Epilepsia, 44(5):688–692, 2003
Blackwell Publishing, Inc.
© 2003 International League Against Epilepsy
EEG-related Functional MRI in Benign Childhood Epilepsy with
Centrotemporal Spikes
Stephan Boor*, Goran Vucurevic*, Christian Pfleiderer*, Peter Stoeter*, Georg Kutschke†,
Rainer Boor†
Institute for Neuroradiology* and University Children’s Hospital, Pediatric Neurology† Johannes Gutenberg-University,
Mainz (Germany)
Abstract: The localization of epileptic foci is an important
issue in children with extratemporal epilepsies. However, the
value of noninvasive methods such as the EEG-assisted func-
tional magnetic resonance imaging (fMRI) has not been suffi-
ciently investigated in children.
As a model of extratemporal epilepsies, we studied 7 patients
aged 5 to 12 (median 10) years with benign childhood epilepsy
and centrotemporal (rolandic) spikes. Interictal spikes were re-
corded during the fMRI acquisition on a MR-compatible
battery-powered digital EEG system with 16 channels. The
fMRI sequences were correlated off-line with the EEG spikes
INTRODUCTION
The localization of epileptic foci is an important issue
in children with focal epilepsies. However, the value of
noninvasive methods of source analysis such as the
EEG-assisted functional magnetic resonance imaging
(fMRI) has not been sufficiently investigated in children.
Using a new aspect of the fMRI-technique, we studied
patients with benign rolandic epilepsy as a model of an
extratemporal epilepsy.
The benign childhood epilepsy with centrotemporal
(rolandic) spikes (BECTS), is an idiopathic focal epi-
lepsy with distinct sleep-enhanced interictal EEG parox-
ysmal activity, age-related onset, and benign course
(Commission on Classification and Terminology of the
ILAE [1]). Characteristic facial sensorimotor seizures of-
ten occur with involvement of oropharyngeal muscles,
but secondary generalization is quite common (2). Focal
seizures involving a limb or unilateral seizures are some-
times observed (3,4). The characteristic EEG pattern are
high–voltage centrotemporal sharp-waves followed by a
Accepted December 28, 2002.
Correspondence to: Dr. Stephan Boor, Institute for Neuroradiology,
Johannes Gutenberg-University, D-55101 Mainz, Germany Tel.: +49
6131 17-7139; Fax: +49 6131 17-6643 E-mail: boor@neuroradio.
klinik.uni-mainz.de
688
and analyzed with the software Statistical Parametrical Map-
ping SPM99.
The fMRI results demonstrated the spike-related activation
in the perisylvian central region in three patients; we could not
demonstrate fMRI activation despite active spiking in 2 pa-
tients, and 2 patients did not produce sufficient spikes for fMRI
analysis. We currently consider the spike-related fMRI as a
research tool that localizes epileptic activity in selected pa-
tients. Further improvements of the technique are necessary to
allow a clinical application of this method. Key Words:
fMRI—Functional imaging—BOLD—Epilepsy—BECTS.
slow wave and may include a tangential dipole formation
with a centrotemporal negativity and simultaneous fron-
tal positivity (5,6,7,8,9). The pathology underlying the
EEG foci is probably transmitted as an autosomal dom-
inant trait, and appear to be controlled by a single auto-
somal dominant gene with age-dependent penetrance and
expression (10,11,12). Neubauer et al. (13) found evi-
dence of linkage linkage to chromosome 15q14. The
clinical syndrome is considered as genetically heterog-
enous. Interictal EEG and MEG findings indicate that the
epileptic focus of BECTS is located in the inferior ro-
landic (face) or superior rolandic (hand) regions of the
sensorimotor cortex (5,9,14).
FMRI has been developed as a noninvasive tool for
the detection of focal changes of the blood supply as a
result of neuronal activity, using the blood oxygenation
level-dependant (BOLD) effect (15). EEG-triggered
fMRI, where the fMRI data is acquired at a fixed interval
following an EEG spike and at rest periods has been used
to detect regions of cerebral activation resulting from
interictal epileptiform activity in patients with focal epi-
lepsies (16,17,18,19,20). However, triggered images are
prone to intensity changes of successive acquisitions, es-
pecially at the beginning of the sequence. The delayed
image acquisition and the lack of pre-activity baseline
images may cause further problems. As the EEG-
 FMRI IN BENIGN CHILDHOOD EPILEPSY
triggered fMRI cannot be used in patients with a high
rate of spontaneous spikes, we used an event-related
fMRI analysis of the epileptiform activity with simulta-
neous and continuous acquisition of EEG and fMRI as
recently described (21,22,23,24,25).
METHODS
We studied 7 consecutive patients aged 5 to 12 (me-
dian 10) years with BECTS. The study was approved by
the local ethics committee and all parents gave their in-
formed consent. The diagnosis of BECTS included a
centrotemporal sharp-wave focus on the EEG, seizure
onset between the ages of 4 and 14 years, normal neu-
rological examination, absence of brain lesions, and
compatible seizure types. Three patients had focal oro-
pharyngeal seizures, one patient had focal motor and
sensory seizures of the hand, and three patients had gen-
eralized seizures, which were supposed to be secondarily
generalized focal seizures. The EEG focus was strictly
unilateral, without side shifting in our patients. The
structural MRI was normal in all children. Five patients
were on antiepileptic medication with Sultiame, one pa-
tient was treated with Sultiame and Valproate, and one
patient was without medication. The patients were stud-
ied after sleep deprivation but were not sedated. The
procedure was well tolerated by all patients.
The MRI studies were conducted at a routine high field
scanner at 1.5 T (Siemens Magnetom Vision, Erlangen,
Germany) using the standard imaging circular polarized
(CP) head coil. “Functional” desoxyhaemoglobin-sensitive
MRI recordings were performed with EPI-T2*-
sequences (BOLD-imaging) based on a blipped gradient-
echo echoplanar imaging sequence (TR⳱6000 ms,
FIG. 1. Interictal stereotyped
sharp-waves with the maximum
over the right centrotemporal re-
gions can be identified in the EEG
between the artifacts (marked with
grey color) of the fMRI gradients.
No reliable EEG and no spikes
can be identified during the
artifacts.
689
TE⳱66 ms, flip angle 90°, FOV 210 mm, 128×128 im-
age matrix) and 6 mm slice thickness. The slice distance
was set to 10% of the slice thickness. A multislice ac-
quisition of 16 ascending sections of the whole head re-
sulted in a total measuring time of 2400 ms leaving a 3600
ms delay before image repetition. One examination com-
prised a total of 300 scans with 30 min of measuring time.
Continuous EEG recordings in the scanner were per-
formed with a commercially available MR-compatible
battery-powered EEG system and 16 sintered Ag/AgCl
scalp electrodes with shielded leads (EMR16, Schwar-
zer, München, Germany). The EEG electrodes were
placed on the scalp according to 10–20 standard posi-
tions using an electrode cap. We used conventional ad-
hesive electrode gel. Skin-electrode impedance was
below 10 k⍀ and the electrode wires were fixed on the
electrode cap, bound together and twisted from the top of
the head to the amplifier to reduce pulse artifacts. The
EEG system used a nonferrous headbox, which was
placed inside the magnet bore about 50 cm from the
patient’s head. The EEG was sampled at a rate of 1 kHz.
The digitized information was transmitted by a fiber op-
tic cable to a personal computer outside the magnet
room. The EEG was displayed using a common refer-
ence. We did not use artifact correction. During the pe-
riods between the gradient artifacts, the continuous EEG
recording remained sufficiently undisturbed to allow an
unambiguous retrospective evaluation off-line. For com-
parison, we started with an EEG-recording outside the
scanner room, on the same MR-compatible EEG system
and the same montage as used inside the magnet. Imme-
diately afterwards, the patient was placed into the mag-
net, with the electrodes still in place. Four of the patients
fell asleep during the fMRI scanning.
Epilepsia, Vol. 44, No. 5, 2003
690 S. BOOR ET AL.

FIG. 2. A-C: SPM{t} maps in the patients 1-3, thresholded at P

< 0.05 (corrected). The color-coded statistical parameter maps
are projected onto coronal and transversal slices of the mean T2*
image volumes, showing activation in the perisylvian right central
regions in all patients, including the face area in all 3 patients and
extending into the hand area in the patient 3 (2C).

The fMRI data were realigned and smoothed (Gauss- off-line and every sharp-wave occurring in the periods
ian kernel; full width at half maximum: 8×8×12 mm) that were free from gradient artifacts, was marked. The
using the software Statistical Parametrical Mapping assignment of these spikes to specific image numbers
(SPM99) (26,27). The rating of the EEG was performed was facilitated by the regular occurrence of EEG artifacts

Epilepsia, Vol. 44, No. 5, 2003

 FMRI IN BENIGN CHILDHOOD EPILEPSY
caused by gradient switching, which were marked and
numbered. The EEG evaluation yielded a vector consist-
ing of zeros for periods without spikes and an accounting
for the number of spikes preceding each fMRI volume
acquisition. This vector was used to create a design ma-
trix for the event-related fMRI analysis. High pass fil-
tering, implemented in SPM99, was used to eliminate
long-term signal changes exceeding 128 seconds. The
resulting SPM{t} maps were thresholded at P < 0.05
(corrected). The statistical parameter maps were color-
coded and coregistered to the mean T2* image volumes.
RESULTS
All patients had frequent interictal stereotype sharp-
waves with their maximum over the centrotemporal re-
gions in their sleep-EEG, consistent with the diagnosis of
BECTS. The combined EEG/fMRI analysis allowed the
identification of the sharp-waves during the scanning
pauses. The distortion of the EEG traces was largely
confined to the actual switching of magnetic field gradi-
ents during the respective imaging periods and the qual-
ity of the EEG during the scanning pauses (3.6 s out of
6 s) was sufficient to identify epileptic activity for sub-
sequent fMRI analysis. Figure 1 illustrates a typical EEG
with sharp-waves over the right centrotemporal region in
patient 2 recorded from inside the magnet during the
fMRI-examination.
Four patients fell asleep during the fMRI examination.
We found extended fMRI activations in the perisylvian
right central regions in three patients, patients 1 and 2
asleep (Fig. 2A-C). The activation includes the face area
in all 3 patients and extended into the hand area in patient
3. We could not identify spikes during the fMRI acqui-
sition, neither in the EEG recording immediately before
the fMRI acquisition (outside the scanner) nor during the
fMRI recording in two patients, who did not fall asleep.
Two other patients, who had fallen asleep had centro-
temporal sharp-waves, but we were not able to demon-
strate fMRI activations. We did not find differences in
the amplitude, localization, or complexity of the spikes
between these two patients and those who produced
fMRI activations.
DISCUSSION
Spontaneous sharp-waves with the typical morphol-
ogy and scalp distribution of rolandic spikes occurred
over the centrotemporal regions of all subjects and rep-
resented one of the inclusion criteria for this study. We
used the event-related fMRI analysis from continuous
and simultaneous EEG/fMRI acquisition to detect the
focal changes of neuronal activity that were correlated
with the EEG events (24). In comparison to triggering,
continuously running MR recording provides a more
691
constant and hardly disruptive auditory environment for
the subjects and many of our patients spontaneously fell
asleep, enhancing the spike activity.
We found large zones of BOLD activation in the
perisylvian right central regions in three patients,
reaching into the hand area in one patient. The high
amplitude, the relatively prolonged duration, and the
bluntness of the rolandic spikes suggest that they are
generated by an extensive neuronal pool (28). Our zones
of BOLD activation are consistent with previous findings
from the literature in BECTS on MEG and EEG di-
pole modelling. Using a multichannel MEG system,
Minami et al. (8) demonstrated that rolandic discharges
often originate from the somatosensory cortex of the
lower lip territory. Kamada et al. (14) divided the
estimated dipole locations into superior (hand motor)
and inferior (oromotor) rolandic epileptogenic foci. The
dipoles of the negative sharp-wave were localized deep
in the rolandic fissure and included the pre- and postcen-
tral gyri.
Unfortunately, we could not demonstrate fMRI acti-
vation in 2 patients with BECTS, who had spike activity
during the fMRI acquisition. Future progress of the tech-
nique is necessary to facilitate a reliable clinical appli-
cation of the spike related fMRI. This might include the
improvement of the EEG quality in the scanner by the
subtraction of the pulse artifacts (29) and the reduction
of the artifacts due to scanner activity (30,31). These
methods were not available for us during the study pe-
riod. If we could detect all spikes that occur during the
fMRI examination, even during the periods of gradient
switching, we might improve the statistical power of
the analysis and possibly the sensitivity of the EEG-
related fMRI.
A still unresolved problem of the fMRI technique is
the degree to which the EEG-correlated activations cor-
respond to the patient’s primary epileptic zone. Since
epileptic activity may propagate to other cortical regions,
it cannot be excluded that BOLD fMRI responses occur
not only in the primary epileptic zone but also in cortical
projections of such regions. It is even unclear, whether
the fMRI activations indicate true foci of epileptic activ-
ity or if they indicate the processing of induced activity
or both (21). The combination of EEG related fMRI with
the technique of source analysis of interictal spikes (32)
might be useful to discriminate the onset of epileptiform
activity from propagated sources and to differentiate be-
tween epileptic foci and induced activity (23). However,
given the small number of patients tested so far, and the
relatively low sensitivity of the method as indicated by
the current study, the fMRI technique should be currently
considered as an experimental method. Further improve-
ments of the technique would be necessary to allow a
clinical application of this method.
Epilepsia, Vol. 44, No. 5, 2003
692 S. BOOR ET AL.
Acknowledgements: This paper includes parts of the thesis
of C. Pfleiderer.
REFERENCES
1. Commission on Classification and Terminology of the ILAE: Pro-
posal for a revised classification of epilepsies and epileptic syn-
dromes. Epilepsia 1989;30:389–99.
2. Stephani U. Typical semiology of benign childhood epilepsy with
centrotemporal spikes (BCECTS). Epileptic Disord 2000;2 (Suppl.
1):S3–S4.
3. Aicardi J. Atypical semiology of rolandic epilepsy in some related
syndromes. Epileptic Disord 2000;2 (Suppl. 1):S5–9.
4. Saint-Martin A de, Carcangiu R, Arzimanoglou A, Massa R,
Thomas P, Motte J, Marescaux C, Metz-Lutz M-N, Hirsch E.
Semiology of typical and atypical rolandic epilepsy: a video-EEG
analysis. Epileptic Disord 2001;3:173–81.
5. Baumgartner C, Graf M, Doppelbauer A, Serles W, Lindinger G,
Olbrich A, Bacher J, Pataraia E, Almer G, Lischka A. The func-
tional organization of the interictal spike complex in benign rolan-
dic epilepsy. Epilepsia 1996;37:1164–74.
6. Gastaut Y. Un élément déroutant de la séméiologie électroén-
cephalographique: les pointes prérolandiques sans signification fo-
cale. Rev Neurol 1952;87:488–9.
7. Graf M and Lischka A. Topographical EEG analysis of rolandic
spikes. Clin Electroencephalogr 1998;29:132–7.
8. Minami T, Gondo K, Yamamoto T, Yanai S, Tasaki K, Ueda K.
Magnetoencephalographic analysis of rolandic discharges in be-
nign childhood epilepsy. Ann Neurol 1996;39:326–34.
9. Van der Meij W, Wieneke GH, van Huffelen AC. Dipole source
analysis of rolandic spikes in benign rolandic epilepsy and other
clinical syndromes. Brain Topography 1993;5:203–13.
10. Bray PF, Wiser WC. Evidence for a genetic etiology of temporal-
central abnormalities in focal epilepsy. N Engl J Med 1964;271:
926–33.
11. Doose H, Brigger-Heuer B, Neubauer BA. Children with focal
sharp-waves: clinical and genetic aspects. Epilepsia 1997;38:788–96.
12. Heijbel J, Blom S, Rasmuson M. Benign epilepsy of childhood
with centrotemporal EEG foci. A genetic study. Epilepsia 1975;
16:285–93.
13. Neubauer BA, Fiedler B, Himmelein B, Kämpfer F, Läßker U,
Schwabe G, Spanier I, Tams D, Bretscher C, Moldenhauer K,
Kurlemann G, Weise S, Tedroff C, EEG-Olofsson O, Wadelius C,
Stephani U. Centrotemporal spikes in families with rolandic epi-
lepsy: Linkage to chromosome 15q14. Neurology 1998;51:1608–12.
14. Kamada K, Möller M, Saguer M, Kassubek J, Kaltenhäuser M,
Kober H, Überall M, Lauffer H, Wenzel D, Vieth J. Localization
analysis of neuronal activities in benign rolandic epilepsy using
magnetoencephalography. J Neurol Sci 1998;154:164–72.
15. Ogawa, S., Lee, T. M., Kay, A. R., Tank, D. W. Brain magnetic
resonance imaging with contrast dependent on blood oxygenation.
Proc Natl Acad Sci USA 1990;87,9868–72.
16. Jackson GD, Connelly A, Cross JH, Gordon I, Gadian DG. Func-
tional magnetic resonance imaging of focal seizures. Neurology
1994;44:850–6.
17. Warach S, Ives JR, Schlaug G, Patel MR, Darby DG, Thangaray V,
Epilepsia, Vol. 44, No. 5, 2003
Edelman RR, Schomer DL. EEG-triggered echo-planar functional
MRI in epilepsy. Neurology 1996;47:89–93.
18. Seeck M, Lazeyras F, Michel CM, Blanke O, Gericke CA, Ives J,
Delavelle J, Golay X, Haenggeli CA, de-Tribolet N, Landis T.
Non-invasive epileptic focus localization using EEG-triggered
functional MRI and electromagnetic tomography. Electroencepha-
logr Clin Neurophysiol 1998;106:508–12.
19. Krakow K, Woermann FG, Symms MR, Allen PJ, Lemieux L,
Barker GJ, Duncan JS, Fish DR. EEG-triggered functional MRI of
interictal epileptiform activity in patients with partial seizures.
Brain 1999;122:1679–88.
20. Lazeyras F, Blanke O, Perrig S, Zimine I, Golay X, Delavelle J,
Michel CM, de Tribolet N, Villemure JG, Seeck M. EEG-triggered
functional MRI in Patients with pharmacoresisatant epilepsy. J
Magn Reson Imaging 2000;12:177–85.
21. Baudewig J, Bittermann HJ, Paulus W, Frahm J. Simultaneous
EEG and functional MRI of epileptic activity: a case report. Clin
Neurophysiol 2001;112:1196–1200.
22. Boor S, Vucurevic G, Kutschke G, Boor R, Nedelmann M, Stoeter
P. FMRI unter Nutzung einer simultanen EEG-Ableitung bei Epi-
lepsie, Vergleich mit einer EEG-Quellenanalyse. Röntgen
Fortschr. 2000;172:Suppl. 1:28.
23. Boor R, Vucurevic G, Kutschke G, Boor S. EEG source analysis
and EEG-assisted functional MRI in benign childhood epilepsy
with centrotemporal spikes. Clin Neurophysiol 2001;112:Suppl 1:
S10.
24. Lemieux L, Salek-Haddadi A, Josephs O, Allen P, Toms N, Scott
C, Krakow K, Turner R, Fish DR. Event-related fMRI with simul-
taneous and continuous EEG: Description of the method and initial
case report. Neuroimage 2001;14:780–7.
25. Vucurevic G, Boor S, Boor R, Kutschke G, Stoeter P. fMRI and
concurrent EEG in patients with focal epilepsy: Processing of EEG
signals for fMRI analysis, Proc. Intl. Soc. Mag. Reson. Med 2001;
9:1745.
26. Friston KJ, Holmes AP, Worsley KJ, Poline JP, Frith CD, Frack-
owiak RSJ. Statistical parametrical maps in fuctional imaging: A
general linear approach. Hum Brain Mapping 1995;2:189–210.
27. Ashburner J, Friston KJ. Nonlinear Spatial normalization using
basic functions. Hum Brain Mapping 1999;7:254–66.
28. Kellaway P. The electroencephalographic features of benign cen-
trotemporal (rolandic) epilepsy of childhood. Epilepsia 2000;
41:1053–6.
29. Allen PJ, Polizzi G, Krakow K, Fish DR, Lemieux L. Identification
of EEG events in the MR scanner: The problem of pulse artifact
and a method for its subtraction. Neuroimage 1998;8:229–39.
30. Allen PJ, Josephs O, Turner R. A method for removing imaging
artifact from continuous EEG recorded during functional MRI.
Neuroimage 2000;12:230–9.
31. Hoffmann A, Jäger L, Werhahn KJ, Jaschke M, Noachtar S, Reiser
M. Electroencephalography during functional echo-planar imag-
ing: detection of epileptic spikes using post-processing methods.
Magn Reson Med 2000;44:791–8.
32. Scherg M, Bast T, Berg P. Multiple source analysis of interictal
spikes: goals, requirements, and clinical value. J Clin Neurophysiol
1999;16:214–24.