- Email: [email protected]
Focal cortical high-frequency oscillations trigger epileptic spasms: Confirmation by digital video subdural EEG
Clinical Neurophysiology 116 (2005) 2819–2825 www.elsevier.com/locate/clinph
Focal cortical high-frequency oscillations trigger epileptic spasms: Confirmation by digital video subdural EEG Tomoyuki Akiyama, Hiroshi Otsubo*, Ayako Ochi, Taichi Ishiguro, Gemmu Kadokura, Rajesh RamachandranNair, Shelly K. Weiss, James T. Rutka, O. Carter Snead III Divisions of Neurology, Neurosurgery, The Hospital for Sick Children, 555 University Avenue, Toronto, Ont., Canada M5G 1X8 Accepted 23 August 2005 Available online 25 October 2005
Abstract Objective: To localize high-frequency oscillations (HFOs) on the cortex during epileptic spasms using video subdural EEG and Multiple Band Frequency Analysis (MBFA). Methods: Using video subdural EEG sampled at 1 kHz, we studied a 14-year-old boy with asymmetric epileptic spasms of possible left frontal origin. We identified HFOs, then analyzed and localized their distributions by MBFA. We correlated HFO distribution to clinical spasm intensity. Results: Ictal subdural EEG recorded HFOs at 60–150 Hz lasting 0.3–4 s. MBFA showed extensive but noncontiguous distribution of HFOs predominantly over the left frontal and temporal regions. HFOs began and became quasiperiodic before manifestation of clinical spasms. As clinical spasms intensified, HFOs persisted in regions where they initiated subclinically but were of higher frequency and greater power than HFOs in other regions. We performed cortical resections over the left frontal and temporal regions with predominant HFOs. Six months after surgery, the patient remained seizure free. Conclusions: HFOs were present over the ictal onset zone during epileptic spasms. Periodic spasms in this patient had the characteristics of partial seizures. Significance: We show that HFOs occurred over the cerebral cortex during epileptic spasms, and we suggest that these focal cortical HFOs triggered the spasms. q 2005 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. Keywords: High-frequency oscillation; Epileptic spasms; Subdural EEG; Multiple band frequency analysis; Partial seizure
1. Introduction Epileptic spasms in clusters commonly occur in generalized epilepsies that are generated mainly in the subcortex, such as the pediatric epileptic syndromes known as Ohtahara, West, and Lennox-Gastaut (Ohtsuka et al., 2001). In 1987, Gobbi et al. proposed the term periodic spasms for epileptic spasms that occur in older patients with localization-related epilepsy. They suggested that periodic spasms, characterized by asymmetric clinical and electrographical features, were partial seizures that originated in the cortex with a particular type of secondary generalization * Corresponding author. Tel.: C1 416 813 6660. fax: C1 416 813 6334. E-mail address: [email protected] (H. Otsubo).
(Gobbi et al., 1987, 1991). Although West syndrome with infantile spasms has been classified as generalized epilepsy, Chugani et al. found that focal cortical resection can control seizures in some patients (Chugani et al., 1993). In addition, when measuring glucose metabolism with positron emission tomography in West syndrome patients, they found abnormal rates in both the cortex and the brainstem (Chugani et al., 1992). Their findings suggest an underlying abnormal interrelationship between cortical and subcortical structures in infantile spasms. On ictal scalp EEG, epileptic spasms in clusters consist of high-voltage slow waves with superimposed fast activity in the beta frequency followed by an electrodecremental activity (Gobbi et al., 1987; Fusco and Vigevano, 1993). Fusco and Vigevano reported a medium-amplitude fast
1388-2457/$30.00 q 2005 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2005.08.029
2820
T. Akiyama et al. / Clinical Neurophysiology 116 (2005) 2819–2825
activity without clinical spasms during the preictal period and at the end of a cluster (Fusco and Vigevano, 1993). Until newer technologies were developed, the limitations of analog scalp EEG prevented detailed analysis of the fast activity of clinical spasms. In 2004, using digital scalp EEG with high sampling frequency, Kobayashi et al. discovered very fast rhythmic activities at 51–98 Hz during episodes of spasms in 11 patients with West syndrome and related disorders (Kobayashi et al., 2004). From these very fast activities in the posterior brains of thin-skulled young children, they inferred that the cortex played a major role in generating epileptic spasms. Using intracranial EEG recordings, others (Alarcon et al, 1995; Allen et al., 1992; Fisher et al., 1992; Traub et al., 2001; Worrell et al., 2004) observed very fast activities ranging from 20 to 120 Hz at the beginning of partial seizures in the ictal onset zone. In addition, Bragin et al. recorded 250–500 Hz high-frequency oscillations (HFOs), termed fast ripples, from the epileptogenic entorhinal cortex and hippocampus during the interictal state (Bragin et al., 2002). This finding suggested that the distribution of HFOs associated with epileptic seizures may localize the epileptogenic zone. We had one patient with periodic spasms, who underwent epilepsy surgery after extraoperative video subdural EEG recording. We hypothesized that the video intracranial digital EEG with high sampling rate could record focal HFOs during epileptic spasms and show that the propagation of HFOs over the cortex was associated with the initiation of spasms.
2. Case history 2.1. Patient Our patient was a 14-year-old left-handed boy who had experienced seizures since he was 8 years, 9 months of age. He had no family history of seizure disorders and no remarkable past history. Findings from his physical and neurological examinations were normal. His seizures consisted of grimacing for about 1 s followed, at times, by turning his head to the left and raising his shoulders. These brief spasms occurred quasiperiodically every 3–30 s in clusters; cluster episodes lasted 30–60 min. Between each spasm, the boy looked confused but was able to converse. Cluster episodes occurred once or twice a day, usually on waking or falling asleep. He experienced no other types of seizures. His seizures were refractory to carbamazepine, lamotrigine, clobazam, oxcarbazepine and gabapentin. At admission for surgical evaluation, he was being treated with oxcarbazepine and gabapentin. Neuropsychological tests revealed slightly impaired visual reasoning and some deficits in visual memory and verbal productivity.
2.2. EEG When 13 years old, the patient underwent prolonged (48 h) video scalp EEG with electrodes placed in the International 10–20 scalp-electrode position and a single reference electrode at Oz (HARMONY 5.4, Stellate, Montreal, PQ, Canada). Interictal scalp EEG showed high-amplitude spikeor sharp-slow wave discharges over the left frontal region (Fig. 1A). Focal interictal discharges became, at times, rhythmic trains of 2–2.5 Hz and extended to the centrotemporal regions. Background activity consisted of a posterior-dominant symmetric 8–9 Hz alpha rhythm, and sleep features were normal. We captured 2 cluster episodes, each containing spasms that lasted 1–3 s in 5–25 s intervals. Ictal video scalp EEG showed high-amplitude sharplycontoured slow waves with superimposed 15–30 Hz fast waves, predominantly over the left fronto-centro-temporal and midline regions, associated with each spasm (Fig. 1B). 2.3. MRI MR images of the patient at 13 years were normal. (GE 1.5T Signa MRI, GE Medical Systems, Milwaukee, WI, USA) 2.4. MEG At age 13, the patient also underwent MEG studies with a whole-head gradiometer Omega system (151 channels, VSM MedTech Ltd., Port Coquitlam, BC, Canada). During MEG, he also underwent simultaneous EEG that was recorded from 19 electrodes (International 10–20 system). We recorded 2 min periods of spontaneous MEG data 15 times. The sampling rate for data acquisition was 625 Hz. We visually identified MEG epileptic discharges, spikes and sharp waves (referred to as spikes), by examining the MEG recordings and cross-referencing them with the simultaneous EEG recording using a band pass filter of 3–70 Hz and a notch filter of 60 Hz. We applied a single moving dipole analysis with a single-shell, whole-head spherical model. We defined the MEG spike dipole for each spike as a single dipole fit from the earliest phase of each spike with the criteria of a residual error of less than 30%. The patient did not have spasms during MEG recording. MEG revealed a cluster of 59 MEG spike sources in the left inferior to superior frontal gyri accompanied by left frontocentro-temporal EEG interictal epileptic discharges (Fig. 1C). Temporal discharges consistently succeeded frontal discharges in the complex of MEG spikes, and we found no dipoles satisfying our criteria in the temporal region. 2.5. Extraoperative video subdural EEG We constructed a 106-channel grid, based on 3D MR images, interictal and ictal scalp EEG results, MEG spike source locations, and MEG-somatosensory evoked fields. The grid electrodes were 5 mm in diameter, embedded in
T. Akiyama et al. / Clinical Neurophysiology 116 (2005) 2819–2825
2821
Fig. 1. (A) A–P bipolar montage scalp interictal EEG shows intermittent high-amplitude spike- or sharp-slow wave complexes over the left frontal region. (band pass filter, 0.5–100 Hz; notch filter, 60 Hz). (B) A–P bipolar montage ictal scalp EEG shows high-amplitude sharply-contoured slow waves with superimposed 15–30 Hz fast activities and muscle artifacts, predominantly over the left frontal and midline regions at the time of clinical epileptic spasm (*). (band pass filter, 0.5–100 Hz; notch filter, 60 Hz). (C) Sagittal T1-weighted MRI shows a cluster of MEG dipoles with random orientations (closed triangles, locations of spike sources; tails, orientation and strength of dipole moments) over the left inferior to superior frontal gyri accompanied by left fronto-centrotemporal EEG interictal epileptic discharges. (D) Intraoperative picture shows the position of left hemispheric subdural grid, electrodes of interest (open circles) and resection margins (white lines). The selected numbers of electrodes relate to those of EEGs and MBFA spectrograms in Figs. 2 and 3.
a silastic sheet with slits between rows and spaced at interelectrode distances of 10 mm. We placed the grid over the left frontal, temporal, and parietal regions (Fig. 1D). The video subdural EEG was sampled at 1 kHz with an averaged reference of channels 55 and 106, which were located in the relatively silent area of the grid. We captured 8 cluster episodes of spasms over 72 h of recording. Interictal discharges on subdural EEG were high-amplitude spikeor multiple spike-waves over the left superior frontal gyrus and the left middle temporal gyrus. 2.6. Analysis of frequency and distribution of HFOs We exported video subdural EEG data sections, including ictal patterns corresponding to spasms, as
EDFCformat files (Kemp and Olivan, 2003). These EDFCfiles were converted to text format files by Insight (Persyst, Prescott, AZ, USA). We performed Multiple Band Frequency Analysis (MBFA) using Short Spectrum Eye software (Gram, Saitama, Japan) for frequency and distribution analyses of HFOs. The algorithm of MBFA was described in detail elsewhere (Shimoyama et al., 2000). The power spectrograms of frequency bands between 40 and 200 Hz were calculated with a frequency resolution of 1 Hz and a temporal resolution of 10 ms for all electrodes individually. We arranged all power spectrograms in the same order as the subdural grid electrodes and visually evaluated the distribution of HFOs. We also compared their distributions during spasms of different clinical intensity.
2822
T. Akiyama et al. / Clinical Neurophysiology 116 (2005) 2819–2825
All 8 clusters were identical. The low amplitude HFOs at 80–120 Hz consistently started 4–15 min before clinical spasms were apparent on the video. The initial appearance, location, and propagation of HFOs in the left frontotemporal region were consistent during and before clinical manifestations, i.e. in preictal periods and during clinical spasms, for the 8 clusters. Intermittent, short (lasting less than 300 ms), independent HFOs were recorded interictally from the fronto-temporal regions. Interictal HFOs never built up and synchronized. In cluster 1, 244 episodes of electrographic ictal events, consisting of 153 episodes before clinical spasms and 91
with clinical spasms, occurred. Low-amplitude HFOs at 80–120 Hz lasting between 300 and 600 ms started 15 min before clinical spasms were apparent on the video. HFOs appeared quasiperiodically with intervals of 3–15 s and were located over the middle to posterior portion of the superior frontal gyrus, the posterior portion of the inferior frontal gyrus, and anterior to the middle portion of the middle temporal gyrus (Fig. 2A and B). After the clinical spasms of facial grimacing began, the HFOs became longer in duration, were accompanied by slow waves, and extended to the posterior portion of the middle frontal gyrus and the superior temporal gyrus. The longest HFO was 4 s and
Fig. 2. Selected channels on referential montage of subdural EEG (see location on Fig. 1D) show ictal findings (band pass filter, 1.5–300 Hz; notch filter, 60 Hz; using channels 55 and 106 for an averaged reference). (A) At the beginning of a cluster of epileptic spasms without clinical manifestation, high-frequency oscillations (HFOs) at 80–120 Hz were observed lasting 600 ms predominantly over the left superior frontal gyrus (Channels 1–5), and the inferior frontal and middle temporal gyri. (B) Close up of channel 2 showed 60–100 Hz fast waves lasting 450 ms. (C) Later, when epileptic spasms presented as facial grimacing, head turning to the left, and shoulder-raising, HFOs became higher in amplitude, extended diffusely, and lasted longer (up to 4 s) than those at the beginning. Note that diffuse HFOs were sustained predominantly over the superior, inferior frontal and middle temporal gyri (Channels 111–115). No prominent HFOs were recorded outside the ictal onset zone (Channels 76–80). (D) Close up of channel 112 showed 75–145 Hz fast waves.
T. Akiyama et al. / Clinical Neurophysiology 116 (2005) 2819–2825
occurred when the clinical spasms of head turning to the left side and shoulder-raising reached maximum intensity (Fig. 2C and D). The initial location and propagation of HFOs on MBFA were consistent among the series of preictal EEG events and clinical spasms in all 8 clusters. MBFA showed an extensive but noncontiguous distribution of HFOs at 60–120 Hz over the left inferior to superior frontal gyri and the middle temporal gyrus at the beginning of a cluster of spasms during preictal periods (Fig. 3A). As spasms progressed from face to head and shoulders, the frequency of initial HFOs reached 150 Hz, HFOs of 60–100 Hz occurred and lasted up to 4 s, and the HFOs propagated to the posterior portion of the middle frontal gyrus and the superior temporal gyrus (Fig. 3B). As clinical spasms intensified, regions that had HFOs during the preictal period continued to have HFOs; however, the HFOs in these regions were of higher frequency (150 Hz) and greater power than HFOs recorded in other regions. Functional mapping did not reveal language representation in the left fronto-temporal region. We performed a partial lobectomy of the left anterior portion of the frontal lobe, cortical excisions of the superior and inferior frontal gyri, and topectomy of the middle portion of the left middle
2823
temporal gyrus that correlated to the area of preclinical HFOs and of consistent predominant HFOs during clinical spasms. Examination of pathology specimens revealed microdysgenesis. Six months after surgery the patient was seizure free but continued his presurgical antiepileptic medications.
3. Discussion We performed video subdural EEG recordings on an adolescent patient with periodic spasms. We captured HFOs in the left fronto-temporal cortices during spasms. Focal cortical resection of the area with prominent HFOs eliminated the spasms. The frequency of HFOs during a cluster of epileptic spasms ranged from 60 to 150 Hz above gamma frequency. No previous reports of intracranial HFOs recorded with high sampling frequency during epileptic spasms have appeared in the literature. Our work is the first to record and localize HFOs both before and during clinical spasms on digital video subdural EEG. The high frequency and temporal resolutions provided by MBFA enabled us to recognize the existence,
Fig. 3. The results of Multiple Band Frequency Analysis (MBFA) using a referential montage. (A) MBFA for the same time and channels of EEG on Fig. 2A. Selected MBFA power spectra (Channels 1–5) show brief high-frequency oscillations (HFOs) at 60–120 Hz with high spectral power lasting for 300–500 ms. (B) MBFA shows power spectra of the all electrodes during a clinical spasm of facial grimacing, head turning to the left and shoulder raising (see Fig. 2B). The HFOs with high spectral power became longer in duration and more extensive but noncontiguous in distribution. Initial faster HFOs, up to 150 Hz with higher and sustained spectral powers, continued to be recorded mainly by the 25 electrodes (light blue squares) covering the region where HFOs began prior to clinical spasms. We tailored cortical excision based on the data from these electrodes. The 3 blank squares represent the two channels (55 and 106) for an averaged reference, and a poor contact channel (25).
2824
T. Akiyama et al. / Clinical Neurophysiology 116 (2005) 2819–2825
distribution, and localization of short-lasting HFOs during a cluster of spasms. Video subdural EEG recorded the quasiperiodic appearance of focal HFOs before the clinical manifestation of spasms in clusters started. During this preictal period, HFOs localized to 3 discrete cortices: superior frontal, inferior frontal, and middle temporal gyri without slow waves. When HFOs became longer in duration, extended to the frontal and temporal lobes, and were accompanied by slow waves, clinical spasms became obvious on the face and subsequently progressed to the head and shoulders. This buildup of HFOs in the preictal phase may lead to synchronization of sustained ictal HFOs. We suggest that these focal cortical HFOs triggered the epileptic spasms. Recording epileptic spasms in West syndrome with ictal scalp EEG, Fusco and Vigevano reported that medium-amplitude fast activity at 14–16 Hz was associated with a motionless stare before the clinical manifestation of spasms (Fusco and Vigevano, 1993). Although diffuse high-amplitude slow waves corresponded to the clinical manifestation of spasms, fast activity was clearly not associated with clinical spasms on scalp EEG. A low-pass characteristic between the cortex and the scalp attenuates high frequency activities on scalp EEG (Pfurtscheller and Cooper, 1975). As a result, ictal scalp EEG during spasms showed focal slow waves superimposed by fast waves that had a lower frequency and were less evident than the fast waves recorded by ictal subdural EEG. Reported frequencies of HFOs on intracranial EEG recordings at the onset of partial seizures have varied and included 20–80 Hz (Alarcon et al., 1995), 40–120 Hz (Fisher et al., 1992), 60–100 Hz (Worrell et al., 2004), 70–90 Hz (Traub et al., 2001), and 80–110 Hz (Allen et al., 1992). Despite the variety of HFO frequencies, these reports demonstrated that HFOs were related to seizure onset. Alarcon et al. reported that removal of the ictal onset zone from which HFOs were localized on intracranial EEG resulted in good seizure control for localization-related epilepsy patients (Alarcon et al., 1995). Worrell et al. found that the ictal onset zone showed a significant increase in HFOs during the interictal period, the preictal period, and at seizure onset in patients with neocortical epilepsy (Worrell et al., 2004). They concluded that HFOs are signatures of an epileptic brain and provide spatial localization of the ictal onset zone. In our patient, the HFOs at 60–150 Hz on the focal cerebral cortex triggered epileptic spasms, and the distribution of HFOs defined the ictal onset zone of spasms. We tailored our patient’s cortical excision to the distribution of his preictal and ictal HFOs. Since our patient’s periodic spasms stopped after removal of the fronto-temporal cortices, we suggest that these periodic spasms represented partial seizures. Gobbi et al. proposed that periodic spasms were ‘partial epileptic seizures with a
particular type of secondary generalization’ because of asymmetrical clinical and electrographical features (Gobbi et al., 1991). The asymmetric clinical spasms, the focal interictal epileptic discharges on scalp EEG and MEG, and the focal cortical ictal HFOs support their proposal and corroborate that this subset of epileptic spasms resulted from a partial epileptic disorder in our adolescent patient. Our data suggested that the left temporal lobe should also be included in the surgery because of the presence of HFOs. However, since no MEG spike dipoles in the left temporal lobe met our criteria, it remains an open question whether a left frontal lobectomy alone may have resulted in seizure control. Various etiologies cause epileptic spasms. After reviewing the current clinical and experimental data on West syndrome, Frost and Hrachovy concluded that two or more unbalanced developmental processes in the central nervous system from multiple causative factors can result in infantile spasms, hypsarrhythmia, and psychomotor deterioration (Frost and Hrachovy, 2005). We cannot extend our results of ictal electrocorticography to the pathophysiology of infantile spasms in West syndrome; however, our electroclinical findings might reveal some common neurophysiological processes during epileptic spasms. The epileptic spasms in West syndrome seem to be a final manifestation of various processes, but spasms are believed to start on a cortical level (Vigevano et al., 2001). The ictal subdural EEG in our patient showed widespread but noncontiguous distribution of fast waves. Subtle orientation differences in the gyral topography underlying grid contacts may produce relative amplitude differences for such high-frequency rhythms. Subdural electrodes are not able to record in the same way that microelectrodes at cellular levels do (Bragin et al., 2002) even though our subdural grid array has splits between rows to ensure that each electrode contacts the brain surface equally. Subdural EEG recordings are rarely contaminated by muscle artifacts (Hamer and Morris, 2001). In addition, the consistent appearance, distribution, and propagation of preictal and ictal HFOs in cluster episodes indicated that epileptic cortices produced the HFOs. We show that HFOs occurred over the cerebral cortex before and during epileptic spasms. Initial higher frequencies and greater power of HFOs defined the ictal onset zone of the spasms. The consistent distribution of focal HFOs triggering epileptic spasms suggests that, in this patient, periodic spasms have the characteristics of partial seizures with secondary generalization.
Acknowledgements We thank Mrs Carol L. Squires for her editorial assistance.
T. Akiyama et al. / Clinical Neurophysiology 116 (2005) 2819–2825
References Alarcon G, Binnie CD, Elwes RD, Polkey CE. Power spectrum and intracranial EEG patterns at seizure onset in partial epilepsy. Electroencephalogr Clin Neurophysiol 1995;94:326–37. Allen PJ, Fish DR, Smith SJ. Very high-frequency rhythmic activity during SEEG suppression in frontal lobe epilepsy. Electroencephalogr Clin Neurophysiol 1992;82:155–9. Bragin A, Wilson CL, Staba RJ, Reddick M, Fried I, Engle Jr J. Interictal high-frequency oscillations (80–500 Hz) in the human epileptic brain: entorhinal cortex. Ann Neurol 2002;52:407–15. Chugani HT, Shewman DA, Sankar R, Chen BC, Phelps ME. Infantile spasms: II. Lenticular nuclei and brain stem activation on positron emission tomography. Ann Neurol 1992;31:212–9. Chugani HT, Shewman DA, Shields WD, Sankar R, Comair Y, Vinters HV, Peacock WJ. Surgery for intractable infantile spasms: neuroimaging perspectives. Epilepsia 1993;34:764–71. Fisher RS, Webber WR, Lesser RP, Arroyo S, Uematsu S. High-frequency EEG activity at the start of seizures. J Clin Neurophysiol 1992;9:441–8. Frost Jr JD, Hrachovy RA. Pathogenesis of infantile spasms: a model based on developmental desynchronization. J Clin Neurophysiol 2005;22:25–36. Fusco L, Vigevano F. Ictal clinical electroencephalographic findings of spasms in West syndrome. Epilepsia 1993;34:671–8. Gobbi G, Bruno L, Pini A, Giovanardi Rossi P, Tassinari CA. Periodic spasms: an unclassified type of epileptic seizure in childhood. Dev Med Child Neurol 1987;29:766–75. Gobbi G, Parmeggiani A, Pini A, Tassinari CA. Periodic spasms: a peculiar type of infantile spasm. In: Ohtahara S, Roger J, editors. New trends in
2825
pediatric epileptology. Okayama: Department of Child Neurology, Okayama University; 1991. p. 111–8. Hamer HM, Morris III HH. Indications for invasive video-electroencephalographic monitoring. In: Luders HO, Comair YG, editors. Epilepsy surgery. Philadelphia, PA: Lippincott/Williams and Wilkins; 2001. p. 559–66. Kemp B, Olivan J. European data format ‘plus’ (EDFC), an EDF alike standard format for the exchange of physiological data. Clin Neurophysiol 2003;114:1755–61. Kobayashi K, Oka M, Akiyama T, Inoue T, Abiru K, Ogino T, Yoshinaga H, Ohtsuka Y, Oka E. Very fast rhythmic activity on scalp EEG associated with epileptic spasms. Epilepsia 2004;45:488–96. Ohtsuka Y, Kobayashi K, Ogino T, Oka E. Spasms in clusters in epilepsies other than typical West syndrome. Brain Dev 2001;23:473–81. Pfurtscheller G, Cooper R. Frequency dependence of the transmission of the EEG from cortex to brain. Electroencephalogr Clin Neurophysiol 1975;38:93–6. Shimoyama I, Kasagi Y, Kaiho T, Shibata T, Nakajima Y, Asano H. Flashrelated synchronization and desynchronization revealed by a multiple band frequency analysis. Jpn J Physiol 2000;50:553–9. Traub RD, Whittington MA, Buhl EH, LeBeau FE, Bibbig A, Boyd S, Cross H, Baldeweg T. A possible role for gap junctions in generation of very fast EEG oscillations preceding the onset of, and perhaps initiating, seizures. Epilepsia 2001;42:153–70. Vigevano F, Fusco L, Pachatz C. Neurophysiology of spasms. Brain Dev 2001;23:467–72. Worrell GA, Parish L, Cranstoun SD, Jonas R, Baltuch G, Litt B. Highfrequency oscillations and seizure generation in neocortical epilepsy. Brain 2004;127:1496–506.
Focal cortical high-frequency oscillations trigger epileptic spasms: Confirmation by digital video subdural EEG Tomoyuki Akiyama, Hiroshi Otsubo*, Ayako Ochi, Taichi Ishiguro, Gemmu Kadokura, Rajesh RamachandranNair, Shelly K. Weiss, James T. Rutka, O. Carter Snead III Divisions of Neurology, Neurosurgery, The Hospital for Sick Children, 555 University Avenue, Toronto, Ont., Canada M5G 1X8 Accepted 23 August 2005 Available online 25 October 2005
Abstract Objective: To localize high-frequency oscillations (HFOs) on the cortex during epileptic spasms using video subdural EEG and Multiple Band Frequency Analysis (MBFA). Methods: Using video subdural EEG sampled at 1 kHz, we studied a 14-year-old boy with asymmetric epileptic spasms of possible left frontal origin. We identified HFOs, then analyzed and localized their distributions by MBFA. We correlated HFO distribution to clinical spasm intensity. Results: Ictal subdural EEG recorded HFOs at 60–150 Hz lasting 0.3–4 s. MBFA showed extensive but noncontiguous distribution of HFOs predominantly over the left frontal and temporal regions. HFOs began and became quasiperiodic before manifestation of clinical spasms. As clinical spasms intensified, HFOs persisted in regions where they initiated subclinically but were of higher frequency and greater power than HFOs in other regions. We performed cortical resections over the left frontal and temporal regions with predominant HFOs. Six months after surgery, the patient remained seizure free. Conclusions: HFOs were present over the ictal onset zone during epileptic spasms. Periodic spasms in this patient had the characteristics of partial seizures. Significance: We show that HFOs occurred over the cerebral cortex during epileptic spasms, and we suggest that these focal cortical HFOs triggered the spasms. q 2005 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. Keywords: High-frequency oscillation; Epileptic spasms; Subdural EEG; Multiple band frequency analysis; Partial seizure
1. Introduction Epileptic spasms in clusters commonly occur in generalized epilepsies that are generated mainly in the subcortex, such as the pediatric epileptic syndromes known as Ohtahara, West, and Lennox-Gastaut (Ohtsuka et al., 2001). In 1987, Gobbi et al. proposed the term periodic spasms for epileptic spasms that occur in older patients with localization-related epilepsy. They suggested that periodic spasms, characterized by asymmetric clinical and electrographical features, were partial seizures that originated in the cortex with a particular type of secondary generalization * Corresponding author. Tel.: C1 416 813 6660. fax: C1 416 813 6334. E-mail address: [email protected] (H. Otsubo).
(Gobbi et al., 1987, 1991). Although West syndrome with infantile spasms has been classified as generalized epilepsy, Chugani et al. found that focal cortical resection can control seizures in some patients (Chugani et al., 1993). In addition, when measuring glucose metabolism with positron emission tomography in West syndrome patients, they found abnormal rates in both the cortex and the brainstem (Chugani et al., 1992). Their findings suggest an underlying abnormal interrelationship between cortical and subcortical structures in infantile spasms. On ictal scalp EEG, epileptic spasms in clusters consist of high-voltage slow waves with superimposed fast activity in the beta frequency followed by an electrodecremental activity (Gobbi et al., 1987; Fusco and Vigevano, 1993). Fusco and Vigevano reported a medium-amplitude fast
1388-2457/$30.00 q 2005 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2005.08.029
2820
T. Akiyama et al. / Clinical Neurophysiology 116 (2005) 2819–2825
activity without clinical spasms during the preictal period and at the end of a cluster (Fusco and Vigevano, 1993). Until newer technologies were developed, the limitations of analog scalp EEG prevented detailed analysis of the fast activity of clinical spasms. In 2004, using digital scalp EEG with high sampling frequency, Kobayashi et al. discovered very fast rhythmic activities at 51–98 Hz during episodes of spasms in 11 patients with West syndrome and related disorders (Kobayashi et al., 2004). From these very fast activities in the posterior brains of thin-skulled young children, they inferred that the cortex played a major role in generating epileptic spasms. Using intracranial EEG recordings, others (Alarcon et al, 1995; Allen et al., 1992; Fisher et al., 1992; Traub et al., 2001; Worrell et al., 2004) observed very fast activities ranging from 20 to 120 Hz at the beginning of partial seizures in the ictal onset zone. In addition, Bragin et al. recorded 250–500 Hz high-frequency oscillations (HFOs), termed fast ripples, from the epileptogenic entorhinal cortex and hippocampus during the interictal state (Bragin et al., 2002). This finding suggested that the distribution of HFOs associated with epileptic seizures may localize the epileptogenic zone. We had one patient with periodic spasms, who underwent epilepsy surgery after extraoperative video subdural EEG recording. We hypothesized that the video intracranial digital EEG with high sampling rate could record focal HFOs during epileptic spasms and show that the propagation of HFOs over the cortex was associated with the initiation of spasms.
2. Case history 2.1. Patient Our patient was a 14-year-old left-handed boy who had experienced seizures since he was 8 years, 9 months of age. He had no family history of seizure disorders and no remarkable past history. Findings from his physical and neurological examinations were normal. His seizures consisted of grimacing for about 1 s followed, at times, by turning his head to the left and raising his shoulders. These brief spasms occurred quasiperiodically every 3–30 s in clusters; cluster episodes lasted 30–60 min. Between each spasm, the boy looked confused but was able to converse. Cluster episodes occurred once or twice a day, usually on waking or falling asleep. He experienced no other types of seizures. His seizures were refractory to carbamazepine, lamotrigine, clobazam, oxcarbazepine and gabapentin. At admission for surgical evaluation, he was being treated with oxcarbazepine and gabapentin. Neuropsychological tests revealed slightly impaired visual reasoning and some deficits in visual memory and verbal productivity.
2.2. EEG When 13 years old, the patient underwent prolonged (48 h) video scalp EEG with electrodes placed in the International 10–20 scalp-electrode position and a single reference electrode at Oz (HARMONY 5.4, Stellate, Montreal, PQ, Canada). Interictal scalp EEG showed high-amplitude spikeor sharp-slow wave discharges over the left frontal region (Fig. 1A). Focal interictal discharges became, at times, rhythmic trains of 2–2.5 Hz and extended to the centrotemporal regions. Background activity consisted of a posterior-dominant symmetric 8–9 Hz alpha rhythm, and sleep features were normal. We captured 2 cluster episodes, each containing spasms that lasted 1–3 s in 5–25 s intervals. Ictal video scalp EEG showed high-amplitude sharplycontoured slow waves with superimposed 15–30 Hz fast waves, predominantly over the left fronto-centro-temporal and midline regions, associated with each spasm (Fig. 1B). 2.3. MRI MR images of the patient at 13 years were normal. (GE 1.5T Signa MRI, GE Medical Systems, Milwaukee, WI, USA) 2.4. MEG At age 13, the patient also underwent MEG studies with a whole-head gradiometer Omega system (151 channels, VSM MedTech Ltd., Port Coquitlam, BC, Canada). During MEG, he also underwent simultaneous EEG that was recorded from 19 electrodes (International 10–20 system). We recorded 2 min periods of spontaneous MEG data 15 times. The sampling rate for data acquisition was 625 Hz. We visually identified MEG epileptic discharges, spikes and sharp waves (referred to as spikes), by examining the MEG recordings and cross-referencing them with the simultaneous EEG recording using a band pass filter of 3–70 Hz and a notch filter of 60 Hz. We applied a single moving dipole analysis with a single-shell, whole-head spherical model. We defined the MEG spike dipole for each spike as a single dipole fit from the earliest phase of each spike with the criteria of a residual error of less than 30%. The patient did not have spasms during MEG recording. MEG revealed a cluster of 59 MEG spike sources in the left inferior to superior frontal gyri accompanied by left frontocentro-temporal EEG interictal epileptic discharges (Fig. 1C). Temporal discharges consistently succeeded frontal discharges in the complex of MEG spikes, and we found no dipoles satisfying our criteria in the temporal region. 2.5. Extraoperative video subdural EEG We constructed a 106-channel grid, based on 3D MR images, interictal and ictal scalp EEG results, MEG spike source locations, and MEG-somatosensory evoked fields. The grid electrodes were 5 mm in diameter, embedded in
T. Akiyama et al. / Clinical Neurophysiology 116 (2005) 2819–2825
2821
Fig. 1. (A) A–P bipolar montage scalp interictal EEG shows intermittent high-amplitude spike- or sharp-slow wave complexes over the left frontal region. (band pass filter, 0.5–100 Hz; notch filter, 60 Hz). (B) A–P bipolar montage ictal scalp EEG shows high-amplitude sharply-contoured slow waves with superimposed 15–30 Hz fast activities and muscle artifacts, predominantly over the left frontal and midline regions at the time of clinical epileptic spasm (*). (band pass filter, 0.5–100 Hz; notch filter, 60 Hz). (C) Sagittal T1-weighted MRI shows a cluster of MEG dipoles with random orientations (closed triangles, locations of spike sources; tails, orientation and strength of dipole moments) over the left inferior to superior frontal gyri accompanied by left fronto-centrotemporal EEG interictal epileptic discharges. (D) Intraoperative picture shows the position of left hemispheric subdural grid, electrodes of interest (open circles) and resection margins (white lines). The selected numbers of electrodes relate to those of EEGs and MBFA spectrograms in Figs. 2 and 3.
a silastic sheet with slits between rows and spaced at interelectrode distances of 10 mm. We placed the grid over the left frontal, temporal, and parietal regions (Fig. 1D). The video subdural EEG was sampled at 1 kHz with an averaged reference of channels 55 and 106, which were located in the relatively silent area of the grid. We captured 8 cluster episodes of spasms over 72 h of recording. Interictal discharges on subdural EEG were high-amplitude spikeor multiple spike-waves over the left superior frontal gyrus and the left middle temporal gyrus. 2.6. Analysis of frequency and distribution of HFOs We exported video subdural EEG data sections, including ictal patterns corresponding to spasms, as
EDFCformat files (Kemp and Olivan, 2003). These EDFCfiles were converted to text format files by Insight (Persyst, Prescott, AZ, USA). We performed Multiple Band Frequency Analysis (MBFA) using Short Spectrum Eye software (Gram, Saitama, Japan) for frequency and distribution analyses of HFOs. The algorithm of MBFA was described in detail elsewhere (Shimoyama et al., 2000). The power spectrograms of frequency bands between 40 and 200 Hz were calculated with a frequency resolution of 1 Hz and a temporal resolution of 10 ms for all electrodes individually. We arranged all power spectrograms in the same order as the subdural grid electrodes and visually evaluated the distribution of HFOs. We also compared their distributions during spasms of different clinical intensity.
2822
T. Akiyama et al. / Clinical Neurophysiology 116 (2005) 2819–2825
All 8 clusters were identical. The low amplitude HFOs at 80–120 Hz consistently started 4–15 min before clinical spasms were apparent on the video. The initial appearance, location, and propagation of HFOs in the left frontotemporal region were consistent during and before clinical manifestations, i.e. in preictal periods and during clinical spasms, for the 8 clusters. Intermittent, short (lasting less than 300 ms), independent HFOs were recorded interictally from the fronto-temporal regions. Interictal HFOs never built up and synchronized. In cluster 1, 244 episodes of electrographic ictal events, consisting of 153 episodes before clinical spasms and 91
with clinical spasms, occurred. Low-amplitude HFOs at 80–120 Hz lasting between 300 and 600 ms started 15 min before clinical spasms were apparent on the video. HFOs appeared quasiperiodically with intervals of 3–15 s and were located over the middle to posterior portion of the superior frontal gyrus, the posterior portion of the inferior frontal gyrus, and anterior to the middle portion of the middle temporal gyrus (Fig. 2A and B). After the clinical spasms of facial grimacing began, the HFOs became longer in duration, were accompanied by slow waves, and extended to the posterior portion of the middle frontal gyrus and the superior temporal gyrus. The longest HFO was 4 s and
Fig. 2. Selected channels on referential montage of subdural EEG (see location on Fig. 1D) show ictal findings (band pass filter, 1.5–300 Hz; notch filter, 60 Hz; using channels 55 and 106 for an averaged reference). (A) At the beginning of a cluster of epileptic spasms without clinical manifestation, high-frequency oscillations (HFOs) at 80–120 Hz were observed lasting 600 ms predominantly over the left superior frontal gyrus (Channels 1–5), and the inferior frontal and middle temporal gyri. (B) Close up of channel 2 showed 60–100 Hz fast waves lasting 450 ms. (C) Later, when epileptic spasms presented as facial grimacing, head turning to the left, and shoulder-raising, HFOs became higher in amplitude, extended diffusely, and lasted longer (up to 4 s) than those at the beginning. Note that diffuse HFOs were sustained predominantly over the superior, inferior frontal and middle temporal gyri (Channels 111–115). No prominent HFOs were recorded outside the ictal onset zone (Channels 76–80). (D) Close up of channel 112 showed 75–145 Hz fast waves.
T. Akiyama et al. / Clinical Neurophysiology 116 (2005) 2819–2825
occurred when the clinical spasms of head turning to the left side and shoulder-raising reached maximum intensity (Fig. 2C and D). The initial location and propagation of HFOs on MBFA were consistent among the series of preictal EEG events and clinical spasms in all 8 clusters. MBFA showed an extensive but noncontiguous distribution of HFOs at 60–120 Hz over the left inferior to superior frontal gyri and the middle temporal gyrus at the beginning of a cluster of spasms during preictal periods (Fig. 3A). As spasms progressed from face to head and shoulders, the frequency of initial HFOs reached 150 Hz, HFOs of 60–100 Hz occurred and lasted up to 4 s, and the HFOs propagated to the posterior portion of the middle frontal gyrus and the superior temporal gyrus (Fig. 3B). As clinical spasms intensified, regions that had HFOs during the preictal period continued to have HFOs; however, the HFOs in these regions were of higher frequency (150 Hz) and greater power than HFOs recorded in other regions. Functional mapping did not reveal language representation in the left fronto-temporal region. We performed a partial lobectomy of the left anterior portion of the frontal lobe, cortical excisions of the superior and inferior frontal gyri, and topectomy of the middle portion of the left middle
2823
temporal gyrus that correlated to the area of preclinical HFOs and of consistent predominant HFOs during clinical spasms. Examination of pathology specimens revealed microdysgenesis. Six months after surgery the patient was seizure free but continued his presurgical antiepileptic medications.
3. Discussion We performed video subdural EEG recordings on an adolescent patient with periodic spasms. We captured HFOs in the left fronto-temporal cortices during spasms. Focal cortical resection of the area with prominent HFOs eliminated the spasms. The frequency of HFOs during a cluster of epileptic spasms ranged from 60 to 150 Hz above gamma frequency. No previous reports of intracranial HFOs recorded with high sampling frequency during epileptic spasms have appeared in the literature. Our work is the first to record and localize HFOs both before and during clinical spasms on digital video subdural EEG. The high frequency and temporal resolutions provided by MBFA enabled us to recognize the existence,
Fig. 3. The results of Multiple Band Frequency Analysis (MBFA) using a referential montage. (A) MBFA for the same time and channels of EEG on Fig. 2A. Selected MBFA power spectra (Channels 1–5) show brief high-frequency oscillations (HFOs) at 60–120 Hz with high spectral power lasting for 300–500 ms. (B) MBFA shows power spectra of the all electrodes during a clinical spasm of facial grimacing, head turning to the left and shoulder raising (see Fig. 2B). The HFOs with high spectral power became longer in duration and more extensive but noncontiguous in distribution. Initial faster HFOs, up to 150 Hz with higher and sustained spectral powers, continued to be recorded mainly by the 25 electrodes (light blue squares) covering the region where HFOs began prior to clinical spasms. We tailored cortical excision based on the data from these electrodes. The 3 blank squares represent the two channels (55 and 106) for an averaged reference, and a poor contact channel (25).
2824
T. Akiyama et al. / Clinical Neurophysiology 116 (2005) 2819–2825
distribution, and localization of short-lasting HFOs during a cluster of spasms. Video subdural EEG recorded the quasiperiodic appearance of focal HFOs before the clinical manifestation of spasms in clusters started. During this preictal period, HFOs localized to 3 discrete cortices: superior frontal, inferior frontal, and middle temporal gyri without slow waves. When HFOs became longer in duration, extended to the frontal and temporal lobes, and were accompanied by slow waves, clinical spasms became obvious on the face and subsequently progressed to the head and shoulders. This buildup of HFOs in the preictal phase may lead to synchronization of sustained ictal HFOs. We suggest that these focal cortical HFOs triggered the epileptic spasms. Recording epileptic spasms in West syndrome with ictal scalp EEG, Fusco and Vigevano reported that medium-amplitude fast activity at 14–16 Hz was associated with a motionless stare before the clinical manifestation of spasms (Fusco and Vigevano, 1993). Although diffuse high-amplitude slow waves corresponded to the clinical manifestation of spasms, fast activity was clearly not associated with clinical spasms on scalp EEG. A low-pass characteristic between the cortex and the scalp attenuates high frequency activities on scalp EEG (Pfurtscheller and Cooper, 1975). As a result, ictal scalp EEG during spasms showed focal slow waves superimposed by fast waves that had a lower frequency and were less evident than the fast waves recorded by ictal subdural EEG. Reported frequencies of HFOs on intracranial EEG recordings at the onset of partial seizures have varied and included 20–80 Hz (Alarcon et al., 1995), 40–120 Hz (Fisher et al., 1992), 60–100 Hz (Worrell et al., 2004), 70–90 Hz (Traub et al., 2001), and 80–110 Hz (Allen et al., 1992). Despite the variety of HFO frequencies, these reports demonstrated that HFOs were related to seizure onset. Alarcon et al. reported that removal of the ictal onset zone from which HFOs were localized on intracranial EEG resulted in good seizure control for localization-related epilepsy patients (Alarcon et al., 1995). Worrell et al. found that the ictal onset zone showed a significant increase in HFOs during the interictal period, the preictal period, and at seizure onset in patients with neocortical epilepsy (Worrell et al., 2004). They concluded that HFOs are signatures of an epileptic brain and provide spatial localization of the ictal onset zone. In our patient, the HFOs at 60–150 Hz on the focal cerebral cortex triggered epileptic spasms, and the distribution of HFOs defined the ictal onset zone of spasms. We tailored our patient’s cortical excision to the distribution of his preictal and ictal HFOs. Since our patient’s periodic spasms stopped after removal of the fronto-temporal cortices, we suggest that these periodic spasms represented partial seizures. Gobbi et al. proposed that periodic spasms were ‘partial epileptic seizures with a
particular type of secondary generalization’ because of asymmetrical clinical and electrographical features (Gobbi et al., 1991). The asymmetric clinical spasms, the focal interictal epileptic discharges on scalp EEG and MEG, and the focal cortical ictal HFOs support their proposal and corroborate that this subset of epileptic spasms resulted from a partial epileptic disorder in our adolescent patient. Our data suggested that the left temporal lobe should also be included in the surgery because of the presence of HFOs. However, since no MEG spike dipoles in the left temporal lobe met our criteria, it remains an open question whether a left frontal lobectomy alone may have resulted in seizure control. Various etiologies cause epileptic spasms. After reviewing the current clinical and experimental data on West syndrome, Frost and Hrachovy concluded that two or more unbalanced developmental processes in the central nervous system from multiple causative factors can result in infantile spasms, hypsarrhythmia, and psychomotor deterioration (Frost and Hrachovy, 2005). We cannot extend our results of ictal electrocorticography to the pathophysiology of infantile spasms in West syndrome; however, our electroclinical findings might reveal some common neurophysiological processes during epileptic spasms. The epileptic spasms in West syndrome seem to be a final manifestation of various processes, but spasms are believed to start on a cortical level (Vigevano et al., 2001). The ictal subdural EEG in our patient showed widespread but noncontiguous distribution of fast waves. Subtle orientation differences in the gyral topography underlying grid contacts may produce relative amplitude differences for such high-frequency rhythms. Subdural electrodes are not able to record in the same way that microelectrodes at cellular levels do (Bragin et al., 2002) even though our subdural grid array has splits between rows to ensure that each electrode contacts the brain surface equally. Subdural EEG recordings are rarely contaminated by muscle artifacts (Hamer and Morris, 2001). In addition, the consistent appearance, distribution, and propagation of preictal and ictal HFOs in cluster episodes indicated that epileptic cortices produced the HFOs. We show that HFOs occurred over the cerebral cortex before and during epileptic spasms. Initial higher frequencies and greater power of HFOs defined the ictal onset zone of the spasms. The consistent distribution of focal HFOs triggering epileptic spasms suggests that, in this patient, periodic spasms have the characteristics of partial seizures with secondary generalization.
Acknowledgements We thank Mrs Carol L. Squires for her editorial assistance.
T. Akiyama et al. / Clinical Neurophysiology 116 (2005) 2819–2825
References Alarcon G, Binnie CD, Elwes RD, Polkey CE. Power spectrum and intracranial EEG patterns at seizure onset in partial epilepsy. Electroencephalogr Clin Neurophysiol 1995;94:326–37. Allen PJ, Fish DR, Smith SJ. Very high-frequency rhythmic activity during SEEG suppression in frontal lobe epilepsy. Electroencephalogr Clin Neurophysiol 1992;82:155–9. Bragin A, Wilson CL, Staba RJ, Reddick M, Fried I, Engle Jr J. Interictal high-frequency oscillations (80–500 Hz) in the human epileptic brain: entorhinal cortex. Ann Neurol 2002;52:407–15. Chugani HT, Shewman DA, Sankar R, Chen BC, Phelps ME. Infantile spasms: II. Lenticular nuclei and brain stem activation on positron emission tomography. Ann Neurol 1992;31:212–9. Chugani HT, Shewman DA, Shields WD, Sankar R, Comair Y, Vinters HV, Peacock WJ. Surgery for intractable infantile spasms: neuroimaging perspectives. Epilepsia 1993;34:764–71. Fisher RS, Webber WR, Lesser RP, Arroyo S, Uematsu S. High-frequency EEG activity at the start of seizures. J Clin Neurophysiol 1992;9:441–8. Frost Jr JD, Hrachovy RA. Pathogenesis of infantile spasms: a model based on developmental desynchronization. J Clin Neurophysiol 2005;22:25–36. Fusco L, Vigevano F. Ictal clinical electroencephalographic findings of spasms in West syndrome. Epilepsia 1993;34:671–8. Gobbi G, Bruno L, Pini A, Giovanardi Rossi P, Tassinari CA. Periodic spasms: an unclassified type of epileptic seizure in childhood. Dev Med Child Neurol 1987;29:766–75. Gobbi G, Parmeggiani A, Pini A, Tassinari CA. Periodic spasms: a peculiar type of infantile spasm. In: Ohtahara S, Roger J, editors. New trends in
2825
pediatric epileptology. Okayama: Department of Child Neurology, Okayama University; 1991. p. 111–8. Hamer HM, Morris III HH. Indications for invasive video-electroencephalographic monitoring. In: Luders HO, Comair YG, editors. Epilepsy surgery. Philadelphia, PA: Lippincott/Williams and Wilkins; 2001. p. 559–66. Kemp B, Olivan J. European data format ‘plus’ (EDFC), an EDF alike standard format for the exchange of physiological data. Clin Neurophysiol 2003;114:1755–61. Kobayashi K, Oka M, Akiyama T, Inoue T, Abiru K, Ogino T, Yoshinaga H, Ohtsuka Y, Oka E. Very fast rhythmic activity on scalp EEG associated with epileptic spasms. Epilepsia 2004;45:488–96. Ohtsuka Y, Kobayashi K, Ogino T, Oka E. Spasms in clusters in epilepsies other than typical West syndrome. Brain Dev 2001;23:473–81. Pfurtscheller G, Cooper R. Frequency dependence of the transmission of the EEG from cortex to brain. Electroencephalogr Clin Neurophysiol 1975;38:93–6. Shimoyama I, Kasagi Y, Kaiho T, Shibata T, Nakajima Y, Asano H. Flashrelated synchronization and desynchronization revealed by a multiple band frequency analysis. Jpn J Physiol 2000;50:553–9. Traub RD, Whittington MA, Buhl EH, LeBeau FE, Bibbig A, Boyd S, Cross H, Baldeweg T. A possible role for gap junctions in generation of very fast EEG oscillations preceding the onset of, and perhaps initiating, seizures. Epilepsia 2001;42:153–70. Vigevano F, Fusco L, Pachatz C. Neurophysiology of spasms. Brain Dev 2001;23:467–72. Worrell GA, Parish L, Cranstoun SD, Jonas R, Baltuch G, Litt B. Highfrequency oscillations and seizure generation in neocortical epilepsy. Brain 2004;127:1496–506.