- Email: [email protected]
Magnetoencephalography as a Prognostic Tool in Patients with Medically Intractable Temporal Lobe Epilepsy
Original Article
Magnetoencephalography as a Prognostic Tool in Patients with Medically Intractable Temporal Lobe Epilepsy Chang Kyu Park1, Su Jeong Hwang2, Na Young Jung2,3, Won Seok Chang2,3, Hyun Ho Jung2,3, Jin Woo Chang2,3
-
BACKGROUND: Most surgical treatments for medically intractable temporal lobe epilepsy are helpful. When a patient has persistent symptoms after surgery, there are no tests that accurately predict whether a patient will have remnant epileptic foci. The aim of this study was to evaluate the usefulness of magnetoencephalography (MEG) as a prognostic tool in patients with temporal lobe epilepsy.
symptoms; GOs on postoperative MEG were especially correlated with epileptic recurrence. Our data show that GOs on postoperative MEG may have prognostic value.
METHODS: From July 2012 to July 2016, 21 patients underwent preoperative and postoperative MEG at our center. Postoperative MEG was performed within 2 weeks after surgery. We analyzed MEG by estimating the timefrequency component of the signal to define gamma oscillations (GOs), which are an indicator of epileptogenic foci. We analyzed the relationship between GOs on MEG and surgical outcomes.
INTRODUCTION
-
RESULTS: Mean follow-up period was 28.3 months (range, 13e44 months). At the last follow-up visit, patients were divided into 2 groups according to surgical outcome. All patients showed spike waves and GOs on preoperative electroencephalography and MEG. In the seizure control group (16 patients), spike waves (2 patients) and GOs (2 patients) were seen postoperatively despite absence of symptoms. In the recurrent seizure group (5 patients), whereas 3 patients showed spike waves, all 5 patients showed GOs on MEG postoperatively. There was a significant association between presence of GOs on postoperative MEG and surgical outcome (P [ 0.01).
-
CONCLUSIONS: MEG can provide valuable postsurgical information on epileptic foci in patients with recurrent
-
Key words Anterior temporal lobectomy - Magnetoencephalography - Prognosis - Temporal lobe epilepsy -
Abbreviations and Acronyms ECoG: Electrocorticography EEG: Electroencephalography GO: Gamma oscillation MEG: Magnetoencephalography MRI: Magnetic resonance imaging
WORLD NEUROSURGERY -: e1-e7, - 2019
E
valuation of prognosis after surgery is important for patients with medically refractory epilepsy. Various studies have used electroencephalography (EEG) for prognostic planning. Results have shown that among the electrical activity seen on EEG the fast oscillation (>30 Hz, i.e., gamma oscillation [GO]) exhibited by the epileptic brain is of prognostic interest. Recent studies have indicated that fast oscillation can be a good indicator of the epileptogenic zone and may be related to patient outcomes after epilepsy surgery.1,2 However, significant results were not obtained in EEG studies, possibly owing to the limitations of EEG.3,4 The resolution is not adequate for evaluation because rhythmic activity is affected by the conductivity of brain structures; even accurate anatomic distinction is difficult owing to deformation of postoperative anatomic location. Also, evaluation is subjective. Moreover, in some cases, EEG is unable to detect significant pathologic activity postoperatively in patients with epilepsy.5 Magnetoencephalography (MEG) is a functional neuroimaging technique that uses a very sensitive magnetometer to record the magnetic field generated by currents naturally occurring in the brain.6,7 It is a noninvasive method used to detect interictal discharges, such as spikes, to identify epileptic foci with appropriate source localization techniques.8 Single-dipole modeling, which is commonly used to analyze spikes in epilepsy, has been
From the 1Department of Neurosurgery, Kyung Hee University College of Medicine; and 2 Brain Institute, Severance Hospital, and 3Department of Neurosurgery, Yonsei University College of Medicine, Seoul, Korea To whom correspondence should be addressed: Jin Woo Chang, M.D., Ph.D. [E-mail: [email protected]] Chang Kyu Park and Su Jeong Hwang are coefirst authors. Citation: World Neurosurg. (2019). https://doi.org/10.1016/j.wneu.2018.12.024 Journal homepage: www.journals.elsevier.com/world-neurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
www.journals.elsevier.com/world-neurosurgery
e1
ORIGINAL ARTICLE CHANG KYU PARK ET AL.
PROGNOSTIC VALUE OF MEG FOR TEMPORAL LOBE EPILEPSY
Table 1. Classification of Outcomes of Epilepsy Surgery Class
Definition
I
Free of disabling seizures
II
Almost seizure-free (rare disabling seizures)
III
Worthwhile improvement
IV
No worthwhile improvement No significant seizure reduction, no appreciable change, or seizures getting worse
used to successfully position MEG during interictal discharges.9 Single-dipole modeling is an established procedure for analyzing single or spatially and temporally restricted epileptic activity. However, it has limitations for analyzing spatially propagated and temporally prolonged rhythmic magnetologic activity, including ictal data during secondary generalization.10 EEG and MEG time-frequency analysis has been used to investigate rhythmic activity, which is often an important feature of epileptogenesis.11 The short-time Fourier transform applies a window to the signal and performs a series of Fourier transforms within this window to slide across the recorded data. This technique can be used to estimate the time-frequency components of the signal and to visualize the spectral distribution.10 In the present study, we evaluated preoperative and postoperative MEG data for the presence of GOs to determine prognostic value of this technique using the short-time Fourier transform method.
Figure 1. Visualization of a power spectrum using time-frequency analysis. Red indicates that the power is relatively strong. The white square represents the position of the gamma oscillation band at 30e70 Hz. This was measured in the usual resting state.
Hz and digitized at 1000 Hz. Based on automatic search and visual inspection of the sensory evoked field data, responses contaminated by excessive MEG signals or eye movements were discarded. Postoperative MEG was measured using regions of interest near the surgical resection site. All EEG data were independently read by neurologists.
MATERIALS AND METHODS Patients This study was approved by the Institutional Review Board of Severance Hospital, Seoul, Korea (Institutional Review Board No. 4-2007-0335), and all participants provided written informed consent. Between July 2012 and July 2016, 21 patients with medically intractable temporal lobe epilepsy underwent preoperative and postoperative MEG at Severance Hospital. All patients underwent surgery. MEG was performed within 2 weeks after surgery. Outcomes were prospectively evaluated in patients who either had their seizures controlled or had persistent or recurrent symptoms after surgery at the last follow-up visit. Classification of symptoms after surgery was performed using the 4 categories of Engel (Table 1).12 The seizure control group comprised patients who were Engel class IeII, and the persistent or recurrent seizure group comprised patients who were Engel class IIIeIV. All patients were independently assessed by a neurologist for prognosis at last follow-up after surgery. EEG and MEG Recordings MEG and EEG activities were simultaneously recorded using a 306-channel, whole-head MEG system developed by the Korean Research Institute of Standards and Science (Daejeon, Korea). Spontaneous MEG and EEG signals were acquired in seated patients in a magnetically shielded room for approximately 60 minutes. Signals were recorded with a bandpass filter of 0.1e100
e2
www.SCIENCEDIRECT.com
Source Estimations of GOs To reconstruct the neuronal sources of GOs, we referred to the methodology used in the study by Schneider et al.13 This source
Table 2. Patient Demographics (N ¼ 21 Patients) Variable Age, years, mean SD Sex, male/female Follow-up, months, mean SD
Value 38.9 12.4 11/10 28.3 14.3
Surgical outcome Engel class I
11
Engel class II
5
Engel class III
3
Engel class IV
2
Diagnosis Hippocampal sclerosis
10
Cavernous malformation
7
Ganglioglioma
3
Cortical dysplasia
1
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.12.024
ORIGINAL ARTICLE CHANG KYU PARK ET AL.
PROGNOSTIC VALUE OF MEG FOR TEMPORAL LOBE EPILEPSY
Table 3. Analysis of Potential Prognostic Factors Surgical Outcome Seizure Control Group (n [ 16)
Recurrent Seizure Group (n [ 5)
P Value
40.8 14.8
43.2 5.9
0.606
10/6
1/4
0.149
Preoperative EEG
16
5
Postoperative EEG
2
3
16
5
Age, years, mean SD Sex, male/female Epileptiform discharge on EEG
0.063
GO on MEG Preoperative MEG Postoperative MEG Diagnosis (hippocampal sclerosis/others)
2
5
0.001*
6/10
4/1
0.149
EEG, electroencephalography; GO, gamma oscillation; MEG, magnetoencephalography. *Statistically significant.
reconstruction technique uses an adaptive spatial filter, which passes activity from 1 specific location of interest with unit gain and maximally suppresses the other activity that is present in the data. Briefly, the source activity was estimated by constructing a spatial filter using a lead field and the crossspectral density matrix. The data were first Fourier-transformed using a wavelet analysis, and the cross-spectral density matrix was subsequently constructed using all 126 scalp EEG channels for the 40- to 50-Hz response in each trial; this was determined separately for a baseline time point (from 500 ms to 524 ms) and for average time points of epileptiform discharge for each individual subject. For postoperative MEG, we analyzed the moment when the most suspicious EEG waveforms appeared without a definite epileptiform discharge. Spectral Analysis of GOs Spectral changes in oscillatory activity were analyzed using a wavelet transform, which provides a good compromise between time and frequency resolution.14 The period from before spike to after spike (from 500 ms to 524 ms) served as the baseline for all spectral analyses. Grand mean time-frequency representations were computed over all subjects. A repeated-measures analysis of variance with factors of congruency (congruent, incongruent) and region of interest was used to statistically analyze the total power activity in the GOs (Figure 1). Statistical Analysis Relationships between postoperative EEG or MEG and surgical outcome were assessed using the Mann-Whitney test, c2 test, and McNemar test. In addition, to analyze the relationship between other potential factors and prognosis, we assessed the following parameters: age, sex, and histologic characteristics. IBM SPSS Version 21.0 software (IBM Corp., Armonk, New York, USA) was used for all calculations, with a P value of < 0.05 considered statistically significant.
WORLD NEUROSURGERY -: e1-e7, - 2019
RESULTS Patient Demographics The mean SD age of patients was 38.9 12.4 years (range, 22e 67 years); there were 11 men and 10 women. The mean SD follow-up period was 28.3 14.3 months (range, 13e44 months). All patients underwent an anterior temporal lobectomy with or without amygdalohippocampectomy. All patients underwent preoperative scalp EEG to confirm epileptiform discharge at the site where the lesion was suspected. Magnetic resonance imaging (MRI) of the brain showed hippocampal sclerosis in 10 patients, cavernous malformation in 7 patients, tumorous lesion with calcification in 3 patients, and cortical dysplasia in 1 patient.
Surgical Outcomes Seizures were well controlled immediately after surgery in all patients. However, at the last follow-up visit, seizures remained controlled in only 16 patients (Engel class IeII), and 5 patients were experiencing recurrent symptoms (Engel class IIIeIV). Of patients, 11 were Engel class I, 5 were Engel class II, 3 were Engel class III, and 2 were Engel class IV. A pathologist diagnosed hippocampal sclerosis on histology in 10 patients; type 1A was diagnosed in 8 patients, type 2A was diagnosed in 1 patient, and type 4 was diagnosed in 1 patient. Histologic findings of cavernous angioma with hemosiderin-laden macrophages suggestive of cavernous malformation were observed in 7 patients. Ganglioglioma was diagnosed in 3 patients. Focal cortical dysplasia type IIIa was diagnosed in 1 patient (Table 2). In the postoperative seizure control group, mean patient age was 40.8 14.8 years. There were 10 men and 6 women. Diagnoses included hippocampal sclerosis in 6 patients, cavernous malformation in 7 patients, and ganglioglioma in 3 patients. All patients had epileptiform discharge on EEG and GOs on MEG preoperatively at the measured regions of interest. Only 1 patient showed epileptiform discharge on postoperative EEG, 1 patient
www.journals.elsevier.com/world-neurosurgery
e3
ORIGINAL ARTICLE CHANG KYU PARK ET AL.
PROGNOSTIC VALUE OF MEG FOR TEMPORAL LOBE EPILEPSY
Figure 2. Electroencephalography and magnetoencephalography analysis in case 1. (A) Basal view of the brain. The lesion was found in the left temporal lobe. To analyze the magnetoencephalography data, we divided the temporal lobe into 8 parts and analyzed the magnetoencephalography signal for each part. The white squares are the areas where each analysis was performed. (B) Preoperative electroencephalography showed spike waves
showed GOs on postoperative MEG, and 1 patient showed both epileptiform discharge and GOs. In the recurrent seizures group, mean patient age was 43.2 5.9 years. One patient was a man, and 4 patients were women. Hippocampal sclerosis was diagnosed in 4 patients, and cortical dysplasia was diagnosed in 1 patient. All patients were symptomfree immediately after surgery. However, 1 month after surgery, 1 patient experienced recurrence of seizures. Two patients presented with recurrent symptoms 6 months after surgery, and 2 patients presented with recurrent symptoms 12 months after surgery. All patients had epileptiform discharge on EEG and GOs on MEG preoperatively. At 2 weeks after surgery, although all patients were symptom-free at that time, 3 patients still had epileptiform discharge on EEG, and all patients showed GOs on MEG at the regions of interest. In addition, when symptoms recurred, all patients experiencing recurrence showed epileptiform discharge on EEG and GOs on MEG. Assessment of Potential Prognostic Factors In statistical analyses comparing both groups, the P values for differences in age, sex, and diagnosis were 0.606, 0.149, and 0.149. Epileptiform discharge on EEG was not a significant
e4
www.SCIENCEDIRECT.com
in the T3 and F7 regions. Postoperative electroencephalography still showed abnormal waves in the T3 and F7 regions. (C) Before surgery, strong power was measured at the gamma oscillation point of the tracing (asterisks). After surgery, the gamma band observed in the previous (asterisks) test was absent. The patient had no symptoms after surgery. EEG, electroencephalography; MEG, magnetoencephalography.
prognostic indicator for surgical outcome, but it showed a trend toward significance (P ¼ 0.063). However, the presence of GOs on MEG was significantly associated with occurrence of recurrent epilepsy compared with a control group (P < 0.001) (Table 3). Additionally, we analyzed the association between EEG and MEG findings using McNemar test. Our results showed a P value of 0.688, indicating that there was no significant difference between EEG and MEG for detection of epileptiform discharge. Illustrative Cases Case 1. A 28-year-old man experienced intermittent seizures with loss of consciousness. The seizures were refractory to various antiepileptic drugs. MRI including T1-weighted low-intensity, T2-weighted high-intensity, and fluid attenuated inversion recovery sequences showed typical hippocampal sclerosis in the left temporal lobe. Interictal EEG revealed epileptiform spikes at electrodes T3 and T5. The 306-channel MEG showed rhythmic fast activity in the left temporal region. GOs on MEG were observed in the left temporal area preoperatively. The patient underwent a temporal lobectomy with amygdalohippocampectomy. He has remained seizure-free for 24 months. Although there was
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.12.024
ORIGINAL ARTICLE CHANG KYU PARK ET AL.
Figure 3. Electroencephalography and magnetoencephalography analysis of case 2. (A) Basal view of the brain. The lesion was located in the right temporal lobe, which was divided into nine parts. Magnetoencephalography analysis was performed in each part (white squares). (B) Preoperative electroencephalography showed spike waves in the T4 and T6 regions. Postoperative electroencephalography still showed abnormal waves in the
epileptiform discharge on postoperative EEG, there were no GOs on MEG postoperatively (Figure 2). Case 2. A 40-year-old woman experienced occasional complex partial seizures with motion arrest and loss of consciousness. MRI with T1-weighted low-intensity, T2-weighted high-intensity, and high-intensity fluid attenuated inversion recovery sequences showed hippocampal sclerosis in the right temporal lobe. EEG and MEG showed intermittent but prolonged rhythmic activity in the right temporal region. Preoperative short-time Fourier transform analysis demonstrated widespread rhythmic activity ipsilateral to the lateral aspect of the temporal lobe on MEG. The patient underwent a temporal lobectomy with amygdalohippocampectomy. The rhythmic activity was still present postoperatively on EEG and MEG. She presented with recurrent symptoms 1 month after surgery (Figure 3). DISCUSSION Rhythmic activities on scalp EEG and electrocorticography (ECoG) indicate ictal onset zones.15 However, scalp EEG has limited value in showing the precise location of ictal onset owing to voltage
WORLD NEUROSURGERY -: e1-e7, - 2019
PROGNOSTIC VALUE OF MEG FOR TEMPORAL LOBE EPILEPSY
T4 and T6 region. (C) The preoperative (asterisk and section mark) points showed gamma oscillation bands. Magnetoencephalography performed after surgery confirmed that the gamma band disappeared in the marked area (asterisk), but gamma oscillations were still observed in the marked area (section mark). Seizures recurred 1 month after surgery. EEG, electroencephalography; MEG, magnetoencephalography.
declines caused by the slowing of electrical conduction through the cerebrospinal fluid and cranium.16 ECoG is the gold standard for defining the location of seizure origin, but it cannot be used for routine evaluation. In patients with focal cortical dysplasia, rhythmic epileptic discharges seen on EEG have been shown to correspond well to the continuous epileptic discharges seen on ECoG.17 Dalal et al.18 also reported good concordance between beta bands (12e30 Hz) and high gamma bands (65e90 Hz) on MEG and ECoG, respectively, using selfpaced finger-movement tasks. Their work indicated that MEG is equivalent to ECoG for evaluating epileptic rhythmic activities. Sueda et al.10 reported that the localization of rhythmic activities was consistent with the ECoG findings, and surgical resection that included the high-magnitude areas of ictal MEG rendered the patient seizure-free. Their results suggest the usefulness of MEG for determining the location of rhythmic activities. Previous MEG-based studies reported that sources of beta oscillation and GO of interictal spikes were consistent with the epileptogenic zone delineated by intracranial EEG.1,19 Moreover, the sources were localized to the epileptogenic zone, which corresponded to the surgically resected areas in patients with good
www.journals.elsevier.com/world-neurosurgery
e5
ORIGINAL ARTICLE CHANG KYU PARK ET AL.
PROGNOSTIC VALUE OF MEG FOR TEMPORAL LOBE EPILEPSY
surgical outcomes.20 Other MEG studies also showed that source localization of fast frequency components, including GOs, was concordant with the epileptic focus. These studies demonstrate that GOs detected on MEG are correlated with the epileptogenic zone and surgical outcomes.2,21 However, none of the previous MEG studies on GOs compared postoperative interictal spikes and GOs with surgical outcomes. In the present study, we evaluated the significance of GOs on postoperative MEG. We detected GOs in all patients who experienced postoperative seizures. This indicates that there was a significant relationship between the remnant epileptogenic zone and GO after surgery. High-frequency oscillatory activity, especially in the gamma band, has been frequently observed in EEG data both before and during epileptic events.22 Invasive intracranial and depth electrode studies have suggested that an increase in gamma band power before ictal onset is highly localized to the ictal onset zone.23 Although abnormal increases in gamma activity in the ictal state have been frequently reported in previous studies,1,22,24 the underlying mechanisms have not been fully elucidated. One animal study reported increases in power level and phase synchrony in gamma (30e100 Hz) frequencies between 8 seconds and 2 seconds before ictal spiking and stated that these were mediated by a dynamic state in putative interneurons.25 In human epilepsy, many researchers agree with the opinion that reduced synchrony in a preictal state could contribute to the functional isolation of pathologically discharging neurons at the epileptic focus from the brain’s functional network.26 Our results show specific roles of gamma activity in the seizure focus; however, the underlying mechanisms remain unclear. Persistent seizures occur in 20%e60% of patients after resection for intractable epilepsy.27,28 Most patients have recurrent or persistent seizures emanating from the same location of the brain owing to insufficient initial resection, activation of mesiotemporal
REFERENCES 1. Perry G, Brindley LM, Muthukumaraswamy SD, et al. Evidence for increased visual gamma responses in photosensitive epilepsy. Epilepsy Res. 2014;108:1076-1086. 2. Rampp S, Kaltenhauser M, Weigel D, et al. MEG correlates of epileptic high gamma oscillations in invasive EEG. Epilepsia. 2010;51:1638-1642. 3. Zhang N, Qiao H, Wang YJ, et al. The application of magnetoencephalography versus scalp electroencephalography in intractable temporal lobe epilepsy. Zhonghua Nei Ke Za Zhi. 2007;46:370-372. 4. Gorji A, Speckmann EJ. Epileptiform EEG spikes and their functional significance. Clin EEG Neurosci. 2009;40:230-233. 5. Guy CN, Walker S, Alarcon G, et al. MEG and EEG in epilepsy: is there a difference? Physiol Meas. 1993;14(suppl 4A):A99-A102. 6. Papanicolaou AC, Simos PG, Castillo EM, et al. Magnetocephalography: a noninvasive alternative to the Wada procedure. J Neurosurg. 2004;100: 867-876.
e6
www.SCIENCEDIRECT.com
structures in temporal lobe epilepsy, or development of a nearby independent zone of epileptogenesis.27 Brain shifts following resection may cause gross distortions of topographic anatomy in these patients. In this situation, MEG is superior to scalp EEG because the magnetic field is not distorted by skull defects and collections of cerebrospinal fluid, which may cause false localization on scalp EEG.29 Thus, MEG was used to successfully localize the residual epileptogenic zone in patients with late-onset recurrent seizures in whom MRI revealed no residual abnormality and ictal scalp EEG was not lateralizing. The present study has a few limitations. First, its retrospective nature and the fact that it was carried out at a single institution are major caveats that should be taken into account. Second, the relationship between administration of antiepileptic drugs and MEG has not yet been analyzed. Nevertheless, in terms of the prognostic value for patients with epilepsy, this MEG study analyzed preoperative and postoperative data according to surgical outcomes. In particular, postoperative MEG may be helpful to identify epileptic zones that may remain in patients with postoperative symptoms. Moreover, according to our results, disappearance or persistence of gamma oscillatory activity adjacent to the epileptogenic zone is likely an important prognostic factor for surgical outcome.
CONCLUSIONS In the present study, MEG provided valuable information for postsurgical evaluation to define epileptic foci in patients with persistent symptoms after surgery. Our data suggest that GOs on MEG may be correlated with surgical outcome. Therefore, ascertaining the presence of GOs on MEG after epilepsy surgery may be helpful for predicting recurrent seizures.
7. Takanashi Y, Nakajima K. Magnetoencephalograph. Nihon Rinsho. 1997;55(suppl 1):389-392. 8. Takusa Y, Sugai K, Sasaki M, et al. Magnetoencephalographic and electrocorticographic studies on the epileptogenic focus in a patient with Kojewnikow syndrome. No To Hattatsu. 1999; 31:48-53. 9. Ebersole JS. Magnetoencephalography/magnetic source imaging in the assessment of patients with epilepsy. Epilepsia. 1997;38(suppl 4):S1-S5. 10. Sueda K, Takeuchi F, Shiraishi H, et al. MEG time-frequency analyses for pre- and post-surgical evaluation of patients with epileptic rhythmic fast activity. Epilepsy Res. 2010;88:100-107.
13. Schneider TR, Debener S, Oostenveld R, et al. Enhanced EEG gamma-band activity reflects multisensory semantic matching in visual-toauditory object priming. Neuroimage. 2008;42: 1244-1254. 14. Tallon-Baudry C, Bertrand O. Oscillatory gamma activity in humans and its role in object representation. Trends Cogn Sci. 1999;3:151-162. 15. Baumgartner C. MEG, EEG and ECoG: discussion. Acta Neurol Scand Suppl. 1994;152:91-92. 16. Barkley GL, Baumgartner C. MEG and EEG in epilepsy. J Clin Neurophysiol. 2003;20:163-178. 17. Galicia E, Imai K, Mohamed IS, et al. Changing ictal-onset EEG patterns in children with cortical dysplasia. Brain Dev. 2009;31:569-576.
11. Bosnyakova D, Gabova A, Kuznetsova G, et al. Time-frequency analysis of spike-wave discharges using a modified wavelet transform. J Neurosci Methods. 2006;154:80-88.
18. Dalal SS, Guggisberg AG, Edwards E, et al. Spatial localization of cortical time-frequency dynamics. Conf Proc IEEE Eng Med Biol Soc. 2007;2007: 4941-4944.
12. Engel J Jr. Update on surgical treatment of the epilepsies. Summary of the Second International Palm Desert Conference on the Surgical Treatment of the Epilepsies (1992). Neurology. 1993;43: 1612-1617.
19. Heers M, Hirschmann J, Jacobs J, et al. Frequency domain beam forming of magnetoencephalographic beta band activity in epilepsy patients with focal cortical dysplasia. Epilepsy Res. 2014;108: 1195-1203.
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.12.024
ORIGINAL ARTICLE CHANG KYU PARK ET AL.
20. Assaf BA, Karkar KM, Laxer KD, et al. Magnetoencephalography source localization and surgical outcome in temporal lobe epilepsy. Clin Neurophysiol. 2004;115:2066-2076. 21. Doesburg SM, Ibrahim GM, Smith ML, et al. Altered Rolandic gamma-band activation associated with motor impairment and ictal network desynchronization in childhood epilepsy. PLoS One. 2013;8:e54943. 22. Uhlhaas PJ, Pipa G, Neuenschwander S, et al. A new look at gamma? High- (>60 Hz) gammaband activity in cortical networks: function, mechanisms and impairment. Prog Biophys Mol Biol. 2011;105:14-28. 23. Zhigalov A, Arnulfo G, Nobili L, et al. Relationship of fast- and slow-timescale neuronal dynamics in human MEG and SEEG. J Neurosci. 2015; 35:5385-5396. 24. von Ellenrieder N, Pellegrino G, Hedrich T, et al. Detection and magnetic source imaging of fast oscillations (40-160 Hz) recorded with
PROGNOSTIC VALUE OF MEG FOR TEMPORAL LOBE EPILEPSY
magnetoencephalography in focal epilepsy patients. Brain Topogr. 2016;29:218-231. 25. Srejic LR, Valiante TA, Aarts MM, et al. Highfrequency cortical activity associated with postischemic epileptiform discharges in an in vivo rat focal stroke model. J Neurosurg. 2013;118: 1098-1106. 26. Fisher RS, Scharfman HE, deCurtis M. How can we identify ictal and interictal abnormal activity? Adv Exp Med Biol. 2014;813:3-23. 27. Ramey WL, Martirosyan NL, Lieu CM, et al. Current management and surgical outcomes of medically intractable epilepsy. Clin Neurol Neurosurg. 2013;115:2411-2418. 28. West S, Nolan SJ, Newton R. Surgery for epilepsy: a systematic review of current evidence. Epileptic Disord. 2016;18:113-121. 29. Yoshinaga H, Kobayashi K, Hoshida T, et al. Magnetoencephalogram in a postoperative case
WORLD NEUROSURGERY -: e1-e7, - 2019
with a large skull defect. Pediatr Neurol. 2008;39: 48-51.
Conflict of interest statement: This work was supported by the Korea Research Council of Fundamental Science and Technology through a Basic Research Project managed by the Korea Research Institute of Standards and Science. This study was supported by a grant entitled ‘Association study between biomarkers of the brain degeneration based on functional connectivity imaging’ (No. K18172) awarded from the Korea Institute of Oriental Medicine. Received 17 September 2018; accepted 7 December 2018 Citation: World Neurosurg. (2019). https://doi.org/10.1016/j.wneu.2018.12.024 Journal homepage: www.journals.elsevier.com/worldneurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
www.journals.elsevier.com/world-neurosurgery
e7
Magnetoencephalography as a Prognostic Tool in Patients with Medically Intractable Temporal Lobe Epilepsy Chang Kyu Park1, Su Jeong Hwang2, Na Young Jung2,3, Won Seok Chang2,3, Hyun Ho Jung2,3, Jin Woo Chang2,3
-
BACKGROUND: Most surgical treatments for medically intractable temporal lobe epilepsy are helpful. When a patient has persistent symptoms after surgery, there are no tests that accurately predict whether a patient will have remnant epileptic foci. The aim of this study was to evaluate the usefulness of magnetoencephalography (MEG) as a prognostic tool in patients with temporal lobe epilepsy.
symptoms; GOs on postoperative MEG were especially correlated with epileptic recurrence. Our data show that GOs on postoperative MEG may have prognostic value.
METHODS: From July 2012 to July 2016, 21 patients underwent preoperative and postoperative MEG at our center. Postoperative MEG was performed within 2 weeks after surgery. We analyzed MEG by estimating the timefrequency component of the signal to define gamma oscillations (GOs), which are an indicator of epileptogenic foci. We analyzed the relationship between GOs on MEG and surgical outcomes.
INTRODUCTION
-
RESULTS: Mean follow-up period was 28.3 months (range, 13e44 months). At the last follow-up visit, patients were divided into 2 groups according to surgical outcome. All patients showed spike waves and GOs on preoperative electroencephalography and MEG. In the seizure control group (16 patients), spike waves (2 patients) and GOs (2 patients) were seen postoperatively despite absence of symptoms. In the recurrent seizure group (5 patients), whereas 3 patients showed spike waves, all 5 patients showed GOs on MEG postoperatively. There was a significant association between presence of GOs on postoperative MEG and surgical outcome (P [ 0.01).
-
CONCLUSIONS: MEG can provide valuable postsurgical information on epileptic foci in patients with recurrent
-
Key words Anterior temporal lobectomy - Magnetoencephalography - Prognosis - Temporal lobe epilepsy -
Abbreviations and Acronyms ECoG: Electrocorticography EEG: Electroencephalography GO: Gamma oscillation MEG: Magnetoencephalography MRI: Magnetic resonance imaging
WORLD NEUROSURGERY -: e1-e7, - 2019
E
valuation of prognosis after surgery is important for patients with medically refractory epilepsy. Various studies have used electroencephalography (EEG) for prognostic planning. Results have shown that among the electrical activity seen on EEG the fast oscillation (>30 Hz, i.e., gamma oscillation [GO]) exhibited by the epileptic brain is of prognostic interest. Recent studies have indicated that fast oscillation can be a good indicator of the epileptogenic zone and may be related to patient outcomes after epilepsy surgery.1,2 However, significant results were not obtained in EEG studies, possibly owing to the limitations of EEG.3,4 The resolution is not adequate for evaluation because rhythmic activity is affected by the conductivity of brain structures; even accurate anatomic distinction is difficult owing to deformation of postoperative anatomic location. Also, evaluation is subjective. Moreover, in some cases, EEG is unable to detect significant pathologic activity postoperatively in patients with epilepsy.5 Magnetoencephalography (MEG) is a functional neuroimaging technique that uses a very sensitive magnetometer to record the magnetic field generated by currents naturally occurring in the brain.6,7 It is a noninvasive method used to detect interictal discharges, such as spikes, to identify epileptic foci with appropriate source localization techniques.8 Single-dipole modeling, which is commonly used to analyze spikes in epilepsy, has been
From the 1Department of Neurosurgery, Kyung Hee University College of Medicine; and 2 Brain Institute, Severance Hospital, and 3Department of Neurosurgery, Yonsei University College of Medicine, Seoul, Korea To whom correspondence should be addressed: Jin Woo Chang, M.D., Ph.D. [E-mail: [email protected]] Chang Kyu Park and Su Jeong Hwang are coefirst authors. Citation: World Neurosurg. (2019). https://doi.org/10.1016/j.wneu.2018.12.024 Journal homepage: www.journals.elsevier.com/world-neurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
www.journals.elsevier.com/world-neurosurgery
e1
ORIGINAL ARTICLE CHANG KYU PARK ET AL.
PROGNOSTIC VALUE OF MEG FOR TEMPORAL LOBE EPILEPSY
Table 1. Classification of Outcomes of Epilepsy Surgery Class
Definition
I
Free of disabling seizures
II
Almost seizure-free (rare disabling seizures)
III
Worthwhile improvement
IV
No worthwhile improvement No significant seizure reduction, no appreciable change, or seizures getting worse
used to successfully position MEG during interictal discharges.9 Single-dipole modeling is an established procedure for analyzing single or spatially and temporally restricted epileptic activity. However, it has limitations for analyzing spatially propagated and temporally prolonged rhythmic magnetologic activity, including ictal data during secondary generalization.10 EEG and MEG time-frequency analysis has been used to investigate rhythmic activity, which is often an important feature of epileptogenesis.11 The short-time Fourier transform applies a window to the signal and performs a series of Fourier transforms within this window to slide across the recorded data. This technique can be used to estimate the time-frequency components of the signal and to visualize the spectral distribution.10 In the present study, we evaluated preoperative and postoperative MEG data for the presence of GOs to determine prognostic value of this technique using the short-time Fourier transform method.
Figure 1. Visualization of a power spectrum using time-frequency analysis. Red indicates that the power is relatively strong. The white square represents the position of the gamma oscillation band at 30e70 Hz. This was measured in the usual resting state.
Hz and digitized at 1000 Hz. Based on automatic search and visual inspection of the sensory evoked field data, responses contaminated by excessive MEG signals or eye movements were discarded. Postoperative MEG was measured using regions of interest near the surgical resection site. All EEG data were independently read by neurologists.
MATERIALS AND METHODS Patients This study was approved by the Institutional Review Board of Severance Hospital, Seoul, Korea (Institutional Review Board No. 4-2007-0335), and all participants provided written informed consent. Between July 2012 and July 2016, 21 patients with medically intractable temporal lobe epilepsy underwent preoperative and postoperative MEG at Severance Hospital. All patients underwent surgery. MEG was performed within 2 weeks after surgery. Outcomes were prospectively evaluated in patients who either had their seizures controlled or had persistent or recurrent symptoms after surgery at the last follow-up visit. Classification of symptoms after surgery was performed using the 4 categories of Engel (Table 1).12 The seizure control group comprised patients who were Engel class IeII, and the persistent or recurrent seizure group comprised patients who were Engel class IIIeIV. All patients were independently assessed by a neurologist for prognosis at last follow-up after surgery. EEG and MEG Recordings MEG and EEG activities were simultaneously recorded using a 306-channel, whole-head MEG system developed by the Korean Research Institute of Standards and Science (Daejeon, Korea). Spontaneous MEG and EEG signals were acquired in seated patients in a magnetically shielded room for approximately 60 minutes. Signals were recorded with a bandpass filter of 0.1e100
e2
www.SCIENCEDIRECT.com
Source Estimations of GOs To reconstruct the neuronal sources of GOs, we referred to the methodology used in the study by Schneider et al.13 This source
Table 2. Patient Demographics (N ¼ 21 Patients) Variable Age, years, mean SD Sex, male/female Follow-up, months, mean SD
Value 38.9 12.4 11/10 28.3 14.3
Surgical outcome Engel class I
11
Engel class II
5
Engel class III
3
Engel class IV
2
Diagnosis Hippocampal sclerosis
10
Cavernous malformation
7
Ganglioglioma
3
Cortical dysplasia
1
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.12.024
ORIGINAL ARTICLE CHANG KYU PARK ET AL.
PROGNOSTIC VALUE OF MEG FOR TEMPORAL LOBE EPILEPSY
Table 3. Analysis of Potential Prognostic Factors Surgical Outcome Seizure Control Group (n [ 16)
Recurrent Seizure Group (n [ 5)
P Value
40.8 14.8
43.2 5.9
0.606
10/6
1/4
0.149
Preoperative EEG
16
5
Postoperative EEG
2
3
16
5
Age, years, mean SD Sex, male/female Epileptiform discharge on EEG
0.063
GO on MEG Preoperative MEG Postoperative MEG Diagnosis (hippocampal sclerosis/others)
2
5
0.001*
6/10
4/1
0.149
EEG, electroencephalography; GO, gamma oscillation; MEG, magnetoencephalography. *Statistically significant.
reconstruction technique uses an adaptive spatial filter, which passes activity from 1 specific location of interest with unit gain and maximally suppresses the other activity that is present in the data. Briefly, the source activity was estimated by constructing a spatial filter using a lead field and the crossspectral density matrix. The data were first Fourier-transformed using a wavelet analysis, and the cross-spectral density matrix was subsequently constructed using all 126 scalp EEG channels for the 40- to 50-Hz response in each trial; this was determined separately for a baseline time point (from 500 ms to 524 ms) and for average time points of epileptiform discharge for each individual subject. For postoperative MEG, we analyzed the moment when the most suspicious EEG waveforms appeared without a definite epileptiform discharge. Spectral Analysis of GOs Spectral changes in oscillatory activity were analyzed using a wavelet transform, which provides a good compromise between time and frequency resolution.14 The period from before spike to after spike (from 500 ms to 524 ms) served as the baseline for all spectral analyses. Grand mean time-frequency representations were computed over all subjects. A repeated-measures analysis of variance with factors of congruency (congruent, incongruent) and region of interest was used to statistically analyze the total power activity in the GOs (Figure 1). Statistical Analysis Relationships between postoperative EEG or MEG and surgical outcome were assessed using the Mann-Whitney test, c2 test, and McNemar test. In addition, to analyze the relationship between other potential factors and prognosis, we assessed the following parameters: age, sex, and histologic characteristics. IBM SPSS Version 21.0 software (IBM Corp., Armonk, New York, USA) was used for all calculations, with a P value of < 0.05 considered statistically significant.
WORLD NEUROSURGERY -: e1-e7, - 2019
RESULTS Patient Demographics The mean SD age of patients was 38.9 12.4 years (range, 22e 67 years); there were 11 men and 10 women. The mean SD follow-up period was 28.3 14.3 months (range, 13e44 months). All patients underwent an anterior temporal lobectomy with or without amygdalohippocampectomy. All patients underwent preoperative scalp EEG to confirm epileptiform discharge at the site where the lesion was suspected. Magnetic resonance imaging (MRI) of the brain showed hippocampal sclerosis in 10 patients, cavernous malformation in 7 patients, tumorous lesion with calcification in 3 patients, and cortical dysplasia in 1 patient.
Surgical Outcomes Seizures were well controlled immediately after surgery in all patients. However, at the last follow-up visit, seizures remained controlled in only 16 patients (Engel class IeII), and 5 patients were experiencing recurrent symptoms (Engel class IIIeIV). Of patients, 11 were Engel class I, 5 were Engel class II, 3 were Engel class III, and 2 were Engel class IV. A pathologist diagnosed hippocampal sclerosis on histology in 10 patients; type 1A was diagnosed in 8 patients, type 2A was diagnosed in 1 patient, and type 4 was diagnosed in 1 patient. Histologic findings of cavernous angioma with hemosiderin-laden macrophages suggestive of cavernous malformation were observed in 7 patients. Ganglioglioma was diagnosed in 3 patients. Focal cortical dysplasia type IIIa was diagnosed in 1 patient (Table 2). In the postoperative seizure control group, mean patient age was 40.8 14.8 years. There were 10 men and 6 women. Diagnoses included hippocampal sclerosis in 6 patients, cavernous malformation in 7 patients, and ganglioglioma in 3 patients. All patients had epileptiform discharge on EEG and GOs on MEG preoperatively at the measured regions of interest. Only 1 patient showed epileptiform discharge on postoperative EEG, 1 patient
www.journals.elsevier.com/world-neurosurgery
e3
ORIGINAL ARTICLE CHANG KYU PARK ET AL.
PROGNOSTIC VALUE OF MEG FOR TEMPORAL LOBE EPILEPSY
Figure 2. Electroencephalography and magnetoencephalography analysis in case 1. (A) Basal view of the brain. The lesion was found in the left temporal lobe. To analyze the magnetoencephalography data, we divided the temporal lobe into 8 parts and analyzed the magnetoencephalography signal for each part. The white squares are the areas where each analysis was performed. (B) Preoperative electroencephalography showed spike waves
showed GOs on postoperative MEG, and 1 patient showed both epileptiform discharge and GOs. In the recurrent seizures group, mean patient age was 43.2 5.9 years. One patient was a man, and 4 patients were women. Hippocampal sclerosis was diagnosed in 4 patients, and cortical dysplasia was diagnosed in 1 patient. All patients were symptomfree immediately after surgery. However, 1 month after surgery, 1 patient experienced recurrence of seizures. Two patients presented with recurrent symptoms 6 months after surgery, and 2 patients presented with recurrent symptoms 12 months after surgery. All patients had epileptiform discharge on EEG and GOs on MEG preoperatively. At 2 weeks after surgery, although all patients were symptom-free at that time, 3 patients still had epileptiform discharge on EEG, and all patients showed GOs on MEG at the regions of interest. In addition, when symptoms recurred, all patients experiencing recurrence showed epileptiform discharge on EEG and GOs on MEG. Assessment of Potential Prognostic Factors In statistical analyses comparing both groups, the P values for differences in age, sex, and diagnosis were 0.606, 0.149, and 0.149. Epileptiform discharge on EEG was not a significant
e4
www.SCIENCEDIRECT.com
in the T3 and F7 regions. Postoperative electroencephalography still showed abnormal waves in the T3 and F7 regions. (C) Before surgery, strong power was measured at the gamma oscillation point of the tracing (asterisks). After surgery, the gamma band observed in the previous (asterisks) test was absent. The patient had no symptoms after surgery. EEG, electroencephalography; MEG, magnetoencephalography.
prognostic indicator for surgical outcome, but it showed a trend toward significance (P ¼ 0.063). However, the presence of GOs on MEG was significantly associated with occurrence of recurrent epilepsy compared with a control group (P < 0.001) (Table 3). Additionally, we analyzed the association between EEG and MEG findings using McNemar test. Our results showed a P value of 0.688, indicating that there was no significant difference between EEG and MEG for detection of epileptiform discharge. Illustrative Cases Case 1. A 28-year-old man experienced intermittent seizures with loss of consciousness. The seizures were refractory to various antiepileptic drugs. MRI including T1-weighted low-intensity, T2-weighted high-intensity, and fluid attenuated inversion recovery sequences showed typical hippocampal sclerosis in the left temporal lobe. Interictal EEG revealed epileptiform spikes at electrodes T3 and T5. The 306-channel MEG showed rhythmic fast activity in the left temporal region. GOs on MEG were observed in the left temporal area preoperatively. The patient underwent a temporal lobectomy with amygdalohippocampectomy. He has remained seizure-free for 24 months. Although there was
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.12.024
ORIGINAL ARTICLE CHANG KYU PARK ET AL.
Figure 3. Electroencephalography and magnetoencephalography analysis of case 2. (A) Basal view of the brain. The lesion was located in the right temporal lobe, which was divided into nine parts. Magnetoencephalography analysis was performed in each part (white squares). (B) Preoperative electroencephalography showed spike waves in the T4 and T6 regions. Postoperative electroencephalography still showed abnormal waves in the
epileptiform discharge on postoperative EEG, there were no GOs on MEG postoperatively (Figure 2). Case 2. A 40-year-old woman experienced occasional complex partial seizures with motion arrest and loss of consciousness. MRI with T1-weighted low-intensity, T2-weighted high-intensity, and high-intensity fluid attenuated inversion recovery sequences showed hippocampal sclerosis in the right temporal lobe. EEG and MEG showed intermittent but prolonged rhythmic activity in the right temporal region. Preoperative short-time Fourier transform analysis demonstrated widespread rhythmic activity ipsilateral to the lateral aspect of the temporal lobe on MEG. The patient underwent a temporal lobectomy with amygdalohippocampectomy. The rhythmic activity was still present postoperatively on EEG and MEG. She presented with recurrent symptoms 1 month after surgery (Figure 3). DISCUSSION Rhythmic activities on scalp EEG and electrocorticography (ECoG) indicate ictal onset zones.15 However, scalp EEG has limited value in showing the precise location of ictal onset owing to voltage
WORLD NEUROSURGERY -: e1-e7, - 2019
PROGNOSTIC VALUE OF MEG FOR TEMPORAL LOBE EPILEPSY
T4 and T6 region. (C) The preoperative (asterisk and section mark) points showed gamma oscillation bands. Magnetoencephalography performed after surgery confirmed that the gamma band disappeared in the marked area (asterisk), but gamma oscillations were still observed in the marked area (section mark). Seizures recurred 1 month after surgery. EEG, electroencephalography; MEG, magnetoencephalography.
declines caused by the slowing of electrical conduction through the cerebrospinal fluid and cranium.16 ECoG is the gold standard for defining the location of seizure origin, but it cannot be used for routine evaluation. In patients with focal cortical dysplasia, rhythmic epileptic discharges seen on EEG have been shown to correspond well to the continuous epileptic discharges seen on ECoG.17 Dalal et al.18 also reported good concordance between beta bands (12e30 Hz) and high gamma bands (65e90 Hz) on MEG and ECoG, respectively, using selfpaced finger-movement tasks. Their work indicated that MEG is equivalent to ECoG for evaluating epileptic rhythmic activities. Sueda et al.10 reported that the localization of rhythmic activities was consistent with the ECoG findings, and surgical resection that included the high-magnitude areas of ictal MEG rendered the patient seizure-free. Their results suggest the usefulness of MEG for determining the location of rhythmic activities. Previous MEG-based studies reported that sources of beta oscillation and GO of interictal spikes were consistent with the epileptogenic zone delineated by intracranial EEG.1,19 Moreover, the sources were localized to the epileptogenic zone, which corresponded to the surgically resected areas in patients with good
www.journals.elsevier.com/world-neurosurgery
e5
ORIGINAL ARTICLE CHANG KYU PARK ET AL.
PROGNOSTIC VALUE OF MEG FOR TEMPORAL LOBE EPILEPSY
surgical outcomes.20 Other MEG studies also showed that source localization of fast frequency components, including GOs, was concordant with the epileptic focus. These studies demonstrate that GOs detected on MEG are correlated with the epileptogenic zone and surgical outcomes.2,21 However, none of the previous MEG studies on GOs compared postoperative interictal spikes and GOs with surgical outcomes. In the present study, we evaluated the significance of GOs on postoperative MEG. We detected GOs in all patients who experienced postoperative seizures. This indicates that there was a significant relationship between the remnant epileptogenic zone and GO after surgery. High-frequency oscillatory activity, especially in the gamma band, has been frequently observed in EEG data both before and during epileptic events.22 Invasive intracranial and depth electrode studies have suggested that an increase in gamma band power before ictal onset is highly localized to the ictal onset zone.23 Although abnormal increases in gamma activity in the ictal state have been frequently reported in previous studies,1,22,24 the underlying mechanisms have not been fully elucidated. One animal study reported increases in power level and phase synchrony in gamma (30e100 Hz) frequencies between 8 seconds and 2 seconds before ictal spiking and stated that these were mediated by a dynamic state in putative interneurons.25 In human epilepsy, many researchers agree with the opinion that reduced synchrony in a preictal state could contribute to the functional isolation of pathologically discharging neurons at the epileptic focus from the brain’s functional network.26 Our results show specific roles of gamma activity in the seizure focus; however, the underlying mechanisms remain unclear. Persistent seizures occur in 20%e60% of patients after resection for intractable epilepsy.27,28 Most patients have recurrent or persistent seizures emanating from the same location of the brain owing to insufficient initial resection, activation of mesiotemporal
REFERENCES 1. Perry G, Brindley LM, Muthukumaraswamy SD, et al. Evidence for increased visual gamma responses in photosensitive epilepsy. Epilepsy Res. 2014;108:1076-1086. 2. Rampp S, Kaltenhauser M, Weigel D, et al. MEG correlates of epileptic high gamma oscillations in invasive EEG. Epilepsia. 2010;51:1638-1642. 3. Zhang N, Qiao H, Wang YJ, et al. The application of magnetoencephalography versus scalp electroencephalography in intractable temporal lobe epilepsy. Zhonghua Nei Ke Za Zhi. 2007;46:370-372. 4. Gorji A, Speckmann EJ. Epileptiform EEG spikes and their functional significance. Clin EEG Neurosci. 2009;40:230-233. 5. Guy CN, Walker S, Alarcon G, et al. MEG and EEG in epilepsy: is there a difference? Physiol Meas. 1993;14(suppl 4A):A99-A102. 6. Papanicolaou AC, Simos PG, Castillo EM, et al. Magnetocephalography: a noninvasive alternative to the Wada procedure. J Neurosurg. 2004;100: 867-876.
e6
www.SCIENCEDIRECT.com
structures in temporal lobe epilepsy, or development of a nearby independent zone of epileptogenesis.27 Brain shifts following resection may cause gross distortions of topographic anatomy in these patients. In this situation, MEG is superior to scalp EEG because the magnetic field is not distorted by skull defects and collections of cerebrospinal fluid, which may cause false localization on scalp EEG.29 Thus, MEG was used to successfully localize the residual epileptogenic zone in patients with late-onset recurrent seizures in whom MRI revealed no residual abnormality and ictal scalp EEG was not lateralizing. The present study has a few limitations. First, its retrospective nature and the fact that it was carried out at a single institution are major caveats that should be taken into account. Second, the relationship between administration of antiepileptic drugs and MEG has not yet been analyzed. Nevertheless, in terms of the prognostic value for patients with epilepsy, this MEG study analyzed preoperative and postoperative data according to surgical outcomes. In particular, postoperative MEG may be helpful to identify epileptic zones that may remain in patients with postoperative symptoms. Moreover, according to our results, disappearance or persistence of gamma oscillatory activity adjacent to the epileptogenic zone is likely an important prognostic factor for surgical outcome.
CONCLUSIONS In the present study, MEG provided valuable information for postsurgical evaluation to define epileptic foci in patients with persistent symptoms after surgery. Our data suggest that GOs on MEG may be correlated with surgical outcome. Therefore, ascertaining the presence of GOs on MEG after epilepsy surgery may be helpful for predicting recurrent seizures.
7. Takanashi Y, Nakajima K. Magnetoencephalograph. Nihon Rinsho. 1997;55(suppl 1):389-392. 8. Takusa Y, Sugai K, Sasaki M, et al. Magnetoencephalographic and electrocorticographic studies on the epileptogenic focus in a patient with Kojewnikow syndrome. No To Hattatsu. 1999; 31:48-53. 9. Ebersole JS. Magnetoencephalography/magnetic source imaging in the assessment of patients with epilepsy. Epilepsia. 1997;38(suppl 4):S1-S5. 10. Sueda K, Takeuchi F, Shiraishi H, et al. MEG time-frequency analyses for pre- and post-surgical evaluation of patients with epileptic rhythmic fast activity. Epilepsy Res. 2010;88:100-107.
13. Schneider TR, Debener S, Oostenveld R, et al. Enhanced EEG gamma-band activity reflects multisensory semantic matching in visual-toauditory object priming. Neuroimage. 2008;42: 1244-1254. 14. Tallon-Baudry C, Bertrand O. Oscillatory gamma activity in humans and its role in object representation. Trends Cogn Sci. 1999;3:151-162. 15. Baumgartner C. MEG, EEG and ECoG: discussion. Acta Neurol Scand Suppl. 1994;152:91-92. 16. Barkley GL, Baumgartner C. MEG and EEG in epilepsy. J Clin Neurophysiol. 2003;20:163-178. 17. Galicia E, Imai K, Mohamed IS, et al. Changing ictal-onset EEG patterns in children with cortical dysplasia. Brain Dev. 2009;31:569-576.
11. Bosnyakova D, Gabova A, Kuznetsova G, et al. Time-frequency analysis of spike-wave discharges using a modified wavelet transform. J Neurosci Methods. 2006;154:80-88.
18. Dalal SS, Guggisberg AG, Edwards E, et al. Spatial localization of cortical time-frequency dynamics. Conf Proc IEEE Eng Med Biol Soc. 2007;2007: 4941-4944.
12. Engel J Jr. Update on surgical treatment of the epilepsies. Summary of the Second International Palm Desert Conference on the Surgical Treatment of the Epilepsies (1992). Neurology. 1993;43: 1612-1617.
19. Heers M, Hirschmann J, Jacobs J, et al. Frequency domain beam forming of magnetoencephalographic beta band activity in epilepsy patients with focal cortical dysplasia. Epilepsy Res. 2014;108: 1195-1203.
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.12.024
ORIGINAL ARTICLE CHANG KYU PARK ET AL.
20. Assaf BA, Karkar KM, Laxer KD, et al. Magnetoencephalography source localization and surgical outcome in temporal lobe epilepsy. Clin Neurophysiol. 2004;115:2066-2076. 21. Doesburg SM, Ibrahim GM, Smith ML, et al. Altered Rolandic gamma-band activation associated with motor impairment and ictal network desynchronization in childhood epilepsy. PLoS One. 2013;8:e54943. 22. Uhlhaas PJ, Pipa G, Neuenschwander S, et al. A new look at gamma? High- (>60 Hz) gammaband activity in cortical networks: function, mechanisms and impairment. Prog Biophys Mol Biol. 2011;105:14-28. 23. Zhigalov A, Arnulfo G, Nobili L, et al. Relationship of fast- and slow-timescale neuronal dynamics in human MEG and SEEG. J Neurosci. 2015; 35:5385-5396. 24. von Ellenrieder N, Pellegrino G, Hedrich T, et al. Detection and magnetic source imaging of fast oscillations (40-160 Hz) recorded with
PROGNOSTIC VALUE OF MEG FOR TEMPORAL LOBE EPILEPSY
magnetoencephalography in focal epilepsy patients. Brain Topogr. 2016;29:218-231. 25. Srejic LR, Valiante TA, Aarts MM, et al. Highfrequency cortical activity associated with postischemic epileptiform discharges in an in vivo rat focal stroke model. J Neurosurg. 2013;118: 1098-1106. 26. Fisher RS, Scharfman HE, deCurtis M. How can we identify ictal and interictal abnormal activity? Adv Exp Med Biol. 2014;813:3-23. 27. Ramey WL, Martirosyan NL, Lieu CM, et al. Current management and surgical outcomes of medically intractable epilepsy. Clin Neurol Neurosurg. 2013;115:2411-2418. 28. West S, Nolan SJ, Newton R. Surgery for epilepsy: a systematic review of current evidence. Epileptic Disord. 2016;18:113-121. 29. Yoshinaga H, Kobayashi K, Hoshida T, et al. Magnetoencephalogram in a postoperative case
WORLD NEUROSURGERY -: e1-e7, - 2019
with a large skull defect. Pediatr Neurol. 2008;39: 48-51.
Conflict of interest statement: This work was supported by the Korea Research Council of Fundamental Science and Technology through a Basic Research Project managed by the Korea Research Institute of Standards and Science. This study was supported by a grant entitled ‘Association study between biomarkers of the brain degeneration based on functional connectivity imaging’ (No. K18172) awarded from the Korea Institute of Oriental Medicine. Received 17 September 2018; accepted 7 December 2018 Citation: World Neurosurg. (2019). https://doi.org/10.1016/j.wneu.2018.12.024 Journal homepage: www.journals.elsevier.com/worldneurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
www.journals.elsevier.com/world-neurosurgery
e7