Brain Advance Access originally published online on May 20, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Brain, Vol. 127, No. 7, 1496-1506,
July 2004
© 2004 Guarantors of Brain
doi: 10.1093/brain/awh149
High-frequency oscillations and seizure generation in neocortical epilepsy
1 Departments of Neurology and Bioengineering, and 2 Department of Neurosurgery, Hospital of the University of Pennsylvania, Philadelphia, PA and 3 Department of Neurology and Division of Epilepsy, Mayo Clinic, Rochester, MN, USA
Correspondence to: Gregory A. Worrell, MD, PhD, Department of Neurology and Division of Epilepsy, Mayo Clinic, Rochester, MN 55905, USA E-mail: Worrell.Gregory{at}mayo.edu
Received June 13, 2003. Revised December 3, 2003. Accepted January 27, 2004.
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionReferences |
|---|
Neocortical seizures are often poorly localized, explosive and widespread at onset, making them poorly amenable to epilepsy surgery in the absence of associated focal brain lesions. We describe, for the first time in an unselected group of patients with neocortical epilepsy, the finding that high-frequency (60–100 Hz) epileptiform oscillations are highly localized in the seizure onset zone, both before and temporally removed from seizure onset. These findings were observed in all six patients with neocortical epilepsy out of 23 consecutive patients implanted with intracranial electrodes for pre-surgical evaluation during the study period. The majority of seizures (62%) in these patients were anticipated by an increase in high-frequency activity in the 20 min prior to neocortical seizure onset. Contrary to observations in normal brain, high-frequency activity was strongly modulated by behavioural state, and was maximal during slow-wave sleep, which may explain the propensity for neocortical onset seizures to begin during sleep. These findings point to an important role for neuromodulatory circuits, probably involving the thalamus, in mechanisms underlying seizure generation in neocortical epilepsy. These findings demonstrate that high-frequency epileptiform oscillations may prove clinically useful in localizing the seizure onset zone in neocortical epilepsy, for identifying periods of increased probability of seizure onset, and in elucidating mechanisms underlying neocortical ictogenesis. Confirmation that prolonged bursts of high-frequency activity may predict focal onset neocortical seizures will require prospective validation on continuous, prolonged recordings in a larger number of patients. Importantly, the results show that the dynamic range utilized in current clinical practice for localization of epileptogenic brain largely ignores fundamental oscillations that are signatures of an epileptogenic brain. It may prove that many currently available clinical EEG systems and EEG analysis methods utilize a dynamic range that discards clinically important information.
Key Words: Keywords: EEG; epilepsy; high-frequency oscillations; ictogenesis; sleep
Abbreviations: Abbreviations: HFEO= high-frequency epileptiform oscillation
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
The range of EEG oscillation frequencies thought to be physiologically important for human brain function extends well beyond the frequency bandwidth originally considered by its discoverer, Hans Berger (Berger, 1929
; Curio, 2000a). Recent studies describe gamma oscillations (
30–60 Hz) that are believed to play a role in learning and memory (Lisman and Idiart, 1995; Llinas, 1988), ripple oscillations (100–200 Hz) that may be important for memory consolidation (Buzsaki, 1989, 1996; Bragin et al., 1999a; Grenier et al., 2001), and evoked potential oscillations (600 Hz) (Curio, 2000b) produced by somatosensory stimulation. In addition to normal brain function, high-frequency oscillations (>60 Hz) have been described in human and animal epileptic foci (Allen et al., 1992; Fisher et al., 1992; Alarcon et al., 1995; Bragin et al., 1999b; Traub et al., 2001; Spencer and Lee, 2000; Grenier et al., 2003).
Nonetheless, the majority of human studies using intracranial EEG, performed as part of pre-surgical evaluation in patients with intractable epilepsy, usually report a limited frequency range (0.1–30 Hz) of EEG activity (Quesney et al., 1992; Schiller et al., 1998; Quesney, 2000). This limited frequency range may reflect the fact that oscillations with frequencies above 30 Hz are relatively low amplitude and are obscured by lower frequency activity (Nunez, 1981). For this reason, and the fact that most commercial EEG systems employ low-pass filters with cut-off frequencies from 70 to 100 Hz, we suggest that high-frequency oscillations are probably under-recognized in human intracranial EEG studies.
In fact, recent studies in selected groups of patients show that neocortical seizures can begin with low-amplitude high-frequency oscillations (Allen et al., 1992; Fisher et al., 1992; Alarcon et al., 1995; Traub et al., 2001); and in the interictal (between seizures) period there are high-frequency epileptiform oscillations (HFEOs) present (Bragin et al., 1999a,b; Traub et al., 2001), suggesting that HFEOs may be involved in seizure generation (ictogenesis) (Traub et al., 2001). Previous studies, however, report on highly selected patients, and the actual prevalence of HFEO in neocortical epilepsy remains unknown. Whether interictal HFEOs can spatially localize epileptogenic brain, and the role of HFEOs in seizure generation are unclear. For these reasons, we investigated patients with neocortical onset seizures undergoing monitoring with subdural–intracranial electrodes as part of their evaluation for possible epilepsy surgery. In contrast to mesial temporal lobe onset seizures, neocortical onset seizures are often poorly localized, explosive and widespread at onset, making individuals with these seizures poor candidates for epilepsy surgery. We hypothesized that HFEOs might be useful in localizing the functional seizure onset zone in these difficult to treat patients with neocortical epilepsy, which is critical for successful epilepsy surgery, and for identifying periods of increased probability of seizure activity.
The presence of a reliable seizure precursor, a harbinger of focal neocortical seizures, would open a therapeutic window possibly allowing interruption of ictogenesis and preventing seizures (Litt et al., 2001; Litt and Lehnertz, 2002). Better localization of the seizure onset zone and epileptogenic brain by mapping HFEOs in neocortical epilepsy might lead to better techniques for surgically disrupting the epileptic network in these patients.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Subjects
Twenty-three consecutive patients who underwent long-term continuous video and intracranial EEG monitoring at Emory University and University of Pennsylvania as part of their evaluation for epilepsy surgery were studied. Each patient gave their informed consent for participation in these research studies under approval of the Emory University and University of Pennsylvania Internal Review Boards. Two clinical epileptologists (B.L. and G.A.W.) reviewed and marked each patient’s entire video-EEG record for seizures (Fig. 1) and behavioural state.
|
EEG electrodes
Each patient had intracranial electrodes placed according to standard pre-surgical evaluation protocols (Engel, 1987
; Quesney et al., 1992; Quesney, 2000). Ad-Tech subdural grids, strips and depth electrodes (AD-TECH Medical Instrument Corporation, Racine, WI) were used for all studies. Subdural electrodes had varying numbers of contacts, depending on the clinical requirements. The individual electrode contacts of the subdural grids are 4.0 mm diameter platinum discs with a centre-to-centre electrode distance of 10 mm. Patients requiring mesial temporal recordings had 8-contact depth electrodes stereotaxically placed into each temporal lobe via occipital or temporal burr holes.
Intracranial EEG data collection and storage
Continuous video and intracranial EEG were collected using a digital, 64-channel, 12-bit, Nicolet BMS-5000 (Nicolet Biomedical, Madison, WI) epilepsy monitoring system, and stored on videotape. Referentially recorded EEG was band-pass filtered from 0.1 to 100 Hz, digitized at 200 Hz and archived to compact disc (CD-ROM) for later processing. Bipolar electrode montages were used to eliminate common mode artefact, and a digital 60 Hz notch filter was employed to eliminate line noise. Data were processed using custom MATLAB (Mathworks, Natick, MA) programs. Hospital stays varied from 3 to 27 days, yielding 1.5 Gbytes/day of raw EEG data.
Behavioural state
Videotapes were viewed in their entirety for clinical seizures, and approximation of sleep/wake periods. Finer resolution of sleep/wake state was achieved by examining simultaneous EEGs to identify periods of slow-wave sleep characterized by widely distributed delta frequency activity (0.1–4 Hz). Because the clinical recording montage did not include electrodes for recording eye movements or muscle activity, complete sleep staging (Rechtschaffen and Kales, 1968) was not possible. Therefore, we characterized the entire hospital stay into four possible behavioural states: wakefulness; slow-wave sleep; seizure; or indeterminate. Only slow-wave sleep, wake and seizure records were selected for further analysis.
Determination of seizures, seizure onset zone and interictal epileptiform activity
Seizure onset times were determined by visual identification of a clear electrographic seizure discharge, and then looking backwards in the record for the earliest EEG change from baseline associated with the seizure. The earliest EEG change was selected as the seizure onset time. The seizure onset zone, defined by the electrode(s) with the earliest onset of seizure activity, was documented for each seizure, and classified as mesial temporal, neocortical temporal or neocortical extra-temporal. In patients with independent mesial temporal and neocortical onset seizures, only the neocortical onset seizures were selected for analysis. The electrode contacts exhibiting interictal spikes, sharp waves (Engel, 1987) and/or HFEOs (>60 Hz) were identified by visual review (Fig. 3). Lastly, electrodes far removed from the seizure onset zone were chosen as normal/control/non-epileptogenic brain. If available, the homologous region contralateral to the seizure onset zone was selected as the normal control for that patient.
|
Data analysis
Neocortical seizures were analysed provided the seizure onset was preceded by at least 2 h of uninterrupted artefact-free interictal EEG. Under these strict criteria for inclusion, only two neocortical seizures were excluded, except in patient 21 who had multiple seizure clusters. In this case, only the initial seizure of each cluster was analysed. For each patient’s analysis, EEG seizure records (2 h records containing the seizure onset at 1 h and 50 min), and 10 each of randomly selected 15 min slow-wave sleep and wake baseline records were analysed. Seizure onset times and the seizure onset zone, both determined by visual review, were confirmed by time–frequency analysis (Figs 1 and 2). The frequency of oscillation at seizure onset was then determined by spectral analysis, selecting the frequency of the dominant spectral peak obtained from the initial 3–5 s at seizure onset (Fig. 2). The spatial location and frequency of interictal HFEOs were identified visually and confirmed using time–frequency and spectral analysis (Fig. 3).
|
In addition to HFEOs, the level of broadband high-frequency activity was quantified by band-pass filtering (60–100 Hz) raw EEG with a 6-pole Butterworth filter and then creating an average energy time series for each channel defined by:
where V[n] is the bipolar voltage between adjacent contacts at discrete time tn, M a 0.100 second window (M = 20 data points), and D a 0.045 second window overlap (D = nine points). Average high-frequency energy time series were created for seizure records, and the 150 min of randomly selected interictal slow-wave sleep and wake records.
Statistical analysis
A large degree of variability in the high-frequency energies was observed between patients due to the wide range of effective impedance values for subdural electrodes. This was the case despite attempts to optimize electrode impedance and to minimize artefacts. For this reason, we performed statistical analysis on each patient individually rather than across all patients. For each patient, all the EEG time series, 10 randomly selected 15 min EEG baseline records from slow-wave sleep and wake, and the 120 min seizure records were first converted to a running average energy time series (1) for each electrode site. In order to best probe for statistically significant differences in high-frequency energy at different electrode sites and/or during different behavioural states, we first obtain a normal distribution of the average energy data using the log transform of the high-frequency energy (Gasser et al., 1982). The spatial distribution of broadband high-frequency activity in the seizure onset zone and normal brain regions in both wake and sleep periods was investigated using analysis of variance (ANOVA). For post hoc analysis, Tukey’s multiple comparison tests was used. For each patient, the ‘within-group’ measures were repeated high-frequency energies obtained from 10 randomly selected 15 minute records. The ‘between-group’ measures included the seizure onset zone (epileptogenic brain), normal brain, and sleep and wake behavioural states for each brain region. The F statistic is used to test the null hypothesis that there is no difference in the high-frequency energy between behavioural states, wake and slow-wave sleep, in the epileptogenic and control/normal/non-epileptogenic brain for a given subject.
To determine if the high-frequency energy is increased within epileptogenic brain in the time period prior to seizure onset, we investigate the null hypothesis that the average high-frequency energy within the seizure onset zone is unchanged in the 20 min period prior to seizure onset. The hypothesis is tested by creating random samples of high-frequency energies in the pre-seizure period and in the interictal period (temporally far removed from seizures and the same behavioural state). A 20 min period is required to allow sufficient statistical power, i.e. to obtain enough samples of high-frequency energy to test the hypothesis. If the time period is significantly shortened under 20 min, in an effort to obtain greater temporal resolution, the statistical power needed to refute the null hypothesis is lost. A pre-seizure period, significantly different from the interictal period, was identified when the t test demonstrated that the null hypothesis could be rejected with P < 0.05 confidence.
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
After the entire video and intracranial EEG of each patient was reviewed, six (26%) patients from the 23 consecutive patients studied had either exclusively neocortical, or neocortical and mesial temporal lobe seizures. The remaining 17 patients had exclusively mesial temporal lobe onset seizures. Of the six patients with neocortical onset seizures, three had seizures strictly from the dorsal lateral frontal convexity, and the remaining three had mesial temporal lobe and either neocortical temporal or frontal lobe onset seizures (Table 1). A total of 34 neocortical onset seizures met the inclusion criteria for analysis. In all the seizures analysed, a single (contiguous electrodes) neocortical seizure onset zone was identified, except for patient 7 who had independent frontal and temporal neocortical seizures.
|
Seizures and seizure onset zone
As described above, seizure onset times, seizure onset zone location and frequency of the dominant seizure onset oscillation were determined (Table 1). Patient 7 had an independent frontal lobe neocortical focus and temporal lobe neocortical focus [denoted in Table 1 as Pt 7(1) and Pt 7(2), respectively]. The remaining five patients had a single seizure onset zone focus from which all neocortical seizures were initiated. In each patient, the seizure onset zone spanned between three and five contiguous electrodes, except patient 21 who had a large seizure onset zone involving seven contacts. All 34 neocortical seizures had prominent high-frequency spectral components (60–100 Hz) at seizure onset, which were the dominant spectral peaks in frontal lobe onset foci. In neocortical temporal lobe onset seizures, there were prominent high-frequency spectral components present, but the dominant onset oscillation was at a lower gamma frequency (Table 1).
Interictal discharges and behavioural state
HFEOs were found to be localized to only the seizure onset zone in all the patients studied. The frequency characteristics of interictal HFEOs are shown in Table 1. In all patients, the pathological HFEOs were stereotypic and clearly distinguishable from normal background activity (e.g. Fig. 3). HFEOs were brief in duration (average duration = 75 ± 33 ms, n = 100) with a low relative amplitude (average amplitude = 65 ± 40 µV, n = 100) compared with oscillations in the 1–30 Hz frequency band. The HFEOs were intermittently present throughout the interictal record, and often occurred asynchronously at different electrodes within the seizure onset zone (Fig. 3). In addition to these prominent interictal HFEOs, there was also an increase in the broadband (60–100 Hz) high-frequency activity in the seizure onset zone compared with control/normal/non-epileptogenic brain, which was apparent even during records that did not contain clear discrete HFEOs. The increase in broadband activity was quantified by comparing the average energy of filtered (60–100 Hz) EEG in the seizure onset zone and control/normal/non-epileptogenic brain regions. In Fig. 4, the average high-frequency energies from epileptogenic (seizure onset zone) and non-epileptogenic/normal/control brain are shown. The energy in epileptogenic brain was greater than in the normal control region in all patients, and significant at the level P < 0.05 in all seizure onset zone foci, except for the temporal neocortical focus of patient 7(2) [the F statistics are Fpt 2(3,150) = 6.48, Fpt 3(3,150) = 6.87, Fpt 7(1)(5,150) = 4.31, Fpt 7(2)(5,150) = 2.25, Fpt 9(3,150) = 3.86, Fpt 20(3,150) = 6.412, Fpt 21(3,150) = 3.89].
|
Within the seizure onset zone of all six patients there was increased high-frequency energy and HFEO activity during slow-wave sleep compared with wakefulness, which reached statistical significance in three of six patients [Fpt 2(3,150) = 2.77, Fpt 3(3,150) = 3.51, Fpt 7(1)(5,150) = 0.56, Fpt 7(2)(5,150) = 0.15, Fpt 9(3,150) = 3.31, Fpt 20(3,150) = 0.08, Fpt 21(3,150) = 14.3]. Outside the seizure onset zone, in normal control brain, there was no significant difference in broadband high-frequency activity between sleep and wakefulness. In addition to HFEOs, there were epileptiform spikes and sharp waves identified in all patients. In all the patients studied, HFEOs and the region of increased (sleep and wake) broadband high-frequency activity were localized to the seizure onset zone, whereas epileptiform spikes and sharp waves had a much wider distribution involving regions outside the seizure onset zone. As this analysis was performed retrospectively, after patients underwent resective surgery based upon the results of standard intracranial video-EEG monitoring, the data required to determine if resection of the interictal HFEO zone is associated with excellent surgical outcome is limited in our study. Unfortunately, only in patients 20 and 21 was detailed information available regarding which electrode sites were actually resected. In the remaining four patients, the exact extent of surgical resection relative to the positions of the intracranial electrodes was not clear, and correlation of surgery outcome with degree of resection of the epileptogenic brain defined by interictal HFEOs is not possible. In the two patients where the information was available, we found that resection of electrodes exhibiting HFEOs was predictive of excellent post-surgical outcome (patient 20 is seizure free 1 year after surgery that removed the epileptogenic zone defined by ictal and interictal HFEOs. Patient 21 continues with frequent seizures after surgery that did not remove the seizure onset zone determined by ictal and interictal HFEOs).
Pre-seizure high-frequency activity
To investigate whether seizures are preceded by precursors of increased high-frequency activity and HFEOs, we examined the discrete average high-frequency energy time series in the period prior to seizure onset within the seizure onset zone. The time series consists of sequential high-frequency energies obtained as defined in equation (1). The average high-frequency energy time series provides multiple measurements of high-frequency neuronal activity. To compare the amount of high-frequency activity in the pre-seizure period (just prior to seizure onset) with times distant from seizure activity, we use the Student t test. The statistics of multiple measures of high-frequency activity in the seizure onset zone during the pre-seizure period (defined for this study as 20 min prior to the onset of a seizure on the EEG) were compared with randomly selected 15 min records (with the same behavioural state) that were temporally far removed from the seizure onset time (>2 h from any seizures). An identifiable pre-seizure period was considered to be present if the average high-frequency energy in the pre-seizure period was significantly different (P < 0.05) from the average high-frequency energy in the randomly selected samples.
In 21 (62%) of the 34 neocortical seizures analysed, there was a statistically significant increase in the broadband high-frequency activity within the seizure onset zone in the 20 min period prior to the seizure onset (Fig. 5 and Table 1). The increase in average high-frequency activity prior to seizure onset appears to be due to both broadband asynchronous activity and HFEOs. This result supports the concept that high-frequency activity and HFEOs within the seizure onset zone may serve as seizure precursors, useful for identifying periods of increased probability of clinical seizures.
|
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
In all six (100%) patients with neocortical seizures, the seizure onset on intracranial EEG demonstrated significant high-frequency activity and HFEOs, as determined by spectral analysis (Table 1). Thus, unlike seizure onset activity commonly reported for mesial temporal lobe onset seizures (Schiller et al., 1998
), this finding in an unselected group of patients (non-lesional and lesional MRI) supports the notion that high-frequency activity is common at the onset of focal neocortical seizures.
In all six (100%) patients with neocortical seizures, the seizure onset zone shows a significant increase in interictal high-frequency activity, both at baseline and during the pre-seizure period. This increase in high-frequency energy, unique to the seizure onset zone, comes from brief duration HFEOs, which have the same spectral characteristics as the seizure onset oscillations, as well as an increase in overall broadband high-frequency activity (Fig. 4). Even in the absence of distinct HFEOs during the interictal record, there is more broadband high-frequency energy within the seizure onset zone compared with control/normal/non-epileptogenic regions. Since the macroscopic field potentials recorded from subdural electrodes (Fig. 3) represent summated synaptic potentials of many neurons (Traub, 1999; Traub et al., 2001), the increased broadband energy in the seizure onset zone probably represents an overall increase in asynchronous high-frequency neuronal activity, and the intermittent HFEOs are likely to be a network manifestation of increased neuronal synchrony. This is consistent with single neuron multiunit recordings obtained from human entorhinal cortex, which show an increase in single neuron bursting and neuronal synchrony in epileptic foci (Bragin et al., 2002).
Importantly, we did not find HFEOs outside the seizure onset zone, supporting that these oscillations are pathological epileptic oscillations unique to epileptic brain. The frequencies of interictal HFEOs (Table 1) are well above the Berger EEG frequency bands (0.1–30 Hz), and are of such brief duration and low amplitude that utilizing common clinical EEG viewing parameters (e.g. 10 s page and 0.5 mV/mm), these oscillations are obscured by lower frequency range activity. In fact, even when occurring as a sustained discharge at the seizure onset, the HFEO is not always clearly apparent (Fig. 2) (Allen et al., 1992; Fisher et al., 1992; Alarcon et al., 1995; Traub et al., 2001).
The increase in high-frequency energy within the seizure onset zone during sleep compared with wakefulness was found in all six patients, but only reached statistical significance in three patients. This is a robust finding, given that the probability of finding increased high-frequency energy in the seizure onset zone in all six subjects by chance would be of the order of 1.5%. The increase of HFEOs and broadband high-frequency activity within the seizure onset zone during slow-wave sleep demonstrates that neocortical networks in the seizure onset zone are strongly modulated by the behavioural state. The increase of high-frequency activity within a neocortical seizure onset zone is consistent with single and multiunit recordings in epileptic entorrhinal cortex which show an increase in high-frequency neuronal bursting and increased synchrony during slow-wave sleep (Steriade et al., 1995; Staba et al., 2002a, b). The significant increase in high-frequency activity in the seizure onset zone during slow-wave sleep may explain the well-known clinical observation that neocortical seizures frequently occur during sleep (Herman et al., 2001).
Evidence for a fundamental role for high-frequency activity in the initiation of neocortical seizures (ictogenesis) is suggested by the localized spatial distribution of HFEOs and the frequent increase in interictal high-frequency activity within the seizure onset zone prior to the onset of seizures. The increase in high-frequency activity prior to seizures was found in 21 (62%) of the 34 neocortical seizures analysed. The seizures preceded by a statistically significant increase in high-frequency activity are shown in the Table 1. It is possible that limited spatial sampling due to clinical necessity (clinical morbidity rises with increasing number of implanted electrodes) may have precluded picking up this activity during the other seizures, in which the seizure onset zone may not have been completely covered with intracranial electrodes. A limitation of the present method is the relative lack of temporal resolution for identifying significant changes in the high-frequency energy in the pre-seizure period. In order to obtain statistical significance and refute the null hypothesis, that there is no difference in mean high-frequency energy between the pre-seizure and interictal periods, interictal and pre-seizure periods of the order of 20 min are required. This precludes the possibility of actual ‘seizure prediction’, but may provide a useful measure of increased probability of seizure occurrence.
Whether increased broadband high-frequency activity in the seizure onset zone is the manifestation of a dynamic state where the threshold for seizure onset is lowered (possibly in the same way that the behavioural state modulates high-frequency activity) is an important question that cannot be answered directly by these results. Animal (Bragin et al., 1995; Maloney et al., 1997) data have demonstrated increased gamma frequency (30–60 Hz) activity during wakefulness and paradoxical sleep compared with slow-wave sleep. Similarly, magnetoencephalography (MEG) (Llinas and Ribary, 1993) and EEG (Gross and Gotman, 1999) studies in humans have demonstrated that gamma activity is highest during wakefulness and rapid eye movement sleep compared with slow-wave sleep. Gross and Gotman (1999) demonstrated that gamma activity during mental activation was increased further when performing cognitive tasks compared with quiet wakefulness. Many studies have examined gamma frequency activity at the intracellular level (Llinas et al., 1991; Steriade and Amzica, 1996; Steriade et al., 1996, 1998; Penttonen et al., 1998) and demonstrated gamma frequency activity in all behavioural states, but enhanced during brain activation (rapid eye movement sleep and wakefulness).
Fast oscillations (>100 Hz), called ripple oscillations, have been described in rat hippocampus and peririhinal cortex, where they are associated with the bursts of sharp potentials during anaesthesia, behavioural immobility and sleep (Buzsaki et al., 1992; Chrobak and Buzsaki, 1996). Ripple oscillations were also described in cat cortex during anaesthesia and natural states of vigilance (Grenier et al., 2001) and recently have been implicated in an experimental model of seizures in cats (Grenier et al., 2003). During slow-wave sleep and anaesthesia, ripple oscillations are selectively correlated to the depolarizing phase of the slow field oscillation (1.0 Hz). During natural states of vigilance, ripple oscillations are more prominent during the depolarizing phase of slow wave sleep than during wakefulness and rapid eye movement sleep. Ripple oscillations have been described recently in human hippocampus and entorhinal cortex (Bragin et al., 1999a,b; Bragin et al., 2002), and demonstrated to be modulated similarly by the behavioural state, i.e. more prominent during slow-wave sleep compared with wakefulness or rapid eye movement sleep. Based on the intracellular recordings during different behavioural states, Steriade (Steriade and Amzica, 1996; Steriade et al., 1996, 1998) has suggested that high-frequency EEG activity (>30 Hz) reflects the degree of depolarization of the neuronal membrane.
We might speculate that the increased high frequency activity seen in the minutes prior to onset of some neocortical seizures is the result of an aberration of the same physiological mechanisms underlying memory consolidation, possibly synaptic modification, and progressive reinforcement of local epileptogenic circuits. The HFEOs are similar to ripple oscillations, in that they are increased during the relatively depolarized state of slow-wave sleep.
It is important to note that while slow-wave sleep does result in an increase of broadband high-frequency activity and HFEOs in the seizure onset zone, a simple change of behavioural state, i.e. from wakefulness to sleep, does not explain our findings prior to seizure onset. The majority of seizures analysed occurred without clear evidence of any behavioural state change by video and clinical EEG (1–20 Hz) criteria. However, more subtle state changes involving the same thalamocortical networks that modulate the behavioural state (Steriade et al., 1993; Buzsaki et al., 1992; Steriade and Contreras, 1995; Chrobak and Buzsaki, 1996; Staba et al., 2002a, b), gamma oscillations and ripple oscillations are interesting potential contributors to these findings that warrant further study.
Previous studies have examined the spatiotemporal distribution of interictal epileptiform spikes and sharp waves prior to seizure onset and found that there were no changes in this activity prior to seizures (Gotman and Marciani, 1985; Katz et al., 1991). In addition, it is well established that spike and sharp wave discharges are broadly distributed, often occurring well outside the focal seizure onset zone and even in the contralateral hemisphere (Engel, 1987; Blume et al., 2001). These findings are in sharp contrast to our results, and suggest that the neuronal network mechanisms responsible for interictal HFEOs, and spikes and sharp waves are likely to be different.
We have been careful to eliminate high-frequency artefact from our records, and it is unlikely that any of the results reported here are due to measurement artefacts. All calculations were performed using a bipolar montage, which should help eliminate referential and other global artefacts. We find that the HFEOs are unique to the seizure onset zone in all patients, and that the high-frequency activity is increased during sleep, both of which argue against muscle or environmental artefacts. During the immediate post-ictal period, when the background EEG activity is severely suppressed, we did not observe HFEO or high-frequency activity in the seizure onset zone, which also supports a physiological basis for HFEOs.
Conclusions
HFEOs demonstrate a high degree of spatial localization to the seizure onset zone, contrary to epileptic spikes and sharp waves, indicating that HFEOs could prove clinically useful for more accurate localization of the seizure onset zone. Unfortunately, only in two patients was there adequate information available about the relationship of electrode sites and the brain region resected at surgery.
Broadband high-frequency activity and HFEOs in the seizure onset zone are strongly modulated by behavioural state, and are most prominent in slow-wave sleep, which provides insight into how sleep states differentially modulate abnormal epileptic neuronal networks in human neocortical epilepsy. The most important finding is that focal neocortical seizures frequently occur during periods of increased high-frequency activity in the seizure onset zone, supporting the idea that broadband high-frequency activity and HFEOs are involved in neocortical ictogenesis.
These findings indicate that increased broadband high-frequency activity and HFEOs are signatures of epileptic brain, and that a more complete understanding of their role in ictogenesis may lead to a new understanding of mechanisms underlying epileptogenesis in the neocortex and novel treatments for these frequently medicine-resistant patients (Litt et al., 2001; Litt and Lehnertz, 2002).
We have initiated a prospective clinical study to investigate further the clinical utility of using HFEOs to localize the seizure onset zone for epilepsy surgery, and to evaluate the possibility of real-time identification of periods of increased seizure probability in neocortical epilepsy.
|
Acknowledgements |
|---|
We wish to thank Marc Dichter and Leif Finkel for review of this manuscript. This work was supported by a Mayo Foundation Scholar Grant (G.A.W.), a National Science Foundation Graduate Research Fellowship and NIH grant T32-GM07517 (S.D.C.), and The Whitaker Foundation, The American Epilepsy Society, The Dana Foundation and National Institutes of Health Grants RO1NS041811-01 and MH-62298RO1 (B.L.).
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Alarcon G, Binnie CD, Elwes RD, Polkey CE. Power spectrum and intracranial EEG patterns at seizure onset in partial epilepsy. Electroencephalogr Clin Neurophysiol 1995; 94: 326–37.[CrossRef][Web of Science][Medline]
Allen PJ, Fish DR, Smith SJ. Very high-frequency rhythmic activity during SEEG suppression in frontal lobe epilepsy. Electroencephalogr Clin Neurophysiol 1992; 82: 155–9.[CrossRef][Web of Science][Medline]
Berger H. Über das Elektrenkephalogramm des Menschen. I. Mitteilung. Arch Psychiat Nervkrankh 1929; 87: 527–70.[CrossRef][Web of Science]
Blume WT, Holloway GM, Wiebe, S. Temporal epileptogenesis: localizing value of scalp and subdural interictal and ictal EEG data. Epilepsia 2001; 42: 508–14.[CrossRef][Web of Science][Medline]
Bragin A, Jando G, Nadasdy Z, Hetke J, Wise K, Buzsaki G. Gamma (40–100 Hz) oscillations in the hippocampus of the behaving rat. J Neurosci 1995; 15: 47–60.[Abstract]
Bragin A, Engel J Jr, Wilson CL, Fried I, Buzsaki G. High-frequency oscillations in human brain. Hippocampus 1999a; 9: 137–42.[CrossRef][Web of Science][Medline]
Bragin A, Engel J Jr, Wilson CL, Fried I, Mathern GW. Hippocampal and entorhinal cortex high-frequency oscillations (100–500 Hz) in human epileptic brain and in kainic acid-treated rats with chronic seizures. Epilepsia 1999b; 40: 127–37.[CrossRef][Web of Science][Medline]
Bragin A, Mody I, Wilson CL, and Engel J Jr. Local generation of fast ripples in epileptic brain. J Neurosci 2002; 22: 2012–21.
Buzsaki G. Two-stage model of memory trace formation: a role for ‘noisy’ brain states. Neuroscience 1989; 31: 551–70.[CrossRef][Web of Science][Medline]
Buzsaki G. The hippocampo-neocortical dialogue. Cereb Cortex 1996; 6: 81–92.
Buzsaki G, Horvath Z, Urioste R, Hetke J, Wise K. High-frequency network oscillation in hippocampus. Science 1992; 256: 1025–7.
Chrobak J, Buzsaki G. High-frequency oscillations in the output networks of the hippocampal–entorhinal axis of the freely behaving rat. J Neurosci 1996; 16: 3056–66.
Curio G. Ain’t no rhythm fast enough: EEG bands beyond beta. J Clin Neurophysiol 2000a; 17: 339–40.[CrossRef][Web of Science][Medline]
Curio G. Linking 600-Hz ‘spike like’ EEG/MEG wavelets (‘sigma-bursts’) to cellular substrates: concepts and caveats. J Clin Neurophysiol 2000b; 17: 377–96.[CrossRef][Web of Science][Medline]
Engel J Jr. Surgical treatment of the epilepsies. New York: Raven Press; 1987.
Fisher RS, Webber WR, Lesser RP, Arroyo S, Uematsu S. High- frequency EEG activity at the start of seizures. J Clin Neurophysiol 1992; 9: 441–8.[Web of Science][Medline]
Gasser T, Bacher P, Mocks J. Transformations towards the normal distribution of broadband spectral parameters of the EEG. Electroencephalogr Clin Neurophysiol 1982; 53: 119–24.[CrossRef][Web of Science][Medline]
Gotman J, Marciani MG. Electroencephalographic spiking activity, drug levels, and seizure occurrence in epileptic patients. Ann Neurol 1985; 17: 597–603.[CrossRef][Web of Science][Medline]
Grenier F, Timofeev I, Steriade M. Focal synchronization of ripples (80–200 Hz) in neocortex and their neuronal correlates. J Neurophysiol 2001; 86: 1884–98.
Grenier F, Timofeev I, Steriade M. Neocortical very fast oscillations (ripples, 80–120 Hz) during seizures: intracellular correlates. J Neurophysiol; 2003; 89: 841–52.
Gross DW, Gotman, J. Correlation of high frequency oscillations with the sleep–wake cycle and cognitive activity in humans. Neuroscience 1999; 94: 1005–18.
Herman ST, Walczak TS, Bazil CW. Distribution of partial seizures during the sleep–wake cycle: differences by seizure onset site. Neurology 2001; 56: 1453–9.
Katz A, Marks DA, McCarthy G, Spencer SS. Does interictal spiking change prior to seizures? Electroencephalogr Clin Neurophysiol 1991; 79: 153–6.[CrossRef][Web of Science][Medline]
Lisman JE, Idiart MA. Storage of 7 ± 2 short-term memories in oscillatory subcycles. Science 1995; 267: 1512–5.
Litt B, Lehnertz K. Seizure prediction. Curr Opin Neurol 2002; 15: 173–7.[CrossRef][Web of Science][Medline]
Litt B, Esteller R, Echauz J, D’Alessandro M, Shor R, Henry T, et al. Epileptic seizures may begin hours in advance of clinical onset: a report of five patients. Neuron 2001; 30, 51–64.
Llinas RR. The intrinsic electrophysiological properties of mammalian neurons: insights into central nervous system function. Science 1988; 242: 1654–64.
Llinas R, Ribary U. Coherent 40-Hz oscillation characterizes dream states in humans. Proc Natl Acad Sci USA 1993; 90: 2078–81.
Llinas RR, Grace AA, Yarom Y. In vitro neurons in mammalian cortical layer 4 exhibit intrinsic oscillatory activity in the 10- to 50-Hz frequency range. Proc Natl Acad Sci USA 1991; 88: 897–901.
Maloney KJ, Cape EG, Gotman J, Jones BE High-frequency gamma electroencephalogram activity in association with sleep–wake states and spontaneous behaviors in the rat. Neuroscience 1997; 76: 541–55.[CrossRef][Web of Science][Medline]
Nunez P. L. Electric fields of the brain: the neurophysics of EEG. New York: Oxford University Press; 1981.
Penttonen M, Kamondi A, Acsady L, Buzsaki G. Gamma frequency oscillation in the hippocampus of the rat: intracellular analysis in vivo. Eur J Neurosci 1998; 10: 718–28.[CrossRef][Web of Science][Medline]
Quesney LF. Intracranial EEG investigation in neocortical epilepsy. Adv Neurol 2000; 84: 253–74.[Medline]
Quesney LF, Constain M, Rasmussen T, Olivier A, Palmini A. Presurgical EEG investigation in frontal lobe epilepsy. Epilepsy Res Suppl 1992; 5: 55–69.[CrossRef]
Rechtschaffen A, Kales A. A manual for standardized terminology, techniques and scoring system for sleep stages of human subjects. Bethesda (MD): US Dept. of Health, Education and Welfare; 1968.
Schiller Y, Cascino GD, Busacker NE, Sharbrough FW. Characterization and comparison of local onset and remote propagated electrographic seizures recorded with intracranial electrodes. Epilepsia 1998; 39: 380–8.[CrossRef][Web of Science][Medline]
Spencer SS, Lee SA. Invasive EEG in neocortical epilepsy: seizure onset. Adv Neurol 2000; 84: 275–85.[Medline]
Staba RJ, Wilson CL, Bragin A, Fried I, Engel J Jr. Quantitative analysis of high-frequency oscillations (80–500 Hz) recorded in human epileptic hippocampus and entorhinal cortex. J Neurophysiol 2002a; 88: 1743–52.
Staba RJ, Wilson CL, Bragin A, Fried I, Engel J Jr. Sleep states differentiate single neuron activity recorded from human epileptic hippocampus, entorhinal cortex, and subiculum. J Neurosci 2002b; 22: 5694–704.
Steriade M, Amzica F. Intracortical and corticothalamic coherency of fast spontaneous oscillations. Proc Natl Acad Sci USA 1996; 93: 2533–8.
Steriade M, Contreras D. Relations between cortical and thalamic cellular events during transition from sleep patterns to paroxysmal activity. J Neurosci 1995; 15: 623–42.[Abstract]
Steriade M, McCormick DA, Sejnowski TJ. Thalamocortical oscillations in the sleeping and aroused brain. Science 1993; 262: 679–85.
Steriade M, Amzica F, Contreras D. Synchronization of fast (30–40 Hz) spontaneous cortical rhythms during brain activation. J Neurosci 1996; 16: 392–417.
Steriade M, Timofeev I, Durmuller N, Grenier F. Dynamic properties of corticothalamic neurons and local cortical interneurons generating fast rhythmic (30–40 Hz) spike bursts. J Neurophysiol 1998; 79: 483–90.
Steriade M., Timofeev I, Greiner C. Natural waking and sleep states: a view from inside neocortical neurons. J Neurosci 2001; 85: 1969–85.
Traub RD, Jefferys JGR, Whittington MA. Fast oscillations in cortical circuits. Cambridge (MA): MIT Press; 1999.
Traub RD, Whittington MA, Buhl EH, LeBeau FE, Bibbig A, Boyd S, et al. A possible role for gap junctions in generation of very fast EEG oscillations preceding the onset of, and perhaps initiating, seizures. Epilepsia 2001; 42: 153–70.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Aubert, F. Wendling, J. Regis, A. McGonigal, D. Figarella-Branger, J.-C. Peragut, N. Girard, P. Chauvel, and F. Bartolomei Local and remote epileptogenicity in focal cortical dysplasias and neurodevelopmental tumours Brain, November 1, 2009; 132(11): 3072 - 3086. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. C. Stacey, M. T. Lazarewicz, and B. Litt Synaptic Noise and Physiological Coupling Generate High-Frequency Oscillations in a Hippocampal Computational Model J Neurophysiol, October 1, 2009; 102(4): 2342 - 2357. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Lasztoczi, G. Nyitrai, L. Heja, and J. Kardos Synchronization of GABAergic Inputs to CA3 Pyramidal Cells Precedes Seizure-Like Event Onset in Juvenile Rat Hippocampal Slices J Neurophysiol, October 1, 2009; 102(4): 2538 - 2553. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Dichter Emerging Concepts in the Pathogenesis of Epilepsy and Epileptogenesis Arch Neurol, April 1, 2009; 66(4): 443 - 447. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. H. Higgs and W. J. Spain Conditional Bursting Enhances Resonant Firing in Neocortical Layer 2-3 Pyramidal Neurons J. Neurosci., February 4, 2009; 29(5): 1285 - 1299. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Bartolomei, P. Chauvel, and F. Wendling Epileptogenicity of brain structures in human temporal lobe epilepsy: a quantified study from intracerebral EEG Brain, July 1, 2008; 131(7): 1818 - 1830. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. A. Worrell, A. B. Gardner, S. M. Stead, S. Hu, S. Goerss, G. J. Cascino, F. B. Meyer, R. Marsh, and B. Litt High-frequency oscillations in human temporal lobe: simultaneous microwire and clinical macroelectrode recordings Brain, April 1, 2008; 131(4): 928 - 937. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Clemens, M. Molle, L. Eross, P. Barsi, P. Halasz, and J. Born Temporal coupling of parahippocampal ripples, sleep spindles and slow oscillations in humans Brain, November 1, 2007; 130(11): 2868 - 2878. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Mormann, R. G. Andrzejak, C. E. Elger, and K. Lehnertz Seizure prediction: the long and winding road Brain, February 1, 2007; 130(2): 314 - 333. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. D. Jirsch, E. Urrestarazu, P. LeVan, A. Olivier, F. Dubeau, and J. Gotman High-frequency oscillations during human focal seizures Brain, June 1, 2006; 129(6): 1593 - 1608. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||