9th Workshop on the Neurobiology of Epilepsy (WONOEP IX): The transition from the interictal to the ictal state (Teluk Nibong, Langkawi Island, Malaysia, July 4–7, 2007)


The Workshop on the Neurobiology of Epilepsy (WONOEP) is a satellite meeting of the International Epilepsy Congress organized since 1989 by the Commission on Neurobiology of the International League Against Epilepsy (ILAE), devoted to the discussion of the basic neuroscience of epilepsies in an informal setting (see reports of previous WONOEPs in the References). The theme for the 9th edition of WONOEP (Langkawi Island, Malaysia, July 4–7, 2007) was “The transition from the interictal to the ictal state.” The identification of the mechanisms that control the transition into the ictal state is crucial for the understanding of seizure genesis (ictogenesis) and, ultimately, for developing new strategies to cure epilepsy. Several recent technological advances in the diagnosis of human epilepsies have provided an opportunity to improve our understanding of ictogenesis. Intracranial EEG recordings performed in patients with drug-resistant epilepsies, and selected for surgical ablation of the epileptogenic region, has allowed investigators to study human brain activity during seizures. The most important outcome from the analysis of presurgical and intraoperative intracerebral recordings is the identification of interictal and ictal EEG patterns generated in the region of seizure onset. Presurgical recordings also provide an experimental window in which to study micro- and macrocircuits in supposedly normal human brain structures, explored en passage while aiming at nearby epileptic areas. Another area of research that has promoted and reinforced the need to understand ictal and interictal patterns is the expanding field of seizure prediction and detection. Anticipation of seizure occurrence relies on the analysis of the temporal evolution of EEG signal changes that precede the ictal discharge. Developing a precise and possibly automated protocol for identification of ictal and interictal patterns in different forms of epilepsies is crucial for detecting preictal states. Finally, the recent use of functional imaging in epilepsy has provided a novel approach to the study of ictogenesis, and uncovered the role of the interactions between neuronal and vascular brain compartments in the generation of interictal and ictal discharge patterns.

Several controversial issues on ictal transition are still open for discussion, including (1) the definition of “ictal” and “interictal” states, (2) the characterization of ictal onset patterns and the correlation between a particular seizure pattern and a specific brain region, (3) the identification of peculiar ictal patterns for given experimental models/conditions. Several elements contribute to confusion in the dialogue between basic and clinical scientists working in the epilepsy field, because the terminology utilized in the clinical setting to define activity patterns may differ from that in use among basic neuroscientists. To enhance interactions between clinical and basic sciences, a larger participation of scientists with clinical expertise was promoted in WONOEP IX.

The meeting was opened by two lectures, by György Buzsáki and by Jean Gotman, on the role of fast oscillations in neuronal synchronization during physiological processes and in epileptogenesis. During a panel session, the features of focal ictal patterns in different epileptic experimental conditions and in human epilepsy were discussed, in an attempt to draw a correlation between the experimental data and the intracranial EEG findings observed in humans. Ictal patterns in mesial temporal lobe and neocortical human epilepsies were reviewed by Anatol Bragin and Stefano Francione, respectively, based on intracranial presurgical recordings performed in drug-resistant patients. Experimental ictal patterns observed in animal models of focal epilepsy in vivo and in vitro were reviewed by Ed Bertram and Marco de Curtis. A definition of the interictal and ictal patterns and the prevalent ictal onset features in different forms of focal epilepsies were discussed. The existence of a correlation between a specific region/brain alteration and a particular seizure pattern was stressed for both experimental and human findings.

Paroxysmal high-frequency activity (HFA) is one of the common interictal patterns observed in focal epilepsy. HFA on the scalp electroencephalogram (EEG) colocalizes with seizure onset in children and were proposed as predictive surrogate marker of epileptogenesis by Joyce Wu, since their presence showed a strong correlation with seizure exacerbation in sequential video-EEG recordings. Hiroshi Otsubo confirmed that in pediatric patients, interictal HFA correlates with the ictal onset zone, identified by intracranial video-EEG recordings, and can be detected during interictal spikes with event-related analysis performed with magnetoencephalography. In an in vitro experimental model of seizure, preictal strengthening in power and synchronization of HFA generated by the collective, asynchronous firing of multiple mini-aggregates of CA1 neurons (extending over ∼150 μm) was reported by John Jefferys. These observations suggest that the generation and propagation of seizures could be mediated by a progressive preictal synaptic synchronization of HFA across neuronal assemblies. Michel Le Van Quyen showed, using an intact immature septohippocampal formation in vitro, that hippocamposeptal long-range gamma aminobutyric acid (GABA)ergic interneurons are necessary to synchronize local hippocampal interneurons into coherent high-frequency firing. These long-range neurons, by providing fast synchronization among distantly functional networks, may also participate in the construction of hippocampal networks during development.

The existence of a preictal state in patients with intractable focal epilepsy was discussed by Bela Weiss, who showed dramatic changes in brain electrical activity analyzed by estimation of a self-similarity pattern, the Hurst exponent, before the onset of epileptic seizures. Paolo Federico proposed that a preictal state could be detected at the site of presumed seizure focus as well as in remote brain regions up to 20 min prior to seizure onset by analyzing BOLD signal changes during functional MRI analysis. Kevin Staley described the correlation between interictal spikes and seizures, with the view that seizures were the consequence of a relatively improbable spike ignition pathway. Optical recordings of interictal activity in slices and computer modeling demonstrated that seizures may arise as a consequence of the looping/regenerative activation of interictal spikes.

Anatol Bragin described two seizure onset patterns in patients with mesial temporal lobe epilepsy: hypersynchronous and low-voltage fast pattern. The hypersynchronous pattern could be reproduced experimentally in freely moving animals by injection of bicuculline into the hippocampal CA3 area. The low-voltage fast pattern was not observed in acute experiments, suggesting that it may result from chronic structural abnormalities in neuronal and glial networks. Igor Timofeev presented data suggesting that cortical ictogenesis and epileptogenesis are based on mechanisms independent from the proepileptic trigger. He demonstrated the presence of long-lasting hyperpolarizing potentials in both trauma and kindling models of epileptogenesis, during all states of vigilance that induce homeostatic upregulation of neuronal excitability, eventually leading to seizure generation. John Swann presented an animal model of infantile spasms induced by chronic infusion of the sodium channel blocker, TTX, in which frequent clusters of spasms correlate to high-voltage slow transients followed by an electrodecremental episode of fast activity. Between “spasms” EEG recordings showed a hypsarrhythmic pattern consisting of asynchronous high-amplitude slow waves intermixed with multifocal spikes. Regional or generalized EEG alterations may predict transition from hypsarrhythmia to the electrodecremental events correlated with spasms. Luiz Mello described the ictal patterns in the marmoset pilocarpine model of epilepsy, characterized by recurrent spontaneous epileptic seizures not associated with intense degenerative process often seen in rodents. Jeff Noebels discussed gene-specific epilepsy syndromes observed in mice with targeted deficiencies in ion channel subunit genes. Video-EEG seizure recordings from transgenic mice that exhibit elevated thalamic T-type currents revealed a stereotyped pattern of abrupt, generalized thalamocortical seizures with no evidence of interictal abnormalities. In contrast, potassium channel subunit knockout mice showed transition patterns of partial-onset seizures with subsequent generalization.

The role of GABAergic function in interictal-ictal transition was discussed by Gary Mathern, who examined tissue changes in human and rodent models of cortical dysplasia. By comparing the number of GABAergic neurons in human tissue, he showed that the expression of GAD67 mRNA and GAD protein is not reduced in pediatric cortical dysplasia cases. These results are consistent with cellular electrophysiology studies presented by Carlos Cepeda that support the concept that GABA is an important neurotransmitter in pediatric cortical dysplasia. In slices from cortical dysplasia pediatric patients, ictal discharges were observed only when the proconvulsant 4-aminopyridine (4-AP) was coperfused with the GABAB receptor antagonist, phaclofen, but not in the presence of GABAA receptor antagonists. These findings indicate a permissive role of GABAB receptors in pediatric dysplastic tissue in vitro. René Pumain presented evidence showing a reduced efficacy of glycolysis-dependent GABAAR phosphorylation and of GABAergic inhibition in postsurgical cortical tissue samples from epileptic patients, favoring seizure initiation and propagation. These findings suggest a functional link between the epileptic pathology and the regional cerebral glucose hypometabolism observed in patients with partial epilepsies. Gilles Huberfeld showed that in human postoperative epileptogenic subiculum, ictal discharges can be induced by increasing cellular excitability or by changing external pH. The transition period is characterized by the emergence of large population preictal events that are dependent on glutamatergic transmission. Once established, preictal discharges trigger ictal events through a GABAergic mechanism. Ed Dudek showed that repetitive activation of synaptic inputs in minislices of the dentate gyrus obtained from kainate-treated epileptic rats, causes depression of GABA-mediated inhibition and facilitation of intracellular EPSPs in granule cells, thus unmasking new recurrent excitatory circuits and triggering seizures. To examine neural dynamics of the seizure transition, Peter Carlen presented an in vitro model of recurrent seizures induced by perfusing the isolated mouse hippocampus with low magnesium solution. Spontaneous seizure transitions recorded in pyramidal neurons and interneurons in the CA1 and CA3 regions were preceded by an initial “substate” wherein inhibitory synaptic input dominated, followed by a phase in which inhibition was maintained and increasing, which in turn entrained “overpowering” excitation prevailed. Paul Rutecki showed that transient group I metabotropic glutamate receptor activation in the rat hippocampal slice causes a long-lasting suppression of the slow afterhyperpolarization (sAHP) activated by neuronal activity, that converts the interictal pattern into ictal activity. These results suggest that enhancement of the sAHP may be target for antiepileptic drugs. Jeffrey Loeb reported on the molecular analysis of human neocortical tissue removed from the focal seizure onset zone. Using microarrays, a core group of induced genes—referred to as the human “epileptic transcriptome”—was identified. Many of these genes and associated signaling pathways are induced in pyramidal neurons and have sharply demarcated borders that may relate to the gyral pattern of the cortex. The presentations of Albert Becker and Heinz Beck focused the contribution of increases of the potassium M-current (KCNQ2 and 3) and persistent sodium current (Nav1.6) in CA1 neurons of chronically epileptic animals to enhanced propensity for seizures. The voltage-dependence of activation of M-currents was shifted in a depolarizing direction, rendering the channel less available at subthreshold potentials. Intriguingly, these findings are similar to the functional consequences of KCNQ channel mutations that give rise to some hereditary forms of epilepsy. Feng Ru Tang showed that voltage-dependent calcium channels, Cav1.2, Cav1.3 or Cav2.1, up- or down-regulate in functionally different groups of CA1 interneurons during and after pilocarpine-induced SE. The induced expression of Cav1.3 or Cav2.1 in reactive astrocytes 7 and 60 days after SE suggests that calcium signaling in these cells may be involved in initiation, maintenance, or spread of seizure activity.

During experimental status epilepticus (SE), synaptic GABAA receptors become internalized in hippocampal granule cells, while extrasynaptic delta-subunit-containing GABAA receptors that mediate the tonic current increase. David Naylor demonstrated that a decrease in amplitude of IPSCs coupled with an increase in GABAA tonic current was also observed following exogenous extracellular GABA application, suggesting that an activity-dependent increase in extracellular GABA occurs during SE, and may be sufficient to erode synaptic inhibition and favor the transition from an interictal to an ictal state. Moreover, simulation experiments showed that high-frequency GABA release approximating “fast ripples” (100–200 ms potentials at 40–160 Hz) can temporarily degrade synaptic inhibition. Claude Wasterlain reported that 30–60 min of SE increases N-methyl-d-aspartate (NMDA)- and alpha-amino-3-hydroxy-5-methyl-4-isoxazole proprionic acid (AMPA)-receptor-mediated miniature postsynaptic currents in granule cells, due to a migration of NR1 subunits of NMDA receptors from cytoplasm to synapses. These changes in receptor trafficking enhance excitation, and may play a role in the transition from single seizures to SE. Nikolaus Sucher demonstrated that pyramidal neurons in layer V of the developing rodent somatosensory cortex express Mg2+-insensitive NMDA receptors that induce excitatory postsynaptic currents directly without the need for coincident membrane depolarization. These noncanonical NMDARs are upregulated in a mouse model of tuberous sclerosis and could promote long-lasting plasticity changes that occur during epileptogenesis.

The role of neurosteroids in determining whether or not a seizure occurs during different behavioral conditions was discussed by Michael Rogawski. Neurosteroids synthesized in the periphery and in the brain are powerful regulators of brain excitability as positive allosteric modulators of extrasynaptic GABAA receptors. Endogenous neurosteroids play a role in regulating seizure susceptibility during the menstrual cycle (catamenial epilepsy), in stressful conditions, and in the hypoandrogenic state that is common in the setting of uncontrolled temporal lobe epilepsy. Filiz Onat reported on the temporary (3 months) suppression of spike-and-wave discharges (SWD) after kainic acid injections in genetic absence epilepsy rats from Strasbourg (GAERS). These data demonstrate a transient breakdown of mechanisms underlying corticothalamocortical oscillatory circuits involved in SWD generation and suggest an interaction between the limbic circuits and corticothalamocortical networks in epileptogenesis. The generation of rhythmic SWDs in absence epilepsy may be due to an increase in cortical firing synchronization. By using a new method for calculation of extracellular field potentials, based on a single pyramidal cell compartmental model, Gilles van Luijtelaar showed that weak population synchronization of cortical pyramidal neurons generates field responses similar to a sleep spindle, while stronger synchronization clearly correlated with SWD. These simulation studies suggest that changes in the degree of firing synchronization in large population of pyramidal cells reflect changes in synaptic strength. In a mathematical model of the nonsynaptic mechanisms of epileptiform activity, Antonio-Carlos Almeida hypothesized that (1) the sodium-potassium pump activity is responsible for the duration of the interictal and ictal states; (2) the interictal-ictal transition is characterized by an increasing excitability mediated by the chloride Nernst potential “overcoming” the membrane potential (as in immature neurons).

Intrahippocampal pathways may also play a role in promoting ictogenesis in temporal lobe epilepsy. By performing simultaneous voltage-sensitive dye imaging and patch-clamp recordings in hipopcampal slices from epileptic pilocarpine-treated rats, Douglas Coulter showed that the temporoammonic pathway exhibits a large increase in effectiveness that was stronger in animals with a high seizure frequency, suggesting that this pathway mediates the generation and/or propagation of seizure activity in the hippocampus. Another structure that may be importantly involved in epileptogenesis is the subiculum. By immunohistochemistry and in situ hybridization in the kainic acid model, Guenther Sperk demonstrated degeneration of parvalbumin-containing GABAergic neurons and afferent (calretinin-positive) thalamic neurons, coupled to sprouting of somatostatin/NPY/GABAergic interneurons (OML cells) and over-expression of NPY in pyramidal cells of the subiculum, 1–3 months after the initial SE. Aristea Galanopoulou showed that the switch of GABAA receptor signaling from depolarizing to hyperpolarizing in CA1 pyramidal neurons of the rat hippocampus appears earlier in female than in male animals. Furthermore, early life kainic acid-induced SE, show sex-specific effects on GABAA receptor signaling, which may have potential repercussions for subsequent brain development. Michele Simonato demonstrated that overexpression of FGF-2 in transgenic mice exerts two different, apparently contrasting effects: it favors ictogenesis and prevents seizure-induced cell death. The mechanisms underlying these contrasting effects are critical to control seizures and their deleterious consequences through modulation of the FGF-2 system. The endocannabinoid (2AG)-synthesizing enzyme, colocalizes with mGluR5 in the perisynaptic annulus on dendritic spines, and is supposed to induce retrograde inhibition of glutamate release via presynaptic cannabiniod receptors CB1. Tamas Freund demonstrated that CB1 receptors are selectively reduced in the epileptic human hippocampus, even when cell death is minimal. This reduction may result in a loss of negative feedback on glutamate release, and thereby could contribute to ictogenesis in the epileptic temporal lobe. Yoshiya Murashima showed that in the brain of EL mice, proinflammatory cytokines are increased progressively with both development and seizure activity. A periodic increase of cytokines prior to the next seizure episode may work together with neuronal factors during epileptogenesis and in the transition from interictal to ictal state. Luisa Rocha showed that 24 h after electrical amygdala kindling in rats, the after-discharge threshold and susceptibility to subsequent seizures were reduced by administration of nociceptin, a peptide activated by seizure activity, while seizure parameters at threshold were not modified. In fully kindled rats, nociceptin enhanced after-discharge threshold and decreased susceptibility to subsequent seizures. The data suggest that nociceptin induces inhibitory effects on after-discharge threshold and enhances the postictal period.

In several animal models of temporal lobe epilepsy, the chronic phase is preceded by a seizure-free latent period. Interestingly, interictal-like activity appears soon after the initial insult, suggesting that it could play a role in building an epileptogenic network and could be causally linked to cognitive deficits. Christophe Bernard reported that epileptic animals begin to display spatial memory deficits during the latent period. These deficits were not correlated to interictal-like activity, but to modifications in the endogenous theta rhythm. When present, interictal-like activity disrupted theta activity. Pavel Mares discussed the ontogenetic development of interictal-ictal transition in cortical foci elicited by epidural application of bicuculline in freely moving 7-, 12-, 18- and 25-day-old rats. All the age groups studied exhibited transition of interictal activity of unilateral foci into ictal phases. Surprisingly, ictal activity that started focally and irregularly spread over the cortex of the two hemispheres was recorded in 7-day-old rat pups. Raman Sankar described the application of rapid kindling as a method for studying the antiepileptogenic potential of candidate drugs, such as topiramate, in rat pups at different developmental stages. This method has also been effectively utilized to develop an animal model of epilepsy-associated depression, an important comorbidity of epilepsy. Increases in ATP turnover with elevation of adenosine may be very important in causing postictal depression of brain activity. Maria-Jose Fernandes demonstrated an increase in the expression of P2 × 7 receptors in glial cells during acute SE, while in the interictal period there was a reduction in the concentration of all anticonvulsant purines that may contribute to the triggering of spontaneous seizures. ATP may reduce glutamate release by activating P2 × 7 receptors, and form adenosine through its breakdown by ecto-5-nucleotidase. These two mechanisms inhibit hyperexcitability and spontaneous seizures occurrence.


The workshop was sponsored by ILAE and by the generous support of UCB Pharma International, UCB Pharma Japan, Janssen Pharma Japan, Valeant Pharmaceutical International, Ortho-McNeill Neurologics, Teikoku Pharmaceutical Co. Ltd, Cyberonics and Kyouwahakkou Pharmaceutical Co. Ltd. We are grateful to Pascale Quilichini for the administrative organization of the meeting.