.C., 2001. Transcranial

magnetic stimulation can be used to test connections to primary motor

areas from frontal and medial cortex in humans. NeuroImage 14,

1444–1453.

Goebel, R., Linden, D.E., Lanfermann, H., Zanella, F.E., Singer, W., 1998.

Functional imaging of mirror and inverse reading reveals separate

coactivated networks for oculomotion and spatial transformations.

NeuroReport 9, 713–719.

Herwig, U., Kolbel, K., Wunderlich, A.P., Thielscher, A., von Tiesen-

hausen, C., Spitzer, M., Schonfeldt-Lecuona, C., 2002. Spatial

congruence of neuronavigated transcranial magnetic stimulation and

functional neuroimaging. Clin. Neurophysiol. 113, 462–468.

Johansen-Berg, H., Rushworth, M.F., Bogdanovic, M.D., Kischka, U.,

Wimalaratna, S., Matthews, P.M., 2002. The role of ipsilateral premotor

cortex in hand movement after stroke. Proc. Natl. Acad. Sci. U. S. A.

99, 14518–14523.

Kemna, L.J., Gembris, D., 2003. Repetitive transcranial magnetic

stimulation indices different responses in different cortical areas: a

functional magnetic resonance study in humans. Neurosci. Lett. 336,

85–88.

Komssi, S., Kahkonen, S., Ilmoniemi, R.J., 2004. The effect of stimulus

intensity on brain responses evoked by transcranial magnetic stimula-

tion. Hum. Brain Mapp. 21, 154–164.

Lotze, M., Kaethner, R.J., Erb, M., Cohen, L.G., Grodd, W., Topka, H.,

2003. Comparison of representational maps using functional magnetic

resonance imaging and transcranial magnetic stimulation. Clin. Neuro-

physiol. 114, 306–312.

Marconi, B., Genovesio, A., Giannetti, S., Molinari, M., Caminiti, R., 2003.

Callosal connections of dorso-lateral premotor cortex. Eur. J. Neurosci.

18, 775–788.

S. Bestmann et al. / NeuroImage 28 (2005) 22–29 29

Mochizuki, H., Huang, Y.Z., Rothwell, J.C., 2004. Interhemispheric

interaction between human dorsal premotor and contralateral primary

motor cortex. J. Physiol. 561, 331–338.

Nahas, Z., Lomarev, M., Roberts, D.R., Shastri, A., Lorberbaum, J.P.,

Teneback, C., McConnell, K., Vincent, D.J., Li, X., George, M.S.,

Bohning, D.E., 2001. Unilateral left prefrontal transcranial magnetic

stimulation (TMS) produces intensity-dependent bilateral effects as

measured by interleaved BOLD fMRI. Biol. Psychiatry 50, 712–720.

Oxley, T., Fitzgerald, P.B., Brown, T.L., de Castella, A., Daskalakis, Z.J.,

Kulkarni, J., 2004. Repetitive transcranial magnetic stimulation reveals

abnormal plastic response to premotor cortex stimulation in schizo-

phrenia. Biol. Psychiatry 56, 628–633.

Pascual-Leone, A., Valls-Sole, J., Wassermann, E.M., Hallett, M., 1994.

Responses to rapid-rate transcranial magnetic stimulation of the human

motor cortex. Brain 117, 847–858.

Paus, T., Jech, R., Thompson, C.J., Comeau, R., Peters, T., Evans, A.C.,

1997. Transcranial magnetic stimulation during positron emission

tomography: a new method for studying connectivity of the human

cerebral cortex. J. Neurosci. 17, 3178–3184.

Paus, T., Jech, R., Thompson, C.J., Comeau, R., Peters, T., Evans, A.C.,

1998. Dose-dependent reduction of cerebral blood flow during rapid-

rate transcranial magnetic stimulation of the human sensorimotor

cortex. J. Neurophysiol. 79, 1102–1107.

Picard, N., Strick, P.L., 2001. Imaging the premotor areas. Curr. Opin.

Neurobiol. 11, 663–672.

Rizzo, V., Siebner, H.R., Modugno, N., Pesenti, A., Munchau, A.,

Gerschlager, W., Webb, R.M., Rothwell, J.C., 2004. Shaping the

excitability of human motor cortex with premotor rTMS. J. Physiol.

554, 483–495.

Sanger, T.D., Garg, R.R., Chen, R., 2001. Interactions between two

different inhibitory systems in the human motor cortex. J. Physiol.

530, 307–317.

Schluter, N.D., Rushworth, M.F., Passingham, R.E., Mills, K.R., 1998.

Temporary interference in human lateral premotor cortex suggests

dominance for the selection of movements. A study using transcranial

magnetic stimulation. Brain 121, 785–799.

Schluter, N.D., Rushworth, M.F., Mills, K.R., Passingham, R.E., 1999.

Signal-, set-, and movement-related activity in the human premotor

cortex. Neuropsychologia 37, 233–243.

Shastri, A., George, M.S., Bohning, D.E., 1999. Performance of a system

for interleaving transcranial magnetic stimulation with steady-state

magnetic resonance imaging. Electroencephalogr. Clin. Neurophysiol.,

Suppl. 51, 55–64.

Siebner, H.R., Filipovic, S.R., Rowe, J.B., Cordivari, C., Gerschlager, W.,

Rothwell, J.C., Frackowiak, R.S., Bhatia, K.P., 2003. Patients with focal

arm dystonia have increased sensitivity to slow-frequency repetitive

TMS of the dorsal premotor cortex. Brain 126, 2710–2725.

Speer, A.M., Willis, M.W., Herscovitch, P., Daube-Witherspoon, M.,

Shelton, J.R., Benson, B.E., Post, R.M., Wassermann, E.M., 2003a.

Intensity-dependent regional cerebral blood flow during 1-Hz repetitive

transcranial magnetic stimulation (rTMS) in healthy volunteers studied

with H215O positron emission tomography: II. Effects of prefrontal

cortex rTMS. Biol. Psychiatry 54, 826–832.

Speer, A.M., Willis, M.W., Herscovitch, P., Daube-Witherspoon, M.,

Shelton, J.R., Benson, B.E., Post, R.M., Wassermann, E.M., 2003b.

Intensity-dependent regional cerebral blood flow during 1-Hz repetitive

transcranial magnetic stimulation (rTMS) in healthy volunteers studied

with H215O positron emission tomography: I. Effects of primary motor

cortex rTMS. Biol. Psychiatry 54, 818–825.

Stewart, L.M., Walsh, V., Rothwell, J.C., 2001. Motor and phosphene

thresholds: a transcranial magnetic stimulation correlation study.

Neuropsychologia 39, 415–419.

Talairach, J., Tournoux, P., 1988. Coplanar Stereotaxic Atlas of the Human

Brain. Thieme, Stuttgart.

Wassermann, E.M., 1998. Risk and safety of repetitive transcranial

magnetic stimulation: report and suggested guidelines from the Interna-

tional Workshop on the Safety of Repetitive Transcranial Magnetic

Stimulation, June 5–7, 1996. Electroencephalogr. Clin. Neurophysiol.

108, 1–16.

Wu, T., Sommer, M., Tergau, F., Paulus, W., 2000. Lasting influence of

repetitive transcranial magnetic stimulation on intracortical excitability

in human subjects. Neurosci. Lett. 287, 37–40.