1296

Case Reports / Journal of Clinical Neuroscience 15 (2008) 1296–1298

Tuberous sclerosis: Localizing the epileptogenic tuber with
synthetic aperture magnetometry with excess kurtosis analysis
Leonides Canuet a,b,*, Ryouhei Ishii a, Masao Iwase a, Ryu Kurimoto a, Koji Ikezawa a,
Michiyo Azechi a, Mari Wataya-Kaneda c, Masatoshi Takeda a
a

Departments of Clinical Neuroscience and Psychiatry, Osaka University Graduate School of Medicine, Osaka, Japan
b
Department of Neurology, ‘‘Saturnino Lora’’ Provincial Hospital, Santiago de Cuba, Cuba
c
Department of Dermatology, Osaka University Graduate School of Medicine, Osaka, Japan
Received 19 January 2007; accepted 18 March 2007

Abstract
The hallmark of tuberous sclerosis is the presence of multiple cortical tubers. Identifying the epileptogenic tubers is difficult and often
requires invasive intracranial electroencephalograph (EEG) monitoring. We report on a patient with tuberous sclerosis upon whom the
novel magnetoencephalography (MEG) technique of synthetic aperture magnetometry (SAM) with excess kurtosis (g2) analysis was performed for localization of the epileptogenic tuber. Simultaneous electroencephalography (EEG) was also performed. MEG data, as analyzed by SAM(g2), were superimposed on the patient’s MRIs. In the fluid attenuated inversion recovery MRIs, several tubers and
subependymal nodules could be identified, with the largest tubers being located in the right frontal and left anteriotemporal regions.
Despite multiple cortical lesions existing, the SAM(g2) images showed a single large tuber and surrounding epileptogenic tissue in the
left temporal cortex. We suggest that MEG with SAM(g2) analysis may be clinically useful for the accurate identification of epileptogenic
tubers in patients with tuberous sclerosis.
Ó 2007 Elsevier Ltd. All rights reserved.
Keywords: Tuberous sclerosis; Magnetoencephalography; Synthetic aperture magnetometry; Kurtosis; Epileptogenic tuber

1. Introduction
Multimodality neuroimaging has improved the detection
of epileptogenic foci, allowing an increasing number of individuals to be evaluated for surgery.1 However, because
localization of epileptic foci remains challenging, there is
an urgent need to improve the methods by which epileptogenic tubers are detected. Synthetic aperture magnetometry
(SAM)2,3 with excess kurtosis (g2)4,5 is a recently developed
analysis technique that transforms magnetoencephalography (MEG) data into a functional image of spike-like activity, giving the source waveforms for those locations. This
technique seems to be appropriate for localizing epileptogenic tubers in tuberous sclerosis. In the current study we
illustrate the value of the combination of MEG and
SAM(g2) for the accurate interictal localization of an epileptogenic tuber in a patient with focal epilepsy as a symptom of tuberous sclerosis.
2. Clinical presentation
A 29-year-old female patient was diagnosed with tuberous sclerosis, in accordance with the revised clinical
*
Corresponding author. Present address: Department of Psychiatry,
Osaka University Graduate School of Medicine, Yamadaoka 2–2, D-3,
Suita City 565–0871, Japan. Tel.: +81 6 6879 3051; fax: +81 6 6879 3059.
E-mail address: leocanon2002@yahoo.com.mx (L. Canuet).

diagnostic criteria for tuberous sclerosis complex.6 Symptomatic focal epilepsy was also diagnosed, with age of onset being 14 years. After epilepsy appeared, the patient was
followed up at the outpatient clinic of psychiatry, Osaka
University Hospital, Japan. The patient experienced frequent partial complex seizures, suggestive of a temporal focus. Seizures were characterized by behavioral arrest and
staring, usually accompanied by fear, nausea, mild cyanosis, and, rarely, secondary generalized seizures. Previous
electroencephalography (EEG) recordings often showed
bilateral epileptic discharges in the temporal regions. The
patient underwent an MEG measurement using a 64-channel whole head magnetometer (NeuroSQUID Model 100,
CTF Systems Inc., Port Coquitlam, BC, Canada). The data
sample rate was 250 Hz per channel. To assist in spike
detection, simultaneous EEG data were recorded using 19
electrodes (International 10–20 system). MEG demonstrated clear spike activity over the left temporal region
(Fig. 1).
MEG data were analyzed by using synthetic SAM(g2).
This functional imaging analysis technique combines a
three-dimensional functional image of excess kurtosis, g2,
considered a measure of spike-like activity, and the estimation of source waveforms at the coordinates identified in
the functional image by SAM virtual sensors. The presence
of spike activity yields a large positive kurtosis, whereas
normal cortical activity exhibits a small kurtosis, and noise

Case Reports / Journal of Clinical Neuroscience 15 (2008) 1296–1298

1297

Fig. 1. Magnetoencephalography (MEG) waveforms and topographic maps. Frequent sharp waves were identified on the MEG recording. Threedimensional whole-head topographic maps show the corresponding source localization of this epileptic activity over the left anterior temporal region.

has an excess kurtosis of zero. Epileptic activity was imaged by SAM(g2) for the 20–70 Hz frequency band, which
is known to provide optimal image contrast for interictal
spike activity.
SAM(g2) images were superimposed on T1-weighted
MRIs, and compared with the patient’s fluid-attenuated
inversion recovery (FLAIR) MRI slices, as FLAIR has
proved to be more sensitive for detection of cortical tubers
than conventional T1- and T2-weighted images.7 The
FLAIR MRI sequences revealed several cortical tubers distributed widely in the brain, and subependymal nodules
(Fig. 2). The largest cortical tubers were located in the right
frontal and left anterior temporal regions.
Despite the patient having multiple and bihemispheric
cortical lesions, the SAM(g2) images clearly showed a single large tuber and surrounding epileptogenic tissue in the
left anterior temporal cortex (Fig. 3). This finding was consistent with the seizure semiology, and with that of the
simultaneous EEG, even though several MEG epileptic
spikes without corresponding EEG spikes were observed.

Fig. 2. Fluid attenuated inversion recovery (FLAIR) axial MRI. Widespread cortical tubers were visualized. (A) Subependymal nodules can also
be seen. (B–C) Large tubers were identified in the right frontal and left
temporal regions. The arrow indicates the location of the epileptogenic
tuber in the left anterior temporal cortex.

Fig. 3. Synthetic aperture magnetometry with excess kurtosis [SAM(g2)]
images superimposed on the three plane T1-weighted MRI. Epileptic
activity was imaged by SAM(g2) in the left anterior temporal region that
corresponded with the location of a large tuber and surrounding
epileptogenic tissue.

3. Discussion
To date, accurate localization of small and widespread
epileptogenic lesions is still challenging. MRI cannot be
used to determine whether a tuber has epileptic activity,
and it is often unable to detect all tubers.8 SPECT often
fails to detect epileptic foci in tuberous sclerosis.9 IctalEEG monitoring has an important role in identification
of the primary epileptogenic zone in patients with multiple
brain lesions, and it has been used with functional neuronavigation for the removal of epileptogenic lesions.10 Nevertheless, invasive EEG monitoring with intracranial
electrodes is generally required. Hence the interest in applying the recently developed MEG analysis technique,
SAM(g2), in tuberous sclerosis.
Several studies9,11 have shown that MEG is of value in
the noninvasive localization of interictal spike sources in
partial epilepsy, including tuberous sclerosis. Most of these
studies have used dipole analysis. This method is laborintensive and time-consuming, as it requires the manual
identification of spikes and considerable skill to minimize
errors. Although SAM(g2) has not yet been extensively

1298

Case Reports / Journal of Clinical Neuroscience 15 (2008) 1298–1301

used in symptomatic epilepsy, SAM automated analysis
appears to offer improved detection of irritable zones and
beneficial volumetric and frequency descriptions compared
to conventional dipole modeling.2
Our findings are consistent with those of other investigations,12 which have provided evidence suggesting that large
tubers correlate better with the EEG foci than small tubers;
our SAM(g2) findings revealed that the epileptic activity in
this patient arose from a large tuber and surrounding tissue. However, a large tuber involving the right frontal area
failed to exhibit epileptogenicity. Frontal tubers have been
reported to have the weakest correlation with epileptogenic
foci.13
SAM(g2) analysis was a useful tool for localizing the
epileptogenic tuber in this patient. This suggests that the
combination of the brain anatomical data provided by
MRI with the functional images of MEG and SAM(g2)
analysis may help identify epileptogenic zones, and differentiate between epileptogenic and nonepileptogenic tubers
in patients suffering from tuberous sclerosis. This may be
particularly important when evaluating intractable epileptic patients for surgical treatment. Further application of
SAM(g2) in large samples of patients with tuberous sclerosis and other focal epilepsies with multiple brain structural
lesions may confirm this argument.
References
1. Asano E, Chugani DC, Muzik O, et al. Multimodality imaging for
improved detection of epileptogenic foci in tuberous sclerosis
complex. Neurology 2000;54:1976–84.

2. Xiao Z, Xiang J, Holowka S, et al. Volumetric localization of epileptic
activities in tuberous sclerosis using synthetic aperture magnetometry.
Pediatr Radiol 2006;36:16–21.
3. Ishii R, Shinosaki K, Ukai S, et al. Medial prefrontal cortex generates
frontal midline theta rhythm. Neuroreport 1999;10:675–9.
4. Robinson SE, Nagarajan SS, Mantle M, et al. Localization of
interictal spikes using SAM(g2) and equivalent current dipole fit. J
Clin Neurophysiol 2005;22:362–5.
5. Ishii R, Canuet L, Iwase M, et al. Right parietal activation during
delusional state in episodic interictal psychosis of epilepsy: A report of
two cases. Epilepsy Behav 2006;9:367–72.
6. Roach ES, Gomez MR, Northrup H. Tuberous sclerosis complex
consensus conference: revised clinical diagnostic criteria. J Child
Neurol 1998;13:624–8.
7. Takanashi J, Sugita K, Fujii K, et al. MR Evaluation of tuberous
sclerosis: increased sensitivity with fluid-attenuated inversion recovery
and relation to severity of seizures and mental retardation. Am J
Neuroradiol 1995;16:1923–8.
8. Peressona M, Lopez L, Narici L, et al. Magnetic source imaging and
reactivity to rhythmical stimulation in tuberous sclerosis. Brain Dev
1998;20:512–8.
9. Kamimura T, Tohyama J, Oishi M, et al. Magnetoencephalography
in patients with tuberous sclerosis and localization-related epilepsy.
Epilepsia 2006;47:991–7.
10. Karatas A, Erdem A, Savas A, et al. Identification and removal of an
epileptogenic lesion using Ictal-EEG, functional-neuronavigation and
electrocorticography. J Clin Neurosci 2004;11:343–6.
11. Oishi M, Kameyama S, Masuda H, et al. Single and multiple clusters
of magnetoencephalographic dipoles in neocortical epilepsy: significance in characterizing the epileptogenic zone. Epilepsia 2006;47:
355–64.
12. Ohmori I, Ohtsuka Y, Ohno S, et al. Analysis of ictal EEGs of
epilepsy associated with tuberous sclerosis. Epilepsia 1998;39:1277–83.
13. Tamaki K, Okuno T, Ito M, et al. Magnetic resonance imaging in
relation to EEG epileptic foci in tuberous sclerosis. Brain Dev
1990;12:316–20.

doi:10.1016/j.jocn.2007.03.030

Primary intracranial low-grade fibromyxoid sarcoma (Evans tumor)
Kagan Tun a,*, Ozlem Ozen b, Erkan Kaptanoglu a, Oktay Gurcan a,
Ethem Beskonakli a, Bulent Celasun b
a

Department of Neurosurgery, Ankara Numune Education and Research Hospital, Ankara, Turkey
b
Department of Pathology, Baskent University, Ankara, Turkey
Received 28 June 2007; accepted 24 July 2007

Abstract
Low-grade fibromyxoid sarcoma was first described in 1987 as a rare soft tissue neoplasm characterized by a bland and deceptively
benign histological appearance but with aggressive behavior. A 20-year-old male patient presented with a recent history of headache and
seizure. A right frontal mass was detected on MRI and he was operated upon to remove the intracranial mass. Histological examination
revealed mildly atypical fibroblastic cells embedded within a myxoid matrix. Nuclear atypia and pleomorphism were minimal, and necro-

*

Corresponding author. Present address: Haymana yolu, Karsıyaka mah. Gorkem Sitesi, 355 sok. No: 8, 06830 Golbasi, Ankara, Turkey. Tel.: +90 312
4840933.
E-mail address: kagantun@yahoo.com (K. Tun).