- Email: [email protected]
Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy
Clinical Neurophysiology xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Clinical Neurophysiology journal homepage: www.elsevier.com/locate/clinph
Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy Lu Tang a, Jing Xiang c, Shuyang Huang a, Ailiang Miao a, Huaiting Ge a, Hongxing Liu a, Di Wu a, Qingshan Guan a, Ting Wu b, Qiqi Chen b, Lu Yang b, Xiaopeng Lu d, Zheng Hu d, Xiaoshan Wang a,⇑ a
Department of Neurology, Nanjing Medical University, Affiliated Nanjing Brain Hospital, Nanjing, Jiangsu 210029, China MEG Center, Nanjing Brain Hospital, Nanjing, Jiangsu 210029, China MEG Center, Division of Neurology, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Avenue, Cincinnati, OH 45220, USA d Department of Neurology, Nanjing Children’s Hospital, Nanjing, Jiangsu 210029, China b c
a r t i c l e
i n f o
Article history: Accepted 26 August 2015 Available online xxxx Keywords: Magnetoencephalography Childhood absence epilepsy High frequency oscillations
h i g h l i g h t s Neuromagnetic changes from interictal to ictal periods predominantly occurred in medial prefrontal
cortex and parieto-occipito-temporal junction in absence seizures. Significant differences between interictal and ictal activities were found in low-frequency bands
(<30 Hz). The strength of ictal high-frequency oscillations (200–1000 Hz) significantly correlated with the
severity of absence seizures.
a b s t r a c t Objective: This study quantified the clinical correlation of interictal and ictal neuromagnetic activities from low- to very-high-frequency ranges in childhood absence epilepsy (CAE). Methods: Twelve patients with clinically diagnosed drug-naïve CAE were studied using a 275-channel whole-head magnetoencephalography (MEG) system. MEG data were digitized at 6000 Hz and analyzed at both sensor and source levels with multi-frequency analyses. Results: Neuromagnetic changes from interictal to ictal periods predominantly occurred in medial prefrontal cortex and parieto-occipito-temporal junction in absence seizures. The changes were statistically significant in low-frequency bands only (<30 Hz, p < 0.0001). There was a significant correlation between the source strength of ictal high-frequency oscillations (HFOs) in 200–1000 Hz and the number of daily seizures (r = 0.734, p < 0.01). Conclusions: CAE has focal neuromagnetic sources. The transition from interictal to ictal periods is associated with the elevation of low-frequency brain activities. The strength of HFOs reflects the severity of absence seizures. Significance: Low- and high-frequency MEG signals reveal distinct brain activities in CAE. HFOs is a new biomarker for the study of absence seizures. Ó 2015 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.
1. Introduction High-frequency oscillations (HFOs, 80–500 Hz) and very-HFOs (VHFOs, >1000 Hz) have been reported in epileptic patients with ⇑ Corresponding author at: Department of Neurology, Nanjing Brain Hospital, Nanjing Medical University, 264 Guangzhou Road, Nanjing, Jiangsu 210029, China. Tel.: +86 25 82296208; fax: +86 25 83719457. E-mail address: [email protected] (X. Wang).
clinical intracranial electroencephalography (EEG) in previous studies (Usui et al., 2010; Kerber et al., 2014). HFOs is considered new biomarker for epilepsy surgery because they are more sensitive and specific to seizure onset zone (SOZ) than epileptic spikes (Usui et al., 2010; Xiang et al., 2010; Jacobs et al., 2012; Kerber et al., 2014). In patients with lesional epilepsy, removal of HFOgenerating regions is largely a good indicator of surgical outcomes (Jacobs et al., 2010; Wu et al., 2010). Some studies have also demonstrated that HFO rates can be inhibited by antiepileptic
http://dx.doi.org/10.1016/j.clinph.2015.08.016 1388-2457/Ó 2015 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
2
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
The inclusion criteria of CAE patients were as follows: (1) clinically diagnosed CAE according to the International League Against Epilepsy Proposal for Revised Classification of Epilepsies and Epileptic Syndromes (1989), (2) routine clinical EEG recordings with bilateral synchronous symmetrical 3 Hz SWDs on a normal background with at least one burst lasting P3 s, (3) no treatment with AEDs, (4) no other type of seizure or other neurological disorder. This study was approved by the Institutional Review Board at NBH, Nanjing Children’s Hospital, and Nanjing Medical University. Patients and their guardians were informed in detail of the research purposes and procedures. Written informed consent was obtained from the parent of each subject, and informed assent was obtained from each child.
drugs (AEDs) and that HFOs may be markers of epileptogenicity and seizure severity (Jacobs et al., 2012; Zijlmans et al., 2012). In contrast to intracranial studies, magnetoencephalography (MEG) studies allow non-invasive HFO analysis in a wider range of patients (Xiang et al., 2013, 2014a). MEG also presents higher temporal resolution than other neuromagnetic modalities, such as functional magnetic resonance imaging (fMRI). A newly developed MEG methodology provides spatial, frequency, and volumetric descriptions of abnormalities in brain activity in multi-frequency ranges at the source level; these descriptions are generally more objective than previous analyses of brain wave forms using one or more dipoles (Xiang et al., 2013, 2014a). Childhood absence epilepsy (CAE) is characterized by multiple brief impairments of consciousness (Tenney and Glauser, 2013); these episodes provide a unique opportunity for interictal and ictal brain activity capture by MEG. According to previous reports, the cerebral cortex performs a critical and perhaps fundamental function in the pathophysiology of absence seizures (Stefan and Rampp, 2009; Gupta et al., 2011). Increasing evidence indicates that CAE is associated with focal seizures (Holmes et al., 2004; Tenney and Glauser, 2013). Studies focusing on early pre-ictal sources have demonstrated that spike and wave discharges (SWDs) do not suddenly increase but gradually build up over a dynamic network and are preceded by low-frequency occipital and frontal sources (Gupta et al., 2011). Children with CAE commonly present with psychological problems, even if their seizures are controlled (Cerminara et al., 2013; Glauser et al., 2013). Considering that HFOs occurring in different patterns are likely to have different neurophysiological mechanisms and clinical relevance (Jacobs et al., 2012; Kobayashi et al., 2013), this study quantitatively characterized the neuromagnetic signatures and clinical correlations of interictal and ictal activities from low- to very-high-frequency ranges in patients with CAE.
2.2. MEG recording MEG data were recorded in a magnetically shielded room using a whole-head CTF MEG system with 275 channels (VSM Medical Technology Company, Canada) at MEG Center at NBH. To identify system and environmental noises, we routinely recorded one background MEG dataset without patients just before the experiment. Three small coils were placed to the left and right pre-auricular points and nasion to measure head positions relative to the MEG sensors during recording. The system allowed for head localization at 1 mm accuracy; head movement during each recording was limited to 5 mm. All MEG data were recorded with noise cancellation of third-order gradients. The sampling rate was 6000 Hz. The recording was done in supine position. An audio-visual system was used to monitor the patient during recording. For each subject, at least five epochs were recorded, and the duration of one MEG epoch was 2 min. Seizures that occurred during recording were marked and stored as MEG data. 2.3. MRI scan
2. Materials and methods 3D MR images were acquired on a 3.0-T Siemens MAGNETOM Verio MR imaging scanner (Siemens, Erlangen, Germany). Three fiduciary marks were placed in locations identical to the positions of the three coils used during MEG recording with the aid of digital photographs to allow accurate co-registration of the MEG and MRI datasets. All anatomical landmarks digitized in the MEG study were made identifiable in the MR images.
2.1. Patients CAE patients were recruited at Nanjing Brain Hospital (NBH) and the Neurology Department of the Nanjing Children’s Hospital from March 2011 to September 2013. Subjects were prescreened by pediatric neurologists specializing in epilepsy. While 25 CAE patients were recruited, only 12 met the inclusion criteria and were included in this research; other patients were excluded because of excessive head/body motion, non-standard treatment history, and lack of SWDs. The clinical details of the patients are shown in Table 1.
2.4. MEG data analysis Similar to previous reports (Xiang et al., 2010, 2014a), MEG waveforms were visually inspected to identify magnetic noise
Table 1 Clinical characteristics of CAE patients. Patients
Sex/age (F/M)/ (year)
Seizure onset age (year)
Head circumferences (cm)
Epilepsy duration (months)
Seizure frequency (times/day)
Semiology
History/family history
Seizure captures (times)
Ictal duration (s)
1 2 3 4 5 6 7 8 9 10 11 12
M/12 F/12 F/6 F/5 M/8 F/6 M/8 F/5.5 F/10 F/9 F/10 M/6.5
11.5 10.5 6 5 8 6 8 5 9 9 9 6
54.5 53.8 49.2 49.0 50.8 49.3 51.0 49.3 50.5 50.2 51.0 50.8
6 20 1 5 4 4.5 5 6 12 4 10 7.5
10–15 10 10–15 10 10 5–6 6–7 1–2 3–4 15 4–5 4–5
Absence Absence Absence Absence Absence Absence Absence Absence Absence Absence Absence Absence
–/– –/– –/– –/– –/– –/– –/– –/– –/– –/– –/– –/–
3 4 3 6 6 2 2 2 3 4 4 2
18.1 15.4 12.7 11 13.6 28.2 10.7 25 23 5.2 33.3 10.7
Average
–/8.17
7.75
50.78
7.08
8
–
–
–
17.24
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
3
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
and artifacts. MEG data without noise or artifacts (>6 pT) were then filtered with a band pass filter of 1–4 Hz to distinguish interictal and ictal MEG segments because ictal MEG data, which are recorded during absence seizures, are characterized by 3 Hz SWDs in MEG waveforms (Fig. 1). We verified ictal periods with audio– visual system recordings by finding that patients began to stare at the same period of time (impaired consciousness). Thus, to analyze both low- and high-frequency MEG signals, MEG data preprocessing was conducted in two steps. First, MEG data were digitally filtered to 1–4 Hz. Ictal 3 Hz SWDs were clearly identifiable in the waveforms after filtering. Second, seizure start and end times were recorded and applied to all frequency bands for further ictal segment analysis. Of note, selection of the ictal segment was based on both clinical observation of absence and 3-Hz spike-and-slow-wave complexes in MEG. We selected interictal segments that did not contain any visual rhythmic spike wave activity. In other words, if there were long periods of rhythmic spike wave activity without obvious clinical manifestations, we would not select it as interictal segments or ictal segments. In addition, interictal segments were at least 1 min away from ictal segments. 2.4.1. Source localization Magnetic sources of MEG signals in frequency ranges of delta (1–4 Hz), theta (4–8 Hz), alpha (8–12 Hz), beta (12–30 Hz), low gamma (30–45 Hz), high gamma (55–90 Hz), ripple (90–200 Hz), HFOs (200–1000 Hz), and VHFOs (1000–2000 Hz) were localized. We analyzed low gamma (30–45 Hz) and high gamma (55– 90 Hz) MEG data to avoid power-line noise, which was approximately 50 Hz. Building on previous studies (Xiang et al., 2014a), the present study used accumulated source imaging to continuously accumulate total source strengths and source properties without assuming source numbers or time windows. Accumulated source imaging
integrated spatial filter and accumulation into a systematic approach to summarize and visualize brain activity over a long period of time as a volumetric image. The aforementioned method for reconstruction of brain activity was implemented in the MEG Processor with C/C++ in the Windows platform (Xiang et al., 2014a). More than one time slice was used to reveal the fluctuation of source activity in space and time. To increase the reliability of source measurements, we computed ‘‘corrected source strength” according to the following equation:
C s ¼ ST
ð1Þ
where Cs represents the corrected source strength; S represents the strength of source activity and T represents the t-value associated with the correlation coefficient of the measured magnetic field data and the estimated forward solution data MEG. We used accumulated source imaging to continuously accumulate the total source strength and source properties without making any assumption about the number of sources and the time windows. The whole brain was scanned at 6 mm resolution (approximately 17,160 brain areas for each frequency band for each dataset). In the present study, the predominant neuromagnetic activities were typically localized to up to 3 brain areas. In order to segment brain areas, individual MRI was integrated into MEG source imaging with three fiducial points (the left and right pre-auricular points, and the nasion). The MEG Processor automatically marked source peaks. The minimal distance between any two-source peaks was 10 mm. In other words, if the distance between two source peak voxels was 610 mm, they were considered as one source. In this case, the center of the source would be at the voxel with higher spectral power. If the source strength of the two source voxels was equal, the center of the source would be at the middle of the two voxels. Therefore, significant and consistent spatial activities were visually identified in the MRI images. MEG Processor allowed us to visualize and segment brain areas in both 2D and 3D viewers (Xiang et al., 2014a,b). The locations of the sources, such as the occipital cortex, was identified with individual MRI. Accumulated source imaging could localize correlated sources by using node-beam lead fields (Xiang et al., 2014a). Since each node-beam lead field represented a form of ‘‘source-beamfor mer” or ‘‘sub-space solution”, accumulated source imaging had multiple source beamformer to separate correlated sources. Detailed mathematical algorithms and validations related to this approach have been described in a recent paper (Xiang et al., 2014a). 2.4.2. Time–frequency transform To characterize the frequency and spectral signatures of MEG data, complex Morlet continuous wavelet transform was used to transform waveform data into a time–frequency spectrogram. The time–frequency analysis was conducted with Morlet continuous wavelet transform (CWT) in this study, which can be described by the following equation:
GðtÞ ¼ C r p1=4 eð1=2Þt ðeirt jr Þ 2
Fig. 1. MEG waveforms recorded from a child with absence seizures showing selection of interictal MEG data. Two minutes of interictal MEG data was selected after excluding 3 Hz SWDs and visual epileptiform activity or spikes.
ð2Þ
In this equation, t indicates time and f indicates frequency. jr represents admissibility and C r represents normalized constant. r represents the standard deviation of the Gaussian curve in the time domain. The mathematical algorithms, procedures, and parameters employed in this study have been described in detail in previous reports (Xiang et al., 2004, 2013; Kotecha et al., 2009). Six hundred frequency bins and 600 time bands were applied to all MEG spectrograms. We computed the sum of all sensor spectrograms (or global spectrogram) to quantify the spectral power of the entire brain (Xiang et al., 2010). The spectrograms of all MEG sensors are displayed as contour maps to visualize the spatial distribution of brain activities.
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
4
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
Fig. 2. Accumulated source images showing spatial activities in five low-frequency bands (1–4, 4–8, 8–12, 12–30, and 30–45 Hz) recorded from CAE subjects. 3D images are displayed in axial, coronal and oblique sagittal positions. During absence seizures (right column, ‘‘Ictal”), the brain activity source demonstrated large variations and propagated quickly towards adjacent regions in low-frequency bands. The blue and green arrows indicate the predominant neuromagnetic activities of interictal and ictal activities, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
5
Fig. 3. Accumulated source images showing spatial activities in four high-frequency bands (55–90, 90–200, 200–1000, and 1000–2000 Hz) recorded from CAE subjects. 3D images are displayed in axial, coronal and oblique sagittal positions. Interictal and ictal activities are spatially concordant in high-frequency ranges but not in low-frequency ranges. Green arrows indicate predominant neuromagnetic activities during absence seizures (‘‘Ictal”). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
2.5. Statistics Pearson correlation was used to analyze correlation between the number of daily seizure episodes and the strength of each source for each frequency band. Student’s t-test was used to compare source strength in interictal and ictal activities. Fisher’s exact test was performed on predominant neuromagnetic source locations between interictal and ictal activities. The threshold of statistical significance for differences was set at p < 0.05 for each test. For multiple comparisons, Bonferroni correction for multiple comparison was applied (e.g., for nine frequency bands, p < 0.005; up to four brain areas for each frequency band, p < 0.001). To solve the problem of type I errors, we use a statistical procedure for controlling the false discovery rate (FDR) (Genovese et al., 2002; Jacobs et al., 2015). Statistical analyses
were performed using SPSS version 16.0 for Windows (SPSS Inc., Chicago, IL, USA). 3. Results The range of ictal durations was 5.2–33.3 s, and the average ictal duration was 17.24 s. Magnetic source imaging (Figs. 2 and 3) revealed significant differences between interictal and ictal activities in low-frequency bands (<30 Hz, p < 0.0001), not in highfrequency bands. In addition, spectrograms (Fig. 4) also showed that neuromagnetic changes from interictal to ictal periods predominantly occurred in low-frequency bands (<30 Hz, p < 0.0001). Our data showed that each subject typically had 2–3 sources, which were very strong compared with the rest of the brain activity. As a group of patients, interictal and ictal activities in absence
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
6
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
Fig. 4. Global spectrograms (a summation of all MEG sensors) and contour maps showing distributions of epileptic activity in low- and high-frequency bands recorded from subjects with CAE. The global spectrograms (upper row) show central frequencies of interictal and ictal activities. The contour maps (bottom row) show the positions of corresponding spectrograms over the entire brain area. The ictal criterion maps (middle column) show ictal contour maps with spectral power scales identical to those described for the interictal maps. The sample coordinates (last row) indicate the orientations of the contour maps.
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
7
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
seizures predominantly occurred in four brain areas in each frequency band. Although individual variations among the subjects were noted, significant and consistent differences between interictal and ictal periods could still be visually identified.
Table 2 Predominant neuromagnetic activities of MOC, PC, MPFC and POT in interictal and ictal MEG segments.
3.1. Source localization Analysis of conventional low-frequency brain activity in 1–4 Hz (delta) and 4–8 Hz (theta) bands revealed that the medial prefrontal cortex (MPFC) and parieto-occipito-temporal (POT) junction showed high odds of activity during seizures, whereas the precuneus (PC) showed consistent activity in both interictal and ictal activities. Interictal neuromagnetic activity in 8–12 Hz (alpha) was localized to the middle occipital cortex (MOC). In the 12–30 Hz (beta) band, POT showed high odds of activity during seizures. In the 30–45 Hz (low gamma) band, MPFC demonstrated predominant activity during seizures. No significant difference in predominant brain area was found in the 55–90 Hz band. Neuromagnetic MEG signals above 90 Hz were analyzed in three frequency bands, such as 90–200 Hz (ripple), 200–1000 Hz (HFOs), and 1000–2000 Hz (VHFOs). Accumulated source imaging strongly suggested that, in all three bands, subjects with CAE present spatial patterns of source imaging in both interictal and ictal activities in POT and MPFC. No significant difference in predominant brain activity was found between interictal and ictal activities in these three bands. Detailed statistical results for these bandwidths are shown in Table 2. Representative source images from subjects with CAE are shown in Figs. 2 and 3.
Frequency band (Hz)
Source locations
Interictal (times)
Ictal (times)
p value
FDR threshold
1–4
MOC PC MPFC POT
0 11 2 3
0 8 12 12
1.000 0.158 0.0001⁄⁄ 0.0001⁄⁄
0.044 0.013 0.003 0.003
4–8
MOC PC MPFC POT
2 9 1 5
2 6 11 11
0.705 0.200 0.0001⁄⁄ 0.014⁄
0.037 0.018 0.003 0.009
8–12
MOC PC MPFC POT
5 7 6 1
1 6 9 7
0.077 0.500 0.200 0.014⁄
0.012 0.028 0.018 0.009
12–30
MOC PC MPFC POT
2 9 6 2
0 4 9 11
0.239 0.050 0.200 0.0001⁄⁄
0.022 0.010 0.018 0.003
30–45
MOC PC MPFC POT
0 4 4 1
0 2 12 3
1.000 0.320 0.001⁄ 0.295
0.044 0.025 0.006 0.023
55–90
MOC PC MPFC POT
0 0 5 1
0 1 8 1
1.000 0.500 0.207 1.000
0.044 0.028 0.020 0.044
90–200
MOC PC MPFC POT
0 3 1 3
0 3 2 3
1.000 0.680 0.500 0.680
0.044 0.034 0.028 0.034
200–1000
MOC PC MPFC POT
0 0 3 2
0 0 3 3
1.000 1.000 0.680 0.500
0.044 0.044 0.034 0.028
1000–2000
MOC PC MPFC POT
0 0 2 3
0 0 5 3
1.000 1.000 0.185 0.680
0.044 0.044 0.015 0.034
3.2. Clinical correlate We found that the strength of ictal source peak in the 4–8 Hz (theta), 90–200 Hz (ripple), and 200–1000 Hz (HFOs) bands positively correlated with clinical measurements. The strength of ictal source peak in the 200–1000 Hz band (HFOs) showed the strongest correlation with the number of daily seizure episodes (p < 0.01, r = 0.734). Though the strength of ictal source peak in the 4–8 Hz band (p < 0.05, r = 0.587) and 90–200 Hz (p < 0.05, r = 0.640) also correlated with the number of daily seizure episodes, the correlations were relatively weak. Fig. 5 shows the results of the correlations between MEG measurements and clinical seizures. Detailed statistical analysis of neuromagnetic strength of source peak between interictal and ictal segments are shown in Table 3.
MOC: the middle occipital cortex; MPFC: the medial prefrontal cortex; PC: the precuneus; POT: parieto-ocipito-temporal junction. p < 0.05 after Bonferroni multiple comparisons; **p < 0.001 after Bonferroni multiple comparisons. Italic, the result is still significant after correction for multiple comparisons using the FDR (corrected for 4 9 tests) with FDR = pFDR * i/N. Here, pFDR = 0.05 (usual significance value in most statistical tests), i refers to the ranked index of the p-values that are computed and N to the number of tests.
4. Discussion
4.1. Low-frequency neuromagnetic activities
Previous reports on absence epilepsy using MEG/EEG typically focused on 3 Hz ictal SWDs (Sakurai et al., 2010). This study improves our understanding of absence epilepsy by identifying abnormalities in both interictal and ictal activities over a wider frequency range of 1–2000 Hz. Furthermore, accumulated source imaging can visualize epileptic activity of the entire brain volumetrically (Xiang et al., 2014a). Since children’s head circumference may be smaller than those of adults may, it is necessary to take into account the issue of MEG signal attenuation by head-channel distance. Specifically, when the recording position is supine, signal attenuation in the frontal area can be significant, which result in low signal-to-noise ratio in MEG data. We used accumulated source imaging in the present study since this method can increase the signal-to-noise ratio by accumulating MEG signals and spatially filtering noise (Xiang et al., 2014a). However, in our opinion, the best solution to solve this problem is to use a pediatric MEG system. A pediatric MEG system that utilizes a small helmet optimized for children may minimize this problem (Hiraishi et al., 2015).
Source analysis of conventional low-frequency brain activity in the 1–4 Hz band (delta) and the 4–8 Hz (theta) band revealed that MPFC and POT perform different functions in interictal and ictal activities. Whereas PC showed consistent activity in both interictal and ictal activities. The cerebral mechanism of increased lowfrequency neuromagnetic activities (1–4 Hz and 4–8 Hz) in absence epilepsy remains controversial. A study using MEG and diffusion tensor imaging has revealed that abnormal MEG delta waves may be linked to disconnection between the gray and white matters (Huang et al., 2009). Another EEG study suggests that the low-frequency brain activity abnormality maybe a result of thalamic dysfunction. It seems that the reciprocally connected neurons in the thalamus and cortex are responsible for the SWDs (Hughes et al., 2011). In our study, low gamma oscillation (30–45 Hz) showed predominant neuromagnetic activities of MPFC during absence seizures. Previous studies consider that the anterior activity in MPFC is very important in the initialization and propagation of SWDs (Stefan and Rampp, 2009; Sakurai et al., 2010; Gupta et al.,
*
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
8
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
Fig. 5. Charts of Pearson’s correlation showing relationships between source strength of interictal/ictal activities and daily seizure episodes. Ictal source strength was found to be positively correlated with daily seizure episodes in the 4–8 Hz (p < 0.05, r = 0.587), 90–200 Hz (p < 0.05, r = 0.640) and 200–1000 Hz (p < 0.01, r = 0.734) bands. Correlation between the two variables is plotted in the sample coordinate (last row) with source strength as abscissa and daily seizure episodes as ordinate. The unit of source strength has been statistically transformed.
2011). An increase of source strength in 30–45 Hz may reflect frequent abnormal discharges due to chronic disorganization of the functional architecture due to neurophysiologic dysfunction. Other studies have demonstrated that high gamma oscillatory (55– 90 Hz) activity is related to signal transmission in cortical microcir-
cuits and that increased high gamma amplitude leads to multi-unit neural firing in the ictal core from microelectrode and subdural recordings (Sohal, 2012; Weiss et al., 2013). Our results support the notion that subjects with absence seizures demonstrate significant elevation of low-frequency
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx Table 3 Analysis of neuromagnetic peak source strength between interictal and ictal segments. Frequency band (Hz)
Interictal
Ictal
t value
p value
FDR threshold
1–4 4–8 8–12 12–30 30–45 55–90 90–200 200–1000 1000–2000
6.82 ± 1.89 4.31 ± 1.66 3.16 ± 1.35 2.47 ± 0.53 1.45 ± 1.67 1.88 ± 3.14 1.15 ± 1.61 1.56 ± 1.88 0.27 ± 0.21
61.01 ± 17.80 19.34 ± 8.98 9.10 ± 3.20 8.84 ± 4.17 3.37 ± 2.72 2.93 ± 4.05 1.73 ± 1.55 2.15 ± 1.88 0.26 ± 0.21
10.274 5.673 7.300 5.602 2.838 3.140 2.113 2.242 0.437
0.0001⁄⁄ 0.0001⁄⁄ 0.0001⁄⁄ 0.0001⁄⁄ 0.016⁄ 0.009⁄ 0.058 0.047⁄ 0.671
0.013 0.013 0.013 0.013 0.033 0.027 0.044 0.038 0.050
*
p < 0.05 after Bonferroni multiple comparisons; **p < 0.001 after Bonferroni multiple comparisons. Italic, the result is still significant after correction for multiple comparisons using the FDR (corrected for 9 tests).
activities. In addition, the source of low-frequency brain activity presented large variations and propagated quickly towards adjacent regions. This indicates that visual waveform inspection does not always result in delineation of a potential epileptogenic zone (EZ). In the case of refractory epilepsy, this may lead to incomplete resection of ictal onset zone and result in residual seizures (Kim et al., 2010; Weiss et al., 2013; Lee et al., 2014). 4.2. Source strength of HFOs and clinical correlation Increasing evidence suggests that HFOs and VHFOs exist in epilepsy patients and can be used to localize epileptogenic origins before surgery (Jacobs et al., 2008; Usui et al., 2010; Kerber et al., 2014). Although our understanding of how specific information is carried by HFOs is poor, many experiments have verified that spikes and HFOs present different pathophysiologic mechanisms and HFOs are more clinically related to seizures than lowfrequency brain activities. HFOs can occur with or without epileptic spikes (Jacobs et al., 2012; Kobayashi et al., 2013). Studies on pre-ictal and ictal HFOs have suggested that only channels with HFOs during seizure onset show HFOs during seizure evolution. Compared with spikes, these HFOs seem more specific for SOZ and could be helpful in determining the EZ in the clinical practice of epilepsy surgery (Zijlmans et al., 2012; Weiss et al., 2013). In other studies, interictal HFOs obtained during sleep, but not those based on the record of an ictal event, were also found to provide reliable information regarding SOZ localization (Andrade-Valenca et al., 2012). Larger numbers of interictal HFOs indicate greater risk of seizure; thus interictal HFOs are a sign of seizure susceptibility (Zijlmans et al., 2011; Jacobs et al., 2012). Using an animal model, Bragin et al. found a strong correlation between decreased times of HFO detection and increased rates of spontaneous seizures (Bragin et al., 2004). Zijlmans et al. found that HFOs increase after seizures and reduction of medication in intractable focal epilepsy patients (Zijlmans et al., 2009). Therefore, we postulate that HFOs is a useful clinical marker for epilepsy. Previous reports on HFOs in epilepsy relied mainly on intracranial EEG recordings (Usui et al., 2010; Weiss et al., 2013). Considering that intracranial EEG recordings are extremely invasive and ethically prohibited in the study of patients not requiring brain surgery, the noninvasive MEG approach utilized in the present study with excellent results may be appropriate for future studies of HFOs and VHFOs. Based on the MEG data recorded from CAE in the 90–200 Hz (ripple), 200–1000 Hz (HFOs), and 1000–2000 Hz (VHFOs) bands, accumulated source imaging strongly suggested that interictal and ictal activities are spatially concordant in high-frequency ranges but not in low-frequency ranges. Strength of ictal peak source was notably higher than strength of interictal peak source in the 200–1000 Hz (HFOs) band. Hence, HFOs may present
9
noticeable high-frequency source strength changes during seizure but fewer positional variations compared with low-frequency MEG signals in CAE subjects. Low-frequency activities are propagated epileptic activities, as those activities have no close relationship with clinical severity (number of daily seizure episodes). In other words, high-frequency epileptic activities may have closer relationship with EZ than low-frequency epileptic activities, and may provide precise spatial information about cortical dysfunction (Xiang et al., 2013). This observation is consistent with previous reports on medically intractable epilepsy (Xiang et al., 2009, 2010). MPFC and POT are the primary nodes of the default mode network. The default mode network has been hypothesized on the basis of the observation that specific regions of the brain are consistently activated during the resting state but deactivated when engaged with tasks (Danielson et al., 2011). This network is assumed to be selectively impaired during epileptic seizures associated with loss of consciousness, such as absence seizures. The initial portion of ictal activation of absence seizures is predominantly localized to the MPFC and POT regions. 5. Conclusion In this study involving CAE patients, distinct neuromagnetic signatures were found to be associated with interictal and ictal activities in both low- and high-frequency ranges. The magnetic sources observed in this work support the notion that focal cortical areas are responsible for absence seizures. Though low-frequency neuromagnetic activities showed greater changes from interictal to ictal periods, the strength of neuromagnetic HFOs (200–1000 Hz) significantly correlated with clinical severity of absence seizures. Therefore, low- and high-frequency neuromagnetic signals may reveal distinct brain activities in CAE. Neuromagnetic HFOs (200–1000 Hz) is a new imaging biomarker for the study of absence seizures. Acknowledgments We would like to thank the physicians and fellows at NBH and Nanjing Children’s Hospital. We thank the MEG Center at Cincinnati Children’s Hospital for helping with MEG data analysis. We also thank all participants and their families for their time and support. The work was supported by the National Natural Science Foundation of China (Grant No. 81471324, http://npd.nsfc.gov.cn/), the Key Project of Medical Science and Technology Development Foundation (Grant No. ZKX11002, http://www.njh.gov.cn/html/list_83. shtml), the Fourth Phase of Jiangsu ‘‘Project 333” Scientific Research Funding Schemes, 2013, the Health Department of Jiangsu Province (Grant No. H201443) and the Nanjing Medical University General Program (Grant No. 2014NJMU050). Conflict of interest: We declare that all authors have no financial and personal relationships with other people or organizations that could have inappropriately influenced this work. None of the authors has potential conflicts of interest to be disclosed. References Andrade-Valenca L, Mari F, Jacobs J, Zijlmans M, Olivier A, Gotman J, et al. Interictal high frequency oscillations (HFOs) in patients with focal epilepsy and normal MRI. Clin Neurophysiol 2012;123:100–5. Bragin A, Wilson CL, Almajano J, Mody I, Engel Jr J. High-frequency oscillations after status epilepticus: epileptogenesis and seizure genesis. Epilepsia 2004;45: 1017–23. Cerminara C, D’Agati E, Casarelli L, Kaunzinger I, Lange KW, Pitzianti M, et al. Attention impairment in childhood absence epilepsy: an impulsivity problem? Epilepsy Behav 2013;27:337–41. Danielson NB, Guo JN, Blumenfeld H. The default mode network and altered consciousness in epilepsy. Behav Neurol 2011;24:55–65.
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
10
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
Genovese CR, Lazar NA, Nichols T. Thresholding of statistical maps in functional neuroimaging using the false discovery rate. NeuroImage 2002;15:870–8. Glauser TA, Cnaan A, Shinnar S, Hirtz DG, Dlugos D, Masur D, et al. Ethosuximide, valproic acid, and lamotrigine in childhood absence epilepsy: initial monotherapy outcomes at 12 months. Epilepsia 2013;54:141–55. Gupta D, Ossenblok P, van Luijtelaar G. Space-time network connectivity and cortical activations preceding spike wave discharges in human absence epilepsy: a MEG study. Med Biol Eng Comput 2011;49:555–65. Hiraishi H, Kikuchi M, Yoshimura Y, Kitagawa S, Hasegawa C, Munesue T, et al. Unusual developmental pattern of brain lateralization in young boys with autism spectrum disorder: power analysis with child-sized magnetoencephalography. Psychiatry Clin Neurosci 2015;69:153–60. Holmes MD, Brown M, Tucker DM. Are ‘‘generalized” seizures truly generalized? Evidence of localized mesial frontal and frontopolar discharges in absence. Epilepsia 2004;45:1568–79. Huang MX, Theilmann RJ, Robb A, Angeles A, Nichols S, Drake A, et al. Integrated imaging approach with MEG and DTI to detect mild traumatic brain injury in military and civilian patients. J Neurotrauma 2009;26:1213–26. Hughes SW, Lorincz ML, Parri HR, Crunelli V. Infraslow (<0.1 Hz) oscillations in thalamic relay nuclei basic mechanisms and significance to health and disease states. Prog Brain Res 2011;193:145–62. Jacobs J, LeVan P, Chander R, Hall J, Dubeau F, Gotman J. Interictal high-frequency oscillations (80–500 Hz) are an indicator of seizure onset areas independent of spikes in the human epileptic brain. Epilepsia 2008;49:1893–907. Jacobs J, Zijlmans M, Zelmann R, Chatillon CE, Hall J, Olivier A, et al. High-frequency electroencephalographic oscillations correlate with outcome of epilepsy surgery. Ann Neurol 2010;67:209–20. Jacobs J, Staba R, Asano E, Otsubo H, Wu JY, Zijlmans M, et al. High-frequency oscillations (HFOs) in clinical epilepsy. Prog Neurobiol 2012;98:302–15. Jacobs J, Vogt C, LeVan P, Zelmann R, Gotman J, Kobayashi K. The identification of distinct high-frequency oscillations during spikes delineates the seizure onset zone better than high-frequency spectral power changes. Clin Neurophysiol 2015. http://dx.doi.org/10.1016/j.clinph.2015.04.053. Kerber K, Dumpelmann M, Schelter B, Le Van P, Korinthenberg R, Schulze-Bonhage A, et al. Differentiation of specific ripple patterns helps to identify epileptogenic areas for surgical procedures. Clin Neurophysiol 2014;125:1339–45. Kim DW, Kim HK, Lee SK, Chu K, Chung CK. Extent of neocortical resection and surgical outcome of epilepsy: intracranial EEG analysis. Epilepsia 2010;51:1010–7. Kobayashi K, Miya K, Akiyama T, Endoh F, Oka M, Yoshinaga H, et al. Cortical contribution to scalp EEG gamma rhythms associated with epileptic spasms. Brain Dev 2013;35:762–70. Kotecha R, Xiang J, Wang Y, Huo X, Hemasilpin N, Fujiwara H, et al. Time, frequency and volumetric differences of high-frequency neuromagnetic oscillation between left and right somatosensory cortices. Int J Psychophysiol 2009;72: 102–10. Lee C, Kim JS, Jeong W, Chung CK. Usefulness of interictal spike source localization in temporal lobe epilepsy: electrocorticographic study. Epilepsy Res 2014;108: 448–58.
Proposal for revised classification of epilepsies and epileptic syndromes. Commission on Classification and Terminology of the International League against Epilepsy. Epilepsia 1989;30:389–99. Sakurai K, Takeda Y, Tanaka N, Kurita T, Shiraishi H, Takeuchi F, et al. Generalized spike-wave discharges involve a default mode network in patients with juvenile absence epilepsy: a MEG study. Epilepsy Res 2010;89:176–84. Sohal VS. Insights into cortical oscillations arising from optogenetic studies. Biol Psychiatry 2012;71:1039–45. Stefan H, Rampp S. Current clinica–neurophysiological findings in absence epilepsies. Nervenarzt 2009;80:378–85. Tenney JR, Glauser TA. The current state of absence epilepsy: can we have your attention? Epilepsy Curr 2013;13:135–40. Usui N, Terada K, Baba K, Matsuda K, Nakamura F, Usui K, et al. Very high frequency oscillations (over 1000 Hz) in human epilepsy. Clin Neurophysiol 2010;121: 1825–31. Weiss SA, Banks GP, McKhann Jr GM, Goodman RR, Emerson RG, Trevelyan AJ, et al. Ictal high frequency oscillations distinguish two types of seizure territories in humans. Brain 2013;136:3796–808. Wu JY, Sankar R, Lerner JT, Matsumoto JH, Vinters HV, Mathern GW. Removing interictal fast ripples on electrocorticography linked with seizure freedom in children. Neurology 2010;75:1686–94. Xiang J, Holowka S, Qiao H, Sun B, Xiao Z, Jiang Y, et al. Automatic localization of epileptic zones using magnetoencephalography. Neurol Clin Neurophysiol 2004;2004:98. Xiang J, Liu Y, Wang Y, Kirtman EG, Kotecha R, Chen Y, et al. Frequency and spatial characteristics of high-frequency neuromagnetic signals in childhood epilepsy. Epileptic Disord 2009;11:113–25. Xiang J, Wang Y, Chen Y, Liu Y, Kotecha R, Huo X, et al. Noninvasive localization of epileptogenic zones with ictal high-frequency neuromagnetic signals. J Neurosurg Pediatr 2010;5:113–22. Xiang J, deGrauw X, Korostenskaja M, Korman AM, O’Brien HL, Kabbouche MA, et al. Altered cortical activation in adolescents with acute migraine: a magnetoencephalography study. J Pain 2013;14:1553–63. Xiang J, Luo Q, Kotecha R, Korman A, Zhang F, Luo H, et al. Accumulated source imaging of brain activity with both low and high-frequency neuromagnetic signals. Front Neuroinform 2014a;8:57. Xiang J, Tenney JR, Korman AM, Leiken K, Rose DF, Harris E, et al. Quantification of interictal neuromagnetic activity in absence epilepsy with accumulated source imaging. Brain Topogr 2014. http://dx.doi.org/10.1007/s10548-014-0411-5. Zijlmans M, Jacobs J, Zelmann R, Dubeau F, Gotman J. High-frequency oscillations mirror disease activity in patients with epilepsy. Neurology 2009;72: 979–86. Zijlmans M, Jacobs J, Kahn YU, Zelmann R, Dubeau F, Gotman J. Ictal and interictal high frequency oscillations in patients with focal epilepsy. Clin Neurophysiol 2011;122:664–71. Zijlmans M, Jiruska P, Zelmann R, Leijten FS, Jefferys JG, Gotman J. High-frequency oscillations as a new biomarker in epilepsy. Ann Neurol 2012;71:169–78.
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
Contents lists available at ScienceDirect
Clinical Neurophysiology journal homepage: www.elsevier.com/locate/clinph
Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy Lu Tang a, Jing Xiang c, Shuyang Huang a, Ailiang Miao a, Huaiting Ge a, Hongxing Liu a, Di Wu a, Qingshan Guan a, Ting Wu b, Qiqi Chen b, Lu Yang b, Xiaopeng Lu d, Zheng Hu d, Xiaoshan Wang a,⇑ a
Department of Neurology, Nanjing Medical University, Affiliated Nanjing Brain Hospital, Nanjing, Jiangsu 210029, China MEG Center, Nanjing Brain Hospital, Nanjing, Jiangsu 210029, China MEG Center, Division of Neurology, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Avenue, Cincinnati, OH 45220, USA d Department of Neurology, Nanjing Children’s Hospital, Nanjing, Jiangsu 210029, China b c
a r t i c l e
i n f o
Article history: Accepted 26 August 2015 Available online xxxx Keywords: Magnetoencephalography Childhood absence epilepsy High frequency oscillations
h i g h l i g h t s Neuromagnetic changes from interictal to ictal periods predominantly occurred in medial prefrontal
cortex and parieto-occipito-temporal junction in absence seizures. Significant differences between interictal and ictal activities were found in low-frequency bands
(<30 Hz). The strength of ictal high-frequency oscillations (200–1000 Hz) significantly correlated with the
severity of absence seizures.
a b s t r a c t Objective: This study quantified the clinical correlation of interictal and ictal neuromagnetic activities from low- to very-high-frequency ranges in childhood absence epilepsy (CAE). Methods: Twelve patients with clinically diagnosed drug-naïve CAE were studied using a 275-channel whole-head magnetoencephalography (MEG) system. MEG data were digitized at 6000 Hz and analyzed at both sensor and source levels with multi-frequency analyses. Results: Neuromagnetic changes from interictal to ictal periods predominantly occurred in medial prefrontal cortex and parieto-occipito-temporal junction in absence seizures. The changes were statistically significant in low-frequency bands only (<30 Hz, p < 0.0001). There was a significant correlation between the source strength of ictal high-frequency oscillations (HFOs) in 200–1000 Hz and the number of daily seizures (r = 0.734, p < 0.01). Conclusions: CAE has focal neuromagnetic sources. The transition from interictal to ictal periods is associated with the elevation of low-frequency brain activities. The strength of HFOs reflects the severity of absence seizures. Significance: Low- and high-frequency MEG signals reveal distinct brain activities in CAE. HFOs is a new biomarker for the study of absence seizures. Ó 2015 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.
1. Introduction High-frequency oscillations (HFOs, 80–500 Hz) and very-HFOs (VHFOs, >1000 Hz) have been reported in epileptic patients with ⇑ Corresponding author at: Department of Neurology, Nanjing Brain Hospital, Nanjing Medical University, 264 Guangzhou Road, Nanjing, Jiangsu 210029, China. Tel.: +86 25 82296208; fax: +86 25 83719457. E-mail address: [email protected] (X. Wang).
clinical intracranial electroencephalography (EEG) in previous studies (Usui et al., 2010; Kerber et al., 2014). HFOs is considered new biomarker for epilepsy surgery because they are more sensitive and specific to seizure onset zone (SOZ) than epileptic spikes (Usui et al., 2010; Xiang et al., 2010; Jacobs et al., 2012; Kerber et al., 2014). In patients with lesional epilepsy, removal of HFOgenerating regions is largely a good indicator of surgical outcomes (Jacobs et al., 2010; Wu et al., 2010). Some studies have also demonstrated that HFO rates can be inhibited by antiepileptic
http://dx.doi.org/10.1016/j.clinph.2015.08.016 1388-2457/Ó 2015 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
2
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
The inclusion criteria of CAE patients were as follows: (1) clinically diagnosed CAE according to the International League Against Epilepsy Proposal for Revised Classification of Epilepsies and Epileptic Syndromes (1989), (2) routine clinical EEG recordings with bilateral synchronous symmetrical 3 Hz SWDs on a normal background with at least one burst lasting P3 s, (3) no treatment with AEDs, (4) no other type of seizure or other neurological disorder. This study was approved by the Institutional Review Board at NBH, Nanjing Children’s Hospital, and Nanjing Medical University. Patients and their guardians were informed in detail of the research purposes and procedures. Written informed consent was obtained from the parent of each subject, and informed assent was obtained from each child.
drugs (AEDs) and that HFOs may be markers of epileptogenicity and seizure severity (Jacobs et al., 2012; Zijlmans et al., 2012). In contrast to intracranial studies, magnetoencephalography (MEG) studies allow non-invasive HFO analysis in a wider range of patients (Xiang et al., 2013, 2014a). MEG also presents higher temporal resolution than other neuromagnetic modalities, such as functional magnetic resonance imaging (fMRI). A newly developed MEG methodology provides spatial, frequency, and volumetric descriptions of abnormalities in brain activity in multi-frequency ranges at the source level; these descriptions are generally more objective than previous analyses of brain wave forms using one or more dipoles (Xiang et al., 2013, 2014a). Childhood absence epilepsy (CAE) is characterized by multiple brief impairments of consciousness (Tenney and Glauser, 2013); these episodes provide a unique opportunity for interictal and ictal brain activity capture by MEG. According to previous reports, the cerebral cortex performs a critical and perhaps fundamental function in the pathophysiology of absence seizures (Stefan and Rampp, 2009; Gupta et al., 2011). Increasing evidence indicates that CAE is associated with focal seizures (Holmes et al., 2004; Tenney and Glauser, 2013). Studies focusing on early pre-ictal sources have demonstrated that spike and wave discharges (SWDs) do not suddenly increase but gradually build up over a dynamic network and are preceded by low-frequency occipital and frontal sources (Gupta et al., 2011). Children with CAE commonly present with psychological problems, even if their seizures are controlled (Cerminara et al., 2013; Glauser et al., 2013). Considering that HFOs occurring in different patterns are likely to have different neurophysiological mechanisms and clinical relevance (Jacobs et al., 2012; Kobayashi et al., 2013), this study quantitatively characterized the neuromagnetic signatures and clinical correlations of interictal and ictal activities from low- to very-high-frequency ranges in patients with CAE.
2.2. MEG recording MEG data were recorded in a magnetically shielded room using a whole-head CTF MEG system with 275 channels (VSM Medical Technology Company, Canada) at MEG Center at NBH. To identify system and environmental noises, we routinely recorded one background MEG dataset without patients just before the experiment. Three small coils were placed to the left and right pre-auricular points and nasion to measure head positions relative to the MEG sensors during recording. The system allowed for head localization at 1 mm accuracy; head movement during each recording was limited to 5 mm. All MEG data were recorded with noise cancellation of third-order gradients. The sampling rate was 6000 Hz. The recording was done in supine position. An audio-visual system was used to monitor the patient during recording. For each subject, at least five epochs were recorded, and the duration of one MEG epoch was 2 min. Seizures that occurred during recording were marked and stored as MEG data. 2.3. MRI scan
2. Materials and methods 3D MR images were acquired on a 3.0-T Siemens MAGNETOM Verio MR imaging scanner (Siemens, Erlangen, Germany). Three fiduciary marks were placed in locations identical to the positions of the three coils used during MEG recording with the aid of digital photographs to allow accurate co-registration of the MEG and MRI datasets. All anatomical landmarks digitized in the MEG study were made identifiable in the MR images.
2.1. Patients CAE patients were recruited at Nanjing Brain Hospital (NBH) and the Neurology Department of the Nanjing Children’s Hospital from March 2011 to September 2013. Subjects were prescreened by pediatric neurologists specializing in epilepsy. While 25 CAE patients were recruited, only 12 met the inclusion criteria and were included in this research; other patients were excluded because of excessive head/body motion, non-standard treatment history, and lack of SWDs. The clinical details of the patients are shown in Table 1.
2.4. MEG data analysis Similar to previous reports (Xiang et al., 2010, 2014a), MEG waveforms were visually inspected to identify magnetic noise
Table 1 Clinical characteristics of CAE patients. Patients
Sex/age (F/M)/ (year)
Seizure onset age (year)
Head circumferences (cm)
Epilepsy duration (months)
Seizure frequency (times/day)
Semiology
History/family history
Seizure captures (times)
Ictal duration (s)
1 2 3 4 5 6 7 8 9 10 11 12
M/12 F/12 F/6 F/5 M/8 F/6 M/8 F/5.5 F/10 F/9 F/10 M/6.5
11.5 10.5 6 5 8 6 8 5 9 9 9 6
54.5 53.8 49.2 49.0 50.8 49.3 51.0 49.3 50.5 50.2 51.0 50.8
6 20 1 5 4 4.5 5 6 12 4 10 7.5
10–15 10 10–15 10 10 5–6 6–7 1–2 3–4 15 4–5 4–5
Absence Absence Absence Absence Absence Absence Absence Absence Absence Absence Absence Absence
–/– –/– –/– –/– –/– –/– –/– –/– –/– –/– –/– –/–
3 4 3 6 6 2 2 2 3 4 4 2
18.1 15.4 12.7 11 13.6 28.2 10.7 25 23 5.2 33.3 10.7
Average
–/8.17
7.75
50.78
7.08
8
–
–
–
17.24
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
3
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
and artifacts. MEG data without noise or artifacts (>6 pT) were then filtered with a band pass filter of 1–4 Hz to distinguish interictal and ictal MEG segments because ictal MEG data, which are recorded during absence seizures, are characterized by 3 Hz SWDs in MEG waveforms (Fig. 1). We verified ictal periods with audio– visual system recordings by finding that patients began to stare at the same period of time (impaired consciousness). Thus, to analyze both low- and high-frequency MEG signals, MEG data preprocessing was conducted in two steps. First, MEG data were digitally filtered to 1–4 Hz. Ictal 3 Hz SWDs were clearly identifiable in the waveforms after filtering. Second, seizure start and end times were recorded and applied to all frequency bands for further ictal segment analysis. Of note, selection of the ictal segment was based on both clinical observation of absence and 3-Hz spike-and-slow-wave complexes in MEG. We selected interictal segments that did not contain any visual rhythmic spike wave activity. In other words, if there were long periods of rhythmic spike wave activity without obvious clinical manifestations, we would not select it as interictal segments or ictal segments. In addition, interictal segments were at least 1 min away from ictal segments. 2.4.1. Source localization Magnetic sources of MEG signals in frequency ranges of delta (1–4 Hz), theta (4–8 Hz), alpha (8–12 Hz), beta (12–30 Hz), low gamma (30–45 Hz), high gamma (55–90 Hz), ripple (90–200 Hz), HFOs (200–1000 Hz), and VHFOs (1000–2000 Hz) were localized. We analyzed low gamma (30–45 Hz) and high gamma (55– 90 Hz) MEG data to avoid power-line noise, which was approximately 50 Hz. Building on previous studies (Xiang et al., 2014a), the present study used accumulated source imaging to continuously accumulate total source strengths and source properties without assuming source numbers or time windows. Accumulated source imaging
integrated spatial filter and accumulation into a systematic approach to summarize and visualize brain activity over a long period of time as a volumetric image. The aforementioned method for reconstruction of brain activity was implemented in the MEG Processor with C/C++ in the Windows platform (Xiang et al., 2014a). More than one time slice was used to reveal the fluctuation of source activity in space and time. To increase the reliability of source measurements, we computed ‘‘corrected source strength” according to the following equation:
C s ¼ ST
ð1Þ
where Cs represents the corrected source strength; S represents the strength of source activity and T represents the t-value associated with the correlation coefficient of the measured magnetic field data and the estimated forward solution data MEG. We used accumulated source imaging to continuously accumulate the total source strength and source properties without making any assumption about the number of sources and the time windows. The whole brain was scanned at 6 mm resolution (approximately 17,160 brain areas for each frequency band for each dataset). In the present study, the predominant neuromagnetic activities were typically localized to up to 3 brain areas. In order to segment brain areas, individual MRI was integrated into MEG source imaging with three fiducial points (the left and right pre-auricular points, and the nasion). The MEG Processor automatically marked source peaks. The minimal distance between any two-source peaks was 10 mm. In other words, if the distance between two source peak voxels was 610 mm, they were considered as one source. In this case, the center of the source would be at the voxel with higher spectral power. If the source strength of the two source voxels was equal, the center of the source would be at the middle of the two voxels. Therefore, significant and consistent spatial activities were visually identified in the MRI images. MEG Processor allowed us to visualize and segment brain areas in both 2D and 3D viewers (Xiang et al., 2014a,b). The locations of the sources, such as the occipital cortex, was identified with individual MRI. Accumulated source imaging could localize correlated sources by using node-beam lead fields (Xiang et al., 2014a). Since each node-beam lead field represented a form of ‘‘source-beamfor mer” or ‘‘sub-space solution”, accumulated source imaging had multiple source beamformer to separate correlated sources. Detailed mathematical algorithms and validations related to this approach have been described in a recent paper (Xiang et al., 2014a). 2.4.2. Time–frequency transform To characterize the frequency and spectral signatures of MEG data, complex Morlet continuous wavelet transform was used to transform waveform data into a time–frequency spectrogram. The time–frequency analysis was conducted with Morlet continuous wavelet transform (CWT) in this study, which can be described by the following equation:
GðtÞ ¼ C r p1=4 eð1=2Þt ðeirt jr Þ 2
Fig. 1. MEG waveforms recorded from a child with absence seizures showing selection of interictal MEG data. Two minutes of interictal MEG data was selected after excluding 3 Hz SWDs and visual epileptiform activity or spikes.
ð2Þ
In this equation, t indicates time and f indicates frequency. jr represents admissibility and C r represents normalized constant. r represents the standard deviation of the Gaussian curve in the time domain. The mathematical algorithms, procedures, and parameters employed in this study have been described in detail in previous reports (Xiang et al., 2004, 2013; Kotecha et al., 2009). Six hundred frequency bins and 600 time bands were applied to all MEG spectrograms. We computed the sum of all sensor spectrograms (or global spectrogram) to quantify the spectral power of the entire brain (Xiang et al., 2010). The spectrograms of all MEG sensors are displayed as contour maps to visualize the spatial distribution of brain activities.
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
4
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
Fig. 2. Accumulated source images showing spatial activities in five low-frequency bands (1–4, 4–8, 8–12, 12–30, and 30–45 Hz) recorded from CAE subjects. 3D images are displayed in axial, coronal and oblique sagittal positions. During absence seizures (right column, ‘‘Ictal”), the brain activity source demonstrated large variations and propagated quickly towards adjacent regions in low-frequency bands. The blue and green arrows indicate the predominant neuromagnetic activities of interictal and ictal activities, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
5
Fig. 3. Accumulated source images showing spatial activities in four high-frequency bands (55–90, 90–200, 200–1000, and 1000–2000 Hz) recorded from CAE subjects. 3D images are displayed in axial, coronal and oblique sagittal positions. Interictal and ictal activities are spatially concordant in high-frequency ranges but not in low-frequency ranges. Green arrows indicate predominant neuromagnetic activities during absence seizures (‘‘Ictal”). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
2.5. Statistics Pearson correlation was used to analyze correlation between the number of daily seizure episodes and the strength of each source for each frequency band. Student’s t-test was used to compare source strength in interictal and ictal activities. Fisher’s exact test was performed on predominant neuromagnetic source locations between interictal and ictal activities. The threshold of statistical significance for differences was set at p < 0.05 for each test. For multiple comparisons, Bonferroni correction for multiple comparison was applied (e.g., for nine frequency bands, p < 0.005; up to four brain areas for each frequency band, p < 0.001). To solve the problem of type I errors, we use a statistical procedure for controlling the false discovery rate (FDR) (Genovese et al., 2002; Jacobs et al., 2015). Statistical analyses
were performed using SPSS version 16.0 for Windows (SPSS Inc., Chicago, IL, USA). 3. Results The range of ictal durations was 5.2–33.3 s, and the average ictal duration was 17.24 s. Magnetic source imaging (Figs. 2 and 3) revealed significant differences between interictal and ictal activities in low-frequency bands (<30 Hz, p < 0.0001), not in highfrequency bands. In addition, spectrograms (Fig. 4) also showed that neuromagnetic changes from interictal to ictal periods predominantly occurred in low-frequency bands (<30 Hz, p < 0.0001). Our data showed that each subject typically had 2–3 sources, which were very strong compared with the rest of the brain activity. As a group of patients, interictal and ictal activities in absence
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
6
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
Fig. 4. Global spectrograms (a summation of all MEG sensors) and contour maps showing distributions of epileptic activity in low- and high-frequency bands recorded from subjects with CAE. The global spectrograms (upper row) show central frequencies of interictal and ictal activities. The contour maps (bottom row) show the positions of corresponding spectrograms over the entire brain area. The ictal criterion maps (middle column) show ictal contour maps with spectral power scales identical to those described for the interictal maps. The sample coordinates (last row) indicate the orientations of the contour maps.
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
7
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
seizures predominantly occurred in four brain areas in each frequency band. Although individual variations among the subjects were noted, significant and consistent differences between interictal and ictal periods could still be visually identified.
Table 2 Predominant neuromagnetic activities of MOC, PC, MPFC and POT in interictal and ictal MEG segments.
3.1. Source localization Analysis of conventional low-frequency brain activity in 1–4 Hz (delta) and 4–8 Hz (theta) bands revealed that the medial prefrontal cortex (MPFC) and parieto-occipito-temporal (POT) junction showed high odds of activity during seizures, whereas the precuneus (PC) showed consistent activity in both interictal and ictal activities. Interictal neuromagnetic activity in 8–12 Hz (alpha) was localized to the middle occipital cortex (MOC). In the 12–30 Hz (beta) band, POT showed high odds of activity during seizures. In the 30–45 Hz (low gamma) band, MPFC demonstrated predominant activity during seizures. No significant difference in predominant brain area was found in the 55–90 Hz band. Neuromagnetic MEG signals above 90 Hz were analyzed in three frequency bands, such as 90–200 Hz (ripple), 200–1000 Hz (HFOs), and 1000–2000 Hz (VHFOs). Accumulated source imaging strongly suggested that, in all three bands, subjects with CAE present spatial patterns of source imaging in both interictal and ictal activities in POT and MPFC. No significant difference in predominant brain activity was found between interictal and ictal activities in these three bands. Detailed statistical results for these bandwidths are shown in Table 2. Representative source images from subjects with CAE are shown in Figs. 2 and 3.
Frequency band (Hz)
Source locations
Interictal (times)
Ictal (times)
p value
FDR threshold
1–4
MOC PC MPFC POT
0 11 2 3
0 8 12 12
1.000 0.158 0.0001⁄⁄ 0.0001⁄⁄
0.044 0.013 0.003 0.003
4–8
MOC PC MPFC POT
2 9 1 5
2 6 11 11
0.705 0.200 0.0001⁄⁄ 0.014⁄
0.037 0.018 0.003 0.009
8–12
MOC PC MPFC POT
5 7 6 1
1 6 9 7
0.077 0.500 0.200 0.014⁄
0.012 0.028 0.018 0.009
12–30
MOC PC MPFC POT
2 9 6 2
0 4 9 11
0.239 0.050 0.200 0.0001⁄⁄
0.022 0.010 0.018 0.003
30–45
MOC PC MPFC POT
0 4 4 1
0 2 12 3
1.000 0.320 0.001⁄ 0.295
0.044 0.025 0.006 0.023
55–90
MOC PC MPFC POT
0 0 5 1
0 1 8 1
1.000 0.500 0.207 1.000
0.044 0.028 0.020 0.044
90–200
MOC PC MPFC POT
0 3 1 3
0 3 2 3
1.000 0.680 0.500 0.680
0.044 0.034 0.028 0.034
200–1000
MOC PC MPFC POT
0 0 3 2
0 0 3 3
1.000 1.000 0.680 0.500
0.044 0.044 0.034 0.028
1000–2000
MOC PC MPFC POT
0 0 2 3
0 0 5 3
1.000 1.000 0.185 0.680
0.044 0.044 0.015 0.034
3.2. Clinical correlate We found that the strength of ictal source peak in the 4–8 Hz (theta), 90–200 Hz (ripple), and 200–1000 Hz (HFOs) bands positively correlated with clinical measurements. The strength of ictal source peak in the 200–1000 Hz band (HFOs) showed the strongest correlation with the number of daily seizure episodes (p < 0.01, r = 0.734). Though the strength of ictal source peak in the 4–8 Hz band (p < 0.05, r = 0.587) and 90–200 Hz (p < 0.05, r = 0.640) also correlated with the number of daily seizure episodes, the correlations were relatively weak. Fig. 5 shows the results of the correlations between MEG measurements and clinical seizures. Detailed statistical analysis of neuromagnetic strength of source peak between interictal and ictal segments are shown in Table 3.
MOC: the middle occipital cortex; MPFC: the medial prefrontal cortex; PC: the precuneus; POT: parieto-ocipito-temporal junction. p < 0.05 after Bonferroni multiple comparisons; **p < 0.001 after Bonferroni multiple comparisons. Italic, the result is still significant after correction for multiple comparisons using the FDR (corrected for 4 9 tests) with FDR = pFDR * i/N. Here, pFDR = 0.05 (usual significance value in most statistical tests), i refers to the ranked index of the p-values that are computed and N to the number of tests.
4. Discussion
4.1. Low-frequency neuromagnetic activities
Previous reports on absence epilepsy using MEG/EEG typically focused on 3 Hz ictal SWDs (Sakurai et al., 2010). This study improves our understanding of absence epilepsy by identifying abnormalities in both interictal and ictal activities over a wider frequency range of 1–2000 Hz. Furthermore, accumulated source imaging can visualize epileptic activity of the entire brain volumetrically (Xiang et al., 2014a). Since children’s head circumference may be smaller than those of adults may, it is necessary to take into account the issue of MEG signal attenuation by head-channel distance. Specifically, when the recording position is supine, signal attenuation in the frontal area can be significant, which result in low signal-to-noise ratio in MEG data. We used accumulated source imaging in the present study since this method can increase the signal-to-noise ratio by accumulating MEG signals and spatially filtering noise (Xiang et al., 2014a). However, in our opinion, the best solution to solve this problem is to use a pediatric MEG system. A pediatric MEG system that utilizes a small helmet optimized for children may minimize this problem (Hiraishi et al., 2015).
Source analysis of conventional low-frequency brain activity in the 1–4 Hz band (delta) and the 4–8 Hz (theta) band revealed that MPFC and POT perform different functions in interictal and ictal activities. Whereas PC showed consistent activity in both interictal and ictal activities. The cerebral mechanism of increased lowfrequency neuromagnetic activities (1–4 Hz and 4–8 Hz) in absence epilepsy remains controversial. A study using MEG and diffusion tensor imaging has revealed that abnormal MEG delta waves may be linked to disconnection between the gray and white matters (Huang et al., 2009). Another EEG study suggests that the low-frequency brain activity abnormality maybe a result of thalamic dysfunction. It seems that the reciprocally connected neurons in the thalamus and cortex are responsible for the SWDs (Hughes et al., 2011). In our study, low gamma oscillation (30–45 Hz) showed predominant neuromagnetic activities of MPFC during absence seizures. Previous studies consider that the anterior activity in MPFC is very important in the initialization and propagation of SWDs (Stefan and Rampp, 2009; Sakurai et al., 2010; Gupta et al.,
*
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
8
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
Fig. 5. Charts of Pearson’s correlation showing relationships between source strength of interictal/ictal activities and daily seizure episodes. Ictal source strength was found to be positively correlated with daily seizure episodes in the 4–8 Hz (p < 0.05, r = 0.587), 90–200 Hz (p < 0.05, r = 0.640) and 200–1000 Hz (p < 0.01, r = 0.734) bands. Correlation between the two variables is plotted in the sample coordinate (last row) with source strength as abscissa and daily seizure episodes as ordinate. The unit of source strength has been statistically transformed.
2011). An increase of source strength in 30–45 Hz may reflect frequent abnormal discharges due to chronic disorganization of the functional architecture due to neurophysiologic dysfunction. Other studies have demonstrated that high gamma oscillatory (55– 90 Hz) activity is related to signal transmission in cortical microcir-
cuits and that increased high gamma amplitude leads to multi-unit neural firing in the ictal core from microelectrode and subdural recordings (Sohal, 2012; Weiss et al., 2013). Our results support the notion that subjects with absence seizures demonstrate significant elevation of low-frequency
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx Table 3 Analysis of neuromagnetic peak source strength between interictal and ictal segments. Frequency band (Hz)
Interictal
Ictal
t value
p value
FDR threshold
1–4 4–8 8–12 12–30 30–45 55–90 90–200 200–1000 1000–2000
6.82 ± 1.89 4.31 ± 1.66 3.16 ± 1.35 2.47 ± 0.53 1.45 ± 1.67 1.88 ± 3.14 1.15 ± 1.61 1.56 ± 1.88 0.27 ± 0.21
61.01 ± 17.80 19.34 ± 8.98 9.10 ± 3.20 8.84 ± 4.17 3.37 ± 2.72 2.93 ± 4.05 1.73 ± 1.55 2.15 ± 1.88 0.26 ± 0.21
10.274 5.673 7.300 5.602 2.838 3.140 2.113 2.242 0.437
0.0001⁄⁄ 0.0001⁄⁄ 0.0001⁄⁄ 0.0001⁄⁄ 0.016⁄ 0.009⁄ 0.058 0.047⁄ 0.671
0.013 0.013 0.013 0.013 0.033 0.027 0.044 0.038 0.050
*
p < 0.05 after Bonferroni multiple comparisons; **p < 0.001 after Bonferroni multiple comparisons. Italic, the result is still significant after correction for multiple comparisons using the FDR (corrected for 9 tests).
activities. In addition, the source of low-frequency brain activity presented large variations and propagated quickly towards adjacent regions. This indicates that visual waveform inspection does not always result in delineation of a potential epileptogenic zone (EZ). In the case of refractory epilepsy, this may lead to incomplete resection of ictal onset zone and result in residual seizures (Kim et al., 2010; Weiss et al., 2013; Lee et al., 2014). 4.2. Source strength of HFOs and clinical correlation Increasing evidence suggests that HFOs and VHFOs exist in epilepsy patients and can be used to localize epileptogenic origins before surgery (Jacobs et al., 2008; Usui et al., 2010; Kerber et al., 2014). Although our understanding of how specific information is carried by HFOs is poor, many experiments have verified that spikes and HFOs present different pathophysiologic mechanisms and HFOs are more clinically related to seizures than lowfrequency brain activities. HFOs can occur with or without epileptic spikes (Jacobs et al., 2012; Kobayashi et al., 2013). Studies on pre-ictal and ictal HFOs have suggested that only channels with HFOs during seizure onset show HFOs during seizure evolution. Compared with spikes, these HFOs seem more specific for SOZ and could be helpful in determining the EZ in the clinical practice of epilepsy surgery (Zijlmans et al., 2012; Weiss et al., 2013). In other studies, interictal HFOs obtained during sleep, but not those based on the record of an ictal event, were also found to provide reliable information regarding SOZ localization (Andrade-Valenca et al., 2012). Larger numbers of interictal HFOs indicate greater risk of seizure; thus interictal HFOs are a sign of seizure susceptibility (Zijlmans et al., 2011; Jacobs et al., 2012). Using an animal model, Bragin et al. found a strong correlation between decreased times of HFO detection and increased rates of spontaneous seizures (Bragin et al., 2004). Zijlmans et al. found that HFOs increase after seizures and reduction of medication in intractable focal epilepsy patients (Zijlmans et al., 2009). Therefore, we postulate that HFOs is a useful clinical marker for epilepsy. Previous reports on HFOs in epilepsy relied mainly on intracranial EEG recordings (Usui et al., 2010; Weiss et al., 2013). Considering that intracranial EEG recordings are extremely invasive and ethically prohibited in the study of patients not requiring brain surgery, the noninvasive MEG approach utilized in the present study with excellent results may be appropriate for future studies of HFOs and VHFOs. Based on the MEG data recorded from CAE in the 90–200 Hz (ripple), 200–1000 Hz (HFOs), and 1000–2000 Hz (VHFOs) bands, accumulated source imaging strongly suggested that interictal and ictal activities are spatially concordant in high-frequency ranges but not in low-frequency ranges. Strength of ictal peak source was notably higher than strength of interictal peak source in the 200–1000 Hz (HFOs) band. Hence, HFOs may present
9
noticeable high-frequency source strength changes during seizure but fewer positional variations compared with low-frequency MEG signals in CAE subjects. Low-frequency activities are propagated epileptic activities, as those activities have no close relationship with clinical severity (number of daily seizure episodes). In other words, high-frequency epileptic activities may have closer relationship with EZ than low-frequency epileptic activities, and may provide precise spatial information about cortical dysfunction (Xiang et al., 2013). This observation is consistent with previous reports on medically intractable epilepsy (Xiang et al., 2009, 2010). MPFC and POT are the primary nodes of the default mode network. The default mode network has been hypothesized on the basis of the observation that specific regions of the brain are consistently activated during the resting state but deactivated when engaged with tasks (Danielson et al., 2011). This network is assumed to be selectively impaired during epileptic seizures associated with loss of consciousness, such as absence seizures. The initial portion of ictal activation of absence seizures is predominantly localized to the MPFC and POT regions. 5. Conclusion In this study involving CAE patients, distinct neuromagnetic signatures were found to be associated with interictal and ictal activities in both low- and high-frequency ranges. The magnetic sources observed in this work support the notion that focal cortical areas are responsible for absence seizures. Though low-frequency neuromagnetic activities showed greater changes from interictal to ictal periods, the strength of neuromagnetic HFOs (200–1000 Hz) significantly correlated with clinical severity of absence seizures. Therefore, low- and high-frequency neuromagnetic signals may reveal distinct brain activities in CAE. Neuromagnetic HFOs (200–1000 Hz) is a new imaging biomarker for the study of absence seizures. Acknowledgments We would like to thank the physicians and fellows at NBH and Nanjing Children’s Hospital. We thank the MEG Center at Cincinnati Children’s Hospital for helping with MEG data analysis. We also thank all participants and their families for their time and support. The work was supported by the National Natural Science Foundation of China (Grant No. 81471324, http://npd.nsfc.gov.cn/), the Key Project of Medical Science and Technology Development Foundation (Grant No. ZKX11002, http://www.njh.gov.cn/html/list_83. shtml), the Fourth Phase of Jiangsu ‘‘Project 333” Scientific Research Funding Schemes, 2013, the Health Department of Jiangsu Province (Grant No. H201443) and the Nanjing Medical University General Program (Grant No. 2014NJMU050). Conflict of interest: We declare that all authors have no financial and personal relationships with other people or organizations that could have inappropriately influenced this work. None of the authors has potential conflicts of interest to be disclosed. References Andrade-Valenca L, Mari F, Jacobs J, Zijlmans M, Olivier A, Gotman J, et al. Interictal high frequency oscillations (HFOs) in patients with focal epilepsy and normal MRI. Clin Neurophysiol 2012;123:100–5. Bragin A, Wilson CL, Almajano J, Mody I, Engel Jr J. High-frequency oscillations after status epilepticus: epileptogenesis and seizure genesis. Epilepsia 2004;45: 1017–23. Cerminara C, D’Agati E, Casarelli L, Kaunzinger I, Lange KW, Pitzianti M, et al. Attention impairment in childhood absence epilepsy: an impulsivity problem? Epilepsy Behav 2013;27:337–41. Danielson NB, Guo JN, Blumenfeld H. The default mode network and altered consciousness in epilepsy. Behav Neurol 2011;24:55–65.
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016
10
L. Tang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx
Genovese CR, Lazar NA, Nichols T. Thresholding of statistical maps in functional neuroimaging using the false discovery rate. NeuroImage 2002;15:870–8. Glauser TA, Cnaan A, Shinnar S, Hirtz DG, Dlugos D, Masur D, et al. Ethosuximide, valproic acid, and lamotrigine in childhood absence epilepsy: initial monotherapy outcomes at 12 months. Epilepsia 2013;54:141–55. Gupta D, Ossenblok P, van Luijtelaar G. Space-time network connectivity and cortical activations preceding spike wave discharges in human absence epilepsy: a MEG study. Med Biol Eng Comput 2011;49:555–65. Hiraishi H, Kikuchi M, Yoshimura Y, Kitagawa S, Hasegawa C, Munesue T, et al. Unusual developmental pattern of brain lateralization in young boys with autism spectrum disorder: power analysis with child-sized magnetoencephalography. Psychiatry Clin Neurosci 2015;69:153–60. Holmes MD, Brown M, Tucker DM. Are ‘‘generalized” seizures truly generalized? Evidence of localized mesial frontal and frontopolar discharges in absence. Epilepsia 2004;45:1568–79. Huang MX, Theilmann RJ, Robb A, Angeles A, Nichols S, Drake A, et al. Integrated imaging approach with MEG and DTI to detect mild traumatic brain injury in military and civilian patients. J Neurotrauma 2009;26:1213–26. Hughes SW, Lorincz ML, Parri HR, Crunelli V. Infraslow (<0.1 Hz) oscillations in thalamic relay nuclei basic mechanisms and significance to health and disease states. Prog Brain Res 2011;193:145–62. Jacobs J, LeVan P, Chander R, Hall J, Dubeau F, Gotman J. Interictal high-frequency oscillations (80–500 Hz) are an indicator of seizure onset areas independent of spikes in the human epileptic brain. Epilepsia 2008;49:1893–907. Jacobs J, Zijlmans M, Zelmann R, Chatillon CE, Hall J, Olivier A, et al. High-frequency electroencephalographic oscillations correlate with outcome of epilepsy surgery. Ann Neurol 2010;67:209–20. Jacobs J, Staba R, Asano E, Otsubo H, Wu JY, Zijlmans M, et al. High-frequency oscillations (HFOs) in clinical epilepsy. Prog Neurobiol 2012;98:302–15. Jacobs J, Vogt C, LeVan P, Zelmann R, Gotman J, Kobayashi K. The identification of distinct high-frequency oscillations during spikes delineates the seizure onset zone better than high-frequency spectral power changes. Clin Neurophysiol 2015. http://dx.doi.org/10.1016/j.clinph.2015.04.053. Kerber K, Dumpelmann M, Schelter B, Le Van P, Korinthenberg R, Schulze-Bonhage A, et al. Differentiation of specific ripple patterns helps to identify epileptogenic areas for surgical procedures. Clin Neurophysiol 2014;125:1339–45. Kim DW, Kim HK, Lee SK, Chu K, Chung CK. Extent of neocortical resection and surgical outcome of epilepsy: intracranial EEG analysis. Epilepsia 2010;51:1010–7. Kobayashi K, Miya K, Akiyama T, Endoh F, Oka M, Yoshinaga H, et al. Cortical contribution to scalp EEG gamma rhythms associated with epileptic spasms. Brain Dev 2013;35:762–70. Kotecha R, Xiang J, Wang Y, Huo X, Hemasilpin N, Fujiwara H, et al. Time, frequency and volumetric differences of high-frequency neuromagnetic oscillation between left and right somatosensory cortices. Int J Psychophysiol 2009;72: 102–10. Lee C, Kim JS, Jeong W, Chung CK. Usefulness of interictal spike source localization in temporal lobe epilepsy: electrocorticographic study. Epilepsy Res 2014;108: 448–58.
Proposal for revised classification of epilepsies and epileptic syndromes. Commission on Classification and Terminology of the International League against Epilepsy. Epilepsia 1989;30:389–99. Sakurai K, Takeda Y, Tanaka N, Kurita T, Shiraishi H, Takeuchi F, et al. Generalized spike-wave discharges involve a default mode network in patients with juvenile absence epilepsy: a MEG study. Epilepsy Res 2010;89:176–84. Sohal VS. Insights into cortical oscillations arising from optogenetic studies. Biol Psychiatry 2012;71:1039–45. Stefan H, Rampp S. Current clinica–neurophysiological findings in absence epilepsies. Nervenarzt 2009;80:378–85. Tenney JR, Glauser TA. The current state of absence epilepsy: can we have your attention? Epilepsy Curr 2013;13:135–40. Usui N, Terada K, Baba K, Matsuda K, Nakamura F, Usui K, et al. Very high frequency oscillations (over 1000 Hz) in human epilepsy. Clin Neurophysiol 2010;121: 1825–31. Weiss SA, Banks GP, McKhann Jr GM, Goodman RR, Emerson RG, Trevelyan AJ, et al. Ictal high frequency oscillations distinguish two types of seizure territories in humans. Brain 2013;136:3796–808. Wu JY, Sankar R, Lerner JT, Matsumoto JH, Vinters HV, Mathern GW. Removing interictal fast ripples on electrocorticography linked with seizure freedom in children. Neurology 2010;75:1686–94. Xiang J, Holowka S, Qiao H, Sun B, Xiao Z, Jiang Y, et al. Automatic localization of epileptic zones using magnetoencephalography. Neurol Clin Neurophysiol 2004;2004:98. Xiang J, Liu Y, Wang Y, Kirtman EG, Kotecha R, Chen Y, et al. Frequency and spatial characteristics of high-frequency neuromagnetic signals in childhood epilepsy. Epileptic Disord 2009;11:113–25. Xiang J, Wang Y, Chen Y, Liu Y, Kotecha R, Huo X, et al. Noninvasive localization of epileptogenic zones with ictal high-frequency neuromagnetic signals. J Neurosurg Pediatr 2010;5:113–22. Xiang J, deGrauw X, Korostenskaja M, Korman AM, O’Brien HL, Kabbouche MA, et al. Altered cortical activation in adolescents with acute migraine: a magnetoencephalography study. J Pain 2013;14:1553–63. Xiang J, Luo Q, Kotecha R, Korman A, Zhang F, Luo H, et al. Accumulated source imaging of brain activity with both low and high-frequency neuromagnetic signals. Front Neuroinform 2014a;8:57. Xiang J, Tenney JR, Korman AM, Leiken K, Rose DF, Harris E, et al. Quantification of interictal neuromagnetic activity in absence epilepsy with accumulated source imaging. Brain Topogr 2014. http://dx.doi.org/10.1007/s10548-014-0411-5. Zijlmans M, Jacobs J, Zelmann R, Dubeau F, Gotman J. High-frequency oscillations mirror disease activity in patients with epilepsy. Neurology 2009;72: 979–86. Zijlmans M, Jacobs J, Kahn YU, Zelmann R, Dubeau F, Gotman J. Ictal and interictal high frequency oscillations in patients with focal epilepsy. Clin Neurophysiol 2011;122:664–71. Zijlmans M, Jiruska P, Zelmann R, Leijten FS, Jefferys JG, Gotman J. High-frequency oscillations as a new biomarker in epilepsy. Ann Neurol 2012;71:169–78.
Please cite this article in press as: Tang L et al. Neuromagnetic high-frequency oscillations correlate with seizure severity in absence epilepsy. Clin Neurophysiol (2015), http://dx.doi.org/10.1016/j.clinph.2015.08.016