Robot-assisted image-guided transcranial magnetic stimulation for somatotopic mapping of the motor cortex: a clinical pilot study.

Abstract

Shape and exact location of motor cortical areas varies among individuals. The exact knowledge of these locations is crucial for planning of neurosurgical procedures. In this study, we have used robot-assisted image-guided transcranial magnetic stimulation (Ri-TMS) to elicit MEP response recorded for individual muscles and reconstruct functional motor maps of the primary motor cortex.
One healthy volunteer and five patients with intracranial tumors neighboring the precentral gyrus were selected for this pilot study. Conventional MRI and fMRI were obtained. Transcranial magnetic stimulation was performed using a MagPro X100 stimulator and a standard figure-of-eight coil positioned by an Adept Viper s850 robot. The fMRI activation/Ri-TMS response pattern were compared. In two cases, Ri-TMS was additionally compared to intraoperative direct electrical cortical stimulation.
Maximal MEP response of the m. abductor digiti minimi was located in an area corresponding to the "hand knob" of the precentral gyrus for both hemispheres. Repeated Ri-TMS measurements showed a high reproducibility. Simultaneous registration of the MEP response for m. brachioradialis, m. abductor pollicis brevis, and m. abductor digiti minimi demonstrated individual peak areas of maximal MEP response for the individual muscle groups. Ri-TMS mapping was compared to the corresponding fMRI studies. The areas of maximal MEP response localized within the "finger tapping" activated areas by fMRI in all six individuals.
Ri-TMS is suitable for high resolution non-invasive preoperative somatotopic mapping of the motor cortex. Ri-TMS may help in the planning of neurosurgical procedures and may be directly used in navigation systems.

Full text

Page 1

EXPERIMENTAL RESEARCH
Robot-assisted image-guided transcranial magnetic
stimulation for somatotopic mapping of the motor cortex:
a clinical pilot study
Sven Rainer Kantelhardt & Tommaso Fadini & Markus Finke & Kai Kallenberg &
Jakob Siemerkus & Volker Bockermann & Lars Matthaeus & Walter Paulus &
Achim Schweikard & Veit Rohde & Alf Giese
Received: 14 September 2009 /Accepted: 3 November 2009 /Published online: 27 November 2009
# The Author(s) 2009. This article is published with open access at Springerlink.com
Abstract
Purpose Shape and exact location of motor cortical areas
varies among individuals. The exact knowledge of these
locations is crucial for planning of neurosurgical proce-
dures. In this study, we have used robot-assisted image-
guided transcranial magnetic stimulation (Ri-TMS) to elicit
MEP response recorded for individual muscles and recon-
struct functional motor maps of the primary motor cortex.
Methods One healthy volunteer and five patients with
intracranial tumors neighboring the precentral gyrus were
selected for this pilot study. Conventional MRI and fMRI
were obtained. Transcranial magnetic stimulation was
performed using a MagPro X100 stimulator and a standard
figure-of-eight coil positioned by an Adept Viper s850
robot. The fMRI activation/Ri-TMS response pattern were
compared. In two cases, Ri-TMS was additionally com-
pared to intraoperative direct electrical cortical stimulation.
Results Maximal MEP response of the m. abductor digiti
minimi was located in an area corresponding to the “hand
knob” of the precentral gyrus for both hemispheres.
Repeated Ri-TMS measurements showed a high reproduc-
ibility. Simultaneous registration of the MEP response for
m. brachioradialis, m. abductor pollicis brevis, and m.
abductor digiti minimi demonstrated individual peak areas
of maximal MEP response for the individual muscle
groups. Ri-TMS mapping was compared to the
corresponding fMRI studies. The areas of maximal MEP
response localized within the “finger tapping” activated
areas by fMRI in all six individuals.
Conclusions Ri-TMS is suitable for high resolution non-
invasive preoperative somatotopic mapping of the motor
cortex. Ri-TMS may help in the planning of neurosurgical
procedures and may be directly used in navigation systems.
Keywords FunctionalMRI.Transcranialmagnetic
stimulation.Motorcortex.Brainmapping.Robotized
neuronavigation.Electricalcortexstimulation
Introduction
The movements of the human body are controlled by the
motor cortex located within the precentral gyrus, an
anatomical structure readily identified on MRI [1] and by
S. R. Kantelhardt (*):V. Bockermann:V. Rohde:
A. Giese (*)
Department of Neurosurgery,
Georg-August University of Göttingen,
Robert-Koch-Strasse 40,
37075 Göttingen, Germany
e-mail: sven.kantelhardt@med.uni-goettingen.de
e-mail: alf.giese@med.uni-goettingen.de
T. Fadini:W. Paulus
Department of Clinical Neurophysiology,
Georg-August University of Göttingen,
Göttingen, Germany
K. Kallenberg
Department of Neuroradiology,
Georg-August University of Göttingen,
Göttingen, Germany
M. Finke:L. Matthaeus:A. Schweikard
Institute for Robotics und Cognitive Systems,
University of Lübeck,
Lübeck, Germany
K. Kallenberg:J. Siemerkus
MR-Research in Neurology and Psychiatry,
Georg-August University of Göttingen,
Göttingen, Germany
Acta Neurochir (2010) 152:333–343
DOI 10.1007/s00701-009-0565-1

Page 2

intraoperative mapping techniques during neurosurgical
procedures [2]. The shape and exact location of the cortical
motor areas varies among individuals and pathological
conditions may cause distortion of functional motor areas
and even shift of motor function to other regions [3].
However, the individual location of cortical motor areas is
crucial for planning of neurosurgical procedures and
preservation of motor function.
Direct electrical cortical stimulation (ECS) is well
established and essential for localizing functional motor
areas during neurosurgical procedures [2, 4]. For the
planning of surgical approaches and procedures, preopera-
tive functional maps of the motor cortex are required and in
some cases the ability to precisely locate functional areas
can become a determinant for feasibility of the neurosur-
gical intervention.
Blood-oxygen-level-dependent functional magnetic res-
onance imaging (BOLD-fMRI) is an established method in
clinical neuroscience applied to visualize changes in
magnetic field (in-)homogeneity caused by the oxygenation
status of hemoglobin related to “brain activity” associated
with specific tasks [5]. It can be used for mapping of the
functional brain anatomy, which for motor- sensory- and
speech areas have been widely used in neurosurgical
decision-making [6]. The fMRI is the most widely used
technique for non-invasive somatotopic mapping and is
regarded as the gold standard today. The validity of this
technique has been shown in several studies comparing
fMRI with direct ECS [3, 7–9]. However, the spatial
resolution of fMRI is limited because function is visualized
indirectly by determination of changes in blood oxygen
levels (BOLD), which are susceptible to alterations of the
signal by large draining vessels or highly metabolically
active tumors near the region of interest. Because changes
in blood oxygen level require a larger number of active
neurons, typically functional areas displayed by fMRI
represent activity of relatively complex movements involv-
ing several muscle groups. The mapping of eloquent
functional areas representing single muscles remains
difficult.
Other techniques for non-invasive functional imaging
include functional PET [10, 11], neuromagnetic recording
[12], and conventional TMS [13]. To reduce inter-
investigator variations and improve reproducibility, as well
as the spatial accuracy, the latter has been assisted by
stereotactical systems [14, 15] and frameless image-
guidance navigation systems for positioning of TMS coils
[16]. Denslow et al. compared an image-guided TMS
technique to conventional hand-held function-guided coil
placement and reported high precision and effectiveness of
the image-guided TMS in reference to fMRI [17].
In this study, we investigate whether robot-assistance can
be used as a feasible and safe technique for further
improvement of image-guided transcranial magnet stimula-
tion, and whether it allows the identification of cortical
motor areas of individual muscles.
Methods
One healthy volunteer and five patients aged 53±11 years
(mean±SD, three female) diagnosed with intracranial
tumors (one meningeoma WHO grade I, one astrocytoma
WHO grade II, one astrocytoma WHO grade III, and two
glioblastomas WHO grade IV), neighboring the precentral
gyrus but without apparent paresis at clinical examination
were selected for this pilot study.
MRI-imaging, processing, and visualization
MRI studies were performed at 3 T (Magnetom Trio,
Siemens Medical Solutions, Erlangen, Germany) using
an eight-channel head coil. A T1-weighted whole-brain
3D-dataset (time to repetition (TR)/time to echo (TE)=
11/4 ms, flip angle 15?;1 ? 1 ? 1 mm3isotropic resolu-
tion) and for fMRI-analysis a T2-weighted dataset (EPI,
frequency-selective fat suppression, time to repetition TR/
time to echo TE=2,000/36 ms, flip angle 70°, 2×2 mm2
resolution, 4 mm section thickness, 22 slices, top slice
adjusted to the most superior proportion of the cortex; 129
acquisitions; total acquisition time=4:20 min) were
obtained.
The fMRI-paradigm consisted of 9x18s of resting state
as control-condition alternating with 8×12s of an active
state starting with the control-condition. For the active state,
subjects were instructed to successively touch the fingertips
II–V with the thumb of either right or left hand as fast as
possible (finger-tapping).
To avoid the T1 saturation effect the first two volumes of
the T2-datasets were excluded from analysis. Preprocessing
included mean intensity adjustment, slice scan time
correction (cubic spline ascending interleaved), 3D motion
correction (trilinear interpolation), and temporal filtering
(high-pass/GLM-Fourier at 2 sines/cosines). The functional
datasets were semi-automatically co-registered to the
Talairach-transformed same subject T1-weighted whole-
brain scan.
For the finger-tapping condition, a two-Gamma hemo-
dynamic response function was used to calculate a general
linear model (GLM).
For visualization, an appropriate statistical threshold for
the GLM of each subject was defined. The positive supra-
threshold voxels were overlayed as a color-coded map with
higher t values in blue and lower t values in yellow on a
semi-automatically calculated 3D-reconstruction of the cor-
tex–white matter boundary of the T1-weighted dataset of the
334 S.R. Kantelhardt et al.

Page 3

corresponding subject including all suprathreshold cortex
voxels at a range of 3 mm to the cortex–white matter
boundary.
Robot-assisted image-guided TMS
The system for robot-assisted image-guided TMS contains
six main components (Fig. 1): A: TMS coil, B: robot, C:
robot controller, D: personal computer, E: infrared tracking
camera, and F: headband with reflective markers. Experi-
ments were carried out using a standard figure-of-eight coil
(Medtronic, MC-B70, Skovlunde, Denmark) attached to the
articulated arm of an Adept Viper s850 robot (Adept
Technology, Inc. Livermore, CA, USA), which was driven
by a standard PC at 2.8 GHz and software developed at the
Institute for Robotics and Cognitive Systems, University of
Lübeck, for image-guided robot-control [18]. A rubber
headband carrying three passive Polaris reflective markers
on a plastic frame was fixed to the head and tracked by a
Polaris infrared stereo-optical tracking system (Northern
Digital Inc., Waterloo, Ontario, Canada). The robot and the
tracking system were registered to a common coordinate
system using standard algorithms [19]. The coordinate
system for the robot-guided TMS coil was defined by
registration of three reference positions on the coil using a
tracked pointer. This also determined the coordinate
transformation from the robot end-effector (end of the
robot arm) to the coil.
Registration of the spatial position of the cranium to the
reconstruction of the cranial surface from MRI data was
done by pointer acquisition of three standard landmarks
(lateral orbital rims and tip of the nose) and up to 300
additional surface points [20].
For the transcranial magnetic stimulation, the coil target
position was defined using the three-dimensional recon-
struction of cranial MRI data. A stimulation target on the
surface of the virtual head was selected and the orientation
of the coil was calculated tangentially to cortical surface
and 45° to a saggital plane based on the reconstructed MRI
data. The virtual coil position was then transformed to robot
coordinates, which defined the movement of the robot to
the corresponding target position relative to the patient’s
head. The position of the cranium was continuously tracked
and the trajectory of the coil’s path was adapted to
movements of the head by a motion compensation module
in real-time, which guaranteed a precise coil to target
position and allowed free head movements during the study
[21].
Transcranial magnetic stimulation was performed using a
MagPro X100 MagOption stimulator (Medtronic, Dantec
S.A., Skovlunde, Denmark) connected to a monitoring unit
(Endeavor CR, Nicolet Biomedical, Dublin/Ohio, USA).
The trigger signal for the stimulator and the recording unit
of the Endeavor CR was generated by a script run on an
external personal computer. The signal from stimulated
muscles was registered by the Endeavor CR, band-pass-
A
C
D
E
B
F
E
B
A
G
F
H
Fig. 1 The upper panel shows
the six main components of
robot-guided TMS: (A) TMS
coil, (B) robot, (C) robot con-
troller, (D) computer, (E) Polaris
tracking system, (F) headband.
The lower left panel shows the
examination set up and the Ri-
TMS components A, B, E, F and
the MEP registration electrodes
(G). The right lower panel
shows a surface reconstruction
of a T1 MRI scan used for grid
definition and robot control. The
grid points stimulated by TMS
are displayed as color-coded
dots. The color indicates the
magnitude of the MEP response
evoked by Ri-TMS stimulation
(red none, blue maximal)
Ri-TMS for somatotopic mapping of the motor cortex335

Page 4

filtered between 2 and 2000 Hz and sampled at 5000 Hz.
The m. brachioradialis, m. abductor pollicis brevis, and m.
abductor digiti minimi were recorded using Ag/AgCl bipolar
surface electrodes. Both hemispheres were sequentially
investigated by delivering ten biphasic pulses per target with
a 5-s interstimulation interval (0.2 Hz). The analysis of the
peak-to-peak amplitudes of the motor-evoked potentials
(MEPs) was performed online using commercial software
(Viasys, Nicolet Biomedical, Dublin/Ohio, USA). The final
MEP response values used for construction of the motor map
were calculated from the average of ten valid MEP response
curves for each target. For selection of the initial stimulation
intensity up to five targets corresponding to the position of the
anatomically identified motor cortex (hand knob) were
stimulated. The intensity was determined on the basis of
reliable activation for at least one of the recorded muscles,
peak-to-peak amplitude ofthe MEP≥0.5 mV, without causing
discomfort to the patients.
The distribution of the electric field of the TMS coil was
determined in a previous experiment using the robot,a copper
wireprobe,andaPCS1008-bitdigitaloscilloscope(Velleman
Components N.V., Gavere, Belgium) with a sampling
frequency of 800 kHz [21]. The coil position data and the
field characteristics of the coil allowed an approximation of
the electric field distribution in the brain for each stimulation
target. Assuming a functional relationship between the
electric field strength at the anatomical representation of a
muscle within the motor cortex and its MEP response, we
calculated the most likely three-dimensional representation
for the recorded muscle using an adapted version of the
correlation ratio statistic (for details see [22]). The calculated
field distributions were projected on the surface map of the
brain, segmented from the diagnostic MRI.
Procedure and evaluation
For all patients and the healthy volunteer, conventional MRI
and fMRI were obtained including protocol sequences for
neuronavigation according to the manufactures instructions
(VectorVision 2, BrainLAB, Feldkirchen, Germany) [23]. Ri-
TMS was performed for both hemispheres. Maps of the
brain surface were based on the segmented cortex structure
from T1-weighted 3D MR sequence, acquired as described
above. The fMRI and Ri-TMS data were projected on these
maps, resulting in individual motor cortex maps.
The motor cortex maps generated by both techniques
were analyzed by a neuroradiologist and a neurosurgeon.
The location of the maximal activity/response was pro-
jected to the gyral pattern and compared relative to
anatomical landmarks. The fMRI activation/Ri-TMS re-
sponse pattern were scored in one of three categories:
“Match” if the regions of fMRI activation and maximal Ri-
TMS response located to the same anatomical site; “non-
match” if the regions did not overlap; “match/nonmatch”
for an partial overlap of areas of fMRI activation and Ri-
TMS response.
Integration of Ri-TMS response into intraoperative
neuronavigation
For patients scheduled for surgery, the regions of the
maximal MEP response were transferred to the neuro-
navigation (VectorVision 2, BrainLAB, Feldkirchen,
Germany). A direct transformation of maximal MEP
response by TMS to intraoperative neuronavigation data
was not possible because the navigation system used for Ri-
TMS (Northern Digital Inc.) was based on a different
coordinate system than the intraoperative neuronavigation
(VectorVision 2). The data transformation was achieved by
registration of scalp fiducial markers in both navigation
systems regardless of the different registration protocols.
This defined the marker positions in both coordinate
systems (×17). The regions of maximal MEP response
were calculated by the divergence vectors from three
different scalp fiducial markers. This mathematical opera-
tion allowed the location of the region of interest relatively
to the fiducial markers and their position data of the
intraoperative neuronavigation device.
Intraoperative monitoring
In three patients scheduled for surgery, standard
neuronavigation-guided craniotomies were performed. Prior
to and during the resection of the tumors, neuronavigation-
tracked ECS was carried out. For stimulation and registra-
tion of MEP response, an Endeavor CR neuromonitoring
unit (Viasys, Nicolet Biomedical, Dublin/Ohio, USA) with
a monopolar brain-stimulation electrode (Dr. Langer Med-
ical GmbH, Waldkirch, Germany) was used. The regions of
maximal response to cortical stimulation were directly
compared to the calculated position of maximal MEP
response by TMS.
Results
Patients and stimulation parameters
At the time of Ri-TMS, one of five patients required
anticonvulsant medication because of a history of focal
seizures. Medication of the other patients included dexa-
methasone and non-steroidal antiphlogistic therapy, which
are not known to interfere with TMS. During the Ri-TMS,
the patients were allowed to position and move their heads
freely. Both hemispheres were examined, with the excep-
tion of one patient who developed a focal seizure and
336S.R. Kantelhardt et al.

Page 5

discontinued the mapping study. The stimulation intensity
used for the mapping was selected between 40% and 45%
(43±2, mean±SD) of maximum stimulator output (MSO).
This corresponds well with the intensities other authors
applied for the same task [24]. Per hemisphere, an average
of 38 grid points (range 32–60) was examined (Table 1).
Measurements were started at a grid point directly above
the anatomical hand knob identified on the surface
reconstruction of MRI and continued in all planar directions
at a grid width of 8 mm until no MEP response could be
elicited in the muscle group of interest (Fig. 1, panel H).
Cortical motor mapping
The areas of maximal MEP response by Ri-TMS and fMRI
activation were projected as maps on the surface reconstruc-
tions from MRI for each individual (Fig. 2). In a healthy
volunteer, Ri-TMS was performed for both hemispheres. The
maps of the calculated maximal MEP response demonstrated
a maximal MEP response of the m. abductor digiti minimi in
an area corresponding to the “hand knob” of the precentral
gyrus (Fig. 2a). The area of maximal MEP response of the
m. abductor digiti minimi was similar for the left and right
hemispheres and corresponded well to the expected anatom-
ical region (Fig. 2a and b). Repeated measurements using de
novo registration and de novo definition of a grid system for
Ri-TMS demonstrated a high reproducibility with the MEP
response of the m. abductor digiti minimi localizing to
exactly the same anatomical site (Fig. 2a and c).
Simultaneous registration and discrimination of individual
cortical motor areas
Simultaneous registration of the MEP response for m.
brachioradialis, m. abductor pollicis brevis, and m.
abductor digiti minimi demonstrated individual peak
areas of maximal MEP response for the individual
muscle groups, which corresponded to their expected
somatotopic representation within the precentral gyrus
(Fig. 3 a, b and c). fMRI using a “finger tapping”
paradigm showed activation of larger areas of the
precentral gyrus and parts of the postcentral gyrus. The
areas of maximal MEP response for m. brachioradialis, m.
abductor pollicis brevis, and m. abductor digiti minimi
matched with the area of fMRI activation (Fig. 3d). In four
of five patients with intracranial tumors near the motor
cortex, Ri-TMS mapping of the effected hemisphere
allowed identification of clearly distinct areas of maximal
MEP response for the three muscles. In one patient
(patient #2) with a large postcentral histologically con-
firmed glioblastoma, Ri-TMS mapping localized the
maximal MEP response for all three muscles to a
relatively large area of the precentral gyrus. However,
the areas of maximal MEP response for each of the three
muscles overlapped, which did not allow discrimination of
the individual peak areas of MEP response for the
individual muscles. In a case of a low-grade glioma
(patient #5) involving the left central region and displac-
ing the precentral gyrus rostrally, Ri-TMS mapping
Table 1 The table summarizes patient data, technical details, and results for all examined individuals examined by Ri-TMS (1 volunteer, 5
patients)
DiagnosisRi-TMSfMRI/Ri-TMSIOP-ECS/Ri-TMS
%
MSO
Grid Time Right
hemisphere
Left
hemisphere
Right
hemisphere
Left
hemisphere
Distance of peak areas
in mm
Volunteer 45-year-
old male
37-year-
old male
58-year-
old female
68-year-
old female
49-year-
old female
48-year-
old male
No pathologies4648 50 Complete
mapping
Complete
mapping
Incomplete
mapping
Complete
mapping
Complete
mapping
Complete
mapping
Complete
mapping
Complete
mapping
Complete
mapping
Complete
mapping
Complete
mapping
Incomplete
mapping
MatchMatchNo surgery performed
Patient #
1
Patient #
2
Patient #
3
Patient #
4
Patient #
5
Pre central
glioblastoma
Post central
glioblastoma
Meningeoma
central
Prae central of
oligoastrocytoma
Post central
astrocytoma
4433 28Match
–
26.3 after resection
4448 40Match/
nonmatch
Match
–
Not exposed during
surgery
<5 prior to resection,
16.9 after resection
No surgery performed
40 40 33
–
42 36 30
–
Match
44 3227
–
MatchNo surgery performed
Technical details include: intensity of stimulation (%MSO), number of grid points stimulated per hemisphere (grid), time required for mapping of
one hemisphere (time). The results for the hemispheres affected by the tumor are shaded in gray. “Complete mapping” stands for a successful
identification of distinct areas of maximal MEP response for mm. brachioradialis, abductor pollicis brevis, and abductor digiti minimi within the
precentral gyrus; while “incomplete mapping” describes an identification of areas of maximal MEP response with insufficient discrimination of
the three muscle areas. If the areas of maximal MEP response to Ri-TMS are completely located within the area activated on fMRI, the result is
rated as “match”; while “match/nonmatch” indicates an only partial overlap of fMRI and Ri-TMS
Ri-TMS for somatotopic mapping of the motor cortex337

Page 6

demonstrated highly circumscript areas of maximal MEP
response for the m. brachioradialis, m. abductor pollicis
brevis, and m. abductor digiti minimi, but the three peak
areas localized to a small area of the distorted precentral
gyrus (Fig. 4). In both of these cases, identification of
distinct cortical motor areas for the three muscles was
possible for the unaffected contralateral hemisphere
(Table 1).
AB
C
D
Fig. 3 Simultaneous registra-
tion of the maximal MEP re-
sponse for m. brachioradialis
(a), m. abductor pollicis brevis
(b), and m. abductor digiti min-
imi (c) by Ri-TCM.
Corresponding fMRI for a
finger-tapping paradigm (d)
A
C
B
D
Central sulcus
hand knob
Fig. 2 Projection of the MEP
response on a segmented surface
map of the brain (based on T1
MRI images). Red shows no
excitability, blue shows the re-
gion of strongest MEP response.
a Shows the area of maximal
MEP response for the right m.
abductor digiti minimi. b Shows
the corresponding region for the
left m. abductor digiti minimi. c
Shows a repeat examination
illustrated for the area of maxi-
mal MEP response for the right
m. abductor digiti minimi
(compare a) after 3 days using
de novo registration and defini-
tion of a new grid system.
Anatomical landmarks (central
sulcus and hand knob) are
highlighted in d for the right
hemisphere
338S.R. Kantelhardt et al.

Page 7

Comparison of fMRI and Ri-TMS
Using a “finger tapping” paradigm, fMRI for all individuals
tested showed a large area of activation within the lateral
precental gyrus and, to a varying extent, activation of the
ipsilateral SMA, the contralateral precentral gyrus, some
postcentral areas and bilateral cerebellar areas. When Ri-
TMS mapping was compared to the corresponding fMRI
studies, the areas of maximal MEP response for the three
muscles localized within the “finger tapping”-activated areas
by fMRI in four of five patients and the healthy volunteer
(compare Table 1). For patient #2, a partial overlap of the
area of maximal MEP response and fMRI was found.
Comparison of direct ECS and Ri-TMS
Two of the five patients examined by Ri-TMS had
previously undergone biopsy and did present for follow
up examination only, the other three were scheduled for
surgery and intraoperative ECS was performed during
tumor resection in two of these (in the third case, the
central region was not exposed during surgery). In two of
these patients prior to surgery, the Ri-TMS spatial coor-
dinates of the maximal MEP response were integrated into
the neuronavigational device (VectorVision 2) to allow a
reference for intraoperative acquisition of ECS data. Intra-
operative direct electrical stimulation of the motor cortex
successfully located areas of maximal MEP for different
muscles in hand and forearm in both patients. Areas of
maximal MEP response were registered using the neuro-
navigation (Fig. 5). Measurements of the cortical distance
between the areas of peak MEP response to preoperative
Ri-TCM and intraoperative ECS were performed. This
analysis demonstrated distances between 5 and 26.3 mm.
For the latter case, access to the precentral gyrus could only
be obtained after removal of a large tumor and significant
volume shifts had occurred (compare Table 1.).
Case illustration
Patient #3
A 68-year-old female presented with focal seizures of the
left arm. MRI-imaging was done and a lesion with
radiological characteristics of a small meningeoma
immediately above the right central sulcus was diag-
nosed. An fMRI demonstrated that left-sided “finger
tapping” resulted in activation within the motor cortex
immediately adjacent, lateral, and anterior to the tumor
(Fig. 5). Ri-TMS was performed and demonstrated clear
delineation of individual areas of maximal MEP response
for the m. brachioradialis, m. abductor pollicis brevis, and
m. abductor digiti minimi. The areas of maximum MEP
response by Ri-TMS were transferred to the neuronaviga-
tion (Fig. 5b). A neuronavigation-guided craniotomy
exposed the tumor and a small area of adjacent cortex
which was examined by direct electrical cortex stimula-
tion. The area of maximal MEP response by ECS for the
BC
DE
A
Fig. 4 Astrocytoma WHO grade II of the left central region (patient
#5). a T1- (upper panel) and T2- (lower panel) weighted MRI. Areas
of maximal MEP response for m. brachioradialis (b), m. abductor
pollicis brevis (c), and m. abductor digiti minimi (d) by Ri-TCM of
the left hemisphere. d: Corresponding fMRI (finger-tapping paradigm)
Ri-TMS for somatotopic mapping of the motor cortex 339

Page 8

m. digiti was identified just lateral and anterior to the
tumor. The cortical area was registered using the neuro-
navigation. Comparison of data acquired for the maximal
MEP response for Ri-TMS and ECS showed that both
areas localized to a morphologically distinct area of the
proximal right “hand knob” within 5 mm. However, due to
the small craniotomy, more lateral aspects of the precental
gyrus could only be exposed following resection of the
tumor. Repeat cortical mapping by ECS now identified the
cortical motor area of the m. abductor pollicis brevis.
Comparison of the spatial data of maximal MEP response
by Ri-TCM and ECS now showed a mismatch of
16.9 mm, possibly due to brain shift during the course of
the operation as a consequence of loss of CSF and the
resected tumor volume.
Discussion
TMS for somatotopic brain mapping
TMS is a reliable technique for somatotopic mapping of the
motor cortex, which has been compared to fMRI imaging
[13, 15, 25–27] and direct ECS [14, 25]. The technique was
much improved by introduction of image-guidance [17]. In
this study, we investigated whether a further automatization
by introduction of robot-assistance likewise results in a
further improvement of the technique.
Coil positioning and reproducibility
Ri-TMS offers some advantages to conventional function-
guided, stereotactic-, or image-guided-TMS. Robotized
positioning of the coil was tightly controlled and defined
by a constant grid width and a controlled position based on
the surface reconstruction of the brain. It operated at a
constant distance to the brain surface and the angulations of
the coil was calculated relative to the individual shape of
the brain surface. This minimizes inter-investigator varia-
tions. Repeated measurements on the same individual at a
study interval of several days using de novo registration
and planning resulted in identical results.
Spatial resolution of Ri-TMS
The cortical areas of maximal MEP response by Ri-TMS
appeared to localize above the central sulcus but frequently
involved areas within the pre- and postcentral gyrus,
Sag
TraCor
A
B
C
Fig. 5 Meningeoma above the right central sulcus causing focal
seizures (patient #3). a T2-weighted MRI and fMRI using a finger-
tapping paradigm. b Preoperative Ri-TCM (right) and integration of
spatial Ri-TCM coordinates into the neuronavigation data set. c
Intraoperative ECS mapping and comparison of data acquired for the
maximal MEP response for ECS and Ri-TMS after tumor resection
340 S.R. Kantelhardt et al.

Page 9

correlating very well with the fMRI findings in these
patients. Accordingly, anatomical and physiological studies
demonstrate that most motoneurons are located at the
bottom of the central sulcus and the dorsal aspect of the
precentral gyrus [28]. Although newer findings in primates
suggest that also frontal non-primary motor areas were
directly connected to the effector organs [29]. Whether this
is true in humans has not yet been shown.
In one patient, where intraoperative exposure of the motor
cortex allowed ECS prior to tumor resection, the mismatch
was found to be less than 5 mm. Lotze et al., in comparison,
found in their study with five healthy volunteers, an average
spatial difference between fMRI and TMS activation maxima
of 13.9 mm [13] and Krings et al. found in an investigation
of two patients with space-occupying tumors close to the
central sulcus, more than 75% of the sites of maximal MEP
response to TMS to be located within 1 cm of the sites of
maximal ECS response [14].
Unfortunately, the size of our study (five patients and
one volunteer) does not yet allow general assessment of
maximum spatial resolution of Ri-TMS; however, the
results of this pilot study already suggest a high accuracy.
Identification of cortical motor areas of individual muscle
groups
Multi-channel registration of MEP allowed simultaneous
analysis of the areas of maximal MEP for different muscle
groups. Conceptually, this is not limited to three muscles
(like in our study) and potentially allows generation of
complex functional motor maps of a whole hemisphere by
stimulation of a single grid. While this cannot be achieved
by fMRI, as it visualizes activation of cortical areas induced
by relatively complex tasks such as finger-tapping, involving
groups of agonistic and antagonistic muscles; conventional
image-guided TMS in combination with multi-channel MEP
registration might allow discrimination of cortical motor areas
for individual muscles as well. However, the anatomic
differentiation of cortical motor areas is highly dependent to
the spatial resolution of the technique. Therefore robot-
assistance might enhance discrimination by enhancing the
spatial resolution. In our pilot study, differentiation and
identification of cortical motor areas for individual muscle
groups was achieved in all patients/probands who completed
the Ri-TMS examination (at least in the non-affected
hemisphere).
Investigation time and patient compliance
Besides technical issues like correct registration and coil
placement, the accuracy of (image-guided) TMS is further
limited by the ability of the patient to comply with the
investigation. This is especially important for patients who
suffer from severe comorbidities and/or neurological
deficits which might restrict compliance. Dynamic robot
positioning control allows the patient to move his head at
will, and avoids the need for rigid fixation. This not only
increases patient’s comfort during the mapping procedure,
but also increases safety by minimizing the risk of collision
with the device by uncontrolled patient movements, for
example, associated with seizures. Furthermore, it shortens
investigation times significantly. The time required for Ri-
TMS mapping largely depends on the number of grid points
and the stimulation process itself, as coil positioning is
extremely fast and no longer consumes significant time,
even though the speed of the robot was limited to 3% of the
maximal velocity (228 mm/s) only.
Conceptually, after determination of the stimulation
threshold, the mapping of a grid could be done automati-
cally. In industrial settings, the type of robot used here
operates at up to 7,600 mm/s and maximal rotation: axis 1–
3 250°/s, axis 4–5 375°/s, and axis 6 600°/s.
Outlook
fMRI offers the possibility to localize non-motor functional
areas,suchas language [9, 30, 31] and sensory function [31].
Likewise, TMS can be used to localize these cortical areas
by reversible suppression of neurological functions, such as
cutaneous [32] and visual perception [33, 34], speech and
counting [35], modulation and suppression of pain sensation
[36], and therapeutic modulation of depression [37]. Some of
these applications may benefit from application of Ri-TMS,
especially in patients with restricted compliance. However,
before Ri-TMS can be applied for these tasks, a large study
has to be performed in order to exactly define the accuracy
of Ri-TMS in comparison to ECS.
Conclusions
Ri-TMS is a feasible and safe technique for non-invasive
somatotopic mapping of the motor cortex. The localization
of functional areas within the motor cortex showed high
intraindividual reproducibility and due to a high degree of
automatization, was investigator-independent. Ri-TMS re-
solved the areas of maximal MEP response for individual
muscles and therefore allowed identification of the cortical
representation of motor control with high resolution. This
suggests that the functional and spatial data acquired by Ri-
TMS may help in the planning of neurosurgical procedures
and may be directly used in navigation systems.
Acknowledgements
Neuroscience Early Stage Research Training (NEUREST, MEST-CT-
2004-504193).
T.F. was supported by an EC fellowship for
Ri-TMS for somatotopic mapping of the motor cortex341

Page 10

Open Access
Commons Attribution Noncommercial License which permits any
noncommercial use, distribution, and reproduction in any medium,
provided the original author(s) and source are credited.
This article is distributed under the terms of the Creative
References
1. Missir O, Dutheil-Desclercs C, Meder JF, Musolino A, Fredy D
(1989) Central sulcus patterns at MRI. J Neuroradiol 16(2):133–144
2. Penfield W, Boldry E (1937) Somatic motor and sensory
representation in the cerebral cortex of man as studied by
electrical stimulation. Brain 60:389–443
3. Lehéricy S, Duffau H, Cornu P, Capelle L, Pidoux B, Carpentier A,
Auliac S, Clemenceau S, Sichez JP, Bitar A, Valery CA, Van
EffenterreR,FaillotT,SrourA,FohannoD,PhilipponJ,LeBihanD,
Marsault C (2000) Correspondence between functional magnetic
resonance imaging somatotopy and individual brain anatomy of the
central region: comparison with intraoperative stimulationin patients
with brain tumors. J Neurosurg 92(4):589–598
4. Sala F, Lanteri P (2003) Brain surgery in motor areas: the
invaluable assistance of intraoperative neurophysiological moni-
toring. J Neurosurg Sci 47(2):79–88
5. Frahm J, Bruhn H, Merboldt KD, Hänicke W (1992) Dynamic
MR imaging of human brain oxygenation during rest and photic
stimulation. J Magn Reson Imaging 2(5):501–505
6. Rao SM, Binder JR, Hammeke TA, Bandettini PA, Bobholz JA,
Frost JA, Myklebust BM, Jacobson RD, Hyde JS (1995)
Somatotopic mapping of the human primary motor cortex with
functional magnetic resonance imaging. Neurology 45(5):919–924
7. Bizzi A, Blasi V, Falini A, Ferroli P, Cadioli M, Danesi U, Aquino
D, Marras C, Caldiroli D, Broggi G (2008) Presurgical functional
MR imaging of language and motor functions: validation with
intraoperative electrocortical mapping. Radiology 248(2):579–589
8. Fandino J, Kollias SS, Wieser HG, Valavanis A, Yonekawa Y
(1999) Intraoperative validation of functional magnetic resonance
imaging and cortical reorganization patterns in patients with brain
tumors involving the primary motor cortex. J Neurosurg 91
(2):238–250
9. Mueller WM, Yetkin FZ, Hammeke TA, Morris GL 3rd, Swanson
SJ, Reichert K, Cox R, Haughton VM (1996) Functional magnetic
resonance imaging mapping of the motor cortex in patients with
cerebral tumors. Neurosurgery 39(3):515–520
10. Grafton ST, Woods RP, Mazziotta JC, Phelps ME (1991)
Somatotopic mapping of the primary motor cortex in humans:
activation studies with cerebral blood flow and positron emission
tomography. J Neurophysiol 66(3):735–743
11. Viñas FC, Zamorano L, Mueller RA, Jiang Z, Chugani H, Fuerst
D, Muzik O, Mangner TJ, Diaz FG (1997) [15O]-water PET and
intraoperative brain mapping: a comparison in the localization of
eloquent cortex. Neurol Res 19(6):601–608
12. Cheyne D, Kristeva R, Deecke L (1991) Homuncular organization
of human motor cortex as indicated by neuromagnetic recordings.
Neurosci Lett 122(1):17–20
13. Lotze M, Kaethner RJ, Erb M, Cohen LG, Grodd W, Topka H
(2003) Comparison of representational maps using functional
magnetic resonance imaging and transcranial magnetic stimula-
tion. Clin Neurophysiol 114(2):306–312
14. Krings T, Buchbinder BR, Butler WE, Chiappa KH, Jiang HJ,
Rosen BR, Cosgrove GR (1997) Stereotactic transcranial mag-
netic stimulation: correlation with direct electrical cortical stimu-
lation. Neurosurgery 41(6):1319–1325
15. Neggers SF, Langerak TR, Schutter DJ, Mandl RC, Ramsey NF,
Lemmens PJ, Postma A (2004) A stereotactic method for image-
guided transcranial magnetic stimulation validated with fMRI and
motor-evoked potentials. Neuroimage 21(4):1805–1817
16. Sparing R, Buelte D, Meister IG, Paus T, Fink GR (2008)
Transcranial magnetic stimulation and the challenge of coil
placement: a comparison of conventional and stereotaxic neuro-
navigational strategies. Hum Brain Mapp 29(1):82–96
17. Denslow S, Bohning DE, Bohning PA, Lomarev MP, George MS
(2005) An increased precision comparison of TMS-induced motor
cortex BOLD fMRI response for image-guided versus function-
guided coil placement. Cogn Behav Neurol 18(2):119–126
18. Matthäus L, Trillenberg P, Bodensteiner C, Giese A, Schweikard
A (2006) Robotized TMS for motion compensated navigated
brain stimulation. Presented at the 20th International Computer
Assisted Radiology and Surgery Congress, Osaka, Japan, June
28–July 1.
19. Horn B (1987) Closed-form solution of absolute orientation using
unit quaternions. J Opt Soc Am A 4(4):629–642
20. Noirhomme Q, Ferrant M, Vandermeeren Y, Olivier E, Macq B,
Cuisenaire O (2004) Registration and real-time visualization of
transcranial magnetic stimulation with 3-d MR images. IEEE T
Biomed E 51(11):1994–2005
21. Matthäus L, Giese A, Wertheimer D, Schweikard A (2006)
Planning and analyzing robotized TMS using virtual reality. Stud
Health Technol Inform 119:373–378
22. Matthäus L, Trillenberg P, Fadini T, Finke M, Schweikard A
(2008) Brain mapping with transcranial magnetic stimulation
using a refined correlation ratio and Kendall’s tau. Stat Med 27
(25):5252–5270
23. Gumprecht H, Widenka D, Lumenta C (1999) BrainLab vectorvi-
sion neuronavigation system: technology and clinical experiences
in 131 cases. Neurosurgery 44(1):97–104
24. Kammer T, Beck S, Thielscher A, Laubis-Herrmann U, Topka H
(2001) Motor thresholds in humans: a transcranial magnetic
stimulation study comparing different pulse waveforms, current
directions and stimulator types. Clin Neurophysiol 112(2):250–258
25. Krings T, Buchbinder BR, Butler WE, Chiappa KH, Jiang HJ,
Cosgrove GR, Rosen BR (1997) Functional magnetic resonance
imaging and transcranial magnetic stimulation: complementary
approaches in the evaluation of cortical motor function. Neurol-
ogy 48(5):1406–1416
26. Macdonell RA, Jackson GD, Curatolo JM, Abbott DF, Berkovic
SF, Carey LM, Syngeniotin A, Fabinyi GC, Scheffer IE (1999)
Motor cortex localization using functional MRI and transcranial
magnetic stimulation. Neurology 53(7):1462–1467
27. Niyazov DM, Butler AJ, Kadah YM, Epstein CM, Hu XP (2005)
Functional magnetic resonance imaging and transcranial magnetic
stimulation: effects of motor imagery, movement and coil
orientation. Clin Neurophysiol 116(7):1601–1610
28. Rathelot JA, Strick PL (2006) Muscle representation in the
macaque motor cortex: an anatomical perspective. Proc Natl Acad
Sci USA 103(21):8257–8262
29. Teitti S, Määttä S, Säisänen L, Könönen M, Vanninen R, Hannula
H, Mervaala E, Karhu J (2008) Non-primary motor areas in the
human frontal lobe are connected directly to hand muscles.
Neuroimage 40(3):1243–1250
30. Epstein CM, Meador KJ, Loring DW, Wright RJ, Weissman JD,
Sheppard S, Lah JJ, Puhalovich F, Gaitan L, Davey KR (1999)
Localization and characterization of speech arrest during trans-
cranial magnetic stimulation. Clin Neurophysiol 110(6):1073–1079
31. Roux FE, Boulanouar K, Lotterie JA, Mejdoubi M, LeSage JP,
Berry I (2003) Language functional magnetic resonance imaging
in preoperative assessment of language areas: correlation with
direct cortical stimulation. Neurosurgery 52(6):1335–1345
32. Seyal M, Masuoka LK, Browne JK (1992) Suppression of
cutaneous perception by magnetic pulse stimulation of the human
brain. Electroencephalogr Clin Neurophysiol 85(6):397–401
342S.R. Kantelhardt et al.

Page 11

33. Amassian VE, Cracco RQ, Maccabee PJ, Cracco JB, Rudell A,
Eberle L (1989) Suppression of visual perception by magnetic coil
stimulation of human occipital cortex. Electroencephalogr Clin
Neurophysiol 74(6):458–462
34. Corthout E, Hallett M, Cowey A (2003) Interference with vision
by TMS over the occipital pole: a fourth period. NeuroReport 14
(4):651–565
35. Pascual-Leone A, Gates JR, Dhuna A (1991) Induction of speech
arrest and counting errors with rapid-rate transcranial magnetic
stimulation. Neurology 41(5):697–702
36. Borckardt JJ, Smith AR, Reeves ST, Weinstein M, Kozel FA,
Nahas Z, Shelley N, Branham RK, Thomas KJ, George MS
(2007) Fifteen minutes of left prefrontal repetitive transcranial
magnetic stimulation acutely increases thermal pain thresholds in
healthy adults. Pain Res Manag 12(4):287–290
37. Dell’Osso B, Mundo E, D’Urso N, Pozzoli S, Buoli M, Ciabatti
M, Rosanova M, Massimini M, Bellina V, Mariotti M, Carlo
Altamura A (2009) Augmentative repetitive navigated transcranial
magnetic stimulation (rTMS) in drug-resistant bipolar depression.
Bipolar Disord 11(1):76–81
Ri-TMS for somatotopic mapping of the motor cortex 343