Clinical Neurophysiology 127 (2016) 1794–1802
Contents lists available at ScienceDirect
Clinical Neurophysiology journal homepage: www.elsevier.com/locate/clinph
When spikes are symmetric, ripples are not: Bilateral spike and wave above 80 Hz in focal and generalized epilepsy Francesca Pizzo a,b,⇑, Taissa Ferrari-Marinho a,c, Mina Amiri a, Birgit Frauscher a, Francois Dubeau a, Jean Gotman a a b c
Montreal Neurological Institute and Hospital, McGill University, Montreal, QC, Canada Epilepsy Unit, Careggi Hospital, University of Florence, Florence, Italy Department of Clinical Neurophysiology, Hospital Israelita Albert Einstein, São Paulo, Brazil
See Editorial, pages 1759–1761
a r t i c l e
i n f o
Article history: Accepted 27 November 2015 Available online 22 December 2015 Keywords: Bilateral synchrony Scalp EEG Ripple Fast oscillations Idiopathic generalized epilepsy
h i g h l i g h t s Scalp ripples can be used as an additional tool to lateralize the epileptic focus in secondary bilateral
synchrony. In idiopathic generalized epilepsy scalp ripples are recordable and show an anterior dominance. To differentiate focal patients with secondary bilateral synchrony from patients with idiopathic gen-
eralized epilepsy scalp ripples are not useful.
a b s t r a c t Objective: To evaluate scalp ripples distribution in secondary bilateral synchrony as a tool to lateralize the epileptic focus and to differentiate focal from generalized epilepsy. Methods: Seventeen EEG recordings with bilateral synchronous discharges of focal (focal group-FG: 10) and generalized (generalized group-GG: 7) epilepsy patients were selected for spikes and ripples marking; the spike-normalized ripple rate was calculated in each hemisphere (right/left – anterior/posterior) and a ripple-dominant hemisphere (the one with the highest rate) was identified. Concordance in FG between the ripple dominant hemisphere and the hemisphere of clinical lateralization was evaluated. The ripple-dominant/ripple-nondominant spike-normalized ripple rate ratio was studied to compare groups. Results: In FG the hemisphere of clinical lateralization and the ripple-dominant hemisphere were 100% concordant. In GG only 3/7 patients showed ripples (vs 10/10 FG), all with anterior dominance. No difference in hemisphere ripple dominance between groups was found. Conclusions: Ripples in secondary bilateral synchrony help to lateralize the epileptic focus but do not help to differentiate between focal and generalized epilepsy. This is the first report of visually identified ripples in idiopathic generalized epilepsy. Significance: Ripples confirm the clinical lateralization of the epileptic focus in secondary bilateral synchrony but cannot distinguish between focal and generalized epilepsy. Ó 2015 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.
⇑ Corresponding author. Present address: Institut de Neurosciences des Systèmes, INSERM UMR 1106, 27 Bd Jean Moulin, 13005 Marseille, France. Tel.: +39 3409388165; fax: +33(0)4 91 78 99 14. E-mail address: [email protected]
Bilateral synchronous spike and wave activity can be found on the scalp EEG of patients with focal and generalized epilepsy. The expression ‘‘secondary bilateral synchrony” (SBS) (Tukel and Jasper, 1952) implies that in focal epilepsy the generalized synchronous spike and wave discharge is triggered by an epileptic
http://dx.doi.org/10.1016/j.clinph.2015.11.451 1388-2457/Ó 2015 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.
F. Pizzo et al. / Clinical Neurophysiology 127 (2016) 1794–1802
discharge in a localized area of the brain, whereas ‘‘primary bilateral synchrony” refers to a discharge triggered by bilateral subcortical-cortical interactions. On scalp EEG, secondary and primary bilateral synchrony can look similar; many studies tried to find differences by investigating specific features of the standard frequency EEG, such as time difference in interictal epileptiform discharges (IEDs) between hemispheres (Gotman, 1981; Kobayashi et al., 1992) and spike and wave morphological features (Tukel and Jasper, 1952; Blume and Pillay, 1985), but it remains still challenging for clinicians to distinguish on the basis of standard EEG if there is a focality or not. The neurophysiological basis of SBS is still not completely understood. The thalamic (Penfield and Jasper, 1954) and callosal connections (Musgrave and Gloor, 1980; Gotman, 1981) are thought to have a role in the propagation of the discharge from the cortical foci. Diffuse encephalopathy (Gloor, 1969) and multiple foci (Blume and Pillay, 1985) were suggested to facilitate bisynchronous epileptic discharge. To differentiate between patients with focal and generalized epilepsy is fundamental regarding the medical therapeutic options and regarding surgery in pharmacoresistant patients (Engel et al., 2003). In the last decade, studies on frequencies above the standard scalp EEG range (0.1–70 Hz) opened a promising research area. Scalp ripples (>80 Hz) have been recorded in generalized epilepsy (Kobayashi et al., 2010, 2015), and are emerging as a new biomarker of the seizure onset zone (SOZ) in focal epilepsy, being more specific markers than interictal spikes (Andrade-Valenca et al., 2011; Melani et al., 2013). These findings, as demonstrated earlier in intracranial EEG studies (Jirsch et al., 2006; Urrestarazu et al., 2007; Jacobs et al., 2008, 2010; Wu et al., 2010; Akiyama et al., 2011), suggest a potential role of scalp HFOs in identifying an epileptic focus in patients affected by focal epilepsy. Our goal is to determine if fast oscillations, namely ripples, could lateralize the origin of SBS. We also aim to assess if ripples could be recorded on an adult population of patients affected by IGE and if their distribution in bilateral synchronous IEDs differ between IGE and focal epilepsy patients.
2. Methods 2.1. Patient selection From January 2011 all patients admitted in the EEG-telemetry Unit of the Montreal Neurological Hospital had one EEG recording sampled at 1000 Hz which was performed during the second night of telemetry. These EEGs are collected in a database from which we selected consecutive recordings of patients affected by focal or generalized epilepsy with predominantly bilateral synchronous spike and wave activity on a previous scalp EEG recording. Patients were divided into two groups: focal group (FG), with focal epilepsy, and generalized group (GG), with generalized epilepsy. The diagnosis was made based on clinical, neurophysiological and neuroimaging data by attending staff physicians. Exclusion criteria were unclear diagnosis, severe encephalopathy, artefacts interfering with the scoring of ripples, or absence of IEDs in the EEG recording sampled at 1000 Hz.
P3, P3-O1, Fp1-F7, F7-T3, T3-T5, T5-O1, and symmetrically in the right hemisphere). The analysis in each patient was made during 30 min of N2 and N3 sleep, manually scored according to AASM 2.0 (Berry et al., 2012) by an electrophysiologist with board certification in sleep medicine. To equalize as much as possible IED duration throughout all the patients (as it is known that HFOs on scalp EEG occur most often at the time of IEDs – Melani et al., 2013) we applied this criterion: for each EEG we analyzed the first consecutive 10 min of the N2 and N3 sleep and if 10 IEDs (see ‘‘marking of IEDs and fast oscillations” section) were present we also selected for the ripple analysis the following consecutive 20 min of N2 and N3 sleep (this was valid for all patients included in the FG and for two patients in the GG); if there were less than 10 IEDs in the first 10 min we further selected, in the same night recording, N2 and N3 sleep EEG sections where IEDs were identified (with a duration of a few minutes around each IED) to reach a total time of 30 min per patient. Sleep recordings were chosen because they: (1) present less movement artefacts, (2) have higher ripple rate (Staba et al., 2004; Bagshaw et al., 2009; Dümpelmann et al., 2015) and (3) show higher incidence of SBS (Blume and Pillay, 1985). We excluded from the analysis sections with arousals, awakenings and artefacts. When selecting sections, we also excluded 5 s before and after each major artefact seen in the unfiltered EEG. Channels with a considerable amount of artefacts interfering with ripple identification were also excluded. To minimize the effects of seizure on the EEG we analyzed only recordings at least 2 h before and after a clinical seizure (electrographic seizures were not considered, but they were excluded from the analysis). 2.3. Marking of IEDs and fast oscillations We applied a similar methodology for ripple marking as described in our previous scalp EEG papers (Andrade-Valenca et al., 2011; Melani et al., 2013). IEDs were identified using a bipolar montage with 10s/page time scale, and an amplification of 7.50 lV/mm. Spikes, sharp waves, spikes and slow waves complexes, polyspikes and polyspikes and slow wave complexes were marked. The mark was made from the beginning of the negative or positive deflection to the return at baseline in all channels where IEDs were identified. If consecutive IEDs were present, without a background activity interval, we considered the event as a single IED (Fig. 1). Considering the diffuse IED distribution in our patients, we selected for further analysis all channels for the period of the longest IED identified in the bipolar montage (Fig. 1). Ripples (>80 Hz) were then marked with the IED marking invisible. We used a high-pass filtering at 80 Hz and a low-pass filtering at 200 Hz; a finite-impulse response filter was applied to minimize ringing. Ripples were defined as at least four consecutive fast oscillations (> 80 Hz and < 200 Hz) that clearly stand out of the background. After the EEGs were reviewed twice by one reviewer, each mark was cross-checked by a second reviewer. In ambiguous cases, a consensus was reached by both reviewers. 2.4. Artefact identification
2.2. Recording methods EEGs were obtained using the Harmonie monitoring system (Stellate, Montreal, QC, Canada), with a 300 Hz low pass filter and a 1000 Hz sampling rate. Electrodes were placed according to the 10–20 system and recorded with CPz as reference; a bipolar anterio–posterior montage was used for review (Fp1-F3, F3-C3, C3-
A three-step analysis was carried out in order to avoid contamination with artefacts: (1) we excluded from the analysis all the channels with malfunctions or continuous artefacts (AndradeValenca et al., 2011) and in the channels selected for analysis we excluded each visible artefact (muscle, electrodes artefacts) on the raw EEG at 10 s/page time scale prior to beginning the ripple analysis; (2) we took the morphology of the oscillations into
F. Pizzo et al. / Clinical Neurophysiology 127 (2016) 1794–1802
1s Fig. 1. Example of spike marking: (A) the mark covers all the channels from the first positive or negative deflection to the last return to the baseline; (B) if consecutive IEDs were present without return to the baseline the mark includes all the IEDs and was considered as a single IED.
account. We considered as artefacts the oscillations with irregular morphology, very high amplitude compared to the background, or great variation of amplitude and frequency during the train of oscillation (Bénar et al., 2010); and (3) when there was a doubtful fast oscillation on the filtered EEG we looked at the raw EEG and only if fast activity consistent with the ripple was visible we considered it a true ripple, otherwise we discarded the event (Fig. 2). 2.5. Data analysis All data analysis was performed with MATLAB (The Mathworks Inc.). The total number of ripples and the number of ripples cooccurring with spikes (defined as the overlap of the marking of ripples and IEDs independently of the duration or sequence of both events – Andrade-Valenca et al., 2011) were extracted for each channel (see Fig. 3 for the illustration of ripples co-occurring with spikes); the proportion of patients showing ripples per group was calculated and compared between groups with a two-tailed Fisher exact test. Then the ‘‘spiking duration”, defined as total time of spiking activity in minutes, was calculated, evaluated with the Kolmogorov–Smirnov test that confirmed the normal distribution of the variable, and compared with the student t-test between the FG and GG. The ‘‘total ripple rate” (total number of ripples divided by total data duration 30 min) and the normalized rate of ripple-spike cooccurrence, named ‘‘spike-normalized ripple rate” (number of rip-
ples co-occurring with spikes divided by spiking duration), were computed in each channel. These rates were calculated only in channels having at least one ripple, excluding the channels without ripples. As demonstrated by Melani et al. (2013), HFOs on scalp EEG occur most often at the time of spikes; because of a significant difference in spiking duration between the GG and the FG in our study, we calculated the spike-normalized ripple rate as described above. Without such a normalization, ripple rates could simply reflect spiking rates. We then evaluated the mean spikenormalized ripple rate (in channels with ripples) per patient and we calculated the ripple – IED co-occurrence ratio (number of ripple-IED co-occurrences/total number of ripple) per patient and its mean. We divided the channels in belonging to the left (Fp1-F7, F7-T3, Fp1-F3, F3-C3, T3-T5, T5-O1, C3-P3, P3-O1), to the right (Fp2-F8, F8-T4, Fp2-F4, F4-C4, T4-T6, T6-O2, C4-P4, P4-O2) to the anterior (Fp1-F7, F7-T3, Fp1-F3, F3-C3, Fp2-F8, F8-T4, Fp2-F4, F4-C4) and to the posterior (T3-T5, T5-O1, C3-P3, P3-O1, T4-T6, T6-O2, C4P4, P4-O2) hemispheres (Supplementary Fig. S1). The mean spike-normalized ripple rate was calculated in each hemisphere and a ripple-dominant (the hemisphere with the highest value) and a ripple-nondominant hemisphere were identified between right/left (RL) and anterior/posterior (AP) hemispheres in each patient, then a ripple-dominant/ripple non-dominant spike-normalized ripple rate ratio was calculated for each patient. To assess lateralization in the FG, the RL ripple-dominant hemisphere in each patient was compared with the clinically suspected
F. Pizzo et al. / Clinical Neurophysiology 127 (2016) 1794–1802
Filtered EEG (80-200 Hz)
100 ms Fig. 2. Ripple (A – channel F7-T3) and artefact (B – channel F8-T4) in the filtered (column on the left) and unfiltered (column on the right) EEG signal. In the unfiltered signal, the fast activity is only visible during the ripple event.
hemisphere of the epileptic focus (identified by the staff physicians based on clinical, neuroimaging, and neurophysiological data). The following variables were identified for the statistical analysis: (1) lD (mean of spike normalized ripple rates in the rippledominant hemispheres) in focal (FlD) and generalized group (GlD); (2) lND (mean of spike normalized ripple rates in the ripple-nondominant hemispheres) in focal (FlND) and generalized group (GlND); (3) l(D/ND)n (mean of the ratio between the values of ripple-dominant/ripple-nondominant spike normalized ripple rates) in focal (Fl(D/ND)n) and generalized (Gl(D/ND)n) group. The Kolmogorov–Smirnov test was applied to check if these variables had a normal distribution, and since normality was confirmed for all variables, we used a parametric test for the analysis. We calculated the following: (1) FlD/FlND and GlD/GlND ratios to test the null hypothesis of hemispheric ripplenondominance; (2) FlND/GlND ratio to verify the comparability between the FG and GG; (3) Fl (D/ND)n/Gl (D/ND)n ratio to evaluate the difference in the hemispheric ripple-dominance between the GG and FG. The t-student test was applied to check statistical significance. All analyses were repeated excluding patient 9 who was an outlier with a particularly high ripple rate.
3. Results 3.1. Population and channels characteristics Between January 2011 and August 2014, 50 of the 473 patients with a 1000 Hz night scalp EEG recording presented in a previous EEG report a predominant bilateral synchronous epileptic activity. The EEGs were analyzed and the patients’ clinical information was collected. Thirty EEGs were excluded for the following reasons: lack of IEDs during sleep (11), severe encephalopathy (7), considerable artefacts interfering with ripple scoring (7), no clear diagnosis (4), absence of 2 h seizure freedom (1). Three patients affected by focal epilepsy were not included so that the sizes of the focal and generalized groups would not be too different. Seventeen patients were included (12 women, mean age at evaluation 31 ± 11 years, mean age at seizure onset 12 ± 8 years, mean epilepsy duration 23 ± 15 years), 10 affected by focal epilepsy and 7 by generalized epilepsy. All patients with generalized epilepsy suffered from idiopathic epilepsy. The MRI was normal in 7 patients (2 in the FG and 5 in the GG); the remaining patients’ MRI findings are described in Table 1. The MRI abnormalities in the
F. Pizzo et al. / Clinical Neurophysiology 127 (2016) 1794–1802
Fig. 3. Ripple co-occurring with spikes in a focal epilepsy patient. Diffuse spike activity on the left and ripple only on the Fp2-F4 channel in the right (HPF 80 Hz) co-occurring with the spike. *Indicates an event on Fp2-F8 that was not marked as ripple because it had only 2 waves.
GG were considered incidental findings unrelated to the epileptic syndrome. We analyzed 268 channels (16 patients with 16 channels, 1 patient with 12 channels); the channels Fp1-F7, Fp2-F8, Fp1-F3, Fp2-F4 in patient 13 (GG) were excluded due to artefacts. 3.2. Ripple and spiking rates Ripples were found in 13 of the 17 patients (76%), and in 119 of 268 channels (44%); the proportion of patients per group showing ripples (10/10 in FG and 3/7 in GG) was significantly different (p < 0.05) between FG and GG. Average spiking duration across all patients was 2.8 ± 2.7 min; in the FG it was 3.8 ± 2.8 min and in the GG it was 1.2 ± 2.1 min; the difference in spiking duration between FG and GG was marginally significant (p 0.07). Ripple co-occurring with spikes were present in 8/10 patients in the FG (patients 7 and 8 showed respectively 1 and 2 events not associated with spikes) and in the 3 patients that showed ripples in the GG (for an example of ripples in GG see Fig. 4); the mean ripple – spike co-occurrence ratio was 0.87 ± 0.16 in all patients studied, 0.84 ± 0.19 in the FG and 0.93 ± 0.02 in the GG. The mean spike-normalized ripple rates per group are reported in Table 2. We also calculated the mean total ripple rate: in all patients and in all channels analyzed (268) it was 0.38 ± 1.66/min, considering only channels with ripples (119), it was 0.85 ± 2.41/min. In the FG, the total ripple rate in all channels (160) was 0.53 ± 2.09/min and in channels with ripple (92) 0.93 ± 2.70/min. In the GG, the total ripple rate for all channels (108) was 0.14 ± 0.48/min and for channels with ripples (27), it was 0.56 ± 0.84 /min.
GlD/GlND, p < 0.01), demonstrating a significant hemispheric ripple dominance in both groups. Concerning the FG, all patients had a suspected epileptic focus (data are reported in Table 3) and 8 of them showed ripples cooccurring with IEDs. Concordance between the RL ripple-dominant hemisphere and the clinically suspected hemisphere of the epileptic focus was 100% (8/8) (Table 3). Regarding the GG, of note, all patients showed an anterior ripple-dominant hemisphere (Table 3). Focal and generalized groups were comparable (FlND/GlND, p > 0.05) and they did not differ in AP and RL hemispheric dominance (Fl(D/ND)n/Gl(D/ND)n, p > 0.05), suggesting a similar hemispheric dominance entity in the two groups. All results remained unchanged when excluding patient 9. 4. Discussion The study objectives were (1) to determine if ripples can lateralize the epileptic focus in focal epileptic patients with SBS, (2) to asses if ripples are recordable on the scalp EEG of patients affected by IGE and (3) to analyze if there are differences in scalp EEG ripple distributions between patients affected by focal epilepsy with bilateral synchronous IEDs and patients affected by IGE. We showed that (1) in patients affected by focal epilepsy, ripples are concordant with the lateralization of the epileptic focus, (2) ripples are found in IGE patients, but only in a minority of them compared to focal epilepsy patients and always with an anterior hemisphere dominance and (3) comparing ripple distributions (right/left and anterior/posterior hemispheric dominance) between focal and generalized epilepsy does not help to distinguish between the two syndromes.
3.3. Right-left and antero-posterior hemispheric ripple distribution: characteristics of focal and generalized group and their comparison
4.1. Scalp EEG ripples in focal epilepsy with SBS
The null hypothesis of hemispheric non-dominance in both RL and AP hemispheres in FG and GG was rejected (FlD/FlND and
Ripples showed a lateralized distribution in patients with SBS; the hemisphere of ripple dominance was always concordant with
F. Pizzo et al. / Clinical Neurophysiology 127 (2016) 1794–1802 Table 1 Population characteristics. Group
Scalp interictal EEG pattern (IEDs focality)
1 2 3 4 5 6 7 8 9 10
21/F/15 y 23/F/19 y 26/M/8 m 30/F/13 y 33/M/9 y 27/F/4 y 30/F/9 m 50/M/2 y 21/F/12 y 45/F/31 y
FP lobe F lobe TP lobe F lobe Bil TPO F lobe T lobe F lobe T lobe T lobe
Subtle atrophy of L med-post-C gyr Normal L TP encephalitis sequalae L FCD deep preC gyrus Bil PTO H Normal R MTS R giant cell astrocytoma, bil MTS Bil TPO NH Normal
GSW and GPSW Bil F SW PSW Bil F SW ShW GSW and bil post ShW and SW GSW Bil F spikes Bil F SW (L F) GSW an GPSW (R T) GSW, GPSW (L post)
11 12 13 14 15 17 17
23/M/7 y 23/M/3 y 59/F/14 y 38/F/5 y 29/F/12 y 23/F/18 y 60/F/4 y
IGE Atypical IGE JME IGE JME JME IGE
Normal L Hip atrophy Post traumatic R T gliosis Normal Normal Normal Cerebellum atrophy
GS, GSW, GPSW GSW, GPSW GS, GSW GS, GSW GSW (R T predominance) GS, GSW GS, GSW
IED – interictal epileptiform discharge, R – right, L – left, bil – bilateral, F – frontal, T – temporal, P – parietal, O – occipital, C – central, ant – anterior, nid – middle, post – posterior, med – medial, lat – lateral, GSW – generalized spike and wave, PSW – polyspike and wave, GPSW – generalized polyspike and wave, ShW – sharp wave, S – spike, GS – generalized spike, H – heterotopia, MTS – mesial temporal sclerosis, Hipp – hippocampus, JME – juvenile myoclonic epilepsy, IGE – idiopathic generalized epilepsy, Y – year, M – months.
the clinically suspected side of the epileptic focus when it was identified. We also found a significant ripple dominance in anterior/posterior brain regions that, together with the lateralization, defines a quadrant of ripple dominance for each patient that might help to localize the epileptic focus supporting the clinical data. The total ripple rate and the ripple spike co-occurrence rate were higher compared to our previous studies on scalp HFOs (Andrade-Valenca et al., 2011; Melani et al., 2013). This could be explained by the fact that in the current study we looked at focal patients with a widespread spike distribution, whereas the previous studies investigated patients with focal epilepsy in general. This explanation is supported by a study in childhood sleep induced electrical status epilepticus (Kobayashi et al., 2010) which showed a higher rate of ripple co-occurring with spikes with continuous spike-wave during slow wave sleep than with focal spikes; the authors suggested that generalized spikes in that condition are due to a SBS phenomenon. The analysis of ripples during SBS might also provide a better understanding of the neurophysiological mechanisms underlying this phenomenon and, particularly, about the role of the brain midline structures. How can we explain higher rates of ripples during SBS compared to focal spikes? What role do the midline brain structures play? An insight about the relationship between HFOs and the thalamus comes from a fMRI study by Fahoum et al. (2014). The authors showed that thalamic metabolic involvement was more common in focal epileptic patients when IEDs were accompanied by more ripples. Aghakhani et al. (2006) showed in a EEG-fMRI study (2006) that the thalamus is sometimes involved in partial epilepsy during interictal spikes, and more prominently during SBS. The thalamus might play a role in generating ripples in the neocortex, and it can be even more involved during ripples occurring in SBS discharges than during ripples with focal spikes. 4.2. Scalp EEG ripples in generalized epilepsy Others studies already demonstrated that ripples can be recorded on the scalp EEG of patients affected by symptomatic generalized epilepsy (Iwatani et al., 2012; Kobayashi et al., 2010, 2015), but the present research illustrates, for the first time, that
ripples can be recorded on the scalp EEG of adult patients affected by IGE. We showed that ripples in IGE patients are present (3/7 patients in the GG) but only in a minority of patients compared to the FG. Ripples in IGE have an asymmetric distribution: in particular they have an anterior dominance and are not right/left symmetrical, showing predominance in one hemisphere in some patients and in the other hemisphere in other patients. In IGE, generalized IEDs showed the highest amplitude on the anterior EEG channels (Weir, 1965); we demonstrated that ripple co-occurring with spikes are predominant in the same brain region. Recent findings on HFOs distribution during ictal MEG–EEG recording in patients affected by childhood absence epilepsy are concordant with our results. Tenney et al. (2014) showed that HFOs recorded on MEG during absence seizure are lateralized and almost exclusively confined to the prefrontal and orbitofrontal cortex. The orbitofrontal dominance in children affected by childhood absence epilepsy was also observed in a MEG study performed by Miao et al. (2014). In the last decade a growing number of functional neuroimaging data showed focal aspects in generalized epilepsy: EEG-fMRI studies demonstrated focal cortical activation in the medial frontal cortex, precuneus, lateral parietal and frontal cortex during generalized spike and wave discharges in absence epilepsy (Moeller et al., 2010; Carney et al., 2010). Boundaries between generalized and partial epilepsy are becoming less marked and the classification of the two different entities has been revisited (Berg et al., 2010). In this perspective, to find asymmetries in ripples distribution in IGE is not as unexpected, and it simply stresses that the dichotomy between generalized and focal epilepsy is less pronounced than we thought. However, our findings refer exclusively to the neurophysiological aspects of the two epilepsy syndromes; from a clinical point of view focal and generalized epilepsy still represents two very different entities. Analyzing the unfiltered scalp EEG of patient 12 (in the GG) we found some train of widespread oscillations (10–20 Hz) usually before IEDs, which we decided not to mark as IED. This EEG pattern is known in the NREM sleep of generalized patients affected by Lennox Gastaut syndrome, and it is also reported in IGE patients
F. Pizzo et al. / Clinical Neurophysiology 127 (2016) 1794–1802
10μV 100μV 1s
Fig. 4. Ripples co-occurring with spikes in an idiopathic generalized epilepsy patient. Ripple distribution is predominant on the anterior EEG channels.
Table 2 Mean spike-normalized ripple rate calculated as number of ripples co-occurring with spikes divided by spiking duration. The number of channels is given in brackets.
ALL FG GG
Channels with event
All channels excluding pt 9
Channels with event excluding pt 9
1.96 ± 7.6 (268) 2.66 ± 9.5 (160) 0.91 ± 2.47 (108)
4.40 ± 10.9 (109) 4.62 ± 12.3 (83) 3.77 ± 3.88 (26)
0.95 ± 2.2 (252) 0.99 ± 1.9 (144)
2.53 ± 2.9 (95) 2.06 ± 2.4 (69)
Table 3 Distribution of ripples across right-left and anterior-posterior hemispheres and comparison to the suspected epileptic focus. Group
Suspected epileptic focus
Ripple-dominant/ripple-nondominant RL hemisphere (RL ripple-dominant hemisphere)
Ripple-dominant/ripple-nondominant AP hemisphere (AP ripple-dominant hemisphere)
1 2 3 4 5 6 7 8 9 10
L FP LF L TP LF Bil TP (>R) LF RT LF R PT L post T
3 (L) 1.75 (L) 1.2 (L) 3.8 (L) 1.29 (R) 13.31 (L) – – 50 (R) 8.34 (L)
3.67 (P) 1.2 (P) 3.30 (A) 47 (A) 8.26 (P) 1.12 (P) – – 2.85 (P) 5.25 (P)
11 12 13 14 15 16 17
– – – – RT prevalence – –
1.21 (L) 1.75 (R) – – 5 (R) – –
5.25 (A) 4.5 (A) – – 11 (A) – –
R – right, L – left, F – frontal, T – temporal, P – parietal, ripple-dominant/ripple-nondominant RL hemisphere. (Mean of RL ripple-dominant hemisphere spike normalized ripple rate/mean of RL ripple-nondominant hemisphere spike normalized ripple rate), ripple-dominant/ripplenon dominant AP hemisphere (mean of AP ripple-dominant hemisphere spike normalized ripple rate/mean of ripple-nondominant hemisphere spike-normalized ripple rate).
(Halász et al., 2004), and was related to bad prognosis (Guye et al., 2001). Patient 13 has a pharmacoresistant IGE and a complex mood disorder: looking at the time of this activity for frequencies above
80 Hz we found a high rate of ripples that are not reported in the analysis. The role of fast oscillations on scalp EEG related to prognosis in IGE awaits further exploration.
F. Pizzo et al. / Clinical Neurophysiology 127 (2016) 1794–1802
4.3. Differences between focal and generalized patients
Two differences were found between the two groups: the proportion of ripple occurrence and the IED duration. Only few IGE patients presented ripples compared to the focal epilepsy patients (43% in GG vs 100% in FG, p < 0.05), but in IGE ripples had a higher spike co-occurrence rate compared to the focal patients (0.93 vs. 0.84 respectively). This means that almost each ripple in IGE was detected at the time of IED and very few ripples were detected without spikes. Could the paucity of ripples without spikes in IGE reflect the absence of a defined epileptic cortical focus? Why some generalized IEDs had ripples and others not? Is the thalamus involved in ripple generation in IGE? Further studies are needed to answer these questions. The difference in IED duration (p 0.07) is probably multifactorial and due to the different epileptic syndromes, to the better pharmacological control in IGE, and to the sleep staging selection (N2 and N3, with exclusion of awakenings and transition periods). IGE patients tend to have more IEDs during the sleep-wake transition or unstable sleep (Terzano et al., 1989; Gigli et al., 1992; Bonakis and Koutroumanidis, 2009; Zambrelli and Canevini, 2011), periods that were excluded from the analysis, while the FG presents more IEDs during slow wave sleep (Sammaritano et al., 1991) with highest spike rate during high amplitude slow waves (Frauscher et al., 2015).
Aghakhani Y, Kobayashi E, Bagshaw AP, Hawco C, Bénar CG, Dubeau F, et al. Cortical and thalamic fMRI responses in partial epilepsy with focal and bilateral synchronous spikes. Clin Neurophysiol 2006;117:177–91. Akiyama T, McCoy B, Go CY, Ochi A, Elliott IM, Akiyama M, et al. Focal resection of fast ripples on extraoperative intracranial EEG improves seizure outcome in pediatric epilepsy. Epilepsia 2011;52:1802–11. Andrade-Valenca LP, Dubeau F, Mari F, Zelmann R, Gotman J. Interictal scalp fast oscillations as a marker of the seizure onset zone. Neurology 2011; 77:524–31. Bagshaw AP, Jacobs J, Levan P, Dubeau F, Gotman J. Effect of sleep stage on interictal high-frequency oscillations recorded from depth macroelectrodes in patients with focal epilepsy. Epilepsia 2009;50:617–28. Bénar CG, Chauvière L, Bartolomei F, Wendling F. Pitfalls of high-pass filtering for detecting epileptic oscillations: a technical note on ‘‘false” ripples. Clin Neurophysiol 2010;121:301–10. http://dx.doi.org/10.1016/ j.clinph.2009.10.019. Berg AT, Berkovic SF, Brodie MJ, Buchhalter J, Cross JH, Van Emde Boas W, et al. Revised terminology and concepts for organization of seizures and epilepsies: report of the ILAE Commission on Classification and Terminology, 2005–2009. Epilepsia 2010;51:676–85. Berry RB, Brooks R, Gamaldo CE, Harding SM, Marcus CL, Vaughn BV, for the American Academy of Sleep Medicine. The AASM manual for the scoring of sleep and associated events: rules, terminology and technical specifications, version 2.0. Darien Illinois: American Academy of Sleep Medicine; 2012. www. aasmnet.org. Blume WT, Pillay N. Electrographic and clinical correlates of secondary bilateral synchrony. Epilepsia 1985;26:636–41. Bonakis A, Koutroumanidis M. Epileptic discharges and phasic sleep phenomena in patients with juvenile myoclonic epilepsy. Epilepsia 2009;50:2434–45. Carney PW, Masterton RA, Harvey AS, Scheffer IE, Berkovic SF, Jackson GD. The core network in absence epilepsy: differences in cortical and thalamic BOLD response. Neurology 2010;75:904–11. Dümpelmann M, Jacobs J, Schulze-Bonhage A. Temporal and spatial characteristics of high frequency oscillations as a new biomarker in epilepsy. Epilepsia 2015;56:197–206. . Engel JJ, Wiebe S, French J, Sperling M, Williamson P, Spencer D, et al. Practice parameter: temporal lobe and localized neocortical resections for epilepsy: report of the quality standards subcommittee of the American Academy of Neurology, in association with the American Epilepsy Society and the American Association of Neurological Surgeons. Neurology 2003;60:538–47. Fahoum F, Melani F, Andrade-Valença L, Dubeau F, Gotman J. Epileptic scalp ripples are associated with corticothalamic BOLD changes. Epilepsia 2014;55:1611–9. http://doi.wiley.com/10.1111/epi.12760. Frauscher B, von Ellenrieder N, Ferrari-Marinho T, Avoli M, Dubeau F, Gotman J. Facilitation of epileptic activity during sleep is mediated by high amplitude slow waves. Brain 2015;138:1629–41. Gigli GL, Calia E, Maciani MG, Mazza S, Mennuni G, Diomedi M, Terzano MG, Janz D. Sleep microstructure and EEG epileptiform activity in patients with juvenile myoclonic epilepsy. Epilepsia 1992;33:799–804. Gloor P. Neurophysiological bases of generalized seizures termed centrencephalic. In: Gastaut H, Jasper H, Bancaud J, Waltregny A, editors. The physiopathogenesis of the epilepsies. Spingfield, Illinois: Chales C. Thomas; 1969. p. 209–36. Gotman J. Interhemispheric relations during bilateral spike-and-wave activity. Epilepsia 1981;22:453–66. Guye M, Bartolomei F, Gastaut JL, Chauvel P, Dravet C. Absence epilepsy with fast rhythmic discharges during sleep: An intermediary form of generalized epilepsy? Epilepsia 2001;42:351–6. Halász P, Janszky J, Barcs G, Szucs A. Generalised paroxysmal fast activity (GPFA) is not always a sign of malignant epileptic encephalopathy. Seizure 2004;13:270–6. Iwatani Y, Kagitani-Shimono K, Tominaga K, Okinaga T, Kishima H, Kato A, et al. Ictal high-frequency oscillations on scalp EEG recordings in symptomatic West syndrome. Epilepsy Res 2012;102:60–70. http://dx.doi.org/10.1016/j. eplepsyres.2012.04.020. 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 CÉ, Hall J, Olivier A, et al. High-frequency electroencephalographic oscillations correlate with outcome of epilepsy surgery. Ann Neurol 2010;67:209–20. Jirsch JD, Urrestarazu E, LeVan P, Olivier A, Dubeau F, Gotman J. High-frequency oscillations during human focal seizures. Brain 2006;129:1593–608. Kobayashi K, Ohtsuka Y, Oka E, Ohtahara S. Primary and secondary bilateral synchrony in epilepsy: differentiation by estimation of interhemispheric small time differences during short spike-wave activity. Electroencephalogr Clin Neurophysiol 1992;83:93–103. Kobayashi K, Watanabe Y, Inoue T, Oka M, Yoshinaga H, Ohtsuka Y. Scalp-recorded high-frequency oscillations in childhood sleep-induced electrical status epilepticus. Epilepsia 2010;51:2190–4. Kobayashi K, Akiyama T, Oka M, Endoh F, Yoshinaga H. A storm of fast (40–150 Hz) oscillations during hypsarrhythmia in West syndrome. Ann Neurol 2015;77:58–67.
4.4. Strength and potential limitation of the study The strength of the study is the careful design, including interrater agreement on HFO marking. Limits of the study are the retrospective design, the relatively small cohort of patients and the selection of EEG sections that were non-consecutive in some of the patients (in 5 GG patients who did not present 10 IEDs in the first 10 min of the first sleep stage, EEG section around IEDs were selected to obtain a spiking duration as similar as possible to that of the FG). A major challenge remains the differentiations between ‘‘true” HFOs and artefacts of muscle origin or spike filtering, although we applied a careful three step analysis combined with a review of ripple events by two raters. 5. Conclusion Analyzing ripple distribution on the scalp EEG in SBS patients could help to lateralize the epileptic focus in focal epileptic patients but it is not useful to discriminate between focal and generalized epilepsy. Ripples can be found in IGE but their role remains to be elucidated. Acknowledgements The authors thank Natalja Zazubovits for the EEG database and Dr Ilan Goldberg for help in collecting and analyzing clinical data (Montreal Neurological Institute). BF was supported by the Austrian Science Fund (Schrödinger fellowship abroad J3485-B24). This work was supported by a grant of the Canadian Institutes of Health Research (MOP 102710). Conflict of interest: None of the authors have potential conflicts of interest to be disclosed. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.clinph.2015.11. 451.
F. Pizzo et al. / Clinical Neurophysiology 127 (2016) 1794–1802
Melani F, Zelmann R, Dubeau F, Gotman J. Occurrence of scalp-fast oscillations among patients with different spiking rate and their role as epileptogenicity marker. Epilepsy Res 2013;106:345–56. Miao A, Xiang J, Tang L, Ge H, Liu H, Wu T, et al. Using ictal high-frequency oscillations (80–500 Hz) to localize seizure onset zones in childhood absence epilepsy: a MEG study. Neurosci. Lett. 2014;566:21–6. http://dx.doi.org/ 10.1016/j.neulet.2014.02.038. Moeller F, Levan P, Muhle H, Stephani U, Dubeau F, Siniatchkin M, et al. Absence seizures: individual patterns revealed by EEG-fMRI. Epilepsia 2010;51:2000–10. Musgrave J, Gloor P. The role of the corpus callosum in bilateral interhemispheric synchrony of spike and wave discharge in feline generalized penicillin epilepsy. Epilepsia 1980;21:369–78. Penfield W, Jasper H. Epilepsy and the functional anatomy of the human brain. Boston: Little, Brown; 1954. Sammaritano M, Gigli GL, Gotman J. Interictal spiking during wakefulness and sleep and the localization of foci in temporal lobe epilepsy. Neurology 1991;41:290–7. Staba RJ, Wilson CL, Bragin A, Jhung D, Fried I, Engel Jr J. High-frequency oscillations recorded in human medial temporal lobe during sleep. Ann Neurol 2004;56:108–15.
Tenney JR, Fujiwara H, Horn PS, Vannest J, Xiang J, Glauser TA, et al. Low- and highfrequency oscillations reveal distinct absence seizure networks. Ann Neurol 2014;76:558–67. . Terzano MG, Parrino L, Anelli S, Halasz P. Modulation of generalized spike and wave discharges during sleep by cyclical alternating pattern. Epilepsia 1989;30:772–81. Tukel K, Jasper H. The electroencephalogram in parasagittal lesions. Electroencephalogr Clin Neurophysiol 1952;4:481–94. Urrestarazu E, Chander R, Dubeau F, Gotman J. Interictal high-frequency oscillations (100–500 Hz) in the intracerebral EEG of epileptic patients. Brain 2007;130:2354–66. Weir B. The morphology of the spike-wave complex. Electroencephalogr Clin Neurophysiol 1965;19:284–90. 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. Zambrelli E, Canevini MP. Pre- and post-dormitum epilepsies: Idiopathic generalized epilepsies. Sleep Med 2011;12:S17–21. http://dx.doi.org/10.1016/ j.sleep.2011.10.0.