Epilepsy Research (2014) 108, 257—266
journal homepage: www.elsevier.com/locate/epilepsyres
Causal influence of epileptic network during spike-and-wave discharge in juvenile myoclonic epilepsy Chany Lee a, Sung-Min Kim a, Young-Jin Jung b, Chang-Hwan Im b, Dong Wook Kim c, Ki-Young Jung a,∗ a
Department of Neurology, College of Medicine, Korea University, Seoul, Republic of Korea Department of Biomedical Engineering, Hanyang University, Seoul, Republic of Korea c Department of Neurology, College of Medicine, Konkuk University, Seoul, Republic of Korea b
Received 24 May 2013; received in revised form 7 October 2013; accepted 3 November 2013 Available online 16 November 2013
KEYWORDS Juvenile myoclonic epilepsy; Spike-and-wave discharge; Network; Effective connectivity; Precuneus
Summary Electroencephalographic (EEG) characteristic of juvenile myoclonic epilepsy (JME) is spike-and-wave discharge (SWD), which is dominant in the frontal region. However, activity in the parietal area, including the precuneus, has also been documented for several seconds before and during SWD. The aim of this study was to identify the role of the parietal region, especially the precuneus, and to clarify the causal dynamics among cortical regions during SWD. EEGs were obtained from seven patients with JME. Each SWD was divided into six distinct temporal phases: spike onset, spike peak, slow-wave onset, slow-wave ascending, slow-wave peak, and slow-wave descending phases. Based on the cortical current source distribution and the results of a previous study, we selected the medial frontal, orbitofrontal, anterior cingulate, and mesial temporal cortices and the precuneus as regions of interest (ROIs). To assess epileptic networks and the causal relationships among ROIs during SWD, the directed transfer function (DTF), a measure of multivariate causality, was calculated for each phase of SWD. During spike onset, the maximal outdegree region in all patients was the precuneus. The spike-peak and slow-wave onset phases did not show a consistently dominant outflow region. Outflow from the anterior cingulate cortex was dominant in four patients during the slow-wave ascending phase, and the precuneus showed the maximal outdegree in six patients during the slow-wave peak. In the slow-wave descending phase, four patients showed maximal outflow from the temporal
∗ Corresponding author at: Department of Neurology, Korea University Medical Center, Korea University College of Medicine, #126-1, Anam-Dong 5Ga, Seongbuk-Gu, Seoul 136-705, Republic of Korea. Tel.: +82 2 920 6649; fax: +82 2 925 2472. E-mail address: [email protected] (K.-Y. Jung).
0920-1211/$ — see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.eplepsyres.2013.11.005
258
C. Lee et al. cortex. Our findings suggest that the precuneus is likely a key region for SWD despite the small amount of neural activity observed. The precuneus was the region with the maximal outdegree during both the spike onset and slow-wave peak phases, indicating that SWD in JME is initiated and sustained by a network involving the frontal cortex, precuneus, and thalamus. © 2013 Elsevier B.V. All rights reserved.
Introduction Juvenile myoclonic epilepsy (JME) is the most common idiopathic generalized epilepsy (IGE) syndrome, accounting for 5—10% of all cases of epilepsy (Janz and Christian, 1957; Janz, 1985; Vollmar et al., 2011). The characteristic features of JME are myoclonic jerks on awakening, generalized tonic—clonic seizures, and less frequent absence seizures. In terms of electroencephalography (EEG), the hallmark of JME consists of bilateral, synchronous, widespread spike or polyspike and wave complexes, which are commonly assumed to occur without lateralizing or localizing features (Nordli, 2005). Combined EEG and functional magnetic resonance imaging (fMRI) studies in patients with IGE, including JME, have shown a clear spike-and-wave discharge (SWD)-associated increase in the blood oxygenation level-dependent (BOLD) signal in the thalamus and a decrease in the BOLD signal in the parietal and frontal cortical regions, which is consistent with the default mode network (DMN) of the human brain (Archer et al., 2003; Gotman et al., 2005; Hamandi et al., 2006; Labate et al., 2005). Time-course analyses of the BOLD signal during SWD have shown BOLD increments starting approximately 5—10 s before the onset of SWD in the precuneus (PCN)/posterior cingulate cortex (PCC) and medial/lateral parietal cortex (Benuzzi et al., 2012) as well as in the orbital/medial frontal cortex (Bai et al., 2010). Additionally, a recent magnetoencephalograpihc (MEG) study showed that low-frequency sources were sequentially activated in the frontal and occipital regions prior to the first generalized spikes (Gupta et al., 2011). An fMRI study using dynamic causal modeling suggested that activity in the PCN gates SWD in IGE in the thalamocortical network (Vaudano et al., 2009). The posteromedial cortical region, including the PCN and PCC, showed the highest degree of interactions with the rest of the DMN, suggesting that this area plays a pivotal role in the network. This finding indicates that the posterior cortex may influence SWD generation such that it drives the SWD network under pathological circumstances. Although EEG-fMRI studies have been used for the localization of SWD in IGE (Aghakhani et al., 2004; Gotman et al., 2005; Moeller et al., 2008a), fMRI reflects the BOLD signal, which is indirectly related to the underlying neuronal activity and cannot describe instantaneous neural activities of SWD because the BOLD signal is an accumulation of signals over several seconds (Archer et al., 2003; Blumenfeld, 2005; Sullivan and Dlugos, 2004). Additionally, most EEG-fMRI studies oversimplify the hemodynamic response function (HRF) related to brain activity, as most activity cannot be measured by the standard HRF used for conventional fMRI analyses of SWD in IGE (Blumenfeld, 2012).
Recently, as epilepsy has been considered network disorder, epileptic brain networks have been evaluated using mathematical measures of connectivity (Amor et al., 2009; Gupta et al., 2011; Killory et al., 2011; Kramer and Cash, 2012; Zhang et al., 2011). In our study, we adopted a directed transfer function (DTF), which is a type of multivariate application of Granger causality (Kami´ nski and Blinowska, 1991), since DTF can efficiently estimate causal interactions among multiple EEG signals. DTFs have been extensively used in the analysis of epileptic networks (Ding et al., 2007; Franaszczuk and Bergey, 1998; Lu et al., 2012; Wilke et al., 2011). A series of studies have demonstrated that the DTF technique can be used to identify ictal onset zones from intracranial EEG recordings in cases of mesial and lateral temporal lobe epilepsies as well as in cases of neocortical extra-temporal lobe epilepsies. In addition, it was recently reported that seizure activity in patients with symptomatic generalized epilepsy, such as Lennox—Gastaut syndrome, can be localized using DTF analysis (Jung et al., 2011). In the present study, to assess epilepsy networks and the causal relationship among cortical regions during SWD in patients with JME, we first identified cortical current sources and applied DTF on these current sources during SWD. We also tried to elucidate the role of the PCN in gating SWD during the ictogenesis of JME.
Methods Subjects Seven patients with a clinical diagnosis of JME were included in this study. The diagnosis of JME was based on electroclinical criteria as per the International League Against Epilepsy (ILAE) classification (Commission of ILAE, 1989). Inclusion criteria were as follows: (1) a typical clinical history of JME with onset of myoclonic jerks and generalized seizures in adolescence; (2) no evidence of neurological abnormality and intellectual decline; (3) apparent spike or polyspike and wave discharges on normal background rhythm on previous standard international 10—20 EEG; and (4) normal MRI finding by visual inspection. The study protocol was approved by the Institutional Review Board of Korea University Medical Center. Clinical data of the included patients are summarized in Table 1.
EEG recording A 64-channel EEG signal and two electrooculogram channels were recorded at a sampling rate of 1600 samples/s using an EEG recording system (Grass Technologies, Quincy, MA, USA). An electrode cap with sintered Ag/AgCl electrodes was
Effective connectivity in SWD Table 1
259
Clinical characteristics of patients.
Patient #
Sex
Age (year)
Onset (year)
EEG
1 2 3 4 5 6 7
M M F M M M F
42 19 24 19 24 32 17
13 16 22 16 18 30 12
4—5 Hz 4—5 Hz 3—4 Hz 3—4 Hz 2—3 Hz 3—4 Hz 3—4 Hz
25.3 8.9
18.1 6.2
Mean SD
AEDs polyspike and wave SWD SWD SWD SWD SWD SWD
LMT, VPA LEV LEV VPA LEV ZNS VPA, LEV
AED, antiepileptic drug; LTG, lamotrigine; VPA, valproate; LEV, levetriracetam; ZNS, zonisamide; SD, standard deviation.
used (Quick-Cap, Compumedics Neuroscan, Charlotte, NC, USA). The impedances of all electrodes were reduced below 10 k. The reference electrode was set to linked mastoid electrodes, and the bandpass filter was set at 0.3—70 Hz. The patient was seated comfortably in a chair, and EEG recordings were taken during waking, sleep, and hyperventilation for approximately 1 h.
Epoching of SWD Whole EEG recordings were inspected visually, and typical SWDs were identified (KJ). All patients showed bilateral widespread SWD, with runs at frequencies of either 3—4 Hz or 4—6 Hz (Fig. 1). The onset of SWD was defined as the first large transition in SWD signals (Holmes et al., 2010), and
SWDs were extracted from −100 ms to 900 ms, where 0 ms was the spike onset time. The preictal state was defined as the period from -100 ms to 0 ms. As the early phase of SWD may have a more consistent network than do the later phases of the seizure (Meeren et al., 2002), only the first SWD of the ictal discharges was considered for analysis. In order to assess the time-varying causality of SWD, a single SWD was divided into six phases: spike onset, spike peak, slow-wave onset, slow-wave ascending, slow-wave peak, and slow-wave descending phases (Fig. 2).
Current density distribution of SWD Forward problem was solved using Brainstorm (http://neuroimage.usc.edu/brainstorm) with default
Figure 1 An example of spike-and-wave discharge (SWD) on an EEG recording. The vertical dashed line indicates the onset of SWD. The reference was set to linked mastoid electrodes, and negativity is directed upward. Electrodes in the figure were selected according to the international 10—20 system.
260
C. Lee et al.
Figure 2 Overlay map of current density distribution from 98 epochs in all patients. After calculating the current density for each epoch, the threshold was set at 30% of the maximal current density, and binary results were acquired and accumulated. Thus, the distribution in this figure indicates the frequency of activation from all epochs instead of the actual current density.
anatomy data (Tadel et al., 2011). In Brainstorm, the default numerical models of the cortex, scalp, and the inner and outer skull boundaries included 15,028, 1028, 642, and 642 vertices and 29,899, 2160, 1280, and 1280 triangles, respectively. The electrical conductivities of the scalp, skull, and cerebrospinal fluid/brain were set as default values of 1, 0.0125, and 1 S/m, respectively (Ahrens et al., 2012). The boundary element method by OpenMEEG embedded in Brainstorm was used (Gramfort et al., 2010). For the inverse algorithm, standardized low-resolution brain electromagnetic tomography (sLORETA) was adopted (Pascual-Marqui, 2002). An inverse operator of the leadfield matrix calculated by Brainstorm was extracted and applied to each time slice of each epoch so that the current density distribution was obtained for any given time point during each phase.
Selection of ROIs To locate the frequently activated regions during each phase of SWD across epochs in each patient, current densities greater than 30% of maximal value (approximately at a signal-to-noise ratio (SNR) of 5 dB) remained as binary distributions and were summed over all epochs. The resultant distribution of cortical activation refers to overlay map. The common distribution of current densities across patients was then obtained for each phase of SWD using the same method (Fig. 2). Based on the current source distribution, the orbitofrontal cortex (OFC), medial frontal cortex (MFC),
mesial temporal cortex (MTC), and anterior cingulate cortex (ACC) were selected as regions of interest (ROIs). Despite its relative lack of activation, we also included the PCN as a ROI, as previous studies have suggested that it may play a key role in SWD (Benuzzi et al., 2012; Vaudano et al., 2009). The areas and locations of the ROIs in MNI coordinates are presented in Table 2. In this study, we considered only the left hemisphere because SWD has bilateral characteristics in JME (Meeren et al., 2002; Sullivan and Dlugos, 2004), although it is likely that some asymmetric distribution also occurs. The current source density was averaged over each ROI, so each ROI had its own representative current source density.
Directed transfer function analysis DTF is a measure of the information flow through multichannel signals (Blinowska et al., 2010; Ding et al., 2007). In order to obtain information flow during each phase of SWD, the window length was set to 50 data points (31.25 ms). Because multivariate autoregressive modeling (MVAR) can be applied only to stationary data (Chatfield, 2004; Cryer and Chan, 2008), we tested the stationarity of windowed data by dividing them into four temporal segments. Each segment had 20 time slices with 50% overlap. After calculating the autocorrelation function for each segment, we found that autocorrelation was a function of time-lag only (Fig. S1), which means that the representative current sources of the five ROIs were stationary in a window and that MVAR could be applied.
Effective connectivity in SWD Table 2
261
Locations and areas of selected ROIs.
ROIs
Center location in MNI coordinates (mm)
MFC ACC OFC MTC PCN
x
y
0 −2 −6 −18 −1
21 16 62 −2 −73
Brodmann area
Area (mm2 )
6, 8 24 11 20, 28, 38 7
315 167 390 400 130
z 53 27 −10 −29 50
MFC, medial frontal cortex; ACC, anterior cingulate cortex; OFC, orbitofrontal cortex; MTC, mesial temporal cortex; PCN, precuneus.
Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.eplepsyres.2013.11.005. The MVAR order was determined by the Akaike information criterion (AIC) (Akaike, 1974; Ding et al., 2000; Philiastides and Sajda, 2006). The model order of MVAR was determined to be 7 because AIC values did not significantly change above order 7 (Fig. S2), and similar results were obtained in all windows. Hence, a model order of 7 was applied to all windows. Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/ 10.1016/j.eplepsyres.2013.11.005. For each window, 500 surrogate data were employed and p-values < 0.01 were considered to be statistically significant. To construct a binary adjacency matrix, causal influence from region j to region i was 1 if DTF had statistical meaning or 0 if it did not. In a given patient, an individual adjacency matrix was obtained after adjacency matrices for each phase across epochs were averaged (Fig. 3). Finally, a connectivity matrix for each phase was generated by
Figure 3
averaging all individual adjacency matrices for each phase across patients.
Results A total of 98 epochs (mean, 14.0 ± 9.6) of SWD were selected from seven patients. The average duration and standard deviation of the first SWD were 425.6 ms and 112.3 ms, respectively. An example of SWD in patient no. 3 is shown in Fig. 1.
Current source distribution In each patient, the distribution of current density at each phase of SWD was usually maximal over the medial frontal regions, including the ACC and OFC, and the anteromesial temporal regions. Although the distribution did not change significantly across phases, the magnitude of the current density did differ significantly between phases.
Schematic diagram of directed transfer function (DTF) analysis with a short window for a single patient.
262 The overlay map for each phase across patients is demonstrated in Fig. 2. Similar to the results of the individual analysis, the most commonly involved areas were the frontotemporal cortical regions, including the medial frontal, anterior cingulate, and temporal cortices. Of them, the ACC was the most consistently activated region (61.2%, 60 of 98 epochs) during the spike period (spike onset and spike peak). The anteromesial temporal region (59.2%, 59 of 98 epochs) was the next most frequently activated area during these phases. During slow wave phases, the medial frontal and ACC were most commonly involved, showing activation in 90 of 98 epochs.
C. Lee et al.
Causal influence among ROIs The DTF was estimated for all phases. For an instance, the DTF results of patient no. 6 are shown in Fig. 4. Fig. 5 shows the mean outdegree across epochs and patients for each phase of SWD. Despite intraindividual variability, noticeable outflow characteristics were observed. The DTF during spike onset showed that the PCN exhibited maximal outflow of information for all patients (63 epochs), which is the most consistent finding across individuals. Spike peak and slowwave onset phases did not have a consistently dominant outflow region. Outflow from the ACC was dominant in four patients (53 epochs) during slow-wave ascending phase, and
Figure 4 DTF results for patient no. 6. (A) Spike onset, (B) spike peak, (C) slow-wave onset, (D) slow-wave ascending, (E) slowwave peak, and (F) slow-wave descending phases. The left panel of each phase indicates information flow (from tail to head) during the representative epoch. The middle panel is an individual adjacency matrix. M, A, O, T, and P refer to the medial frontal cortex, anterior cingulate cortex, orbitofrontal cortex, mesial temporal cortex, and precuneus, respectively. The color represents the mean outdegree of each region. The right panel shows the indegree and outdegree of each region. Inflow and outflow are the vertical and horizontal summations of the adjacency matrix, respectively.
Effective connectivity in SWD
263
Figure 5 Mean outdegree across epochs and patients for each phase of spike wave discharge. The outdegree of the precuneus (PCN) is slightly over 2 in spike onset phase, which means that at least two regions receive information from the PCN. The medial frontal cortex and the anterior cingulate cortex in slow-wave ascending phase, the PCN in slow-wave peak phase, and the mesial temporal cortex in slow-wave descending phase are dominant. In the preictal state, the outdegree of all ROIs did not exceed 1.
the PCN was the maximal outdegree region during slow-wave peak in six patients (68 epochs). In slow-wave descending phase, four patients (51 epochs) showed outflow from the MTC that was superior to that of other regions. No noticeable outdegree was noted just before SWD onset or during the preictal baseline period compared with other phases.
Discussion In the present study, we investigated the epilepsy network of SWD and the causal influence among cortical regions that account for its network in patients with JME. Source localization revealed that the frontotemporal network, including the medial frontal gyri, anterior cingulate, orbitofrontal area, and anteromesial temporal region, were primarily involved in the generation of SWD in JME. Frontal cortical involvement during SWD has been consistently reported in various neuroimaging studies (Benuzzi et al., 2012; Kim et al., 2012; Santiago-Rodríguez et al., 2002; Stefan et al., 2009; Westmijse et al., 2009; Woermann et al., 1999). This frontal involvement appears to relate thalamo-cortical network, which is an important feature of SWD in IGE (Anderson and Hamandi, 2011; Bernhardt et al., 2009). In order to investigate functional connectivity and causal influence within the JME network, we selected four ROIs in the frontotemporal regions in which the current source distribution was most active during SWD. Additionally, although precuneal activity exceeding the threshold was observed in only one patient, the PCN was selected as an additional ROI based on the following recent EEG-fMRI findings; (1) SWD-related BOLD signal changes showed both activation and deactivation not only in the thalamus and frontal regions but also in the posterior cortical regions (Gotman, 2008; Hamandi et al., 2006); (2) activation in the PCN/posterior cingulate region precedes that in the thalamo-frontal regions (Moeller et al., 2008a); and (3) a dynamic causal model of BOLD during SWD suggested that
the PCN may act on the thalamic-frontal cortical network in facilitating the development of SWD (Vaudano et al., 2009). In our study, the outflow of information during spike onset was most dominant in the PCN in all patients, whereas the current source was maximal over the frontotemporal regions. In contrast, no significant outflow of information was observed in the medial frontal regions, which was the predominant activated region during spike peak. This finding suggests that the PCN may provide information to the medial frontal regions for the initiation of SWD in the thalamo-frontotemporal network, which is conceptually consistent with the previous EEG-fMRI study (Vaudano et al., 2009). This information flow may be supported by direct anatomical connections between the PCN and the frontal lobe (Goldman-Rakic, 1988; Leichnetz, 2001; Petrides and Pandya, 1984; Vaudano et al., 2009). In addition, there are subcortical connections between the PCN and the thalamus (Cavanna and Trimble, 2006; Yeterian and Pandya, 1985). Although recent research has proposed that a focal cortical region was found to lead the thalamus (Meeren et al., 2002), the thalamus is still thought to play an essential role in the thalamo-cortical network in IGE (Avoli, 2012; Meeren et al., 2005), and increments in the BOLD signal in the thalamus during SWD have been consistently reported (Aghakhani et al., 2004; Archer et al., 2003; Benuzzi et al., 2012; Gotman et al., 2005; Hamandi et al., 2006). Therefore, it can be speculated that the PCN may influence medial frontal regions via a direct white matter pathway and/or thalamocortical projections. Furthermore, the outdegree of the PCN was dominant during slow-wave peak phase as well as and spike onset phases. This finding implies that the PCN acts not only as an initiator but also as a mediator of SWD. Taken together, our data suggest that SWD is initiated and sustained by a frontal-precuneal-thalamic network in JME. Previous EEG-fMRI studies demonstrated that SWD was strongly related to the DMN (Laufs et al., 2006; McGill et al., 2012; Moeller et al., 2011, 2008b). Both activation and deactivation of the BOLD signal within the DMN were
264 observed during SWD (Laufs et al., 2006). It has also been reported that the structural and functional connectivity of the DMN is changed in IGE (Bernhardt et al., 2009; Luo et al., 2011; O’Muircheartaigh et al., 2011; Yang et al., 2013). The changes in the neuronal state within the PCN may reflect spontaneous fluctuations in awareness and may affect the thalamo-frontal network of SWD (Vaudano et al., 2009). The PCN is one of the key regions in the DMN, as it showed the highest resting metabolic rates, indicating that a high level of information processing was occurring in this region during the resting state of the brain (Gusnard and Raichle, 2001). Structural abnormalities, such as changes in gray matter volume, have also been found in the DMN regions of patients with JME (O’Muircheartaigh et al., 2011; Ronan et al., 2012; Woermann et al., 1999). As shown in Fig. 5, the preictal phase which was defined as the 100-ms period just before the onset of SWD did not show any significant information flow among the ROIs. In other words, we observed no noticeable interaction among ROIs during the preictal period. In contrast, EEG-fMRI studies have reported that the parietal region was involved several seconds before SWD onset (Benuzzi et al., 2012; Moeller et al., 2008a). Hence, it could be interpreted that parietal region does not interact with other regions before SWD onset and starts sending information to other regions at SWD onset.
Conclusion In conclusion, our study supports the notion that the PCN may play an important role for onset of SWD in the thalamo-frontal cortical network. Moreover, the PCN is also a mediator of SWD so that epileptic network is sustained during SWD.
Conflict of interest None of the authors has any conflict of interest to disclose.
Acknowledgments This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2012R1A6A3A01019371) and by a grant from the Korea Healthecare Technology R&D Project, Ministry for Health, Welfare & Family Affairs, Republic of Korea (A090794).
References Aghakhani, Y., Bagshaw, A.P., Bénar, C.G., Hawco, C., Andermann, F., Dubeau, F., Gotman, J., 2004. fMRI activation during spike and wave discharges in idiopathic generalized epilepsy. Brain 127, 1127—1144. Ahrens, H., Argin, F., Klinkenbusch, L., 2012. A comparison of minimum norm and MUSIC for a combined MEG/EEG sensor array. Adv. Radio Sci. 10, 99—104. Akaike, H., 1974. A new look at the statistical model identification. IEEE Trans. Automat. Contr. 19, 716—723.
C. Lee et al. Amor, F., Baillet, S., Navarro, V., Adam, C., Martinerie, J., Quyen, M.L. Van, 2009. Cortical local and long-range synchronization interplay in human absence seizure initiation. Neuroimage 45, 950—962. Anderson, J., Hamandi, K., 2011. Understanding juvenile myoclonic epilepsy: contributions from neuroimaging. Epilepsy Res. 53, 779—789. Archer, J., Abbott, D., Waites, A., Jackson, G., 2003. fMRI deactivation of the posterior cingulate during generalized spike and wave. Neuroimage 20, 1915—1922. Avoli, M., 2012. A brief history on the oscillating roles of thalamus and cortex in absence seizures. Epilepsia 53, 779—789. Bai, X., Vestal, M., Berman, R., Negishi, M., Spann, M., Vega, C., Desalvo, M., Novotny, E.J., Constable, R.T., Blumenfeld, H., 2010. Dynamic time course of typical childhood absence seizures: EEG, behavior, and functional magnetic resonance imaging. J. Neurosci. 30, 5884—5893. Benuzzi, F., Mirandola, L., Pugnaghi, M., Farinelli, V., Tassinari, C.A., Capovilla, G., Cantalupo, G., Beccaria, F., Nichelli, P., Meletti, S., 2012. Increased cortical BOLD signal anticipates generalized spike and wave discharges in adolescents and adults with idiopathic generalized epilepsies. Epilepsia 53, 622—630. Bernhardt, B.C., Rozen, D.A., Worsley, K.J., Evans, A.C., Bernasconi, N., Bernasconi, A., 2009. Thalamo-cortical network pathology in idiopathic generalized epilepsy: insights from MRI-based morphometric correlation analysis. Neuroimage 46, 373—381. Blinowska, K., Kus, R., Kaminski, M., Janiszewska, J., 2010. Transmission of brain activity during cognitive task. Brain Topogr. 23, 205—213. Blumenfeld, H., 2005. Cellular and network mechanisms of spikewave seizures. Epilepsia 46, 21—33. Blumenfeld, H., 2012. Impaired consciousness in epilepsy. Lancet Neurol. 11, 814—826. Cavanna, A.E., Trimble, M.R., 2006. The precuneus: a review of its functional anatomy and behavioural correlates. Brain 129, 564—583. Chatfield, C., 2004. The Analysis of Time Series: An Introduction, Chapman Hall CRC Texts in Statistical Science. Chapman & Hall/CRC. Commission of ILAE, 1989. Proposal for revised classification of epilepsies and epileptic syndromes. Epilepsia 30, 389—399. Cryer, J.D., Chan, K.-S., 2008. Time Series Analysis. Springer, New York, NY. Ding, L., Worrell, G.A., Lagerlund, T.D., He, B., 2007. Ictal source analysis: localization and imaging of causal interactions in humans. Neuroimage 34, 575—586. Ding, M., Bressler, S.L., Yang, W., Liang, H., 2000. Short-window spectral analysis of cortical event-related potentials by adaptive multivariate autoregressive modeling: data preprocessing, model validation, and variability assessment. Biol. Cybern. 83, 35—45. Franaszczuk, P.J., Bergey, G.K., 1998. Application of the directed transfer function method to mesial and lateral onset temporal lobe seizures. Brain Topogr. 11, 13—21. Goldman-Rakic, P.S., 1988. Topography of cognition: parallel distributed networks in primate association cortex. Annu. Rev. Neurosci. 11, 137—156. Gotman, J., 2008. Epileptic networks studied with EEG-fMRI. Epilepsia 49, 42—51. Gotman, J., Grova, C., Bagshaw, A., Kobayashi, E., Aghakhani, Y., Dubeau, F., 2005. Generalized epileptic discharges show thalamocortical activation and suspension of the default state of the brain. Proc. Natl. Acad. Sci. U. S. A. 102, 15236—15240. Gramfort, A., Papadopoulo, T., Olivi, E., Clerc, M., 2010. OpenMEEG: opensource software for quasistatic bioelectromagnetics. Biomed. Eng. Online 9, 45.
Effective connectivity in SWD Gupta, D., Ossenblok, P., van Luijtelaar, G., 2011. Space-time network connectivity and cortical activations preceding spike wave discharges in human absence epilepsy: a MEG study. Med. Biol. Eng. Comput. 49, 555—565. Gusnard, D.A., Raichle, M.E., 2001. Searching for a baseline: functional imaging and the resting human brain. Nat. Rev. Neurosci. 2, 685—694. Hamandi, K., Salek-Haddadi, A., Laufs, H., Liston, A., Friston, K., Fish, D.R., Duncan, J.S., Lemieux, L., 2006. EEG-fMRI of idiopathic and secondarily generalized epilepsies. Neuroimage 31, 1700—1710. Holmes, M.D., Quiring, J., Tucker, D.M., 2010. Evidence that juvenile myoclonic epilepsy is a disorder of frontotemporal corticothalamic networks. Neuroimage 49, 80—93. Janz, D., 1985. Epilepsy with impulsive petit mal (juvenile myoclonic epilepsy). Acta Neurol. Scand. 72, 449—459. Janz, D., Christian, W., 1957. Impulsiv-Petit mal. Dtsch. Zeitschrift f. Nervenheilkd. 176, 346—386. Jung, Y.-J., Kang, H.-C., Choi, K.-O., Lee, J.S., Kim, D.-S., Cho, J.-H., Kim, S.-H., Im, C.-H., Kim, H.D., 2011. Localization of ictal onset zones in Lennox—Gastaut syndrome using directional connectivity analysis of intracranial electroencephalography. Seizure 20, 449—457. Kami´ nski, M.J., Blinowska, K.J., 1991. A new method of the description of the information flow in the brain structures. Biol. Cybern. 65, 203—210. Killory, B.D., Bai, X., Negishi, M., Vega, C., Spann, M.N., Vestal, M., Guo, J., Berman, R., Danielson, N., Trejo, J., Shisler, D., Novotny, E.J., Constable, R.T., Blumenfeld, H., 2011. Impaired attention and network connectivity in childhood absence epilepsy. Neuroimage 56, 2209—2217. Kim, J.H., Suh, S.-I., Park, S.-Y., Seo, W.-K., Koh, I., Koh, S.-B., Seol, H.Y., 2012. Microstructural white matter abnormality and frontal cognitive dysfunctions in juvenile myoclonic epilepsy. Epilepsia 53, 1371—1378. Kramer, M.A., Cash, S.S., 2012. Epilepsy as a disorder of cortical network organization. Neuroscientist 18, 360—372. Labate, A., Briellmann, R.S., Abbott, D.F., Waites, A.B., Jackson, G.D., 2005. Typical childhood absence seizures are associated with thalamic activation. Epileptic Disord. 7, 373—377. Laufs, H., Lengler, U., Hamandi, K., Kleinschmidt, A., Krakow, K., 2006. Linking generalized spike-and-wave discharges and resting state brain activity by using EEG/fMRI in a patient with absence seizures. Epilepsia 47, 444—448. Leichnetz, G.R., 2001. Connections of the medial posterior parietal cortex (area 7 m) in the monkey. Anat. Rec. 263, 215—236. Lu, Y., Yang, L., Worrell, G.A., He, B., 2012. Seizure source imaging by means of FINE spatio-temporal dipole localization and directed transfer function in partial epilepsy patients. Clin. Neurophysiol. 123, 1275—1283. Luo, C., Li, Q., Lai, Y., Xia, Y., Qin, Y., Liao, W., Li, S., Zhou, D., Yao, D., Gong, Q., 2011. Altered functional connectivity in default mode network in absence epilepsy: a resting-state fMRI study. Hum. Brain Mapp. 32, 438—449. McGill, M.L., Devinsky, O., Kelly, C., Milham, M., Castellanos, F.X., Quinn, B.T., DuBois, J., Young, J.R., Carlson, C., French, J., Kuzniecky, R., Halgren, E., Thesen, T., 2012. Default mode network abnormalities in idiopathic generalized epilepsy. Epilepsy Behav. 23, 353—359. Meeren, H., Pijn, J., Van Luijtelaar, E., Coenen, A., Lopes da Silva, F., 2002. Cortical focus drives widespread corticothalamic networks during spontaneous absence seizures in rats. J. Neurosci. 22, 1480—1495. Meeren, H., van Luijtelaar, G., Lopes da Silva, F., Coenen, A., 2005. Evolving concepts on the pathophysiology of absence seizures: the cortical focus theory. Arch. Neurol. 62, 371—376.
265 Moeller, F., Maneshi, M., Pittau, F., Gholipour, T., Bellec, P., Dubeau, F., Grova, C., Gotman, J., 2011. Functional connectivity in patients with idiopathic generalized epilepsy. Epilepsia 52, 515—522. Moeller, F., Siebner, H.R., Wolff, S., Muhle, H., Boor, R., Granert, O., Jansen, O., Stephani, U., Siniatchkin, M., 2008a. Changes in activity of striato-thalamo-cortical network precede generalized spike wave discharges. Neuroimage 39, 1839—1849. Moeller, F., Siebner, H.R., Wolff, S., Muhle, H., Granert, O., Jansen, O., Stephani, U., Siniatchkin, M., 2008b. Simultaneous EEG-fMRI in drug-naive children with newly diagnosed absence epilepsy. Epilepsia 49, 1510—1519. Nordli, D.R., 2005. Idiopathic generalized epilepsies recognized by the International League Against Epilepsy. Epilepsia 46 (Suppl. 9), 48—56. O’Muircheartaigh, J., Vollmar, C., Barker, G.J., Kumari, V., Symms, M.R., Thompson, P., Duncan, J.S., Koepp, M.J., Richardson, M.P., 2011. Focal structural changes and cognitive dysfunction in juvenile myoclonic epilepsy. Neurology 76, 34—40. Pascual-Marqui, R.D., 2002. Standardized low-resolution brain electromagnetic tomography (sLORETA): technical details. Methods Find. Exp. Clin. Pharmacol. 24, 5—12. Petrides, M., Pandya, D.N., 1984. Projections to the frontal cortex from the posterior parietal region in the rhesus monkey. J. Comp. Neurol. 228, 105—116. Philiastides, M.G., Sajda, P., 2006. Causal influences in the human brain during face discrimination: a short-window directed transfer function approach. IEEE Trans. Biomed. Eng. 53, 2602—2605. Ronan, L., Alhusaini, S., Scanlon, C., Doherty, C.P., Delanty, N., Fitzsimons, M., 2012. Widespread cortical morphologic changes in juvenile myoclonic epilepsy: evidence from structural MRI. Epilepsia 53, 651—658. Santiago-Rodríguez, E., Harmony, T., Fernández-Bouzas, A., Hernández-Balderas, A., Martínez-López, M., Graef, A., Carlos García, J., Fernández, T., 2002. Source analysis of polyspike and wave complexes in juvenile myoclonic epilepsy. Seizure 11, 320—324. Stefan, H., Paulini-Ruf, A., Hopfengärtner, R., Rampp, S., 2009. Network characteristics of idiopathic generalized epilepsies in combined MEG/EEG. Epilepsy Res. 85, 187—198. Sullivan, J.E., Dlugos, D.J., 2004. Idiopathic generalized epilepsy. Curr. Treat. Options Neurol. 6, 231—242. Tadel, F., Baillet, S., Mosher, J.C., Pantazis, D., Leahy, R.M., 2011. Brainstorm: a user-friendly application for MEG/EEG analysis. Comput. Intell. Neurosci., 879716. Vaudano, A.E., Laufs, H., Kiebel, S.J., Carmichael, D.W., Hamandi, K., Guye, M., Thornton, R., Rodionov, R., Friston, K.J., Duncan, J.S., Lemieux, L., 2009. Causal hierarchy within the thalamocortical network in spike and wave discharges. PLoS ONE 4, e6475. Vollmar, C., O’Muircheartaigh, J., Barker, G.J., Symms, M.R., Thompson, P., Kumari, V., Duncan, J.S., Janz, D., Richardson, M.P., Koepp, M.J., 2011. Motor system hyperconnectivity in juvenile myoclonic epilepsy: a cognitive functional magnetic resonance imaging study. Brain 134, 1710—1719. Westmijse, I., Ossenblok, P., Gunning, B., van Luijtelaar, G., 2009. Onset and propagation of spike and slow wave discharges in human absence epilepsy: a MEG study. Epilepsia 50, 2538—2548. Wilke, C., Worrell, G., He, B., 2011. Graph analysis of epileptogenic networks in human partial epilepsy. Epilepsia 52, 84—93. Woermann, F.G., Free, S.L., Koepp, M.J., Sisodiya, S.M., Duncan, J.S., 1999. Abnormal cerebral structure in juvenile myoclonic epilepsy demonstrated with voxel-based analysis of MRI. Brain 122, 2101—2108.
266 Yang, T., Luo, C., Li, Q., Guo, Z., Liu, L., Gong, Q., Yao, D., Zhou, D., 2013. Altered resting-state connectivity during interictal generalized spike-wave discharges in drug-naïve childhood absence epilepsy. Hum. Brain Mapp. 34, 1761—1767. Yeterian, E.H., Pandya, D.N., 1985. Corticothalamic connections of the posterior parietal cortex in the rhesus monkey. J. Comp. Neurol. 237, 408—426.
C. Lee et al. Zhang, Z., Liao, W., Chen, H., Mantini, D., Ding, J.-R., Xu, Q., Wang, Z., Yuan, C., Chen, G., Jiao, Q., Lu, G., 2011. Altered functional—structural coupling of large-scale brain networks in idiopathic generalized epilepsy. Brain 134, 2912—2928.
journal homepage: www.elsevier.com/locate/epilepsyres
Causal influence of epileptic network during spike-and-wave discharge in juvenile myoclonic epilepsy Chany Lee a, Sung-Min Kim a, Young-Jin Jung b, Chang-Hwan Im b, Dong Wook Kim c, Ki-Young Jung a,∗ a
Department of Neurology, College of Medicine, Korea University, Seoul, Republic of Korea Department of Biomedical Engineering, Hanyang University, Seoul, Republic of Korea c Department of Neurology, College of Medicine, Konkuk University, Seoul, Republic of Korea b
Received 24 May 2013; received in revised form 7 October 2013; accepted 3 November 2013 Available online 16 November 2013
KEYWORDS Juvenile myoclonic epilepsy; Spike-and-wave discharge; Network; Effective connectivity; Precuneus
Summary Electroencephalographic (EEG) characteristic of juvenile myoclonic epilepsy (JME) is spike-and-wave discharge (SWD), which is dominant in the frontal region. However, activity in the parietal area, including the precuneus, has also been documented for several seconds before and during SWD. The aim of this study was to identify the role of the parietal region, especially the precuneus, and to clarify the causal dynamics among cortical regions during SWD. EEGs were obtained from seven patients with JME. Each SWD was divided into six distinct temporal phases: spike onset, spike peak, slow-wave onset, slow-wave ascending, slow-wave peak, and slow-wave descending phases. Based on the cortical current source distribution and the results of a previous study, we selected the medial frontal, orbitofrontal, anterior cingulate, and mesial temporal cortices and the precuneus as regions of interest (ROIs). To assess epileptic networks and the causal relationships among ROIs during SWD, the directed transfer function (DTF), a measure of multivariate causality, was calculated for each phase of SWD. During spike onset, the maximal outdegree region in all patients was the precuneus. The spike-peak and slow-wave onset phases did not show a consistently dominant outflow region. Outflow from the anterior cingulate cortex was dominant in four patients during the slow-wave ascending phase, and the precuneus showed the maximal outdegree in six patients during the slow-wave peak. In the slow-wave descending phase, four patients showed maximal outflow from the temporal
∗ Corresponding author at: Department of Neurology, Korea University Medical Center, Korea University College of Medicine, #126-1, Anam-Dong 5Ga, Seongbuk-Gu, Seoul 136-705, Republic of Korea. Tel.: +82 2 920 6649; fax: +82 2 925 2472. E-mail address: [email protected] (K.-Y. Jung).
0920-1211/$ — see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.eplepsyres.2013.11.005
258
C. Lee et al. cortex. Our findings suggest that the precuneus is likely a key region for SWD despite the small amount of neural activity observed. The precuneus was the region with the maximal outdegree during both the spike onset and slow-wave peak phases, indicating that SWD in JME is initiated and sustained by a network involving the frontal cortex, precuneus, and thalamus. © 2013 Elsevier B.V. All rights reserved.
Introduction Juvenile myoclonic epilepsy (JME) is the most common idiopathic generalized epilepsy (IGE) syndrome, accounting for 5—10% of all cases of epilepsy (Janz and Christian, 1957; Janz, 1985; Vollmar et al., 2011). The characteristic features of JME are myoclonic jerks on awakening, generalized tonic—clonic seizures, and less frequent absence seizures. In terms of electroencephalography (EEG), the hallmark of JME consists of bilateral, synchronous, widespread spike or polyspike and wave complexes, which are commonly assumed to occur without lateralizing or localizing features (Nordli, 2005). Combined EEG and functional magnetic resonance imaging (fMRI) studies in patients with IGE, including JME, have shown a clear spike-and-wave discharge (SWD)-associated increase in the blood oxygenation level-dependent (BOLD) signal in the thalamus and a decrease in the BOLD signal in the parietal and frontal cortical regions, which is consistent with the default mode network (DMN) of the human brain (Archer et al., 2003; Gotman et al., 2005; Hamandi et al., 2006; Labate et al., 2005). Time-course analyses of the BOLD signal during SWD have shown BOLD increments starting approximately 5—10 s before the onset of SWD in the precuneus (PCN)/posterior cingulate cortex (PCC) and medial/lateral parietal cortex (Benuzzi et al., 2012) as well as in the orbital/medial frontal cortex (Bai et al., 2010). Additionally, a recent magnetoencephalograpihc (MEG) study showed that low-frequency sources were sequentially activated in the frontal and occipital regions prior to the first generalized spikes (Gupta et al., 2011). An fMRI study using dynamic causal modeling suggested that activity in the PCN gates SWD in IGE in the thalamocortical network (Vaudano et al., 2009). The posteromedial cortical region, including the PCN and PCC, showed the highest degree of interactions with the rest of the DMN, suggesting that this area plays a pivotal role in the network. This finding indicates that the posterior cortex may influence SWD generation such that it drives the SWD network under pathological circumstances. Although EEG-fMRI studies have been used for the localization of SWD in IGE (Aghakhani et al., 2004; Gotman et al., 2005; Moeller et al., 2008a), fMRI reflects the BOLD signal, which is indirectly related to the underlying neuronal activity and cannot describe instantaneous neural activities of SWD because the BOLD signal is an accumulation of signals over several seconds (Archer et al., 2003; Blumenfeld, 2005; Sullivan and Dlugos, 2004). Additionally, most EEG-fMRI studies oversimplify the hemodynamic response function (HRF) related to brain activity, as most activity cannot be measured by the standard HRF used for conventional fMRI analyses of SWD in IGE (Blumenfeld, 2012).
Recently, as epilepsy has been considered network disorder, epileptic brain networks have been evaluated using mathematical measures of connectivity (Amor et al., 2009; Gupta et al., 2011; Killory et al., 2011; Kramer and Cash, 2012; Zhang et al., 2011). In our study, we adopted a directed transfer function (DTF), which is a type of multivariate application of Granger causality (Kami´ nski and Blinowska, 1991), since DTF can efficiently estimate causal interactions among multiple EEG signals. DTFs have been extensively used in the analysis of epileptic networks (Ding et al., 2007; Franaszczuk and Bergey, 1998; Lu et al., 2012; Wilke et al., 2011). A series of studies have demonstrated that the DTF technique can be used to identify ictal onset zones from intracranial EEG recordings in cases of mesial and lateral temporal lobe epilepsies as well as in cases of neocortical extra-temporal lobe epilepsies. In addition, it was recently reported that seizure activity in patients with symptomatic generalized epilepsy, such as Lennox—Gastaut syndrome, can be localized using DTF analysis (Jung et al., 2011). In the present study, to assess epilepsy networks and the causal relationship among cortical regions during SWD in patients with JME, we first identified cortical current sources and applied DTF on these current sources during SWD. We also tried to elucidate the role of the PCN in gating SWD during the ictogenesis of JME.
Methods Subjects Seven patients with a clinical diagnosis of JME were included in this study. The diagnosis of JME was based on electroclinical criteria as per the International League Against Epilepsy (ILAE) classification (Commission of ILAE, 1989). Inclusion criteria were as follows: (1) a typical clinical history of JME with onset of myoclonic jerks and generalized seizures in adolescence; (2) no evidence of neurological abnormality and intellectual decline; (3) apparent spike or polyspike and wave discharges on normal background rhythm on previous standard international 10—20 EEG; and (4) normal MRI finding by visual inspection. The study protocol was approved by the Institutional Review Board of Korea University Medical Center. Clinical data of the included patients are summarized in Table 1.
EEG recording A 64-channel EEG signal and two electrooculogram channels were recorded at a sampling rate of 1600 samples/s using an EEG recording system (Grass Technologies, Quincy, MA, USA). An electrode cap with sintered Ag/AgCl electrodes was
Effective connectivity in SWD Table 1
259
Clinical characteristics of patients.
Patient #
Sex
Age (year)
Onset (year)
EEG
1 2 3 4 5 6 7
M M F M M M F
42 19 24 19 24 32 17
13 16 22 16 18 30 12
4—5 Hz 4—5 Hz 3—4 Hz 3—4 Hz 2—3 Hz 3—4 Hz 3—4 Hz
25.3 8.9
18.1 6.2
Mean SD
AEDs polyspike and wave SWD SWD SWD SWD SWD SWD
LMT, VPA LEV LEV VPA LEV ZNS VPA, LEV
AED, antiepileptic drug; LTG, lamotrigine; VPA, valproate; LEV, levetriracetam; ZNS, zonisamide; SD, standard deviation.
used (Quick-Cap, Compumedics Neuroscan, Charlotte, NC, USA). The impedances of all electrodes were reduced below 10 k. The reference electrode was set to linked mastoid electrodes, and the bandpass filter was set at 0.3—70 Hz. The patient was seated comfortably in a chair, and EEG recordings were taken during waking, sleep, and hyperventilation for approximately 1 h.
Epoching of SWD Whole EEG recordings were inspected visually, and typical SWDs were identified (KJ). All patients showed bilateral widespread SWD, with runs at frequencies of either 3—4 Hz or 4—6 Hz (Fig. 1). The onset of SWD was defined as the first large transition in SWD signals (Holmes et al., 2010), and
SWDs were extracted from −100 ms to 900 ms, where 0 ms was the spike onset time. The preictal state was defined as the period from -100 ms to 0 ms. As the early phase of SWD may have a more consistent network than do the later phases of the seizure (Meeren et al., 2002), only the first SWD of the ictal discharges was considered for analysis. In order to assess the time-varying causality of SWD, a single SWD was divided into six phases: spike onset, spike peak, slow-wave onset, slow-wave ascending, slow-wave peak, and slow-wave descending phases (Fig. 2).
Current density distribution of SWD Forward problem was solved using Brainstorm (http://neuroimage.usc.edu/brainstorm) with default
Figure 1 An example of spike-and-wave discharge (SWD) on an EEG recording. The vertical dashed line indicates the onset of SWD. The reference was set to linked mastoid electrodes, and negativity is directed upward. Electrodes in the figure were selected according to the international 10—20 system.
260
C. Lee et al.
Figure 2 Overlay map of current density distribution from 98 epochs in all patients. After calculating the current density for each epoch, the threshold was set at 30% of the maximal current density, and binary results were acquired and accumulated. Thus, the distribution in this figure indicates the frequency of activation from all epochs instead of the actual current density.
anatomy data (Tadel et al., 2011). In Brainstorm, the default numerical models of the cortex, scalp, and the inner and outer skull boundaries included 15,028, 1028, 642, and 642 vertices and 29,899, 2160, 1280, and 1280 triangles, respectively. The electrical conductivities of the scalp, skull, and cerebrospinal fluid/brain were set as default values of 1, 0.0125, and 1 S/m, respectively (Ahrens et al., 2012). The boundary element method by OpenMEEG embedded in Brainstorm was used (Gramfort et al., 2010). For the inverse algorithm, standardized low-resolution brain electromagnetic tomography (sLORETA) was adopted (Pascual-Marqui, 2002). An inverse operator of the leadfield matrix calculated by Brainstorm was extracted and applied to each time slice of each epoch so that the current density distribution was obtained for any given time point during each phase.
Selection of ROIs To locate the frequently activated regions during each phase of SWD across epochs in each patient, current densities greater than 30% of maximal value (approximately at a signal-to-noise ratio (SNR) of 5 dB) remained as binary distributions and were summed over all epochs. The resultant distribution of cortical activation refers to overlay map. The common distribution of current densities across patients was then obtained for each phase of SWD using the same method (Fig. 2). Based on the current source distribution, the orbitofrontal cortex (OFC), medial frontal cortex (MFC),
mesial temporal cortex (MTC), and anterior cingulate cortex (ACC) were selected as regions of interest (ROIs). Despite its relative lack of activation, we also included the PCN as a ROI, as previous studies have suggested that it may play a key role in SWD (Benuzzi et al., 2012; Vaudano et al., 2009). The areas and locations of the ROIs in MNI coordinates are presented in Table 2. In this study, we considered only the left hemisphere because SWD has bilateral characteristics in JME (Meeren et al., 2002; Sullivan and Dlugos, 2004), although it is likely that some asymmetric distribution also occurs. The current source density was averaged over each ROI, so each ROI had its own representative current source density.
Directed transfer function analysis DTF is a measure of the information flow through multichannel signals (Blinowska et al., 2010; Ding et al., 2007). In order to obtain information flow during each phase of SWD, the window length was set to 50 data points (31.25 ms). Because multivariate autoregressive modeling (MVAR) can be applied only to stationary data (Chatfield, 2004; Cryer and Chan, 2008), we tested the stationarity of windowed data by dividing them into four temporal segments. Each segment had 20 time slices with 50% overlap. After calculating the autocorrelation function for each segment, we found that autocorrelation was a function of time-lag only (Fig. S1), which means that the representative current sources of the five ROIs were stationary in a window and that MVAR could be applied.
Effective connectivity in SWD Table 2
261
Locations and areas of selected ROIs.
ROIs
Center location in MNI coordinates (mm)
MFC ACC OFC MTC PCN
x
y
0 −2 −6 −18 −1
21 16 62 −2 −73
Brodmann area
Area (mm2 )
6, 8 24 11 20, 28, 38 7
315 167 390 400 130
z 53 27 −10 −29 50
MFC, medial frontal cortex; ACC, anterior cingulate cortex; OFC, orbitofrontal cortex; MTC, mesial temporal cortex; PCN, precuneus.
Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.eplepsyres.2013.11.005. The MVAR order was determined by the Akaike information criterion (AIC) (Akaike, 1974; Ding et al., 2000; Philiastides and Sajda, 2006). The model order of MVAR was determined to be 7 because AIC values did not significantly change above order 7 (Fig. S2), and similar results were obtained in all windows. Hence, a model order of 7 was applied to all windows. Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/ 10.1016/j.eplepsyres.2013.11.005. For each window, 500 surrogate data were employed and p-values < 0.01 were considered to be statistically significant. To construct a binary adjacency matrix, causal influence from region j to region i was 1 if DTF had statistical meaning or 0 if it did not. In a given patient, an individual adjacency matrix was obtained after adjacency matrices for each phase across epochs were averaged (Fig. 3). Finally, a connectivity matrix for each phase was generated by
Figure 3
averaging all individual adjacency matrices for each phase across patients.
Results A total of 98 epochs (mean, 14.0 ± 9.6) of SWD were selected from seven patients. The average duration and standard deviation of the first SWD were 425.6 ms and 112.3 ms, respectively. An example of SWD in patient no. 3 is shown in Fig. 1.
Current source distribution In each patient, the distribution of current density at each phase of SWD was usually maximal over the medial frontal regions, including the ACC and OFC, and the anteromesial temporal regions. Although the distribution did not change significantly across phases, the magnitude of the current density did differ significantly between phases.
Schematic diagram of directed transfer function (DTF) analysis with a short window for a single patient.
262 The overlay map for each phase across patients is demonstrated in Fig. 2. Similar to the results of the individual analysis, the most commonly involved areas were the frontotemporal cortical regions, including the medial frontal, anterior cingulate, and temporal cortices. Of them, the ACC was the most consistently activated region (61.2%, 60 of 98 epochs) during the spike period (spike onset and spike peak). The anteromesial temporal region (59.2%, 59 of 98 epochs) was the next most frequently activated area during these phases. During slow wave phases, the medial frontal and ACC were most commonly involved, showing activation in 90 of 98 epochs.
C. Lee et al.
Causal influence among ROIs The DTF was estimated for all phases. For an instance, the DTF results of patient no. 6 are shown in Fig. 4. Fig. 5 shows the mean outdegree across epochs and patients for each phase of SWD. Despite intraindividual variability, noticeable outflow characteristics were observed. The DTF during spike onset showed that the PCN exhibited maximal outflow of information for all patients (63 epochs), which is the most consistent finding across individuals. Spike peak and slowwave onset phases did not have a consistently dominant outflow region. Outflow from the ACC was dominant in four patients (53 epochs) during slow-wave ascending phase, and
Figure 4 DTF results for patient no. 6. (A) Spike onset, (B) spike peak, (C) slow-wave onset, (D) slow-wave ascending, (E) slowwave peak, and (F) slow-wave descending phases. The left panel of each phase indicates information flow (from tail to head) during the representative epoch. The middle panel is an individual adjacency matrix. M, A, O, T, and P refer to the medial frontal cortex, anterior cingulate cortex, orbitofrontal cortex, mesial temporal cortex, and precuneus, respectively. The color represents the mean outdegree of each region. The right panel shows the indegree and outdegree of each region. Inflow and outflow are the vertical and horizontal summations of the adjacency matrix, respectively.
Effective connectivity in SWD
263
Figure 5 Mean outdegree across epochs and patients for each phase of spike wave discharge. The outdegree of the precuneus (PCN) is slightly over 2 in spike onset phase, which means that at least two regions receive information from the PCN. The medial frontal cortex and the anterior cingulate cortex in slow-wave ascending phase, the PCN in slow-wave peak phase, and the mesial temporal cortex in slow-wave descending phase are dominant. In the preictal state, the outdegree of all ROIs did not exceed 1.
the PCN was the maximal outdegree region during slow-wave peak in six patients (68 epochs). In slow-wave descending phase, four patients (51 epochs) showed outflow from the MTC that was superior to that of other regions. No noticeable outdegree was noted just before SWD onset or during the preictal baseline period compared with other phases.
Discussion In the present study, we investigated the epilepsy network of SWD and the causal influence among cortical regions that account for its network in patients with JME. Source localization revealed that the frontotemporal network, including the medial frontal gyri, anterior cingulate, orbitofrontal area, and anteromesial temporal region, were primarily involved in the generation of SWD in JME. Frontal cortical involvement during SWD has been consistently reported in various neuroimaging studies (Benuzzi et al., 2012; Kim et al., 2012; Santiago-Rodríguez et al., 2002; Stefan et al., 2009; Westmijse et al., 2009; Woermann et al., 1999). This frontal involvement appears to relate thalamo-cortical network, which is an important feature of SWD in IGE (Anderson and Hamandi, 2011; Bernhardt et al., 2009). In order to investigate functional connectivity and causal influence within the JME network, we selected four ROIs in the frontotemporal regions in which the current source distribution was most active during SWD. Additionally, although precuneal activity exceeding the threshold was observed in only one patient, the PCN was selected as an additional ROI based on the following recent EEG-fMRI findings; (1) SWD-related BOLD signal changes showed both activation and deactivation not only in the thalamus and frontal regions but also in the posterior cortical regions (Gotman, 2008; Hamandi et al., 2006); (2) activation in the PCN/posterior cingulate region precedes that in the thalamo-frontal regions (Moeller et al., 2008a); and (3) a dynamic causal model of BOLD during SWD suggested that
the PCN may act on the thalamic-frontal cortical network in facilitating the development of SWD (Vaudano et al., 2009). In our study, the outflow of information during spike onset was most dominant in the PCN in all patients, whereas the current source was maximal over the frontotemporal regions. In contrast, no significant outflow of information was observed in the medial frontal regions, which was the predominant activated region during spike peak. This finding suggests that the PCN may provide information to the medial frontal regions for the initiation of SWD in the thalamo-frontotemporal network, which is conceptually consistent with the previous EEG-fMRI study (Vaudano et al., 2009). This information flow may be supported by direct anatomical connections between the PCN and the frontal lobe (Goldman-Rakic, 1988; Leichnetz, 2001; Petrides and Pandya, 1984; Vaudano et al., 2009). In addition, there are subcortical connections between the PCN and the thalamus (Cavanna and Trimble, 2006; Yeterian and Pandya, 1985). Although recent research has proposed that a focal cortical region was found to lead the thalamus (Meeren et al., 2002), the thalamus is still thought to play an essential role in the thalamo-cortical network in IGE (Avoli, 2012; Meeren et al., 2005), and increments in the BOLD signal in the thalamus during SWD have been consistently reported (Aghakhani et al., 2004; Archer et al., 2003; Benuzzi et al., 2012; Gotman et al., 2005; Hamandi et al., 2006). Therefore, it can be speculated that the PCN may influence medial frontal regions via a direct white matter pathway and/or thalamocortical projections. Furthermore, the outdegree of the PCN was dominant during slow-wave peak phase as well as and spike onset phases. This finding implies that the PCN acts not only as an initiator but also as a mediator of SWD. Taken together, our data suggest that SWD is initiated and sustained by a frontal-precuneal-thalamic network in JME. Previous EEG-fMRI studies demonstrated that SWD was strongly related to the DMN (Laufs et al., 2006; McGill et al., 2012; Moeller et al., 2011, 2008b). Both activation and deactivation of the BOLD signal within the DMN were
264 observed during SWD (Laufs et al., 2006). It has also been reported that the structural and functional connectivity of the DMN is changed in IGE (Bernhardt et al., 2009; Luo et al., 2011; O’Muircheartaigh et al., 2011; Yang et al., 2013). The changes in the neuronal state within the PCN may reflect spontaneous fluctuations in awareness and may affect the thalamo-frontal network of SWD (Vaudano et al., 2009). The PCN is one of the key regions in the DMN, as it showed the highest resting metabolic rates, indicating that a high level of information processing was occurring in this region during the resting state of the brain (Gusnard and Raichle, 2001). Structural abnormalities, such as changes in gray matter volume, have also been found in the DMN regions of patients with JME (O’Muircheartaigh et al., 2011; Ronan et al., 2012; Woermann et al., 1999). As shown in Fig. 5, the preictal phase which was defined as the 100-ms period just before the onset of SWD did not show any significant information flow among the ROIs. In other words, we observed no noticeable interaction among ROIs during the preictal period. In contrast, EEG-fMRI studies have reported that the parietal region was involved several seconds before SWD onset (Benuzzi et al., 2012; Moeller et al., 2008a). Hence, it could be interpreted that parietal region does not interact with other regions before SWD onset and starts sending information to other regions at SWD onset.
Conclusion In conclusion, our study supports the notion that the PCN may play an important role for onset of SWD in the thalamo-frontal cortical network. Moreover, the PCN is also a mediator of SWD so that epileptic network is sustained during SWD.
Conflict of interest None of the authors has any conflict of interest to disclose.
Acknowledgments This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2012R1A6A3A01019371) and by a grant from the Korea Healthecare Technology R&D Project, Ministry for Health, Welfare & Family Affairs, Republic of Korea (A090794).
References Aghakhani, Y., Bagshaw, A.P., Bénar, C.G., Hawco, C., Andermann, F., Dubeau, F., Gotman, J., 2004. fMRI activation during spike and wave discharges in idiopathic generalized epilepsy. Brain 127, 1127—1144. Ahrens, H., Argin, F., Klinkenbusch, L., 2012. A comparison of minimum norm and MUSIC for a combined MEG/EEG sensor array. Adv. Radio Sci. 10, 99—104. Akaike, H., 1974. A new look at the statistical model identification. IEEE Trans. Automat. Contr. 19, 716—723.
C. Lee et al. Amor, F., Baillet, S., Navarro, V., Adam, C., Martinerie, J., Quyen, M.L. Van, 2009. Cortical local and long-range synchronization interplay in human absence seizure initiation. Neuroimage 45, 950—962. Anderson, J., Hamandi, K., 2011. Understanding juvenile myoclonic epilepsy: contributions from neuroimaging. Epilepsy Res. 53, 779—789. Archer, J., Abbott, D., Waites, A., Jackson, G., 2003. fMRI deactivation of the posterior cingulate during generalized spike and wave. Neuroimage 20, 1915—1922. Avoli, M., 2012. A brief history on the oscillating roles of thalamus and cortex in absence seizures. Epilepsia 53, 779—789. Bai, X., Vestal, M., Berman, R., Negishi, M., Spann, M., Vega, C., Desalvo, M., Novotny, E.J., Constable, R.T., Blumenfeld, H., 2010. Dynamic time course of typical childhood absence seizures: EEG, behavior, and functional magnetic resonance imaging. J. Neurosci. 30, 5884—5893. Benuzzi, F., Mirandola, L., Pugnaghi, M., Farinelli, V., Tassinari, C.A., Capovilla, G., Cantalupo, G., Beccaria, F., Nichelli, P., Meletti, S., 2012. Increased cortical BOLD signal anticipates generalized spike and wave discharges in adolescents and adults with idiopathic generalized epilepsies. Epilepsia 53, 622—630. Bernhardt, B.C., Rozen, D.A., Worsley, K.J., Evans, A.C., Bernasconi, N., Bernasconi, A., 2009. Thalamo-cortical network pathology in idiopathic generalized epilepsy: insights from MRI-based morphometric correlation analysis. Neuroimage 46, 373—381. Blinowska, K., Kus, R., Kaminski, M., Janiszewska, J., 2010. Transmission of brain activity during cognitive task. Brain Topogr. 23, 205—213. Blumenfeld, H., 2005. Cellular and network mechanisms of spikewave seizures. Epilepsia 46, 21—33. Blumenfeld, H., 2012. Impaired consciousness in epilepsy. Lancet Neurol. 11, 814—826. Cavanna, A.E., Trimble, M.R., 2006. The precuneus: a review of its functional anatomy and behavioural correlates. Brain 129, 564—583. Chatfield, C., 2004. The Analysis of Time Series: An Introduction, Chapman Hall CRC Texts in Statistical Science. Chapman & Hall/CRC. Commission of ILAE, 1989. Proposal for revised classification of epilepsies and epileptic syndromes. Epilepsia 30, 389—399. Cryer, J.D., Chan, K.-S., 2008. Time Series Analysis. Springer, New York, NY. Ding, L., Worrell, G.A., Lagerlund, T.D., He, B., 2007. Ictal source analysis: localization and imaging of causal interactions in humans. Neuroimage 34, 575—586. Ding, M., Bressler, S.L., Yang, W., Liang, H., 2000. Short-window spectral analysis of cortical event-related potentials by adaptive multivariate autoregressive modeling: data preprocessing, model validation, and variability assessment. Biol. Cybern. 83, 35—45. Franaszczuk, P.J., Bergey, G.K., 1998. Application of the directed transfer function method to mesial and lateral onset temporal lobe seizures. Brain Topogr. 11, 13—21. Goldman-Rakic, P.S., 1988. Topography of cognition: parallel distributed networks in primate association cortex. Annu. Rev. Neurosci. 11, 137—156. Gotman, J., 2008. Epileptic networks studied with EEG-fMRI. Epilepsia 49, 42—51. Gotman, J., Grova, C., Bagshaw, A., Kobayashi, E., Aghakhani, Y., Dubeau, F., 2005. Generalized epileptic discharges show thalamocortical activation and suspension of the default state of the brain. Proc. Natl. Acad. Sci. U. S. A. 102, 15236—15240. Gramfort, A., Papadopoulo, T., Olivi, E., Clerc, M., 2010. OpenMEEG: opensource software for quasistatic bioelectromagnetics. Biomed. Eng. Online 9, 45.
Effective connectivity in SWD Gupta, D., Ossenblok, P., van Luijtelaar, G., 2011. Space-time network connectivity and cortical activations preceding spike wave discharges in human absence epilepsy: a MEG study. Med. Biol. Eng. Comput. 49, 555—565. Gusnard, D.A., Raichle, M.E., 2001. Searching for a baseline: functional imaging and the resting human brain. Nat. Rev. Neurosci. 2, 685—694. Hamandi, K., Salek-Haddadi, A., Laufs, H., Liston, A., Friston, K., Fish, D.R., Duncan, J.S., Lemieux, L., 2006. EEG-fMRI of idiopathic and secondarily generalized epilepsies. Neuroimage 31, 1700—1710. Holmes, M.D., Quiring, J., Tucker, D.M., 2010. Evidence that juvenile myoclonic epilepsy is a disorder of frontotemporal corticothalamic networks. Neuroimage 49, 80—93. Janz, D., 1985. Epilepsy with impulsive petit mal (juvenile myoclonic epilepsy). Acta Neurol. Scand. 72, 449—459. Janz, D., Christian, W., 1957. Impulsiv-Petit mal. Dtsch. Zeitschrift f. Nervenheilkd. 176, 346—386. Jung, Y.-J., Kang, H.-C., Choi, K.-O., Lee, J.S., Kim, D.-S., Cho, J.-H., Kim, S.-H., Im, C.-H., Kim, H.D., 2011. Localization of ictal onset zones in Lennox—Gastaut syndrome using directional connectivity analysis of intracranial electroencephalography. Seizure 20, 449—457. Kami´ nski, M.J., Blinowska, K.J., 1991. A new method of the description of the information flow in the brain structures. Biol. Cybern. 65, 203—210. Killory, B.D., Bai, X., Negishi, M., Vega, C., Spann, M.N., Vestal, M., Guo, J., Berman, R., Danielson, N., Trejo, J., Shisler, D., Novotny, E.J., Constable, R.T., Blumenfeld, H., 2011. Impaired attention and network connectivity in childhood absence epilepsy. Neuroimage 56, 2209—2217. Kim, J.H., Suh, S.-I., Park, S.-Y., Seo, W.-K., Koh, I., Koh, S.-B., Seol, H.Y., 2012. Microstructural white matter abnormality and frontal cognitive dysfunctions in juvenile myoclonic epilepsy. Epilepsia 53, 1371—1378. Kramer, M.A., Cash, S.S., 2012. Epilepsy as a disorder of cortical network organization. Neuroscientist 18, 360—372. Labate, A., Briellmann, R.S., Abbott, D.F., Waites, A.B., Jackson, G.D., 2005. Typical childhood absence seizures are associated with thalamic activation. Epileptic Disord. 7, 373—377. Laufs, H., Lengler, U., Hamandi, K., Kleinschmidt, A., Krakow, K., 2006. Linking generalized spike-and-wave discharges and resting state brain activity by using EEG/fMRI in a patient with absence seizures. Epilepsia 47, 444—448. Leichnetz, G.R., 2001. Connections of the medial posterior parietal cortex (area 7 m) in the monkey. Anat. Rec. 263, 215—236. Lu, Y., Yang, L., Worrell, G.A., He, B., 2012. Seizure source imaging by means of FINE spatio-temporal dipole localization and directed transfer function in partial epilepsy patients. Clin. Neurophysiol. 123, 1275—1283. Luo, C., Li, Q., Lai, Y., Xia, Y., Qin, Y., Liao, W., Li, S., Zhou, D., Yao, D., Gong, Q., 2011. Altered functional connectivity in default mode network in absence epilepsy: a resting-state fMRI study. Hum. Brain Mapp. 32, 438—449. McGill, M.L., Devinsky, O., Kelly, C., Milham, M., Castellanos, F.X., Quinn, B.T., DuBois, J., Young, J.R., Carlson, C., French, J., Kuzniecky, R., Halgren, E., Thesen, T., 2012. Default mode network abnormalities in idiopathic generalized epilepsy. Epilepsy Behav. 23, 353—359. Meeren, H., Pijn, J., Van Luijtelaar, E., Coenen, A., Lopes da Silva, F., 2002. Cortical focus drives widespread corticothalamic networks during spontaneous absence seizures in rats. J. Neurosci. 22, 1480—1495. Meeren, H., van Luijtelaar, G., Lopes da Silva, F., Coenen, A., 2005. Evolving concepts on the pathophysiology of absence seizures: the cortical focus theory. Arch. Neurol. 62, 371—376.
265 Moeller, F., Maneshi, M., Pittau, F., Gholipour, T., Bellec, P., Dubeau, F., Grova, C., Gotman, J., 2011. Functional connectivity in patients with idiopathic generalized epilepsy. Epilepsia 52, 515—522. Moeller, F., Siebner, H.R., Wolff, S., Muhle, H., Boor, R., Granert, O., Jansen, O., Stephani, U., Siniatchkin, M., 2008a. Changes in activity of striato-thalamo-cortical network precede generalized spike wave discharges. Neuroimage 39, 1839—1849. Moeller, F., Siebner, H.R., Wolff, S., Muhle, H., Granert, O., Jansen, O., Stephani, U., Siniatchkin, M., 2008b. Simultaneous EEG-fMRI in drug-naive children with newly diagnosed absence epilepsy. Epilepsia 49, 1510—1519. Nordli, D.R., 2005. Idiopathic generalized epilepsies recognized by the International League Against Epilepsy. Epilepsia 46 (Suppl. 9), 48—56. O’Muircheartaigh, J., Vollmar, C., Barker, G.J., Kumari, V., Symms, M.R., Thompson, P., Duncan, J.S., Koepp, M.J., Richardson, M.P., 2011. Focal structural changes and cognitive dysfunction in juvenile myoclonic epilepsy. Neurology 76, 34—40. Pascual-Marqui, R.D., 2002. Standardized low-resolution brain electromagnetic tomography (sLORETA): technical details. Methods Find. Exp. Clin. Pharmacol. 24, 5—12. Petrides, M., Pandya, D.N., 1984. Projections to the frontal cortex from the posterior parietal region in the rhesus monkey. J. Comp. Neurol. 228, 105—116. Philiastides, M.G., Sajda, P., 2006. Causal influences in the human brain during face discrimination: a short-window directed transfer function approach. IEEE Trans. Biomed. Eng. 53, 2602—2605. Ronan, L., Alhusaini, S., Scanlon, C., Doherty, C.P., Delanty, N., Fitzsimons, M., 2012. Widespread cortical morphologic changes in juvenile myoclonic epilepsy: evidence from structural MRI. Epilepsia 53, 651—658. Santiago-Rodríguez, E., Harmony, T., Fernández-Bouzas, A., Hernández-Balderas, A., Martínez-López, M., Graef, A., Carlos García, J., Fernández, T., 2002. Source analysis of polyspike and wave complexes in juvenile myoclonic epilepsy. Seizure 11, 320—324. Stefan, H., Paulini-Ruf, A., Hopfengärtner, R., Rampp, S., 2009. Network characteristics of idiopathic generalized epilepsies in combined MEG/EEG. Epilepsy Res. 85, 187—198. Sullivan, J.E., Dlugos, D.J., 2004. Idiopathic generalized epilepsy. Curr. Treat. Options Neurol. 6, 231—242. Tadel, F., Baillet, S., Mosher, J.C., Pantazis, D., Leahy, R.M., 2011. Brainstorm: a user-friendly application for MEG/EEG analysis. Comput. Intell. Neurosci., 879716. Vaudano, A.E., Laufs, H., Kiebel, S.J., Carmichael, D.W., Hamandi, K., Guye, M., Thornton, R., Rodionov, R., Friston, K.J., Duncan, J.S., Lemieux, L., 2009. Causal hierarchy within the thalamocortical network in spike and wave discharges. PLoS ONE 4, e6475. Vollmar, C., O’Muircheartaigh, J., Barker, G.J., Symms, M.R., Thompson, P., Kumari, V., Duncan, J.S., Janz, D., Richardson, M.P., Koepp, M.J., 2011. Motor system hyperconnectivity in juvenile myoclonic epilepsy: a cognitive functional magnetic resonance imaging study. Brain 134, 1710—1719. Westmijse, I., Ossenblok, P., Gunning, B., van Luijtelaar, G., 2009. Onset and propagation of spike and slow wave discharges in human absence epilepsy: a MEG study. Epilepsia 50, 2538—2548. Wilke, C., Worrell, G., He, B., 2011. Graph analysis of epileptogenic networks in human partial epilepsy. Epilepsia 52, 84—93. Woermann, F.G., Free, S.L., Koepp, M.J., Sisodiya, S.M., Duncan, J.S., 1999. Abnormal cerebral structure in juvenile myoclonic epilepsy demonstrated with voxel-based analysis of MRI. Brain 122, 2101—2108.
266 Yang, T., Luo, C., Li, Q., Guo, Z., Liu, L., Gong, Q., Yao, D., Zhou, D., 2013. Altered resting-state connectivity during interictal generalized spike-wave discharges in drug-naïve childhood absence epilepsy. Hum. Brain Mapp. 34, 1761—1767. Yeterian, E.H., Pandya, D.N., 1985. Corticothalamic connections of the posterior parietal cortex in the rhesus monkey. J. Comp. Neurol. 237, 408—426.
C. Lee et al. Zhang, Z., Liao, W., Chen, H., Mantini, D., Ding, J.-R., Xu, Q., Wang, Z., Yuan, C., Chen, G., Jiao, Q., Lu, G., 2011. Altered functional—structural coupling of large-scale brain networks in idiopathic generalized epilepsy. Brain 134, 2912—2928.
