The improvement of digital EEG enabled us to record and analyze wideband EEG. Especially in epilepsy patients for surgery, at least potentially it has delineated epileptogenic area more reliably.1,2
High-frequency oscillations (HFOs), usually defined as oscillatory activities higher than 80 Hz, are divided to two types of activities, i.e., ripples (80–250 Hz) and fast ripples (250–500 Hz). Recent studies revealed that interictal HFOs is closely related to the seizure-onset zone, and surgical resection of the regions generating HFOs rather well correlates with good seizure control.2,3
However, our previous studies in intractable focal patients with epilepsy revealed that ictal direct current (DC) shifts and HFOs were observed within the epileptic areas, being more restricted than those defined by conventional ictal EEG finding. Interictally, even in the invasive EEG recording, focal slow activity is usually prominent in the epileptogenic area, but it remains to be solved whether this slow represents nonepileptic focal hypofunction or epileptic activity although clear spike components were not accompanied.4
In the present patient with intractable focal motor seizures with impaired awareness, we detected ictal DC shifts followed by ictal HFOs in the epileptic foci by chronically implanted subdural electrodes. In addition, we also found interictal slow accompanied by HFOs during both pre- and post-ictal periods. Therefore, we have systematically analyzed co-occurrence of interictal slow and HFOs (IIS + HFO). Co-occurrence of IIS + HFO delineated the same behavior as a combination of ictal DC shifts and HFOs do. We hypothesized that this type of interictal slow arising from epileptic focus could be useful as “epileptic slow.” The purpose of this study is to clarify temporal change of co-occurrence of IIS + HFO and to elucidate clinical usefulness of IIS + HFO as a surrogate marker of epileptogenic zone. In this article, we used the term “IIS + HFO” for co-occurrence of interictal slow and HFOs as defined in the method in detail.
CASE PRESENTATION AND METHODS
The patient was a 61-year-old right-handed man with left parietal tumor and intractable focal motor seizures with impaired awareness since the age of 28 years. Ictal scalp EEG showed rhythmic theta activities in the left frontocentral area.
He finally decided to have epilepsy surgery by chronic implantation of subdural electrode grids and strips to identify seizure-onset zone and functional areas for two weeks (Figs. 1 and 2). Subdural grid electrodes were made of platinum. Electrodes of A and C-H (Fig. 1A) with a recording diameter of 2.3 mm and a center-to-center interelectrode distance of 10 mm were used (Ad-Tech Medical Instrument, Racine, WI). Electrodes of B (Fig. 1A) with a recording diameter of 1.5 mm and a center-to-center interelectrode distance of 5 mm (Unique Medical, Tokyo, Japan) were also used.
Electrocorticogram (ECoG) was recorded with EEG 1,100 (AC amplifier with input impedance of 200 MΩ; Nihon Kohden, Tokyo, Japan) in the following two conditions (1): the bandpass filter of 0.016 to 600 Hz, the sampling rate of 2,000 Hz (3 seizures) and (2) the bandpass filter of 0.016 to 300 Hz, the sampling rate of 1,000 Hz (3 seizures). The system reference was set to the scalp electrode (made of platinum) placed on the skin of the contralateral mastoid bone.1 However, because it contained a lot of EMG artifacts, the reference electrode was set to one of the electrodes under the epicranial aponeurosis.1,5,6
He had 17 seizures recorded during two weeks of monitoring. Conventional initial ictal pattern seen in all seizures started from electrode B07, 08, 14 to 16 placed on the posterior edge of the tumor (Fig. 1A) and spread to the surrounding electrodes. Clear ictal HFO changes were observed at only 5 electrodes (B07, 14–16, C11) and especially HFOs at B07 was the most prominent (Fig. 1B). We considered B07 was the core electrode of epileptogenicity. However, interictal spikes and sharp waves were widely distributed in plates A and B including electrodes far from resection area and the frequency of interictal spikes and sharp waves was also low (see Figure 1, Supplemental Digital Content 1, http://links.lww.com/JCNP/A36).
At B07, interictal slow activity frequently associated with HFO (Fig. 2A filled circles) was observed before the seizure onset, and clearly decreased after the seizure end. Thus, we decided to analyze the co-occurrence of IIS + HFO at B07 and selected 6 seizures that contained less EMG artifacts in the pre- and post-segment of 30 minutes to seizures.
Analyses of Co-occurrence of IIS + HFO
We used the following four methods to define co-occurrence of IIS + HFO. Data analyses were accomplished by using in-house Matlab scripts applicable for offline analyses (Matlab version 8.1.0; the MathWorks Inc, Natick, MA). The number of co-occurrence was counted by one to two board-certified epileptologists (T.I. and M.I.).
1. Visual Analyses of HFOs
First, HFOs were evaluated by time–frequency analyses using the short-time Fourier transform (Fig. 2B-1). The spectral power (μV2) was calculated as 10-Hz resolutions using a 100-ms time window and 50-ms overlapping. Color coordinate was shown in logarithmic scale and baseline was set as all time windows (Fig. 2B-1). By visual inspection of time window of 30 seconds, we took only oscillations visible as “blobs” or an “island”.7,8 In addition, we excluded not only obvious artifacts such as EMG but also HFOs occurring with spikes and slow wave after spike in the ECoG trace.
2. Calculation of Power of Ripple Band
First of all, by visual inspection of time–frequency analysis, the HFO band of 80 to 200 Hz was recognized (Fig. 2B-2). We calculated time series of mean and SD of ripple band (80–200 Hz) power in the pre- and post-segment of 30 minutes to each seizure also by excluding the ictal period in the logarithmic scale. Among HFOs candidate data based on the criteria of the method (1), we finally defined HFOs once the degree of power exceeded the mean + 2 SD. of whole analyses window (Fig. 2.3.-2).
3. Assessment of Slow Wave
We visually analyzed slow activity under the setting of time constant 10 seconds and high-frequency filter of 6 Hz (Fig. 2B-3). We adopted interictal slow in duration of 0.33 to 10 seconds, amplitude ≥50 μV, and co-occurring with HFO.
4. Counting the Number of IIS + HFO
We divided the adopted slow samples (3) into 5 groups depending on the amplitude of the slow wave, i.e., 50 to 99, 100 to 199, 200 to 299, 300 to 399, and 400 μV (Fig. 2C). About 6 seizures, we counted the number of co-occurrence of IIS + HFO every 10 minutes (−30 to −20, −20 to −10, −10 minutes to the beginning of seizure, the end of seizure to +10, +10 to +20, and +20 to +30 minutes).
Statistical analyses were performed using commercially available statistical software (SPSS 15.0 J for Windows; SPSS, Chicago, IL). Wilcoxon signed-rank test was used to compare the differences between the pre- and post-segment of 30 minutes to each seizure. Multiple comparison with Tamhane's T2 measures was made to compare the differences among every 10 minutes and among every amplitude group. The significance level was set at P < 0.05.
The Ethical Committee of Kyoto University approved the experimental protocol (IRB #79) and the patient provided written informed consent for participating in the study.
Co-occurrence of IIS + HFO
The number of co-occurrence of IIS + HFO was 2,890 in the core epileptogenic zone at B07 (see Table 1, Supplemental Digital Content 1, http://links.lww.com/JCNP/A37). The number of IIS + HFO with each seizure every 10 minutes was 86.8 ± 16.1 (mean ± SD, −30 to −20 minutes), 86.3 ± 25.2 (−20 to −10 minutes), 91.7 ± 24.4 (−10 minutes to seizure), 68.8 ± 34.6 (seizure to +10 minutes), 79.7 ± 18.3 (+10 to +20 minutes), and 68.3 ± 26.2 (+20 to +30 minutes). The number of IIS + HFO among all 10-minute segments was not different statistically (Fig. 2C).
However, by comparing between the pre- and post-segment of 30 minutes to each seizure, the number of IIS + HFO significantly decreased after seizure, i.e., pre-segment (88.3 ± 21.1) and post-segment (72.3 ± 26.0) of 30 minutes (Wilcoxon signed-rank test, P < 0.05) (Fig. 2C and D).
Thereafter, we analyzed the amplitude of IIS + HFO between the pre- and post-segment of 30 minutes to each seizure. The number of IIS + HFO with the amplitude of 200 to 299 and 300 to 399 μV, namely 200 to 399 μV, significantly decreased after seizures (Wilcoxon signed-rank test, P < 0.05) (Fig. 2D).
Seizure Outcome after Surgery
Based on the results of clinical and multimodal date, we finally resected the tumor and surrounding area including the ictal onset zone as revealed by ictal, wideband ECoG finding (Fig. 1A). The patient manifested rare disabling seizures after surgery; then, he has been seizure-free for the past 4 years (Engel class I-C). Pathological diagnosis was WHO grade 2 oligoastrocytoma.
The location of IIS + HFO was concordant with ictal onset zone but not irritative zone. The results of our study may suggest that this IIS + HFO was called as interictal “epileptic” slow and that it may delineate the core part of epileptogenic area as similarly as ictal DC shifts and ictal HFOs do.
The IIS + HFO significantly decreased after seizures, especially the activity of 200–399 μV. In general, as the classic EEG interpretation, intermittent regional slow activity is regarded as focal abnormality by the nonepileptic, hypofunction but not epileptic, and usually occurred more frequently immediately after seizures, as opposed to the present findings. Because HFOs reportedly tended to decrease after seizures,9 the present finding rather supports the note that IIS + HFO has the characteristics of epileptic activity but not hypofunction.
Ictal DC shifts can be divided into the passive and active DC shifts based on the generator mechanisms. It could explain the reason why the ictal DC shifts occur not only before but also after conventional ictal EEG pattern. For the passive DC shifts, the generator are indicated as 1) passive sustained depolarization of astrocytes following massive ictal neuronal firing causing increase of extracellular K+,10,11 and (2) sustained paroxysmal depolarization shift occurred during the ictal period.12 But, its contribution is small because ictal DC shifts are usually very large in amplitude such as several millivolts in the subdural EEG recording, much larger than ongoing spike amplitude.
For the active DC shifts, by contrast, astrocyte plays the important role in K+ homeostasis. Once the extracellular K+ increased around the synaptic area within the tripartite synapse, it is buffered by Kir 4.1 channel and then swept out through the gap junction within the functional syncytium of the astrocytes, and finally moved out into the vascular system. Therefore, its K+ homeostasis mechanism was mainly stressed after seizure.13 However, astrocytes spontaneously generate slow oscillation being partly coupled with neuronal activity and also often precede the neuronal activity. In this study, because IIS + HFO showed the slow onset followed by HFOs, the mechanism of active DC shifts was more plausible.
However, pathological HFOs reflect abnormal bursts of population spikes or action potentials of pyramidal cell.14 Most studies have analyzed interictal HFOs, or ictal HFOs to a lesser degree, mainly overriding on the sharp wave or spike.15
IIS + HFO explicitly shows the similar behavior to the co-occurrence of ictal DC shifts and ictal HFOs. Once the condition exceeds the threshold to trigger seizure, the transition from the interictal “epileptic” slow and HFOs into the ictal state occurs. However, its mechanism still remains to be solved. Regarding this interictal “epileptic” slow and HFOs, if it is appropriate, we may coin the term “red slow,” which could be conceptually equivalent to the red spike.
In this study, the data were only analyzed for the one core electrode of epileptogenicity in the pre- and post-segment of 30 minutes to each seizure in a single patient. We focused on IIS + HFO in the selected electrodes, and spikes and sharp wave are usually interictally abundant, although it was infrequent in this selected electrode. Direct comparison between IIS + HFO and spike/sharp wave + HFOs in the whole electrodes is also important. We did not investigate IIS + HFO from the view point of its duration of either of slow or HFO in this study. We rather focused on the duration of slow in duration of 0.33 to 10 seconds (0.1–3 Hz) in this study. In the near future, we hope to be able to conduct further studies.
We focused on co-occurrence of interictal epileptic slow and HFOs (IIS + HFO) in the particular patient. We found that IIS + HFO was observed in the ictal onset area. Those with amplitude of 200 to 399 μV was influenced by and thus decreased after seizures. It might reflect the core part of epileptogenic area as similarly as ictal DC shifts and ictal HFOs. This epileptic slow, “red slow,” may be useful to delineate at least a part of the epileptogenic zone.
1. Ikeda A, Terada K, Mikuni N, et al. Subdural recording of ictal DC shifts in neocortical seizures in humans. Epilepsia 1996;37:662–674.
2. Jacobs J, Staba R, Asano E, et al. High-frequency oscillations (HFOs) in clinical epilepsy. Prog Neurobiol 2012;98:302–315.
3. Ochi A, Otsubo H, Donner EJ, et al. Dynamic changes of ictal high-frequency oscillations in neocortical epilepsy: using multiple band frequency analysis. Epilepsia 2007;48:286–296.
4. Steriade M, Gloor P, Llinas RR, Lopes de Silva FH, Mesulam MM. Report of IFCN Committee on Basic Mechanisms. Basic mechanisms of cerebral rhythmic activities. Electroencephalogr Clin Neurophysiol 1990;76:481–508.
5. Imamura H, Matsumoto R, Inouchi M, et al. Ictal wideband ECoG: direct comparison between ictal slow shifts and high frequency oscillations. Clin Neurophysiol 2011;122:1500–1504.
6. Kanazawa K, Matsumoto R, Imamura H, et al. Intracranially recorded ictal direct current shifts may precede high frequency oscillations in human epilepsy. Clin Neurophysiol 2015;126:47–59.
7. Benar CG, Chauviere L, Bartolomei F, Wendling F. Pitfalls of high-pass filtering for detecting epileptic oscillations: a technical note on “false” ripples. Clin Neurophysiol 2010;121:301–310.
8. Amiri M, Lina JM, Pizzo F, Gotman J. High Frequency Oscillations and spikes: separating real HFOs from false oscillations. Clin Neurophysiol 2016;127:187–196.
9. Pearce A, Wulsin D, Blanco JA, Krieger A, Litt B, Stacey WC. Temporal changes of neocortical high-frequency oscillations in epilepsy. J Neurophysiol 2013;110:1167–1179.
10. Kuffler SW, Nicholls JG, Orkand RK. Physiological properties of glial cells in the central nervous system of amphibia. J Neurophysiol 1966;29:768–787.
11. Bordey A, Lyons SA, Hablitz JJ, Sontheimer H. Electrophysiological characteristics of reactive astrocytes in experimental cortical dysplasia. J Neurophysiol 2001;85:1719–1731.
12. Ayala GF, Matsumoto H, Gumnit RJ. Excitability changes and inhibitory mechanisms in neocortical neurons during seizures. J Neurophysiol 1970;33:73–85.
13. Devinsky O, Vezzani A, Najjar S, De Lanerolle NC, Rogawski MA. Glia and epilepsy: excitability and inflammation. Trends Neurosci 2013;36:174–184.
14. Le Van Quyen M, Bragin A. Analysis of dynamic brain oscillations: methodological advances. Trends Neurosci 2007;30:365–373.
15. Crepon B, Navarro V, Hasboun D, et al. Mapping interictal oscillations greater than 200 Hz recorded with intracranial macroelectrodes in human epilepsy. Brain 2010;133:33–45.